Epistemology in Classical India

Epistemology in Classical India

The Knowledge Sources of the Nyāya School

Stephen Phillips

NEW YORK LONDON

First published 2012
by Routledge
711 Third Avenue, New York, NY 10017
Simultaneously published in the UK
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint of the Taylor & Francis Group,
an informa business
© 2012 Taylor & Francis
The right of Stephen Phillips to be identified as author of this work has
been asserted by him in accordance with sections 77 and 78 of the Copy-
right, Designs and Patents Act 1988.
Typeset in Sabon by IBT Global.
Printed and bound in the United States of America on acid-free paper by
IBT Global.
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in
any information storage or retrieval system, without permission in writing
from the publishers.
Trademark Notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and explanation
without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Phillips, Stephen H., 1950–
Epistemology in classical India : the knowledge sources of the Nyaya
school / Stephen Phillips.
p. cm.
Includes bibliographical references (p. ) and index.
1. Knowledge, Theory of (Hinduism) 2. Nyaya. I. Title.
B132.K6P49 2011
121.0954—dc23
2011023967

ISBN13: 978-0-415-89554-5 (hbk)

ISBN13: 978-0-203-15238-6 (ebk)
For Colin and Prashant
Contents

Acknowledgments ix

1 Historical and Conceptual Introduction 1

Nyāya within Classical Indian Philosophy 2
Knowledge: Truth, Belief, and Justification 4
Internalism and Externalism 13

2 Certification 17
The Justification Regress 17
Fallible Foundations 20
Epistemic Excellences and Defects 23
The Generality Problem 24
Belief-Warranting tarka, ‘‘Suppositional Reasoning” 30

3 Perception 33
Concept-Laden vs. Concept-Free Perception 35
Recognition 38
Perceptual Error (Pseudo-Perception) 41
The Generality Problem Revisited: Types of Sensory Connection 44
Apperception 48

4 Inference 51
Inference for Oneself and Inference for Another (Formal
Demonstration) 53
From Extrapolation to Generalization 56
The Ontology of Pervasion 60
Philosophical Proofs of Self, God, and mukti, ‘‘Liberation’’ 65
Fallacies and Debate Theory 71

5 Analogy 74
Learning What Words Mean 76
viii Contents
“Indirect Indication,’’ upalakṣaṇa 78
The Ontology of Similarity 80

6 Testimony 82
Testimony Not a Form of Inference 85
Statements and Facts 87
“Figurative Meaning,’’ lakṣaṇā 91
Speaker’s Intention 93

7 Lessons for Analytic Epistemology 96

Appendix: The Analogy Chapter of “(Wish-Fulfi lling) Jewel of

Reflection on the Truth (about Epistemology),” Tattva-cintā-maṇi 105
gaṅgeśopādhyāya-viracitas tattva-cintā-manau upamāna-khaṇḍaḥ 107

Notes 133
Sanskrit Glossary 159
Text and Translation 173
Bibiography 179
Index 187
Acknowledgments

This book began in the epistemological reconstructions of a previous gen-

eration of Nyāya philosopher/scholar, Bimal Matilal, Jiten Mohanty, Karl
Potter, and Sibajian Bhattacharyya in particular, along with the “radically
empiricist” and “tracking the truth” approaches to epistemology of my
teachers Roderick Firth and Robert Nozick. More fundamentally, I owe
my understanding of Nyāya to my teacher and colleague, N. S. Ramanuja
Tatacharya, a traditionally trained Sanskrit-speaking scholar/philosopher
whose mastery of the pramāṇa-śāstra is matched only by his graciousness
and generosity. We worked on Nyāya over a period of years (with six dif-
ferent trips by me to India) at the Institut Français de Pondichèry in Pondi-
cherry in South India and later in Bangalore.
Let me also acknowledge Arindam Chakrabarti, with whom I discussed
Nyāya epistemology at length in connection with papers the two of us read at
a special session of the American Philosophical Association, Eastern Division,
New York, December 2000, “Justification, Mental Causation, and Defeat-
ers: An Introduction to Indian Philosophy for Analytic Philosophers.” My
paper, “What Analytic Philosophers Can Learn from Indian Reliabilism,”
and Chakrabarti’s, “Knowledge from Trusted Tellings and Its Preventers,”
were critiqued by Philip Quinn, Matthew Steup, and Steven Luper.
More immediate predecessors to the chapters here are four papers I pre-
sented in 2009 and one in 2011: to the Philosophy Department of Baylor
University, “The Nyāya Theory of Knowledge Sources: Perception, Infer-
ence, and Testimony as Factive, Prolific, and Trusted,” “The Epistemologi-
cal Question: Nyāya on the Purpose of Pramāṇa-śāstra,” at the conference
“Self, Knowledge, and Reality in Indian Traditions,” University of Cali-
fornia, Berkeley, “Nyāya’s pramāṇa (Knowledge-Generators) as Natural
Kinds,” at the conference, “Thinking Inside the Box: The Idea of a Cat-
egory in Indian Philosophy,” Center for Hindu Studies, Oxford University,
“Cannibalizing Nyāya Epistemology,” at the “International Conference
for Indian Philosophy: World Views,” Barcelona (read in my unfortunate
absence by Parimal Patil), and, most recently, “Nyāya’s Internalist The-
ory of Justification Applied to Its mukti-vāda,” at the University of New
Mexico, at “Life, Death, and Liberation: A Conference of Comparative
x Acknowledgments
Philosophy.” Of all the comments and criticisms I received at these places,
those by Kisor Chakrabarti and Jonardon Ganeri in particular I know have
influenced my exposition. And John Taber in particular helped me avoid
some mistakes. Also informing this book is an earlier paper, “Internalism
as well as Externalism in Nyāya Epistemology,” which I read at the invita-
tion of the Philosophy Department, Jadavpur University, Kolkata, January
2008, with excellent feedback from Sukharanjan Saha.
A portion of chapter four is taken from my contribution to Piotr Balcerow-
itz, ed., Logic and Belief in Indian Philosophy, Warsaw Indological Studies,
Volume 3 (Delhi: Motilal Banarsidass, 2011), “From Gaṅgeśa’s Tattva-cintā-
maṇi, the kevala-vyatireki-prakaraṇam, ‘Negative-Only Inference.’” I wish
to thank Piotr for his editing. The entry in the on-line Stanford Encyclopedia
of Philosophy, “Epistemology in Classical Indian Philosophy” <http://plato.
stanford.edu/entries/epistemology-india> although appearing on the Web
earlier than this book’s date of publication is in many places derivative.
At the University of Texas at Austin, I must thank fi rst of all the Phi-
losophy Department, whose teaching arrangements made possible my
long stays in India and most immediately whose grant recommendation,
approved and funded by the College of Liberal Arts, for the Fall semester
2010, enabled me to compose these pages. Among current and past gradu-
ate students, the book has benefited from discussions with Joel Feldman,
Alex Catlin, Neil Dalal, Ellen Stansell, Matthew Dasti, Malcolm Keating,
and Paul J. Williams. Williams read a late draft suggesting several improve-
ments. Matthew Dasti deserves special acknowledgment and his disserta-
tion which skillfully links Nyāya epistemology with Vedāntic theism. I and
all these students (five now professors elsewhere) have utilized resources of
the indological community as well as that of philosophy at the University
of Texas at Austin, and I feel fortunate to live in an academic environment
with both dimensions.
At Routledge, I wish to thank Andrew Beck, Michael Andrews, Erica
Wetter, Felisa Salvago-Keyes, Michael Watters (Integrated Book Technology
Global), and Sherri Linsenbach, a most admirable copy-editor.

Austin, Texas
May 2011
1 Historical and Conceptual Introduction

This book is written for philosophers and students of philosophy, not for
specialists in classical Indian thought. An historical stage is set in this first
chapter for a philosophic spectacle to follow. And even here the main aim is
not historical but to introduce key concepts of the theory of knowledge that
will be engaged throughout, that of the Nyāya school. Nyāya’s epistemologi-
cal concepts were pretty much shared across school or system, but my inten-
tion in this book is not to present a common classical Indian epistemology
but rather only the Nyāya version. Competing Buddhist and other theories
will be surveyed, but mainly from the Nyāya perspective. I try to avoid using
Sanskrit words, but those I do use are defined in an appended glossary, the
first part of which gives dates of texts and authors as well as the names of the
major schools together with their most important positions.
In this book I attempt to draw on as much of Nyāya history as I com-
mand to creatively reconstruct Nyāya epistemology in the philosophic terms
of English. While an individual author may or may not have had his own dis-
tinct views and arguments, every Nyāya philosopher takes himself to speak
for the school. Although there are “camps” in contemporary philosophy,
there is little that is similar to the classical Indian school. The established
view of a school is called siddhānta, as is too the portion of a text where an
author presents his own views and arguments as opposed to pūrva-pakṣa,
text devoted to prima facie views or opponents’ reasonings. My point is that
textual siddhānta is always to be taken as expressing more than the views of
an individual thinker, as expressing what the author takes to be the truth as
discerned in Nyāya in general. Although some of the later philosophers have
important differences with earlier Nyāya positions, differences they point
out, usually these are not very radical and more a matter of refinement than
revolution. Originality is downplayed, as authors strive to perfect the system
and answer objections from other schools. Thus to try to find a single coher-
ent theory, which is admittedly an abstraction from a long series of texts, is in
accord with the dominant attitude within Nyāya itself. To be sure, the school’s
later history is marked by innovative arguments and a few novel positions,
mainly in ontology. And to some extent, every major author, including those
belonging to what comes to be called Old Nyāya and of early New Nyāya
2 Nyāya Epistemology
(Gaṅgeśa and company, fourteenth century), is creative in representing the
philosophy and responding to challenges from Buddhists and others, some-
times fellow realists of distinct schools. But the Nyāya mainstream changes
hardly at all. Moreover, innovations and divergences should be understood
against that background, it seems to me. This mainstream the tradition takes
to be defined by the Nyāya-sūtra and its core commentaries (c. 200–1000)
and then in the New Nyāya period by those works plus Gaṅgeśa’s Tattva-
cintā-maṇi (c. 1325).
The outline of the Nyāya theory of knowledge should be plain by the end
of this chapter, but let me say right away that all knowledge is produced by
delineable knowledge sources according to Nyāya, and it is through theo-
rizing about how we know that the sources are in place that Nyāya philoso-
phers provide a theory of justification. It is my contention that, details aside,
Nyāya’s project can be generalized as an epistemological theory according
to which, fi rst, there are two types of knowledge, unreflective and certified,
and, second, there are signs of knowledge sources (and their imitators such
as fallacious reasoning) recognizing which we turn our unreflective knowl-
edge into knowledge certified. By knowing the sources of our beliefs, we
come to have epistemic justification.

NYĀYA WITHIN CLASSICAL INDIAN PHILOSOPHY

Sanskrit philosophical literature as defi ned by a prominence of self-con-

scious argument (a criterion ruling out very old proto-philosophical texts
such as the Upanishads) runs from about 200 bce to 1900+. Some philoso-
phy is still written today in Sanskrit by the traditionally learned. Nyāya
is one of a dozen or so prominent classical schools using Sanskrit as its
medium. Sanskrit was an intellectual lingua franca across India for more
than two thousand years, with extensive literatures in grammar, medicine,
poetry, drama, aesthetics, astrology, jurisprudence, and other fields in addi-
tion to more well-known traditions of religion and philosophy (Buddhist as
well as Hindu and Jaina). In 1835, the British Parliament declared that no
government funds would go to schools using Sanskrit, and Sanskrit tradi-
tions and literatures other than those such as Vedānta sustained by popular
religion have given way to modern cultures and the media of the regional
languages and English.
In Sanskrit, the word ‘nyāya’ is a proper name, the name of the school,
but the word can also mean “logical or investigative procedure.” Sometimes
translators say the school of “Logic.” In the system itself, the word is given
a technical meaning as the method or methods, both evidential and concep-
tual, to be employed to end controversy and alleviate doubt, in a word, philo-
sophic procedure.1 Nyāya proves its appellation apt by probing the concept of
knowledge, repelling skepticism, and championing right procedures in both
debate and inquiry. Its methods were picked up by other schools, and were
 Historical and Conceptual Introduction 3
used in other areas and literatures such as jurisprudence and aesthetics. And
there was a long mutual exchange and development between Nyāya and Bud-
dhist Yogācāra in particular on topics of logic and informal reasoning.
Nyāya emerges as a school of philosophy, a worldview to be sustained
by generations of contributors, together with six or seven competing phi-
losophies, around the second century ce (some of the classical schools are
not so old). It has a root text, the Nyāya-sūtra, which is attributed to Gau-
tama, a legendary but entirely human figure about whom we know nothing
much in particular. Scholars have claimed to fi nd a manual of debate and
informal logic as a subset of the whole.2 Since the Nyāya-sūtra includes
much more than logic, probably the text should be regarded as having been
gradually filled out in its earliest years, with the logical portions the oldest.
But for our purposes there is no point in dividing it and assuming other
authors than Gautama, the “sūtra-kāra.”
Alongside Nyāya’s celebrated epistemology is a complex ontology (con-
cerning what is real and interrelations among realities), ethics, philosophy
of language (to be surveyed by us under the epistemological category of
testimony), and an extensive arsenal of arguments, both constructive and
destructive, aimed at opponent positions, on metaphysical topics such as
personal identity, the reality of universals, the relationship of properties and
property-bearers, and so on, almost all of which were debated across school.
Historically, Nyāya comes to be, for example, the target of Buddhist polem-
ics (very highly refi ned Buddhist polemics, I might add, in Dharmakīrti,
Śāntarakṣita, Ratnakīrti, and other Buddhist Sanskrit authors). However,
Mīmāṃsā, “Exegesis,” a school bent on defending Vedic revelation with its
own subtle positions in the epistemology of testimony, assumes the role of
the main adversary, and not the Buddhist, in most later Nyāya treatises.
Absolutely central to the Buddhist debate is Nyāya’s commitment to
realism, that is, to the metaphysical thesis that things are what they are
independently of our knowings and perceivings, which are themselves con-
ceived in objectivist terms. Nyāya embraces an empiricism that connects
with its realism: without perception, which is our principal link with the
world, none of the other sources could operate. In all periods—and no mat-
ter who the opponent targeted—Nyāya philosophers try to make plain the
connections between the operations of knowledge sources and the things
and facts known. Sometimes its principles of epistemology are defensible
independently of ontology, it seems to me, but sometimes not. Similarly,
some of the ontological theses are suitably abstract and plausible, while
others are inadequate. For example, similarity is analyzed as a property
supervenient on other properties and defined as one thing having many
properties in common with something else—a view that has merit indepen-
dently of Nyāya’s theory of four types of atom, for instance, or sound as a
quality. In any case, the worldview is progressively refi ned and expanded
by Nyāya-sūtra commentators and by philosophers writing non-commen-
tarial works over almost a score of centuries. A series of commentaries
4 Nyāya Epistemology
written from about 400 to around 1000, with four contributors (in addi-
tion to Gautama, the “sūtra-maker,” they are Vātsyāyana, Uddyotakara,
Vācaspati Miśra, and Udayana) form a few thousand pages of classical
Nyāya-sūtra (NyS) literature. During the period—called Old Nyāya—there
are also a few non-commentarial texts by Jayanta Bhaṭṭa and Bhāsarvajña
among others. These rarely diverge from views of the NyS commentaries,
but some of the outlier positions and arguments are interesting in their own
right and will be taken up by us in later chapters.
So-called New Nyāya, Navya Nyāya, emerges in the thirteenth and four-
teenth centuries and flourishes in the later part of the classical or pre-modern
age, well into the period of Muslim rule in North India. New techniques of
analysis become standardized and more sophisticated positions ironed out.
The New Naiyāyikas profited not only from study of Old Nyāya but also, in
particular, of the competing positions of Buddhist Yogācāra and Mīmāṃsā.
In later years, Navya-Nyāya texts become numerous, and the philosophic
sophistication of its advocates—who are learned, we should stress, not only
in Nyāya tradition but in controversies across school—is extraordinary by
any measure. Along with new techniques of logical and linguistic analysis,
there are new arguments and argument strategies. But there is little change
in basic outlook or core positions, that is, at least not until Raghunātha
(sixteenth century) of late Navya Nyāya, who does considerably modify,
not to say overhaul, the inherited ontological system.
Still, Nyāya does not start with but comes to develop in the late com-
mentarial literature and in a long and complex text by Udayana in the
eleventh century a rational theology rich with arguments (an omniscient
Creator is established by a dozen formal proofs several of which are
unknown in the West), as will be explained in the chapter on inference.
Nevertheless, despite Udayana’s arguments, Nyāya’s strong vein of onto-
logical reflection is developed independently of its all-told rather minimal
theology. This is because the ontology is thought out, in the earlier centu-
ries, mainly within the Vaiśeṣika school, which is atheistic, and Vaiśeṣika
arguments and commitments are carried over into the later periods without
much if any theological influence. A layered view is put forth of individual
things as particular substances that bear different types of property and
relation to other things, including causal relations. At least this much of a
core remains even with innovators such as Raghunātha of the later Navya
period and apparent mavericks such as Bhāsarvajña (tenth century). Nyāya
and Vaiśeṣika are sister schools in the early period, but from about 1000
and Udayana they become a single system.3

KNOWLEDGE: TRUTH, BELIEF, AND JUSTIFICATION

Several words in Sanskrit are used much in the same way as ‘know’ and
company in English.4 There are, however, some differences. For example,
 Historical and Conceptual Introduction 5
the word ‘jñāna’, which has about the same semantic range as ‘knowledge’
in English (philologically the two are Indo-European cognates), is better
translated “cognition” since there can be false jñāna whereas there cannot
be “false knowledge” as the words are used in (analytic) English. Most
importantly, a cognition is a mental event, not a belief. Better to under-
stand how knowledge is viewed as an occurrent mental event, as a “cogni-
tion,” jñāna, in Nyāya as throughout classical Indian philosophy, let us
take an overview.
We may say then provisionally and painting in broad strokes that there
are two levels to the Nyāya theory, pramā, raw animal knowledge, so to
say, and knowledge self-consciously certified, nirṇaya and siddhānta.5 “Ani-
mal” is not quite the right word since Nyāya views inference and testimony,
and not only perception, as generating knowledge without the reflection
required for epistemic justification. Furthermore, reflective knowledge is
not fundamentally different from the unreflective variety adults share with
children and animals who cannot formulate an argument. The one derives
from the other, and self-consciously attending to knowledge sources is one
way we achieve reflective knowledge, according to Nyāya.
Like other schools of classical epistemology, Nyāya views all knowledge
as pramāṇa-born, i.e., generated by a veritable “knowledge source” (per-
ception, inference, analogy, or testimony).6 In more technical terms, we
may say that for Nyāya occurrent knowledge (knowing and thinking right
now “I am writing for a worthy reader”) forms non-occurrent knowledge
retrievable by memory (“I was writing for a worthy reader”). Occurrent
cognitions, jñāna, form beliefs, and beliefs that are true are born of cogni-
tions, pramā, that are themselves born of knowledge sources, pramāṇa,
the citing of which constitutes justification (“I saw it,” “There is this infer-
ence,” “S told me”).
Nyāya epistemology is fi rst of all an externalism, since it is a causal
theory of knowledge according to which self-conscious justification by a
subject S of a proposition p is not required for S to know that p. Instead, S
needs to stand in the right causal relation to the truth that p. Furthermore,
S may well know that p without knowing that she knows that p—this the
so-called KK thesis, which is a defi ning mark of analytic internalism, is
rejected by Nyāya. Instead of a matter of right relation between a knower
and a fact out there in the world, the fact captured by the proposition or
belief generated, internalism insists that a knower fulfill epistemic duties,
that there is more to knowledge than passively receiving information, that
a person does not know what she has not consciously considered and tried
to justify. There is clearly internalism in Nyāya’s theory of knowledge, since
the school has a robust theory of justification along just such lines, that is
to say, reflective certification. And a thorough-going externalism struggles,
unlike Nyāya, to explain justification.
Let us go over a few technical items. A cognition’s knowledge status entails
its truth or veridicality (yathārthatva, pramātva). The seventeen-century
6 Nyāya Epistemology
textbook-writer Laugākṣi Bhāskara echoes Gaṅgeśa, the New Nyāya sys-
tematizer (fourteenth century), providing a defi nition that is pretty much
standard throughout later Nyāya:

Veridicality is the having of φ predication content about an object that

is φ, e.g., the bit of knowledge, “That’s a pot,” in reference to a pot.
And this defi nition is common to both memory and immediate expe-
rience, so we should conclude. And here the truth of the knowledge
makes it fit for producing successful action.7

If a cognition, for example, that a is F is veridical, then a is F, and if a

cognition presents an object not as it is, it is not veridical. A veridical cog-
nition, no matter what its source, hits a fact, or, more properly, an object
(viṣaya), its object, and there can be cognitions that are accidentally veridi-
cal. For instance, a fi re-cognition produced from a person’s mistaking dust
for smoke on a mountain where there is actually fi re is veridical but is not
pramāṇa-generated. Not being the product of a knowledge source, it does
not get the highest epistemic award, which is prāmāṇya, “truth generated
by a knowledge source,” which is something more than mere truth or verid-
icality, pramātva. A merely veridical cognition is a cognition of an object
as it is (yathārtha), and may or may not be prāmāṇya, which entails that
there is a true cognition that is pramāṇa-born.8 The non-genuine inferen-
tial cognition of fi re from seeing dust is veridical though accidentally. The
veridicality of cognitions is explained by the cognition-producing process,
but veridicality itself is a relation that is independent of cognitive origin.
Nyāya philosophers embrace a correspondence theory of truth or veridi-
cality, although about the equivalent term in Sanskrit, yathārthatva (“as
the object, so the cognition”), there is some unhappiness. This is due to a
sense that the term is thought to imply a kind of similarity, and there is,
Nyāya philosophers say, little similarity between, for example, a pot and a
cognition of a pot.9 Nevertheless, facts make cognitions true, we are told
again and again by Vātsyāyana and company.10
Now knowledge for Nyāya is, as mentioned, a species of mental event,
not of belief, although a certain range of mental states—thoughts, testi-
monial comprehensions, inferences, perceptions—have belief—i.e., prop-
ositional—content. Perhaps the closest concept in Indian traditions to the
“belief” of Western philosophers is “trust in cognition,” for which there
are common words.11 It is axiomatic that we trust knowledge (pramā) to
guide us in what we do. Epistemic trust is an important part of the Nyāya
picture. Nyāya focuses, however, on occurrent cognitions or awarenesses.
These are psychological events or properties that have verbal (in the case
of testimony) or verbalizable indications or content. To bridge traditions of
epistemology, we have to move back and forth between beliefs and cogni-
tions that are identified by content types and by mode of generation (that
a cognition is perceptual or inferential, etc., in putative type is said to be
 Historical and Conceptual Introduction 7
introspectable). Nyāya’s veridical cognitions are psychological events that
inform us about things, and so may be thought of as occurrent beliefs in
that a cognition’s content or indication is to be expressed in propositional
form. What we might call standing beliefs are for Nyāya preserved in mem-
ory, which works through the fi ring of saṃskāra, “mental dispositions.”
Central to the Nyāya epistemology is to view the truth of standing beliefs
to depend upon the truth of the occurrent cognitions that form them. Thus
Nyāya can be said to presuppose the two highly defensible theses; fi rst, that
all of our standing knowledge has resulted from genuine knowledge sources,
and, second, that all of our justified beliefs have resulted from knowledge
sources (objectively justified belief) or their close pretenders (subjectively
justified belief), the bogus pramāṇa that we sometimes mistake for the real
McCoy, more about which in a moment.
With regard to cognitive “content” (viṣayatā), Nyāya holds that even
the simplest verbalizable cognition is of an entity as qualified by a quali-
fier. That is to say, even a perception’s indication is said to have minimally
the structure of a qualificandum qualified by a property or qualifier, cog-
nized all at once.12 For example, the cognition, “That’s a pot,” indicates an
entity as qualified by pothood or being-a-pot. Nyāya embraces the notion
of implicit occurrent beliefs. Efforts, for example, which are another kind
of psychological property, are thought to have cognitive content or, more
precisely, an object-directedness towards something that might be picked
up, a pot, for instance.
There seems to me to be the advantage of parsimony in talk of cognitions
rather than beliefs. For cognitions are immediately introspectable proper-
ties of a person and something such seems uneliminable, whereas beliefs
seem otiose once we have cognitions and the dispositional properties they
etch in memory. Nyāya philosophers account for the non-occurrent qual-
ity of standing beliefs by a theory of memory, as indicated. Indeed, all
thinkers on the classical scene, talk about memory in terms of “mental
dispositions,” saṃskāra.13 There is an elaborate view both of dispositional
formation in memory and dispositional determination of occurrent cogni-
tion and action, as we shall see in the fi rst section of the chapter on percep-
tion. In brief, perception or another cognition-generator indicating a as an
F produces a disposition that when aroused (by environmental stimulus
or internal “awakener,” bodhaka) is itself a causal factor helping to bring
about a remembering, an effort, or another psychological event or property
also indicating a as an F.
Nyāya thinks of memory in terms of dispositional properties. But it is
not a reductionism, and is unlike Western dispositional theories of belief
that would reduce beliefs to dispositions to act. The targets of epistemic
evaluation are cognitions, which are psychological events inscribing dispo-
sitional properties in memory. All true and reliable memories are formed
by cognitions that are bits of occurrent knowledge generated by knowledge
sources. Memory is then not an independent source. Remembering operates
8 Nyāya Epistemology
depending on the effects of the real knowledge-generators, the processes
responsible for the origination of our true beliefs. So instead of evaluating
memories—which of course we could do if need be—as epistemologists we
focus on occurrent knowledge, the source of all true and reliable memory.
What is the object or range of objects of a bit of knowledge according to
Nyāya? And what underpins the intersubjectivity of cognition and belief,
such that two people S and T can have the same belief? A common Western
view is of course that propositions are the objects of belief, their content,
though it is mysterious how S and T can both access propositions as objects.
So again I think we find the merit of parsimony in Nyāya’s ontological pic-
ture. S and T each have their own cognitions, but cognitions fall into types
according to their propositional content, which S and T may well indeed both
be apprised of. The intentionality or having-an-object (viṣayatā) of cogni-
tions is a direct relation between cognitions and things in the world, with no
intermediate realm of sense data, ideal forms, or propositions. A veridical
cognition indicates something’s being some way that it is. That is its object.
Strictly speaking, cognitions do not have content. They have intentionality
which by nature is the hitting of things. The relation between a cognition and
its object is talked about in two ways, as objecthood, which is a relational
property of an object cognized with the cognizing cognition as the “succes-
sor” or second term, and having-an-object, which is a relational property of
a cognition with the object cognized as the successor term.14
What about cognitions that are non-veridical, whose indications do not
hit the facts? According to Nyāya, a non-veridical cognition also has an
intentionality that is directed towards a reality or realities. When we mis-
perceive a rope as a snake, taking a to be F when it is not F, an F-hood bit
of intentionality, which originates in previous experience of Fs—of snakes
in our example—directs one towards that property, which is thoroughly
real, among the world’s furniture, though it is not experienced where it
exists in fact, which is in snakes, not in the rope at hand. At work would
be a mental disposition (saṃskāra) which under normal conditions would
prompt a remembering but which in the deviant conditions of perceptual
error fuses a snakehood bit of intentionality into a current pseudo-percep-
tion. A non-veridical cognition is said to have misplaced intentionality. It
presents something (the rope) not as what it is (anyathā-khyāti).15 But what
it presents it as is something that exists full-bloodedly elsewhere, and as
its object under normal circumstances is intended also in the abnormal
circumstances of perceptual error. It is the real-world locus of part of the
objecthood whose second term is the cognition whose object it is.
The crucial cognitive relation considered from the side of cognition, a
cognition’s having-an-object, is said to have three “parts”: a cognition’s
qualificandum portion, its qualifier portion, and the portion of qualifica-
tive relationality, matching the object’s objecthood, i.e., its being a proper-
ty-bearer and mere object cognized (a), its having a property that is being
cognized (F), and its being related to that property by an ontologically
 Historical and Conceptual Introduction 9
qualificative relation. An illusion’s qualificandum portion, the “a” part in
“a is F,” successfully hits the object in front, the rope, as were someone to
say “That’s a snake,” pointing to a rope, being right that there is something
there, ontologically a qualificandum cognized as qualified by a qualifier,
though the qualifier is “misplaced.” (This hitting of the a would indicate
the error to be pseudo-perceptual, i.e., a result of a process that in some
way is a deviation from the operation of perception as a veritable “knowl-
edge source,” pramāṇa.) Nyāya’s theory of perceptual illusion will be spelt
out further in Chapter 3.
Typing cognitions by their intentionality allows Nyāya philosophers to
carry out their textbook logic and epistemology (often using pronouns for
variables as well as the technical terms of inference, “prover,” hetu, and so
on), but “having-an-object,” viṣayatā, is itself just a property of cognitions,
a relational property whose second term is, I repeat, the object cognized.
Again, there appears to be parsimony, no cluttering up of Nyāya ontolo-
gy—indeed simplicity is a touchstone often cited by classical philosophers,
both Naiyāyikas and their adversaries (this is a variety of tarka, “supposi-
tional reasoning”: see the section in Chapter 2). In sum, a person’s life as
an epistemic agent is comprised of a series of cognition tokens the types of
which as specified by their having-an-object Nyāya philosophers discuss in
the ways Western analytic epistemologists discuss propositions and people
having the same belief.
Cognitions are, then, moments of consciousness, not species of belief,
but we may say that cognitions form beliefs in forming dispositions and
that veridical cognitions form true beliefs. A knowledge episode—to speak
in the Nyāya way—is a cognition generated in the right fashion, with the
right origins in fact. Knowledge episodes form non-occurrent knowledge,
and so an examination of what is crucial to the arising of a knowledge epi-
sode is crucial to the evaluations of epistemology, as Nyāya sees things.
Now knowledge cannot arise by accident. A lucky guess, though true or
veridical, would not count as knowledge because it would not have been
generated in the right fashion, would not have the right pedigree or etiology.
The central notion in the Nyāya view—as of all classical Indian epistemolo-
gy—is the “knowledge source” (pramāṇa), a commonly recognized process
of veridical-cognition generation. Nyāya’s is an epistemology of knowledge
sources, the most important of which are perception, inference, and testi-
mony (the fourth source, analogy, is directed solely to word-meanings).
Briefly let me mention what looks to be to Western eyes a strange
assumption Nyāya makes (along with the everyday speaker of Sanskrit, it
seems to me, though not as self-consciously). The word ‘pramāṇa’ (“knowl-
edge source”) along with the words for the individual knowledge sources,
‘pratyakṣa’ (“perception”), and so on, are used such that the truth of the
resultant cognition is implied. This runs counter to the meaning of the
corresponding English words, along with broad philosophic supposition,
which is different with ‘perception’ and company. I think the factivity of
10 Nyāya Epistemology
perception and so on is a common presupposition in classical Indian epis-
temology, running across school, though it is perhaps motivated differently
in different traditions. Yogācāra Buddhists, for example, apparently out
of allegiance to the metaphysical view known as momentariness, which is
a presentism (only things existing right now are real), claim that there is
no difference between source and result, process of knowledge and effect,
pramāṇa and pramā. Thus there can be no wedge driven between cause
and effect such that there could possibly be knowledge by accident. The
Vedic schools (Mīmāṃsā, Vedānta, Nyāya, Vaiśeṣika, Sāṃkhya, Yoga) do
distinguish knowledge as result and knowledge-producing process but also
see the concepts as wedded in that, as indicated, no genuine knowledge
source ever produces a false belief. Only pseudo-sources do. That is to say,
no non-veridical cognition is knowledge-source generated.
A knowledge source may be defi ned as the lawful connection between
the fact that p and the belief that p. Deviant functioning of a process or
method, however truth-conducive, is not a truth-guaranteeing pramāṇa.
Being merely reliable does not fit the bill. A veritable “knowledge source”
according to Nyāya has a truth logic, like ‘knowledge’ in English. Maybe we
should say perception*, inference*, testimony* to render the Nyāya ideas.
False testimony, for example, does not count as a knowledge-generator;
the Sanskrit word for testimony is used only for what would be termed
in English “epistemically successful testimony,” i.e., with a hearer having
knowledge in virtue of a speaker telling the truth. A non-veridical percep-
tion is not really a perception at all but a “pseudo-perception,” pratyakṣa-
ābhāsa, “apparent perception,” a perception imitator or perception solely
from a fi rst-person point of view. You don’t really see an illusory snake; you
only think you see one. An apparent perception P2 may be indistinguishable
from the subject’s own perspective from a bit of genuine perceptual knowl-
edge P1, both forming a-as-F type dispositions. But Nyāya insists they are
different, taking a “disjunctivist” position. The two do not in fact have the
same intentionality, and the one results from a veritable knowledge source
(no mere method) and the other does not. According to Nyāya, nothing that
a person knows would have been formed by a cognitive process other than
a knowledge source, pramāṇa. Part of the Nyāya project is to specify the
connections between the world and cognizers that result nomologically in
cognitions that are true.
The three sources of knowledge of the world as opposed to meanings,
namely, perception, inference, and testimony, are thought of, each one, as fac-
tive, prolific, and trusted; factive, as generating cognition that is true, prolific
in operating throughout our daily lives, and trusted at the first level by people
ordinarily without thought or express justification. At the second level, the
three are trustworthy because they deliver the truth. Habits of speech are
reinforced by success in action, classical theorists across school recognize in
accepting the presumptive authority of common opinion (vyavahāra). But “a
knowledge source” probably should be thought of as a technical term, one
 Historical and Conceptual Introduction 11
that entails factivity, as we have seen, as a matter of definition. Similarly with
justification, the having of which, if veritable (or objective), as opposed to the
apparent (ābhāsa), means that the justified cognition is true.16 Nevertheless,
according to Nyāya justification can be conscious or alternatively implicit,
certified belief being distinct from beliefs belonging to first-order knowledge-
generators, perception and the rest. Certification, which will be our primary
focus in Chapter 2, is apperceptive knowledge that a target cognition as spec-
ified by its objecthood is true.
A knowledge source can be identified both by intrinsic features and in
relation to a particular result which becomes then certified—the founda-
tional way a cognition can be known to be true.17 In brief, Nyāya presents
causal paradigms whereby beliefs can be evaluated supplementing a coher-
entist picture. For non-foundational beliefs, certification consists of argu-
ment: a conclusion may be shown, formally, by inference, to follow from a
set of known premises. A special kind of modus tollens inference has been
interpreted as foundational: a cognition can be certified with respect to its
fruit, success of effort and action (“I would not have been successful had
the cognition been false”).18 However, this is an instance of tarka, “sup-
positional reasoning,” and presupposes the certificational status of having
for one’s belief, and recognizing, a genuine knowledge source. Source iden-
tification is where the true foundations lie. And once a type of cognition as
specified by its objecthood has been certified, a later cognition known to
be a token of that type would be certified, too.19 Certificational knowledge
can in some cases result from tarka, but, to repeat, to be effective tarka
presupposes the operation of a genuine knowledge source which has to be
identified (“My cognition of a river in the distance proved to be a genuine
perception and not an illusion, since I would not now be bathing, etc., had
it been non-veridical”). In other cases, identification of a source appercep-
tively serves as certification without tarka. Sometimes source identification
is not immediately apperceptive, however, but rather a matter of inference,
as we identify signs, concomitant marks, of knowledge processes such as
proper organ/object connection in the case of a putative perception or the
trustworthiness of the testifier in the case of putative testimonial knowl-
edge. 20 That is, we certify that a cognition in focus is knowledge-source-
produced by an inference that takes as its prover term (hetu) an indicator
(guṇa, an “excellence,” or doṣa, a “flaw”) that differentiates the properly
pedigreed from the close pretender.
Aside from foundational considerations, in general we come to know
that a given cognition is true by inferring that it is true by following from
other propositions established in our knowledge system (siddhānta), as
mentioned. Inferential certification gets a bit complicated, however, because
inference proper has to be distinguished from “suppositional reasoning,”
tarka, used to tip the scales of justification one way or another when there
are competing claims complete with competing etiologies, two apparently
good inferences with contradictory conclusions, for instance. We shall take
12 Nyāya Epistemology
up tarka in Chapter 2. Philosophy, by the way, which is all reflective knowl-
edge (since all philosophic proposals get challenged), proceeds generally by
inference and tarka. New Nyāya texts are little other than a series of infer-
ences and bits of suppositional reasoning. But certification also goes on,
Naiyāyikas recognize, in everyday life.
In sum, certified knowledge, nirṇaya, contrasts with mere knowledge,
pramā, by involving a kind of “conscious justification” as opposed to a
cognition’s being simply justified. A knowledge source provides a non-in-
ferential bottom level of implicit justification that does not itself have to be
known, or, in the case of inference as a knowledge source, a bottom level of
justification that is inferential but is also about the world, i.e., about some-
thing other than the veridicality of the cognition generated by inference.
On the meta-level of certification, it is that veridicality that is arrived at.
A person S knows a proposition that p just in case the belief that p has
been generated in the right way by the truth that p, and this includes knowl-
edge of the knowledge that p. Nyāya would assimilate knowledge of knowl-
edge to its overall causal picture, and to know a cognition, including its
objecthood (what it is about, its content, whether or not true), certification is
not required. 21 But to certify a cognition that p is of course something more
than mere knowledge that a cognition in focus indicates that p. Certifica-
tion invariably strengthens trust (viśvāsa), fortifying beliefs against doubt.
With respect to a question, our being able to cite the sources of our beliefs
is sometimes crucial. Given practical motivation to eliminate doubt and
fi nd out the real truth, we self-consciously try to make sure that processes
of factual connection are really in place. Nevertheless, knowledge does not
require certification. Knowledge does not require “conscious justification”
in the sense explained. It requires only the proper tie between cognition and
fact, which amounts to a cognition’s being knowledge-source-generated.
If justification is required for knowledge, then justification is simply the
fact that a cognition is source-generated whether anyone knows that fact,
according to Nyāya.
All this is perhaps clearest in the Nyāya view of testimony. The central
position is that uptake is fused with a modicum of positive epistemic warrant
such that a hearer is justified normally in believing a statement without a
background check of the trustworthiness of the testifier. There are circum-
stances where we may legitimately ask: “Does my informant know the truth?
Could she be deceiving me?” The speaker’s knowing the truth and having
no reason to deceive are identified as certification conditions (“epistemic
excellences,” guṇa) for testimony, and it is a commonplace to certify testi-
monial knowledge with reference to these. But here, too, we can have knowl-
edge without certification. Doubt-free understanding—entailing warranted
acceptance—is the cognitive default, how a hearer normally learns.22
Consider now how Nyāya looks in the light of post-Gettier defi nitions
of knowledge (Edmund Gettier having shown that the three conditions of
belief, truth, and justification are together insufficient, that there are cases
 Historical and Conceptual Introduction 13
of justified true belief that are not knowledge). 23 Thus we come to a fourth
condition proposed by John Pollock and others in response to Gettier cases,
justification-plus or objective justification. 24 Such examples were known on
the classical scene but were not given extensive analysis. 25 Consider again
the scenario of a line of dust blown up from a mountain across the way,
a mountain where there is indeed fi re though unrelated to the dust that a
subject S from a distance mistakes for smoke.26 From this apparent percep-
tion of smoke, by apparent inference, S has a cognition that prompts effort
on his part towards fi re, real fi re at precisely the point that S hits upon by
tracing the mistaken perceptual evidence.
My reading of Nyāya here is that the cognition is to be regarded as
veridical but not as knowledge certified (nirṇaya). The cognition could be
objectively defeated although S has at the time no reason to suspect that
the cognition has not been source-generated.27 To the deluded subject, the
cognition would seem to deserve certification. But just about any fact can
become concealed. The belief that there is fi re on the mountain, although
true, pramā, cannot be objectively certified (nirṇaya = prāmāṇya). The
subject may have subjective (“pseudo-” or bogus) justification, but there is
no objective knowledge (arthavat—see the quotation from Vātsyāyana at
the beginning of the next chapter). The proper functioning that constitutes
inference—as opposed to apparent inference, which is what is going on
with S—requires true perceptual premises, and in the dust/fi re apparent
inference a premise is false. Considering the fi rst-person perspective, the
subject has an excuse, there is prima facie warrant. But there is not the cor-
rect tie between cognition and fact required for knowledge.
No cognitive process involving acceptance of a false premise would be a
genuine knowledge source, since a knowledge source is a type of cognitive
process—not a one-time happening—a type of process that invariably gener-
ates true beliefs. Nyāya’s pramāṇa typology is discussed further in Chapter 2,
especially in the section, “The generality problem,” and further Gettier-style
problems not involving false premises are taken up in Chapter 7.

INTERNALISM AND EXTERNALISM

Going over the internalist/externalist debate in analytic epistemology

should help us better grasp the distinction between certified and uncerti-
fied knowledge in the Nyāya system. Internalists insist, in the tradition of
Descartes, that in order to know a proposition that p we must have and
be aware of a good reason for believing that p. From Plato is inherited the
account of knowledge as justified true belief, and justification amounts to
having a good reason that the believer can articulate. Externalists deny that
in order to know we have to have conscious access to our reasons for the
belief, arguing that knowledge is a matter of an objective or natural relation
between knower and known, in the case of perceptual knowledge a causal
14 Nyāya Epistemology
relation. Animals and children have knowledge of various things while they
cannot give reasons. Nyāya for its part lines up on the externalist side of
the controversy in admitting so-called animal knowledge (including from
inference and testimony). However, in the notion of epistemic justification,
internalism gets with it traction: despite “unreflective knowledge” where
the knower is not aware of, and does not have access to, her reasons for
believing, still there is the fundamental epistemological question of whether
we have good reasons for believing the things we do.
On such a fundamental question, Nyāya looks to be internalist, embrac-
ing in its notion of nyāya, “critical reasoning,” or “empirical and philosophic
scrutiny,” its own internalist concept of epistemic justification. Reasons
or causes of belief to which we have no cognitive access cannot count as
constituting epistemic justification. Western internalists seem rightly more
interested in the concept of epistemic justification than they are in the con-
cept of knowledge, since in their view it is in the having of good reasons for
our beliefs that we have epistemic merit. Such merit does not lie in simply
believing the truth. Even if our beliefs are false in that we are being deceived
by a powerful demon or are a brain in a vat made to experience and believe
the ways we do by evil scientists on Alpha Centura twisting dials, still we
may have epistemic merit if we have good reasons. 28
To go a little deeper, we may say that in the Nyāya externalism knowl-
edge flows at the fi rst level out of causal connections, natural processes
that generate true beliefs, perception, inference, and testimony. As we
shall see, inference and testimony (analogy, too) are dependent in their
workings (but not reducible to) perception, which is said to be the
“eldest,” jyeṣṭa, knowledge source. But when it comes to certification,
Nyāya concerns itself either with arguments or with properties that are
signs of pramāṇa operation and thus criteria for warranted beliefs of a
basic sort. This “modest” foundationalism of Nyāya amounts to source-
identification providing warrant that is not conditional. Indeed, putative
source-identification is usually enough.
Thus concerning certification, Nyāya lines up with the internalist.
Remarkably, Nyāya may be said to side in some ways with both major camps
of internalism, the old foundationalists and their coherentist adversaries.
The common commitment, also by Nyāya endorsed, is that as epistemic
agents we have a duty to believe responsibly, checking our beliefs against
the standards of logic and science (śāstra). Whether our beliefs actually hit
the facts is, so it is argued by the internalists, not the main point, but rather
our reasons for thinking they do. Therein lies epistemic merit.
Now Nyāya agrees but with the important addendum that by attend-
ing to the nature of perception, inference, and testimony, which at the
fi rst level operate with us unselfconsciously, we at the second level self-
consciously certify what we know and believe. The internalism flows out
of the externalism. And while certification is unnecessary for many bits of
knowledge, it is of course entirely necessary for theses of philosophy, for
 Historical and Conceptual Introduction 15
example, which are all controversial. Where it is not necessary is in rap-
idly acquired and rapidly lost bits of perceptual knowledge to which one is
mainly indifferent.
Controversy is the spur to philosophy, Naiyāyikas tell us, the spur to
what is named nyāya, “critical inquiry and reasoning,” directed to the
resolution of doubt and dispute. Perhaps the biggest difference between
Nyāya and Cartesian epistemology is that we do not “begin” with foun-
dational knowledge and work out to knowledge of the world. Rather, we
end doubt and controversy once it arises by employing—as best we can and
self-consciously as opposed to non-self-consciously before the challenge—
knowledge sources to ascertain the truth. These are to be supplemented,
the tradition insists from the earliest, by tarka, “suppositional reasoning,”
drawing out untoward consequences, etc., of an opponent’s view, very much
in the spirit of Socratic elenchus, as we will discuss. And at the second level,
pramāṇa are now “methods of knowledge” as opposed to simply natural
processes, as indeed is tarka, the “reasoning” that is the philosopher’s tool.
Nyāya does not recognize the internalist’s epistemological question as fun-
damental although it does in its own way recognize foundational consider-
ations (ending doubt or dispute). For example, for a subject S to point out
that p is a matter of perception is for S to show the epistemic justification S
has for p, which is then known self-consciously at the second level. 29
In sum, the Nyāya concept of epistemic justification, or, as I prefer, certi-
fication, centers on arguments whose premises if challenged can be justified
in the end by foundational considerations. But these are not the founda-
tions of the Cartesians in self-evident reasons or beliefs that are justified in
virtue of their content or even non-doxastically. No, for Nyāya, our beliefs
form a coherent system of mutual support, and anything that we know can
be a premise in an argument. Then in addition to this coherentism, there
is also a modest foundationalism: beliefs identified as the results of percep-
tion, inference, or testimony are certified, presumptively, without further
argument, unless there emerges a defeater bringing the identification into
doubt (e.g., “You are too far away to see clearly”).
Were there a (meta-) challenge to the presupposition that inference, for
example, generates beliefs that are true, the Nyāya response would be to
use tarka, “suppositional reasoning” meant to expose the falsity of an
opponent’s argument or to support one’s own. We shall look at the answer
more closely in Chapter 2. But simply put it is that we would not act as
we do. Nyāya’s project is limited to understanding the knowledge that we
trust which is cognition that has been produced in the right way. There is a
correlation between knowledge and success in action. The reason we turn
to epistemology is to be able to resolve doubt or dispute. Doubt or dispute
prevents “unhesitating action,” niṣkampa-pravṛtti. And the same defaults
granted everyday instances of perception, inference, and testimony are
granted to any investigation we might self-consciously undertake spurred
by doubt or controversy. That is to say, if we self-consciously engage in
16 Nyāya Epistemology
an inference whose conclusion is p, then we are reflectively warranted in
believing that p without certifying the certificational capacity of inference.
And this is shown in our action, which becomes unhesitating, confident—
and rightly so, since, as our philosophers remind us, knowledge helps us get
what we want and avoid what we want to avoid.
2 Certification

The fi rst word of the fi rst sūtra of the Nyāya-sūtra is ‘pramāṇa’, “knowl-
edge source.” The oldest extant commentary, by Vātsyāyana, explains the
emphasis and the importance of the concept in an opening statement:

Success of action undertaken on the basis of an object-directed cogni-

tion arising from a knowledge source (pramāṇa) shows such a source to
be objective (artha-vat, “having a (real) object”). Without a knowledge
source, there would not be knowledge of the object (artha, the conative
as well as cognitive object), and without knowledge of the object, suc-
cessful action (i.e., successful voluntary action, pravṛtti) would not be.1

This is a crucial bit of “suppositional reasoning” (tarka) used to explain the

point of a system of certification targeting the origins of our beliefs.2 We do
not act unhesitatingly if there is doubt. Epistemology, pramāṇa-śāstra, is
concerned with the knowledge possessing which a person acts confidently.
There is a response to skepticism here which will be developed throughout
the chapter.
Genuine doubt or challenge bumps us up to the second level of reflective
knowledge, that is, if we can state reasons or justification. This is often not
easy to do, since justification requires either argument or source identifica-
tion. By “source identification,” I mean a foundational reason that indi-
cates how a belief has been acquired, as visual beliefs have been generated
by vision, inferential beliefs by inference, and testimonial beliefs generated
by the statement of an expert. Thus doubts are commonly expressed by
questioning how a person acquired information asserted, “Did you see it?”
“How do you arrive at that conclusion?” “Who told you that?”

THE JUSTIFICATION REGRESS

Classical Western foundationalism, from Descartes through Russell and

Ayer, is motivated by the worry that a series of reasons, reasons for reasons,
and so on, could be unending. This would land us in a form of skepticism
18 Nyāya Epistemology
since justification is thought to take the form of conditional arguments. Belief
B1 is justified on the condition that the further beliefs B2 through Bn are justi-
fied in that the truth of B2 through Bn would either entail or inductively sup-
port B1. But why should we believe any or all of B2 through Bn? Where do the
justifying beliefs get their justifying charge, so to say, to be able to light up B1?
If they get it from still other beliefs, we seem to be descending stepwise into
an infinite darkness where no belief is really justified.
In the classical Indian context, a problem of a regress of reasons is
pointed out by the famous Buddhist Nāgārjuna (c. 100 ce). There appears
to be an exchange between Nāgārjuna and Gautama on the issue: “What
is the pramāṇa for a pramāṇa? What is the justification for taking a
claim’s being generated by a knowledge source to be a claim we should
believe?” Nāgārjuna asks. He explicitly alleges that infi nite regress
looms, “no stopping point,” anavasthā, for the project that proposes
to resolve doubt and controversy by citing pramāṇa. 3 Probably he has
Nyāya in mind although the name is not mentioned.4 In any case, several
passages in the Nyāya-sūtra, as interpreted by Vātsyāyana, address argu-
ments of Nāgārjuna’s. The crucial passage for Nāgārjuna’s challenge is
Nyāya-sūtra 2.1.16–20. There we have an analogy drawn to the light of
a lamp. Sūtra 2.1.19: “No (justification regress is not stopped by a mere
absence of an additional pramāṇa as making known a pramāṇa cited to
justify a claim). For pramāṇa are certified in the way that the light of a
lamp is used in certifi cation.”5 The lamp is actually a good metaphor to
make one miss Gautama’s point—which is that certifiers may themselves
be certified. The sūtra-kāra, Vātsyāyana says emphatically, is not endors-
ing the Mīmāṃsaka and Vedāntic position of self-certification, svataḥ-
prāmāṇya—those other philosophers talk about a lamp as not requiring
another light to be illumined, i.e., as self-illumining. Self-illumination is
not a Nyāya position. Rather, Gautama is pointing out that an instrument
like the light of a lamp can be both a means and, non-concurrently for
an individual subject S, an object of knowledge. A perceptual cognition
provides S knowledge of its object, typically a thing in the world such as
a pot. But a perceptual cognition may also be an object, as when we say,
“I see it,” or perhaps better expressed, “I see my seeing it.” And by such
apperceptions verbalized as observation statements, or as conclusions of
inferences self-consciously drawn, or as purposeful citings of authori-
ties, controversy normally comes to an end. In other words, unless a
candidate defeater is cognized with respect to conditions governing the
operation of the cited source (bādhaka-abhāvāt), presumption of truth
is restored for a cognition having been brought into doubt. The burden
of proof, the requirement of having good reasons for further doubt, is
on the doubter’s shoulders. Thus there is no regress in a need to cite one
after another pramāṇa.6
A passage near the end of Vātsyāyana’s commentary on Nyāya-sūtra
2.1.20 expressly blunts the regress charge:
 Certification 19
If comprehension of perception or another (knowledge source) landed
us in infi nite regress, then everyday action and discourse would not go
on through comprehension of self-consciously known objects and their
known causes. (However) everyday action and discourse do proceed
for someone comprehending self-consciously known objects and their
known causes: when (self-consciously) I grasp by perception an object
(such as a pot) or I grasp one by inference or I grasp one by analogy
or I grasp one by tradition or testimony (the four knowledge sources),
the (after-) cognition that occurs goes like this: “My knowledge is per-
ceptual” or “My knowledge is inferential” or “My knowledge is from
analogy” or “My knowledge is testimonial.”
And motivation to seek righteousness (dharma), wealth, pleasure,
and liberation proceeds through these comprehensions (whereas if
there is doubt, no such goal-directed activity would occur), as like-
wise motivation to reject their opposites. Such discourse and action
would cease (to be possible for such a subject) if what is alleged (jus-
tificational regress) were indeed to hold. And it is not the case that
there is action and discourse other than this (that proceeds on the
basis of such comprehensions, such fi nal comprehensions) that would
land us in infi nite regress whereby the alleged no-stopping-point
would really obtain.7

The tarka, “suppositional reasoning,” at the end is the argument that other-
wise we would not proceed unhesitatingly but we do.8 It is also interesting
that the practical pursuits mentioned as guided by second-order, reflective
knowledge are: “righteousness (dharma), wealth, pleasure, and liberation.”
In shoring up confidence on such weighty matters, on the big questions of
life, philosophy has purpose.
Nyāya has much in common with both of the two most prominent forms
of what is now called internalism in analytic epistemology. The school
embraces both a kind of foundationalism, although not the Cartesian vari-
ety, and coherentism as shown in the concept of siddhānta, which is a clus-
ter of bits of knowledge within one’s worldview.9 “Established tenets” serve
as premises for inferences throughout the literature, or are explicitly con-
clusions of arguments formally given. They are also winnowing devises for
testing new views, whether of an opponent’s school or a new discovery. The
presumption is that we accept testimony, but not that which is “defeated in
advance” (bādhita) by siddhānta. Thus we encounter the school’s coherent-
ism, a view whose counterpart in the West is criticized for its acceptance
of circular reasoning. But a circle means that we have the resources, poten-
tially, to answer challenges that are themselves only potential. The reason
that there are just certain (quasi-) basic beliefs and not simply a circle of
mutually supporting opinions is that only certain beliefs are putatively the
results of perception and the other sources. So there is also a moderate
foundationalism in Nyāya’s epistemology. Let me elaborate.
20 Nyāya Epistemology

FALLIBLE FOUNDATIONS

Nyāya takes veridicality as a cognitive default: a presentational experience,

anubhava, whether perceptual, inferential, analogical, or testimonial, and
whether veridical or non-veridical in fact (pseudo-perceptual, etc.), is to be
taken to be veridical unless counterconsiderations are evident. The position
is pragmatically motivated but that is not my present point which is rather
the defeasibility thesis. Nyāya subscribes to the principle, “Innocent until
reasonably challenged.”10 There is a fallibilism implicit here that seems to
be motivated by the school’s realism and empiricism. Objects, excepting
cognitions themselves as known apperceptively, are known through pro-
cesses that are not immediately introspectable. In perception, for example,
the working of the process of object-organ connection is complex and sub-
ject to environmental factors such as lighting. All the other sources depend
on the pramāṇa status of perception. Sometimes environmental or other
abnormalities occur, and produced are non-genuine, false (in two senses)
perception-like cognitions, etc., as indeed sometimes we find out, being
able to explain the error (“Well, tin looks a lot like silver”).
Thus although Nyāya embraces an infallibilist concept of a knowledge
source, neither knowledge which has not been certified nor that which
apparently has is immune from the possibility of doubt, of meaningful
doubt. With the exception of an apperception cognizing a target’s content,
any cognition could prove to be false/non-veridical, no matter how dili-
gently we have checked, although with some repeated perceptions and com-
monly performed inferences the barrier to doubt would be high.
Nyāya philosophers are fallibilists about occurrent cognitions with the
one exception. The possibility of error about some indicated external thing
seems ineliminable. Apperception, the “introspection” of some Westerners,
is the sole exception (and not all Nyāya philosophers agree that an apper-
ception has to be true).11 The mainstream view, which becomes prominent
only with Gaṅgeśa and New Nyāya but which is implicit in Vātsyāyana
and the early Naiyāyikas,12 is that the peculiar objects of apperceptions are
cognitions themselves and other psychological properties. It is thought that
there are no intermediaries to consciousness becoming aware of an imme-
diately preceding bit of consciousness including what it is about, its inten-
tionality. An example of an apperception that makes the point plain would
be a cognition C2 whose object or target is an immediately prior cognition
C1 apperceptively cognized as a cognition with a-as-F objecthood whether
or not a is an F in fact. Other types of object cognized exist independently
of our consciousnesses, and the processes of knowing are mysteriously indi-
rect (relying on uncognized physical elements). The objects of cognition
and action are causes of—and not just objects of—different individuals’
truth-hitting mental events. They are external to anyone’s individual cog-
nition, whereas cognitions as objects are not in this way intersubjective
and external. In sum, processes of cognitive generation are not open to
 Certification 21
direct cognitive grasp except in the case of apperception. Many factors in
the generational process are not directly cognizable. This transcendence
of objects to consciousness leads Nyāya to embrace cognitive fallibilism, it
seems to me. The whole range of cognition that is truth-directed (anubhava
= “freshly informative mental event”) has as its normal nature to be veridi-
cal in fact, but only with apperceptions are our cognitions/beliefs guaran-
teed to be true or veridical.
The direct results of perception, inference, and testimony count as basic,
potentially resolving disagreement without requiring further justification.
And just like any new bit of information, putative identification of a belief’s
pedigree in a knowledge source also enjoys a default status of a presump-
tion of truth. This modest foundationalism is supported, as we have seen,
by the argument (tarka) that we couldn’t get along in the world without
relying on the results of perception and the rest. On the second level, where
we are called to check a belief’s pedigree or to prove it by inference, the
principle also holds. Certification we can do, fortunately, when some bit of
knowledge is drawn into question. We have working defi nitions of percep-
tion, inductive procedure, and credential-checking in the case of testimony.
We employ them as the occasion demands.
That Nyāya embraces such a subjective variety of justification is evi-
dent in its concept of pseudo-certification, certification that looks right
from a fi rst-person point of view but that is misleading in fact.13 Appar-
ent certification can be defeated by S’s coming to learn something that
undermines or rebuts a putatively certificational pseudo-inference, but
genuine certification requires that there be no ultimate defeater in fact,
i.e., that S’s evidence for regarding a cognition as veridical would hold
no matter what else she comes to know (bādhaka-abhāvāt). In particu-
lar, a pseudo-perceptual claim can be trumped (“That is yellow” said by
the subject affl icted with hepatitis, a stock example along with the two
moons “seen” by the astigmatic) and corrected by an explanation itself
built up from perceptual evidence, and just what defeats what in what cir-
cumstances is a large topic for Nyāya and other classical epistemologists.
It is recognized that when faced both with a cognition “Fa” for whose
knowledge-source generation there is positive evidence and with “not-Fa”
for which there is also evidence we may not know which is true. Veritable
doubt calls for investigation, and we do not have knowledge if we have
doubt that is genuine. Nevertheless, an unchallenged belief does not have
to be certified to count as knowledge.
It does, however, have to be in principle certifiable. Nyāya holds that
any bit of knowledge is in principle certifiable (that it is knowledge is
knowable),14 though a subject S might not have the resources at a time t
to certify a cognition C of hers that has been source-generated. Thus “cer-
tifiable” here is to be understood in the weak sense that S could certify
C under optimal conditions. The prospect of certification may be remote
while a cognition is nevertheless a knowledge episode. This is because we
22 Nyāya Epistemology
naturally trust certain types of cognition to present the world as it is and
couldn’t get along in it otherwise.
Trust in cognition is understood dispositionally. And it can be built
up—let me expand on the point—by certification. Not only does certifica-
tion fi rm up a person’s confidence, a certified belief that a is F would nor-
mally prevent one from accepting a pseudo-inference that a is not F. To
repeat, depending on other factors, new information could defeat “per-
ceptual” warrant, undermine apparent inferences, and draw into question
previous testimony. But typically by knowing, for example, that fi re is
hot, we reject an apparent proof that fi re is not hot. Our dispositions in
this and similar cases would be, if not reinforced by actual certificational
inferences, so ripe for reinforcement that it is conceded by some Nyāya
philosophers—although self-certification is by them strictly denied—that
the second-order veridicality of a veridicality-confi rming inference may
appear self-evident.15
We fortify dispositions by practice (“training”) as when by frequently
recalling a certain event we make its future recall comparatively easier.
In other words, our acceptance of the conclusion that fi re is not hot is
prevented—defeated in advance, so to say—by a fi rm disposition to act
(and speak) as though it is hot. Defeaters are mental causes. They are bits
of knowledge and dispositions formed therefrom that prevent or correct
mistakes, counterconsiderations that prevent acceptance of views or lead
to their relinquishment. It is a causal law that a cognition that p blocks or
prevents a cognition that ∼p. The one could not be the other’s successor.
New information would have to intervene bringing p into doubt. Similarly,
the perceptual cognition required in the paradigm case of inferring fi re
from smoke has causal relevance. The veridical inferential cognition that
there is fi re on yonder mountain is prompted by a smoke experience along
with a dispositional association of fi re with smoke, fi rm memory of a per-
vasion or invariable connection between the two types of occurrence, fi re
and smoke. The dispositional association of fi re and smoke gets aroused by
a smoke perception and infused into a “consideration” (parāmarśa) that
grasps not just smoke but smoke-as-qualified-by-smokehood-as-pervaded-
by-fi rehood, triggering occurrent inferential knowledge. Statements may
also have causal relevance, generating knowledge in a hearer or, in a debate
context or a scientific treatise (śāstra), blocking acceptance of an opponent’s
views and establishing right positions (siddhānta). In other words, Nyāya’s
causal picture extends to the second level of certification, where there, too,
there may be error. And although certification involves a voluntary act,
mental causes are perfectly legitimate according to Nyāya’s Humean or
interactionist dualism, it bears pointing out.
A few words more on the principle of default un-self-conscious justifi-
cation as holding on the second level of self-conscious certification which
depends on apperception. Apperceptively we know when we are perceiv-
ing something. That is to say, we know when we attend to the perceiving.
 Certification 23
Similarly, we know when we have made an inference if we attend to the
question (at least with sufficient patience having learned to reconstruct
inferences), and so on. Thus we try to answer a question, resolve a doubt
or bit of controversy by saying, “I saw it,” “There is this inference,” and
“S told me so.” Veridicality is assumed at the second level, because, to give
the quick answer, apperception is a form of perception. There is also, as on
the fi rst level, always a possibility of error, not that a scoped cognition has
occurred (a fact that is immediately and indubitably known), but that the
scoped is pramāṇa-born. We cannot arbitrarily call an end to controversy.
But it can and does end, and in these ways supplemented as the case may be
by tarka, “suppositional reasoning.”

EPISTEMIC EXCELLENCES AND DEFECTS

My interpretation finds Nyāya’s epistemology to have internalist features

despite its talk of causal processes. One is the concept of pseudo-certifica-
tion, as pointed out. More importantly, however, there are the ways we come
to recognize the results of genuine pramāṇa. Much theoretical attention is
devoted to special epistemic properties called excellences (guṇa) and defects
(doṣa) which, given doubt or desire to know, are said to be signs of knowl-
edge-source operation and thus key to certification and the establishing of
right positions in philosophy as in everyday life. Now these properties have
to be cognized. They are labelled from an epistemic perspective. They are
“excellences” and “defects” from an epistemic point of view. For instance,
one may make an inference and act on its basis, but to certify that the con-
clusion drawn is the result of inference as a knowledge source is to check
the process to make sure that it meets certain conditions including being
based on knowledge of a pervasion (vyāpti) of F-hood by G-hood, consider-
ing an inference from something a as an F (Fa) to its being a G (Ga), a fact
confirmed with reference to positive correlations—other things both F and
G—or negative correlations—things not-G and not-F—or both. To be sure,
such epistemic excellences are themselves supposed to have causal relevance,
even in inference. They are both properties figuring in causal laws and signs
to us of knowledge sources. For example, our having cognized other things as
F and G is causally responsible (in part) for our new inferential knowledge.
In other words, when doubt arises, identification of knowledge sources
and “excellences” and “deficiencies” as epistemic properties becomes rele-
vant, answering questions, restoring confidence, and ending dispute. Much
about the process, for example, of perceptual generation of knowledge is
extraneous to certification in that certain causal factors—atoms and their
configurations, say Nyāya philosophers in illustration—are not readily dis-
cerned.16 Epistemic excellences and flaws are factors in belief-generating-
processes that make it evident to us that it is a genuine knowledge source
responsible for a belief in question and no mere pseudo-pramāṇa.
24 Nyāya Epistemology
Each separate knowledge source has its own peculiar excellences and
deficiencies; indeed there are several subspecies. We shall look at many
of these in chapters to come, on the individual sources. Before taking the
plunge, let us consider a problem that haunts the Nyāya project overall,
the question of how to identify perception (etc.) in all of its epistemically
relevant subspecies. For example, Nyāya recognizes that the necessary con-
ditions for perception of an absence, “There is no pot on the floor,” are
different from those for perception of a presence, “The pot is on the floor.”
Once special conditions are identified within subspecies of knowledge-
generator, we have to wonder about the boundaries of the perceptual, etc.
What counts as a sensible characteristic and why?17

THE GENERALITY PROBLEM

Nyāya faces a problem of individuation of cognitive processes, a problem it

shares with analytic reliabilism. Analytic epistemologists identify a so-called
generality problem affl icting reliabilist theories in particular: How are we
to differentiate doxastic, i.e., belief-forming processes in an epistemically
relevant fashion?18 How do we chose the candidates whose track records
show them able to confer justification? We should not choose perception
in general, as we normally speak in English, since it is fallible and indeed
too prolific in generating false beliefs, being reliable only in specific circum-
stances. As mentioned in Chapter 1, perception as commonly understood
in English is not Nyāya’s pratyakṣa, which as factive I am tempted to render
“perception*.” In the broader English sense, perception is reliable in the
sense of “truth-conducive” only under certain conditions such as proper
lighting in the case of vision. Only then are sight and on on trustworthy.
What are the salient types of knowledge source? Are sight and hearing
separate sources? What is the source of knowledge of psychological prop-
erties such as pleasures and pains? Is there a special source for knowledge
of absences, as claimed by the Bhāṭṭa Mīmāṃsaka? What are the types of
inference? Is arthāpatti, “circumstantial implication,” a separate knowl-
edge source, as claim Vedāntins as well as Mīmāṃsakas? Are there invari-
ant criteria for testimonial expertise? With regard to these questions, which
drive the research of several prominent Nyāya philosophers, the main les-
son, it seems to me, is that specific rules have to be formulated for sub-
types of knowledge source such as vision. Naturally Nyāya philosophers do
not fold up their tents in the face of their generality problem but work at
specifying rules and subrules. They stress that the special excellences that
indicate knowledge sources do vary—in the case of perception according to
sensory modality and other factors—as do common epistemic deficiencies.
There are nevertheless common rules and common signs.
Nyāya does indeed share with reliabilism the problem of specifying
appropriate conditions governing subspecies of perception and the rest of
 Certification 25
the pramāṇa. But it seems to me Nyāya’s problem is distinct and not as
severe as the reliabilist’s because of the understanding of pramāṇa as fac-
tive. By holding candidates to the highest standard—one-hundred percent
reliability—entries to the field are severely restricted. The sources as con-
ceived by Nyāya may well be too restricted—there seem to be bits of knowl-
edge not generated by any of Nyāya’s four (a topic for our last chapter)—but
Nyāya does not have the opposite problem of letting the actually false and
unreliable count as source-generated. Furthermore, Nyāya begins with the
assumption that normally presentational experiences are veridical as well
as with the idea that there are perfectly reliable knowledge sources. It does
not have the problem of how to get started in carving up the pie in that,
with the exception of analogy, its three sources are commonly so regarded
in everyday discourse (vyavahāra).
It is worth dwelling on this last point. We do in English, too, commonly
recognize perception and the others as certificational. That this is a com-
mon human practice (not just a language game in Sanskrit) Vātsyāyana and
company may be interpreted as intent to show in their comments on Nyāya-
sūtra 2.1.20. Both certification and discertification in the Nyāya fashion are
commonly expressed in English, for sometimes we say, for example, “S is
indeed over there, since I see him,” and “You couldn’t really have perceived
S because condition Y does not hold” (“You can’t see anyone from this
distance”).
Nyāya philosophers bring to bear on their generality problem theoriz-
ing about universals and quasi-universals (upādhi) and the signs whereby
such critters are recognized. Gaṅgeśa and his followers embrace the view
that perceptionhood, inferencehood, and so on are, well, not strictly speak-
ing true universals, jāti or sāmānya, but rather kinds that are themselves
known in the ways universals are known, namely, by concrete indicators of
instances in invariable relationships, such as, for cowhood, possession of a
dewlap. Other characteristics that make the cowhood universal manifest in
what is called “recurrent cognition,” anugama-pratyaya, are a cow’s shape
and face and horns. Recurrent cognition is a kind of perception informed by
previous cognition of the same type of thing, “A cow, another, and another,
and so on,” in the example of the Vaiśeṣika philosopher Praśastapāda
(sixth century).19 Verbalizable perception is concept-laden, savikalpaka-
pratyakṣa, and concepts are formed by prior experience. Universals (and
quasi-universals) are posited as the unifiers of the unity among a group of
individual objects perceived.
There are technical, nitpicking reasons why being-a-perceptual-cogni-
tion, being-an-inferential-conclusion, and so on are not to be counted genu-
ine jātis; veridicality is not itself a universal, Gaṅgeśa takes great pains to
show as nevertheless he provides defi nitions that cover all cognitions that
are veridical.20 Raghunātha says explicitly that the unanalyzable upādhi
or “surplus property” is recognized through recurrence just as is the true
universal or jāti. 21 The point is that the unanalyzable upādhi is not entirely
26 Nyāya Epistemology
mind-imposed—maybe in its precise contours, or maybe it is a thoroughly
objective property—but in any case it is recognizable by others, behaving
mostly like a universal. Veridicality is not a true universal because it is not
“locus-pervading,” vyāpya-vṛtti, whereas cowhood is present in every part
of a living cow, horns to hoofs. A cognition can be in one part veridical
and one part not, like a monkey in contact with the branches of a tree but
not in contact with the roots. A cognition of “This is silver” in the face of
mother-of-pearl is veridical in presenting a “this” (there is something there
to be picked up) but non-veridical in its silverness portion. All verbalizable
cognitions have both qualifier and qualificandum indications within their
intentionality or objecthood (eka-vṛtti-vedyatā).
Let us not get bogged down now in technical considerations that are
mainly irrelevant to epistemology. The point is that being-a-perception and
the others are properties that like true universals are known by certain
repeating characteristics, such as, in the case of perception, being-imme-
diate or presenting-its-object-immediately as opposed to being-an-inferen-
tial-result which concerns an object known mediatedly, unperceived fi re on
yonder mountain, for instance. Some pramāṇa-indicative properties may
take some work to uncover, such as “being-produced-by-contact-of-sense-
organ-and-object.” But though unnecessary, identification of a cognition
as being so produced would be sufficient to generate the recognition of the
type, namely, that the cognition is perceptual. And this is, as we have seen,
one way a justification regress can terminate. A cognition’s being found
to be perceptual, inferential, analogical, or testimonial ends the matter,
unless there is a good reason to doubt the second-order judgment that the
fi rst-level is pramāṇa-born. Good reasons, of course, sometimes arise. But
unless they do, we may go about our business.
I have said that perceptionhood is not in the technical terms of Nyāya a
true sāmānya or jāti but only a quasi-universal. Nevertheless, this abstract
character is thought of, as we might say in less technical English, as the
character it is as the result of a natural process involving a sense faculty in
relation to an object presented, a process involving a sense organ working
in optimal conditions, we might say, and it is really irrelevant the precise
place of the entity in the ontological system. The point of calling or at least
treating it as a jāti is that although only quasi it is known as a universal
is known, as I have explained. But some Nyāya philosophers do appear to
take it to be a genuine natural kind, for example, the late Navya textbook-
writer Viśvanātha of the seventeenth century. 22
Also on point is the famous Tattva-cintā-maṇi commentary of
Mathurānātha, likewise of the seventeenth century, on a couple of sentences
from Gaṅgeśa about pramā, “veridical cognition,” dividing into four types,
the perceptual, inferential, analogical, and testimonial. These are mutually
exclusive but not exhaustive of all pramā since there can be cognitions that
are veridical in part as well as accidental pramā that are not generated by
the four types of knowledge source and thus not knowledge (prāmāṇya).
 Certification 27
Gaṅgeśa says, immediately after defi ning veridicality in general, that
there are four subkinds: “And such veridical cognition (pramā) occurs in
four varieties, in that the perceptual, the inferential, the analogical, and the
verbal are distinct. In this way, there are proximate instrumental causes
for the four—i.e., ‘means to veridical cognition,’ pramāṇa—in four variet-
ies, in that perception, inference, analogy, and testimony are distinct.”23
Mathurānātha comments:

He (Gaṅgeśa) divides that which he has defi ned with the words “And
such veridical experience (comes in four varieties)”—that is to say,
veridical experience (pramā), in the form of an “(informative) experi-
ence whose predication content hits that which has it” (tadvati tat-
prakāraka-anubhava)—this is the meaning. The meaning of his use
of “in four varieties” is that these four are basic kinds of (presenta-
tional or informative) experience (anubhava) non-mediatedly, with no
universal intervening (between being-a-perception, etc., and being-an-
experience, anubhavatva). That these four basic, non-mediated kinds
of experience are veridical experiences, well, this is the implication
(although veridicality is no intermediate kind). Thus that some individ-
ual veridical experience fall outside the quartet would cause no harm
(to our view).
Against the objection that this quartet of (so-called) natural kinds
(jāti) not being ordered as pervaded and pervader (with respect to
being-an-experience, anubhavatva, unlike being-earthen and being-a-
substance) there could be the fallacy of cross-classification in any single
veridical experience (which could be both, e.g., a perception and an
inference), like being-an-element and having-a-fi nite-form (which are
also not ordered as pervaded and pervader and thus are not true jāti,
natural kinds, since ākāśa is an element but does not have fi nite form
and manas has fi nite form but is not an element), he (Gaṅgeśa) says,
“Perception, (inferential cognition, and so on),” (spelling out explicitly)
“perceptual knowledge, inferential knowledge, analogical knowledge,
and testimonial knowledge,” because individual veridical cognitions do
mutually divide up in this way (i.e., if the one, then not the other). And
in this way there is no cross-classification because individual veridical
cognitions, which support (the abstract character, being-a-perception
and so on), do mutually divide up in this way. This is the upshot.
Having divided up veridical cognitions, he goes on to divide up
knowledge sources. With the words, “in this way,” and “in four
varieties,” he implies that a quartet of properties (the perceptual as
a pramāṇa, the inferential as a process of knowledge, the analogi-
cal, and the testimonial) splits up the knowledge sources, that to
the quartet of properties splitting up the knowledge sources can be
attributed instrumental causality for knowledge (i.e., for veridical
cognition, pramā). 24
28 Nyāya Epistemology
In other words, perceptionhood and the rest are basic kinds of presenta-
tional or informative experience (anubhava), all pervaded by veridicality
although veridicality is not a natural kind and there can be accidentally
veridical cognitions. Another kind of anubhava is suppositional reason-
ing (tarka) as well as imagination (vikalpa) and a miscellaneous group of
false cognitions, some misleading, and some, like my pseudo-sight of a
double moon (I have astigmatism), not misleading in that they are known
to be illusions when they occur. Being-a-memory, smṛtitva, contrasts with
anubhavatva as a second kind of cognition, whose instances divide into the
correct, yathārtha, and the incorrect, ayathārtha. 25
Personally, my impression is that New Nyāya philosophers often use
the terms ‘jāti’ and ‘upādhi’ loosely, almost interchangeably, although
the official doctrine, due to Udayana, is that a candidate that fails the
cross-classification test or another “universal-blocker,” jāti-bādhaka, is
not a true universal but only an upādhi, an accidental or mind-imposed
kind. 26 There is also the blocker of failing to be locus-pervading as cow-
hood of every part of a living cow, as we noted previously. The reason
the upādhi/jāti distinction is often less than sharp is that, one, not many
candidates pass all the tests, and, two, to be a candidate, that is to say,
to be at least an upādhi character that is evidenced by recurrent cogni-
tion and that is either pervaded by or pervades another character, is all
we need to carry out inferences, inferences being underpinned by perva-
sion relationships (vyāpti) between class-character terms. Pervasions do
not have to be between the extensions of two true universals, although
paradigmatically that’s how we may think of the matter. Pervasion of one
upādhi by another is sufficient to permit inference. But if we do want to
stress the depth of a categorial cut, a distinction between upādhi and jāti
can be made, as does Mathurānātha in the previous passage. So if there is
any genuine natural kind here, it is being-a-perception along with the oth-
ers which, as Mathurānātha sees things, are basic kinds of presentational
experience all of which are veridical.
Recurrent experience (anugama-pratyaya) can be of the class character
itself, as in the example of Praśastapāda, “A cow,” “Another,” “Another,”
and so on, or of an indicator of a class character, such as “A dewlap” indi-
cating a cow and cowhood. Note that there is nothing essential to a cow in
having a dewlap. That is simply a reliable indicator of something being a
cow. Similarly, the perceptionhood or inferencehood of a cognition in ques-
tion can be known by indicators, called guṇa, “excellences,” to bring out,
as I have said, their epistemic relevance.
In sum, we are able to fi nd marks correlating with individual knowl-
edge sources (jāti-vyañjana) whereby we sift out pramāṇa pretenders such
as guessing. There are no recurrent indications of a connection between
guessing and the truth of the thing guessed—there is no discernible pattern,
no correlating marks, unlike the case with perception, inference, analogy,
and testimony—and so guessing is no pramāṇa.
 jñƗna, cognition

anubhava, sm‫܀‬ti,
presentational
remembering
experience

pramƗ, the apramƗ, the yathƗrtha, ayathƗrtha,

veridical non-veridical correct incorrect

pratyak‫܈‬a, anumiti, upamiti, ĞƗbda-bodha, samĞaya, doubt

testimonial viparyaya, ƗhƗrya-
perceptual inferential analogical
falsehood, error buddhi,
knowledge knowledge knowledge knowledge fantasy,
imaginative
pratyak‫܈‬a- anumiti- upamiti- ĞƗbda- miscellaneous supposition
ƗbhƗsa, ƗbhƗsa, ƗbhƗsa, ƗbhƗsa,
pseudo- pseudo- pseudo- pseudo-
perception inferential analogical testimonial tarka,
cognition cognition cognition suppositional
reasoning

Figure 2.1 Types and subtypes of “cognition,” jñāna.

Certification
29
30 Nyāya Epistemology
Finally, there are famous issues of whether certain candidate sources are
reducible to one of the four. Nyāya often takes a reductionist position but is
in principle open to recognition of additional sources. This comes out pretty
clearly in, for example, the postulation of analogy, which is said to have
testimony as well as perception as auxiliary causes, as well as in Nyāya’s
defense of testimony against Buddhist and Vaiśeṣika reductionisms.27 This
is an important point which will be expanded in Chapter 7 in connection
with Nyāya’s apparent failure to account for a priori knowledge.
Despite such championing of testimony and analogy, Nyāya argues that
theoretic simplicity and other considerations counter, for example, the
Bhāṭṭa Mīmāṃsaka insistence (1) that non-cognition, anupalabdhi, is the
source of knowledge of absences and (2) that “circumstantial implication,”
arthāpatti, is not a form of inference but a pramāṇa on its own. Both Old
and New Nyāya take the contentions seriously, but object to the adding
of a separate source in each case, arguing that the origination of absential
bits of knowledge, etc., can be explained perfectly well using only the four
sources. We shall take up the topic of circumstantial implication in the
inference chapter (Naiyāyikas say it is the same as negative-only inference)
and non-cognition in the next chapter which is on perception.
Let me emphasize just one further general point, namely, that taxonomi-
cal questions arise aplenty in classical Indian epistemology, as we can see by
taking an overview: Yogācārins accept two sources, perception and infer-
ence, Bhāṭṭa Mīmāṃsakas along with certain Vedāntins accept six, Nyāya’s
four plus non-cognition and circumstantial implication, Prābhākaras five,
rejecting non-cognition but accepting circumstantial implication, the oth-
erwise Naiyāyika Bhāsarvajña fi nds only three, seeing analogy as a form
of testimony, and other positions are taken by Vaiśeṣikas, Jainas and other
players. (One school accepts only perception, rejecting even inference, the
Cārvāka, whose argument that perception is the only source becomes a
(laughing) stock example of self-defeating reasoning.)28

BELIEF-WARRANTING TARKA, “SUPPOSITIONAL REASONING”

Somewhat surprisingly, Nyāya holds from the earliest that apparent certifi-
cation may not be enough to warrant belief in some instances. Even if our
beliefs/cognitions have indeed been generated by processes that would be
counted pramāṇa did they not face counterconsiderations, in facing counter-
considerations—in being reasonably challenged—they are not trustworthy
and do not guide unhesitating effort and action. There is a social dimen-
sion to knowledge, where reasoning reigns resolving controversy in ways
over and above the sources. These are the ways of tarka, “hypothetical” or
“suppositional reasoning.” Paradigmatically, tarka is called for in order to
re-establish a presumption of truth in favor of one thesis that has putative
source support against a rival thesis that also has putative source support,
 Certification 31
a thesis and a counterthesis both backed up by, for example, apparently
genuine inferences (the most common situation) or by competing percep-
tual or testimonial evidence. By supposing the truth of the rival thesis and
(in Socratic style) showing how it leads to unacceptable consequences or
breaks another intellectual norm, one repossesses a presumption of truth,
provided—the classical reasoners never tire of emphasizing—provided
one’s own thesis does indeed have at least the appearance of a pramāṇa in
its corner. 29 Nyāya philosophers join a consensus across schools that such
arguments are not in themselves knowledge-generators although they can
swing the balance concerning what it is rational to believe.
“Suppositional reasoning,” tarka, is what a philosopher is good at,
drawing out of implications of opposed views and testing them against
mutually accepted positions (siddhānta), according to, broadly speaking,
criteria of coherence but also of simplicity. Here we come to the vital cen-
ter of Nyāya, the secret to the life and prosperity of a Nyāya philosopher,
which is reflected in honorific appellations, for example, “tarka-śiromaṇi,”
“Crest-jewel of Reasoning,” and book titles, Tarkāmṛta, “Immortal Nec-
tar of Reasoning,” Tarka-sāra, Tarka-saṃgraha, and so on (I count more
than fifty philosophy titles beginning with the word). “Dialectical reason-
ing” is the translation I have used in earlier works, since tarka presupposes
a context of controversy.30 Tarka focusing on simplicity (“lightness”) or
“theoretic cumbersomeness” (lāghava and gaurvava) is not so much sup-
positional as comparative. To show theoretic “lightness” is to use anukūla-
tarka as opposed to pratikūla-tarka, i.e., suppositional reasoning that is
“favorable” as opposed to “unfavorable,” “with the grain” as opposed to
“against the grain.” The latter is not the sole but is the standard variety,
showing that an opponent’s hypothesis (or an opposite thesis, ~p) violates
an intellectual norm.31 There is at least one other form of “favorable” tarka,
which is having one’s own thesis presupposed by the opponent’s while the
reverse does not hold—see (6) in the next paragraph. Still, usually tarka is
thought of as “unfavorable,” only indirectly supporting a thesis p by estab-
lishing a counterconsideration against a rival, ~p, or, sometimes, against
a rival hypothesis, q, which would explain a set of commonly recognized
truths in different terms.
Udayana (eleventh century) appears to inherit a six-fold division
of tarka according to the nature of the error in an opponent’s position,
and expressly lists five types (the sixth, “contradiction” or “opposition,”
vyāghātatā, either being assumed as the most common variety, or sub-
sumed under Udayana’s fi fth type, “unwanted consequence”). Philosophers
from other schools present distinct but overlapping lists. The Nyāya text-
book writer, Viśvanātha, of the early seventeenth century, mentions ten,
Udayana’s five plus five more, many of which are used by the Advaitin
Śrīharṣa (probably Udayana’s younger contemporary) among other rea-
soners.32 They are: (1) ātmāśraya, “self-dependence” (begging the ques-
tion); (2) anyonyāśraya, “mutual dependence” (mutual presupposition);
32 Nyāya Epistemology
(3) cakraka, “circularity” (reasoning in a circle); (4) anavasthā, “infi nite
regress;” and (5) aniṣṭa-prasaṅga, “unwanted consequence” (including con-
tradiction presumably)—Udayana’s five—plus (6) prathama-upasthitatva,
“being presupposed by the other,” the “fi rst established” (a form of “favor-
able” tarka); (7) utsarga, “(hasty) generalization;” (8) vinigamanā-viraha,
“differentiation failure;” (9) lāghava, “theoretic lightness;” and (10) gau-
rava, “theoretic heaviness.”33 There is a lot to say about some of these, but
the logic of the fi rst three errors is pretty plain: concluding A from A, or
A from B and B from A, or, in a wider circle, A from B, B from C, and C
from A, and so on in little circles (Nyāya siddhānta is a very large circle, it
sometimes seems). Some of the other types will be taken up by us in later
chapters. Let me close this one with a few words about self-contradiction
and its varieties.
Three kinds of self-contradiction come to be recognized: in language
(“My mother is barren”), in thought (“I do not know this”), and in action
(a speaker saying, “I am mute”).34 As we saw at the beginning of the chap-
ter, the tarka that establishes a presumption in favor of Nyāya’s knowl-
edge-source epistemology concerns action. People do unhesitatingly, post
philosophy, pursue dharma, wealth, pleasure, and even liberation, mukti,
having justified, to themselves at least, the pursuit on the basis of certified
inference or testimony. Similarly, the tarka that refutes skepticism in gen-
eral concerns action. Gaṅgeśa:

Were a person P, who has ascertained thoroughgoing positive correla-

tions (F wherever G) and negative correlations (wherever no G, no F),
to doubt that an effect might arise without a cause, then—to take up
the example of smoke and fi re—why should P, as he does, resort to
fi re for smoke (in the case, say, of a desire to get rid of mosquitoes)?
(Similarly) to food to allay hunger, and to speech to communicate to
another person?35

The argument, which is taken from the tarka section of Gaṅgeśa’s chapter
on inference, is essentially the same as that with which Vātsyāyana opens
his Nyāya-sūtra commentary quoted here at the beginning of the chapter.
Without the confidence that presupposes knowledge, we would not act as
we do.
3 Perception

Remarkably, the word ‘perception’ in English shares an ambiguity with its

counterpart in Sanskrit, ‘pratyakṣa’, both being used for (1) a mental event,
a bit of occurrent knowledge that is perceptual in character, as well as (2)
the process that produces it. Consonantly ‘perception’ will be used here,
in English, rendering Nyāya, both as the name of the type of knowledge
and its source. For example, my current visual experience is perception
(fi rst sense) that has been generated by a visual process also called per-
ception (second sense). Nyāya uses two words for inferential knowledge
and the inferential process (the former called anumiti, the latter anumāna),
as, too, with analogy (upamāna), which generates analogical knowledge
(upamiti), and testimony (śabda), which generates testimonial knowledge
(śābda-bodha). But context often has to serve to disambiguate references to
perceptions as bits of knowledge and perceptual processes.
Perceptual cognition or knowledge, like inferential knowledge, etc., is
a type of presentational experience, anubhava, which is defi ned as expe-
rience having intentionality or objecthood, viṣayatā. Perception has both
phenomenological character and an epistemological role. Early Nyāya phi-
losophers move back and forth between thinking about it as richly presenta-
tional without the mediation of concepts or words and thinking about it as
embedding, and justifying, certain conceptually mediated propositions. On
the one hand, perceptual cognition has a phenomenological inside, richer
than anything anyone could say about it or its objects, and on the other,
the perceptual process plays a foundational role in giving us knowledge—
and not just perceptual knowledge but also the inferential, analogical, and
testimonial. As we shall see in later chapters, the other sources depend on
perception as a pramāṇa.
The thesis tying together perception’s phenomenological and epistemo-
logical sides is naïve or direct realism. If perception is phenomenologically
rich, the world is too, indeed, richer, according to Nyāya.
Possibly following Kumārila (eighth century)—the early Mīmāṃsaka
fellow realist who presents a barrage of arguments against Buddhist
positions—Vācaspati (tenth century) uses the NyS definition of perception
to distinguish “concept-laden” from “concept-free” perception, savikalpaka
34 Nyāya Epistemology
and nirvikalpaka pratyakṣa.1 Only the former has full propositional content.
The NyS definition runs: “Perception is cognition generated from sense organ/
object connection, not expressible in words, unerring, and determinate.”2 To
bring out perception’s phenomenological side, Vācaspati follows Uddyotakara
(seventh century) on avyapadeśya, “not expressible in words.” Uddyotakara:

Perception is cognition that varies according to differences in the object

(or objects) perceived. It happens to a subject S for whom the relation
of word and object has not been employed (who does not know the
object’s name). Even if S does know the object as denoted by its name,
at the time of the cognition, it happens just as if S did not know. 3

This statement brings out early Nyāya philosophers’ apparent agreement

with Buddhist epistemologists of the Yogācāra school that in itself per-
ception is concept-free, nirvikalpaka, free from thought and imagination,
vikalpa or kalpanā. According to the Buddhists, we project ideas (vikalpa)
and conceptually color perceptions, constructing convenient fictions
according to our desires. So-called concept-laden perception, which is thus
determinate and verbalizable, is really a form of inference.
Vācaspati wants to sharpen the disagreement with Buddhists he sees
implicit in the way his Nyāya predecessors understand perceptual cognition
as the result of a knowledge source. Thus though he agrees with Uddyo-
takara about the concept-free variety, he takes the last word of the NyS
characterization, vyavasāyātmaka, “determinate,” to refer a distinct type
of perception, the concept-laden, savikalpaka. For the Buddhists, this kind
of cognition is indeed propositional, but, they say, even sensory predication
depends on ideas of unreal generality—all predication involves repeatable
general terms. Thus its propositional content is suspicious just because this
is not raw perception which alone is capable of presenting the real, the sva-
lakṣaṇa, “that which is its own mark,” the “unique” or “particular.” Nyāya
philosophers hold that perception is none the worse for being concept-laden
in that concepts are features of the world as impressed upon the mind or
self. Perception founds true beliefs, and the repeatable predicates and con-
cepts (“cowhood”) perceptually acquired and re-presented and verbalized
pick out constituents of real objects, things that do re-occur (there are lots
of cows in the world). For the New Nyāya philosophers, concept-laden
perception comes so to dwarf in importance the indeterminate, concept-
free variety that the latter becomes problematic. Perception in its episte-
mological role is concept-laden. Otherwise, it could not be a pramāṇa, and
identification of it would not be certificational. Perception as a knowledge
source is a doxastic, belief-generating process. Perceptual beliefs—or any-
how bits of perceptual knowledge and their verbalizations in observation
statements—are dependent on concepts. To believe or say that there is a pot
on the floor, one must possess the concepts of “pot” and “floor.”
 Perception 35
The importance of the Nyāya difference from the Buddhist position is eas-
ily grasped. Concept-laden perception is capable of deciding some weighty
questions, according to Nyāya, for example, the endurance of self or con-
sciousness, according to Nyāya philosophers of all periods. Perceptual rec-
ognition of an individual as the same over time is in principle of a piece with
an observation that something present is currently qualified by a perceptible
characteristic, in that all determinate perception involves projection, so to
say, into current experience of a perceptual concept acquired previously.
Concept formation is dependent on the capacities of the sense organs to
gather property-wise information from objects directly sensed, and it is here
that there is a role for the concept-free variety of perception—according to
Gaṅgeśa and other but not all Navya Nyāya philosophers, for some of whom
all seeing is really “seeing as.”4

CONCEPT-LADEN VS. CONCEPT-FREE PERCEPTION

Kumārila mentions the cognition of an infant as an example of percep-

tion that is concept-free.5 Phenomenologically humans would seem to have
much in common with animals considering this type. However, perception,
according to the great Mīmāṃsaka, does not so much divide into types as
form a process with the concept-free as the fi rst stage. Awareness of the
object is only quasi-propositional in the fi rst moment, and at the second
has its content fi lled out to become the means whereby an individual is
ascertained to have a certain character, to be a certain kind of substance or
to possess a universal or an action, etc.6
Even a baby or an animal has some minimal conceptual capacities,
or quickly develops some. Consider for example the perceptual predicate
“edible.” A dog has a pretty good ability to tell what’s edible from what
is not, and a baby less but still some of the virtue.7 An adult’s experience
is much more textured when it comes to edibility, with a richer palette of
concepts to inform whatever current experience. In any case, for Nyāya
ascertainment of edibility or any other sensible characteristic is proposi-
tional, that is to say, can be expressed in a declarative sentence, “That’s
blue,” for instance. For Kumārila, the pioneer of the theory, the object per-
ceived, the lotus (or whatever), is known in the fi rst stage as an individual
whole, both in its individuality and as having a character. But the character,
the thing’s being blue as opposed to red, and it’s being here right now, are
not known without the mediation of concepts. Seeing is “seeing as” and
is “shot through with words,” to use the expression of the grammarian,
Bhartṛhari.8 Kumārila, unlike Bhartṛhari, is a realist about ideas or con-
cepts, like his Nyāya colleagues, who quote him often in their own fights
with Buddhist subjectivists. Concepts begin as features of things, whether
as substances, universals, actions, or another category.
36 Nyāya Epistemology
The Buddhists’ best argument for their subjectivism (which one suspects
derives more fundamentally from a commitment to the possibility of a uni-
versal nirvāṇa-experience, although this is not said) is perceptual illusion.
Illusion proves that a perception’s object is not a feature of the world but is
contributed somehow from the side of the subject. A rope can be perceived
as a snake, with no difference, from the perspective of the perceiver, between
the illusion and a veridical snake perception. Similarly, dreams are the “per-
ceptions” of a dreamer but do not touch reality. (Our world is a dream, say
Buddhists, and we should try to become buddha, “Awakened.”)
One way to resist the pull of the illusion argument belongs to the
Prābhākara Mīmāṃsaka. Prabhākara was a renegade pupil of Kumārila
with the reputation of outstripping his teacher in philosophic acumen and
coming to be known as “the teacher” (guru) across the classical schools.
The Prābhākara insists that not only is perception invariably veridical but
also cognition in general, jñāna. Nyāya philosophers hold in contrast that
illusion is a false cognition, mithyā-jñāna. Thus they agree with the sub-
jectivists, that is, in the fight between Nyāya’s “misplacement” view and a
Prābhākara “no-illusion” or “omission” theory. For the Prābhākara, illu-
sion is a failure to cognize of a certain sort, an absence of cognition, for
example, an absence of cognition of the difference between a remembering
of silver and a perceiving of mother-of-pearl when holding in hand a piece
of shell and failing to cognize the difference S exclaims, “Silver!” Nyāya
disagrees: sometimes a person S apparently perceives a to be F—has an
apparently perceptual cognition embedding Fa, a single cognition of an
object as qualified (viśiṣṭa-jñāna)—when a is not in fact F, while S cannot
discern from her own fi rst-person perspective that the cognition is non-
veridical. Nevertheless, the predication content, according to Nyāya as also
Mīmāṃsā, the presentation or indication of F-hood, originates in things’
really being F, through previous veridical experience of F-hood.
Here we touch the heart of classical Indian realism. Snakehood is avail-
able to become illusory predication content through previous veridical
experience of snakes. It gets fused into a current perception by means of a
foul-up in the normal causal process through the arousing of a snakehood
memory-disposition (saṃskāra) formed by previous experience of snakes.
The content or intentionality (viṣayatā, “objecthood”) of an illusion is to
be explained causally as generated by real features of real things just as
is genuine perception although they are distinct cognitive types. Illusion
involves the projection into current (determinate) cognition (which would
be pseudo-perception) of predication content preserved in memory. Some-
times the fusion of an element preserved in memory is cross-sensory, tast-
ing sourness, for instance, when perceiving a lemon by sight or smelling a
piece of sandalwood which is seen at too far a distance for actual olfac-
tory stimulation.9 These are cases of veridical perception with an obvious
admixture or tinge of memory. Nyāya philosophers mention them to make
vivid the point that perceptual concepts can be so enlivened as to appear
 Perception 37
purely sensory when we know they cannot be. Illusion, according to Nyāya,
is to be analyzed similarly, but unlike veridical cases of projection illusion
involves taking something to be what it is not, a seeing or perceiving it
through a “misplaced” qualifier. This means that concept-laden perception
is necessarily combinational—a position taken by Gautama himself, the
sūtra-kāra, and much elaborated by Vātsyāyana and the other commenta-
tors, in a stretch of sūtras apparently aimed at an early form of Buddhist
subjectivism.10
Thought-laden perception, determinate perception, gets its content not
only from the object in connection with the sense organ but also from the
classificational power of the mind (or self). With the perceptual cognition,
“That’s a pot,” for instance, the pot as an individual in connection with a
sensory faculty is responsible for the awareness of a property-bearer, for
what is called the qualificandum portion of the perception, without admix-
ture of memory, but the sensory connection is not by itself responsible for
the qualifier portion, the pothood, that is to say, the thing’s classification
as a pot. Now the classificational power of the mind (or self) should not be
understood as innate, however, so much as the product of presentational
experience (anubhava) over the course of a subject’s life. Repeatable fea-
tures of reality get impressed on the mind (or self) in the form of memory
dispositions. For most adults, prior determinate cognition provides the con-
tent predicable of a particular, or a group of things, presented through the
senses. That is, in perceiving a as an F, an F-saṃskāra formed by previous
knowledge-source-produced bits of occurrent cognition of things F would
be a causal factor. The perception’s own content includes the repeatable
nature of the qualifier through the operation of this factor. We see the tree
as a tree.
But sometimes neither a prior determinate cognition nor a memory dis-
position is at all responsible for the predication content, for example, when
a child sees a cow for the very fi rst time. Rather, an “in the raw” perceptual
grasping of the qualifier (cowhood) delivers it to an ensuing concept-laden
and verbalizable perception. In other words, there are cases of determinate
cognition where indeterminate, concept-free perception furnishes the qual-
ifier portion independently and the ensuing concept-laden perception is not
tinged by memory. Normally, saṃskāra, “memory-dispositions,” do play a
causal role in determinate perception, according to Nyāya. But sometimes
an immediately prior concept-free perception of a qualifier plays the role of
the saṃskāra, furnishing by itself the concept, the predication content, the
qualifier portion of an ensuing determinate, proposition-laden perception,
which is the type of cognition that founds our beliefs about the world.
If this were not an “immaculate perception” but itself a grasping of a
property through the grasping of another property, we would be faced with
an infi nite regress and direct perception of the world would be impossible.
Concept-free perception need not provide the classifying not only with sec-
ond and third-time perceptions of something as F but not even, strictly
38 Nyāya Epistemology
speaking, with a fi rst-time perception, since there could be an intervening
cognitive factor (such as in the stock example of the working of analogy
where an analogical statement furnishes the qualifier). But with that fac-
tor again the question would arise how it gets its content, and so since an
indeterminate perception has to be posited at some point to block a regress
it might as well be at the start. This is the main argument of Gaṅgeśa, the
New Nyāya systematizer, in defense of positing nirvikalpaka as a type or
fi rst stage of perception.11
A distinct line of reflection informs the view that determinate, concept-
laden perception is a cognizing of a qualificandum through a qualifier:
things have multiple properties some of which normally go undetected on
any given occasion of experience. I can see Devadatta without seeing what
he has on his back. Wholes are implicit in their parts, the very notion of
which makes no sense cut off from that of the whole: parts are parts of a
whole.12 And if I can touch what I have seen, then when I am only touching
the thing, I will not normally be aware of the thing’s color perceptually. If
the ontological layering of things having qualifiers were not reflected in the
causal ordering of an indeterminate perception feeding, so to say, the deter-
minate, then perception of a qualificandum should entail that the “thick”
particular be presented, i.e., the thing with all of its properties, and, as
Gaṅgeśa argues, a blind person in touching a yellow piece of cloth would
know its color as yellow.13
Nevertheless, for all intents and purposes, perception embodies beliefs.
More accurately, a perceptual belief is the result of the operation of percep-
tion as a knowledge source. Everything that is nameable is knowable and
vice-versa. There is nothing that when we attend to it cannot bear a name,
for we can make up names. We can in principle verbalize the indications
of our experience, though many of them are not named since we are indif-
ferent (pebbles perceived along the road). Concept-free perception is the
Nyāya rendering of our ability to form perceptual concepts by attending
to perception’s phenomenological side. Epistemologically, it plays no role,
since it is itself a posit and is unverbalizable and not directly apperceived.14
We shall return to the point about apperception in the last section of the
chapter. Let us turn now to look at determinate perception’s importance to
what we may justifiably believe and to conclusions of philosophy.

RECOGNITION

Typically verbalizable perception is infused with concepts formed by previ-

ous experience. But presentational experience is not just of qualifiers but
also of individuals, called qualificanda (viśeṣya), i.e., individual property-
bearers, which come into connection with the sense organs and are known
at once through one of another property they bear. Indeed, individuals
 Perception 39
bear whole complex groups of properties through which they are known.
Although the notion of a qualificandum is an abstraction from presenta-
tional experience—a bare particular is never encountered directly, at least
not as bare—perception forms not only qualifier-type dispositions but also
memories of individuals. For example, a subject S can remember Deva-
datta whom S saw yesterday. Such a mental event is called a remembering,
smaraṇa, which occurs in the present but whose object is not only past but
known as past. Perception, in contrast, is about an object qua qualifican-
dum currently in connection with one or more of the sense organs. Sensory
connection with properties is said to work through, or presuppose, connec-
tion with the property-bearer.
Now having acknowledged the combinational role of concept-laden per-
ception as presenting a qualificandum a as bearing a property F, Nyāya
philosophers extend the—let us keep in mind—thoroughly legitimate role
of memory dispositions in perception to include recognition of an object
as the same qua qualificandum as an object experienced in the past. For
example, S’s recognition of Devadatta, “This is that Devadatta (whom I
saw yesterday),” is taken to be perceptual. The apparent co-intentional-
ity of the “this” and the “that” presented—or the co-referentiality of the
two terms used in the characteristic verbalization—is not a statement of
synonymy since the “that” indicates (elliptically) an object as past (i.e.,
temporally distant, since otherwise “this” would be the appropriate term).
The “that” may seem dubiously perceptual, but Nyāya sees little difference
in a saṃskāra playing a causal role here from its playing a role in other
cases of determinate perception. The difference is that in recognition as
in remembering we are aware that the thatness is a matter of memory and
past experience whereas this is not the case with a simple perception of a
sensory characteristic.
The Nyāya position is that a recognition is a single determinate percep-
tual cognition. The Prābhākaras hold in contrast that it is two cognitions,
a perception (accounting for the “this”) and a remembering (accounting
for the “that”). Buddhists argue that the “this” and the “that” indicate
distinct realities, and the apparently recognitional cognition, “This is that
Devadatta,” is strictly speaking false. These are the major positions. With
the stock example, we are to imagine a subject S encountering Devadatta in
the street and identifying him perceptually. Thus “This is that Devadatta”
has to be at least in part perceptual, everyone agrees. It’s the meaning of the
“that” that’s controversial.
The Nyāya view is that just as in the case of a simple perceptible charac-
teristic such as blue (in “It’s blue”) where memory is sparked to project the
mentally retained blue concept into the perception accounting for the full
concept-laden way the thing is perceived along with the possibility of ver-
balization, the same holds for the “thatness” in a recognition. On the ver-
bal level, ‘that’ is an anaphor standing for Devadatta as experienced earlier,
yesterday, let us presume. For Nyāya, Devadatta now currently perceived
40 Nyāya Epistemology
is qualified by thatness in fact. Thus although a recognition depends on
memory for the “thatness” cognitively, the thing recognized is the “that”
in fact. The connection with the qualifier may run through the mind of
S, but there is nothing in principle wrong with such a causal sequence.
Similarly, cognition of an absence, “There is no elephant in this room,” has
the elephant, the absentee, what is called the absence’s “countercorrelate,”
pratiyogin, furnished by the cognizer S. This is an objective truth nonethe-
less, given that there is indeed no elephant.
There are limits on the properties that may be sensorily evident. It is a
big task to specify these. But recognition is commonly considered not only
true but perceptual—it is a matter of vyavahāra, everyday discourse—and
Nyāya philosophers claim that they can account for its veridicality better
than rival views that do not respect the common pattern of speech. It is not
a stretch, they imply, to extend the projection power of memory to include
certain properties as past. In remembering, we are immediately aware that
an object is past, and something similar happens in recognition. An impor-
tant piece of background here is the position of the Bhāṭṭa Mīmāṃsaka, the
follower of Kumārila, on the property “cognizedness” as created in a thing
by the fact of its having been cognized. A second cognition on the part
of a subject S targeting the cognizedness, “The thing has been cognized
(previously by me),” proves the previous cognition’s existence. The Bhāṭṭa
position is taken to be that the cognizedness can be manifest perceptual-
ly.15 Nyāya does not much disagree, although the role of the mind is better
understood, it seems to me. Both schools hold that something perceived can
be known immediately as the same thing encountered in the past (“There’s
the pot for which I have been looking”). Perception thus establishes the
existence of things that endure. More precisely, this is the conclusion of an
inference, but perceptual recognition provides the crucial evidence, namely,
that the thing experienced in the past (and thus qualified by “thatness”) is
this thing here.
Now every Nyāya philosopher that I can think of, from Gautama on,
exploits the presumption of truth for the recognitional, “This is that Deva-
datta,” to infer the existence of an enduring self or cognizer, not just the
plainer fact that Devadatta has endured.16 Uddyotakara rules out rival
explanations with the following considerations: (a) no subject can remem-
ber what another has perceived; (b) no sensory property proper to one sense
organ (e.g., sight) can be identified with a property proper to another sense
organ (e.g., touch) except by way of the identity of a property-bearer, the
qualificandum (“This color is that touch” is disallowed although “This
blue thing is that which is in contact with my hand” is legitimate); and (c)
cognitions produced by distinct instruments cannot be identified (“This
visual cognition of the thing as blue is the same as that testimonial cogni-
tion of the thing as blue” is disallowed as is also a claim of identity between
a visual cognition and a tactile one although “My perception of a blue
thing has as its object—qua qualificandum—the same thing as that which
 Perception 41
Devadatta told me about” is legitimate).17 Thus we get the gist and see the
direction of the Nyāya realist reasoning. But let us desist from examining
further these, shall we say, metaphysical arguments, since they are complex
and have occupied whole monographs both in the classical literature and
modern scholarship.18

PERCEPTUAL ERROR (PSEUDO-PERCEPTION)

Nyāya takes a direct realist position about the objects of perception, and a
disjunctivist position about the ontological nature of illusion versus veridi-
cal perceptual experience. There are no sense-data and what we are aware of
perceptually are intersubjective objects with which we interact. To account
for illusion, which is a non-veridical cognition that seems to be perceptual
(pratyakṣa-ābhāsa, “semblance of perception”), Nyāya is faced with the
problem of explaining illusion’s intentionality without admitting objects
other than those with which we are familiar through veridical experience.
Like analytic disjunctivists who deny that illusion and veridical expe-
rience are the same type of mental state, Nyāya philosophers hold that
an illusion has different intentionality or objecthood, viṣayatā, than has a
genuine perception. “This is silver” cognized in the presence of tin is veridi-
cal in part, concerning the “this,” the perceived object in front. The “silver”
part of its intentionality is however not a matter of perception, not entirely
at least, picking out the universal silver as mediated by memory. Subjec-
tively, it is an overlaying on the presentation of the object at hand. Working
through the mind of the (mis)perceiver, the silverhood part of the cognition’s
intentionality—i.e., the qualifier portion of its objecthood, viṣayatā, what
it is about—hits the property as had by real pieces of silver. Universals such
as silverhood exist only in things, not in a subjective heaven. Thus a veridi-
cal cognition, “Here is a piece of silver,” and a non-veridical one prompting
the same verbalization, “Here is a piece of silver” (when the thing is really
tin), do not have the same content or intentionality—they do not belong
to the same ontological type (see the chart, p. 29), the one being veritable
perception, the other being “appearance of perception,” a common form
of error (viparyaya)—since in the one case the silverhood presented hits
the object that figures cognitively as qualificandum (the thing in S’s hand
is really qualified by silverhood) and in the other the intentionality is split
between the thing in front (the tin, the qualificandum) and something other
than the thing in front, the silverhood inhering in all silver (connected to
the mind of S non-physically, through an “extraordinary relation,” more
about which to follow), figuring as qualifier. For this reason, the Nyāya
theory is called the anyathā-khyāti theory of illusion, the “misplacement”
theory where something appears “other than what it is.”
Now Nyāya does not much consider hallucination (though a lot of
attention is paid to dreaming), where even the qualificandum portion of a
42 Nyāya Epistemology
cognition fails to hit an object (thus this would not be considered a pseu-
do-perception). However, the logic of the explanation can pretty easily be
extended. The material for psychological projection comes from veridi-
cal experience as preserved in memory. The unfamiliar is to be explained
through the familiar.19
Illusion has subjective and objective causes. Phenomenologically tin
looks a lot like silver, and the universals, tinhood and silverhood, have
some of the same “manifesters,” vyañjaka, “indicators,” such as shininess.
That this is an illusion is, moreover, something that is in principle discover-
able though it may be entirely convincing at the time. Errors can be found
out along with their causes, which are various but also usually attributable
to an abnormality in a doxastic process. 20 The processes that are knowl-
edge sources are the standard, 24-carat, the genuine issue which makes the
very concept of the semblance possible. Knowledge sources are natural pro-
cesses working normally, and thus their results fall into natural kinds, or
quasi-natural kinds, according to Nyāya theory, as we discussed in Chapter
2. Close imitators, illusions, incorrect inferences, false testimonial com-
prehensions, and so on, are different kinds of cognition than the veridical,
resulting from different processes in which in principle there is a fault or
flaw (doṣa) with respect to the true pramāṇa. Thus the results of pramāṇa
and of pramāṇa-ābhāsa—non-genuine, fake pramāṇa—are not the same
at all, although a wider uniting kind may be identified higher up in the
categorial system (in the classifiers’ terms, anubhavatva, “being a presen-
tational experience” or “cognition apparently presenting fresh news,” but
more commonly just jñānatā, “being a cognition”: it is cognitions, jñāna,
most generally that are false or true).
This is the short story. The longer story about illusion in Nyāya epis-
temology takes us fi rst into the context of disputes with Buddhists, both
with Nāgārjuna’s Mādhyamika school and with Yogācāra, in particular
Dignāga (fifth century) and Dharmakīrti (seventh century), and then to
Prābhākara Mīmāṃsā. The response to Mādhyamika comes at NyS 4.2.34
in the form of a parasitism argument: the very notion of false cognition
as in dreams and illusions presupposes that of true cognition. Vātsyāyana
makes the argument, and in his commentary on a subsequent sūtra, 4.2.37,
brings out the combinational character of illusion, with an object presented
(later called the qualificandum) as qualified by a qualifier that it does not
have in fact.21
Against Dignāga, Dharmakīrti, and followers, the Nyāya response
is much more complex. Uddyotakara seems principally occupied with
Dignāga throughout his long commentary on Vātsyāyana, and Vācaspati
says at the beginning of his prodigious work on Uddyotakara that to coun-
ter the arguments of these “bad reasoners” (ku-tarkika)—presumably
including Dharmakīrti—is what motivates him to write. War is waged on
several fronts. Dharmakīrti combines a Buddhist version of pragmatism
about concepts and truth with a verificationalist theory of justification. 22
 Perception 43
Relative to a subject’s “web of belief,” a sensory cognition is said to be
illusory when its indications fail to cohere with other bits of information
and further sensory data. Ropes don’t bite (unlike snakes), and sand (or
water in a mirage) does not quench thirst. Dharmakīrti defi nes a veridical
cognition in much the same terms as Vātsyāyana, as a cognition that leads
(normally) to successful action. So far so good, from the Nyāya perspec-
tive. But the Yogācārins advance a pragmatic theory of concept formation
according to which all of our beliefs turn out to be “convenient fictions”
(including Buddhist metaphysical positions as helpful for reaching the sum-
mum bonum). The very same thing that quenches an animal’s thirst is a lot
more to a human who uses it to cook and bathe, etc., whereas to a demon
it’s a river of pus. 23 An amorous young man may see a beautiful, nubile
young woman, but a hyena sees prey, and a sannyāsin (world-renunciant)
sees a corpse.24
The complaint on the part of Nyāya philosophers against Dharmakīrti
in this theatre is circularity, that successful action can only be determined
as successful by cognition that is veridical. Also, things to which we are
indifferent and towards which we do not act (e.g., pebbles in the road) can
still be veridically cognized. Thus the defi nition is too narrow.25 Finally,
there is the objection that desires can be satisfied even in dreams. 26
In New Nyāya, the Prābhākara displaces the Buddhist as the principal
adversary. The key issue becomes the unity of illusion as a cognition, as it
is with recognition. When we cognize “This is silver” given that the thing
we have picked up is tin, do we have a single perception-like cognition, or
two cognitions, perception providing us with the something that we pick
up while simultaneously we remember silver? The Prābhākara argues, as
noted, that all cognition is true. The function of cognition is to present the
truth. So-called illusion is, then, a failure to cognize, a failure to cognize
the difference, e.g., between a perceiving of tin and a remembering of silver.
The error lies in acting on this failure, saying, “There is silver,” when the
thing is tin.
Nyāya of course disagrees. 27 Vācaspati makes the point that the false
indication of an illusion does nevertheless inspire action as though there
were, e.g., silver in fact. We reach out to pick up silver, no matter that it is
tin. No mere remembering would so motivate us to act. This proves that the
silverhood qualifier is taken to be perceptual. Of course, the Prābhākara
need not disagree: it is the “stealing away” of awareness that the silverhood
is the object of a remembering (smṛti-pramoṣa) that constitutes the error,
a failure to be aware that this is a remembering. Nevertheless, Gaṅgeśa
parlays Vācaspati’s point into an inference taken to prove the Nyāya theory
against the Prābhākara.

An effort, directed toward mother-of-pearl, generated by a desire for

silver, is produced by a cognition of an entity as qualified, a cognition
whose object is that of the effort directed toward what is wanted, since
44 Nyāya Epistemology
it is an effort, like an effort directed toward mother-of-pearl on the
part of someone seeking mother-of-pearl.28

All effort directed towards something or other is guided by cognition of

the something as qualified by a qualifier, whether the guiding cognition be
veridical or non-veridical. This is an inductive generalization. Thus an illu-
sion is a single cognition of something as qualified by a qualifier, and not
two simultaneous cognitions.
A last argument comes from apperception, which also shows that the
illusion has a qualificandum as qualified by a qualifier as its object, as we
shall discuss a little later.

THE GENERALITY PROBLEM REVISITED:

TYPES OF SENSORY CONNECTION

The NyS defi nition of perception mentions sense-organ/object connection,

which becomes a general condition for certification of perceptual knowl-
edge of all sorts. But there are also specific conditions governing specific
ranges of the perceptible in that causal complexes responsible for percep-
tion of different kinds of things vary according to sensory modality and in
other ways. Nyāya looks at perception causally, and sufficient causality is
analyzed in terms of complex bundles of necessary factors. Although the
bundles vary, the object or objects indicated in a bit of occurrent percep-
tual knowledge are deemed not just real objects that are intersubjective,
etc., but themselves necessary factors in the perceptual process working
through a connection with a sense organ. A bull of a thesis this causal
realism, which Nyāya grabs by the horns, trying to wrestle the beast down
to the ground wherever it drags and pulls. Although empiricism has often
been tied to idealism in the West, Nyāya weds empiricism and realism: the
world exists independently of our cognition, and perception is not only the
fundamental cognitive link but, in a way, our only link, since all the other
knowledge sources crucially depend on perception in their operations but
not the reverse.
Just how the world can be known through perception, which parts of
it and precisely how, along with the limitations of sensory experience, are
part of Nyāya’s generality problem. For example, we cannot see atoms,
which are known instead by inference from perceptually warranted prem-
ises. However, a sensory connection to a pot entails connection to the
atoms that make it up. Thus there is a need to explain why the visual organ
is incapable of revealing atoms and other very small objects though it is
capable with objects that are large and medium-sized. And, as Vātsyāyana
mentions, we cannot see many medium-sized things at a distance (“You
cannot tell from here whether it is a post or a person”). The science of
 Perception 45
the sense organs—at least parts of it—is relevant to questions of certifi-
cation. Knowledge of things through their sound is also conditioned by
distance, as well as, in some cases, by a hearer’s consciously attending to
one sound stream, the voice of her own child, for instance. The theory of
the medium of sound, ether (ākāśa), including the substance’s relation to
the audial organ in the ear, is proposed to handle some of these questions,
as well as, in the case of vision, a requirement of light (which is not at all
well-understood). Attention, which usually works separately from physical
factors, is, in many cases of perception, an “auxiliary cause,” sahakārin,
a factor that is in the collection, sāmagrī, that is sufficient to bring a bit of
knowledge about. As mentioned, Nyāya, like Hume, sets no preconditions
about what sort of thing can stand as a causal factor for what other sort
of thing as effect (there are mental causes of physical events and physi-
cal causes of mental events and purely mental causes of mental events as
well as purely physical causes of physical events, for example).29 Voluntary
attention is not in all cases, however, a causal factor, as in the case of a
loud sound going off near one’s ear blocking out even the near-by voice of
one’s teacher.30 It bears repeating that Nyāya philosophers recognize that
depending upon the topic—that is to say, according to the type of thing and
property perceptually grasped—causal conditions vary.
I think it is for this reason that Nyāya philosophers for the most part
confi ne themselves to trying to articulate the most general and easily rec-
ognizable conditions governing perception in general and its most obvious
subtypes. They are interested not so much in the mechanics of perception
but in certification conditions, marks whereby we can differentiate veri-
table perception (which generates a true belief) from its semblance (which
generates falsehood). I dare say that the total of what they identify as a
causal factor for some bit or other of perceptual knowledge is woefully
inadequate considering the overall range of the perceptible that they assume
with questions of certification—that is to say, by the standards of their own
epistemology—to consider all that we purport to know perceptually. How-
ever, this is mainly due to the inadequacy of the supporting science, and it is
nevertheless important to acknowledge the direction of the Nyāya effort.
At the beginning of his commentary on Vātsyāyana, Uddyotakara spells
out six types of sensory connection taken by the later tradition to be uni-
form causal conditions covering specific types of perception.31 The theory
is filled out by speculation about environmental factors as well as subjective
factors, such as the focus and attention already mentioned. The operation
of an internal organ, manas, is posited in the NyS itself to account for
perception of emotions, desires, and cognitions (in apperception), as well
as for attention, in which case it is directed by the self to shut off informa-
tional inlets other than that which is by the self selected. The nature and
role of this the “mind” or “mental sense” are hotly debated from Gautama
on both within Nyāya and by critics who oppose the posit outright such as
Buddhists or offer alternative theories such as Mīmāṃsakas.
46 Nyāya Epistemology
Uddyotakara spells out his canonical six sensory connections through
appropriating the Vaiśeṣika ontology where there are six, later seven fun-
damental categories: substance, quality, action, universal, individualizer,
inherence, and absence.32 Within the category of substance, things made of
the atomic substances of earth—likewise water, fi re, and air—are known
differently from the non-atomic substance called “ether,” ākāśa, which is
posited as the medium of sound and known by inference. Moreover, we
can smell, taste, touch, see, and hear earthen things and many of their
properties, but other things composed of other sorts of substance and other
properties are not known through all five senses.
Absences and similarities are important types of relational property
known through a causal process crucially involving the mind of the subject.
S furnishes the idea of the absentee (S knows “There is no elephant in this
room” partly by recalling elephants) and, in the standard analysis of anal-
ogy, furnishes one term of a comparison grounded in something currently
perceived (“This which is called gavaya is similar to a cow”). The absen-
tee part of the cognition’s objecthood—the indication that it is elephants
that are absent—is furnished by the perceiver’s memory. Similarly, memory
informs a judgment of cow-similarity in a present gavaya buffalo by pro-
viding the idea of a cow (or cowhood). In both cases, there is a relational
adjunct or “counterpositive,” pratiyogin, supplied by a (saṃskāra) disposi-
tion of memory.
The paradigm sensory operation is nevertheless apparently taken from
touch, even in the case of the internal organ which somehow “contacts” the
self and, for example, its toothache, which is not really in the tooth but in
the self or consciousness as delimited by the body. The contact is not physi-
cal in this case, it is recognized. It is thought to be physical however in the
case of the visual organ, which is thought to extend out of the eye sockets
and to connect with the object seen at the object’s own location so long as
the object is also in contact with light.33 Visual experience is a property of
a self and we should say that there in the self is where it happens, as with
pain. But vision does not occur inside the head, and the object is seen where
it is, at whatever distance, if indeed it is really seen (visual perception is,
like all perception, factive, let us remember). The self is thought to be of
infi nite size and in the case of a person, which is a self delimited by a living
body, it is as though the self’s body extends to the thing presented visu-
ally. Of course, this is just an image, and mine, please note, not, so far as I
know, used by any classical Nyāya philosopher. I use it to bring out Nyāya’s
assimilation of the idea of the sensory connection operative in vision to that
of touch. We may note fi nally that many of the theories about the opera-
tions of the sense organs were held in common across school and show the
influence of the medical śāstra.34
A further classification of sensory connections groups Uddyotakara’s
six as “ordinary” (physical) as opposed to “extraordinary” (non-physical)
sensory connections, which, according to the Nyāya mainstream, come
 Perception 47
in three types: (a) connection mediated by a universal, as in cognition of
an unobserved instance of a universal (“The calf to be born will have a
dewlap”); (b) connection mediated by memory, as in cognition of sandal-
wood in the distance as having a certain smell, recognition of something
perceived previously, and perceptual illusion; and (c) connection mediated
by yogic power. Jayanta (tenth century) takes the classification for grant-
ed.35 Gaṅgeśa makes it explicit, and though he does not take the group-
ing as a topic for separate treatment, one type, the connection that works
through the universal, is posited by him to explain the possibility of infer-
ence—a position disputed by Maṇikaṇṭha (Gaṅgeśa’s near predecessor),
Raghunātha, and other late Nyāya philosophers, not to mention Buddhists
and those belonging to other schools. We shall take this up in the section on
pervasion in the next chapter. Acceptance of yogic perception is not uncom-
mon among the classical schools. Of particular importance is NyS 4.2.38
and commentaries where the self-knowledge of “yogic trance,” samādhi,
is conceived as a kind of yaugika-pratyakṣa.36 But although an epistemic
parallelism with ordinary perception seems universally endorsed by Nyāya
philosophers as well as by Praśastapāda, the great Vaiśeṣika, 37 few yogic
claims figure as Nyāya tenets. In any case, I have nothing to add here to an
exploration of the topic I have done elsewhere.38 We have already discussed
in effect (b), the connection mediated by memory, by taking up the example
of visual apprehension of the sourness of a lemon, etc., and illusion.
Much of this area of Nyāya theory may be scientifically quaint, but it is
not hard to fi nd its epistemological relevance, considering the Nyāya view
of certification. For, identification of causal factors helps us certify knowl-
edge. Nyāya does not require much science to sketch out its epistemological
theory, since some signs of the operations of the sources are not hard to
discern. Epistemic excellences (guṇa) and common flaws (doṣa) are both
causal factors and signs of knowledge sources, signs that we can readily
recognize (thus, e.g., not involving atoms although atoms are also causal
factors). Consonantly, pseudo-perceptual cognitions are to be explained as
failures to satisfy causal laws.
Thus Nyāya faces up to its generality problem about perception. The
special properties that facilitate certification of perceptual knowledge,
epistemic excellences and defects (guṇa and doṣa), are found to vary with
modality and object. And so conditions governing the operation of the
knowledge source in specific circumstances are delineated. Gaṅgeśa, for
example, devotes a whole section of his perception chapter to the precise
role of light in the case of visual perception. Expert perception is also rec-
ognized, as well as the role of familiarity (abhyāsa-daśā), to mention two
of the more defensible planks of the Nyāya enterprise.39
There is another side to Nyāya’s generality problem, concerning nega-
tive facts or absences. I know that my glasses are not on the table but how?
Dharmakīrti would answer, “By inference,” inferential knowledge of an
absence being one of three fundamental types identified by the Yogācārin:
48 Nyāya Epistemology
“If an elephant were in the room, I (S) would perceive it. I (S) do not per-
ceive an elephant. Therefore, there is no elephant in the room”—similarly
for my glasses not being on the table (presuming it is not so cluttered
that they could be concealed). Gautama and Vātsyāyana, without elabo-
rating, concur that absences are known inferentially.40 But Uddyotakara
and the later tradition argue that we know absences sometimes perceptu-
ally. I cognize immediately my glasses’ absence when I look for them on
the table. Bhāṭṭa Mīmāṃsā, however, says no, there is operative here a
special knowledge source called “non-cognition,” or “non-perception,”
anupalabdhi. The main arguments center on the sufficiency of perception,
or inference, to make known such negative facts, which intuitively we do
know. The Bhāṭṭa argues, for example, that perception makes known only
presences, presents positive characteristics, not absences of characteris-
tics. This seems a good phenomenological observation. Indeed, Nyāya has
a difficult time assimilating such knowledge to its theory of perception.
Its proposal is, to repeat, that an absence has a peculiar structure, relating
a locus (the table) to an absentee (my glasses) and that the absentee idea
is furnished by the cognizer from memory.41 One further problem is that
not only are there no elephants in the room and no Toyotas but also no
demons, piśāca, which are not ordinarily perceptible and so could not be
remembered.
Thus a generality problem opens for Nyāya both above and below, with
resistance to such specific conditions that would make for an additional
knowledge source as well as with the admission that subspecies of percep-
tion are governed by laws that are not universal for all perceptions. Nyāya
philosophers meet the difficulties head-on, however, with effort directed
both above and below, filling out large portions of their treatises.

APPERCEPTION

Finally, a few words about the special type of perception called appercep-
tion. Literally “after cognition,” anuvyavasāya, it is defi ned as a percep-
tion taking a preceding cognition or another psychological property as its
object, typically a scoping cognition in relation to another scoped. Efforts
and desires can be similarly targeted. Through apperception, it is said, one
knows infallibly the intentionality or objecthood, viṣayatā, of the cognition
scoped, in its qualificandum, qualifier, and relational portions, as explained
in Chapter 2. One knows, in other words, what the scoped cognition (or
effort) purports to be about (or directed to), and one knows this indepen-
dently of a determination of the target’s veridicality. Apperception makes
certification possible because it sets up the item to be certified. As discussed
in Chapter 2, since the sensory connection in this case is non-physical, not
mediated by any of the five material organs of perception, apperception is
 Perception 49
generally considered infallible about the contents of the cognition scoped,
but not of course about its veridicality.
Something else that can be known apperceptively is the type of cogni-
tion the scoped is. Such apperception does not embed a judgment that is
always right because Nyāya’s typology of cognition crucially distinguishes
the veridical and the non-veridical, as noted (see the chart, p. 29). Never-
theless, our apperceptive ability to determine whether the scoped cognition
is putatively perceptual, inferential, analogical, testimonial, or of another
character has epistemic relevance, since, as with all perception, there would
be a presumption in favor of the correctness of an identification. Thus there
would also be a restoration of a presumption of truth for a scoped cogni-
tion that had been brought into doubt were it judged to be perceptual,
inferential, analogical, or testimonial in character, though of course fur-
ther examination might prove the identification wrong. The upshot of all
this is, we may say, the Nyāya thesis that we can identify the origins of our
beliefs. By reflection, we can tell where a belief of ours comes from. In some
cases, this is all the justification we need, although in others it would be
only a fi rst step to further examination or argument.
Note that if further examination were called for, its direction would be
determined by the apperceptive identification of type. For, the criteria of
genuine perception, inference, and the rest vary according to origin. The
characteristic “excellences” and “flaws” (guṇa and doṣa) are very different
for each knowledge source and, as we have seen, there are also significant
differences within species of perception in particular. Thus apperceptive
judgment of putative type directs inquiry, telling us what to check.
Corollary to Nyāya’s thesis of the apperceptibility of cognitive types is
its view that illusion can be scoped such that there, too, we fi nd, abstractly,
a cognition of an object as qualified by a qualifier, for example, of a piece
of tin appearing to be silver. There is also the argument that the fact that
we can later see the same thing as really the tin it is shows that the illusion
is a single cognition and not, contra the Prābhākara, two cognitions, a per-
ception and a remembering along with a failure to notice their difference.
But scoping an illusory cognition is supposed to show this, too, and in a
particularly dramatic way when one knows that the cognition is illusory
while it continues, such as apparent sight of a double moon or a person
with hepatitis seeing a wall she knows is white as yellow. Although apper-
ception is said not in itself to be able to determine the truth of a cognition,
in such a case the subject practically sees one thing (the wall) appearing to
be something it is not (yellow), anyathā-khyāti.
We can attend to bits of our experience. By apperception directed by
attention, we can hone in on a single feature of something as presented by
a scoped cognition, not restricted to the entire phenomenology of a cur-
rent moment of consciousness. Anything that we attend to can be named,
according to Nyāya. And indeed to attend to anything is to do so by way
of a concept whether we know a thing’s name or not. The “contents,” so to
50 Nyāya Epistemology
say, of concept-free perception cannot be attended to, consequently. There
can be no apperception of nirvikalpaka, which, strictly speaking, does not
have intentionality, viṣayatā.42 All that we attend to shows the structure of
something (a qualificandum) appearing in some way (prakāra, the “way”
something appears, i.e., its qualifier). Earlier I said that indeterminate cog-
nition plays no epistemic role, because of—to frame the point now in ana-
lytic terms—the “doxastic assumption.” Only beliefs can justify beliefs.
Only determinate perception of something as some one way rather than
another can be either certified or certificational.
Something like indeterminate perception is posited by moderate foun-
dationalists in analytic epistemology as helping to solve the problem of
a justification regress, a mental state not embedding an assertion and so
immune to questions whether it is true or false. Laurence BonJour: “such a
state is, as it were, semi-assertive or semi-judgmental in character: it has a
kind of content or cognitive significance, but not in a way that would raise
a further issue of justification.”43 Such a comparison could be misleading,
however. Nyāya handles the justification regress differently from the foun-
dationalist who conceives in this way experience as a non-doxastic justi-
fier. Perception, according to Nyāya, is a doxastic justifier. The foundations
of our worldly beliefs are doxastic in that determinate perceptions embed
propositions that say at a minimum that something is some way or another.
Concept-free perception may be thought of as Nyāya’s way of rendering of
our ability to form perceptual concepts merely from perception’s phenom-
enological side. It is a theoretical posit made to explain how our concepts
originate in reality—a problem that is epitomized in cases of a fi rst-time
perception of something as an F, the concept F-hood being unavailable to
the perceiver. Indeterminate perception plays thus a metaphysical role in
the Nyāya theory, and is strictly speaking not part of epistemology. That it
cannot be apperceived is a way of making the point.
4 Inference

The Buddhists Dignāga (sixth century) and Dharmakīrti (seventh century)

had more impact on the development of the Nyāya theory of inference than,
I dare say, the NyS itself including the commentary by Vātsyāyana. This is
an overstatement, but there is a quantum leap in refi nement, and system-
atization, among Nyāya philosophers beginning with Uddyotakara, who,
though he criticizes Dignāga’s theory of inference, clearly learned much
from his adversary—as did his commentator Vācaspati from Dharmakīrti.
Claus Oetke says about “Ancient Indian Logic,” which he takes to include
the work of Gautama and Vātsyāyana, that it goes back to three roots: “(1)
common-sense inference, (2) establishment of doctrines in the frame of sci-
entific treatises (śāstra), and (3) justification of tenets in a debate.”1 Under
the “inference sūtra,” NyS 1.1.5, Vātsyāyana provides examples of everyday
reasoning. Several are abductive in character, to use the term popularized
by C. S. Peirce, informal reasoning to the best explanation, from sight of a
swollen river, for example, to the conclusion that it has rained upstream. 2
But instances of deductive, extrapolative, and sometimes properly induc-
tive reasoning on topics of everyday life as well as of philosophy are given
by the two early Nyāya philosophers. Surely it is not true, as is sometimes
claimed, that before Dignāga none had the notion of an inference-under-
pinning “pervasion,” vyāpti, between a prover property and a property to
be proved3 —a notion that later takes center stage both within Nyāya and
outside. However, Uddyotakara feels the constraints of system, let us say,
more than his predecessors, and brings out in their thought a theory meant
to capture how through inference we have knowledge of unobserved facts.
Inference is fi rst of all a natural process—Uddyotakara’s is a psychological
and causal theory—the imitation of which provides a schema for the under-
takings of philosophy and the establishment of tenets in scientific treatises
as well as formal debate. Such a synthesis seems to me to provide the proper
framework or overview for understanding the voluminous Nyāya literature
on inference, particularly the later texts.
There is an almost equally voluminous secondary literature on this
area of classical Indian thought—to include of course Buddhist logic, in
particular Yogācāra, which was often more innovative and in any case
52 Nyāya Epistemology
marched in tandem with Nyāya in developing an epistemology of infer-
ence and related theory of inquiry and debate. My own axe to grind here
is limited to the matter of overview, and I have not much to add in terms
of details.4 In other words, I shall emphasize the incorporation into epis-
temology, into pramāṇa-śāstra, of the topic of inference, where Oetke’s
second and third “roots,” the textbook arguments in favor of or against
various hypotheses and a theory of justification seem to me to become
fused in a theory of inference as a knowledge source. Actually, such inte-
gration is the view of the NyS itself, in chapter 1, as well as of Vātsyāyana.
Uddyotakara is less an innovator than a developer of the theory. In any
case, it is important to see the Nyāya theory of inference as a part of epis-
temology, as explaining how we know certain facts through the media-
tion of our knowledge of other facts. To mention just one advantage, the
subordination of logic to epistemology makes it easy to understand why
Nyāya counts a valid but unsound argument a fallacy: no knowledge is
generated. We have to keep in mind that Nyāya is focused not on logic
per se, but rather on a psychological process whereby we come to know
facts indirectly, by way of a sign, liṅga or hetu, an indication of something
currently beyond the range of the senses, whether at a distance spatially
or temporally or of a sort (such as atoms and God) that by nature cannot
be perceived.
Uddyotakara inherits a confused typology of inference, about which
he complains (shockingly, given his reverence for his Nyāya predeces-
sors). Inferences based on causal relations—of two kinds, reasoning from
effect to cause and from cause to effect—are distinguished from infer-
ences based on relations other than causality—thus three types in all.
Sweeping aside the messiness of a mismatch with some of Vātsyāyana’s
examples, Uddyotakara proceeds to lay out a new typology in which the
concept of a “pervasion” relationship, vyāpti, between a pervaded and a
pervader is central. Pervasion ensures that if the pervaded is present in a
locus or inferential subject then the other, the pervader, is also present.
This is similar but not the same as inclusion in set theory since pervasion
is a relation between properties, not property-bearers. Properties with
the same range, with the same bearers, such as knowability and name-
ability, are nevertheless different. In any case, remembering of examples
of co-presence and/or co-absence—or, in a formal demonstration, exem-
plification (udāharaṇa)—is required to know—or show—this inference-
underpinning relation between the two terms, the pervaded, that is, the
prover property, and the pervader, the probandum property. The per-
vaded (prover) is a sign of the pervader (probandum).
We shall scrutinize Uddyotakara’s three types of inference a little later.
First we need to appreciate their commonality: all three concern the war-
rant or justification for the second line of the following argument pattern
that in a simplified interpretation employs the rules of modus ponens and
universal instantiation to move from premises to conclusion:5
 Inference 53
1. Ha (the subject, a, the pakṣa, is qualified by H, the sādhana or hetu
“prover” property: “The mountain is smoky” or, more perspica-
ciously, “The mountain has smokiness”),
2. (x) (Hx → Sx) (where the H property, there the S, the sādhya “proban-
dum” property: “Whatever is smoky is fiery”). Therefore,
3. Sa (the subject, a, is qualified by S: “The mountain is fiery”).6

Deduction-wise, this is pretty much the whole story. However, inference

for Nyāya includes more than this single argument form. Specifically it
includes the epistemic grounds for the existential and universal premises.
To Western sensibilities, it seems as though there is more than one argu-
ment here in that whether a premise is justified is taken to be a separate
question. However, we should try to appreciate the Nyāya theory on its
own terms: the way that a pervasion is known is thought to be included in
the single operation of the knowledge source.
There is also the requirement that the inferential terms—subject, prover,
and probandum—be “well-known,” prasiddha, to a subject S if S is to
acquire inferential knowledge employing them. Gaṅgeśa draws out the
consequence—regarding the Nyāya effort to formulate an inference prov-
ing the self as a separate subcategory of substance—that to someone, such
as a Buddhist, for whom the property “enselved” is not an item of concep-
tion, the inference does not generate knowledge. It works only for those
who have certain background knowledge or beliefs. It is inference “for one-
self” (svārtha) but not “for another” (parārtha). This division of inferences
is inherited by later Nyāya philosophers from the Buddhists Dignāga and
Dharmakīrti and the Vaiśeṣika philosopher, Praśastapāda (sixth century,
after Dignāga, before Dharmakīrti). It is important for the distinction of
unreflective from certified bits of inferential knowledge.

INFERENCE FOR ONESELF AND INFERENCE FOR

ANOTHER (FORMAL DEMONSTRATION)

In his Nyāya-bindu, Dharmakīrti distinguishes two varieties of inference

to which he devotes separate chapters.7 The early commentator Dharmot-
tara (late eighth century) claims that inference in general cannot be defined
because the difference between the two is fundamental.8 Dharmottara
argues that the former is a matter of cognition (and thus of psychology)
whereas the latter is a matter of communication using words.9 Nyāya phi-
losophers take practically the opposite position, namely, that the inferences
we make in everyday life have the same structure as a formal inference for
others. Usually this is simply assumed.10 Clearly, inference for others gets
the most attention from the NyS on,11 and it is only in the New Nyāya period
that the distinction is at all explicitly drawn. Vātsyāyana indiscriminately
54 Nyāya Epistemology
takes his examples from philosophy and everyday life. Those presented in the
fifth book of the NyS, which is devoted to fallacies and fallacious objections
in a debate context (“Sound is not eternal, since it is a product, like a pot”
and so on), are formally conceived. Nevertheless, it is easy to see from the
examples adduced that Nyāya philosophers think we make inferences all the
time, particularly concerning something to be accomplished, as revealed in
our efforts (which are defined as cognitively guided). That I can make some
smoke now by making a fire is something I know by inference, as well as that
my interlocutors take the word ‘fire’ to mean fire and so on through innumer-
able bits of general knowledge informing speech and other acts.
The early Buddhist theory is superior to Nyāya’s in many respects,
including its talk of the selective power of desire to present candidate con-
comitances.12 These would appear to be acceptable until disproved (by
counterexample) as bases for extrapolation, given the requirements set
forth by Dignāga (see the beginning of the next section) according to which
there need be only a single known instance of co-presence of the prover and
probandum properties so long as there is no known instance of the prover
without the probandum. New Nyāya authors pick up the selection theme
in discussing the requirements for something to be a veritable inferential
subject, pakṣa, the subject property-bearer about which we would like to
know whether it possesses the probandum property.13
From the earliest it is said that a proper demonstration should have five
members (Vātsyāyana considers and rejects a ten-part form). Note these are
not steps in a proof.

1. That mountain is fiery. (The “proposition to be proved,” pratijñā, that

is, that the pakṣa, the inferential subject, is qualified by the sādhya,
the probandum.)
2. That mountain is smoky. (The “reason,” hetu: the inferential subject
is qualified by the prover, sādhana. This premise is taken in the stock
example to be a bit of perceptual knowledge, but it could be testimo-
nial or even inferential instead.)
3. Whatever is smoky is fiery. (Or, “Wherever smoke, there fi re.” The
udāharaṇa, the “exemplification” or statement of the general rule,
vyāpti, “pervasion”: where the prover, there the probandum. The
prover is “pervaded by” the probandum. This premise is generally—
though not by Maṇikaṇṭha—taken to be warranted by wide experi-
ence, bhūyo-darśana, of positive correlations and/or negative, such as
in a kitchen hearth, dṛṣṭānta, the example.)
4. That smoky mountain falls under the “whatever” of the general rule,
“Whatever . . . ” (The upanaya, “application”: The mountain, or
pakṣa, falls within the scope of the universal quantifier, the vyāpti.)
5. That mountain is fiery. (The nigamana, “conclusion”: same as step
one, except now proved.)14
 Inference 55
The five steps are to be construed as a single statement governed by gram-
matical and semantic rules, designed to provoke inferential knowledge
in another. The idea—from Gautama through Navya Nyāya—is that the
five-membered form is an ideal ordering corresponding to requirements
of syntactic binding and semantic “fittingness” (ākāṅkṣā and yogyatā)—
crucial to testimonial knowledge (see Chapter 6)—for provoking inferen-
tial knowledge from a statement that provides the inferential terms in the
proper relationship.15
Natural, unreflective inferences are sometimes cited to illustrate a subject’s
comprehension of the information in one or another of the five “members,”
avayava. But a demonstration as meant for others does not precisely mirror
the psychological process of unreflective inference for oneself. Nevertheless,
both types of inference have to meet the following three conditions, accord-
ing to Uddyotakara and everyone afterwards.16 Let me repeat the main point:
the two kinds of inference are thought of in Nyāya as essentially the same
in principle. There is an epistemic difference—which is sometimes crucial,
determining whether although one oneself knows that p the other knows
that p, too—concerning the availability of the probandum term in an infer-
ence for others. We shall talk more about this the cash value of the distinc-
tion later (in the last section of the chapter). First let us appreciate the three
conditions, which serve as certification conditions and must be met in fact by
both kinds (though it is only at the reflective level that they are identified). In
parentheses I present the usual explanation targeting an inference for oneself,
but it is easy to see how it would go with a demonstration.

(1) pakṣa-dharmatā (The prover has to be known as qualifying the infer-

ential subject or site.)
(2) vyāpti-smaraṇa (The prover as pervaded by the probandum has to be
remembered.)
(3) liṅga-parāmarśa (The subject must connect by reflection the perva-
sion with the subject at hand.)

Now proper reconstruction of the full inferential process that these conditions
address would have, at a minimum, each of three lines (two premises and the
conclusion) prefixed by the epistemic operator ‘(K)’, to be read, “A subject S
knows that. . . . ” To use again a simple representation of pervasion (not sensi-
tive to the precise logic of occurrence/non-occurrence at a location):

1. (K) Ha
2. (K) (x) (Hx → Sx)
3. (K) Sa, provided that parāmarśa has occurred.

I tack the third condition on at the end, since it is no separate operation.

“Inferential reflection,” parāmarśa, is a matter of putting the information
56 Nyāya Epistemology
together from the two premises such that, psychologically speaking, the
inferential conclusion becomes the content of a bit of occurrent knowledge.
The nature of this “reflection” is a hot topic, both within Nyāya and without.
Uddyotakara makes it the central pillar of his theory, and Gaṅgeśa devotes
a long section of his inference chapter to arguing that it is required (contra
Mīmāṃsakas in particular). The upshot is, it seems to me, that knowledge is
not closed under deduction considered in abstraction from the psychological
process of “reflection,” parāmarśa. But through that process, epistemic war-
rant—or “certainty” (niścayatā)—passes from premises to conclusion, and
we act unhesitatingly, for example, to put a fire on yonder mountain out.
Things are yet more complicated. Inferential knowledge is defeasible, or,
more precisely stated, what a subject takes to be inferential knowledge may
turn out to be pseudo, non-genuine, a false cognition imitating a true one,
or even in Gettier-style cases an accidentally true cognition masquerading
as one genuinely inference-born. As mentioned in Chapter 1, knowledge
has a social dimension. Not only would awareness of a counterexample
be a defeater, but also if someone were to present a counterinference to a
conclusion opposed to ours, no longer would we have inferential knowl-
edge. Awareness of any of several kinds of “blocker” of “reflection” can
undermine the generalization on which such reflection depends. There are
potential preventers of inferential awareness, “defeaters,” bādhaka, leading
to belief relinquishment by someone who has hitherto not noticed a coun-
terexample or the like and who has thus drawn a conclusion erroneously.17
Nyāya has an answer to the Cārvāka skepticism that no amount of “wide
experience” could secure knowledge of pervasion since that would include
all things smoky and fiery, for example, past, present, and future.18 The
answer comes in the form of the suppositional reasoning that, for example,
if yonder smoky mountain were not fiery, something’s being smoky would
be a mere accident. Is this alternative plausible? No. Admittedly, there is
something right about the complaint, which is that we are fallible in our
generalizations. But that does not mean that we do not know inference-
underpinning pervasions when we do. What it means is that we have to
have an eye out for counterexamples. As the Advaitin Śrīharṣa (eleventh
century) points out, everything earthen that we have encountered, say more
than a thousand specimens, may have been scratchable by iron. But when
we encounter for the fi rst time a diamond, which is not scratchable by iron,
we relinquish our belief in the concomitance and no longer is earthenness
a sign for us of iron-scatchability.19 The second part of the Nyāya response
is that our action shows that sometimes we do know pervasions. We act
unhesitatingly showing our confidence in inferential knowledge.

FROM EXTRAPOLATION TO GENERALIZATION

Jonardon Ganeri argues that an extrapolative pattern including causal rea-

soning as well as sampling (presuming uniformity in nature) is present in
 Inference 57
some of Vātsyāyana’s examples as well as in cases taken from early Bud-
dhist literature and, indeed, the formalization of the marks of a “(good)
reason” by Dignāga. 20 To take up a Buddhist example, from a drop of ocean
water tasting salty a subject extrapolates to the saltiness of any other bit of
ocean water. The reasoning is from particulars to particulars that are rel-
evantly alike. The dark cloud overhead—in one of Vātsyāyana’s examples
of knowledge by inference—is relevantly similar to the one seen yesterday
that brought rain but not relevantly similar to the white cloud seen the day
before that was rainless. So it is concluded that it will rain. 21
Dignāga, who as a nominalist would want inference to work from
knowledge of particulars to other knowledge of particulars, formulates a
threefold test for a good prover, trairūpya-hetu, which should be compared
to Uddyotakara’s three certification conditions (see p. 55):

(i) that the prover’s occurrence in the site or subject of a proposed infer-
ence be known to a subject S (the same as Uddyotakara’s pakṣa-
dharmatā, condition 1)
(ii) that the prover’s occurrence at least once together with the probandum
be known to S (a version of Uddyotakara’s vyāpti-jñāna, condition 2)
(iii) that no counter-case of a prover occurring without the probandum be
known to S (also corresponding to Uddyotakara’s condition 2).

Ganeri claims that it is better to understand both the Buddhist and Nyāya
theories as “not enthymematic,” not skipping a step of generalization and
then implicitly using universal instantiation in applying the rule to the case
at hand. Case-based reasoning need not be interpreted as relying on uni-
versal quantifiers, and the representation of Schayer and others which uses
them is misleading.22 Theirs is indeed misleading, and Ganeri appears to
be right with regard to the Yogācāra theory. But with late Nyāya Schay-
er’s argument form of UI and MP misleads not for this reason but rather
because it fails to be sufficiently sensitive to the logic of occurrence and
non-occurrence of properties at a location, or qualifying a property-bearer,
as Ganeri among others has brought out. 23 Furthermore, Ganeri is right
that in analyzing the pattern one tends to miss the unity of the causal the-
ory that has one mental state caused by another. To be sure, the unity of
the Nyāya view must be appreciated, everything integrated in the notion of
“reflection,” parāmarśa, as an inference’s “proximate instrumental cause”
or “trigger,” karaṇa. While not the only necessary condition, this is the last
one in place, securing the occurrence of inferential knowledge (anumiti).
Following Matilal, I have framed Gaṅgeśa’s understanding of such
“reflection” as a singular inference. 24 Thus:

(K)(SpHa) → (K)Sa

This says that on the condition that a subject knows that H-as-qualified-
by-being-pervaded-by-S qualifies a, then the subject knows that Sa. 25 The
58 Nyāya Epistemology
arrow should be interpreted as depicting causal sufficiency, following
Uddyotakara and the later tradition. “Reflection” is a complex mental state
that is nevertheless a unity, both as a particular cognition that would be
the decisive factor for the rise of another cognition and as having content,
or “objecthood,” expressible in a single sentence. Ganeri’s attempt to fi nd a
single rule is in consonance with both of these dimensions of the Nyāya the-
ory. But at least in the paradigm case a lot of inductive depth is packed into
the idea of a pervasion being known, and a lot about it is said that shows
that there is generalization according to the Nyāya understanding of infer-
ence as a knowledge source. Maṇikaṇṭha Miśra (thirteenth century) and
other New Naiyāyikas help us to see that grasping the vyāpti relationship
that underpins inference can be other than through collecting instances,
being sometimes immediate, sometimes aided only by suppositional rea-
soning, tarka, as the case may be, that is to say, according to the nature of
the pervader and pervading characteristics. 26 Nevertheless, inductive gen-
eralization stands out as the gold standard for grasping a pervasion relation
permitting inference.
So in the paradigm case, a person P who has inferential knowledge has
discerned a pervasion relationship from a finite set of correlations by P
known (from perception or testimony or even another inference): Hb and
Sb, Hc and Sc, Hd and Sd, . . . (H = prover and S = probandum). P has in
addition to these positive correlations no knowledge of a counterexample
(something H but not S), though, let us say for simplicity’s sake, P does
know of some things that are S but not H (being-an-H and being-an-S are
not co-extensive). Furthermore P has the following negative evidence in a
standard case: ~Se and ~He, ~Sf and ~Hf, ~Sg and ~Hg,. . . . On the basis
of all this, P has generalized, being-an-H is pervaded by being-an-S, and
stores the knowledge in memory. P then comes to know by perception (or
by testimony, etc.) that something is, well, not simply an H (Ha) but an
H-qualified-by-S-pervasion, through the projective power of memory, all of
which we shall abbreviate as ‘H*’. H*a is not a prima facie but a conclusive
reason to believe Sa. Psychologically considered, the reflection that draws
from this information the conclusion Sa is the penultimate cognitive occur-
rence before anumiti, before the inferential knowledge occurs.
This new knowledge is certain but defeasible and uncertified, at least not
yet, as has been explained. 27 The defeasible character of reasoning means
that we have to see the “prover” not as an example we arrive at by extrapo-
lation, but as H*a, as a’s being H as known to be correlated with S through
inductive evidence (as noted earlier). Thus we could be wrong, and as phi-
losophers we know it such that we are sensitive not only to counterexamples
but to the whiff of counterexamples, known in the literature as the upādhi.
An upādhi (U) is defi ned as a property meeting two conditions:

(1) U pervades the probandum S: (x) (Sx → Ux).

(2) U fails to pervade the prover H: (∃x) (Hx . ~Ux).
 Inference 59
The consequence of something’s being an S while there is known to be
something H that is not a U is that an original inference in terms of H and
S is defeated. An upādhi condition entails that there is a counterexample.
Now while the epistemological consequences of a known counterexample
may be obvious, it is not so obvious what Gaṅgeśa claims, to wit, that even
a dubious upādhi undercuts, i.e., blocks and defeats, inference. 28
For example, if P had knowledge of the correlations, Sh and Uh, Si and
Ui, Sj and Uj, . . . along with no knowledge of a counterexample (something
S but not U), and, let us say, P does know of some things that they are U
but not S (being-a-U and being-an-S are not co-extensive, though, Gaṅgeśa
explicitly states, they could be co-extensive and there be an upādhi none-
theless). We may also imagine that P has negative evidence correlating, with
no counterexamples, things ~S and ~U. On the basis of all this, P general-
izes: being-an-S is pervaded by being-a-U: (x) (Sx → Ux). Furthermore, if P
knows, or simply has the suspicion, Gaṅgeśa maintains, that something is
an H but not a U—(∃x) (Hx . ∼Ux)—then P’s awareness of these two con-
ditions together—the U’s pervading the S along with the U’s not pervading
the H—would block and defeat P’s warrant for the otherwise inferentially
based belief (Sa). This shows not only that inductive evidence is taken by
Nyāya to be intrinsic to inference as a knowledge source, but also that
Nyāya philosophers recognize the epistemic relevance of undercutting pos-
sibilities. To know a pervasion as required for inference one has to know
that the relationship between the prover and the probandum is upādhi-free,
according to many Nyāya authors. 29
One of Gaṅgeśa’s examples of a dubious upādhi that undercuts an
inference nonetheless is digestion-of-certain-vegetables, an upādhi for an
attempt to infer that Mitrā’s next child will be dark-complexioned on the
grounds that her other children are dark-complexioned. The inference-
permitting pervasion has to be known, and if doubt is legitimately cast
on that relationship, inferential knowledge is blocked. A dubious upādhi
is counted an inference-blocker, and this despite the pervasion relation-
ship required of the upādhi itself not being certain or the upādhi’s failing
to pervade the prover not being certain or both (see the defi nition of an
upādhi given earlier).
With regard to the Mitrā example, we do not know what Mitrā has
eaten during her current pregnancy. She may or may not have eaten the veg-
etables, and they may or not be responsible for the dark complexion of her
children. Furthermore, it could turn out that not only has she always eaten
the vegetables but will always do so such that the upādhi would not fail to
pervade the prover. Even her future children would be born to a mother
who has eaten the vegetables such that the prover, being-a-child-of-Mitrā’s,
could predict the probandum, being-dark-complexioned, not for a genetic
reason but rather because of the vegetable habit, while the existential condi-
tion for a genuine upādhi would not be met. This, then, is a reason why the
upādhi is dubious. It still undercuts. There is a good chance that although
60 Nyāya Epistemology
Mitrā has eaten the vegetables in the past she has not this time, such that
the upādhi’s failing to pervade the prover is a real possibility.
Similar remarks are to be made about the requirement of the upādhi’s
pervading the probandum. So long as there is evidence for this, the upādhi
defeats the putative inference. The pervasion could be nevertheless dubious
for one or both of two reasons: (i) we do not know whether Mitrā has eaten
the vegetables during each of her previous pregnancies (though there is a
good possibility that she has) and/or (ii) there is a richer complex of condi-
tions, such as Maitra being the father, such that if they were known the
upādhi’s pervading the probandum would be a discarded thesis. And there
is a good chance that there is such pervasion: Mitrā has in fact eaten the veg-
etables during all her previous pregnancies (though not this time), and even if
that is a necessary condition for her children’s being dark-complexioned only
under the further condition of Maitra being the father, or is not a necessary
condition at all, the worry about the vegetable diet (backed up by medical
authority, which Gaṅgeśa also mentions while regarding it as in itself not at
all conclusive) is sufficient to block the original inference. We do not know,
therefore, that her child will be, like her other children, dark-complexioned.
The standards of warrant are thus lower for knowledge of the pervasion
required for an upādhi than with knowledge of the pervasion required for an
original inference in focus. And this seems correct, since the point of infer-
ence is to pass on warrant from premises to conclusion by way of a pervasion
cognized with a degree of justification that does not diminish a person’s over-
all level of confidence and would normally lead to unhesitating effort to act.
Pace Ganeri, knowing a general rule is considered crucial, not just
extrapolation to a next case. From Uddyotakara on, Nyāya philosophers
treat pervasion as the equivalent of a rule stating that—to use the language
of sets and terms—the extension of the probandum term includes that of
the prover term, includes it entirely such that there is nothing that locates
the pervaded property (the prover) that does not also locate the pervader
(the probandum).30 Any new H has then to be an S, because of the rule or
inclusion. One could argue that the centralmost issue and the main line of
effort of the New Nyāya philosophers is to make plain the logic of perva-
sion as well as how we know the universalized items, or entire extensions,
of the terms figuring in our knowledge of such rules, the items that under-
pin our knowledge of such inclusions, such naturally necessary pervasions
of a prover by a probandum property. Here again they learn much from
their Buddhist adversaries, although again they hardly follow them.

THE ONTOLOGY OF PERVASION

New Nyāya philosophers work with an ontology inherited as much from

Vaiśeṣika sources as from Nyāya. Fellow realist Mīmāṃsakas present
 Inference 61
variations and a somewhat richer scheme of ontological categories than
the Vaiśeṣikas. Buddhists reject this work almost in its entirety, and some
Vedāntins see it as representing only the structure of appearance, māyā, the
cosmic illusion, real so far as our everyday concerns go and normal con-
sciousness, not real in a higher or religious sense. Nyāya philosophers them-
selves often express skepticism about one or another of the items on the
Vaiśeṣika list and to various inherited ontological principles. As mentioned,
the New Nyāya philosopher Raghunātha proposes radical changes. Never-
theless, the Vaiśeṣika ontology of substances, qualities, motions, univer-
sals, individualizers, inherence (the ontic glue of the system), and absence
is utilized in various contexts including the ontology of pervasion. New
Nyāya philosophers crystallize a simplification they fi nd in their predeces-
sors’ writings on word meaning (an individual as qualified by a universal)
and the anyathā-khyāti, “misplacement” theory of illusion in particular.31
Cognitions have a certain structure determined by what they are about,
their “objecthood,” viṣayatā, namely, in the simplest form, a qualifican-
dum (a) cognized as qualified by a qualifier (F): Fa. Qualificanda are the
subjects of inferences, the sites whose possession of a probandum/quali-
fier inferences make known, albeit sites are known by way of a qualifying
property, e.g., mountainhood, with the stock example of inference to fi re
on yonder mountain. The property that specifies the subject is epistemically
relevant in certain contexts (e.g., we know that lakes are never fiery no
matter that we see what looks like smoke but is really morning mist). But
this is irrelevant to the present point which is that the traditional scheme
of categories is subsumed into a threefold division of properties (qualifiers),
property-bearers (qualificanda), and qualificative relations (usually simply
occurrence, property-possession, or qualification). Typically a qualifican-
dum and site of an inference will be a substance such as the mountain of the
stock example, which is a material thing composed of earthen atoms. Typi-
cally, a quality, such as blue, a motion, or a universal (such as smokiness or
fieriness) would fi ll each qualifier spot (prover and probandum), the prover
property/qualifier corresponding to a perceptual adjective or predicate in
an observation statement. But particulars can be qualifiers, as when the
floor has a pot on it (and we say, perspicaciously, “The floor is qualified by
a pot,” which would be by the relation of contact or conjunction, saṃyoga).
And a universal or another abstract character can stand in the qualifican-
dum place, as happens in particular in abstruse philosophical arguments.32
So the new ontology may be said to nudge the other out, at least in the
context of the epistemology of inference, although the former informs it at
crucial places, as we shall see.
Probably the most important innovation taken over from Vaiśeṣika is the
concept of absence, abhāva.33 All properties either occur or do not occur at
a specified site, either qualify or fail to qualify a particular property-bearer.
An absence of a property F is another property, F-absence (~F). There is the
following rule capturing part of the logic or general nature of an absence:
62 Nyāya Epistemology
For any qualificandum x, if (a property) F does not qualify x, then ~F
does qualify x (~Fa).
Non-locatable properties are fictions, composites constructed by the
mind out of concepts created by veridical presentational experience. There
are no absences of fictions. There is no absence of a rabbit’s horn on the
table, although there is an absence of horns on a rabbit. “Absence of a rab-
bit’s horn” is ill-conceived, not really a property, the notion standing in
tension with a second rule about absences. This is that the absentee, the
pratiyogin, “countercorrelate,” be a locatable property. “Rabbit-horned”
does not occur. An inference cannot trade on an absence whose countercor-
relate does not occur.
In contrast with this narrowing of the scope of inference by an onto-
logical consideration, another expands it. Universal properties, such as
knowability, qualify every qualificandum in the universe. Thus the nega-
tion of one of these is non-locatable and thus not a true property. We may
infer knowability nevertheless, although in principle there can be no nega-
tive correlations through which to know the pervasion.34
With these ideas in mind, let us turn now to Uddyotakara’s three types
of inference, which are divided according to how a pervasion is known
(vyāpti-graha). Note that from Uddyotakara on, this issue is separated out
for special attention and analyzed prolifically. Suppositional reason (tarka)
is enlisted to the cause. But let us stay focused on the inductive method.
Of the three certification conditions, we are concerned with the second,
vyāpti-smaraṇa, “memory of pervasion.” There are three ways a pervasion
can be known and thus remembered, depending on positive and/or negative
correlations:

(i) inferences based on positive and negative correlations (anvaya-

vyatireka), where both are available, i.e., cases where, for example,
smokiness and fieriness have been known to occur together, kitchen
hearths, campfi res, etc., like (it is claimed) yonder smoky mountain
where being-fiery is to be inferred, taken along with negative examples
where the prover as well as the probandum is known not to occur
(ii) inferences based on positive correlations only (kevala-anvaya), where
there are no known examples of an absence of the probandum prop-
erty, such as would have to be the case with the universally present
property, knowability (there is nothing that is not knowable)
(iii) inferences based on negative correlations only (kevala-vyatireka)
where outside of the subject or site there are no known cases of the
probandum.

As a naturally occurring relation, a pervasion is the same thing no matter

which of these ways it is known. Even negatively formulated as a perva-
sion of the absence-of-the-probandum by the absence-of-the-prover, it is
the same fact.35
 Inference 63
Knowledge of pervasion through inductive generalization is such that nor-
mally examining the vipakṣa (where S is known not to occur) would be rel-
evant: all cases of not-S have to be also not-H. However, with a universally
present property such as knowability, there is no vipakṣa. And so an infer-
ence to something’s knowability cannot in principle be supported by negative
correlations. It would be “positive only,” and “positive only” in principle.
A slightly different story has to be told with the “negative-only,” which
on the face of it seems invalid. Buddhists do not recognize this third type or
categorize what Nyāya philosophers see as its instances as inferences based
upon knowledge of a partial or entire co-referentiality of prover and proban-
dum terms (which comes to be called “internal pervasion,” antar-vyāpti).
Furthermore, Mīmāṃsakas postulate here a distinct source of knowledge
called arthāpatti, sometimes translated “circumstantial implication.” In
any case, the logic hinges on the notion of a genuine pakṣa, inferential
“site” or “subject,” and what can be legitimately assumed in an inferential
context. This is the argument form for several of Nyāya’s philosophical
inferences, and I propose to expand on this third type in the next section in
looking closely at an inference for the existence of self, ātman. First, let us
tie up a couple of loose ends concerning the ontology of pervasion.
Knowing a pervasion entails knowing that a future H-bearer will also be
an S-bearer. Udayana proposes—followed in this by Gaṅgeśa and other but
not all New Nyāya philosophers—that some universals are known percep-
tually. They dwell timelessly and holistically in their instances (cowhood
pervades every part of a living cow), past, present, and future. Thus percep-
tion of, e.g., cowhood in Bessie entails a kind of sensory connection with
cows still to be born. The connection is “extraordinary,” one of three kinds
of extraordinary connection as opposed to the ordinary six, (question-
beggingly) termed sāmānya-lakṣaṇā-pratyāsatti, “relationship that works
through a universal.” This is a type of perception that is very helpful for
inference, providing generalized properties at the perceptual level.
We could say: Well, there is then no role for inductive generalization.
Yes, there is. There is always the possibility we are faced not with a genuine
universal but rather a mind-constructed accident (the metaphysical upādhi),
contoured in an irrelevant way, and so inductive evidence is often relevant.
Induction is the way many universals are certified as genuine although we
grasp the entirety of cowhood in a sense from a single perception of Bessie.
Furthermore, some universals are hidden and known only by indicators,
“manifesters,” vyañjaka, and inference. They are not directly perceived, for
example, atomhood. Still, proper inductive procedure is a certification con-
dition in practically all cases, since universals have temporal spread. Nev-
ertheless, when we perceive one embedded in a locus we get in a sense the
entire thing. Future cows do not change the nature of cowhood. To be sure,
pervasions relate thick particulars, not bare qualificanda, individuals as
qualified by universals (and other characters that are universal-like, “quasi-
universals,” in particular the akhaṇḍa-upādhi, the unanalyzable accident,
64 Nyāya Epistemology
e.g., being-a-quality, as opposed to the analyzable, being-Devadatta’s-
cow36). But this is because universals, etc., inhere in their loci, and do not
exist in the mind. Knowledge of a pervasion does include all instances in
their being instances of the universals it trades on.
Thus on Udayana’s view of perceptible universals there is an answer to
the question of how sometimes we know a pervasion in its full generality—
everything H as an S. It is that some veritable universals are perceptible
characteristics, cowhood, for example, in a cow. Not all universals are hid-
den. Just as we can know Devadatta, who is standing in the door percep-
tually without perceiving his whole body—and all the characteristics that
make him up as a person, as a “thick particular,” a qualificandum together
with its qualifiers—so we can know the universal cowhood in Bessie with-
out perceiving all cows.
Naturally the classical Indian controversy over universals does not end
(nor begin) with this theory, which we shall now table to take up one last
broad area of classical discussion of the ontology of pervasion, causality in
a word. Here let us again go back to the Buddhists and still another typol-
ogy of inference. Dharmakīrti divides inferences into three types based on
different types of pervasion in nature:

(1) sva-bhāva (self-nature: “It is a tree because it is a śiṃśapā”)

(2) tad-utpatti (causality: “Fire is there because smoke is there”)
(3) anupalabdhi (non-perception: “There is no pot here because none is
perceived here”).

Nyāya eliminates two of the three as otherwise explainable. The first, a kind
of a priori knowledge, it would seem, is not counted as inferential by Nyāya
philosophers who dismiss it as mere memory of the meaning of words or
phrases. Or, in certain instances, inferences can be reformulated as “nega-
tive-only” (see note 41 to this chapter for a stock example taken from Bud-
dhist literature). Knowledge of synonymy, whether in whole or in part, is not
really pramā, “knowledge” of the sort we have been discussing, since it is
not “presentational experience,” anubhava, not cognition presenting fresh
news, but rather memory. Or, we should say that since testimony is the main
way we learn the meaning of words (though there are others, for example,
analogy), knowledge of synonymy would typically be a form of testimonial
knowledge. It may well be a shortcoming, but the truth is that the Nyāya
school does not much interest itself in the a priori which is in this way red-
lined (a lack we’ll revisit in Chapter 7). Dharmakīrti’s third kind of inference,
absential knowledge, Nyāya views as a kind of perceptual knowledge, as we
discussed. That leaves causality. And Nyāya philosophers do try to interpret
the pervasions on which inferences hinge as causal relations, generally speak-
ing. They often flounder, it seems to me, in the attempt, but it is important
to recognize that in this way they see pervasions as naturally necessitated.
Without the pervader, the pervaded could not have come to be.
 Inference 65
There are three main types of causal relationship, the instrumental (and
any non-eternal object or event will typically have several instrumental
causes, the most important of which is the “trigger,” karaṇa), the inher-
ent, and the emergent.37 Often we fi nd tedious effort using these notions
to integrate the power of inference and the nature of the facts inferred. But
with one further remark tying our two loose ends together, universals and
causality, let us move on to look at some important philosophical infer-
ences. Universals are relevant to determining causal laws. It is not a match
as a substance—though a match is indeed a substance—that is a cause of
fi re but rather as a wooden stick with a tip of phosphorus. 38 Nyāya philoso-
phers tend to think of natural laws in terms of relations among universals
(and quasi-universals). But again let us put aside the metaphysical issue to
see what we can learn from some famous Nyāya proofs.39

PHILOSOPHICAL PROOFS OF SELF, GOD,

AND MUKTI, “LIBERATION”

Presentation of arguments that meet the requirements of the theory of infer-

ence as a knowledge source are not the only way to establish, or at least
to defend, positions in philosophy as in everyday life. There is also tarka,
suppositional reasoning, which we reviewed in Chapter 2. Nevertheless,
there are some important arguments that proceed according to the stan-
dard inferential form. Let us look fi rst at an inference to self that proceeds
by “negative correlations only,” kevala-vyatireka.40
As mentioned, negative-only inference was highly controversial among
the classical schools. Two co-extensional terms present an epistemological
puzzle when the one is used to pick out the inferential subject (pakṣa, for
example, “being earthen,” or, in a second example, “every living body”)
and the other the prover (hetu, “having smell” or “has breath”). The infer-
ential subject cannot itself be used as inductive support for the required
knowledge of pervasion (“Whatever has smell is a distinct substance” or
“Whatever has breath is enselved”), since otherwise inference would be
pointless. A putative prover’s extensional equivalence to the subject (and
indeed everything that is earthen is thought to have smell and only those
things that are earthen; similarly with the other inference) would mean that
there could be no positive support, no correlations of a positive variety of
the S (sādhya, “being a distinct substance”; “being enselved”) and the H.
So an inference based on a pervasion under these circumstances would have
to be “negative-only.”41
However, the form seems invalid without restriction, since it would
prove too much. Given merely that a is H, with no known H outside the
pakṣa (a as specified by a property such as mountainhood), it would appear
from the correlation of ~S and ~H that we could prove of a any property
66 Nyāya Epistemology
known not to reside outside the pakṣa. For example, “Martian-made” (S)
could be proved of every cow (a) by the prover cowhood (H). Everywhere
we fi nd something that is not a cow, there we find something that is not
Martian-made, e.g., a rock. Thus, since Bessie has cowhood, Bessie is Mar-
tian-made.42 It is important historically that my example would be ruled
a non-genuine inference through its having an “unfamiliar” probandum
(aprasiddha). And it would also fall to the counterinference, sat-pratipakṣa,
“Every cow is non-Martian-made, since it is a cow, unlike a rock.” The
problem of the form’s being too powerful is apparent nonetheless.
Furthermore, negative correlations would seem to be not strictly rel-
evant, at least in some cases, as is brought out by the “ravens paradox”
well-known to students of inductive logic.43 An evidence base supporting
the generalization, “All ravens are black,” benefits from the sight of an
additional black raven. But to flip the pages of a book, noting that one after
another is neither a raven nor black, seems irrelevant. In Nyāya terms, a
pervasion of being-black (H) by being-a-raven (S) should require positive
evidence. Negative correlations, being practically everywhere, shouldn’t
count. Nyāya philosophers would see the pervasion of being-a-raven by
being-black as known by both the positive and negative methods, but again
the problem is plain.
On the other hand, if the absence of the probandum (~S) is grasped as
having the same or an inclusive extension with the absence of the prover
(~H), the presence of the prover (H), which is an absence of the absence
of itself (~~H), proves the absence of the absence of the probandum (~~S),
which is the probandum itself (S). This seems to be how Gaṅgeśa and his
followers understand the logic of the “negative-only” inference. He points
to the logical rule of transposition—(p → q) ≡ (~q → ~p)—as underpinned
by the ontology of pervasion such that the negative-only makes the same
type of thing known as is known by positive correlations.44 A pervasion
made manifest negatively is equivalent to one made manifest positively.
In other words, all things H being things S may be evident only from the
ramification that everything that is not an S is not an H. Positive correla-
tions may be hidden or in principle not exist. A double absence is equivalent
to a positive presence, that is, with respect to two “mutual absences” or
“distinctnesses” (a’s distinctness from being-distinct-from-a is equivalent to
a’s identity). The fundamental truths of things (tattva), which are captured
by philosophical defi nitions, seem accessible only through knowledge of
fundamental distinctions. A defi ning characteristic of earth, for instance, is
having-smell, which can be used as the prover for a negative-only inference
to earth’s being a distinct substance, where the pakṣa is everything earthen
(thus all things earthen are bracketed in the context of the inference): wher-
ever the non-earthen, there the not-having-smell.
To fi rm up our grasp on the form and to appreciate a subtle point of
epistemology, let us focus now on a famous Nyāya inference to “self,”
ātman, as the locus of, in some versions, having-breath, and, in others,
 Inference 67
desire, cognition, and other psychological properties, thence inferred to be
a fundamental subcategory of substance. A simple version:

a (pakṣa) every living body

S (sādhya) has a self
H (sādhana) has breath

Thus, “Every living body has a self, since every living body has breath,
unlike a pot (a pot being qualified by both absence-of-self and absence-of-
breath).” The inferential subject includes all living bodies, and so there is
no sapakṣa, no examples of the probandum known outside the set of things
that are living bodies. Thus, the inference has to be kevala-vyatirekin,
based solely on correlations of absences, “unlike a pot,” a pot having nei-
ther breath nor a self.
Gaṅgeśa wrestles with the problem of how the probandum “has-a-self”
is known, known at least well enough to be available for inference. And
he concedes that his carefully reworked inference will not go through for
everyone, not for someone who does not understand the property to be
proved, for example, the Buddhist.
To fi nd such an epistemic consequence makes Gaṅgeśa’s attitude appear
to be different from that of his teachers who do try to prove the existence of
the self against the Buddhist stream theorist. Buddhist philosophers reject
Nyāya’s whole approach to ontology as a matter of property-bearers endur-
ing through at least some types of property change, pointing to various
conundra or paradoxes concerning relations (such as the so-called Bradley
problem).45 For the Buddhist opponent, “self” is a “convenient fiction,” bun-
dling momentary dharma or properties that have no property-bearers but
are causally ordered in sequences. Gaṅgeśa holds that his Nyāya inference
to self is bona fide only as svārtha, “inference for oneself,” not parārtha,
“inference for others,” not universally at least, having no force against the
Buddhist for whom the probandum does not make sense. Thus the engine
of the negative-only is restricted by Gaṅgeśa to non-controversial predi-
cates, or those for which lots of work has been done in other patches of the
system. As you can see with the following quotation, much in the Nyāya
system is presupposed.
Here is the fi nal version of the argument that Gaṅgeśa defends, which is
really two inferences, the second, negative-only:

With it established that desire resides in a substance, that substance—in

the absence of a locus for it in earth and the rest (of the eight)—is dis-
tinct from the eight substances (on the list) beginning with earth, since
it possesses a property that does not occur in the eight or since there
are counterconsiderations (“defeaters,” bādhaka) that rule out earth
and the rest (of the eight)—thus is there established a substance over
and above the eight. Given that establishment, there is proved—with
68 Nyāya Epistemology
respect to desire (as inferential subject)—a having of (i.e., a resting
in) a substance that is over and above the eight, or, an occurring in a
substance that is over and above the eight. Either probandum (such a
having or such an occurring) is “well-known” (prasiddha) in the case
of substancehood (the universal of substances). The cognition of an
occurrence in a substance that is over and above the eight substances
has as its qualificandum desire. That is proved true by an inference
based on negative correlations.46

The ninth substance is the self, ātman. Desire is had by nothing that is not
a self.
A second philosophical inference is maybe a little less system-bound,
an inference to a creator based on both positive and negative correlations:
“Earth and the like have a conscious agent as an instrumental cause, since
they are effects, like a pot and unlike an atom (whatever is an effect has a
conscious agent as an instrumental cause).”

a (pakṣa) = earth and the like

S (sādhya) = having a conscious agent as an instrumental cause
H (sādhana) = being an effect

Buddhist philosophers try to refute this by pointing to counterexamples, such

as growing grass. Growing grass exhibits the prover property, being-an-ef-
fect, but not the property to be proved, having-an-agent-as-an-instrumental
cause. The Nyāya reply is to point out that growing grass and all such exam-
ples are in dispute, that is to say, fall within the domain of the inferential sub-
ject (pakṣa), here “earth and the like,” to wit, anything that is an effect but
whose agential cause is not apparent, unlike the comparison class, a pot, for
instance, which is clearly both an effect and has an agent as an instrumental
cause. Nothing that falls within the domain of the subject can be used either
as an example supporting the rule of vyāpti inclusion or as a counterexample,
since that would beg the question. The whole point of inference is to make
something known that was not known previously and here, as with the exis-
tence of atoms, the conclusion could not be known perceptually. The theo-
logical inference has the purpose of showing that things like the earth, such
as growing grass, have an agent within the causal complex that brings them
about. This is not something that we know without making the inference.
So, by the rules governing proper inference, the putative counterexample
is rejected, and the proof looks pretty good. For, cleverly it divides all things
into three categories, as Vācaspati remarks.47 There are things that are uncre-
ated, atoms, for instance, created things that clearly have an agential cause,
such as pots, and things such as the earth and growing grass that do not
clearly have an agential cause. This last category becoming the inferential
subject is bracketed such that no examples can be pulled from the set.
 Inference 69
However, another refutation refi ned by Buddhist philosophers over the
centuries is not dismissed so easily.48 The Nyāya argument falls to an “infer-
ential undercutter,” upādhi, in this case the property, having-a-perceptible-
body. As we have seen (see p. 58), an inferential undercutter is defi ned as a
property that pervades the probandum while failing to pervade the prover,
that is to say, that is entailed by the presence of the probandum while not
being instanced in at least one case of the prover. In symbols:

1. (x) (Sx → Ux) (Having-an-agent-as-an-instrumental-cause, which is

the S or probandum property of the target inference, is pervaded by
the agent’s-having-a-perceptible-body, which is the U or undercutter
property: all agents have bodies.)
2. (∃x) (Hx . ~Ux) (There is something that is an H but not a U; some-
thing that is an effect but does not have within its causal complex an
agent with a perceptible body, e.g., growing grass.)

The two conditions being met, it would follow that there is something that
is an H but not an S:

3. (∃x) (Hx . ~Sx) (Something is an H but not an S, something that is an

effect but does not have an agent as an instrumental cause.)

Thus the upādhi defeats the apparent inference by showing that the required
pervasion and entailment do not hold:

4. ~(x) (Hx → Sx) (That every H is an S is false.)

Of course, the Nyāya response is to deny that the putative undercutter, an-
agent’s-having-a-perceptible-body, pervades the probandum of the target
inference, having-an-agent-as-an-instrumental-cause, and several rounds of
objection and response centering on Nyāya’s conception of God as an agent
without a body are aired by Vācaspati and Udayana. This is also Gaṅgeśa’s
main concern. Vācaspati points out that God is like other imperceptible
entities known by inference, atoms, for instance, and that our whole system
of knowledge would break down if we were to reject inference to impercep-
tibles.49 God like other selves (ātman, the category to which God belongs)
is known by God’s acts.
Furthermore, God’s creative activity requires, like that of any agent,
familiarity with the material to be shaped, like the potter’s knowledge of
clay and of what to do to make a pot. God’s knowledge is to be conceived
as appropriate to the tasks to be undertaken, such as originally combining
atoms, and if God like us had to have a body to have knowledge, including
the knowledge required to combine atoms, then, since bodies are combi-
nations of atoms, there would have to be some earlier Creator to make
70 Nyāya Epistemology
the necessary combinations—ad infinitum. Simpler than this conception is
that of a bodiless Creator whose knowledge is appropriate to the material
forming earth and the like but, unlike our knowledge, is not generated.
This seems to be the core of Vācaspati’s reasoning, which is complex tarka,
“suppositional reasoning,” targeting simplicity.50
God may be bodiless in essence but God is capable of assuming a body
for certain purposes. For example, God assumes a body in order to teach.
God is “Guru even of the ancient teachers in not being limited by time”
in the conception of the Yoga-sūtra, which Vācaspati appears to follow.51
Nyāya’s motto in reasoning about God, atoms, ether, and other theoretic
entities in the sense that they are in no instances known immediately by
perception is to assume only so much about a posited cause as is necessary
to account for an effect in view. Nyāya philosophers on mind-body connec-
tions, for example, formulate causal principles on the basis of correlations
without bias about the sorts of things that can be linked so long as the
cause has the character that makes it able to perform the role for which it
is proposed in the fi rst place. Thus selves carry desires and intentions as
qualities, and atoms are colored, according to Nyāya. Furthermore, God is
not omnipotent in that atoms, ether, and individual souls are eternal and
uncreated, and laws of karmic justice, et cetera, are what they are inde-
pendently of God’s creative action. But God is omniscient (sarva-jña), in
knowing all there is to know about everything. Otherwise, God would not
be capable of playing the required causal role. Only an agent thoroughly
familiar with the material with which he or she works is capable of produc-
ing the intended result, like a weaver and thread to be woven. 52
Finally, a brief look at a third philosophical inference, to the possibility
of “liberation,” mukti, from rebirth and future suffering. From the end of
Gaṅgeśa’s chapter on inference:

Then the justification (pramāṇa, the inference to the conclusion that

suffering can come to an end for an individual), well, it is that suffering
in general, or the suffering in general of Devadatta (i.e., of any indi-
vidual self), is the counterpositive of a non-simultaneous destruction
existing in the same locus (the self of Devadatta, namely), since it is a
property that occurs only as an effect (connected to the destruction),
or since it is continuous (until its end), like (the light of) a lamp. And
this continuity is its being a property that occurs only as an effect (with
respect to its destruction) at distinct times (in the same locus). And in
this way the property to be proved (viz., being the counterpositive of a
non-simultaneous destruction existing in the same locus) is also appli-
cable to pleasure in general and the like: liberation is the breaking up
of the causal chain along with its source.53

Here the inferential terms are the following, with the example (dṛṣṭānta) as
the light of a lamp:
 Inference 71
a (pakṣa) the suffering of a particular (transmigrating) self P
S (sādhya) being the counterpositive of a non-simultaneous destruction
existing in P
H1 (hetu) being a property that is exclusively an effect
H 2 (hetu) being continuous (until its end)

The destruction that is mukti is not of current but of future suffering, thus
a “non-simultaneous” destruction. The destruction exists in the same locus
as the suffering that would occur were it not blocked. The suffering does
not really occur—it never comes about—and so its ontological status is
problematic—the flood prevented by the new dam. Clearly, it does not cause
anything, though to prevent it people do set out in action. The blocked
future suffering is an effect only in connection with the bringing about of
the destruction in relation to which it is the absentee or counterpositive,
pratiyogin. Any destruction is of something (the counterpositive) at a place
or locus. The example appears to work only with the second of the two
provers mentioned: like a lamp’s stream of light that can cease absolutely,
the suffering of an individual goes on continuously—in the sense of regu-
larly if not without occasional spells of pleasure—until its absolute end.
Pleasure is similar.
The obvious observation in overview is how much these philosophical
inferences rely on the Nyāya system. But note that the last seems system-
atic all right but also suspiciously informed by an extra-system religious
consideration, thus forced. Of course it seems that way because the whole
discussion has been introduced by Upanishadic testimony. Nyāya philoso-
phers see the Upanishads and other sacred texts as the testimony of yogins
or of God taking a body in order to teach. The other two inferences have
their merits without this deficiency (the theological one is hardly to the
all-powerful God of popular religion and theistic Vedānta), particularly
given the epistemic concession we reviewed concerning the availability of
the probandum with the inference to being-enselved. Tying up tightly with
system is philosophically a good thing, it seems to me, although to recon-
struct the reasoning may call for considerable digging, as, for example,
with the mukti inference.54

FALLACIES AND DEBATE THEORY

The entire fifth and fi nal book of the NyS is devoted to fallacious objections
(jāti) to a cogent inference for others along with points for formal censure
in a debate (nigraha-sthāna) including fallacies of the prover (hetv-ābhāsa).
Let me paint some background. Debate on metaphysical topics is reported
in the oldest Upanishads, and there are debate manuals in Hindu, Buddhist,
and Jaina monastic traditions from very early. In what is probably the old-
est Upanishad, the Bṛhadāraṇyaka (c. 800 bce), King Janaka is depicted
72 Nyāya Epistemology
awarding prizes to one Yājñavalkya who in public debate has vanquished
all comers (about the truths of Brahman and the self, ātman).55 Evidence
abounds that organized philosophical debate was a common practice long
before Nāgārjuna and the fi nal redaction of the NyS.56 In the Mahābhārata
(“Great Indian Epic”), some of the fallacies that come to be minutely ana-
lyzed some ten centuries afterwards are spelled out by one Sulabhā, an
outsider female character speaking in a king’s court on mukti and then,
challenged, on what counts as a good speech.57
B. K. Matilal writes summarily about the origins of logic in India, “Logic
developed in ancient India from the tradition of vāda-vidyā, a discipline
dealing with the categories of debate over various religious, philosophical,
moral, and doctrinal issues.”58 This may well be right, but I should like to
say a few words about debate and inference that tempers the judgment. For
example, it would be wrong to think of Gautama’s treatment of fallacies
and winners and losers in a debate context as at all significantly rhetorical,
concerned with procedure (although failing to state a thesis could be termed
procedural), character of the opponent or the audience, or other irrelevance
from an epistemic point of view. Rather, his treatment in the sūtras of chapter
5 centers on proper inductive procedure as relevant to the standard paradigm
of inference. Such a focus continues through all periods of Nyāya literature.
To be sure, the NyS mentions three types of debate: (a) vāda, which is
aimed at fi nding the truth, along with (b) jalpa and (c) vitaṇḍā, which are
aimed only at victory and where making tricky arguments is appropriate
(so long as you don’t get caught).59 Vātsyāyana implies that one should take
into account personal character (including the nature of the beliefs held by
an opponent) in deciding a debate strategy. True inquiry, vāda, is carried
out among friends and teachers who are all committed to ascertaining the
truth, whereas wrangling and captious argument (jalpa and vitaṇḍā) are
only for the purpose of defeating a malevolent adversary and to protect
right views from public attack, as thorns are used to fence and protect
young shoots.60 And Vātsyāyana says that we should not trust those who
purport to demolish others’ views without trying to establish their own.61
All this I wish to acknowledge while making just one remark: all Nyāya
texts are in effect vāda, “discussion” aimed at discovering and expressing
the truth. Not only does this hold for metaphysical topics but also rhetori-
cal strategies are not a topic of later Nyāya works except to expose them as
fallacies (the exposing amounting to a winner) whereas there are tomes on
why the logical and inductive fallacies do not lead to the truth. The point
is to be aware of tricky arguments so that one is not fooled. There are no
examples given of tricky arguments to be employed against a Buddhist or
another opponent. On the contrary, the positions and arguments of other
schools are taken quite seriously, and the “winners” and “losers” laid out
in the NyS derive their status as winners and losers from the cogency of
the standard pattern of inference including, to be sure, the evidence for a
pervasion or general rule.62
 Inference 73
Adversarial debate may have epistemic relevance nonetheless. As
remarked, a successful counterinference undercuts warrant no matter that
our inductive base seems solid. As Kisor Chakrabarti elaborates, a principle
he designates GAIE, the “general acceptability of inductive examples,” is
embedded in the discussions of NyS chapter 5.63 A counterexample defeats
putative knowledge of a pervasion, and thus should be held to a standard of
evidence that, if lower, as Gaṅgeśa suggests (recall our discussion of the dubi-
ous upādhi), than that of the inductive set of positive and/or negative correla-
tions on which some bit of putative inferential knowledge is based, should
not be much lower. No mere possible candidate is conclusive, but rather that
a counterexample is at least plausible should be acceptable to all parties (the
undercutting of the Mitrā inference relies on real possibilities). The Buddhist
counterexample to the Nyāya inference to God we just discussed, growing
grass, does not count as a veritable counterexample from the Nyāya point of
view since it is off-limits as part of the inferential site, pakṣa. The issue is: Do
we know that growing grass does not have an intelligent maker? Buddhist
philosophers say yes; Nyāya philosophers say no. Similarly, Buddhists do not
accept the examples of a pot and a palace as relevantly similar to earth and
the like, since though both are effects and have intelligent makers those mak-
ers could not make earth and the like and thus are not relevantly similar.
There is thus philosophically a stand-off. Our philosophical inferences do
not go through for everyone—as we saw Gaṅgeśa admit if for the slightly
different reason of the availability of the probandum.
Perhaps, then, the conclusions of these should not count as knowledge,
not as nirṇaya in any case, not as reflective knowledge.64 However, that
these and other philosophical inferences should count as nirṇaya seems the
motivation of Vācaspati, Udayana, and the others, it seems safe to say, driv-
ing some of their most innovative reasoning. There are lots of arguments
that are still largely unexplored by modern scholars, including supposi-
tional reasoning, tarka, both negative (pratikūla) and positive (anukūla),
especially with Udayana.65
5 Analogy

Analogy, upamāna, “comparing,” is in several classical schools the way

we know the similarity between two things (or more), which can be impor-
tant to know for different reasons. Mīmāṃsakas, “(Vedic) Exegetes,” to
provide an interpretation of Vedic injunctions suitable for practice in
actual performances need to be able to designate substitutes, of one type
of grain for another, for example, or one animal for another, depending
upon availability in the fi rst place but in the second place similarity. Simi-
larity is a secondary criterion governing the choice. In Vedānta, analogy is
said to be useful for understanding the Upanishads which make compari-
sons between spiritual or yogic experience and the experiences of ordinary
humans.1 A similar utility is identified in the case of understanding Vedic
words such as ‘svarga’, “heaven,” the gaining of which is said to be the
purpose of certain sacrifices. 2
Logicians, such as the late Yogācārins along with Nyāya philosophers,
Jainas, and others, fi nd similarity, or relevant similarity, to figure in infer-
ence as a knowledge-generating process. It is through cognizing similarity—
and dissimilarity—that we arrive at knowledge of pervasion as required for
inferential knowledge.3 A kitchen hearth counts as an “example” in the
stock inference because of its relevant similarity to the mountain which is
the center of inquiry. It is part of what is called the sapakṣa, the set of posi-
tive correlations, that make us know an inference-underpinning pervasion.
Knowledge of similarity is not viewed in Nyāya as the result of analogy as a
knowledge source—analogy is restricted in scope to a subject’s learning the
meaning of a word. But pervasion is known through generalization from
cases—or even a single case—not only according to the Buddhist Dignāga
but also Maṇikaṇṭha Miśra among New Naiyāyikas4 —presupposing
knowledge of relevant similarity.
Inference aside, Nyāya philosophers think that analogical knowledge
can serve various practical purposes. Vātsyāyana mentions medicine and
the example of someone directed to look for a certain herb with which he is
unfamiliar but which, he is told, resembles certain other herbs in ways laid
out. The subject’s recognizing the desired herb by its similarity to the oth-
ers known beforehand is crucial to his securing the medicine, Vātsyāyana
 Analogy 75
implies.5 One would think that Udayana, among others, would employ the
means theologically to introduce terms like ‘īśvara’, the “Lord,” but I have
been unable to fi nd the Thomist idea of analogical predication in Nyāya.6
The slightly maverick tenth-century Nyāya philosopher Bhāsarvajña jet-
tisons Gautama’s commitment to analogy as a distinct knowledge source,
and the Nyāya position is attacked from many corners.7
What comes to be a stock example of analogical knowledge is indicated
even in the terse sūtras of Gautama and is elaborated in the commentaries and
textbooks well into the New Nyāya period.8 Note that the example is wired to
the Nyāya account. Other schools present different pictures of the knowledge
source and its results. The Nyāya scenario has it that a subject S inquires of
a forester about a gavaya, which is a kind of buffalo, having heard the word
‘gavaya’ used among his schoolmates but not knowing what it means, i.e., not
knowing to what ‘gavaya’ refers. Questioned by S, the forester replies that a
gavaya is like a cow, mentioning certain specifics as also some dissimilarities.
To simplify, Nyāya philosophers say that the forester makes an analogical
statement (“A gavaya is like a cow”), whereby, according to Gaṅgeśa and
followers, our subject S now knows in general (sāmānyataḥ) what the word
means. But S does not yet know how it is used, does not know its reference,
which is deemed its principal meaning.9 Later encountering a gavaya buffalo,
S says, “This, which is similar to a cow, is the meaning of the word ‘gavaya’,”
a statement which expresses S’s new knowledge. Every bit of knowledge has a
source, and this one has been generated by upamāna, “analogy.”
The objection that so-called analogical knowledge can be explained
as knowledge, to be sure, but knowledge resulting from perception and
inference (along with testimony, which itself reduces to a combination
of the workings of perception and inference according to Vaiśeṣika and
Yogācāra), is also mentioned by Gautama, laconically.10 The commenta-
tors provide the answer that the knowledge had by S in the example is not
inferential in that the conditions for inference are not met. These include
having a known inductive base. However, no other gavaya buffalo has been
encountered by S in the stock example. Furthermore, bits of analogical and
inferential knowledge are brought about by different cognitive triggers, the
one by “reflection” (parāmarśa), as explained earlier, and the other by per-
ception of cow-similarity in a present gavaya, i.e., perception informed by
an analogical statement. Of course, the results also differ in kind, as is said
in the last sūtra of the set.11
Nyāya sees meaning as conventional as opposed to a Mīmāṃsaka non-
conventional, intrinsicality view. Conventions have to be learned. S begins
to learn the convention for ‘gavaya’ without knowing it fully. S knows bet-
ter what to look for from the analogical statement of the forester, namely,
an animal that resembles a cow. But S knows only in general, not fully or in
particular, until encountering a referent, an actual gavaya buffalo, and put-
ting the information together, S comes to know and exclaim, “Now that’s
what the word means!”
76 Nyāya Epistemology

LEARNING WHAT WORDS MEAN

Nyāya philosophers pay quite a lot of attention to variations on a stock

example of language-learning on the part of a child hearing an elder’s
command, “Bring the cow,” and observing a cow being brought.12 There
is much controversy about this and in general about word and sentence
meaning. The NyS itself contains a considerable stretch of sūtras about
word meaning, NyS 2.2.55–66, which are almost exclusively (except for
the remarkable sūtra 2.2.59) about reference, abhidhā, as opposed to indi-
rect, figurative meaning, lakṣaṇā, which will be discussed by us in the next
chapter.
Grammarians distinguish three types of reference in a schema that
becomes widely employed (for example, in aesthetic textbooks, alaṃkāra-
śāstra, as well as in the literature of the philosophic schools):13

(1) rūḍhi, “the conventional”

(2) yoga, “the derivative”
(3) yoga-rūḍhi, “the derivative and conventional”

For Nyāya, the conventional, rūḍhi, is a matter of cultural mandate, as

when the name ‘Chaitra’ is stipulated, so to say, to refer to a certain indi-
vidual by Chaitra’s mother.14 There is a direct relation between word and
object unmediated by meaning belonging to the word’s semantic parts, if
the word has them, etymologically considered. The word ‘cow’ for example
has no parts and is straightforwardly conventional in meaning cows. But
‘doghouse’ has its meaning determined by a combination of the meanings
of ‘dog’ and ‘house’. There are other compound words, however, for exam-
ple, ‘windfall’, whose denotation is not dependent on the meaning of the
etymological parts, as a windfall is a “dead metaphor” and no longer car-
ries the meaning of a bounty knocked by the wind to the ground. Similarly,
the Sanskrit ‘maṇḍa-pa’, is used to refer to ascetics without implying that
they are “(alcoholic grain-)scum-drinkers,” an example discussed in the
next chapter (see p. 93). In English, an example of the third type is ‘White
House’ used to refer to the office of the U.S. President. The compound has
been conventionalized and refers as a whole, but the building that houses
the offices of the President and his top officials is indeed white. In Sanskrit,
a stock example of the third type is ‘paṅka-ja’, “mud-born,” which refers to
lotuses. The meaning is conventional in that it is lotuses that are meant and
not some other species of plant or animal that lives in mud. But lotuses are
indeed mud-born, and so the meaning of the word is also derivative.
Verbs as well as nouns have denotation. The verb ‘cook’ by primeval
convention picks out the action of cooking whether in general or in par-
ticular or both—indeed like all words that function as single semantic
units independently of their parts to express a single meaning. The second
category, derivative or etymological meaning, yoga, results not just from
 Analogy 77
compounding but also from grammatical derivation primarily through the
addition of suffi xes and prefi xes. For example, in Sanskrit ‘pācaka’, which
means a cook, is derivative from the verbal root ‘cook’ (‘pāc’) through
affi xation of the agential suffi x ‘aka’, combined (yoga), so to say, to mean
a cook. The reference can be analyzed in terms of the referential functions
of the parts of the word. As the ultimate parts into which a derivative is
analyzed are themselves conventions, rūḍhi emerges as the foundational
type of reference.15
Reference, abhidhā, is thought of as the function of a word with respect
to an object or objects, that which is referred to. Now objects are ana-
lyzed in terms of the Vaiśeṣika ontology from as early as Gautama and
Vātsyāyana, but there is also a simpler schema elaborated by Vātsyāyana
who says that meaning connects words with individuals, universals or gen-
eral characteristics, and form or shape. That is, of three candidates for the
referent of the word ‘cow’, for example, championed by one or another
philosopher or school—(1) the individual (Bessie the cow) or group of indi-
viduals, (2) the universal or class characteristic (cowhood), and (3) a shape
(the shape of a cow as with a pastry or chocolate cow)—Vātsyāyana says
that context of usage determines which is predominant or subordinate but
also that all three are normally meant.16
The New Nyāya view is that individuals are directly referred to in sen-
tences not just simultaneously with a universal or another repeatable char-
acteristic but as qualified by the generality.17 Now enormous controversy
breaks out over generality, just among Mīmāṃsā and Nyāya philosophers,
not to mention Buddhist nominalists. Both of the main camps of Mīmāṃsā
view the universal as the primary meaning of a word. The Bhāṭṭa, as
reported by Gaṅgeśa, holds that the individual is known only by a kind of
inference, which in the Bhāṭṭa’s own view is more precisely the independent
knowledge source, “circumstantial implication” (arthāpatti, discussed in
note 41 to chapter 4). Gautama’s fi nal position seems to make a conces-
sion to the Bhāṭṭa by admitting that context determines whether we are
talking principally, one, about Bessie or, two, the cow as a type or, three,
shape as in a cow-shaped pastry. But Gaṅgeśa argues at length that the
Nyāya siddhānta is that reference is to be understood as to an individual
as qualified, a qualificandum qualified by a universal or another repeatable
character such as shape (eka-vṛtti-vedyatā).18 The main argument is that
the conditions sufficient for cognition of an individual are also sufficient
for cognition of a general characteristic, since we never have a cognition of
a bare particular as bare.
Unfortunately, pragmatics is not much developed in Nyāya, which pretty
much leaves the field to Mīmāṃsā. Still, Nyāya does embrace a holism about
sentence meaning, where speaker’s intention is championed as in some cases
playing a crucial part in the analysis of what is meant. Clearly contextual
factors are recognized, such as prior beliefs of a hearer. As we shall see
in the next chapter on sentence meaning and knowledge from testimony,
78 Nyāya Epistemology
context of usage is also key to some instances of figurative meaning, in
the Nyāya view. Analogical knowledge, however, is not simply a matter of
comprehension through simile or metaphor.
The examples of learning the meaning of words given in Mīmāṃsā look a
lot like Western examples of learning by ostension. But there are also words
like ‘ākāśa’, “ether” (the medium of sound), which are viewed as technical
terms learned in the course of learning a system through defi nitions not evi-
dent in the practices of everyday speech, vyavahāra.19 Grammatical deriva-
tion may, as pointed out, depend ultimately on root conventional meanings.
But that is not the whole story, since there are also combinational rules, at
least some of which one learns by studying grammar. Grammar is looked
upon as an auxiliary art crucial to understanding the Veda—this seems the
pertinent bit of cultural context, as well as to note that there was a tradi-
tion of grammarian literature predating and running alongside Nyāya and
Mīmāṃsā for centuries, to which the philosophers quite often refer. We
might also mention dictionaries, which are commonly listed as one of the
ways we learn the meanings of words, though this is only a special form of
testimony. 20
Testimony, however, is not, in the Nyāya view, the only way we learn
the meaning of words. There are, furthermore, problems with the stan-
dard theory inherited from Mīmāṃsā, as Gaṅgeśa and others bring out.
For example, the child in the stock example learning the meaning of “Bring
the cow” by witnessing a cow being brought when a grown-up uses the
words, is not, at least not when very young, able to differentiate individual
semantic units. 21
In any case, analogy is, while not the only way or even the most impor-
tant, a unique source for learning, for the fi rst time, the meaning of a word.
A referent is indicated not by direct denotation but by a kind of indirect
ostension—as red is said to be the color of ripe apples or of blood exposed
to air. Such a defi nition of ‘red’ tells us where to look (at blood or apples) to
learn the word’s meaning.

“INDIRECT INDICATION,” UPALAKṢAṆA

Analogy appears to work through an appreciation of sense independently

of reference. The analogical statement of the forester seems to convey the
sense of the word ‘gavaya’ which S understands before S understands the
referent. Nyāya philosophers talk about speech acts that employ “indirect
indication,” upalakṣaṇa, in ways that suggest the Fregean theory of sense
determining reference. There are, however, large differences.
In the classical Indian context, the issue with Nyāya is how does our sub-
ject S understand the analogical statement sufficiently well on the one hand
to be able to recognize the newly encountered animal as a gavaya buffalo,
 Analogy 79
and insufficiently well on the other that the meaning of ‘gavaya’ is not
understood by S before encountering one in the flesh? Gaṅgeśa voices the
mainstream view when he insists that the word’s reference is not grasped
without the perception of at least a single referent, but that also (here some
Nyāya philosophers disagree) experience of more than one referent or even
of the same referent on more than one occasion is not needed. We grasp
the reference of ‘Chaitra’ on a single occasion of use, although the word
indicates Chaitra at other times, too.
Meaning is in the fi rst place a matter of reference, a direct relation
between word and thing. Use of a general term like ‘cow’, however, does not
directly refer to all cows, past, present, and future, but only indirectly “sug-
gests” or “indicates” them in such a case as Bessie’s being said to be a cow
(i.e., to have cowhood, in the Nyāya analysis, like all cows, present, past,
or future). Again, learning a word’s reference need not involve ostension. A
general characteristic can be indicated indirectly. A general characteristic
can also indirectly indicate particulars, as in the operation of analogy. In
other words, S does know something in general about what a gavaya is
(namely, similar to a cow) from the forester’s statement. This is not to say
that S grasps the universal, being-a-gavaya-buffalo, since universals dwell
only in particulars and are in the very best way known only as exemplified.
However, similarity can be an indicator—the sharing of characteristics such
as having-horns and so on—such that S has a proto-understanding, has
begun to grasp what a gavaya is. The suggestion (upalakṣaṇa), the indirect
“indication” as opposed to direct reference, accomplished by the forester is
sufficient to allow analogical knowledge to arise upon S’s recognition of the
cow-similarity exhibited in a critter encountered in the flesh.
Gaṅgeśa points out (see the appendix, p. 121) that with words that are
derivatives (yoga) the power of reference is etymologically in conformity
with that from which it is derived. He says explicitly that from “etymology
there is also indication (upalakṣaṇa) that fi xes (general meaning or sense,
though not necessarily a specific reference).” For example, a lotus known
as “mud-born” could guide a person’s identification of lotuses for the fi rst
time, being indirectly indicated by the meaning of the attributive. 22
A word’s having a power of indirect indication seems to be an idea close
to that of a word having connotation, as opposed to denotation, as dis-
cussed, for example, by J. S. Mill (whose nonconnotative/connotative is a
precursor of the Fregean reference/sense distinction).23 But there is at least
one strike against such an interpretation. In indirect indication almost
anything can be used to fi x an expression’s reference. Common examples
are “By the matted hair, it’s an ascetic” and “By the hovering crows, it’s
Devadatta’s house.” With these two instances, both of which presuppose
a conversational context, the referent (the ascetic, Devadatta’s house) are
indirectly indicated by things (matted hair, hovering crows) that have little
or nothing to do with the nature of the thing indicated. Thus sense is not
the only referent-indicator but rather quite extraneous things can serve such
80 Nyāya Epistemology
as the hair and the crows. In his perception chapter, Gaṅgeśa distinguishes
between a true qualifier and an indicator used merely to fi x reference by a
principle of non-opposition. 24 We cannot know Devadatta’s house as with-
out shape, or an ascetic without humanity, but we can know immediately
following the use of crows to indicate it that Devadatta’s house is without
crows. Similarly, the ascetic could shave his head. 25 On the other hand,
like Mill’s “connotation” and its connection with poetry, indirect indica-
tion connects with the independent verbal power called lakṣaṇā, “figura-
tive speech,” which is for Nyāya a second power of words in addition to
reference. Technically, a gavaya is literally like a cow. But our ability to
understand general features indirectly indicated makes possible both com-
munication with figurative speech (including poetry) and the working of
analogy as a knowledge source. (There is a section on figurative speech in
the next chapter.)
In sum, an appreciation of what it is to be similar to a cow is not in itself
sufficient to fi x the reference of ‘gavaya’ for S, our classical urbanite who
has never seen one. But it does make him ready to acquire the word when
he recognizes the animal by its cow-similarity.

THE ONTOLOGY OF SIMILARITY

The working of analogy depends upon real-world similarity between things

compared, but the ontology of similarity is tricky business. In the transla-
tion appended here, Gaṅgeśa presents his own theory along with those of
several others, in particular Mīmāṃsakas.
Similarity according to Gaṅgeśa is a supervenient relational property:
something’s having a lot of the same properties as something else. It is not
a universal, he argues, for similarity relates a correlate (the gavaya buffalo)
and a countercorrelate (the cow), whereas a universal, in contrast, rests as
a unity in, for example, with cowhood, all individual cows. In this way,
it is like contact, saṃyoga, but there are also rather obvious differences.
It is not reducible to any single category among the traditional seven. For
some substances are like one another as are certain qualities and actions.
But similarity also is not, pace the Prābhākara, a category over and above
the recognized seven. Gaṅgeśa’s main argument there is that similarity is
not uniform. It is to an extent a property that is mind-imposed in that the
counterpositive (the cow) is supplied from our side. Moreover, it supervenes
on other properties.
To hold that similarity is not a universal is not to say that there is
anything grammatically or in any other way wrong in pointing out that
one cow, for instance, is similar to another. One cow is indeed similar to
another, and on the basis of cognized similarity we recognize universals
(and other general characteristics) as well as the pervasions that underpin
inference. Similarity and dissimilarity supervene on common properties
 Analogy 81
and absences of common properties. In some cases an underlying similarity
is made manifest by indicator properties such as having-horns for a cow
and a gavaya buffalo alike. But in other cases similarity is directly manifest
in, say, two cows.
The main arguments for the one or another ontological view trade on
considerations of everyday usage, vyavahāra. In particular, a theory of
similarity must be able to explain why we say that some things are only a
little similar to other things whereas sometimes we say that there is a lot
of similarity. An adequate ontology of similarity should explain how the
critter could vary like this, and so, again, similarity cannot be a universal,
Gaṅgeśa argues, since universals do not vary like this. Something is a cow
or is not a cow. It is inadequate to say that the evidence for—or manifesta-
tion of—the universal varies, since a pot does not vary in its pothood by
how evident is its existence. Other views and arguments are presented by
Gaṅgeśa in the appendix. Let us desist and move on to consider testimony
(probably the most controversial of Nyāya’s knowledge sources) with just
one further remark having to do with Nyāya’s two-faced policy of analyz-
ing entities ontologically.
Similarity is an example of a property that can be explained as noth-
ing more than a way of grouping other properties. In contrast, analogy as
a knowledge source although involving the operation of perception and
testimony is nevertheless a unique and irreducible way we get knowledge
of a certain sort. This seems to be Gaṅgeśa’s central point in his analogy
chapter. And his arguments show (along with some other considerations), I
shall maintain in Chapter 7, Nyāya’s willingness to admit pramāṇa in addi-
tion to the canonical four, no matter that they depend in some way or other
on perception or another source.
6 Testimony

In the epistemology of testimony, we have the clearest application of the gen-

eral principles of Nyāya’s externalist view of first-order knowledge-generation
and its internalist view of certification and reflective knowledge. Testimony,
śabda, is a knowledge-generator, pramāṇa, and its result, śābda-bodha, is
testimonial knowledge. By being told that p, we come to know that p, not
what some person has said. We know that the pot is in the kitchen, not, in the
first instance, that S has said that the pot is in the kitchen. We can know that,
too, if we attend to it, but normally we are occupied not with the person’s
saying such and such but with the such and such said, the fact or object. As
early as the NyS, Nyāya resists an inferentialist account of testimony accord-
ing to which we do get knowledge from what someone says but only by an
inference to its truth based on other things we know, in particular about the
trustworthiness of the speaker and the meaning of the words. Nyāya insists
instead that we know directly that the pot is in the kitchen by the power of
the words. We know unless (bādhaka-abhāvāt, “unless there is a defeater”)
we have reasons to suspect the truth of the statement.
Nyāya’s position is misrepresented as a charity view of testimony. We do,
it is assumed, normally give the benefit of the doubt to a testifier. But we
do not if there is a defeater in place (a “preventer” of testimonial transmis-
sion, pratibandhaka, i.e., bādhaka, a “defeater”), such as our knowledge
from other sources that the pot is not in the kitchen or that the testifier has
suspect character. The policy of “innocent unless reasonably challenged
(or blocked beforehand)” is not a gift but common practice, impelled by
our need to get along with our worldly business. That a bit of testimony
is veridical is not only an epistemic default but an assumption required by
our overall language use. As mentioned in Chapter 1, Gaṅgeśa begins his
testimony chapter by pointing out the self-refutation of those who would
testify that testimony is not a source of knowledge.1 In sum, Nyāya’s prin-
ciple is not so much “Give the benefit of the doubt,” since there are many
cases where we should not be charitable. But it is held we do have to be at
least minimally charitable. Otherwise, we wouldn’t know very much, not
even our own names, as has been often remarked. I take it on the testimony
of (especially) my parents that the name I have is indeed mine.
 Testimony 83
Now of course Hinduism with its reverence for the Veda appears to be
the deep context for Nyāya’s generous attitude towards testimony—the
same word in Sanskrit sometimes being used for Vedic revelation or scrip-
ture, śabda, the “Word,” as for the knowledge source. But contrary to the
general policy with testimony, Gautama in his sūtras defends Vedic author-
ity by argument, specifically by arguing that the confi rmation of Vedic
medical teachings in practice, in matters of our own experience, gives us
a reason to trust its teachings about matters not of our experience. Thus
we should trust the teachings about heaven and the like although they are
uncertifiable by us directly while we are alive. 2 This shows that Nyāya takes
its work to be to maintain the rationality of belief, not to defend the Veda
and Vedic practices as a matter of principle.
Nevertheless, despite a difference in attitude towards the undertakings of
philosophy, Nyāya and Mīmāṃsā are often allies with similar agendas when
it comes to knowledge from testimony. Mīmāṃsā is mainly concerned with
Vedic hermeneutics, and Mīmāṃsaka philosophers pioneer many of the best
arguments and counterarguments against Buddhists and others who attack
Vedic authority. Much in the Mīmāṃsā approach to sentence meaning is
crucial to any study of Nyāya views about language, but I wish to trim away
non-essentials to focus on testimony as a knowledge source.3 Thus let us start
with Gautama’s definition at NyS 1.1.7, “Testimony is the (true) statement
of an expert,”4 which indicates certain certification conditions that we shall
look at with respect to everyday knowledge, leaving the application to the
Veda aside. (We’ll come back to the Veda in the last section.)
The later tradition focuses on the conditions wrapped up in what it is
to be a (true) statement, whereas Vātsyāyana and company focus on what
it is to be an “expert,” āpta, which is a trustworthy authority. “Expert” is
indeed not an adequate translation because there is a moral dimension. An
āpta is a person who not only knows the truth but who wants to communi-
cate it without deception. Let’s quote the “commentator” (as interpreted by
Vācaspati in particular, as Uddyotakara goes off on a tangent):

[Testimony (śabda) is the instructive statement (upadeśa) of a (trust-

worthy) expert (āpta). NyS 1.1.7.] An āpta knowing the character of the
things (about which he speaks) by knowledge sources is as an instruc-
tor moved by a desire to communicate the things as they are to him
known. The word ‘āpta’ is employed because expertise (i.e., trustwor-
thy expertise, āpti) is a matter of knowledge of x whereby one acts
(confidently with regard to x). The defi nition applies equally to rishis,
noble compatriots, and foreigners (barbarians, mleccha). Moreover,
everyone proceeds to act in everyday life on this basis.5

Vātsyāyana brings out a certain egalitarianism in the workings of the

knowledge source. Contra the privilege afforded the priestly caste in mat-
ters of Vedic interpretation, he says that even foreigners—“barbarians,”
84 Nyāya Epistemology
mleccha—can be the experts whose statements convey to us testimonial
knowledge, provided, as always, they know the truth and want to commu-
nicate it without deception.
The process of generating testimonial knowledge begins with a speaker S
who knows some proposition p by perception, inference, or testimony (chains
of testimony are okay so long as there are no broken links) and who has a
desire to communicate p to someone or other. A hearer H gains knowledge
through a speech act of S communicating p to H, who has to be competent in
the language in which p is expressed, to know all the words and grammati-
cal forms, which, on the Nyāya account, H has learned also chiefly through
testimony but also in other ways, as we have seen. Vātsyāyana elaborates
the causal process whereby individual letters, then words, then the series of
words that comprises a sentence is understood.6 Thereby he focuses atten-
tion on the hearer. But he and all Naiyāyikas of the Old school seem mainly
concerned with speaker conditions for testimonial knowledge, in particular
what is required to be an āpta, a trustworthy authority.
New Nyāya takes the statement under the interpretation of the hearer H
to be the proximate cause, the trigger of knowledge-acquisition.7 The certi-
fication conditions on which Gaṅgeśa and company focus govern the mean-
ingfulness of a statement from H’s point of view, since, Gaṅgeśa argues, S’s
knowing the truth, etc., cannot be established without H’s understanding
something even if it is false.8 Gaṅgeśa admits speaker’s intention understood
in a general sense as a causal factor. But the immediate trigger of testimonial
knowledge is the statement itself. The knowledge gained is of what the state-
ment says, the fact it indicates, not of the speaker’s intention, though, as we
shall see, determining speaker’s intention is crucial to determining meaning
in cases of ambiguity and in some instances of figurative speech. Neverthe-
less, depending on what is said, we could eliminate doubt through percep-
tion or inference as well as by checking the credentials of the speaker. For
Gaṅgeśa and company, it is no longer the speech act of the expert who knows
and wants to communicate without deceit that is the proximate cause of
acquisition of knowledge through testimony but rather the embedded infor-
mation, the proposition contained in it as understood by H.
The argument that we have to understand fi rst S’s statement in order to
check S’s credentials, which establishes the sentence itself as the trigger of
testimonial knowledge, is, contra Arindam Chakrabarti, why there is the
shift of focus, not the knowledge status of true cognitions with “flaws” in
their etiology.9 Chakrabarti, who does not give a citation though he pur-
ports to be giving Gaṅgeśa’s position, apparently mistakes a Prābhākara
pūrva-pakṣa for Gaṅgeśa’s siddhānta.10 His reading is that a parrot can
make us know something (like a tape-recorder). And even a liar deceived
into believing ~p can communicate p trying to deceive H who neverthe-
less comes to know the truth through S’s statement. However, these are
Prābhākara examples given in a pūrva-pakṣa section, examples designed
to show the falsity of the view that testimonial knowledge depends on
 Testimony 85
knowledge of the absence of epistemic flaws in a speaker S. Gaṅgeśa would
not admit that the true cognitions in these instances were “source-born,”
pramāṇa-ja. He says explicitly that S’s “intention,” tātparya, is crucial for
testimonial knowledge.11 The speaker’s intention is not the trigger of tes-
timonial knowledge on H’s part, but is a slightly upstream causal factor
relevant for certification. If we knew it were a parrot or a liar who was
responsible for the statement, we would no longer believe. It is true that
H has to understand something in the case of the parrot, etc. Otherwise,
there would be nothing to check in fi nding out that the parrot’s speaking
is a case of “apparent testimony,” śabda-ābhāsa—a point that Gaṅgeśa
makes himself.12

TESTIMONY NOT A FORM OF INFERENCE

Gautama devotes a stretch of sūtras to the position (of Vaiśeṣika, later of

Buddhist Yogācāra) that so-called testimonial knowledge, while indeed
knowledge, is a form of inference.13 Although the reply, the siddhānta sūtra
(NyS 2.1.52), is hardly detailed, Vātsyāyana and company have no trouble
refuting the position: the certification conditions are different for the two
knowledge sources.14 In particular, there is no memory of pervasion, vyāpti-
smaraṇa, informing testimonial knowledge. Or, if there is, for example,
between a word used in S’s statement and the word’s meaning (a connection
required for us to understand one another from one occasion to the next),
the memory would be an auxiliary factor included in a bundle of necessary
conditions very different from those required for inference. Both inference
and analogy depend on other knowledge sources in their operation, as does
testimony, but are irreducible all the same, as we have seen.
A second type of refutation is alleged by J. N. Mohanty, and is indeed
implicit in what a host of Naiyāyikas say from Vātsyāyana through Gaṅgeśa
and on. But it is not expressly mentioned by Gaṅgeśa in his long, express
refutation of the Vaiśeṣika position that testimony is a form of inference.15
It is that apperceptively we can tell an inferential cognition from a testi-
monial one. “I know this from having heard it” as opposed to “I know
this by having inferred it.” The reason that Gaṅgeśa does not expressly
mention this argument in his own refutation of the inferentialist position
is that the Vaiśeṣika opponent is considered perfectly capable of appercep-
tion, and what we do to defeat his distinct description of the controversial
bit of knowledge apperceived is to point to one or another feature that
distinguishes it from inferential knowledge, not simply to claim that it is all
self-evident. Nevertheless, Mohanty is right. Through apperception, which
is a form of perception, we know a cognition as qualified by its epistemic
type, which is to us then directly evident, as with all perception, but about
which we could be mistaken. We may know a cognition’s content, its inten-
tionality, infallibly in apperception, but we can be mistaken about cognitive
86 Nyāya Epistemology
type, as with the truth or falsity of any putative perception. That’s why it
is possible to have a debate, taking the Vaiśeṣika to call our position into
question. Furthermore, there may be, we may admit, marginal cases about
which anyone could be confused. However, how normally obvious the dif-
ference! Testimony is a matter of being informed by someone else, whereas
inference is carried out by oneself and at least initially for oneself.
A third type of response is put forth by Vācaspati to undermine a con-
tention he attributes to Dignāga that testimonial knowledge reduces to a
form of inference in that we reason from what S says together with the fact
that S is a trustworthy authority to the conclusion that what S says is true.16
The response is that Dignāga runs together the two levels of knowledge:
we may know p by testimony without having certified p as true. Knowl-
edge of S’s credentials is relevant to certification and second-level reflec-
tive knowledge, but it is not to the fi rst-level—unless there are standing
defeaters such as H’s prior knowledge that p is false. If I yell out in the
middle of a lecture asking whether the water-fountain in the hall, which
is not immediately perceptible by me, is working or not, and someone else
(I know not whom, I hear only the voice) yells back “Yes,” then I have
knowledge that the water fountain is working so long as it is in fact. I do
not have to know whether the speaker has an honest character, unless my
water-fountain knowledge somehow gets challenged or undermined. The
reason given by Dignāga shows a certification condition, not the operation
of testimony. There is indeed testimony in the type of case envisaged: p has
to be already comprehended if there is to be a reason given why we should
believe p. Dignāga is targeting a bit of testimonial knowledge and providing
second-level certification.
Now for these and other reasons (though not, I think, by any study of
Nyāya), recent philosophic literature on testimony has begun to criticize
an inferentialist position similar to Dignāga’s embedded in David Hume’s
discussion of miracles and elsewhere: we have to know that the testifier
is trustworthy on the topic and then infer that what she says is true.17 As
early as the nineteen-sixties, H. H. Price, though overall an internalist and
foundationalist, began to move in the direction of Nyāya’s position that
acceptance and understanding are normally fused.18 Recently P. F. Straw-
son writes, expressing now a widespread opinion:

In any community of language-users, perception, memory, and testi-

mony are not only equally essential to the construction of the belief-
or-knowledge systems of its members. It is also true that all three are
on an equal footing in that there is no possibility of a general reductive
analysis of any one of the three in terms of the others, supplemented
by inference. The interdependence of all does not entail the reducibility
of any. If we (often) know, directly and immediately, what our eyes tell
us, then we (often) know, no less directly and immediately, what other
people tell us.19
 Testimony 87
Of course, inference is taken in analytic philosophy to have much broader
scope than Nyāya’s anumāna. Thus an inferentialist account of testi-
mony might not look inferentialist to Nyāya and be cogent nonetheless.
“The testimony debate is largely over whether testimony-based beliefs are
epistemically inferential or, like perception, memory, and introspection-
based beliefs, epistemically direct,” writes Peter Graham in an overview
paper reviewing recent work on the epistemology of testimony. 20 Graham
lists eleven leading philosophers who argue the “direct” thesis in a couple
of dozen books and papers—citing the Scottish philosopher Thomas Reid
(1710–1796) as the classical (Western) source of the thesis, that, as claimed
by Nyāya quite a bit earlier, we get knowledge directly from being told.

STATEMENTS AND FACTS

As mentioned, a stretch of sūtras in the NyS is devoted to the topic of word

meaning. 21 The principle candidates are the genus (cowhood), the individ-
ual (Bessie), and shape or general form (a cow’s configuration and anatomy,
ākṛti). The NyS’s siddhānta, “established position,” is that all three are
simultaneously intended and understood (although Vācaspati insists that
only the individual is denoted by a word, the universal and shape being
indirectly indicated22). The one or the other becomes the predominant or
subordinate element in different contexts of usage but all are in some sense
meant. According to New Nyāya, every common noun is connected in use
not only with a qualificandum but also a predicate or qualifier place in the
schema “Q(a,φ)” where ‘a’ represents a qualificandum, ‘φ’ a qualifier, and
‘Q’ the qualificative relation. To say in Sanskrit, “gauḥ” (“Cow” in the
nominative case, thus “[Here is a] cow” by ellipsis), is to say that there is
something, a, the thing in front of us, that has, or is qualified by, φ, cow-
hood. The predicate in this case expresses a genuine universal, but in other
cases the “grounds of usage,” pravṛtti-nimitta, is something else, the indi-
vidual’s having a certain (proper) name, “(This is) Devadatta,” for instance,
Devadatta as qualified by having-the-name-‘Devadatta’. Thus for Nyāya
reference hits a thing as qualified by a qualifier. The qualifier in a general
form may be called a word’s sense.
A bit of testimonial knowledge has unity, which is reflected in the unity
of the transmitting sentence. For a hearer H to get knowledge from being
told, at a minimum there has to be a sentence and not just a word or a
list of words. Each sentence spoken by S would be potentially a genera-
tor of a bit of knowledge for H, given that H meets the hearer require-
ments. A single word, a proper name, let us say, which some in India and
in the West (e.g., Frege) view as having a complete or saturated meaning
in isolation, would be, employed in the mouth of S, a sentence according
to Nyāya. The school admits single-word sentences consisting not only of
verbs (“Go!”) but also of nouns, even proper names. Since use in Sanskrit
88 Nyāya Epistemology
(as opposed to mere mention as in spelling a word) requires that a noun be
declined, a single name (“Devadatta”) would, depending on the case sign
(itself counted a semantic unit by Nyāya, thus an inflected noun is really
two words, two semantic units, plus their grammatical relationality), be in
the vocative case (thus “[Hey] Devadatta!”) or the nominative, accusative,
instrumental, etc., where a verb would be elliptically understood according
to syntactical “expectancy” and extralinguistic context (thus, e.g., “[This
is] Devadatta”). 23 All words used have, in the Indian terminology, “expec-
tancy,” ākāṅkṣā, for one another, the fulfilling of which completes a sen-
tence. A bit of testimonial knowledge arises when H hears the last word
of the utterance and recalling the previous words/meanings complete with
their mutual expectancy grasps and at the same time accepts (“learns of”)
the fact expressed by the sentence. Seeing me looking for my glasses, my
wife says, “They’re in the kitchen,” and I know where they are when I hear
and understand “kitchen” in relation to the meaning of the other words.
In classical Indian philosophy, three theories compete to explain senten-
tial unity. There are two Mīmāṃsaka theories, one of which is embraced
by Nyāya, as well as a radically holistic theory propounded by the gram-
marian, Bhartṛhari (c. 450). Bhartṛhari holds that words have no meaning
outside the context of the sentence, which is the basic semantic unit. Words
are abstractions from sentences, and a sentence is understood holistically
“in a flash” (sphoṭa—Bhartṛhari’s theory is called sphoṭa-vāda). This is an
easy target for the Mīmāṃsakas, who point to our abilities to use the same
words in different sentences.24 But the one camp, the Prābhākara, agrees
with the grammarian that words do not convey meaning apart from the full
sentence being understood, that is to say, apart from the full fact indicated
being known “in a flash,” as it were. The other camp, the Bhāṭṭa, whose
theory comes to be taken over by Nyāya, claims that individual words have
reference in isolation, and that in understanding a sentence we understand
the meaning of the individual semantic units which get combined by the
sentence, by the fulfi llment of the syntactic expectancy along with the
meaning of the words, to mean the things denoted in relation. These two
views are termed in Sanskrit anvita-abhidhāna-vāda, “reference of the con-
nected,” which Mark Siderits translates as the “related designation view,”
and abhihita-anvaya-vāda, “connection of the referents,” which Siderits
translates as the “words-plus-relation view.”25
To translate may not be, as in the French expression, to betray (de traduir
est de trahir), but it is surely to interpret, and, I’m afraid, Siderits’ translation
of the second Sanskrit term is badly misleading. The relation is not just an
additional element: it’s not words-plus-relation. There are only a few purely
logical and syntactically binding words in Sanskrit, only a few (mainly
connectives) that are just syncategorematic, since every word is inflected
and there is no need for prepositions, etc.26 Alternatively, we could say that
every word is unsaturated because no word, no single semantic unit, con-
veys the meaning of a sentence by itself alone independently of its relation
 Testimony 89
to at least one other unit. The main difference between the two views is
that the former insists that only a sentence successfully refers, not the indi-
vidual words of which a sentence is composed, whose meanings have to be
connected to one another in order for there to be reference (abhidhā, the
primary mode or power, śakti, of language); whereas the latter holds that
words do have reference individually but not to the connection of the things
mentioned, which is given by the sentence as a whole. In both cases, the fact
or object known by way of a sentence has constituents. On the second view,
the fact is the relatedness of the words’ referents as they are in the world, a
relatedness (anvaya) not indicated by a semantic unit. The connection is to
one another of the things referred to, a connection in the world which we
become aware of because of the order and connectedness (anvaya) of the
words. 27 And on the fi rst, there is, as far as I can tell, the same ontological
critter, i.e., the fact or object that is structured the way the words say but
designated only by a prior combination of meanings.
Siderits considers the Bhāṭṭa/Nyāya view far less worthy of probing than
its historical rival which he connects, creatively, with the Buddhist “exclu-
sion” (apoha) nominalist theory of meaning. But it seems to me that it is the
Prābhākara word meanings apart from the sentence that are mysterious, not
those of Nyāya, where word meaning is the familiar relation of reference, the
primary relation between language and reality according to both camps. Sid-
erits dismisses the Bhāṭṭa/Nyāya view as disregarding the context principle of
sentence meaning (a word’s meaning is determined by its role in a sentence)
as radically as does the Bhartṛharian sphoṭa-vāda holism disregard the com-
position principle (sentence meaning is composed out of word meaning), such
that according to the “words-plus-relation view” it is impossible to distin-
guish a veritable sentence from a word list.28 While Nyāya does dispute the
contention that a sentence refers to something over and above the referents
of the individual words, there is a particular relatedness among the things
denoted that is not designated by the individual semantic units. ‘Bessie’ nev-
ertheless refers to our cow and in countless constructible sentences that bring
out her relations to other things (or even identity to herself).29
Just what a word refers to is sometimes ambiguous not just apart from
sentential context, as Gautama recognizes, but within it. Still, we know
what the word means. We know that a speaker wants salt when S asks for
it even though in Sanskrit the word used for salt, ‘saindhava’, is a homonym
with a word that means horse. S’s intention to communicate p in such a
case is crucial to disambiguation in that S speaks in a context (prakaraṇa).30
Ordinarily, the overall context need not be taken into account, according
to New Nyāya philosophers, to ascertain the meaning of a sentence, which
has to meet only the three conditions of grammaticality, semantic fitting-
ness, and proper presentation, the second of which we will discuss forth-
with. But we do have to take into account the overall context—let us say
“speaker’s intention,” tātparya—Gaṅgeśa and company admit and even
stress, in some cases of ambiguity as also of figurative speech.
90 Nyāya Epistemology
The following three necessary conditions for a meaningful statement are
proposed and discussed not only in Nyāya but throughout the philosophi-
cal and grammatical literatures:31

(A) grammaticality, ākāṅkṣā

(B) semantic fittingness, yogyatā
(C) proper presentation (pronunciation and the like), āsatti

The fi rst we have already discussed. On the Nyāya view, in brief, it accom-
plishes the unity of the sentence though it does not constitute that unity. 32
The words in a sentence have their ākāṅkṣā, (grammatical) “expectancy,”
mutually satisfied in the completion of a sentence as a string of words. The
third condition I propose to ignore, though surely competence in proper
representation is a requirement for communication. The second, the notion
of semantic fittingness, is not well-known in the West, and I think it may
be able to clear up some of the complexities of the a priori as inherited
from early modern philosophy (a topic to be explored further at the end of
the next chapter). In any case, yogyatā is connected to the Nyāya theory
of figurative meaning, since although its violation is not necessary for trig-
gering indirect meaning as comprehended by H, sometimes it is the trigger.
The reason for this seems to be that we assume that a testifier is trying to
say something intelligible, as in the Gricean conversational maxim.33
A stock negative example: “The gardener is watering the plants with
fi re” (agninā siñcati). Watering cannot be done with fi re, and so the mean-
ings of the words do not fit together except possibly figuratively. Some
defi ne “semantic fittingness” in a positive fashion, but it seems easy to fi nd
counterexamples.34 Language has to be flexible so that we can report novel-
ties. There may well be a white leopard. Furthermore, we understand some-
thing when we understand a false statement. Otherwise, again, we would
not know where to look to determine its falsity, or truth, for that matter.
Gaṅgeśa says explicitly that false statements as well as statements of doubt
meet the requirement of semantic fittingness.35 Even statements that are
not just false but that we know are false can pass the semantic-fittingness
test, as, for example (in a quip by Arindam Chakrabarti36) the positions
of one’s opponents! For these and other reasons, Gaṅgeśa, for one, defi nes
yogyatā negatively as “absence of knowledge of a blocker (of testimonial
knowledge).”37 This shows a coherence tie. We cannot even understand tes-
timony way out of whack with what we already know.38
Now normally, with veridical testimonial knowledge not involving figu-
rative meaning where all three conditions are met, we do not notice the
grammaticality, etc., of the transmitting sentence. These factors have to be
present, but we do not have to be aware of them.39 For figurative meaning,
in contrast, we have to notice a blocker,40 which paradigmatically may be
thought of as a violation of yogyatā.41 We know that no one waters plants
with fi re, for instance. Examples of less severe misfit occur, as we shall see
 Testimony 91
in taking up now metaphor and indirect meaning. Violating yogyatā is not
the only way to trigger a second power of words.

“FIGURATIVE MEANING,” LAKṢAṆĀ

“Indirect indication” is one sense of the Sanskrit ‘lakṣaṇā’, which is also the
name of a second power of words beyond reference: “metonomy” or “figu-
rative meaning.” Nyāya admits just two śakti, “semantic powers,” denying
what aesthetes (alaṅkāra-śāstrins) call “suggestion,” dhvani, as a third—a
topic we’ll take up briefly later. Reference, abhidhā, “denotation,” is con-
sidered a direct, unmediated relation between word and thing, as pointed
out. A word directly picks out a qualificandum as qualified by a qualifier.
(Abstracting the qualifier gives us what we may call the word’s sense.) Now
the second power of words is also referential but indirectly so, mediated by
a conceptual link to what S intends to communicate, tātparya. A standard
example: “The cots are crying” where it is not the cots but the people in
them who are crying, so a speaker is saying. The word ‘cots’ here means
things that are not cots.
Clearly we know that cots do not cry, and so semantic fittingness is vio-
lated. However, in the case of “Bring in the sticks,” which means “Bring
in the wandering ascetics carrying walking sticks” and does not mean that
just the walking sticks should be brought in, the sticks nevertheless could
be brought in, and would be, normally, and so there is no semantic unfit-
tingness in consideration of the words’ direct denotation.42 Furthermore,
there is not an entire loss of meaning in that presumably the sticks would
be brought along by the ascetics as they were ushered in. Unlike ‘cots’ in
the previous example, there is retention of the original meaning while there
is also meant something other than the denotation of the word. But even
with ‘cots’ the word’s original meaning (abstractly or in general) is the key
to its figurative use, since it directs us in the right way, so to say, towards
the people.
All figurative meaning involves what Sibajiban Bhattacharyya calls a
chain relation, param-parā, minimally a two-place asymmetrical function,
reference to an ultimately indicated, lakṣya, the beloved’s face, for exam-
ple, by way of an indicating reference, śakya, the moon.43 In all cases, the
ultimate meaning is indirect, mediated by a concept—paradigmatically, a
universal as cognized, its saṃskāra or “trace” part of a person’s memory
bank—but also possibly by any of a whole slew of generalities, ideas of
qualifiers through which we cognize all the many things we are capable of
talking about. Figurative speech depends on words having a primary deno-
tative power in a general, repeatable form. As with the operation of anal-
ogy as a knowledge source, the meaning of the indicator word in figurative
speech may be said to trade on a word’s sense.
92 Nyāya Epistemology
This view is traceable all the way back to Gautama and the NyS, where
the theory on the table is that words mean fi rst and foremost individuals,
like ‘Bessie’ as the name of a cow. Gautama offers examples of words meant
in a figurative and indirect sense to refute such a nominalist view of word
meaning.44 Thereby he gives one of the earliest lists of relations implicated
in figurative speech, association, location, purpose, and so on.45 Gautama’s
point, however, is that if words were exclusively names of individuals, figu-
rative speech would be impossible, since no matter how rich one’s ontology
(admitting, for example, abstract individuals such as cowhood as what is
named by ‘cow’), in figurative speech the normal referent is not what is
meant, not the moon itself, for example, but rather the face of the beloved
being rhapsodized.
Among New Nyāya philosophers, it is, as mentioned, not just violation
of semantic fittingness (yogyatā) that sparks a hearer to understand a word
in a figurative sense. There are plenty of examples recognized of meta-
phors where the words make sense taking them strictly in their denota-
tive meaning. Now it is words that have two “powers” of meaning (śakti),
not sentences, but the second power of figurative meaning presupposes at
a minimum a sentential context. Words’ power or potency of figurative
meaning is activated only in use. The traditional view, from which Gaṅgeśa
demurs but only because of a need for qualification, is that anvaya must
be impossible, which is the overall semantic and grammatical “connected-
ness” of meaning achieved by a sentence over and above the meanings of
the constituent words.46
For example, the word ‘Gaṅgā’ in the locative case (in Sanskrit) convey-
ing the sense of being on or in the midst of the river cannot be construed
denotatively as used in “The village is in the Gaṅgā,” since we know that
no village is actually in the water. This is about the same as violation of
semantic fittingness, and is also incapable of handling what is called ajahal-
lakṣaṇā (metaphor where the original sense of the indicator word is not
abandoned), where the sticks, e.g., not only could be brought, “Bring in
the sticks,” but would be brought, too, as we discussed. (Perhaps a better
example: “The umbrella-bearers are passing” where it is meant that the
people being shaded by umbrellas borne by others as well as the bearers
themselves are passing.) Here it seems we have to talk about H’s knowl-
edge of what S intends, which H grasps from the context (prakaraṇa), or
at least admit that context is sometimes crucial for the precise sense of a
word. For example, by context we know that the verb “bring in” in this
instance means “usher in politely” such that there would be a violation of
the possible relatedness of the words denotatively since you do not usher
in sticks.47 Annaṃbhaṭṭa (seventeenth century) argues that such examples
along with words with more than one sense (such as ‘saindhava’, “salt”
and “horse”) demand that the theory of the trigger for figurative speech
as the impossibility of anvaya, connectedness, be jettisoned in favor of the
view that it is the impossibility of the speaker’s intention, tātparya, that
 Testimony 93
triggers testimonial knowledge in cases where the communicating sentence
contains a word used figuratively.48
We shall move on just a bit below to examine New Nyāya’s notion of
tātparya, “(speaker’s) intention,” and return to a topic closely related to
figurative meaning at the end of the next section, dhvani or vyañjanā, a
proposed third power of words not accepted by Nyāya philosophers, “sug-
gestion.” It turns out that determination of speaker’s intention is crucial to
Nyāya’s dismissal of suggestion as an additional word śakti. As we shall
see, putative instances are explained as the working of inference and/or
testimony analyzed just in terms of the two powers of meaning. It is appro-
priate now, however, to consider one more subtopic regarding figurative
meaning, which looms large in the late philosophy of language of the “aes-
thetes” (alaṅkāra-śāstrins) as well as with luminaries in Vedānta and not
just in Nyāya and Mīmāṃsā.
Recall the division of types of word according to whether the reference
is only by convention (rūḍhi, “stipulation”), only by derivation (yoga, “[ety-
mological] synthesis”), or by a combination of both (yoga-rūḍhi), which we
discussed in the previous chapter (p. 76). Lots of expressions are what we call
in English “dead metaphors.” In Sanskrit, a stock example (which is a little
humorous) is the word ‘maṇḍapa’ (or ‘maṇḍa-pa’, the dash showing that the
word is a compound). By convention the word means an ascetic and by deri-
vation (combining the sense “alcoholic scum of boiled grain,” maṇḍa, with
“drinking,” -pa, thus) “scum-drinker.” At one time the term may have been
used in a figurative sense for an ascetic (maybe even literally), but convention
trumps etymology, Nyāya philosophers argue, and now the word is used to
mean an ascetic denotatively, Gaṅgeśa reports, without a derogatory sense.49
Its second power could be elicited, but it itself is a dead metaphor (or “faded
metaphor,” nirūḍha-lakṣaṇā, in the rendering of Kunjunni Raja50), which
could be re-enlivened but only by some cue, some indication that S has not
spoken using words only with direct reference.

SPEAKER’S INTENTION

Nyāya breaks with Mīmāṃsā over the nature of Vedic authority. The Nyāya
position is that, like all sentences, those of the Veda should be understood
as spoken (or composed, etc.) by someone with an intention to commu-
nicate. The Mīmāṃsā view is that the words of the Veda are not spoken
by anyone originally (that speakers are subject to error would undermine
Vedic authority seems to be the idea). The verses are instead simply chanted
afresh by each new generation of memorizers. Now Nyāya philosophers
are theists, with the possible exception of Gautama, the sūtra-kāra, and
the later writers show more or less enthusiasm for Nyāya’s, let us say, mini-
malist theology. To be sure, the mainstream position from Vātsyāyana on
is that God (the Lord, īśvara) is the author of the Veda, but this is not a
94 Nyāya Epistemology
very central tenet. Theistic ideas, apart from Udayana’s treatise devoted
to proofs of God’s existence, get, all told, small play in Nyāya literature.
Nevertheless, it bears repeating, that God is the author of the Veda is the
mainstream opinion. Mīmāṃsā, in contrast, is atheistic, viewing the Veda
as primordial, and heard, memorized, and passed down from generation
to generation immemorially. 51 It has no author, no beginning. And it is at
the center of the rituals and sacrifices that Mīmāṃsā takes itself to defend,
rituals that have themselves been practiced immemorially. My point is that
Mīmāṃsā comes to the debate about speaker’s intention with an axe to
grind: sentences can be meaningful without having a speaker with an inten-
tion to communicate. Nyāya generalizes from the everyday to assume that,
no, statements and sentences require a speaker, a composer, who in the case
of S’s knowledge passing to H, must be an āpta, i.e., someone who knows
the truth and wants to tell it without deception.
However, we have to be able to understand a spoken sentence to be able
to determine a speaker’s intention, tātparya, which we infer from what is
said supplemented sometimes by contextual cues. Thus knowing the inten-
tion is not an invariable antecedent of testimonial knowledge, says Gaṅgeśa
who is followed here by Annaṃbhaṭṭa and other New Naiyāyikas.52 Under-
standing S’s intention is not a fourth condition on a statement’s meaning-
fulness from the perspective of H—except in some cases of ambiguity and
figurative speech. But in those cases it is indeed crucial, and there is no way
to get around the need to make it out in order to fi x the meaning, which
cannot be gathered at, so to say, a fi rst pass.
But on a second pass, we are able to gather not only indirect meaning but
also more information through inference. In this way, Annaṃbhaṭṭa would
explain what others see as the results of the activation of an alleged third
power of words, namely, dhvani, also called vyañjanā, “suggestion.”53 In
other words, if a sentence contains an ambiguous word or figurative mean-
ing, there may well be no way to tell what it means without considering S’s
intention.54 Now advocates of the third power analyze the stock example
of the village said to be in the Gaṅgā as suggesting that the village has
a meditative atmosphere by association with the sacred river. The whole
point, they argue, of poetic use of figurative speech is to release the third
power of suggestion. Why otherwise not simply say that the village is on
the bank? The speaker uses the figurative speech to suggest the attributes of
coolness and purity. Annaṃbhaṭṭa responds that if one understands from
the statement these attributes then the figurative meaning (lakṣaṇā) of “in
the Gaṅgā” is not just being on the bank of the river but on a bank that
lends coolness and inspires meditation and clean living. This is then just a
more complex case of lakṣaṇā, indicating a cool and purifying location on
the indicated bank (lakṣya).
There is a different kind of putative suggestion that its advocates (in the
so-called dhvani school of alaṅkāra-śāstra) say is not sparked through figu-
rative usage. A complex example of this type is indeed cited by the founder
 Testimony 95
of the dhvani school, Ānandavārdhana, to establish the existence of the
third power of speech near the beginning of his magnum opus.55 But let us
look at a simpler but similar example provided by Annaṃbhaṭṭa’s commen-
tator Nīlakaṇṭha of a verse spoken by a loving wife to her husband upon
his imminent departure (slightly modifying Gopinath Bhattacharya’s trans-
lation): “Go, go, if thou must, my beloved. May thy paths be happy. And
may I be born again just where thou hast gone.”56 Nīlakaṇṭha says that the
suggested meaning is that the husband should not leave (“Beloved, my life
will be over if you depart. So please don’t go”). There is, however, no word
used figuratively to generate this meaning. There is a violation of speaker’s
intention assuming that the wife does not really want to be reborn at the
distant place. But there is no word used figuratively that could generate the
suggested sense. Annaṃbhaṭṭa handles the case by claiming that we, and
the husband, do not grasp the meaning at fi rst pass. He gets it, however, on
a second pass from inference. He reasons that the wife would not have men-
tioned rebirth to be near him had she not wanted him not to go. However,
Nīlakaṇṭha apparently has sympathy, as Professor Bhattacharya points out,
for the position that there is operative here a third power of words: “It is to
be noted, however, that if people actually feel, in an instance like the pres-
ent one, that they do have an immediate cognition of the suggested sense
without the intervention of any inferential relation, then it would be impos-
sible even for the preceptor of the gods to repudiate it.”57
Personally, I must say I share Nīlakaṇṭha’s sympathies. There is, after
all, a significant bulk of literature, on the side of both poets and critics, in
which dhvani is aimed at or is the principal analytic tool. It seems petty for
Nyāya philosophers to try to reduce what many—and many great minds,
such as Abhinava Gupta, Viśvanātha, Jagannātha, the list is long and dis-
tinguished—see as the purpose both of poetry with its figurative speech
and of criticism. More precisely, the dominant view is that suggestion is
the means to aesthetic experience, rasa, which is the goal, or soul (ātman),
as Ānandavardhana says, of poetry.58 It is the critic’s business to reveal
the dhvani and help the reader/hearer grasp it so to share the poetic emo-
tion. Perhaps down deep the Nyāya attitude is that the result is not often
testimonial knowledge. But Annaṃbhaṭṭa may be taken to speak for the
school when he admits there is knowledge, fi rst testimonial knowledge and
then inferential, that seems to trade on a suggestive sense. Nevertheless,
he would explain it away as a function of inference and words’ power of
indirect indication, as we have seen.
7 Lessons for Analytic Epistemology

The number-one lesson from our study for contemporary epistemology

should be rather obvious: classical Indian philosophy includes a rich tradi-
tion of thinking about knowledge, which Nyāya represents in its own real-
ist and common-sense fashion. Hopefully soon Nyāya, Mīmāṃsā-Vedānta,
and Buddhist Yogācāra will take their rightful places as epistemological
resources for students of philosophy. A sign of this will be integration of the
study of the classical Indian schools more broadly into graduate Philosophy
curricula maybe not wherever analytic philosophy is taught but in many
places. There is enough in common between the traditions that it is easy
to see how in principle bringing to bear Nyāya’s views and arguments on
issues of analytic epistemology could well bear fruit.
I should like to spend the rest of this chapter on specific lessons con-
nected to strengths and weaknesses of the Nyāya position as I see them. Let
us start with a common and basic objection to externalism, the so-called
causal fallacy, in a version due to Keith Lehrer, who claims that external-
ists mistake causes of beliefs for reasons.1 Raco, as imagined by Lehrer, is a
racist who initially comes to believe that p (that Raco’s disliked racial group
is more susceptible to some particular disease than is Raco’s own group-
ing) because of his racism which not only generates his belief that p but
implants it with absolute and unshakeable conviction. Later Raco becomes
a doctor and scientist and familiar with real evidence that p. Thus, accord-
ing to Lehrer, Raco is justified in believing that p having reasons that are
entirely distinct from the cause of his belief. Personally, I am surprised at
the sway of the case, since it seems easy to dismiss as a problem, at least for
Nyāya. Causes, or citings of causes, become reasons, not invariably good
ones, when we step back and wonder whether a belief is true. Animals
and children have knowledge without reasons, not reasons at least that
they are aware of as reasons, as do we all, according to Nyāya. To stretch
the meaning of “reason” to include motivations for animal acts such that
there would be animal reasons, the point would remain that we have unre-
flective knowledge where we are not aware of the reasons for our beliefs
as reasons that we could cite. The giving of reasons belongs to reflective
knowledge, where we see causes as reasons. My wife is self-consciously
 Lessons for Analytic Epistemology 97
justified in believing that my glasses are in the kitchen if, challenged by me,
she cites the cause of her belief as her perception, “I am staring at them.”
Raco, by Nyāya’s lights, does not know that p because his belief that p is
not pramāṇa-born, and it seems to me to be a strength of the Nyāya analy-
sis that even though Raco can give good and decisive reasons for believing
that p, reasons that he does in fact believe, intuitively we say he does not
know. His belief that p is not sensitive to the fact that p (it does not “track
the truth,” to use Robert Nozick’s test) in that he would still believe that
p even if p were false and the scientific evidence pointed to ~p. Thus this
is shown to be a great case to reveal, by suppositional reasoning (tarka),
that the having of good reasons is not sufficient for the kind of justification
needed for knowledge. Dr. Raco does not know, because his belief has not
been generated in the right way. It is not objectively justified. If he tran-
scended his racism and came to believe p on the basis of scientific evidence,
then we would say that formerly, as having a belief forged by racism, he did
not know that p whereas later, in believing that p because of the evidence,
he does indeed know.
Similar considerations take care of the “Truetemp” case due to David
Armstrong and Keith Lehrer along with the case of the irrational clairvoy-
ants Amanda and Bertha discussed as counterexamples to reliabilism and
all externalist attempts to explain epistemic or subjective justification by
Laurence BonJour. 2 A subject S who lives in New York is, unbeknownst to
her, wired to a weather station in Sydney in such a way that on the hour it
occurs to her that the temperature is such and such in Sydney and in fact
she is always right. The belief-generating process is thoroughly reliable, in
fact infallible, like Nyāya’s knowledge sources, and involves a causal chain
between the temperature in Sydney and S’s belief. Our intuition is that this
is not a case of knowledge, but it seems that according to the externalism
of Nyāya it would be.
This line of thought fails, however, to take into account key theses in
Nyāya’s understanding of a knowledge source, in particular, that a source
is a type of process that invariably generates true beliefs, not a one-time,
unlawlike generating of a belief that p from the fact that p, and that the
results of the several sources can be apperceptively identified as percep-
tual, inferential, and so on. Thus by Nyāya’s lights too much about the
True-temp case remains obscure. If, for example, there is to be a true-belief
generator whose operation could be detected, then the subject would have
knowledge. That is, if, say, S in New York needed to know what the tem-
perature in Sydney was and certified her knowledge (perhaps by telephon-
ing someone in Sydney) she presumably would come quite reasonably to her
hourly readings. And she would have knowledge indeed. If, on the other
hand, all that S is aware of is the thought that the temperature is such
and such in Sydney—if she were only entertaining it, in Fregean fashion,
akin to what Nyāya calls āhārya-buddhi, “fantasy, imagination,” which
is not considered a “presentational experience,” anubhava—S could not,
98 Nyāya Epistemology
by Nyāya’s lights, be said to believe it. We can tell the difference, apper-
ceptively, between perception and day-dreaming, and we don’t trust the
latter to guide us in action. Cognitions of fictions where it is known that
the objects are not real are deemed āhārya-buddhi, “imagination” (see the
chart, p. 29). A person does not act on the basis of imagination, except,
perhaps, in a play. Trust is built on feedback, and it is relevant to what we
do. Again, if S did come to trust her Sydney-temperature cognition—that
is to say, if she really had belief—she would have knowledge, since we
are imagining a discoverable knowledge source (S could discover the con-
nection or be led to assume (rightly) that there is something like wiring
connecting her brain to Sydney). According to Nyāya, all that is required
for unreflective knowledge (pramā) is a lawful tie to the object cognized.
(Nyāya’s acceptance of analogy as a knowledge source shows particularly
well that the school is open, in principle, to acceptance of sources beyond
the traditional four, as I argued.) As in the case of perception, it would not
matter whether the knowledge source, the generative process, had been
discovered by S in fact. This is especially true of details of its functioning.
We don’t know precisely how sight works, all its mechanisms, so to say. But
sight is a knowledge source, and we can recognize it. S could have reflec-
tive knowledge (nirṇaya) however here, too. What matters for reflective
knowledge is that we understand a source’s operation well enough to dis-
cern “excellences” and “deficiencies” (guṇa and doṣa) as we have discussed.
Imaginably, this S could do.
These reflections should be sufficient to provide the direction for a
response to other cases taken to challenge views similar to Nyāya’s, for
example the clairvoyants Amanda and Bertha who are dreamed up by Lau-
rence BonJour expressly to fit reliabilist accounts of justification but who
intuitively are irrational.3 BonJour imagines that Amanda believes that p
by clairvoyance although she has no evidence whatsoever that clairvoy-
ance is reliable (we are to imagine that for reasons unknown to science it is
100% reliable, again like Nyāya’s pramāṇa). Unlike Bertha, Amanda also
has evidence that the faculty is unreliable (from client feedback, television
reportage, etc.). Bertha has no evidence of either kind, positive or negative,
about the reliability of her clairvoyance, whereas, again, Amanda lacks
positive and has negative evidence. Clearly, Amanda should not believe the
doxastic output of her clairvoyance, despite its reliability. Bertha, BonJour
argues, shows the inadequacy of even the weaker externalism that would
strengthen the requirements for justification beyond merely being the out-
put of a reliable doxastic practice (unqualified externalism) to include S’s
not having reasons for thinking that p is false (there should be no known
defeaters for p, a condition that Bertha meets), a view BonJour designates
qualified externalism. In other words, Bertha lacks positive evidence for
the reliability of her clairvoyance, but, unlike Amanda, also has no evi-
dence that the beliefs it generates are false. Intuitively, Bertha, too, seems
irrational since the reliability of her only source for her belief that p is not
 Lessons for Analytic Epistemology 99
known to her (the terms of the thought experiment dictating that there is
no other source for her belief).
Notice that BonJour stipulates that the clairvoyance’s reliability “is
not cognitively accessible to the believer in question.” This means that,
by Nyāya’s lights, the belief that p could indeed not be certified. Thus,
though true, it would fall easily to a challenge. There is a social dimension
to knowledge. Amanda would be irrational and not have knowledge if she
continued in her belief that p after being faced with negative evidence con-
cerning the source of her conviction. And she would not have knowledge,
even at the fi rst level after the challenge, according to Nyāya. The case
with Bertha is different. In hindsight after, let us imagine, our discovering
the truth-guaranteeing nature of clairvoyance, we might say that she had
unreflective knowledge (pramā) that p since, by the terms of the thought
experiment, the belief faced no defeaters. No reflective knowledge (nirṇaya)
would have been possible, admittedly, at the time for Bertha, as stipulated.
But if we discovered that clairvoyance is a knowledge source, then I think
that we would say that she had knowledge.
Before moving on to a different kind of objection to externalism, the
so-called new evil-demon problem, let us look at two other well-known
cases alleged to be obstacles to externalist accounts of knowledge and jus-
tification. Robert Nozick imagines a grandmother believing (that p) that
her granddaughter is alive and well when she is standing in front her, in a
case that intuitively we would call knowledge but that does not appear to
meet Nozick’s “tracking” standard: If p were not true, S would not believe
that p. Belief that p, Nozick argues, has to be sensitive to the fact that
p if someone has knowledge. It has to vary appropriately as we imagine
possibilities where p is true or false. And here the grandmother’s belief
does not track the truth. This would be because, as Nozick imagines the
case, that everyone but the grandmother knows that she has only weeks
to live and no one would want to upset her by telling her that her grand-
daughter was not alive and well. If the granddaughter were not alive and
well, the grandmother would belief it by way of the inevitable lies. Such
cases force Nozick to “relativize to a method,” as he says, such that the
grandmother’s knowledge would pass the revised test: If p were not true, S
would not believe that p by method M.4 Nozick then proceeds to talk a lot
about “ways and methods” of knowledge. However, much of what he says
is vague and general, not specifying wide-ranging knowledge sources. In
any case, Nyāya handles Nozick’s grandmother case as well as or arguably
better than Nozick himself. The grandmother has knowledge by perception
that the granddaughter is alive and well when she is standing in front of
her. But the grandmother would not have testimonial knowledge that the
granddaughter is alive and well even when she is since her putative source
would not be veritable testimony as a pramāṇa. The testifiers would not be
trustworthy when it comes to this matter (not āpta). They would deceive
her if p were false.5
100 Nyāya Epistemology
Consider, fi nally, the case of the tourist who perceives a real barn and
forms the appropriate belief but who does not know that the barn is real
because the countryside contains many papier-mâché barns that the subject
would have mistaken for a real barn.6 Seeing real barns which are inter-
spersed with fake barns that he takes to be real, S has perceptions that are
not knowledge-generators in that environment. When S says, expressing
his perceptual belief, “Now that’s a fi ne barn,” not only is he right, we
may imagine, but there seems to be absolutely nothing abnormal in the
perceptual belief-generating process that leads him to say what he does.
He’s right and he appears to be justified, but intuitively we say he does not
have knowledge. Furthermore, such cases do not involve a false premise or
apparently any flaw in the belief-generating process and thus present a spe-
cial challenge to Nyāya. The challenge is, I think, nonetheless answerable.
Here I begin to go beyond the classical literature as well as our discussion
in Chapter 1 of a Gettier case of a pseudo-inference to the true conclu-
sion of fi re on a mountain where there is indeed fi re (discussed earlier, p.
13). That case involved a false perceptual premise, a mistaking of dust by
the subject for smoke. Thus the fake-barn case, which does not involve a
false premise, does, I repeat, present a special problem for Nyāya’s analy-
sis. Intuitively, there is no knowledge since S could easily have had a false
belief with the same objecthood or intentionality, that is, had he looked
at the countryside just a moment earlier or later than his lucky glance.
The problem for Nyāya is that p (“Now that’s a fi ne barn”) appears to be
“knowledge-source-generated.”
However, there are resources in Nyāya to deny this, in (a) environmental
conditions required for perception to be genuine and (b) the role assigned
to memory in perceptual experience. Being in an environment where so
many barns are fake could be identified as undermining the operation of
perception with respect to barns and like constructions. It’s a rule for sight
(this is an often cited epistemic “deficiency”) that a middle-sized object
cannot be too far away from the perceiver if its height is to be perceived. So
also, it would seem, for perception of Fs in an environment where so many
F-looking things are not Fs.
Then, slightly changing p and the statement to “Now that’s another
fi ne barn,” or “Now that’s a fi ne barn (compared to others we have been
seeing),” shows that there is a memory-related deficiency. Only a fi rst-time
perception of something as an F is free from the “whiff” of previous experi-
ences of Fs, says Nyāya. A cognitive process misinformed by memory is not
perception as a pramāṇa.
In sum, there are two identifiable deficiencies ruling out genuine percep-
tion in this and similar cases, one in the environment and another in the
subject’s memory. All this accords, moreover, with the “suppositional rea-
soning,” tarka, a Nyāya philosopher is taught to take with such cases: that
if S knew so many barns were fake he would not trust the “perceptual” evi-
dence. The “method” as described from S’s own point of view at the time he
 Lessons for Analytic Epistemology 101
is being deluded may that one time hit the truth, but there is no knowledge
source in fact, no possibility of genuine certification, and no knowledge
even at the fi rst level since there is not the right relation to fact.7
Such a case as this one about fake barns does point up Nyāya’s “generality
problem,” however. The Nyāya response, as I have imagined it, seems ad hoc,
and Nyāya authors do not make up enough subrules for types of perception,
etc. We discussed the generality problem in previous chapters, and I have lit-
tle more to say now than to repeat that the difficultly is faced up to by Nyāya
philosophers. Naiyāyikas specify a few rules, as pointed out, for subspecies
of perception in particular according to the type of object known. Perhaps
the best reply here is, again, that some certification conditions are very gen-
eral (sense-organ/object connection, for instance) and we do not always need
a lot of specificity. Nyāya philosophers tend to focus on the most general and
easy to recognize. But also they clearly are not averse to proposals of specific
rules targeting specific kinds of object, as pointed out.
To move on to the new evil-demon problem, nothing there embarrasses
Nyāya, it seems to me, since in the evil-demon world there would be “pseu-
do-justification,” prāmāṇya-ābhāsa, justification solely from a first-person
point of view, not objective justification which entails truth (discussed earlier,
pp. 13, 135n24, and 136n27). In a possible world where an evil demon is
responsible for all of our experiences (except possibly self-perception) and
where almost all of our beliefs are false, a subject may still be justified in her
beliefs if she is a responsible believer, believing according to the evidence.
However, by externalist accounts of justification—including Pollock’s under-
standing of objective justification which entails truth—in the demon world
there is no justification since factual ties are missing. Nyāya would say that
indeed a responsible subject in the demon world would appear to be justified
with a belief that could not be certified genuinely since it is false. There would
be nevertheless a type of “apparent justification” according to Nyāya, and
that there would be captures our intuitions, it seems to me. There’s a big dif-
ference between being genuinely justified entailing that a belief in question is
true and only appearing to be justified, although this is a point that demands
an ability to appreciate an outside, not just a first-person, point of view.
In sum, the Nyāya analysis of knowledge and justification appears able
to account for cases taken to be unexplainable satisfactorily by similar
Western views. Perhaps when Nyāya is better known, other cases will be
imagined that resist its analysis. However, I know of none. A weakness
is that there are no formula in Nyāya to determine degrees of warrant in
situations where we might like them. Another is that the question of how
disagreement impacts justification is not entirely clear, the epistemology
of conflicting reasons and evidence. However, Nyāya has developed rich
resources in its theory of the subtle ways the scales of justification may be
tipped by unfavorable or favorable tarka, “suppositional reasoning,” the
drawing out of untoward consequences, considerations of simplicity, and
so on, as we discussed in Chapter 2 (pp. 30–32).8
102 Nyāya Epistemology
Finally, I should like to take up briefly what seems to me the greatest
deficiency of the Nyāya approach, which concerns a priori knowledge, for
some types of which Nyāya simply comes up blank. The a priori is not
a Nyāya category, as we have discussed (the closest equivalent in classi-
cal Indian philosophy being probably the later Buddhist concept of “inter-
nal pervasion,” antar-vyāpti, p. 163). We have mathematical and logical
knowledge which Nyāya’s sources seem incapable of generating and pos-
sibly some other types of knowledge a priori as well.
The reason for the lack may be as simple as Nyāya philosophers’ not
countenancing knowledge de dicto but only knowledge de re, although
arithmetic was well-known and the mathematics employed in astrologi-
cal calculations quite sophisticated.9 For Nyāya, knowledge is always of
intersubjective objects, even when these are supervenient, mind-imposed
properties (upādhi), or internal objects, such as desires and emotions
which, consonant with its realism, Nyāya views as by others also know-
able, “Chaitra is angry,” for instance. On the other hand, the ontology
of numbers is an important topic addressed in early Vaiśeṣika literature,
and late Naiyāyikas led by Raghunātha (sixteenth century) make number a
separate category. The relation (paryāpti) whereby it qualifies a property-
bearer or bearers (which can be other properties, such as colors, as well as
substances) is distinct from inherence, in that numbers qualify collections
as wholes whereas properties that inhere in a group of things qualify dis-
tributively (five mangos are not each qualified by the number five but each
individual mango has sweetness).10 However, counting is the only matter of
arithmetic treated epistemologically that I am aware of, and quite amaz-
ingly it is viewed empirically, as a matter of experience, in that an nth num-
ber is considered a mind-imposed result (upādhi) that is dependent upon a
preceding cognition apperceived (i.e., known by internal perception) which
informs the latter by its having n-1 as its qualifier portion or predication
content (viśeṣaṇa), back through a chain of cognitions, each informing its
successor, to the initial cognition in the series that grasps something as pos-
sessing the number one. Other types of a priori knowledge would also be
parsed differently from Western ways.
Quassim Cassam, rehearsing a disagreement between Tyler Burge and
Alvin Goldman over how to defi ne the a priori, writes: “These consider-
ations suggest that what it is and is not acceptable to count as an a priori
warrant is partly a matter of tradition.”11 Taking an overview of Nyāya
underscores the point. By Cassam’s lights, Nyāya’s inferential knowledge
and justification (warrant) count as a priori since they are non-perceptual
although in the West what he calls memory warrant and introspective war-
rant are, he says, non-perceptual but not usually recognized as a priori.12
Reasoning to the conclusion that nothing that is red all over could be at
the same time green all over would not be judged knowledge by Nyāya,
since “Nothing that is red all over could be green all over at the same time”
would not express a presentational experience, anubhava, but would be a
 Lessons for Analytic Epistemology 103
matter of knowing the meaning of words. “The flag that is entirely green
now is red-all-over” violates semantic fittingness, yogyatā, one of three
conditions identified for sentential meaning (pp. 90–91). If we know the
meaning of ‘red’, ‘green’, ‘all over’, etc., we are able to recognize a violation
of semantic fittingness, though I should think Nyāya philosophers would
laugh at the idea that by reflecting carefully about what it is to be red all
over we arrive at the knowledge that something that is red all over is not
green all over. It’s hard to imagine a pakṣa, a veritable subject for the infer-
ence, since that requires our having an interest in whether it possesses the
probandum property.
There are nevertheless “negative-only” inferences that, as we have seen,
though quite abstract and philosophic in nature, are thought to generate
important, factual conclusions. And such inferential knowledge is not a
matter simply of knowing the meaning of words. Knowledge of mean-
ings may be primarily a matter of testimony supplemented by analogy, but
when it comes to categorial slices and defi nitions of fundamental types of
thing, negative-only inference, which seems a lot like a priori reflection as
touted in the West, is employed, as we discussed, which would generate
new knowledge. Perhaps the most interesting feature of this kind of reason-
ing in Nyāya is that what is produced is by its lights not knowledge of ideas
but of facts.
The division between conceptual and factual knowledge in not sharp in
Nyāya. Meaning is thought of as reference, even word meaning, although
in their own way Nyāya philosophers also recognize sense (pp. 78–80 and
87). There is no problem about perception being the way a subject comes
to know that there is no sprinkling with fi re and that the sentence, “The
gardener is sprinkling the plants with fi re,” is either meaningless as violat-
ing yogyatā or an instance of indirect, figurative speech. Nyāya’s thorough
commitment to realism motivates its extreme empiricism which fi nds only
perception as a direct link to the world of mind-independent objects, with,
it bears repeating, the other sources as dependent on perception in some
way. Veridical memory, in turn, is dependent on occurrent knowledge from
perception or another source, and so memory warrant—as also introspec-
tion, i.e., apperception with Nyāya—is not strictly, pace Burge, Cassam,
and others, non-perceptual. Nyāya is in principle not opposed to the admis-
sion of another source, as I have many times stressed. But admission of a
fundamental tie to the world other than perception would violate the spirit
of the school. Of course, inference, analogy, and testimony are processes
that are different from perception albeit they are informed by it. Thus it
is no mystery how justification that is “non-perceptual” (a priori in that
sense) can be justification for a belief about a worldly thing.
 Appendix
The Analogy Chapter of Gaṅgeśa’s
“(Wish-Fulfilling) Jewel of Reflection
on the Truth (about Epistemology)”
Tattva-cintā-maṇi

INTRODUCTION

Gaṅgeśa (Nyāya, fourteenth century, Mithila) uses a dialogic structure for the
Tattva-cintā-maṇi (TCM), his masterpiece of epistemology, although except
rarely he does not provide his interlocutors’ names. He does so in only two
places in this short chapter, on analogy (upamāna), which is the third of four
chapters, each devoted to examination of a “knowledge source,” pramāṇa:
perception, inference, analogy, and testimony in that order. Gaṅgeśa’s dia-
logic style has been addressed by me in several places, but it is worth again
saying a few words—especially about the nature of philosophic pūrva-pakṣa,
the prior or prima facie position or an objection, in relation to an author’s
own views, the siddhānta, the final, established position which answers
the pūrva-pakṣa. Readers have to be acutely mindful of this textual divi-
sion including pūrva-pakṣa within pūrva-pakṣa and sometimes another level
down. Without appreciating the dialogical structure of Gaṅgeśa’s text, no
one could understand more than a few isolated sentences.
A pūrva-pakṣa is topically unified exposition, complete with supporting
arguments, of an opposed position or of an attack relative to a siddhānta,
which is itself unified exposition of an accepted position, complete with sup-
porting arguments and/or correlate responses. Gaṅgeśa probably expected
his immediate audience to be able to identify proponents and practically hear
the voices of individual advocates. I use emboldened labelling to bring out
the players in the dialogue, sometimes no more specific than a nondescript
“Opponent” versus “Gaṅgeśa” expressing his own view, or an objector to a
first or principal objector, labelled then “Opponent.2” versus “Opponent.1.”
With the emboldening of Gaṅgeśa’s own name, however, probably I am a
little over-bold, slightly distorting the text. Let me explain.
Throughout the TCM, Gaṅgeśa uses expressions such as mama tu, “My
view, in contrast,” as discourse markers, a practice that may be general-
ized, attributing siddhānta sections to Gaṅgeśa himself. But such ascription
does not suggest as much as it should. A siddhānta is normally more than
Gaṅgeśa’s own view; it is supposed to be the view of Nyāya tradition, a
view that Gaṅgeśa is interpreting as well as asserting anew (atra brūmaḥ
106 Appendix
also occurs as a discourse marker: “To this, we respond”). Furthermore, a
siddhānta indicator is not meant to introduce a perspective or way things
might be, but a view arrived at by careful considerations and asserted as
true. It is reflective knowledge, nirṇaya. The propositions that comprise a
siddhānta are asserted both as true and as defi nitive of Nyāya even when
they are original. The emboldened labelling of Gaṅgeśa’s name is, then,
best understood as signaling text where Gaṅgeśa gives his own views while
taking himself to speak for the school. To be sure, rival Naiyāyika posi-
tions are sometimes aired, and often he distinguishes his contemporary or
“New” (navya) Nyāya from that of “Old” Nyāya, as he does in one place
in this chapter, even mentioning one philosopher of Old Nyāya by name
(Jayanta Bhaṭṭa). Nevertheless, siddhāntas are asserted as expressing the
positions of a school, not only of an individual philosopher.
The Prābhākara Mīmāṃsaka is the principal opponent here, as the com-
mentators tell us, and throughout the TCM, although many other posi-
tions are taken up. In my opinion, the pūrva-pakṣin interlocutor would
be better thought of as a Naiyāyika with Prābhākara leanings, but I defer
to tradition. Note that he begins by stating the principle that the truth of
everyday patterns of speech (vyavahāra) is to be accepted unless there are
defeaters. In other words, he says that the cognitive default is acceptance
and that the burden of proof is on those who would deny. This is the posi-
tion not only of the Prābhākara, but of Nyāya and practically the whole of
classical Indian philosophy.
The text is taken from the Gaurinath Sastri’s edition of the anal-
ogy chapter of the Tattva-cintā-maṇi with the Pragalbhī commentary by
Pragalbhācārya (Varanasi 1983). There is also an older Calcutta Asiatic
Society edition (1990 reprint), edited by Kamakhanath Tarkavagish, with
the Dīpanī commentary by Kṛṣṇakānta Vidyāvāgīśa. No attempt has been
made to edit the text critically or systematically to compare the two pub-
lished editions. But in a few places I have followed a clearly preferable
phrase of the Calcutta edition, marking the fact in a note. Both Sanskrit
commentaries have proved helpful, particularly Kṛṣṇakānta’s. Typically
ellipses have been supplied following his lead (iti śeṣaḥ), although at places
Pragalbha’s commentary has also proved absolutely crucial for me.
Chapter 5 in this book may serve as an introduction to the concepts
employed by Gaṅgeśa here. Let me add only that it is fortunate for us, or
at least for me as a translator, that the semantics, or grammar in a broad
sense, of ‘similarity’ in English is such that it is easy to map the grammar
of the word ‘sādṛśya’ in Sanskrit. This is particularly evident in the points
made about “conventional usage” (below, Gaṅgeśa’s response to the New
Prābhākaras, p. 115, for instance).
 gaṅgeśopādhyāya-viracitas
tattva-cintā-manau upamāna-khaṇḍaḥ
From the Tattva-cintā-maṇi
“(Wish-Fulfilling) Jewel of Reflection
on the Truth (about Epistemology)”
The Analogy Chapter
By Gaṅgeśa

[Text: p. 1, Varanasi, and p. 1, Calcutta]

Gaṅgeśa: Now analogy is explained. Here some say that analogy is the
source of knowledge of similarity.
Opponent.1 (a Prābhākara Mīmāṃsaka): Similarity is another category
(over and above those recognized by Nyāya and Vaiśeṣika). More spe-
cifically, similarity is (a distinct category) because everyday speech
about it is undefeated. And this is not a single thing throughout the
categories since that supposition entails the unfortunate consequences
that everything would be similar to everything else and everyday
speech about things as a lot or a little similar would be impossible (or
inappropriate) were it a unity.
[Text: p. 2, Varanasi, and p. 8, Calcutta]
Opponent.2 (against the Prābhākara): Thought like that (a little or a lot of
similarity between two things) occurs through a little or a lot of common
characteristics collected such that they make manifest the similarity.
Prābhākara: That should not be said. We do not fi nd the short and the
long (the abundant and the slight) in the case of something like a
pot that is being made manifest abundantly (by being right in front
of our noses) or slightly (by being a long way off). Furthermore, the
concept of the manifester (vyañjaka) is sufficient to account for the
practice in everyday speech (to talk of a little or a lot of similarity)
since nothing additional has been identified. Nor is similarity a single
thing that, like conjunction, exists depending on relata being related
(there is no conjunction without the things conjoined). For this idea
faces the unfortunate consequence that we would have to say of the
(well-known but absent) cow, as we do say of the (recently encoun-
tered and present) gavaya buffalo, that it is very similar (to a cow)
given that the substratum of the similarity is indicated (upalakṣita)
108 Appendix
by cowhood. And if there is (at the time) no sensory relation to a
cow, then there would be the difficulty that, like conjunction (which
is imperceptible when one of its relata is imperceptible, as with an
earth atom in contact with a pot—both relata have to be perceptible
for their conjunction to be perceived), the similarity would be imper-
ceptible (whereas the cow-similarity in the gavaya can be perceived).
Or, if the similarity is to be perceptible, then the difficulty is that inso-
far as the (cow-)similarity specified as in the gavaya buffalo is to be
known (as indeed it is known) by a single cognition, the cow portion
would have to be visible to the eye (whereas in the case at hand there
is no cow but only a cow-resembling buffalo).
[Text: p. 3, Varanasi, and p. 11, Calcutta]
And contrary to what some suggest, similarity is something
different in each of its locales.
Opponent.2: Because each and every thing would deviate and because
there would be no uniformity, there would be in that case no uniform
concept. Furthermore, there would not be the grammar we know for
the word ‘similarity’.
Prābhākara: That should not be said. For the two (similarity as a uni-
form concept and the grammar of the word) are possible (on several
views): because there is a similarity universal to be known by a cogni-
tion of its recurrent form, or because it is something over and above
the seven (recognized) categories, or because like a natural kind (jāti)
or an ultimate atom it is a sui generis particular (sva-lakṣaṇa).
And it (similarity) is not a substance nor a quality nor an action,
because it occurs in qualities and actions.
Moreover, everyday speech on the score does not flow just from
(cognition of) the similarity’s substrata, since the idea of it is formed
even when the substrata are dissimilar.1 Just for this reason, there is
the idea albeit you do not fail to know that there is no relation of simi-
larity between the substrata. For there is no prevailing counterconsid-
eration (to this thesis) and in general such dissimilarity is possible.
And it (similarity) is also not a universal (sāmānya). For (unlike a
universal) it is not a single thing in all its instances, since that is not
accepted (by either of us) and it would make the everyday practice of
speaking of a lot and a little similarity impossible.2
[Text: p. 4, Varanasi, and p. 15, Calcutta]
Objection (by a Bhāṭṭa Mīmāṃsaka to the Prābhākara): Similarity is
(D1) an abundance of common characteristics (not as an independent
category but) in terms of parts, qualities, and actions, given that there
is difference (between the things that are similar). 3
Just for this reason, such an idea of similarity does not arise if the
countercorrelate (the cow or whatever it is to which there is a pur-
ported similarity) is distant (i.e., does not have much in common with
the gavaya buffalo or whatever) since there would be no cognition
 Appendix 109
of an abundance of common characteristics (bhūyaḥ-sāmānya). For
example, with a forest and a palace the cognition and usage would
be that there is minimum similarity since the abundance, that is to
say, the collection (of common characteristics), is slim. Otherwise,
we would have to hold that such a thing as a forest (or tree plantation
which is made up of individual trees) is also something over and above
(the trees). The everyday practice of speaking of a lot and a little simi-
larity is explained by there being a lot or a little in common.
Thus it is said: Just many commonalities among qualities, parts,
and actions is similarity said to be, similarity that is evident through
commonalities belonging to distinct material causes.
Prābhākara: Wrong. As with a pair of lotuses (which are similar in terms
of their qualities but not in terms of their parts), the idea of similar-
ity can arise, too (i.e., your defi nition is too narrow) with respect to
qualities and kinds when there are things of the same kind without
there being parts etc. of the same kind. And (on your view) similar-
ity would exist in a donkey and an elephant calf which are dissimilar
(i.e., your defi nition is too broad).
[Text: p. 4, Varanasi, and p. 18, Calcutta]
Nor is it (D2) an abundance of properties existing in one thing
with respect to the properties of something else. Nor is it (D3) the
possession of properties that are abundant with respect to those had
by x (the countercorrelate, the cow) given that the things (are not
identical, that is) do not have the same uniquely specifying properties
(asādhāraṇa-dharma). Nor is it (D4) the possession of properties that
are abundant with respect to those had by the absentee (pratiyogin) of
an absolute absence that rests in the thing itself (the thing y perceived
to be similar to x the pratiyogin). (These defi nitions are no good for
the same reason, for the reason that) there is no uniformity about
(what counts as) an abundance, since three, four, five, etc. (common
characteristics) are each a different idea.
And it is not the case that, like the “plural” (which is taken to mean
at least three), a triad is sufficient, since this would land us in the pre-
dicament that even two dissimilar things such as an elephant and a
mosquito would be similar.
Thus it has been said:4 In this way, variegatedness (citratā) applies
to something because of differences among groups of commonali-
ties, singular pairs, triads (and so on), propertywise (that is to say)
with respect to kinds, qualities, substances, actions, and powers (or
capacities).
[Text: p. 5, Varanasi, and p. 19, Calcutta]
Objection: Similarity is an abundance of properties with respect to those
had by x (the countercorrelate, e.g., the cow) relative to the exclusion-
ary, determinant properties (of y, the thing perceived to be similar to
x). And an “abundance” is (to be taken as) uniform through three,
110 Appendix
four, and so on (commonalities). In the case of an elephant and a mos-
quito, in contrast, much is excluded, while what is common is slight.
Just for this reason, similarity is said to be something implying there
is just a little difference (from something else).
Prābhākara: No. The everyday practice is to talk of similarity (in some
cases) where the number of exclusionary properties are (for two simi-
lar things) the same, or a few, or an uncountable number.
Furthermore, common characteristics (i.e., universals, sāmānya)
are undivided through differences in their substrata (cowhood, for
example, is the same in all cows), whereas similarity is distinct,
divided. A common characteristic (or universal) does not have a coun-
tercorrelate, and by that it is not made determinant to us. Similarity,
in contrast, does have a countercorrelate, and it is made knowable
with reference to a countercorrelate.
[Text: p. 6, Varanasi, and p. 21, Calcutta]
Objection: Similarity is (D5) the possession of common characteristics
in the abundance had by x (the countercorrelate, e.g., the cow), given
that x and y (the correlate, the buffalo) are distinct.
And distinctness (bheda) is (to be defined as) something else for each
of its substrata (all the many distinct things), and it does have a coun-
tercorrelate (thus accounting for the relational character of similarity).
Prābhākara: If this is what you hold, then since similarity in its distinct-
ness factor would have defi nite boundaries (sāvadhi, like a pot distinct
from a cloth) there would be the locution “Similar from that” but not
(what we have in fact) “Similar with that” or “Similar to that.”
Objection: In that similarity has a countercorrelate, the usage “Similar
from that” is okay.
Prābhākara: Also wrong. For if similarity has defi nite boundaries, then
conception and usage should conform. But they do not with respect
to the countercorrelate (which is then different with similarity as
opposed to what it is with distinctness). Even if there were none (no
similarity), “(Similar) from a pot,” could be conceived—this is the
unfortunate consequence.
Furthermore, if similarity were a possession of x’s properties, then
there would be chaos concerning just what is distinct from what. For
the possession of the properties of x is to be x.
And to have whatever (“to have that,” tadvattā, to have whatever
characteristic) is to be other than the whatever had. Moreover, if “to
have whatever” were applicable to that (the “haver,” the possessor
of whatever characteristic), then “that itself” (tad eva) would be the
equivalent of “having that” (tadvat). (This is the absurdity.)
And likewise there would be (cognition of similarity expressed as),
“This is that,” as we have in the case of a recognition (of something or
someone) by its having certain properties (“This is that Devadatta”),
but not (what is actually the case with similarity), “This has that.”
 Appendix 111
For given that the having of that would be in this way equivalent to
the being of that, the idea and designation of a cow would be appli-
cable, too, to the gavaya buffalo, since, like the cow, it would be a
substratum of the “being that” (tattā) that rightly captures the char-
acteristics (sāmānya) it has in common with cows.
Nor is similarity the same as the individualizer (as a category),
since (unlike individualizers) similarity can be perceived. Nor is it
inherence (the further category), since it exists in a way over and
above the way inherence nests (in its relata).
Opponent.2 (presumably a Naiyāyika): Whether similarity is a presence
or an absence, whether as a presence it also has its own attributes or
is attributeless, and if attributeless then also whether supervenient or
non-supervenient, and if supervenient then also whether it has a com-
mon characteristic or has none, and if it has, then also whether it is
evident or not evident, and if as attributeless, without a common char-
acteristic and yet supervenient, whether it supervenes on a single thing
or on many—such are the questions that lead us to include it within the
seven (established) categories.
Prābhākara: That’s wrong. (Without similarity as a separate category)
the ways we speak (about similarity) would be impossible since they
fall outside those (other categories and the ways we speak about
them). Otherwise, since on this conception inherence, individualizer,
universal would share the properties of the (other) triad (of categories)
substance and the rest, they would be included in them.
[Text: p. 7, Varanasi, and p. 24, Calcutta]
Gaṅgeśa: We answer. Similarity to y is the possession of an abundance
of properties had by y, properties that are other than those that
uniquely specify x (and determine what it is). And what natural
kinds and so on are common (to the two similar things, x and y),
those are (like all universals, all natural kinds, jāti), without defi nite
boundaries—to consider the factor that does not pertain to the dif-
ference between x and y.
(So) similarity is (D6) the possession of an abundance of proper-
ties had by y. Through such an abundance similarity is to be made
known. Thus similarity only as to be made known by that (posses-
sion of an abundance of properties) is such that it has countercor-
relate structure.
For you, too, just this is what its having a countercorrelate amounts
to, not to its having defi nite boundaries as with the distinctness
between two things, being long or tall, etc., since that idea has the
unfortunate consequence of making us expect the thought, “(This is
similar) from that” (as opposed to “with” or “to that” which is what
actually occurs).
Furthermore, such an abundance is not non-uniform (as was alleged)
through sets of three, four, and so on common characteristics.
112 Appendix
Nor does the defi nition extend too far, since even in the case of
an elephant and a mosquito there is similarity in both being living
things, enjoyers of pleasure, subjects of pain, and so on.
Just for this reason, a little and a lot of similarity are accounted for
by something’s having the properties of something else abundantly
or slightly. For example, even given (great) dissimilarity we say “Her
face is like the moon” because of the (common) generating of joy, etc.,
and because of (common) domestication we say “A water buffalo is
like a cow.”
Just for this reason, the person (the forester in the stock example)
questioned (by a subject who is to acquire knowledge by analogy) spells
out the cow-similarity in a gavaya in spelling out just the properties
that it has that the cow has, too (which are many). “Cows resemble a
boar” (is an example where the common characteristics are few but)
thus here in the case of the boar, too, there is (a little) cow-similarity.
The common ways of speech in poetry and literature concerning
analogy and that which is known by analogy proceed by way of one
thing’s having the properties of something else (being made known).
(And) just by having the prover and probandum properties
(together), the inferential example (e.g., a kitchen hearth where smok-
iness and fieriness are known to be together in the case of an inference
from a smoky mountain to a fiery one) is mentioned by logicians (and
reasoners) in statements that indicate the similarity to the inferential
subject (e.g., of the kitchen hearth to the mountain, which is the infer-
ential subject, the pakṣa).
[Text: p. 8, Varanasi, and p. 29, Calcutta]
And it is not the case that these (statements of similarity) are figu-
rative usages, since there is no block (no prevailing counterconsid-
eration) to taking them in their primary sense (indirect, figurative
meaning kicking in only if the primary, denotative is blocked).
Therefore, what it is to be similar is not uniform with respect to
some property or other, not uniform for some possessor (of similar-
ity), not uniform anywhere (that is to say).
[Text: p. 8, Varanasi, and p. 29, Calcutta]
Moreover, knowledge of such a possession of many properties of
the other thing indicates similarity, makes it manifest. Let just such a
cognition be the regulator of everyday speech about it. Why anything
extra? If there were, then the possession of many properties of the
other thing would not be uniform through sets of three, four, and so
on properties even though it is making some similarity manifest.
Objection: The manifesting cognition is also not uniform, as in the case
of fi re being made manifest by perception and so on (inference or tes-
timony, each a distinct source of knowing).
 Appendix 113
Gaṅgeśa: Wrong. For in that case every individual would be different to
the extent of there being no natural kinds. Fire’s pervading smoke,
light, and the like is itself uniform (making inference possible).
Objection: (Okay, then) here, too, there is pervading (and analogy is a
form of inference).
Gaṅgeśa: In that case everyday speech about similarity would arise from
(cognition of) something being pervaded by something else (but it
does not, arising instead from cognition of something’s having prop-
erties had by something else).
And it is not the case that the possession of many properties had
(also) by something else does not alone indicate similarity. For its not
being alone responsible runs against our common experience. And
without cognition of an abundance of the other thing’s properties as a
lot or a little, there would be no cognition of anything as very or only
slightly similar to something else.
Objection: Possession of many properties of something else does not
have a countercorrelate (in the fashion of an absence). Similarity,
however, is not like that (i.e., it does have a countercorrelate though
not in the fashion of an absence). So it sets up (something additional
as) what is to be talked about.
Gaṅgeśa: No. (Just) like similarity, such possession, too, does have a
countercorrelate.
Objection: Then there would be chaos (concerning just what is distinct
from what).
Gaṅgeśa: That’s wrong. Just here (it is further supposed that) to have the
properties of x or to be non-distinct from x is what it is to be x. And
this appears in the case of a perception (a perceptual recognition),
“This is that (Devadatta I saw yesterday).” Given that the distinctness
(between the two things) is evident, to have properties of that y which
is other than whatever x (is the subject of comparison) is to be like
that, whence there is (the expression and cognition) “This is like that;
it is not that itself,” since the two are different.
Objection: When cowhood is grasped (the common characteristic,
being-a-cow) with regard to another individual (cow), then (if you
were right) there should be (the expression and cognition) “This is
like that cow” (which does not actually occur), and not (what does
occur) “This is a cow.”
Gaṅgeśa: Wrong. The collection (of properties required for one thing
to be similar to another) includes (for example) cowhood simply as
resting in individuals that are other than the thing (the subject of the
comparison, the gavaya buffalo), since there is no comprehension of
(the universal as a universal, i.e., as) something singular bound down
to various individuals. Given a cognition of properties that reside
in those (previously experienced) cows, there does, too, occur (the
expression and cognition) “This is like those cows.”
114 Appendix
[Text: p. 10, Varanasi, and p. 34, Calcutta]
Objection: Then the cognition of the collection should be, “This, too,
has horns and so on,” and not “This is similar to a cow.”
Gaṅgeśa: Likewise wrong. With respect to the single thing (the subject of
comparison, the gavaya buffalo), the collection is related to both (the
present buffalo and the remembered cow or cows), since we differenti-
ate (as follows), “Having an abundance of properties had by the other
thing is the fi rst thing’s similarity to the other.”
But the New (Prābhākaras) say the following.5
New Prābhākara: Similarity is experienced as (sui generis, as) excluding
entirely other things when (for example) there is a distinctive pair of
pleasures. It is not, however, a universal (jāti) simply resting in both
at once, since the difficulty would be that in the absence of both the
universal would not have instances nor be eternal (unlike true uni-
versals). Nor is it another generated property. Thus it is to be granted
that similarity is something additional.
Objection (to the New Prābhākara): For you, too, how without an indi-
cator (to make it known) would the possession of the same properties
make similarity manifest?
New Prābhākara: No. It is as you suggest (i.e., there is dependence on
indicators) with respect to substances (that are similar), not with qual-
ities and so on. Just for this reason, it is not made irrelevant (or otiose,
anyathā-siddha) by a (need for an) indicator. Even in the absence of
indicators, we do experience similarity.
[Text: p. 11, Varanasi, and p. 36, Calcutta]
Gaṅgeśa: The proposal is empty. A pair of pleasures unlike one another
(except as pleasures) is not produced by a cause of pleasure in general,
since that would land us in the predicament that another kind of pleasure
would be produced, too.6 Rather, each is produced by distinct Unseen
Force (i.e., karma). And such Unseen Force, which arises from the prac-
tice of particular actions enjoined and causes pleasure, belongs to oth-
ers, too. Thus for them, too, such pleasures come to be. Thus there is a
distinct universal directing particular actions enjoined for everyone.
Therefore, since living beings are innumerable, since death is
beginningless, there is no pleasure that is entirely different from
that (somewhere, somehow) to be produced in the future. This is
true, too, for other effects such as pains. For they would not be
effects of a kind that has not been produced (elsewhere) or will not
be produced.7
[Text: p. 11, Varanasi, and p. 38, Calcutta]
Objection: The destruction of the individual shows that (your) universal
is impermanent (thus there is no such thing).
Gaṅgeśa: This also is wrong. There is no (possible) destroyer of a (true)
universal. Just for this reason, universals remain even during a period
of cosmic dissolution (pralaya).8
 Appendix 115
[Text: p. 11, Varanasi, and p. 38, Calcutta]
Moreover, similarity is distinct (in its instances) because of the
countercorrelate’s distinctness in relation to the single thing (to which
it is similar). Similarity is not itself a single thing or unit, since we
commonly speak of things as a lot or a little similar.
Objection: There would be in this way a grasping of it as a unit.
Gaṅgeśa: Then the diffi culty would be that you would be grasping
everything (since everything is at least a little similar to everything
else). Gradually cognition of an abundance of properties belonging
to the countercorrelate becomes the indicator (of similarity). Just
such indication prompts the everyday speech—so it was said (by
me earlier).
Objection: (On your view) there would be in this way, too, dissimilar-
ity (as a unity), and not (simply a variety of) absences of similarity,
since the opposite of that (defi nition of similarity) is also possible. The
cognition (we actually have) is just of absence of similarity (not of dis-
similarity), in that the sense of the prefi x ‘dis’ (in ‘dissimilarity’, ‘vi’
etymologically responsible for the ‘vai’ in ‘vaisādṛśya) is to negate, to
prohibit (directly and not just by implication).
Gaṅgeśa: Then there would not be dissimilarity with respect to the
gavaya buffalo which is not the locus of the similarity intended with
the statement, “A mahiṣī buffalo is like a cow,” through (mindfulness
of) the common (properties of) being sacred and having milk and so
on. For it is not the case that there is cow-similarity or absence of
cow-similarity there (since nothing about a gavaya has been intended
when a speaker is speaking about a mahiṣī buffalo). Therefore, having
properties in common and not having properties in common ground
similarity and dissimilarity.
[Text: p. 12, Varanasi, and p. 41, Calcutta]
Objection: Similarity is the possession of many properties that reside in
whatever (is the countercorrelate). Thus we have similarity even when
the things compared are non-distinct, as is expressed in such verses as
the following: “The sky has the form of the sky; the ocean is the ana-
logue of the ocean./ The fight between Rāma and Rāvaṇa is like that
between Rāma and Rāvaṇa.” And it is not true that this succumbs to
the difficulty that the reference of the word ‘gavaya’ would include the
cow, too, when there is similarity to a cow (pointed out). The meaning
of the word ‘similarity’ in this case (“There is similarity to a cow there
in the gavaya buffalo”) would be the particular (similarity that rests
in the gavaya buffalo). Otherwise, the defi nition would be too wide in
including the mahiṣī buffalo.
Gaṅgeśa: Wrong. Right usage is to say “It is similar to that” or “similar
with that.” And the relationship (implied) is not present when the
things are non-distinct nor is there common reference. With regard to
the expressions (you quoted), “The sky has the form of the sky,” the
116 Appendix
meaning is just the sky or whatever is such that nothing else has the
relevant properties.
[Text: p. 12, Varanasi, and p. 43, Calcutta]
Objection: One can draw an analogy between the sky and the ocean
in that both involve endurance. Another fight (between Rāma and
Rāvaṇa) can be analogized with respect to a particular fight between
those two as it would be just between the two.
Gaṅgeśa: Also wrong (for the reasons just stated).
Objection: “Whether Devadatta or someone else no matter, this one
(experienced today), is in the fi rst place, similar to that one (expe-
rienced yesterday).” So there may well be doubt about the distinct-
ness or non-distinctness (of two things) along with certainty about
similarity. From this it follows that similarity should be admitted as
a separate category.
Gaṅgeśa: No. We would use the word ‘similarity’ for a part (as similar
to another part) attending only to the properties of Devadatta (not
to Devadatta himself as similar to himself). How (can you account
for the expression) “This is just the same as that, not similar to it”
otherwise? Cognition and usage have it as, “This is similar to that,
not that itself.”
Objection: Here (in your counterexample) the word ‘similarity’ (is used
in a figurative sense, that is to say) means merely the distinctness
(between the two).
Gaṅgeśa: That won’t do. It would be wrong here (as always) to look
for indirect, figurative meaning if the primary meaning is coherent.
Therefore, similarity is not a separate category.
[Text: p. 13, Varanasi, and p. 45, Calcutta]
Objection (by a Bhāṭṭa):9 That may be. Don’t let it stand that similarity is
an additional category. Nevertheless, analogical knowledge (upamiti)
is a cognition, (for example) “A cow is similar to a gavaya buffalo,”
that arises from cognition of cow-similarity in a gavaya buffalo known
by perception or by testimony.10 And this cognition is not perceptual,
because what it is about is not immediately present (the cow, namely,
or the cow portion of the cognition). Nor is it inferential, because
there is no inferential mark (or prover in the cognitive process gener-
ating it). And the similarity present in the gavaya buffalo is not the
mark (or prover), because it is not a property of an inferential subject
(unlike, “There is smoke on yonder mountain”).
Objection: In the case of another similar pair (a mule and a horse, for
instance), that thing G which is the countercorrelate to the similarity
resting in something F is similar to the F—this is known by percep-
tion, given that the pervasion has been grasped (between being-an-F
and similarity-to-things-G). “A cow is similar to a gavaya buffalo” is
(then) a bit of inferential knowledge that arises from (knowledge of)
the countercorrelate of the similarity that rests in the gavaya buffalo.
 Appendix 117
Bhāṭṭa: No. Knowing, without grasping a pervasion, the gavaya buf-
falo as similar to a cow, knowing from the very fi rst, by perception
or by testimony, one has knowledge of gavaya-similarity in cows.
Moreover, the gavaya-similarity in cows is not an (inferential)
probandum property, because it is from the first not cognized like that.
And being the countercorrelate of the similarity resting in the gavaya
buffalo is not known perceptually to rest in cows (given the terms of the
example), since there is no sensory connection with that which would
be the subject of the knowledge (cows, namely). Thus it is not to be
inferred to be a bit of inferential knowledge. There would be an infinite
regress: since its inferential mark (or prover) would not be perceptual, it
would have to be inferred from another mark (ad infinitum).
[Text: p. 14, Varanasi, and p. 47, Calcutta]
Objection: The cow is similar to that gavaya buffalo, since it has an
abundance of commonalities including parts and so on that rest in the
gavaya buffalo, like another gavaya buffalo.
Bhāṭṭa: That should not be said. No other gavaya buffalo being known in
addition (to the one known perceptually in the example), the cognition is
of similarity to the gavaya buffalo applying to a cow.
Just for this reason, we think that without (cognition of) the simi-
larity in the gavaya buffalo to the cow (cognition of) the similarity in
the cow to the gavaya buffalo is not possible. For it is not the case that
we can throw out as expressible (the cognitions) “This gavaya buffalo
is similar to the cow” (at the same time as) “The cow is dissimilar to
that (gavaya buffalo).” If cow-similarity as fi xed by the gavaya buf-
falo is unknown (to a subject S), then without it (becoming known) no
impossible cognition would come to be fi xed.
[Text: p. 14, Varanasi, and p. 48, Calcutta]
Moreover, circumstantial implication (as we say for the separate
and distinct knowledge source) is for you (the Nyāya philosopher)
inference of the sort known as “negative only.”
And no one says that we can know by perception and the rest (of
the knowledge sources excluding analogy) that (the property of) being
the countercorrelate to the similarity in the gavaya buffalo qualifies
the cow.
Objection: Since in this way a camel would be cognized as dissimilar to
a cow, the cognition of a cow as dissimilar to a camel would also be
from another source.
Bhāṭṭa: No. If camel-dissimilarity as applying to a cow amounts to the
possession of a lack of properties qualifying camels, then, like prior
existence which applies to a remembered cow, the dissimilarity is
known just by (the source we recognize as) non-perception.
[Text: p. 15, Varanasi, and p. 49, Calcutta]
Objection: (Such dissimilarity is) the possession of properties that do
not occur in the camel.
118 Appendix
Bhāṭṭa: Then the properties of a cow (which do not occur in a camel)
would be grasped only in a cow (not in relation to a camel). In the
present case, there would be something extra, something which is
merely the absence of recalled cow-properties in the camel. And (cog-
nition of) that comes just from perception.
Objection: The similarity that rests in x having y as countercorrelate,
along with (corresponding) dissimilarity, is simply a cognition, because
the similarity that rests in x having y as countercorrelate is, insofar as
it is appearing, to be known by (one and) the same cognition.
Bhāṭṭa: No. Because of the difference between the qualifier and the
qualificandum as countercorrelate (between, for example, cowhood
and Bessie an individual cow), it is not established that it is to be
known by (one and) the same cognition. Furthermore, a cognition
with that (x) as its qualificandum portion does not arise, “Similar to
that, it has distinct properties,” since sensory relation to that (x) is the
cause in the case of a perception of that (x) as qualificandum.
[Text: p. 15, Varanasi, and p. 51, Calcutta]
Objection: In perception, connection with the (object as) qualificandum
is the (cognitive) cause, but not connection with the qualificandum as
whatever (predicate), since the view is cumbersome (gaurava). Other-
wise, there would be no perception of a pervasion of a qualificandum
by (existence in the three times) the past, the future, and the present
(whatever its properties).
Old Naiyāyika: No. The cognitive connection characteristic of (knowl-
edge of) universals holds in the case of (cognition of) a qualificandum
as past or future.11
Objection: The similarity to a gavaya buffalo that lies in a cow is the
possession of universals (or common characteristics, sāmānya), such
as being-horned, that are had by the gavaya. And if that cow-simi-
larity in a gavaya buffalo is currently appearing (to a subject S), such
possession is quite evident with respect to a cow that (let us imagine)
is on the scene (too).
Alternatively, cow-similarity present in a gavaya buffalo is the belong-
ing to the gavaya of (common characteristics such as) being-horned and
so on which belong to cows. In this way, just that which does really
belong to the gavaya buffalo as belonging to the cow would be what
the similarity to cows amounts to. And this is known just by the senses.
(Thus to know a similarity requires no special knowledge source such as
“analogy.”) Because the object connected with the sense organs is a uni-
versal (sāmānya), which is a single critter, unitary, there is (in cognition
of similarity) something like the “This is that” typical of recognition of
opposed terms.12
Old Naiyāyika: True (to an extent). Knowledge of (characteristics
such as) being-horned and so on as belonging to the gavaya buffalo,
knowledge which is about a cow as qualificandum (“The cow is
 Appendix 119
similar to the gavaya buffalo” when no cow is present), is not gener-
ated by the senses. There is no sensory connection with a cow (in the
terms of the example).
Therefore, the bits of knowledge “This is similar (to that),” “This
has different properties from that,” “This is longer than that,” do not
result from inference. There is no sensory connection with the object
cognized as qualificandum (as there is typically with inference, e.g.,
the smoky mountain seen in the stock inference to fi re), since there is
no cognition of an inferential mark consisting of many (properties)
belonging to x (the “that,” e.g., the cow).
[Text: p. 16, Varanasi, and p. 53, Calcutta]
Objection: And in this way analogy is cognition of something which
has a countercorrelate to whatever—so they say.13 From analogy to
things previously experienced cognition arises of something as other
to something else without perception or the others (the other knowl-
edge sources accounting for it).
[Text: p. 16, Varanasi, and p. 54, Calcutta]
Old Naiyāyika: That’s contemptible. It is a matter of perception that
elephants, for example, are mutually similar.14 Thus that y which is
the countercorrelate to the similarity in x is similar to x (and not just
vice-versa, with x as similar to y). This holds in general. If there is
knowledge of the pervasion (between such countercorrelatehood and
being-similar), the cow is (judged to be) similar to the gavaya buffalo,
since it is the countercorrelate to the similarity in the gavaya, like a
sister with respect to her brother.
Furthermore, that the cow is the countercorrelate to the simi-
larity resting in the gavaya buffalo is known just by the cognition
of similarity in the gavaya buffalo, because, given similarity to the
cow, the cow is known just as the countercorrelate. Otherwise, the
unfortunate consequence would be that similarity to the cow would
be incoherent.
Without grasping a general (sāmānyataḥ) pervasion between two
correlates x and y, there would not be the (cognitive) result, “This is
similar to that.” Grasping a pervasion between two correlates x and y,
there is, for us (of the Nyāya persuasion), negative(-only) inference, for
the others (of the Mīmāṃsā school), circumstantial implication, (and)
thus something else (i.e., circumstantially inferential knowledge) as
result. There is (such) knowledge of similarity, which is a property
that occurs in both correlates, by means of perception due to sensory
connection with some object as qualificandum (as property-bearer).
That there is a sensory connection with the qualificandum as just
so (i.e., as similarity simpliciter) is unmotivated. Otherwise, how
could there be erroneous recognition of opposed terms, “This is that”
(for example)? Furthermore, the countercorrelatehood to the similar-
ity that rests in y is grasped just by the mind (working independently).
120 Appendix
There is no motivation for the view that the qualificandum (with
respect to the cognition, “The cow is similar to the gavaya buffalo,”
the cow, namely) is given perceptually such that inferential reflection
(parāmarśa) can be fed—this has already been stated.
[Text: p. 17, Varanasi, and p. 56, Calcutta]
Jayanta Bhaṭṭa and company: A person (S) desires to know, “What kind
of thing is a gavaya?” Having heard the answer, “As a cow . . . , so
a gavaya . . . ,” S pays close attention to the meaning of the analogi-
cal statement such that when S comes to see a particular individual
(gavaya), which is as stated, S has the thought (the bit of knowledge),
which is the result of analogy (as a knowledge source), “This is what
is meant by the word ‘gavaya’.”
And it does not come from the statement alone, since that would
have the unfortunate consequence that even the individual (gavaya),
before having been perceived, would be known (by S). Nor does it
come from perception alone, since the consequence would be that
a person (T) who had not heard the statement would also know the
meaning of the word.
Nor is it from a combination of the two. For would this be a
combination of knowledge sources or a combination of the results?
On the fi rst option, where it is matter of knowledge source, does
knowledge-source status accrue to the combination or to the com-
bined? Not the fi rst, because there is no possibility of combination
given a multiplicity of result in that there would be an (impossible)
mutuality of auxiliary cause (x in part responsible for y and y in
part responsible for x). Not the last, because the statement and the
perception occur at different times. Although the statement and its
meaning are made occurrent through memory, they are not success-
ful (for our subject S in generating the bit of knowledge) because of
the time delimitation.15 And if the combination is of results, such
that the bit of knowledge is contained within that, then the dif-
ficulty would be that testimony and inference would also be con-
tained within perception.
[Text: p. 18, Varanasi, and p. 59, Calcutta]
Objection: Then does the result (the bit of knowledge) come to be out-
side the working of the knowledge source? Or, if it is within it, just
how far does the thought (the knowledge) extend? Is there to be an
auxiliary cause with each unique sense faculty and so on?
Old Nyāya: In that case, the activity of the open eye (the visual sense
organ) would occur just when similarity is cognized. When there is
such a cognition, then that (activity) is unnecessary (however).
The distinct occasion (for the arising of the bit of knowledge) is
produced for someone (S) for whom the individual gavaya buffalo is
qualified by similarity to cows experienced previously, who has the
memory of the meaning of the (analogical) statement (made by the
 Appendix 121
forester), by force of putting the information together just at the other
(the later) time.
Therefore, the knowledge of similarity is assisted by authoritative
statements and memory, but it is indeed something other than (some-
thing produced by) authority and perception. It is produced by the
knowledge source, analogy.
Gaṅgeśa: This is the view of the Old Naiyāyikas, Jayanta and the rest. It
is wrong. It fails to include (dissimilarity, i.e.) difference in properties.
When it is said by a northerner cursing the camel, “Fie on the camel,
whose neck is too long, whose lips slouch and tremble, eating sharp
bristles (śūka), with a despicable arrangement of limbs, the worst of
domesticated beasts,”16 the southerner (who has not encountered a
camel) having overheard this, traveling in a northerly direction,
encountering such a thing, thinks, “Now this is a camel.” Here how
does he know (the meaning of the word)? It is not by analogy as so
far analyzed (i.e., as the Old Naiyāyikas view the knowledge source),
since there is no similarity (thus no cognition of similarity). And no
other knowledge source is possible (here).
[Text: p. 19, Varanasi, and p. 62, Calcutta]
We answer (giving now the right view).17 There would not be, fi rst
of all, comprehension of the word ‘gavaya’ unless there were uni-
form specification of the reference, or (at least) the indication of it
(upalakṣaṇa), on the part of the person (S) who would recognize (the
gavaya buffalo) through perception which makes immediately evident
the horns and so on. This is like such words as ‘Chaitra’ used with
reference to a particular body (perceptible at the time) by those hav-
ing the view that there is a single body (throughout the changes of
childhood, youth, etc.): the word ‘gavaya’ applies also to other gavaya
buffalos. Nor does it happen that an inexperienced referent gets set in
memory in some other way, (for instance) in the fashion of the word
‘ether’ (ākāśa) through technical (systematic) employment.18 Just for
these reasons—and because (in the case at hand) there is the (required)
perception, and because the indication (upalakṣaṇa) is not excluded
from the power of reference—it is just like that (namely, the word is
understood by S on the occasion of perception of a gavaya buffalo to
mean the critter in front, as well as others of the kind, those others
being “indirectly indicated,” upalakṣita).
[Text: p. 20, Varanasi, and p. 64, Calcutta]
Nor is the grasping of the word’s meaning due to etymological
explanation, as with such nouns as ‘cook’ (‘pācaka’, a cook, which is
derived from the root ‘pac’, “to cook”). Otherwise (i.e., if the meaning
of the word ‘gavaya’ were known etymologically in our paradigm case
where the animal which is a referent is present), ‘the cook’ would not
be used (being semantically inappropriate), if the root “to cook” is not
employed (in the statement, too).19
122 Appendix
Therefore, the power of reference is etymologically in conformity
with, auxiliary to, that (action, e.g., of cooking). And from etymology
there is also indication (upalakṣaṇa) that fi xes (general meaning or
sense, although not the specific referents).
Nor is the grasping of the word’s meaning due to a narrowing of
a meaning set in memory (by previous experience), like the word
‘heaven’ (‘svarga’, which gets its meaning from a narrowing of the
meaning of ‘pleasure’, sukha, or ‘su’ in ‘su’ + ‘arga’), since there is
nothing like that (with the case at hand).
[Text: p. 20, Varanasi, and p. 65, Calcutta]
Therefore, like the word ‘earth’, etc., which is employed because of
earthiness indicated by such (properties) as smell, the words ‘gavaya’
and company have referential power given that being-a-gavaya-buf-
falo, or the like, is indicated by such (properties) as its (cow-)similari-
ty—this, in contrast, is the best (analysis of the) meaning. Otherwise,
there would be opposition between the knowledge source (which pro-
duces knowledge) in the appearance of the result (as false).
And it is not the case that being-a-gavaya-buffalo has been expe-
rientially set in memory (in the terms of our example), because there
is nothing that could have done so. Furthermore, it is not so set just
from (the forester’s use of) the word ‘gavaya’, because that would
involve (impossible) mutual dependence.
[Text: p. 20, Varanasi, and p. 66, Calcutta]
Objection: From the statement alone the occasion (for use of the word) is
fi xed (for S) by that (cow-similarity). At the very beginning, 20 S recog-
nizes, “That which is similar to a cow, (already) understood by me to
be expressed by the word ‘gavaya’, well, this is that.” Thus this is not
the object of analogy (as a separate knowledge source).
Gaṅgeśa: Wrong. For it is not the case that the similarity alone is the
grounds for usage (of the word ‘gavaya’), since this faces the difficulty
that those who have not comprehended the similarity could not use the
word correctly in everyday speech (but they do) and it is a cumbersome
position. Nor is being-a-gavaya buffalo (the exclusive grounds for use
of the word), because that has not been experientially set in memory (at
the time of S’s analogical knowledge). Just for these reasons, it is not
both (the similarity and the universal), since this would render useless
the analogical statement to communicate the occasion for the word’s
employment as understood by S. For the employment is coherent with
respect to something’s being a gavaya buffalo, not with respect to cow-
similarity (which is yet to be comprehended), since the comprehension
is of being-a-gavaya-buffalo, which is the employment’s grounds.21
[Text: p. 21, Varanasi, and p. 68, Calcutta]
Objection (by objector.1): Although the statement has not produced
(knowledge of the word’s meaning) prior to the fi xing of the occa-
sion for use (through perception of the gavaya buffalo), still that very
 Appendix 123
statement, being remembered, will fi x for S the grounds of S’s employ-
ment (of the word). The Vedas are indeed understood on the occasion
of their being taught. The (Vedic) auxiliaries (such as prosody and
grammar) and sub-auxiliaries (such as Mīmāṃsā and Nyāya) may
be completed at another time (without undermining the fact that the
Vedas have been understood).
Objection (by objector.2 against objector.1): It is (all) completed, because
it is understood even prior to the meaning of the statement, so to say.
Objector.1: That should not be alleged, because there’s reasonable
doubt about both (the range of) the indication and the grounds for
employment of (S’s comprehension of) similarity to a cow. Later,
when being-a-gavaya-buffalo is comprehended (upon the encounter
with one), there will be an understanding of the entire connection
(of words according to their meaning, anvaya), when the indication
of the similarity (as comprehended) has become certain—aided (per-
haps) by considerations of simplicity as shown by reasoning (tarka).
Thus here we have something like (the process whereby we under-
stand indirect meaning, as with) “The village is in (the water of)
the Gaṅgā” (which is properly understood as “The village lies on
the bank of the Gaṅgā” while connoting that it has a meditational
atmosphere). 22
Gaṅgeśa: Wrong. Although there may be reasonable doubt about indication
and grounds for employment, (the construction) “That which is similar
to a cow is this which is the meaning of the word ‘gavaya’” is made the
appropriate connection (of words according to their meaning) just by
the words referring to the gavaya being in the same case (and the gavaya
buffalo being the substratum of the qualifications), independently of
another source’s informing (the knowledge). 23 This is like the communi-
cation, “There’s a cloth,” even if there is doubt about whether it is red or
not red. Otherwise, even when a statement has been completely appre-
hended, there would be the difficulty that a distinct statement would be
required to know the first’s connection (of words semantically) because
another source is (thought of as) an auxiliary cause.
Here with “The village is in (the water of) the Gaṅgā,” it is true
that there is dependence on another source with regard to the words’
meaning which is not (literally) fit to express a semantic connection.
And if the meaning also comes under the sway of a cognized relation-
ship, then there is the difficulty of a different statement.
[Text: p. 22, Varanasi, and p. 70, Calcutta]
Objection (by a new fi rst-level pūrvapakṣin, objector.1): Because the
(apparently) intended meaning is impossible, as with “Usher in the
sticks,” let it have indirect, figurative meaning (as with, “Usher in the
ascetics carrying walking sticks”). With respect to the analysis of the
meaning of the word ‘gavaya’, to produce an understanding of it, an
expert (āpta, the forester) makes an (analogical) statement. And that
124 Appendix
would not occur without acquisition (or understanding on S’s part) of
the grounds of the employment.
And the similarity to the cow mentioned, “It’s like that,” which
has the cow as a substratum in common, is not to be thought of
as having indicated being-a-gavaya-buffalo (the class character or
universal) through the words ‘similar to a cow’ (in the forester’s
statement, “A gavaya buffalo is similar to a cow”), to consider the
intended meaning. 24 And it is wrong to think that here, as with
“There’s smoke over there,” the intention to communicate the fi re
over there is accomplished through analogy (upamāna) instead of
figurative meaning (lakṣaṇā).
Furthermore, here there is no other knowledge source whereby
the communication could be accomplished, since analogy has yet to
be shown to be one. Otherwise, there would be mutual dependence.
Moreover, it will not do to say that indirect, figurative meaning
accrues to the statement (as a whole) as that is a property of words.
Figurative meaning applies to a single word although the other words
condition the meaning—this is what we want to say.
Objection (by objector.2 to objector.1): Although it is through indirect,
figurative meaning that being-a-gavaya-buffalo is set in memory (for
S), no reference to the individual (gavaya) is cognized from the words
(of the forester). Thus it is necessarily granted that another source
would make known the reference to the qualified (individual).
Furthermore, there is no rule that says that an x-situated property
(e.g., blue) expressed by a word for x (blue-thing) is (invariably) that
which is expressed by that word, since there is a counterexample in the
case of the word meaning the class character (being-blue or bluehood
as distinct from the particular instance).
Reply (by objector.1): Wrong. It is a matter of indirect, figurative meaning
that indeed the individual gavaya buffalo (yet to be encountered) is cog-
nized as qualified by (the class character) being-a-gavaya-buffalo when
the analogical statement has been understood in its intention to make S
know its express content (which includes all gavaya buffalos).
Objection (by objector.2): If the intended meaning grasped through the
words indicating the cow-similarity were directly (unmediatedly) to
being-a-gavaya-buffalo, then there would have to be indirect, figurative
meaning. But the intended meaning is not understood directly (unme-
diatedly). Otherwise, the same would hold for “There is smoke over
there” (which is intended to say that there is fire over there although
there is no direct fire-reference).
Reply (by objector.1): Wrong. There is nothing else that would set it
in memory (set being-a-gavaya-buffalo in memory). It is understood
directly, because the intended meaning is just what intended meaning is.
And analogy indeed is not like this (and should not be posited as a sepa-
rate knowledge source), since there would be mutual dependence.25
 Appendix 125
[Text: p. 23, Varanasi, and p. 73, Calcutta]
Gaṅgeśa: Don’t think like that. With respect to the understanding of an
object that prompts the employment, “This, which is similar to a cow,
is referred to by the word ‘gavaya’,” it is not the intended meaning (of
the forester’s analogical statement) but just that which should be named
according to its own nature. Here, too, and wherever, according to the
collection of causal factors sufficient (for knowledge), which have been
stated (earlier), there results a precise knowledge of the occasion (i.e.,
when S identifies the gavaya buffalo as the word’s referent).
Furthermore, with respect to the particular (gavaya buffalo) and
the grounds for the (later) activity (of speech on S’s part), it is false
that the intended meaning would be—for the purpose of reflective
analysis of it by someone who wants to—the seed (or spark) for figu-
rative meaning that works through (knowledge of) its impossibility
(as a matter of strict reference).
Moreover, in a case where there is nothing in the intended meaning
directed to (providing) the grounds for (S’s later) activity of speech,
the statement (for instance, given earlier) whose intended meaning is
to curse, “Fie on the camel, whose neck is too long, etc. (whose lips
slouch and tremble, eating sharp bristles, with a despicable arrange-
ment of limbs, the worst of domesticated beasts),” S nonetheless
encountering such an individual and remembering the meaning of
the statement (of the camel-cursing northerner) has the thought (the
knowledge), “This is what is referred to by the word ‘camel’.”
Furthermore, it is always true that it is just the impossibility of
relatedness (of the semantic units understood denotatively, anvaya)
that is the seed (or trigger) of figurative meaning. Even with “Usher
in the sticks” and the like, the point that there be honoring (of the
ascetics) and so on is comprehended by means of the context and
other factors. Since there is in the case of sticks and the like taken
together with an action such as ushering and the like no relatedness
(anvaya), S, not knowing the intended meaning—in part because
(literal) intended meaning is impossible (here) such that S couldn’t
know—comes to know a meaning that is necessarily dependent on
context and other factors.
Objection: Context and the others are not uniform (whereas units
of meaning are uniform). The uniformity of whatever (word or
expression) pervades (i.e., is a necessary condition on) its meaning.
This being the case, let the intended meaning’s impossibility be
the seed (or trigger) of that (fi gurative meaning), since this is the
simplest theory.
Gaṅgeśa: Wrong. There is no opposition between the pervasion (you
claim) and the non-uniform, as with “Smoke” and the like (when
what is intended is fi re). Therefore, the word ‘gavaya’, which is clearly
known as expressing something or other in general since S knows that
126 Appendix
the word is used by fellow students (but does not know precisely or
in particular), comes to be (fully) known in “The thing characterized
as qualified by being-a-gavaya-buffalo is what is meant by the word
‘gavaya’”—it should be concluded because of simplicity. 26
[Text: p. 24, Varanasi, and p. 74, Calcutta]
Further, that (analogy) has other knowledge sources as its auxilia-
ries. And on the particular occasion (as described) there is no other
knowledge source (that could produce the result specified). Therefore,
whatever has knowledge sources as its auxiliaries should be posited as
another knowledge source.
Objection: Let it be inference (that is responsible here). For example,
the word ‘gavaya’ is (known as) expressing the thing characterized
as qualified by being-a-gavaya-buffalo—when other ways it could be
known are lacking, a way such as the word ‘cow’ being used by S’s
elders to refer to a cow (in the presence of S).
Gaṅgeśa: No. Being-a-gavaya-buffalo is not known by S as the grounds
for the (later) activity (of speech) even though S has grasped that what-
ever it is that is qualified by being-a-gavaya-buffalo is what is meant by
the word. Furthermore, without knowing the referential power S would
not be able to use some other way in whatever circumstances to pin it
down, because mere employment of words in the same case (“This is
that”) is common to (all) indication of grounds for employment.
[Text: p. 24, Varanasi, and p. 76, Calcutta]
Objection: One would introduce a cumbersome theory in having cow-
similarity be the grounds of the activity (of speech); one need not intro-
duce analogy. And so by suppositional reasoning (tarka) it becomes
certain that there would be no grounds for the activity otherwise,
such that the word ‘gavaya’ has grounds for its usage in (the idea of)
being-a-gavaya-buffalo. Given that there can be grounds for the usage
in no other fashion, whatever is not the one way (~S), that is not the
other (~H), since it has grounds for usage. 27
Gaṅgeśa: Wrong. Suppositional reasoning does not bring certainty. And
such suppositional reasoning (as you give) is not rooted in (a good
inference where there is) a grasping of a pervasion, in that from a
contrary inference something could be established.
[Text: p. 25, Varanasi, and p. 77, Calcutta]
Objection: The word ‘gavaya’ has grounds for its usage—this is just the
conclusion of an inference based on common characteristics (sāmānyato
dṛṣṭa) aided by suppositional reasoning which establishes that the word
‘gavaya’ would not have grounds for usage otherwise.28 There is, how-
ever, no additional knowledge source. Imagining (the type of animal),
one conceives the suppositional reasoning, which is an auxiliary cause
(for the eventual knowledge).
Gaṅgeśa: No. Although “This has grounds for usage” and “The other
has no grounds for usage” can be conceived, (the proposition) that the
 Appendix 127
word ‘gavaya’ has grounds for usage is not known except through the
additional knowledge source (viz., analogy). The rule is that inferen-
tial knowledge (i.e., the result of inference as a knowledge source) has
predication content (prakāra) precisely in making specific a pervading
by a pervader.
[Text: p. 25, Varanasi, and p. 78, Calcutta]
Objection: As by an inference of the sāmānyato-dṛṣṭa type given desire
(as a property requiring a property-bearer) aided by (the suppositional
reasoning) ruling out (all eight) candidates, its not having any of the
eight substances as its locus (or property-bearer) becomes known (and
thus that a ninth substance, the self, ātman, is to be posited). Other-
wise, how would the fact that desire does not have any of the eight
as locus get grasped? There are reasons that rule out the particular
candidates (the eight substances established on other grounds, earth,
water, fi re, and so on), namely, the substances being determined to be
objects contrary to (having) desire. Nevertheless, aided by the kind of
suppositional reasoning called “the cumbersome” (i.e., considerations
of parsimony or simplicity), the sāmānyato-dṛṣṭa type of inference
makes it known that there would be no grounds for employment in
another fashion. Then, this being known, later there would be an
inference of the “negative-only” type. (In other words) in comparison
with (the cumbersomeness of) positing another knowledge source to
be conceived, it is correct to hold that the well-known (well-conceived)
sources are the causal factors.
Gaṅgeśa: No. With respect to desire (which would be used to prove
a self, ātman, inferentially), it is not by a sāmānyato-dṛṣṭa type of
inference that it is known that it does have any of the eight as locus.
That is determined, fi rst of all, just by blockers of the others aided
by the blocking of each (i.e., all the ruling out taken together) such
that the knowledge that it does not have any of the eight as locus is
qualified by the knowledge that desire is a property of a substance.
The sāmānyato-dṛṣṭa type of inference produces knowledge of (a
self, ātman, as) an entity as qualified by properties precisely through
determination of this pair of qualifiers (namely, its not having any
of the eight as locus and its being a property of a substance). 29 Fur-
thermore, there is no occasion to employ suppositional reasoning
(of the “cumbersome type”) in that it (the self) is made known (in
this way).
[Text: p. 26, Varanasi, and p. 81, Calcutta]
Objection: The statement, “This which is similar to a cow is the meaning
of the word ‘gavaya’,” is about pervasion (pervasion the knowledge of
which underpins inference). We get it from a statement where a pervasion
is understood whose arising comes from inferential knowledge: “This is
what is referred to by the word ‘gavaya’, since it is similar to a cow, like
some other individual (gavaya buffalo).”
128 Appendix
Gaṅgeśa: That should not be alleged. For S, knowing the cow-similarity
(through experience), would not question anyone, “By what word is
this called?” Rather, understanding in general the meaning of the
word ‘gavaya’, S asked, “What is a gavaya like?” And so what would
be the point of an inference if something’s being pervaded were known
by means of usage and precedence?30
[Text: p. 26, Varanasi, and p. 82, Calcutta]
Objection: “What characteristics does it have?” is a statement about
defi nition, which is the meaning or point of the question. And in
this way that which is to be formally spelled out in an inference
(accounting for what you call analogical knowledge) is as follows:
“This is that gavaya buffalo” is to expressed in everyday speech,
since it is similar to a cow; what is not so, is not similar to a cow,
like an elephant.
Gaṅgeśa: Wrong. For there is no knowledge source for the difference
(you cite) with elephants, because the entire formal proof is impos-
sible to understand (without the difference provided by analogy as
a knowledge source) in that the everyday speech of even just a few
people would be equivocal.
Furthermore, “What sort of . . . ” or “Having what characteris-
tics . . . ” is not the meaning of the question (asked by S). For S, not
knowing the inferential mark (the prover, of your inference) nor the
reference of the word ‘gavaya’ as used by anyone, would not have
understood what that one (the forester) wished to say so as to be able
to pose a question having that meaning (of yours). S’s question con-
cerns the individuals that would provide the grounds for the employ-
ment (of the word).
Objection: The question’s meaning and point: What is the inferential
mark for that which would be the grounds for employment of the
word ‘gavaya’?
Gaṅgeśa: No. For that is simply not to be inferred by him such that, it
being made certain for him by that inference, he would come to know
what he does.
Objection: The question concerns only the cause of S’s (eventual) knowl-
edge (qua qualificandum). The answer gives its qualifiers.
Gaṅgeśa: No. Since he does not know the particulars (that would ground
use of the word), he would be answered (by the forester) also with ref-
erence to (his need for a) sensory encounter: “If you go into the forest,
you will see.”
Therefore, S’s questions would concern only the distinct grounds
(of employment): “What is a gavaya like? On what grounds does one
employ the word ‘gavaya’?”31 Since the one asked (viz., the forester)
would not be able to show immediately, directly (sākṣāt) what it is to be
a gavaya buffalo, he would talk about similarity, indirectly indicating
it (what it is to be a gavaya buffalo). And then later when an individual
 Appendix 129
(gavaya buffalo) has been seen by S who remembers the meaning of the
analogical statement, he acquires the knowledge: “This thing qualified
by what it is to be a gavaya is the referent of the word ‘gavaya’.” S’s
knowledge is of a particular that is the grounds for the employment (of
the word). It is the result of analogy (as a knowledge source).
[Text: p. 27, Varanasi, and p. 87, Calcutta]
Objection: It would be just an inference of the negative-only type that
would give the thought (the knowledge) to S that what it is to be a
gavaya buffalo (the universal or class character) provides the grounds
for use (of the word): Given that there is no such grounds in anything
else because of the cumbersomeness that would involve, (whatever
is not the one way, that is not the other), since it does have grounds;
(what is not so is not so).
Gaṅgeśa: No. Insofar as a bit of suppositional reasoning is not rooted in
(a good inference where there is a grasping of) a pervasion, neither a
charge of cumbersomeness nor its opposite would be authoritative.
Objection: If that’s not a knowledge source, then how will your analogy
fare in the face of doubt about the grounds being something else, also
aided by suppositional reasoning?
Gaṅgeśa: Wrong. Just from analogy aided by suppositional reasoning
about cumbersomeness in the form of considering whether the usage
could not have other grounds, S has the knowledge that the usage has
grounds (but not knowledge of the gavaya buffalos that as such actu-
ally are the grounds for use of the word).
Objection: Okay, let us take it for granted that it is just from a knowledge
source, aided by suppositional reasoning of the cumbersome variety,
that it is known that the usage has grounds. It is established (for S)
that it has grounds in what it is to be a gavaya buffalo (the universal
or class character).32
Gaṅgeśa: But it is not so; the defi nite knowledge does not arise through
a negative-only inference. Provided that something’s being an effect
(of something else) has, without a prevailing counterconsideration,
already been concluded, a negative-only inference might come about.
But where from the fi rst there is no negative correlation, the result
comes just from (the consideration that) something is an effect (of
something else); (in other words, here) we would have the ideas that
the usage does not have grounds in something else (something other
than what is called a gavaya) and that it does indeed have grounds.
Nevertheless, the notion that the word ‘gavaya’ has grounds (for its
use) in what it is to be a gavaya buffalo (the universal or class charac-
ter), this notion would not be established (for S until he encountered
an actual gavaya buffalo)—this we have already said (previously).
Moreover, here according to (the pattern of) everyday speech the
grasping of the referential power of the word is with respect to the
object of the (word’s) employment (i.e., the individual), not to its
130 Appendix
(abstract) grounds (the universal), since there is no obstacle to that. Nor
is there grasping of grounds in what it is to be a cow or anything else
in the rejection of other possibilities through considerations of cum-
bersomeness and overextension, since there is (here) no suppositional
reasoning of the cumbersome or any other variety and since (for that)
there is no knowledge source. Therefore, given just a grasping of the
power of reference with respect to the individual, S makes himself come
to understand, by means of auxiliary considerations such as simplic-
ity, that there are grounds in what it is to be a cow and the like (e.g.,
other types of animal) in that there the usage would not have grounds
otherwise.33
[Text: p. 28, Varanasi, and p. 91, Calcutta]
Objection: In all cases just that (the understanding of a word’s referen-
tial power) is the result of analogy.
Gaṅgeśa: Some hold this view. It’s wrong. S does also (if he does, in addi-
tion to having analogical knowledge) make himself come to understand
grounds which do indeed, aided by such considerations as simplicity,
make one comprehend the power of reference—as, for example, with
the unity of the creator with respect to earth and so on.
Objection: Is it that in this way because of simplicity in comparison
with many distinct makers, the effect (earth and so on) makes one
know the unity of the creator? (I say) it does indeed make it known.
Later, the view is (challenged and) rejected because of prevailing
counterconsiderations.
Gaṅgeśa: No. A view appears to be wrong in that its object or content is
blocked (or has been defeated). Furthermore, reflection on simplicity
with respect to the view that fails the test, the non-simple, shows that
it is wrong such that an inference based on it would falsely appear to
be true.
Objection: S’s analogical knowledge amounts to an etymological analy-
sis (vyutpatti). Otherwise, it could not be explained.
Gaṅgeśa: That should not be alleged. If etymological analysis were
an auxiliary cause for comprehending the word, it would rule out
(analogy as) another knowledge source. But according to just what
cognition is considered, there is (afterwards) etymological analysis,
cognition which (here in this case) amounts to knowledge (by anal-
ogy) of a particular thing referred to.34
[Text: p. 29, Varanasi, and p. 95, Calcutta]
But others hold the following.
Objection: Inasmuch as words such as ‘which’ and ‘that’ (‘yat’ and
‘tat’) lack a continuous uniformity (of reference), from grasping
their power of reference with respect to just one thing, (it is real-
ized) they may be used for another. 35 The everyday pattern of speech
(with such pronouns) requires another word (the antecedent or con-
sequent). They are different from the other words (on which their
 Appendix 131
referential power depends). (Similarly) analogy (as non-uniform in
involving distinct similarities and as having other knowledge sources
as auxiliaries) is not an additional knowledge source.
Gaṅgeśa: That’s wrong. The grasping of their referential power is by
indication (upalakṣaṇa) according to what has been set in the mind
and so on, which happens in common (for all things potentially
referred to, a pot, a cloth, and so on). Just from the particular held
in thought, it is determined defi nitely (to be the meaning of the pro-
noun), since the pattern of everyday speech is just so. Otherwise, even
if there were another knowledge source, the fault would remain (of
violating the pattern of speech). For grasping the referential power in
these cases is not just with respect to a uniform idea.
[Text: p. 30, Varanasi, and p. 97, Calcutta]
Objection: Why should the likes of the mother of Maitra be picked out in
the grasping of the referential power of words like ‘mother’ (‘mātṛ’)?
The power works through the likes of her being the wife of Janaka
(being known). That there is a public pattern of everyday speech in
particular (that provides an answer to this question or similarly that
supports a posit of analogy as an additional knowledge source) should
not be said. For the meaning of one word follows upon (and connects
with) the meaning of another, and being-the-wife-of-Janaka is not
like that (not being said). So how through (mention of) Maitra is there
a connection intended with an idea of Janaka? So we would need
another knowledge source (and another and another and another for
other meanings, in the spirit of your posit of analogy).
Gaṅgeśa: That’s wrong. Insofar as dependence on an idea of Janaka is to
be generated (by a statement), (the sentential requirements of) seman-
tic fitness and so on (grammatical expectation and proper representa-
tion) would make the connection with (would make the connection
from the mention of) Maitra.36
[Text: p. 31, Varanasi, and p. 98, Calcutta]
But the defi nition (of analogy) is that it is the instrumental cause
(the trigger) of analogical knowledge.
And what it is to be analogical knowledge (the class character) is a
natural kind (jāti, “universal”). It is not true that it is unestablished
(as such). From (reflection on) what it is to be a cause (or, the idea of a
cause) it is evident that to have a particular cause it is necessary to be
a discrete effect. (Both are here.)
Alternatively, analogical knowledge is (may be defi ned as) knowl-
edge of (i.e., coming to know) the grounds for employment of a word
whose referent as a common substratum (i.e., qua qualificandum)
is fi xed by a property (or properties). What it is to be such grounds
is to be the object of the word’s referential power through some
predication content (prakāra, i.e., some general sense as would be
communicated by the forester in the stock example).
132 Appendix
Or, analogical knowledge is the discriminating knowledge through
some predication content of the grounds for the use of a name (a
noun) that has not been understood previously but that has the same
referent as a common substratum (i.e., qua qualificandum) as the
meaning of a statement (e.g., the forester’s analogical statement).
Or, it is the discriminating knowledge through some predication
content of the grounds for the use of a name that has the same referent
qua common substratum as the meaning of some particular (analogi-
cal) statement.
With “What is dissimilar to a cow is not a gavaya buffalo,” here,
too, making conversions (from “not dissimilar” to “similar” and
the like), we get indeed the (same) intended meaning that we have
elsewhere.—Thus is everything all cleared up.
Notes

NOTES TO CHAPTER 1

1. Vātsyāyana, for example, whose commentary is the oldest extant, explains

the word in this way in his “Bhāṣya” (NySBh) on Nyāya-sūtra (NyS) 1.1.41
(pp. 330 and 38–39), a sūtra addressing the “fi nal ascertainment,” nirṇaya,
that comes about as a result of philosophic inquiry, i.e., nyāya. He also
explains what nyāya is in his commentary on the very fi rst sūtra (NySBh
1.1.1, p. 35). All the later commentators elaborate. In brief, nyāya is the
bringing to bear on a disputed thesis and/or counterthesis factual and hypo-
thetical considerations that help to determine the truth.
2. Meurath 1996 presents the arguments that chapters 1 and 5 of the Nyāya-
sūtra are older. Although chapter 1 covers all the topics mentioned in the
fi rst sūtra, those topics are disproportionately concerned with debate and
reasoning, and chapter 5 is almost entirely dominated with talk of formal
and informal fallacies.
In referring to the text, I shall assume—for convenience, with the tradi-
tion—that Gautama, also known as Akṣapāda (Mr. “Feet-gazing,” so lost in
thought is he), is the author of all the sūtras.
3. There are many materials available now for scholarship on Nyāya as well as a
sizeable secondary literature. Just about everything from manuscripts to sec-
ondary works, through 1985, is listed in Ganguly 1993. An up-to-date, more
complete bibliography—but not just for Nyāya—targeting classical Indian
philosophy as a whole (excluding aesthetics) is available online: Karl Potter,
ed. Encyclopedia of Indian Philosophies, Vol. 1, Bibliography: http://faculty.
washington.edu/kpotter/.
Comparatively little work has been done on the later period of Navya
Nyāya. Groundbreaking is Ganeri 2011 which examines Nyāya develop-
ments in the early modern age, the fi fteenth through seventeenth centuries.
These are largely not covered in this book.
4. The Nyāya-sūtra itself mentions three synonyms (NyS 1.1.15, p. 213) in the
midst of sūtras laying out the skeleton of Nyāya’s philosophy of mind.
5. Both ‘siddhānta’ and ‘nirṇaya’ occur in the long compound listing the topics
of the text in the very fi rst sūtra, NyS 1.1.1 (p. 28). The former is elaborated
in particular under NyS 1.1.26–31 (pp. 260–268), a stretch of sūtras where
four varieties are described. The latter, nirṇaya, differs from siddhānta
mainly in being a single belief as opposed to a web of tenets, siddhānta, self-
consciously held as true.
Although not highlighted by modern scholars, the Nyāya concept of
siddhānta is picked up by the classical textbook authors, e.g., Mādhava,
134 Notes
p. 163: “A ‘tenet’ (siddhānta) is something which is accepted as being author-
itatively settled as true.”
6. Vātsyāyana says as much in his short introduction to NyS 1.1.1 (NySBh, p.
21), for example.
7. The Nyāya-kośa (Jhalakikar 1978: 663) provides this reference to Laugākṣi’s
defi nition (Tarka-kaumudi 1,7), which is influenced by Gaṅgeśa (tadvati tad-
prakārakatva, (a cognition’s) “having predication content φ about a φ-object,”
or “awareness of something there where it is,” Tattva-cintā-maṇi (TCM),
perception chapter, p. 236). The only edition of the Tarka-kaumudi I have
seen, edited by Dvivedi 1886, has a variant reading without the example.
8. That is, as Naiyāyikas see things. Compare the designations of the sides of
the controversy known as parataḥ vs. svataḥ prāmāṇya, which is the ques-
tion whether a bit of knowledge is shown to be so within itself (svataḥ) or
that that requires another (parataḥ), a certifier. The former is the position of
Mīmāṃsā, Yogācāra, and Vedānta (for different reasons); the latter Nyāya’s.
9. For example, Gaṅgeśa, TCM, perception chapter, p. 218.
10. For example, Vātsyāyana introducing NyS 1.1.1 (pp. 26–27) and 3.1.51 (p.
779: pramāṇasya tattva-viṣayatvāt).
11. Such trust, viśvāsa, is shown in action, as will be discussed. The word ‘vāda’,
“opinion” or “position,” is close to English “belief.” But the Sanskrit word is
usually used for an interconnected patch of beliefs or propositions, although
sometimes a position or principle can be stated in a single sentence.
12. This comes to be called eka-vṛtti-vedyatā, an object as known at once as
something qualified by a property. Bhāsarvajña appears to be one of the fi rst
to discern it, although it is implicit in Vātsyāyana’s theory of illusion, it seems
to me, as will be discussed: Bhāsarvajña, Nyāya-sāra, pp. 18, 19.
13. Dispositions, saṃskāra (sometimes called vāsanā, “memory vectors”) are one
of twenty-four types of property recognized in early Vaiśeṣika. They are much
discussed also in the Yoga system as well as in Yogācāra Buddhism. Mental
dispositions are thought to underpin karma and habits along with memory,
and some are said to be carried by an individual into her next birth.
The Vaiśeṣika philosopher Praśastapāda (c. 575) lists saṃskāra as one
of twenty-four qualities, fi nding three subtypes, vega (impetus or speed),
bhāvanā (the experiential, the mental), and sthiti-sthāpaka (elasticity):
Padārtha-dharma-saṃgraha (PDS), pp. 570–574. Later Naiyāyikas debate
about whether the overall category captures a true universal, a natural kind,
as opposed to a mentally projected generality (e.g., Annambhaṭṭa, p. 361),
but it is not too difficult to see the commonality, the dispositional nature of
all three as a kind of self-perpetuation.
Note also that when Yogācāra idealists such as Asaṅga (c. 300 ce) and
Vasubandhu (c. 350) talk of a “storehouse consciousness” (ālaya-vijñāna),
they are thinking about things as webs of saṃskāras, more commonly called
bīja, “seeds,” and vāsanā, “mental vectors” in the broadest sense—responsi-
ble for, in this idealist perspective, not only internal objects but also material
things as, to use J. S. Mill’s phrase, “permanent possibilities of experience.”
14. This is the New Nyāya theory. The match between cognized and cognition
as understood by late Naiyāyikas is elucidated by Sibajiban Bhattacharyya
1987: 226–229.
15. The theory is elaborated by Vācaspati under NyS 1.1.2, which mentions
“wrong cognition,” mithyā-jñāna, understood by Vātsyāyana as viparyaya,
“falsehood,” and glossed by Uddyotakara as “The idea that something that
is not F is F,” atasmins tad iti pratyayaḥ (NySV, p. 72). Under this, Vācaspati
takes the opportunity to explain the anyathā-khyāti theory of perceptual
error (NyVTatp, pp. 73–75; Thakur ed., pp. 67–68). He quotes Kumārila,
 Notes 135
the Mīmāṃsaka, who agrees with Nyāya that cognitions can be false, as
opposed to the Prābhākara Mīmāṃsaka who takes cognition by nature to
be true (Śloka-vārttikā, nirālambana-vāda v. 117–118, pp. 137–138). A little
before Vācaspati but after Kumārila there is also Bhāsarvajña (ninth century)
who uses the label “anyathā-khyāti” for the Nyāya theory which he lays
out after having identified, and refuted, eight competing theories: Nyāya-
bhūṣaṇa, pp. 25–32.
16. The Nyāya idea is thus close to the “objective justification” of John Pollock
(1986: 183–190) and other analytic epistemologists, which also entails truth,
as will be elaborated later near the end of the section.
17. See in particular Vātsyāyana on NyS 2.1.20 quoted and translated in the fi rst
section of the next chapter.
18. Vātsyāyana under NyS 1.1.1., NySBh, p. 21, presents the bit of tarka.
Mohanty 2000: 37–38, for example, misinterprets it as foundational.
19. Gaṅgeśa spells this out in the “Knowing Veridicality” section of the TCM
perception chapter, pp. 126–131.
20. A good example is an inference Uddyotakara makes for the I-cognition being
perceptual because of its independence of a need for memory of a pervasion
as well as its conformity to the distinct nature of the object, “like a cognition
of color”: NySV, p. 704. See also Taber, “Uddyotakara’s Defense of a Self,”
forthcoming.
21. “After-cognition,” anuvyavasāya, is said to be a perception that takes a pre-
ceding cognitive occurrence as object. Matilal 1986: 141–179, “Knowing
That One Knows,” is a lengthy and penetrating discussion.
22. The assumption is that normally we do give the benefit of the doubt to our
informants. But there are also arguments to the effect that not doing so as
a general policy is incoherent. For example, Gaṅgeśa points out that those
who would communicate that testimony is not a knowledge source contradict
themselves: the beginning of the TCM testimony chapter, translation, p. 276.
See also Gautama’s sūtra 4.2.36 and commentaries, NyS, p. 1088, to the
effect that the concept of the non-veridical is parasitic on that of the veridi-
cal. The surrounding stretch of sūtras is directed against a radical illusionism
usually attributed to Mādhyamika Buddhism. Not everything can be a dream
or an illusion because the concept of a dream is parasitical on that of waking
experience. Vātsyāyana says explicitly in his commentary on NyS 4.2.34 that
the concept of the apparent whatever (as an apparent person which is really a
post misperceived in the distance) presupposes the concept of the genuine vari-
ety (formed from previous experiences of persons): NySBh, pp. 1083–1085.
Michael Dummett (1994: 265) writes: “To view the matter otherwise [than
that we always need a particular ground for declining to take things as we are
told that they are] is to subvert the whole institution of language.”
23. Gettier 1963.
24. Pollock 1986: 180–193. Pollock’s predecessors here include Ginet 1975 and,
as Pollock notes (181n4), Kleine 1971. BonJour 2003: 23 argues that the fail-
ure of epistemologists to solve the Gettier problems shows that the concept
of knowledge is obscure. It seems to me, however, that there are at least two
solutions, Pollock’s and Robert Nozick’s (1981), which overlap. I agree with
Pollock’s judgment (187) that the concept of knowledge is itself rather simple
but in requiring what he calls objective justification inherits the complexity
of social norms concerning justification.
25. Saha 2000: 62 identifies a Gettier case due to the Advaitin Śrīharṣa that
is taken up by Gaṅgeśa as well as the (Prābhākara) case of a deceived liar
(2000: 70) that is discussed here in our chapter on testimony. But he over-
looks the case I am about to discuss.
136 Notes
26. Gaṅgeśa and Mathurānātha discuss this under the rubric of the “reflec-
tion,” parāmarśa, required for inference: the TCM chapter on inference, pp.
544–547.
27. Nyāya employs a word, bādhaka, “defeater,” in ways that almost perfectly
match the usage in analytic epistemology. For example, negative coherence is
a belief-eliminator accomplished by bādhaka.
Now it is through the notion of “undefeated defeaters” that John Pollock
(1986) is able to cash out his notion of “objective justification” that he claims
is necessary for knowledge as is shown by the Gettier cases. Pollock’s analy-
sis is amenable to Nyāya’s point of view, it seems to me. In any case, let me
admit that my presentation of the Nyāya position relies on Pollock’s solution
of the Gettier problem. I have also discussed this and other Gettier cases with
traditionally trained Nyāya scholars (“pundits”) in India, including perhaps
the greatest living Naiyāyika, N. S. Ramanuja Tatacharya, who agree with
me in fi lling out the Nyāya analysis here.
28. This is sometimes called the new evil-demon problem for externalism. The
modern-day example of a brain in a vat is apparently due to Hilary Putnam
(1982: 5–7).
29. Vātsyāyana on NyS 2.1.16, p. 433, has a particularly clear statement about
this.

NOTES TO CHAPTER 2

1. pramāṇato ’rtha-pratipattau pravṛtti-sāmārthyād arthavad pramāṇam |

pramāṇam antareṇa na artha-pratipattiḥ | na artha-pratipattim antareṇa
pravṛtti-sāmārthyam | (pp. 1 and 21).
2. All the NyS commentaries develop the argument: pp. 1ff.
3. Nāgārjuna, Vigraha-vyāvārttinī 31, pp. 15–16 and 115.
4. Eli Franco (2004) argues that Nāgārjuna may well not have Nyāya in mind
but rather an early Buddhist epistemology since there is mention of a justifi-
cation regress in an early Buddhist manuscript. This, however, as Franco also
points out, runs against the grain of Nāgārjuna and Nyāya scholarship.
5. na pradīpa-prakāśa-siddhi-vat tat-siddheḥ, NyS, p. 443.
6. Structurally the Nyāya response seems similar to what is called infi nitism.
Potentially we could go on providing reasons for as long as you want. Since
the default is warrant at every level of challenge, the regress is not actual—as
argued by Sosa 1980: 9–11.
7. NySBh, pp. 448–449.
8. The overall strategy is “pragmatic contradiction” (e.g., someone saying, “I
am not speaking”) which is distinct from propositional contradiction (p and
~p) and from conceptual impossibilities (“a child of a barren woman”): see
the discussion at the end of this chapter. Gaṅgeśa says in his inference chap-
ter (Phillips 1995: 163): “It is the doubter’s own behavior that proves the lie
to the doubt, that blocks it (pratibandhaka).” See also Kisor Chakrabarti
1995: 200–203.
Nyāya would have a similar answer, it seems to me, to the problem known
as doxastic ascent, which concerns basic beliefs. Cannot any reason be chal-
lenged? There is a strong intuition to answer “Yes.” Of course, classical
Western foundationalism answers “No” and tries to identify beliefs that
are self-warranted and in principle immune from doubt and challenge. But
however basic beliefs are picked out, we would seem to be able to question
why a belief should count as basic. Thus the regress resumes. If, for exam-
ple, a belief is to be basic because it captures the content of an immediate
 Notes 137
experience, then why is it that capturing the content of immediate experience
should count as basic? Any response would open up the regress again, and so
we would ascend forever to infi nity instead of descending to foundations in
rock certainty.
9. See note 5 to Chapter 1.
10. B. K. Matilal (1986: 314): “Verbal reports, in Nyāya, are innocent until
proven guilty.” The statement is a little too strong, and I have accordingly
modified the principle although I, too, have said the same thing in some
earlier work. Remarkably, Uddyotakara under NyS 1.1.4 (p. 125) applies the
principle even to philosophy: “For it is a rule with systems (of philosophy)
that a position of another that is not expressly disproved (apratiṣiddham) is
(to be regarded as) in conformity (with one’s own).”
11. Partly that seems to depend on what is apperceived. Even Gaṅgeśa, who gen-
erally regards a scoped target’s objecthood to be known beyond the possibil-
ity of doubt, waffles with regard to an example of a long pastry, not with
regard to the objecthood of the scoped cognition or cognitions but with the
questions of cognitive unity and identification of cognitive type. From the
section on the mind’s atomicity from the perception chapter of the TCM, pp.
556 and 558 (translation slightly revised):
(Mīmāṃsaka) objector: How can there be (on your view) five (simultane-
ous) cognitions, as we fi nd with the eating of a long pastry (where at the
same time there is the pastry’s touch, its sight, smell, taste, and the sound
of chewing)?
Gaṅgeśa: The five would occur in (quick) succession.
Objector: No. For, without a prevailing counterconsideration, our sense
of the simultaneity of the cognitions is (to be counted) veridical.
Gaṅgeśa: Attention (to one thing specifically) is a counterconsideration, a
defeater of apparent simultaneity. . . .
Objector: How then (on your view) can there be the apperception, “I am
(simultaneously) cognizing a smell, taste, color, touch, and sound,” as
with the eating of a long pastry (as apperceived)? For (on your view) the
fivefold cognition would be serial such that there could be no simultane-
ous sensory connection with the manas (the internal organ).
Gaṅgeśa: No. The unity of your “apperception” is not demonstrated.
There would be (in the pastry case) precisely five apperceptions (“I am
cognizing a smell,” “I am cognizing a taste,” and so on). Because of the
subtlety (or indistinctness) of (separate apperceptive) occasions, the series
or order among them is (sometimes) not grasped.
Alternatively, due to a memory-impression left behind by (five) apper-
ceptions that have occurred sequentially, a memory whose object is those
five (past) cognitions occurs. When that memory gets misinterpreted
through superimposition of (the property) experiencehood (i.e., when it is
mistaken for an experience), the (pseudo) apperception, “I am cognizing
(five sensory presentations) simultaneously” occurs.
12. Jayanta, for one, is pretty explicit: e.g., Nyāya-mañjarī, pp. 43–44.
13. Udayana, Nyāya-kusumañjali 5.3 (pp. 578–579) shows that the concept of
the merely apparent and misleading, ābhāsa, extends to “suppositional rea-
soning,” tarka-ābhāsa, supplementing the mainstream concept of pramāṇa-
ābhāsa, the “pseudo-source.” The most common usage in philosophy is of
course the hetv-ābhāsa, “fallacy” or pseudo-inference, more literally, “an
inference imitator that appears to hinge on a genuine prover but does not.”
14. This is a corollary to the famous Vaiśeṣika thesis that everything is nameable
and knowable, PDS, p. 37, ably debated in Shaw 1978, Perrett 1999, Krishna
2004a: 282–294, and Ganeri 2011: 175–179.
138 Notes
15. Kisor Chakrabarti (1984: 350–351) draws our attention to a passage in
Vācaspati’s NyS sub-commentary that makes this point, and Matilal (1986:
164–166) has an extended discussion. Vācaspati says that some inferences are
so common as to be in effect self-certifying. We become so familiar in practice
(abhyāsa-daśā-āpanna) with one thing being invariably a sign of another that
we make inferences that are in effect self-confirming. From his Nyāya-vārttika-
tātparya-ṭīkā (p. 9): anumānasya tu pravṛtti-sāmārthya-liṅga-janmano ’nyasya
vā nirasta-samasta-vyabhicāra-śaṅkasya svata eva prāmāṇyam anumeya-
avyabhicāri-liṅga-samutthatvāt. “But an inference born from an inferential
mark proven through (repeated) action or another from which every bit of
doubt about the mark’s being erroneous has disappeared, is known of itself to
generate knowledge, since that it arises from an undeviating mark is itself infer-
able.” I take this to be an exaggeration, as the surrounding discussion makes
clear: pp. 7–10. Vācaspati like other Nyāya philosophers rejects the notion of
self-certification, as mentioned. But his point is that familiarity makes some
inferences practically self-certifying and immune from doubt.
16. For example, Vātsyāyana under NyS 2.1.34 (p. 497): We perceive wholes, not
atomic parts, although the sense organs are in contact with the atoms.
17. Roderick Chisholm (1942) concocts the example of a “speckled hen” (referred
to by Sosa 2003: 121) to bring out a similar problem with Western founda-
tionalist views. Probably we can see at a glance that a hen has three or even
five speckles on its face (this is called samūha-ālambana-jñāna by Nyāya
philosophers, a “group-supported perception”), taking in a precise number
of things in a single cognitive moment. But could we similarly perceive that a
hen has forty-eight speckles? Where to draw the line?
18. Richard Feldman (1985) breaks the problem down into a dilemma which
he calls the No-Distinction Problem, on the one horn, and the Single-Case
Problem, on the other. That is to say, to identify a broad process like infer-
ring would fail to distinguish degrees of justification whereas a specifically
specified process would fail to be repeatable, would not be a token of a reli-
able process type.
19. Padārtha-dharma-saṃgraha (PDS), translation, p. 658. Vātsyāyana has the
idea, too: NySBh 2.1.36, p. 511. That there is a “consecutive character”
(anuvṛtti) in identification of something as a cow is among several argu-
ments proffered against the Buddhist apoha “exclusion” theory of concepts
by Uddyotakara under NyS 2.2.64 (NySV, p. 667).
20. TCM, perception chapter, pp. 210–216 and 238–242.
21. Raghunātha, Padārtha-tattva-nirūpaṇam, text 49.2–4, p. 62.
22. Viśvanātha, Muktāvalī commentary on Kārikāvalī, (Shastri Sanskrit edition)
verse 51, p. 419.
23. TCM, perception chapter, p. 327.
24. Mathurāntāha, Māthurī (commentary on Gaṅgeśa’s TCM), perception chap-
ter (ed. Kamakhyanath Tarkavagish) vol. 1, p. 538.
25. Raghunātha, we may note, attacks the thesis that all veridical cognitions
divide into types of anubhava: Padārtha-tattva-nirūpaṇam, text 52.4–54.2,
pp. 66–67.
26. Udayana, Kiraṇāvalī, pp. 161ff. See also Dravid 1972: 26–33; Kisor
Chakrabarti 1975; and Phillips 1995: 60–63.
27. See, for example, in this book in the appendix (p. 126) Gaṅgeśa’s express
statement about analogy including perception and testimony as auxiliary
causes but being an irreducible knowledge source nonetheless.
28. Udayana, Nyāya-kusumāñjali, pp. 237–238; Gaṅgeśa, TCM inference chap-
ter (Goekoop translation), pp. 57–58; Mādhava, Sarva-darśana-saṃgraha,
pp. 12–13.
 Notes 139
29. For example, Vātsyāyana’s commentary on the tarka sūtra, NyS 1.1.40 (pp.
321–322), mentions belief in rebirth, common to both Buddhism and Hinduism,
as in tension with the (Buddhist) thesis that there is no enduring self (ātman).
Both sides present apparently good inferences. This draws the Nyāya inferential
knowledge into question. But here tarka re-establishes a presumption in favor
of an enduring self in the face of the challenge from the Buddhist anātman
theory: “If there were no enduring self, rebirth would be impossible.”
30. I do so, for instance, in Phillips 1995. Jonardon Ganeri (2001) seems to have
pioneered the translation, “suppositional reasoning,” which I like and have
adopted. Technically, tarka is classified as a kind of false cognition, one that
we realize is false while it is entertained. See the chart, p. 29.
31. This is why, by the way, despite the express mention by Udayana (Nyāya-
kusumāñjali 5.3), some Nyāya philosophers seem reluctant to admit anukūla
“favorable” tarka. There need not be a false supposition. My own supposi-
tional opinion in this regard is that there is supposition in anukūla-tarka,
too, a holding up of one’s own thesis into a subjunctive or suppositional
space: “Now what if p were true?” Compare Nozick 1981: 176 on subjunc-
tive conditionals and the requirements for a belief to track the truth.
32. Jonardon Ganeri (2000: 151–162) has an excellent discussion of tarka. I
translate a long passage from Śrīharṣa and responses by Maṇikaṇṭha and
Gaṅgeśa on tarka (1995: 151–164). Sitansusekhar Bagchi (1953) has a 300-
page book devoted to tarka as understood in all the major schools of classical
Indian philosophy.
33. Viśvanātha presents the list in a commentary on the NyS according to Bagchi
1953: 151.
34. For example, Udayana in his Ātma-tattva-viveka, (Sanskrit, ed. Dvivedin
and Dravida) p. 533; another passage from another edition, sva-vacana-sva-
kriyā-sva-jñāna, is quoted by Bagchi (1953: 178).
35. Translation from Phillips 1995: 160–161 (slightly modified).

NOTES TO CHAPTER 3

1. Under NyS 1.1.4, NyVTatp, pp. 111–120. Although Vācaspati is commonly

credited with introducing the distinction into Nyāya, it is also present in his
near predecessor Bhāsarvajña (also tenth century), Nyāya-sāra, pp. 19–21.
It is sometimes said (e.g., A. Chakrabarti 2000) that Vācaspati introduces
the concept of nirvikalpa into Nyāya under pressure of Buddhist thought.
There may be some truth in this, as will be elaborated. But Kumārila and the
NyS definition itself seem the more important precursors, considering the later
Nyāya concept at least. For, the Yogācāra view of perception as concept-free
is in the end quite distinct from that embraced by Nyāya. The Buddhist view
is tied to the thesis of the object of such perception as the pure particular,
sva-lakṣaṇa, “that which is its own mark.” Nyāya has the opposite view that
nirvikalpa picks up repeatable properties, universals, etc., although there is no
awareness of their predicative role at the time, as will be discussed.
2. NyS 1.1.4, p. 93: indriya-artha-sannikarṣa-utpannaṃ jñānam avyapadeśyam
avyabhicāri vyavasāyātmakaṃ pratyakṣam.
3. NyV, pp. 111–112.
4. Bhattacharyya 1990a: 175; Arindam Chakrabarti (2000) lists seven reasons
why all perception should be viewed by Nyāya philosophers as concept-laden.
5. Kumārila, Śloka-vārttika commentary on the pratyakṣa sūtra, verse 112, of
the Mīmāṃsā-sūtra, p. 94. Matilal (1986: 321–322) nicely illumines the sur-
rounding passage.
140 Notes
6. Kumārila, loc. cit., verse 120, p. 96.
7. I take the example from Sosa 1991: 141–142.
8. Bhartṛhari, Vākya-padīya, ch. 1, verse 123, p. 195 (“all knowledge is twined
with the word”). See also Matilal 1986: 342.
9. Gaṅgeśa, Raghunātha, Gadādhara, Jagadīśa, and company all use the san-
dalwood example: Gaṅgeśa, perception chapter, e.g., p. 124; Jhalakikar,
Nyāya-kośa, p. 163; Jagadīśa, Tarkāmṛtam, p. 31. This is supposed to be a
matter of common experience, vyavahāra, but it runs against the NyS thesis
that restricts the operation of sensory organs to their proper spheres. Of
course, Gautama and his commentators recognize a great deal of overlap,
depending on the characteristic, earthenness, for example, being presented
through all the senses whereas smells are (ordinarily!) grasped only by the
olfactory organ: e.g., NyS 3.1.68, p. 802.
10. NyS 4.2.26–36. The upshot is that, fi rst, the concept of illusion is parasitic
on that of veridical experience (not all coins can be counterfeit), and that,
second, illusion shows a combinational (propositional) structure—this is a
something or other. Perceptual illusion is right in part, that there is some-
thing there, but wrong about what it is.
Colonel Jacob (1925: 82–83) fi nds a “maxim” to this effect, which he
traces in various Vedāntic and Nyāya sources: sarvaṃ jñānaṃ dharmaṇy
abhrāntaṃ prakāre tu vyatyayaḥ.
11. Gaṅgeśa, TCM, perception chapter, pp. 627ff; Phillips 2001.
12. NyS 4.2.15, p. 1056, for instance. Mereology is a major theatre of engage-
ment in the war with Buddhist nominalists.
13. TCM, perception chapter, pp. 391–392.
14. Arindam Chakrabarti (2000) gives this and other reasons for jettisoning the
concept from Nyāya’s own realist point of view.
15. That is, as understood by Gaṅgeśa and other Naiyāyikas. Kumārila lays out
numerous arguments in the śūnya-vāda section of his SV, “On Emptiness,”
to refute the Buddhist position of an identity between cognized and cogniz-
ing cognition, and in verses 113–114 (Jha translation, p. 161) appears to
assert that the cognizing cognition is known partly from memory, partly
from perception, with an infusion—into the manifestation of a currently
cognized object—of the information that is has been cognized previously. At
the beginning of the section, the great Mīmāṃsaka, we might note, concedes
that the self can be known as both subject and object, but not a cognizing
cognition. In the SV section called ātma-vāda, “On the Self,” he asserts that
the self can be known by a direct cognition (SV ātma-vāda verses 125–126,
p. 405), but not the cognizing cognition, knowledge of which requires the
intermediary of an object known as cognized previously.
16. NyS 3.1.7–14 is the locus classicus for the topic.
17. Uddyotakara, NySV, pp. 726–736; Junankar 1978: 375.
18. The premiere Nyāya treatise on the self belongs to Udayana (eleventh century).
A long book with several commentaries, the Ātma-tattva-viveka, “Discrimi-
nation of the Truth about the Self,” is a sustained debate with a Buddhist
stream theorist as the principal adversary, although other players such as
the materialist Cārvāka are also engaged. (I discuss just one of his many
arguments in Phillips 2005.) This may be the most diverse as well as most
well-reasoned area of the whole of classical Indian philosophy. Every school
takes a position replete with a barrage of arguments, many of which seem
to be fi rst aired by the Mīmāṃsaka Kumārila in his SV ātma-vāda. For the
secondary literature, one might begin with Kuznetsova et al., forthcoming.
19. Sympathetic elaboration is to be found in Matilal 1986: 286–291.
20. This is the upshot of NyS 4.2.36 and commentaries.
 Notes 141
21. NySBh, pp. 1089–1090.
22. Dreyfus 1997: 299–315; Dunne 2004: 45–52.
23. Asaṅga, Mahāyāna-saṃgraha, ch. 2, section 14, vol. 2, pp. 104–107. The
point is elaborated in a Tibetan commentary translated in a footnote by Lam-
otte, pp. 105–106. Gods and Bodhisattvas see the “water” still differently.
24. Mādhava, Sarva-darśana-saṃgraha, p. 23.
25. Gaṅgeśa gives both the circularity and indifference arguments: TCM, per-
ception chapter, p. 220.
26. This argument seems taken over from Advaita Vedānta and Śrīharṣa in par-
ticular: Khaṇḍana-khaṇḍa-khādya, p. 40; Jha translation, para. 62; Granoff
1978: 107. As remarked several times in my study of Śrīharṣa and New Logic
(Phillips 1995), the later Naiyāyikas seem to include the Advaitin’s text in their
standard curriculum, relying on it almost like a genuine Nyāya authority.
27. Vācaspati appears to be the fi rst to develop a response, to wit, in his comments
under NyS 1.1.2, pp. 74–75. A more detailed response to the Prābhākara is
presented by him while wearing his Vedāntin hat: in his Bhāmatī commen-
tary on Śaṅkara’s BSB (catussūtrī), pp. 26–31. In the longest section of the
TCM’s perception chapter, Gaṅgeśa follows Vācaspati’s exposition, using
many of his images: pp. 250–326.
28. TCM, perception chapter, p. 259.
29. Kisor Chakrabarti 1999: 19–29 is an excellent discussion. The Nyāya mind/
body position is close to that of Ducasse 1951: 402–404.
30. Non-simultaneity of cognition is a large topic in the NyS. Attention is one
of several causal factors determining just what is cognized given that the
several sense organs are simultaneously in connection with objects whereas
only some of the objects are perceived. Further, any individual self is simul-
taneously in connection (through inherence) with a whole range of desires,
pleasures and pains, and memory dispositions whereas actual remember-
ing, feeling desire, etc., is selective and serial. See especially NyS 2.3.33 and
commentaries, pp. 863–866. Gaṅgeśa uses the example of the loud sound
trumping the power of attention: TCM, perception chapter, p. 563.
31. NyV, pp. 94–95.
32. As Gaṅgeśa puts it:
For different types of perception, different sensory relationships are
indeed required as uniform causal conditions. (1) Substances are grasped
through a connection (to the object perceived). (2) Through inherence-
in-what-is-in-connection, colors (and other qualities) and motions are
grasped. (3) Through inherence-in-what-is-inhering-in-what-is-in-con-
nection, colorhood (the universal of color) and the like are grasped. (4)
Through inherence, sound is grasped. (5) Through inherence-in-what-is-
inhering, soundhood (the universal of sound) and the like are grasped.
(6a) Through being-a-qualifier, absence of sound is grasped. (6b) Through
being-a-qualifier-of-what-is-in-relation-to-a-sense-organ, inherence and
such absences as of a pot are grasped. The sensory grasping (in each case)
results from the appropriate sensory connection, not from sensory con-
nection in general.
TCM perception chapter, p. 343. Slightly revised.
33. Preisendanz 1989 is a detailed discussion of the Nyāya-Vaiśeṣika view of
vision in the context of spelling out the failure of classical Indian civilization
to develop experimental science. Halbfass 1980 shows, as Preisendanz puts
it, that karmic “Unseen Force,” adṛṣṭa, came to “serve as a joker in the expla-
nation of otherwise inexplicable phenomena in physics in general, geology,
and meteorology.” But Preisendanz also herself argues that the philosophers
were the main contributors to optics after a time, breaking out of some of
142 Notes
the constraints of the ancient Āyur-veda and medical literature. Gaṅgeśa, we
may note, devotes an entire section of the TCM perception chapter to vision
and light: pp. 470–488.
34. Caraka, the great medical writer, talks about the sense organs as lining up
with the five material elements recognized by Nyāya: Caraka-saṃhitā, vol.
1, pp. 399, 403, etc. Here Nyāya conceptions seem, like Caraka’s, to start
off especially close to those of Sāṃkhya, although Nyāya has—and Sāṃkhya
does not—an atomic view of earth and the other elements.
35. Karl Potter (1977: 168–169) has a nice discussion, featuring the work of
Jayanta, of the development of the classification in both Old and New
Nyāya.
36. NyS, pp. 1090–1091.
37. Padārtha-dharma-saṃgraha, pp. 392ff.
38. Phillips 2009: 131–135.
39. One of the most important passages of Gaṅgeśa’s TCM turns on the notion
of familiarity: perception chapter (and commentaries), pp. 99–105. The
fi rst time one sees something with which one is unfamiliar (anabhyāsa-
daśāyām), for example, a river in the distance, it is possible to doubt the
thing (counter the Prābhākara position). But not after confi rming that it
is what it is does one on a later occasion doubt that the thing, with which
one is now familiar, is, say, a river. Familiarity builds confidence, and
the second time one sees the river from that spot one acts unhesitatingly
(niṣkampa-pravṛtti).
40. NyS 2.2.2, pp. 575–577.
41. The winning argument for Gaṅgeśa goes like this. It is simpler to view an
absential cognition as a single cognition of something, the absence’s locus,
as qualified by the absence—all information picked up through the sensory
connection with the locus—than it is to view it as the result of a whole new
knowledge source or as a matter of inference. We can also analyze the knowl-
edge as of the absence as qualificandum—which is a relational critter, join-
ing two relata. The locus is one relatum, and the absentee (the elephant or
whatever) is the other.
There are also many other arguments, as well as several interesting exam-
ples aired having to do with a subject Caitra coming to know about the
absence of Maitra at a certain location at a certain time through reflection.
For my own tastes, Gaṅgeśa’s section devoted to absential knowledge and
combatting the arguments of Kumārila is of all sections the most fun to read,
a dialectical masterpiece: TCM, perception chapter, pp. 407–435. See also A.
Chakrabarti 1997.
42. This seems to be implicit in what Gaṅgeśa says at various places: e.g., TCM
perception chapter, pp. 607 and 609. See also Sibajiban Bhattacharyya
1990a: 176. Arindam Chakrabarti (2000) fi nds this to be one more reason
that the concept of indeterminate cognition should be jettisoned (or should
have been jettisoned) by Nyāya philosophers.
43. BonJour 2003: 20.

CHAPTER 4

1. Oetke 2004: 175.

2. NyV 1.1.5, pp. 146–148. Uddyotakara is able to make this fit a common
pattern by considering the river to be made up of parts, upstream and down-
stream parts, such that it may be conceived as a single inferential subject or
 Notes 143
pakṣa qualified by an observed prover, swollenness here, and unobserved
probandum, being-recently-rain-fed-upstream.
3. The notion is evident, among other places, at NyS 1.2.5 along with
Vātsyāyana’s commentary, where the topic is the nature of the fallacy called
“deviation,” savyabhicāra: pp. 373–376.
4. I am grateful that so many have said so much that is illumining about this
portion of Nyāya theory. The studies that have most shaped my understand-
ing (by no means an exhaustive list of the excellent scholarship in the area):
Matilal 1998; Sibajiban Bhattacharyya 1987; Potter 1977 (“Introduction”);
Ingalls 1951; Staal 1973; Schayer 1933; Oetke 1996; Kisor Chakrabarti
1995; and Prasad 2002.
5. That is, according to Stanislaw Schayer (1933) and others in a reconstruc-
tion that is adequate for most purposes but not for the subtler points of New
Nyāya. J. F. Staal, Charlene McDermott, B. K. Matilal, and others present
more precise reconstructions representing properties occurring or not occur-
ring at a location or site: O(x, y) = y is located in x. The universe is composed of
properties and property-bearers and their relations. For later authors, “occur-
rence” should be replaced with “qualification”: Q(x, y) = y qualifies x.
According to Jonardon Ganeri (2001a, 2001b: introduction, and 2003),
instead of a step of inductive generalization and use of the two rules UI and
MP, the pattern can be seen to employ only a single rule of informal reason-
ing, as will be discussed in a later section.
6. In the more precise symbolism of Staal (1973) and others:
1. O(a, H)
2. (x) [O(x, H) → O(x, S)]
3. O(a, S)
7. Nyāya-bindu, p. 45.
8. Dharmottara’s commentary on Dharmakīrti’s Nyāya-bindu was long ago
translated by the great Russian scholar F. I. Shcherbatskoi (Th. Stcherbatsky):
see p. 47 for Dharmottara’s defense of the distinction. Prasad 2002: 26–40 is
thorough and penetrating in critique.
9. In Sanskrit, jñānātmaka vs. śabdātmaka: Nyāya-bindu-ṭīkā, p. 17.
10. But B. K. Matilal makes the summary judgment, “There is . . . no essential
difference in principle between these two types of inference” (1998: 108).
And Rajendra Prasad (2002: 27–28 and 39–40), perhaps noticing the Nyāya
attitude but focused on Dharmakīrti and company, decides against its ten-
ability even given Dharmottara’s own assumptions.
11. Daniel Ingalls (1951: 33) says just the opposite, speaking about New Nyāya.
However, he overlooks all the attention given to the parts or members
(avayava) of a formal demonstration as well as the treatment of fallacies.
12. John D. Dunne (2004: 256–260) explains the importance of “telic func-
tion,” artha-kriyā, a compound that is sometimes badly translated “causal
efficiency” missing the meaning of ‘artha’ as “aim” or “purpose.”
13. The epistemological context demands that a genuine pakṣa cannot already
be known to possess the probandum. Otherwise, we wouldn’t desire to know
whether it has it or not. To allow philosophical inferences to count as hav-
ing a veritable pakṣa when we do already know the conclusion from another
source (say, by perception, or apperception, which comes to be seen as a way
we know the self as a substance, for example, which is also knowable by
inference), Gaṅgeśa qualifies the requirement to suggest that tarkikas (“rea-
soners”) sometimes desire to know something by inference just for the sport:
TCM inference chapter, pp. 338 (p. 424 in the Calcutta edition) and 340
(431–432).
144 Notes
14. A more accurate reconstruction is presented by Ingalls (1951: 33ff) who has
lines two and three as proto-assertions he calls ascripts not requiring full
sentences.
15. Nyāya authors assume that their readers can reconstruct the vyāpti pattern
from a mere mention of the logical terms of pakṣa, “subject” or “site,” hetu
or sādhana, “prover,” sādhya, “probandum,” and dṛṣṭānta, “example” (sug-
gesting the general rule), and commonly use a three-part statement form
where a pervasion is understood (not expressly mentioned).
(part one: Sa) An inferential subject a is qualified by S a probandum
(part two: since Ha) since that same subject a is qualified by H a prover
(part three: like Hb and Sb), like b an example.
In other words, just by indicating a prover and a probandum through such
a statement, an author signals that a pervasion is asserted of the one by the
other (to wit, that all loci or sites of H are loci of S).
16. Uddyotakara, NySV, pp. 134 and 140.
17. The inference of classical epistemologists in general, across school, argues
Oetke (1996), should be understood as non-monotonic for essentially these
reasons.
18. Such Cārvāka skepticism is famously reported in the sixteenth-century text-
book of Mādhava, Sarva-darśana-saṃgraha, pp. 5–9, but is also addressed
in Nyāya texts.
19. Śrīharṣa, Khaṇḍana-khaṇḍa-khādya, p. 345 (translation, para. 417).
20. Ganeri 2001a: 25–35 and 114–121.
21. Vātsyāyana under NyS 1.1.5, a case that Ganeri (2001a) discusses at length.
22. Ganeri 2001b: 20 referring to Schayer 1933.
23. McDermott 1969, Staal 1973, Sibajiban Bhattacharyya 1987, and Matilal
1998.
24. Matilal 1998: 15.
25. A refinement made by Gaṅgeśa in the section of his inference chapter on “reflec-
tion” goes like this (TCM inference chapter, p. 367 (Calcutta edition, p. 493):
(K)(∃H)(SpHa) → (K)Sa
If S knows a site a as having some property or other as pervaded by another,
then S knows that a has that pervading property.
26. Maṇikaṇṭha Miśra, Nyāya-ratna, pp. 62–67. For example, from a single
observation of weaving a subject can infer that the cloth to be woven out of
blue thread will be itself blue: wherever blue thread, there a blue cloth. The
relation of emergent causality accounts for the pervasion and the knowledge
of correlation required for inference.
Maṇikaṇṭha’s example is of a pot inferred to be non-eternal because it has
been produced. One could grasp the rule here just from a single thing known
as both produced and destroyed, along with tarka.
27. Common inferences are practically self-evident, as we discussed in Chapter
2, p. 22.
28. For Gaṅgeśa on the dubious upādhi, see the translation by Phillips and
Ramanuja Tatacharya (2002: 110–119). I now summarize.
29. These include Udayana, Kusumāñjali 3.2, as noted by Ingalls (1951: 104),
as well as Gaṅgeśa, TCM inference chapter (Goekoop translation), p. 134.
Gaṅgeśa nevertheless modifies the pervasion defi nition proffered by his
“teacher” (ācārya), expressly rejecting Udayana’s formulation, pp. 89–90.
30. Kisor Chakrabarti (1995) argues this cogently.
31. Or, possibly they are influenced by Kumārila, whose Śloka-vārttika con-
tains a long passage with several arguments for a view of mutual dependence
between the universal and the particular: (Jha translation) pp. 281–294.
 Notes 145
Gadādhara’s Viṣayatā-vāda, “Discourse on Objecthood” (seventeenth cen-
tury), targets the topic.
32. A couple of good examples are the Buddhist momentariness inference whose
pakṣa is “all these things in dispute” (e.g., Ratnakīrti, Kṣaṇa-bhaṅga-siddhi)
and an inference of Gaṅgeśa’s put forth near the end of the upādhi section
of his TCM’s inference chapter (Phillips and Ramanuja Tatacharya 2002, p.
130) where the pakṣa is all inferences that fall to an upādhi!
33. For a brief history of the concept, see Matilal 1968 or Phillips 1995.
34. Matilal 2002, the last chapter, “Navya-Nyāya: Technical Developments in
the New School since 1300 AD,” contains a lucid discussion.
35. In this way is grounded ontologically the logical rule known as transposi-
tion, which is explicitly recognized by Gaṅgeśa and New Nyāya philoso-
phers, e.g., TCM inference chapter, the section on tarka, Tirupti edition, p.
207 (Calcutta, p. 241).
36. Raghunātha, Padārtha-tattva-nirūpana, pp. 60ff.
37. A substratum of a superstratum is related to it by inherence. A blue pot has
the pot as the inherent cause of its blue color. Bessie is an inherent cause
of cowhood. The inherent causes of a cloth are the threads of which it is
woven. Substances are not the only inherent causes, since universals inhere
in qualities and motions as well as in substances. Thus, a cognition, itself a
quality whose inherent cause is an individual self, is an inherent cause of the
universal, cognitionhood, which has multiple inherent causes, all the cogni-
tions in the universe! An emergent cause, in contrast, is always a quality or
a motion. A whole is said to inhere in its parts, and a quality that inheres
in the inherent causes of something x, a whole, can emerge as, or “cause”
(in this sense), a quality of x itself. Thus it is said to be the emergent cause
(sometimes translated “co-inherent cause”). For example, the blue color of
the threads of a blue piece of cloth is the emergent cause of the blue of the
cloth. All other necessary conditions for a type of effect are said to be instru-
mental causes, an axe for felling a tree, etc., and the operation of a sense
organ for a bit of perceptual knowledge. A cause, kāraṇa, is taken by default
to be a necessary condition. Any necessary condition for an effect of a certain
type may be inferred from the occurrence of an effect of that type, but only
the presence of all the conditions together, along with the most important
instrument cause, called the “trigger,” karaṇa, makes the effect guaranteed.
The trigger, sometimes translated “proximate instrumental cause” or “chief
instrumental cause,” is the unique condition which, on being met—after it
comes along and, so to say, adds itself in operation to a heap of enabling
conditions already in place—the effect necessarily comes about.
38. I take the example from Sibajiban Bhattacharyya (1990a) who takes it from
Donald Davidson. Sibajiban’s excellent essay on the New Nyāya view of cau-
sality is extensive, revealing a complex theory that deserves further explora-
tion: 1990a: 111–138.
39. The characterization of a universal by the early Vaiśeṣika philosopher
Praśastapāda is central to the mountains of later Nyāya discussion (PDS, Jha
edition, pp. 741–742): “A universal is present pervasively in all its instances;
it is identical in all its instances, of whatever number, and is the basis of the
comprehension (pratyaya) of itself recurrently; it is the basis of comprehen-
sion of inclusion inasmuch as it subsists wholly in each of its substrates.”
Among the copious secondary literature targeting the topic are excellent
studies by Dravid (1972) and Kisor Chakrabarti (1975).
40. Here I draw on my comments included with a translation of a section of
Gaṅgeśa’s TCM inference chapter: ed. Balcerowitz 2009.
146 Notes
41. Mīmāṃsā posits a special knowledge source called “circumstantial implica-
tion,” arthāpatti, whose instances Nyāya sees as usually the same as nega-
tive-only inference. Thus about the Mīmāṃsaka’s “Fat Devadatta does not
eat during the day, therefore he eats at night,” where F = “is fat but does not
eat during the day” and G = “eats at night,” Nyāya would say: Whoso F, that
person G, what is not so (F) is not so (G), like Maitra (who eats during the day
and not at night). This would be a “negative-only” inference so long as not
only has Devadatta not been observed to eat at night but also there is no one
else known to be like him in being fat and having been observed to eat only at
night. We do know that he eats at night (though this has not been observed),
and our inductive base is comprised only of negative correlations.
Similarly, what Buddhists following Dharmakīrti call inferential knowl-
edge based on identity, sva-bhāva, “self-nature,” or, later, “internal perva-
sion,” antar-vyāpti, would in many instances be interpreted in Nyāya as
negative-only inference. For example, “These are trees, since they are śiṃśapā
oak”:
pakṣa (subject) = śiṃśapā oak
hetu (prover) = śiṃśapā oak
sādhya (probandum) = tree
Taking the subject to be specified by what it is to be a śiṃśapā oak means that
there are no examples of the prover, which is to-be-a-śiṃśapā-oak, outside of
the subject or site. Thus the inference has to be negative only: whatever is not
a tree, is not a śiṃśapā-oak.
42. Gaṅgeśa makes a similar point concerning the name, “Ḍittha”: Tirupati edi-
tion, p. 500.
43. Hempel 1965: 12–25.
44. TCM inference chapter, Tirupati edition, 442.
45. Phillips 1995 traces many of these through the great Advaitin skeptic Śrīharṣa
and their influence on the development of New Nyāya.
46. TCM inference chapter, Balcerowitz 2009: 491.
47. Vācaspati, Nyāya-vārttika-tātparya-ṭīkā, ed. Thakur, p. 563 (commen-
tary under NyS 4.1.21): trayo hi bhāvā jagati bhavanti | prasiddha-cetana-
kartṛkāś ca . . . prasiddha-tad-viparyayāś ca . . . sandigdha-cetana-kartṛkāś
ca | “For there are three kinds of entity in this universe: those well-known to
have an agent as cause, those well-known not to, and those whose having an
agent as a cause is in doubt.”
48. It appears in Mokṣākāragupta’s twelfth-century textbook, Tarka-bhāṣā, p.
98.
49. Vācaspati, NyVTatp, ed. Thakur, p. 566.
50. ibid., p. 566.
51. Yoga-sūtra 1.26. Camille Bulke (1947: 27) opines that Vācaspati follows the
Yoga-sūtra here.
52. Vācaspati on why it is that it is the īśvara that must be supposed to be the
agent inferred: “(That which is to be accounted for) is the simultaneity of
production of effects throughout immeasurable and unlimited space at every
place and location, effects perceptible and imperceptible in animals and
plants and the organic world as a whole and so on.” NyVTatp on NyS 4.1.21,
ed. Tarkatirtha et al., p. 954; ed. Thakur, pp. 564–565.
53. TCM inference chapter, Tirupati ed., vol. 2, p. 420; Calcutta ed., vol. 2, p.
177.
54. We might note briefly that for Nyāya to be liberated is not to be like a stone,
contrary to the charge of Vedāntins. Nyāya takes a non-cognitivist view of
pleasure and pain or suffering, which have no intentionality or “objecthood,”
viṣayatā. Pleasure and company are instead objects (viṣaya) of internal
 Notes 147
perception. Thus a hedonic event would be simultaneous with a cognition
that grasps it, the two being in different streams of qualities (guṇa) that
inhere in or qualify an individual self. Strikingly, Gaṅgeśa implies that the
liberated somehow know they are free from future suffering. Somehow, one
can know one’s self as qualified by destruction of suffering, etc., as per a defi-
nition he gives. Thus cognition of such an absence of future suffering has to
be compatible with liberation. This is then one instance of positive content.
Presumably there would be more. Gaṅgeśa and Nyāya in general, with a few
exceptions (notably Udayana), subscribe to the possibility of “living libera-
tion.” The living liberated rather obviously have some sort of consciousness.
55. Bṛhadāraṇyaka, the third and fourth brāhmaṇas of the fourth chapter. Phil-
lips 1995: 39–41 takes up the Upanishadic passage and debate in Nyāya.
56. Solomon 1976.
57. Mahābhārata 12.320.78ff. Ester Solomon identifies and discusses the pas-
sage: 1976: 26–28.
58. Matilal 1998: 2.
59. NyS 1.2.2–3 and 4.2.48–51, pp. 356–361 and 1097–1099. This is noticed in
Hamblin 1970, which is one of the fi rst discussions by an analytic philoso-
pher of the NyS list of fallacies and futile rejoinders.
60. Vācaspati under NyS 1.2.1 says that a person of good character may put
forth tricky arguments, with unfounded premises, when an opponent (show-
ing bad character) first uses tricky arguments: NyVTatp, p. 339. Compare
this to a no-fi rst-strike policy with nuclear weapons.
61. NySBh 1.1.1, p. 44. This in itself comes to be counted as an argument against
a presumed “refutational” (prasaṅgika) anti-philosophy of Nāgārjuna and
followers.
62. For detailed support of this judgment with reference to a couple of dozen fal-
lacies commonly identified, see Matilal 1998: 60–87. Other good treatments
include Bandopadhyay 1977 and Gokhale 1992.
63. Kisor Chakrabarti 1999: 8–12.
64. It is significant, I think, that in later Nyāya the inference to self is considered
dispensable in that we know ātman directly, i.e., perceptually, as expressed in
such statements as, “I am having pleasure,” “I am having pain,” “I am per-
ceiving, inferring, etc., that . . . ,” and so on. Later authors make the argument
apparently following Kumārila in the Śloka-vārttika section on ātman, Jha
translation, pp. 401–407. Prior to the copious arguments of the Mīmāṃsaka,
however, Uddyotakara claims against the Buddhist anātman theory that there
is direct perception of the self, which is apparently expressed in such avowals
although not in sentences verbalizing false identifications of self, e.g., with
bodily attributes: see in particular NySV, p. 705, under NyS 3.1.1, Jha transla-
tion, p. 1081, for a summary statement, as well as Taber (forthcoming) who
reconstructs a supporting inference of Uddyotakara’s to the conclusion that
the self is known perceptually in case we have any doubt. The first-person
pronoun refers to the self and the pains, pleasures, and cognitions picked out
are its properties. We do not observe the whole self when we observe it apper-
ceptively as having a particular pleasure, just as we do not observe the backside
of Devadatta when we encounter him on the street while nevertheless knowing
the person perceptually. Similarly, with the theological inference there is sup-
posed to be (Vedic) testimony also in favor of God’s existence. And the mukti
inference, too, is supposed to back up, or be backed up by, yogic testimony,
although this presents a problem on some conceptions of mukti, as Uddyota-
kara points out: NySV introduction to NyS 1.1.2, p. 67.
This is, by the way, called the doctrine of pramāṇa-samplava, “coalescence
of sources,” which Nyāya accepts (with qualifications) but other schools,
148 Notes
especially Mīmāṃsā and Vedānta in their desire to defend the authority of
scripture, reject (each pramāṇa viewed as having its own proper scope or
range). Thus Vedāntins tend to reject rational theology since inference would
compete with scripture.
65. Chemparathy 1972 carefully reconstructs the “clusters of arguments”
Udayana puts forth in his Nyāya-kusumāñjali. But both there and in his
Ātma-tattva-viveka there remain many unmined bits of tarka and formal
inferences.

NOTES TO CHAPTER 5

1. Kumar 1980: 110.

2. The word ‘heaven’ (‘svarga’) is said to be derivative (yoga), involving a
narrowing of the meaning of ‘pleasure’, ‘sukha’, or ‘su’ in ‘su’ + ‘arga’, as
Gaṅgeśa himself tells us in his TCM chapter on analogy, addressing himself
to Mīmāṃsakas (p. 122). Nevertheless, cognizing similarity to earthly plea-
sures is crucial to the concept, it is recognized on all sides.
3. Two sūtras, NyS 1.1.34 and 1.1.35, expressly mention similarity (sādharmya
= sādṛśya) and dissimilarity (vaidharmya) in characterizing a proper prover
in the inferential context. The notion of “relevant similarity” is introduced
a little later, NyS 1.2.18, where a “futile rejoinder” is defi ned as an objec-
tion based on superficial similarity or superficial dissimilarity. See also, in
the appendix to this book, p. 112, an express statement from Gaṅgeśa that
knowledge of similarity is crucial to inference.
4. Not requiring the surety that would come from perceiving all the F-in-
stances through sensory relation with the F-hood universal, we can and do
defeasibly cognize a pervasion relationship, vyāpti, between two proper-
ties F and G, says Maṇikaṇṭha, Nyāya-ratna, pp. 62–67, as mentioned in
a previous note (Chapter 4, note 26). The standard view is that repeated
observations are required, bhūyo-darśana. Maṇikaṇṭha argues that the
bhūyo-darśana notion is vague and misses the main point, which is that
through the aid of “suppositional reasoning,” tarka, the causal relation-
ship between things F and G can be made manifest without a lot of cases
or experience. Gaṅgeśa accepts some but not all of Maṇikaṇṭha’s reasoning
and stresses, contra his predecessor, that a pervasion can be known without
tarka. Otherwise, there would be an infi nite regress, since successful tarka
presupposes knowledge of pervasion: TCM inference chapter (Calcutta),
pp. 210–212.
5. NySBh 1.1.6, pp. 170–171.
6. The term is introduced otherwise, suggests Gaṅgeśa (see in the appendix to
this book, p. 121). Apparently, there can be no analogical knowledge of the
meaning of ‘īśvara’ since that would require direct perception and the Lord’s
existence is atīndriya, “beyond the range of the senses.”
7. Bhāsarvajña, Nyāya-sāra, pp. 103–110. See also Narayanan 1992 and Kumar
1980.
8. NyS 1.1.6 and 2.44–48 and commentaries, pp. 168–172 and 527–533.
9. Compare Michael Dummett (1978: 122), from the essay, “Frege’s Distinc-
tion between Sense and Reference”: “ . . . a speaker may understand a com-
plex expression without knowing its reference—he knows the references of
the component words, and knows how they jointly determine the reference
of the whole, but does not actually know what that is.”
10. NyS 2.1.47, p. 530.
11. NyS 2.1.48, p. 531.
 Notes 149
12. The example becomes famous in the writings of the Mīmāṃsakas. Kunjunni
Raja (1969: 26–28) traces it in both Prabhākara and Kumārila as well as in
the grammarian Nāgeśa Bhaṭṭa (early eighteenth century). Gaṅgeśa disputes
the Prābhākara view in particular of how we learn the meaning of words,
but agrees that the original or primary way we learn is from our elders’
discourse: TCM testimony chapter, Calcutta ed., vol. 2, pp. 460 and 463
(badly translated by Bhatta, vol. 2, p. 645). Viśvanātha changes the example
to a child’s observing an elder bringing a jar when asked, (Madhavananda
translation) p. 152, and there are other variations all of which appear to be
forms of ostension (or training).
13. Kunjunni Raja 1969: 59–63.
14. Ganeri (1999: 31–38) examines at length the Nyāya notion of a verbal con-
vention, which he calls a mandate, rendering ‘icchā’, literally “desire” or
“will.” The standard Nyāya view is that, except for naming and introduction
of technical words in a science or system (paribhāṣika), the Lord, īśvara, sets
the conventions.
15. K. Kunjunni Raja lucidly discusses these matters: 1969: 59–69.
16. NySBh 2.2.62, p. 671.
17. As mentioned earlier, this comes to be called eka-vṛtti-vedya, “that which is
to be known in a single (cognitive) occurrence,” applied here to testimonial
knowledge as mediated by word meaning.
18. TCM testimony chapter, (Bhatta translation) vol. 2, pp. 715–720, in
particular.
19. This is mentioned by Gaṅgeśa in several places, e.g., in the appendix here, p.
121. Raja however, following Mathurānātha, leaves out technical stipulation
in his list of eight ways words are learned: 1969: 26–31.
20. A couple of the ways listed by Mathurānātha seem to be forms of testimony:
(1) usage of elders, (2) testimony, (3) derivative grammar, (4) analogy, (5)
dictionaries, (6) semantic sentential context, (7) written commentaries pro-
viding synonyms, and (8) syntactic sentential context.
21. TCM testimony chapter, (Bhatta translation) vol. 2, pp. 656–659. Here Gaṅgeśa
says that at first the child does not discern the individual words but takes the
sentence as a whole to mean that the cow should be brought. Later, by hearing
the words used in different contexts, the child learns to individuate words.
22. The correlative pronouns ‘that’ and ‘which’ (‘yat’ and ‘tat’) are said by
Gaṅgeśa to work similarly (p. 131): “The grasping of their referential power
is by indication (upalakṣaṇa) according to what has been set in the mind and
so on, which happens in common (for all things potentially referred to pro-
nominally, a pot, a cloth, and so on).”
23. Mill 1843: 30–40.
24. TCM perception chapter, pp. 652–653.
25. Gaṅgeśa quotes Udayana on the difference between a true qualifier and a
mere indicator (TCM perception chapter, p. 654):
Our revered teacher (Udayana) has said, “Whether an existent or a
non-existent, a qualifier is a differentiator that has the same locus as the
thing cognized. An indicator has a different locus.” What he means is as
follows. A qualifier makes known an exclusion occurring in precisely the
locus of that qualifier. That is to say, a qualifier makes known an exclusion
whose locus is specified by that qualifier.
After being cognized, both qualifiers and indicators make known a
differentiation that obtains at a particular locus. But crows mentioned to
point out Devadatta’s house do not distinguish Devadatta’s house with
respect to things that do not have crows but rather with respect to things
that are not houses.
150 Notes
The quotation is from Udayana’s Tātparya-ṭīkā-pariśuddhi commentary
under NyS 1.1.4: p. 145.

NOTES TO CHAPTER 6

1. TCM testimony chapter, p. 276.

2. NyS 2.1.68, p. 565: From the certified status of the Āyur-veda, which con-
cerns health, it is inferred that the other parts of the Veda, which concern
heaven and the like, are also true. Taber 2009 explores the argument.
3. Philosophically, Mīmāṃsā can be reconstructed around the (plausible) argu-
ment (Mīmāṃsā-sūtra 1.1.5) that dharma, “righteousness,” as a matter of
what we ought to do, not of what is being or has been done, falls outside the
province of perception and therefore also of inference. The revealed Word,
which is injunctive in character, fi lls the role, making us know what we
should and should not do.
The key connection between a Mīmāṃsā injunctive theory of meaning
and an indicative theory belonging to Nyāya is brought out by Arindam
Chakrabarti (2006: 35–38). The essential point is that the schema of qualifi-
candum qualified by a qualifier can be used to parse the meaning of injunc-
tions as well as of straightforwardly factual propositions, both embedded in
speech acts. Chakrabarti analyzes the injunction, “Bring the cow”:
Q(you, Q*(agency, Q(bringing, Q(accusativeness, cow))))
where ‘Q*’ is a “special intentional qualification” of the person being asked
to bring the cow.
4. āpta-upadeśaḥ śabdaḥ, NyS p. 173.
5. NySBh, pp. 173–176.
6. Under NyS 3.2.58, which concerns the rapid movement of the sensory mind
or manas as accounting for the apparent simultaneity of sensory awareness,
Vātsyāyana explains cognition of sentence meaning as a similar phenomenon
where we only seem to get it all at once: NySBh, pp. 897–898.
7. J. N. Mohanty (1994: 31) elaborates lucidly overall but says mistakenly that
the utterer (S) has to be known by the hearer (H) to be competent. For cer-
tification, perhaps, S may need to be known by H to be competent, but not
for fi rst-level testimonial knowledge, śābda-bodha. The mistake is repeated
in Mohanty 2000: 25.
8. TCM testimony chapter, vol. 1, p. 329: “Without knowing the sentence-mean-
ing fi rst, it is not possible to ascertain the speaker as not . . . confused.”
9. Chakrabarti 2006. Also, is there really such a shift? The reason that Gaṅgeśa
and company spend more energy on hearer conditions and sentence meaning
than on what it is to be an expert, āptatva, may be that it seems on the topic
of āptatva there remains, after Vātsyāyana et al., not much left to point out.
10. TCM testimony chapter, vol. 1, p. 316.
11. TCM testimony chapter, vol. 1, p. 140, tātparyyādhīnaṃ śabda-prāmāṇyaṃ
(the Sanskrit text, badly translated by Bhatta, and now retranslated:) “The
knowledge-source status of testimony depends on the intention of the
speaker.”
We might remark that if the parrot were like a tape-recorder Nyāya would
accept the knowledge-status of H’s belief. It’s not an accident that the parrot
says what it says. But the deceived deceiver case is a matter of purely acciden-
tal truth which is not pramāṇa-ja and thus not knowledge.
Or, maybe a case could be made that it is not an accident, that the truth,
so to say, breaks through the intention to deceive. Then there could be knowl-
edge (though I think it would be inferential not testimonial).
 Notes 151
Imagine that Desdemona does indeed love Cassio but Iago believes she
does not (her behavior is exactly as though she does not, as in the play). Cas-
sio for his own purposes tries to lie to Othello that Desdemona loves Cassio,
which in this case is true. The truth breaks through to Othello despite its
being through the medium of a deceived deceiver. However, I do not think
that Gaṅgeśa or any Nyāya philosopher would accept this as a case of knowl-
edge. Note that if Othello knew that Iago was out to deceive then belief about
Cassio would be blocked, prevented by that knowledge, but not if Othello
also knew that Iago was deceived (Othello having, so to say, the God’s-eye
point of view).
12. TCM testimony chapter, vol. 1, p. 329.
13. Under NyS 2.1.49–52, pp. 534–539 including the commentaries. Mohanty
1994 illumines the passage.
14. Jayanta Bhaṭṭa sums it up (Nyāya-mañjarī, p. 322): “The conditions that
determine inferential knowledge and those that determine verbal knowledge
are not the same.”
15. Mohanty 1994: 34; TCM testimony chapter, vol. 1, pp. 277–348 (the
Vaiśeṣika refutation); Vātsyāyana, NySBh 2.1.20 (translated here, p. 17).
16. Vācaspati, NySVTatp (Dravid edition), pp. 201–207, referenced by Mohanty
1994: 49.
17. Hume’s focus is on the overall evidence for or against a claim (of a miracle)
attested by testimony, not on the epistemology of testimony itself. But he sees
the trustworthiness of testimony as founded on the inductive generalization
that what people tell us is (normally) true. Consonantly, “a man delirious, or
noted for falsehood and villany, has no manner of authority with us.” Hume
1748: 84–85.
18. Price 1969: 112–129.
19. Strawson 1994: 27.
20. Graham 2006: 93.
21. NyS 2.2.58–69, pp. 654–695, including the commentaries.
22. NySVTatp, p. 683; see also the note by Jha to his NyS translation, vol. 2, p. 1035.
23. “(This is) Devadatta” is analyzed as a bit of knowledge like “(This is a) cow”:
a qualifier, being-named-Devadatta, qualifies an individual qualificandum:
see, e.g., Sibajiban Bhattacharyya (1990a: 162) referring to Gadādhara’s the-
ory. Similarly, cowhood qualifies Bessie. There is of course the metaphysical
difference that Devadatta could have another name and remain the same
person whereas Bessie cannot lose her cowhood while she is alive. But the
present point is not metaphysical but has to do rather with how the cogni-
tion’s intentionality (“objecthood,” viṣayatā) is to be analyzed.
24. For example, Kumārila, Śloka-vārttika (Jha translation), p. 506 (verse 121
of the section on the sentence, under Mīmāṃsā-sūtra 1.1.24), “Words . . . are
capable of making up an endless number of sentences.”
25. Siderits 1991.
26. Adverbs do not carry grammatical terminations but are just in that absence
counted as inflected, too.
27. Gopinath Bhattacharya writes àpropos Annambhaṭṭa’s discussion of the two
Mīmāṃsaka theories (Tarka-saṃgraha, pp. 301–302):
It comes to this then that the understanding of a statement, i.e., of what is
signified by the constituent terms in relation to one another, depends among
other things on the presentation of the terms in the required order. But the
order of arrangement of the terms is not itself a term of the sentence, so that
it cannot be said that this order has its own śakti like the terms.
28. Siderits 1991: 40. Gaṅgeśa expressly addresses the objection that on his view
a sentence would be indistinguishable from a word list: TCM testimony
152 Notes
chapter, pp. 356–357. The answer is again that though a sentence does not
mean separately a relatedness (anvaya)—it means just the things meant by
the individual words—there is a relatedness among the things referred to, a
relatedness reflected in how the sentence structures the meanings. The relat-
edness is supposed to be an ontological free lunch, like the self-linking nature
of inherence and other qualificative relations. Phillips 1995: 134–137 dis-
cusses this; see also Potter 1961.
29. Siderits (1991: 103–110) speculates that Buddhist philosophers, who are
almost silent on these issues, would align themselves with the Prābhākara
holism. But note that on other issues it is Nyāya that is holist, opposed to
nominalist reductions proposed in particular by Buddhists. And indeed there
is, despite the official position that a sentence’s meaning is constituted by the
meaning of its words, a holism here with the Bhāṭṭa/Nyāya theory, the holism
of sentence meaning, which Nyāya views as not just a sum of the referring
parts, the constituent words. The sentence as a whole draws our attention to
a certain relatedness.
30. Kumārila mentions “context,” prakaraṇa, textual context, as crucial to
removing vagueness in understanding mantras: Tantra-vārttika, pt. 2, pp.
1155–1158. As we noted in the previous chapter, Vātsyāyana mentions
prakaraṇa for determining precisely what a word means.
31. Kunjunni Raja 1969: 149–169.
32. This point is expressly made by Gadādhara in his Vyutpatti-vāda according
to Ramanuja Tatacharya 2005: xxxv (and in Sanskrit, p. 425), who does not
give a more specific reference.
33. Grice 1975. “Do not speak nonsense” would appear to be a “maxim of qual-
ity” akin to Grice’s “Do not say what you believe to be false.”
34. Kunjunni Raja 1969: 164–166.
35. TCM testimony chapter, vol. 1, pp. 372–373.
36. Arindam Chakrabarti 2006: 48.
37. TCM testimony chapter, my translation, Bhatta edition, IV.iii.6, vol. 1, p.
136.
38. Compare Wittgenstein’s remarks on Moorean statements of certainty: Wit-
tgenstein 1969, paragraphs 111–153.
39. Gaṅgeśa, TCM testimony chapter, vol. 1, p. 294.
40. Uddyotakara, NySV 2.2.62, p. 664.
41. Kunjunni Raja (1969: 166) reports this as the Nyāya position, but, as we
shall see, that’s not precisely correct.
42. Gaṅgeśa makes the point: TCM testimony chapter, vol. 2, pp. 794–795.
43. S. Bhattacharyya 1990a: 33–34.
Edwin Gerow (2001: vi) points out that a “figure” for the Sanskrit rhetori-
cians is “a mode of thought, a way of formulating or conveying an idea, which
has closer relations with logic than with grammar, or even semantics.”
44. NyS 2.2.62, pp. 663–664 including the commentaries.
45. In the later alaṃkāra-śāstra, figures are divided into those where the relation
is similarity, called gauṇī, “secondary (figurative speech),” and a miscella-
neous collection of relations other than similarity, called śuddhā, “pure (fig-
urative speech),” e.g., in Appaya Dīkṣita, Vṛtti-vārttika, p. 32. The former is
admitted as still another power of words by some Mīmāṃsakas, according
to Kunjunni Raja 1969: 240–241.
46. TCM testimony chapter, my translation, Bhatta edition, IV.ix.42, vol. 1, p.
232: vākya-artha-anvaya-anupapattyā, “by the impossibility of the connect-
edness of the sentence meaning (without understanding a figurative sense).”
47. TCM testimony chapter, vol. 2, p. 794.
48. Annaṃbhaṭṭa, Tarka-saṃgraha, pp. 294–295.
 Notes 153
49. TCM testimony chapter, vol. 2, pp. 802–803.
50. Kunjunni Raja 1969: 262–264.
51. The neo-Vedāntic view popularized by Swami Vivekananda (1863–1902) is
that the Veda resounding in the cosmic ether (ākāśa) was heard and memo-
rized (in this cycle of creation) by great ṛṣis in their pellucid consciousness
(not cluttered by ordinary thought). This seems a concession to the Western
concern with history, but is not out of line with how at least some Naiyāyikas
think. Classical Mīmāṃsā, however, unlike Nyāya, does not endorse the
view that the universe (or the Veda) has an origin. “Just like us, just like
now” is Kumārila’s position on earlier generations’ possession of the Vedic
corpus. They got it, just like us, from their ancestors, in immemorial lin-
eages. See McCrea (2009) who cites Kumārila’s aphorism and explores the
Mīmāṃsaka’s arguments.
52. Gaṅgeśa, TCM testimony chapter, vol. 1, pp. 290–291; Annaṃbhaṭṭa as
interpreted by Lakṣṃīnṛsiṃha (according to a note by Gopinath Bhattacha-
rya), Tarka-saṃgraha, p. 306.
53. Annaṃbhaṭṭa, Tarka-saṃgraha, pp. 289–293.
54. This is clearly evident in the case of irony or “crooked speech,” vakrokti, e.g.,
“Your kindness is celebrated,” said by a lady to her lover whom she knows
has made love to her messenger: Appaya Dīkṣita, p. 68, last line. (Gerow
1971: 286 has this classified as vyāja, “pretense.”)
55. Ānandavardhana, Dhvanyāloka, p. 83: “Go your rounds freely, gentle
monk;/the little dog is gone./Just today from the thickets by the Godā/came
a fearsome lion and killed him.” This is said in a play where the speaker com-
municates (to the audience, not to the monk) that she plans to meet her lover
“in the thickets by the Godā” and does not want the monk to disturb them.
56. Annaṃbhaṭṭa, Tarka-saṃgraha, p. 291.
57. Annaṃbhaṭṭa, Tarka-saṃgraha, p. 293.
58. Ānandavardhana, Dhvanyāloka, p. 132 (dhvani is equated there to “appre-
hension of beauty”).

NOTES TO CHAPTER 7

1. Lehrer 1990: 169–172.

2. Armstrong 1973, Lehrer 1990, and BonJour 2003. Armstrong (1973: 171–
175) worries about cases of veridical hallucination as possibly undermining
his “thermometer” account of non-inferential knowledge, acknowledging
the difficulty of spelling out the lawful relation between the fact that p and
the belief that p. The “Truetemp” objection to Armstong’s causal view is
formulated by Lehrer (1973: 162–164).
3. BonJour 2003: 28–33.
4. Nozick 1981: 179.
5. There is much to say about the relationship of bits of knowledge generated
from different sources. Does apparent inference trump apparent percep-
tion? Surely not in all cases since sometimes we confi rm inferential results
perceptually. Nyāya, unlike Vedānta in particular, accepts the principle of
pramāṇa-samplava, which says that the same fact can be known in mul-
tiple ways. Thus there is the possibility of source dispute, a phenomenon
recognized by Nyāya philosophers who do not come up with, however,
so far as I can tell, general rules. (Maybe there are none.) To be sure, it is
often said that perceptual evidence is the weightiest (but how about for the
movement of the sun, which is mentioned by Vātsyāyana himself?). Further,
one source can certify the output of another source. This is another side
154 Notes
of the pramāṇa-samplava doctrine. But the full ramifications of this plank
of the theory Nyāya philosophers fail to explore, in my opinion. Accepting
pramāṇa-samplava, they are perhaps too sanguine about the unlikelihood
of pramāṇa confl ict.
6. Nozick 1981: 681n9 traces the case to Carl Ginet. Philip Quinn elaborated
it in comments on presentations made by Arindam Chakrabarti and myself
at the American Philosophical Association Eastern Division meeting in New
York, 2000.
7. Another case, due to Gilbert Harman (1968: 172, referred to by Pollock
1986: 190 and reworked by Nozick 1981: 177), seems similar though it is
geared to bring out the “social dimension of knowledge” (a point Harmon
attributes to Ernest Sosa 1964 and with which Nyāya would agree, it seems
to me). By Nyāya lights the case of the “dead dictator,” as it may be called, is
not knowledge however for the simpler reason that there is no genuine testi-
mony. Our subject S, who lives in a country where a practically all-powerful
dictator rules, reads the early-morning edition of the local newspaper whose
headline is: “Dictator Dead” (with some name substituting for ‘dictator’).
The dictator is in fact dead. The dictator’s henchmen realizing that they do
not want the public to know that the dictator is dead force the paper to print
a retraction. S walks into a room where there are twenty people all of whom
have read the second edition and do not believe that the dictator is dead.
Before entering the room and talking with the people, S would seem to have
justified true belief that the dictator is dead, although she will be shaken off
the belief after a minute in the room. Does S know that the dictator is dead
just before she walks into the room? Those who answer no apparently feel
that veritable knowledge should be better rooted psychologically. By Nyāya’s
lights, the newspaper story is not a source of knowledge in the fi rst place
since no newspaper in a dictatorship could meet the requirements of “trust-
worthy authority,” āptatva.
8. A legitimate challenge undermines a position’s status as knowledge even if
it is true and we have good reasons for holding it, according to the NyS and
commentaries. This then becomes an occasion for the employment of tarka,
much like the elenchus of Socrates.
Thomas Kelly (2005) distinguishes epistemic peers from epistemic inferi-
ors and superiors, and defends what he calls epistemic egoism with respect to
disagreements among peers. Peerage seems to be the attitude taken by most
Nyāya philosophers towards their Buddhist and Mīmāṃsaka adversaries in
particular (though not towards everyone, for example, Cārvāka, who is by
tarka humiliated).
9. Pingree 1981: 56–66.
10. Raghunātha, Padārtha-tattva-nirūpaṇam, pp. 86–87. Lucid discussions of
the late Nyāya views of counting and number may be found in S. Bhattacha-
ryya 1987, Perrett 1985, and Ganeri 1996.
11. Cassam 2007: 189; Burge 1998; and Goldman 2002.
12. Cassam 2007: 190.

NOTES TO THE APPENDIX

1. A red truck and a red tomato would have little in common other than color.
2. I translate ‘sāmānya’ as “universal” although the Prābhākara dos not mean
here that similarity is a universal as Nyāya philosophers understand univer-
sals. Other acceptable translations are “natural kind,” “class character,” and
“commonality.”
 Notes 155
3. According to Kṛṣṇakānta (Calcutta, p. 19), D1 is attributable to the Bhāṭṭa
Mīmāṃsaka, distilling, apparently, Kumārila’s Śloka-vārttikā under sūtra
1.1.5, section 7, verses 18–21, (Jha translation) p. 225.
4. “Thus it has been said” is taken by the commentator Pragalbha to imply that
there is agreement between the Prābhākara and his Bhāṭṭa objector about the
verse’s thesis.
5. Occasionally throughout the TCM, Gaṅgeśa expressly mentions “New”
(navya) Prābhākaras who propose views slightly innovative relative to those
of Prabhākara himself and of his close followers such as Śalikanātha Miśra.
6. The idea seems to be that pleasures are too different from one another to
instantiate a single pleasure universal, unlike, say, cows. Gaṅgeśa appears to
hold that there are fundamentally distinct pleasure kinds and that “pleasure-
hood” is not a genuine universal.
7. This is a thorny passage. The commentator Kṛṣṇakānta (Calcutta, p. 37)
understands the compound ‘kāraṇa-viśeṣa-prayojyā’ (“directing a particu-
lar cause”) as vihita-kriyā-viśeṣa-prayojyā (“directing particular ordained
actions”). He glosses Gaṅgeśa as saying that enshrined in everyday speech is
the idea that there is similarity in the actions that are enjoined by scripture
as well as in the pleasures produced. The same goes for the pains produced
by sins. There are universal laws of karma, not mediated by generational
considerations, as we might say today.
8. On an alternative reading of the Calcutta edition, there is ‘dvayor’ instead
of ‘pralaye’: “universals remain even when the two are destroyed.” The
“two” would be the two particulars already mentioned. Universals endure
the destruction of individual instances, even of all instances so long as there
have been instances. Nyāya, it seems to me, embraces a realist view about the
past.
9. Having completed his discussion of the ontology of similarity, Gaṅgeśa now
takes up analogical knowledge (upamiti) as his principal topic through the
voice of a new adversary. The new principal objector, objector.1, may be
understood, following the commentaries, as a Bhāṭṭa Mīmāṃsaka, from the
school of Kumārila and company, who eschews similarity as a separate cat-
egory, like Nyāya, but, unlike Nyāya, accepts non-perception as a separate
knowledge source, the source that makes us know absences.
10. Kumārila himself addresses Nyāya’s view of analogy, Śloka-vārttika, (Jha
translation) pp. 222–230, and the Nyāya-promoted scenario of a subject S
learning from a forester enough to know in his encounter with a gavaya buf-
falo that it is called what it is. Note, however, that the proposition known
will concern, in the Bhāṭṭa’s mouth here, the cow, not the gavaya, as with
Nyāya.
11. See in this book p. 63 for an explanation of the connection to which the
Naiyāyika refers. In brief, this is a type of sensory connection, specifically a
connector that allows us, according to the theory, to have a grasp of future
instances of a universal, cows not yet born. The nature of the connection is
“peculiar to universals,” sāmānya-lakṣaṇā-pratyāsatti, so it is said.
The voice of the main opponent appears now to be different, not that of
a Bhāṭṭa Mīmāṃsaka as just above, but rather a Nyāya philosopher of the
Old school. The commentator Kṛṣṇakānta identifies the voice as Old Nyāya
(Calcutta, p. 55) with respect to Gaṅgeśa’s use of the word “our” (asmāka) in
talking about what “they” (the Mīmāṃsakas) call “circumstantial implica-
tion” (arthāpatti). And defi nitely there is a change of voice for the principal
opponent. The question is precisely where? Earlier, circumstantial implica-
tion was said to be “our” (Bhāṭṭa) view which “they” the Naiyāyikas see as
the form of inference they call “negative only.” So if it is not a new principal
156 Notes
opponent that we have here, then the change occurs with “That’s contempt-
ible” (vigarhitam etat) a little below.
12. The “opposed terms” of which the objector speaks are, for example, Deva-
datta as qualified by thisness, i.e., his being present before our subject S, and
the same Devadatta as qualified by thatness, e.g., his being encountered by S
yesterday in the market.
13. It’s hard to determine who the “they” refers to in “so they say,” appar-
ently fuzzy-thinking Mīmāṃsakas. The objector is probably intended to be
thought of as a Mīmāṃsaka who cites mindlessly a maxim of his teachers.
The objection in the next passage probably belongs to Old Nyāya. Just a little
later, Gaṅgeśa himself (not just his commentators) mentions Jayanta Bhaṭṭa
(ninth century) and Old Nyāya.
14. Here I follow the Calcutta edition as the Varanasi edition drops the word for
elephant such that it would be cognitions that are known perceptually to be
mutually similar.
15. If only as combined A and B can do their work, then their combination can-
not be a causal factor for A or for B.
The time delimitation mentioned is supposed to be a matter of the differ-
ence between (a) S’s hearing the analogical statement at t1 and (b) seeing the
gavaya buffalo at t 2 and having the bit of analogical knowledge.
16. This is to follow the Calcutta edition which has ‘kathora-śūka-āśinaṃ’.
17. Having pointed out what’s wrong with the Old Nyāya theory—all in all a
rather minor oversight—Gaṅgeśa articulates now a fully adequate view in a
siddhānta of New Nyāya.
18. “Ether,” ākāśa, is posited in the technical terms of the Nyāya-Vaiśeṣika
ontology as the medium of sound although we do not directly experience it.
Sound travels and so must have a carrying substratum connecting things that
are making sounds with the audial organ located in the ear canal.
19. Etymological explanation does not entail that the action of the root from
which the word is derived is actually occurring. Thus we can use the word
‘cook’ for a person who is not currently cooking.
20. The Calcutta edition adds ‘api’ after ‘agre’.
21. On Gaṅgeśa’s view, the cow-similarity cognized by S upon encountering a
gavaya buffalo is not known at the time of S’s comprehension of the analogi-
cal statement. This view is tied to the supervenient nature of the similarity
according to him.
22. A stock example of indirect, figurative meaning, “The village is in (the water
of) the Gaṅgā,” contrasts the literal meaning, “The village is built on the
water of the river,” which is known to be impossible, with the intended,
metaphorical meaning, which is to state the village’s proximity to the sacred
river and to suggest that it has a meditational atmosphere, as is discussed in
this book in Chapter 6.
23. Gaṅgeśa points out that it is the gavaya buffalo that is referred to in both the
subject and predicate parts of the sentence that verbalizes S’s knowledge. The
subject expression and the predicate expression, which are appropriately in
the same case, both pick out the gavaya.
24. Intended meaning, in objector.1’s conception, includes unexperienced
instances of universals such that when the class character has been compre-
hended by S, the individual gavaya (or whatever), though not yet encoun-
tered, is known in a way.
25. The mutual dependence would be using analogy to explain a bit of knowl-
edge that can be explained otherwise, and then turning around and using the
example and its explanation as a reason for positing analogy.
 Notes 157
26. S knows that the word ‘gavaya’ is a noun from overhearing “students”—a
word, we may remark, that contrasts with the “elders” of the stock Mīmāṃsaka
example where one learns through the ostensions of one’s elders, the authori-
ties. The Sanskrit word for “student” is ‘śiṣṭa’, “one who is taught.”
27. The objector puts forth a “negative-only inference” (a is S, because it is H,
whatever is not S is not H).
28. In the NyS, three main types of inference are differentiated according to the
relationship between the inferential terms and how that relationship is evi-
dent. The sāmānyato-dṛṣṭa type of inference is said to work on the basis of
knowledge of common characteristics, as opposed to causality with the other
two. It has a conclusion that is in fact not known by current perception, like
all inferences, but also, unlike the other varieties, an imperceptible conclu-
sion, such as “The sun moves,” which is known although no one ever directly
witnesses the motion. Later there is much dispute about the inference type,
and the concept is more or less dropped by New Nyāya philosophers such as
Gaṅgeśa in favor of a different classification. See in this book the third sec-
tion of Chapter 4, “The ontology of pervasion.”
29. The Nyāya inference to a self is given extensive treatment by Gaṅgeśa in the
“negative-only” section of the inference chapter of his Tattva-cintā-maṇi.
See the fourth section of Chapter 4 here, “Philosophical inferences to the
self, God, and mukti.” Gaṅgeśa’s point, however, is simply to show a lack of
parallelism between the way the self is known inferentially and how S knows
what ‘gavaya’ means.
30. Gaṅgeśa’s main point is that once we know something we don’t inquire fur-
ther. But also he says again here, notably, that S would know in general,
sāmānyataḥ, before encountering a gavaya buffalo, what the word means.
31. Understanding something in general about a gavaya would not prevent S,
without an actual encounter, from asking further questions or of appropri-
ately being told by the forester to go and see for himself what a gavaya is.
32. Here Kṛṣṇakānta discerns the end of the pūrva-pakṣa correctly (Calcutta, p.
89). The Varanasi edition does not.
33. Doesn’t S comprehend from the forester’s analogical statement what it is to be a
gavaya buffalo? Gaṅgeśa definitely says no, implying that the universal being-a-
gavaya-buffalo as well as particular buffalos are not known, at least not entirely,
without the actual experience. However, we have to distinguish comprehension
in general and in particular, and Gaṅgeśa has told us earlier that there would
be the former (though not the latter) from S’s grasp of what the forester says.
Thus understanding what it is to be a gavaya buffalo in general (sāmānyataḥ)
is not the same as comprehending the universal, being-a-gavaya buffalo, but
rather a vague or incomplete or proto-understanding through knowing indicator
characteristics common to being a cow (e.g., having-horns). Gaṅgeśa seems to
waffle concerning whether it is the universal or the particular that is the grounds
(nimitta) for a word’s usage. But the truth is that it is both, as Gautama and
Vātsyāyana maintain. Whether the one or the other is predominant is said to
depend on context and other factors (see, in this book, pp. 77 and 87).
34. The standard Nyāya view, stated by Gaṅgeśa earlier, is that the aids to com-
prehension of a word include grammar and dictionaries. Gaṅgeśa is claim-
ing that for the example on the table none of the common aids is sufficient
for comprehending the fact that the thing is front is called gavaya, which is
known instead by analogy.
35. Kṛṣṇakānta supplies the rule upon which the objection draws (Calcutta, p.
96), “As there is knowledge of something according to a uniform (repeatable)
character (F), so there is grasping of referential power to just that (F).”
158 Notes
36. See the discussion of sentential meaning at the end of section, “Statements
and facts,” in Chapter 6. Here Gaṅgeśa relies in particular on semantic fit-
ness, yogyatā. This involves a word’s sense determining real possibilities of
sentential meaning in that a meaningful sentence has to be possibly made
true by a fact.
Sanskrit Glossary

1. PROPER NAMES (PHILOSOPHIC SCHOOLS AND SOME

IMPORTANT CLASSICAL AUTHORS AND TEXTS)

Advaita Vedānta: sub-school of the Upanishadic philosophy of Vedānta that becomes

a whole school to itself, subscribing to a spiritual monism, “All is Brahman,”
including—and especially—the seemingly individual consciousness or self.
Ānandavardhana: (c. 850) author of an influential work of aesthetics, Dhvany-
āloka, where “suggestion,” dhvani, is championed as a third power of words, in
addition to denotation, abhidhā, and indirect, figurative meaning, lakṣaṇā.
Annambhaṭṭa: (c. 1650) Nyāya philosopher whose popular textbook treatment of
Nyāya, Tarka-saṃgraha, is elucidated by an auto-commentary where several
innovative positions and arguments are advanced.
Bhartṛhari: (c. 450) grammarian and philosopher of language celebrated for a
sphoṭa—“comprehension in flash”—theory of sentence comprehension, a Pla-
tonist theory of universals, and an idealist metaphysics of a “word” brahman.
Bhāsarvajña: (c. 950) a slightly maverick Nyāya philosopher whose auto-commen-
tary Nyāya-bhūṣaṇa (“nyāya ornament”) on his Nyāya-sāra (“nyāya essence”) is
a major treatise outside the Nyāya-sūtra commentarial literature of Old Nyāya.
Bhāṭṭa: a follower of the Mīmāṃsaka philosopher Kumārila Bhaṭṭa or a view
belonging to or deriving from Kumārila.
Caraka: (c. 200 bce ?) legendary author of an important medical text.
Cārvāka: classical philosophic school of materialism, religious skepticism, and
hedonism, famous in epistemology for attacking inference as a knowledge source,
alleging the unknowability of inference-underpinning pervasion, vyāpti.
Dharmakīrti: (c. 650) Buddhist logician and epistemologist of the Yogācāra school,
author of Pramāṇa-vārttika and other major works of Buddhist philosophy.
Gadādhara: (c. 1650) New Nyāya author famous for his linguistic theorizing and
analysis of the concept of “intentionality,” viṣayatā, “objecthood.”
Gaṅgeśa: (c. 1325) systematizer if not the founder of “New Nyāya,” navya nyāya,
author of the influential Tattva-cintā-maṇi (q.v.).
Gautama: (c. 200) legendary author of the sūtras of the Nyāya-sūtra, Nyāya’s
founding text, sometimes called Akṣapāda (“lost in thought”).
Jaimini: (c. 100 bce) legendary author of the Mīmāṃsā-sūtra.
Jaina philosophy: a tradition commencing with Mahāvīra (c. 450 bce) with dozens
of texts in all periods, probably most famous for an ethics of ahiṃsā, “non-
injury,” and non-absolutism (anekānta-vāda) in metaphysics.
Jayanta Bhaṭṭa: (c. 875) Old (prācīna) Nyāya author of a non-commentarial trea-
tise reorganizing but ranging over almost all the topics of the Nyāya-sūtra and
disputing rival theories, Nyāya-mañjarī.
160 Sanskrit Glossary
Kaṇāda: (c. 150) legendary author of the Vaiśeṣika-sūtra, Vaiśeṣika’s founding
text.
Kumārila: (c. 650) prominent Mīmāṃsaka philosopher whose metaphysical and
epistemological views Nyāya philosophers often appropriate in part and dispute
in part and sometimes wholesale.
Mādhava: (c. 1350) author of Advaita works as well as of a popular survey of all
the philosophic schools, Sarva-darśana-saṃgraha, an early textbook of classical
Indian philosophies which slightly favors Advaita Vedānta.
Mādhyamika: Buddhist school of skeptical philosophy founded by Nāgārjuna;
sometimes called Buddhist Mysticism or Buddhist Absolutism.
Maṇikaṇṭha Miśra: (c. 1300) author of Nyāya-ratna, a text that precedes several
of Gaṅgeśa’s positions, particularly concerning inference, and that is sometimes
cited by him.
Mathurānātha: (c. 1650) author of an influential commentary on Gaṅgeśa’s TCM.
Mīmāṃsā: a long-running realist school (“Exegesis”) celebrated for its Vedic inter-
pretation, principles of interpretation, philosophy of language, and epistemol-
ogy; the Mīmāṃsā-sūtra (c. 100 bce) is the root text; a commentary by Śabara
(c. 500) is expanded (and sometimes corrected) by Kumārila Bhaṭṭa (c. 650) and
in a second line by Prabhākara (c. 700), the two being the chief philosophic
proponents with followers known as Bhāṭṭas and Prābhākaras.
Mīmāṃsaka: an advocate of Mīmāṃsā.
Mīmāṃsā-sūtra: (c. 100 bce) the founding text of Mīmāṃsā.
Nāgārjuna: (c. 150) prominent Buddhist philosopher understood by Nyāya episte-
mologists to be a skeptic about “knowledge sources,” pramāṇa; founder of the
Mādhyamika school of Buddhist philosophy.
Naiyāyika: an advocate of Nyāya.
Navya Nyāya: “New Nyāya”; the late Nyāya philosophy of, preëminently, Gaṅgeśa
and his followers, pioneered in large part by Udayana, whom, however, Gaṅgeśa
counts as Old (prācīna) Nyāya.
Nyāya: “Logic”; a school of metaphysical realism and “knowledge sources” in epis-
temology prominent throughout the classical period, from the Nyāya-sūtra (c.
200) on; explicitly combined with Vaiśeṣika in later centuries beginning with
Udayana (c. 1000) and sometimes called Nyāya-Vaiśeṣika; Nyāya authors gener-
ally are focused on issues in epistemology but take positions on a wide range of
philosophical topics; see also NYĀYA.
Nyāya-Vaiśeṣika: see NyĀya.
Nyāya-kusumāñjali: (“Flower Offering of Arguments”) a text by Udayana collect-
ing and defending arguments for the existence of God.
Nyāya-ratna: (“Jewel of Nyāya”) a long text by Maṇikaṇṭha that precedes some of
Gaṅgeśa’s TCM particularly concerning inference.
Nyāya-sūtra (NyS): (c. 200) Nyāya’s foundational text attributed to Gautama.
Padārtha-dharma-saṃgraha (PDS): (“Compendium of the Properties of the Fun-
damental Types of Things or Categories”) the sole work of Praśastapāda (c. 575)
and the central text for New Nyāya for ontological categories and supporting
arguments, although there are a few lines of separate development.
Pāṇini: (c. 450 bce) author of a systematic grammar for Sanskrit, sometimes
referred to as Pāṇini’s sūtras, which inaugurated the long and rich grammarian
tradition of classical India including “grammarian philosophy,” a famous later
representative of which is Bhartṛhari.
Prabhākara: (c. 700) prominent Mīmāṃsaka philosopher reputed to be a renegade
pupil of Kumārila, commonly called “Guru” throughout classical philosophy,
whose views are typically taken by Gaṅgeśa to be Nyāya’s principal competitor.
Prācīna Nyāya: “Old Logic”; the philosophy of the Nyāya-sūtra and its commen-
taries and of a few independent treatises; see NyĀya.
 Sanskrit Glossary 161
Praśastapāda: (c. 575) author of the Padārtha-dharma-saṃgraha, the reformula-
tion and explanation of the Vaiśeṣika-sūtra that comes to be the central and
defi ning text of Vaiśeṣika.
Raghunātha: (c. 1500) prominent Gaṅgeśa commentator who championed innova-
tive positions in Nyāya’s realist ontology.
Śabara: (c. 500) author of the oldest extant commentary on the Mīmāṃsā-sūtra.
Śālīkanātha Miśra: (c. 900): author of Prakaraṇa-pañcikā, a good (and available)
presentation of Prābhākara Mīmāṃsaka positions.
Sāṃkhya: “Analysis”; an early school of metaphysical dualism analyzing nature
(prakṛti) in the interests of psychological disidentification.
Śaṅkara: (c. 725) the most prominent philosopher of the Advaita Vedānta school,
espousing a radical monism of the sole reality of Brahman, the Absolute.
Śloka-vārttika: Kumārila’s commentary on a portion of the Mīmāṃsā-sūtra.
Śrīharśa: (c. 1100) Advaita dialectician, whose Khaṇḍana-khaṇḍa-khādya, “Sweet-
meats of Refutation,” directed principally against Nyāya and in particular
Udayana, becomes part of the New Nyāya curriculum.
Tātparya-pariśuddhi: Udayana’s commentary on the Nyāya-sūtra and its earlier
commentaries, in particular Vācaspati’s.
Tattva-cintā-maṇi (TCM): “(Wish-fulfilling) Jewel of Reflection on the Truth
(about Epistemology),” Gaṅgeśa’s sole text, comprised of a chapter on each of the
“knowledge sources,” pramāṇa; focus of much literary activity in later Nyāya.
Udayana: (c. 1000) Nyāya philosopher counted by Gaṅgeśa his “teacher” (ācarya);
prolific author of both commentaries and independent texts and principal uni-
fier of Nyāya epistemology and logic with Vaiśeṣika ontology.
Uddyotakara: (c. 600) author of the core commentary or sub-commentary, Vārttika
(NyV), on Vātsyāyana’s Nyāya-sūtra commentary; adversary of, principally,
Buddhist Yogācāra positions.
Upaniṣad: “secret doctrine”; various prose and verse texts appended to the Veda,
having mystic themes centered on an understanding of self or consciousness in
relation to the Absolute or God, called Brahman; the primary sources for clas-
sical Vedānta philosophy.
Vācaspati Miśra: (c. 950) major Nyāya author whose Tātparya-ṭīkā sub-commen-
tary (NyVTatp) on Uddyotakara’s Vārttika is part of the core Old Nyāya litera-
ture; author also of texts within four other classical schools, Yoga, Sāṃkhya,
Mīmāṃsā, and Advaita Vedānta.
Vaiśeṣika: “Atomism”; a classical philosophy focusing mainly on ontological issues
(“What kinds of things are there?”) and defending a realist view of material
things as composed of atoms as well as a realist ontology of universals or class
characteristics; explicitly combined by Udayana with Nyāya with practically no
separate literature during the New Nyāya period, with its ontological theory
(but not its epistemology) implicitly, and occasionally explicitly, assumed and
endorsed by Nyāya authors before Udayana.
Vātsyāyana: (c. 400) author of the oldest extant commentary, Bhāṣya (NySBh), on
the Nyāya-sūtra.
Veda: “revealed Knowledge”; comprised of four (sometimes three) Vedas which are
collections of hymns to various Indo-European gods and goddesses; the oldest
texts in Sanskrit (some hymns possibly as early as 1500 bce); the most sacred
texts of Hinduism.
Vedānta: originally an epithet for Upaniṣads; in the classical period, the philoso-
phy of the Brahma-sūtra and of several sub-schools defending Upaniṣadic views
often in an idealist or illusionist spirit and opposed to Nyāya’s realism but also
in theistic sub-schools allied with and borrowing from Nyāya.
Yogācāra: Buddhist Idealism; the school of the idealists Asaṅga, Vasubandhu, and
company, as well as of the great Buddhist logicians Dignāga and Dharmakīrti
162 Sanskrit Glossary
who lay out a pragmatist epistemology that is a principal rival of Mīmāṃsā and
Old Nyāya.

2. TERMS

ābhāsa: false semblance, (mere) appearance, the non-genuine; e.g., a hetv-ābhāsa is

a misleading inferential mark, hetu, i.e., a fallacy.
abhāva: absence, negative fact; an absence invariably has a locus and an absentee;
e.g., an absence of a pot on the floor has the floor as its locus and the pot as its
absentee; absences come in four varieties, prior (as of a pot before its produc-
tion), posterior (as of a pot after it is destroyed), absolute (as of a pot on the
floor), and mutual (as between a pot and a cloth).
abhidhā: reference, denotation; the primary power, śakti, of a word in relation to
an object.
abhihita-anvaya-vāda: “theory of the connection of the referents,” the Bhāṭṭa and
Naiyāyika view of sentence meaning as the connection in the object or fact of
the referents of the component words, the words having reference independently
of the context of the sentence.
abhyāsa-daśā-āpanna: familiarity; known through repeated practice.
adhyāsa: superimposition.
adṛṣṭa: “Unseen Force,” i.e., karman or karma.
ajahal-lakṣaṇā: metaphor where a primary, denotative sense of the word meant
figuratively is retained in the figurative, secondary meaning, e.g., “The umbrel-
la-bearers are passing,” where what is meant is that the important people who
walk while being shaded by umbrella-bearers (the brass) are passing and the
umbrella-bearers of course, too.
ākāṅkṣā: syntactic “expectation,” a condition governing the intelligibility of a
statement, according to Naiyāyikas and others; to cite a classical example, no
single word appearing in the accusative case can be understood as a statement
(excepting verbal ellipsis)—somewhat like “to the cow” standing alone in Eng-
lish—because of a violation of “(syntactic) expectation.”
ākāśa: “ether,” the medium of sound, the only non-atomic substance among five
that are material, according to Nyāya-Vaiśeṣika and other classical views.
akhaṇḍa-upādhi: unanalyzable “surplus property” (see UPĀDHI), contrasting with
sakhaṇḍa upādhi.
ākṛti: shape, anatomy; the nature of a universal (jāti) according to Bhāṭṭa
Mīmāṃsā.
alaṃkāra-śāstra: aesthetics, the “science of ornament,” the classical tradition of
literary criticism in particular.
alaukika-sannikarṣa: “extraordinary sensory connection” which, according to the
Nyāya mainstream, comes in three types: (a) connection mediated by a univer-
sal, as in cognition of an unobserved instance of a universal (“The calf to be
born will have a dewlap”); (b) connection mediated by memory, as in cogni-
tion of sandalwood in the distance as having a certain smell, recognition of
something perceived previously, and perceptual illusion—see JÑĀNA-LAK Ṣ A ṆA-
SANNIKAR Ṣ A; and (c) connection mediated by yogic power.
ālaya-vijñāna: “storehouse consciousness,” cosmic mind or mentality where
intersubjective cognitive dispositions are located; a principal concept in early
Yogācāra Buddhism.
anātman: “no self” or “no soul”; an important Buddhist doctrine.
anavasthā: infi nite regress, a defeater and conceptual predicament revealed by
tarka, “suppositional reasoning.”
 Sanskrit Glossary 163
anekānta-vāda: non-absolutism, positive perspectivalism, the “doctrine of many-
sidedness”; the principal metaphysical stance of Jaina philosophers.
antar-vyāpti: “internal pervasion,” pervasion according to something’s partial or
complete identity with itself or something else; a concept of later Buddhist logic
thought to explain inferences based on class inclusions, such as “It is a tree
because it is an oak.”
anubhava: “presentational experience,” awareness, which comes in four veridical
varieties, according to Nyāya, the perceptual, inferential, analogical, and testi-
monial, and many non-veridical varieties (see the chart, p. 29); distinguished as
cognition (jñāna) from remembering as presenting fresh news.
anugama: uniformity, consecutive character; according to Nyāya-Vaiśeṣika, a cri-
terion of a true universal or natural kind, jāti, whose uniformity is taken to
prompt recurrent experience.
anukūla-tarka: reasoning favorable to one’s own position, such as its comparative
simplicity (lāghavatva); “with the grain.”
anumāna: inference, the inferential process; one of four pramāṇa or generators of
knowledge, according to Nyāya.
anumiti: inferential knowledge, the result of inference as a pramāṇa.
anupalabdhi: “non-cognition,” the means of knowledge whereby an absence is
known, according to the Bhāṭṭa branch of Mīmāṃsā.
anuvyavasāya: apperception, “after-cognition,” introspection.
anuyogin: relational correlate to a countercorrelate, pratiyogin.
anvaya: (1) positive correlations (things both H and S) entailed by a natural perva-
sion, vyāpti, constituting positive evidence for the pervasion; see also VYATIREKA;
(2) ontological connection of the meanings of the words in a sentence in a com-
plex object or fact designated, according to Nyāya and Bhāṭṭa Mīmāṃsā; (3) the
total meaning of a sentence considering all the syntactic, semantic, and prag-
matic factors in interrelation.
anvaya-vyatireka: positive correlations (things both H and S) and negative correla-
tions (things not-S and not-H) entailed by a natural pervasion, vyāpti, constitut-
ing both positive and negative evidence for the pervasion.
anvaya-vyatirekin: (1) a property with both a positive and a negative range, i.e.,
both occurring in some places and not occurring in others; (2) an inference
based on both positive and negative correlations; cf. KEVALA-ANVAYIN and
KEVALA-VYATIREKIN.
anvīkṣā: (= ānvīkṣikī) “critical investigation,” a term sometimes used by Naiyāyikas
and others to refer to the enterprises of philosophy in general and Nyāya in
particular.
anvita-abhidhāna-vāda: “theory of reference of the connected,” the Prābhākara
view of reference as belonging to the sentence, not to its component words except
as connected syntactically, with a complex object as that which is referred to,
the words not having reference independently of the context of the sentence.
anyathā-khyāti: the view of perceptual error endorsed by Nyāya that fi nds a
property-bearer presented perceptually as “other than what it is” due to an
epistemic deficiency (doṣa) such as a departure from the normal workings of the
sense organs, etc., and involving a retrieval of previous perceptual information
through excitation of a “memory disposition,” saṃskāra.
anyathā-siddha: “irrelevant,” not regulative; a thing or property that accidentally
precedes an effect in focus or in another way is not relevant to its production as
captured by causal laws (permitting inference); a donkey who happens always
to be used to carry the clay out of which a potter makes all his pots is a stock
example.
anyonyāśraya: “mutual dependence” as a logical, conceptual, or argument flaw
revealed by tarka, “suppositional reasoning.”
164 Sanskrit Glossary
āpta: trustworthy authority or testifier, one who knows the truth and wants to
communicate it without deception, according to Nyāya.
arthāpatti: “postulation,” “presumption,” “circumstantial implication,” an inde-
pendent pramāṇa according to Bhāṭṭa Mīmāṃsakas, a knowledge source viewed
as a form of inference by Naiyāyikas; an argument that something would oth-
erwise be impossible or inexplicable (anyathā-an-upapatti); e.g., from the
information that fat Devadatta does not eat during the day, we know by “pos-
tulation” that he eats at night.
asādhāraṇya: “no similar instance,” an inferential fallacy where there is no com-
parison class, no extrapolative basis for presuming a pervasion or general rule
on which a genuine inference would depend.
asamavāyi-kāraṇa: “emergent” or “co-inherent” cause, e.g., the blue color of the
threads of a blue piece of cloth is an emergent cause of the blue of the cloth.
āsatti: (proper) “proximity” of words in a spoken statement; proper representation;
one of three necessary conditions commonly identified for sentential meaning.
asiddha: the “unshown,” “unestablished,” or “unproved,” an inferential fallacy
that may be likened to an unwarranted premise, normally, the prover’s being
“unestablished” as qualifying the inferential subject.
ātman: individual self, one of nine basic types of substance, according to Nyāya-
Vaiśeṣika; cosmic self, according to Vedānta.
ātmāśraya: “self-dependence”; the defeater, as shown by tarka, of any the-
sis proposing (impossible) self-causation or self-qualification; defi nitional
circularity.
avaccheda: specification, delimitation, “paring down to”; techniques used by late
Nyāya philosophers to remove vagueness and defi nitional overextension.
avacchedaka: specifier, delimiter.
bādha: (1) the fallacy or defeater of “patent falsehood,” where the inferential sub-
ject is known in advance not to exhibit the probandum; (2) experiential “subla-
tion,” as when one sees veridically a real rope previously mistaken for a snake.
bādhaka: counterconsideration, epistemic “defeater,” blocker of a cognitive effect
of a specified type under specified conditions.
bādhaka-abhāva: “absence of defeaters,” a condition on the certifiability of a cog-
nition, according to Nyāya; defeasibility.
bādhita: “defeated”; “defeated in advance” as a type of fallacy.
bhāvanā: a mentally dispositional property such as a memory capacity; a type of
saṃskāra.
bhūyo-darśana: “wide (relevant) experience”; the key condition leading to knowl-
edge of a natural pervasion (vyāpti) of things H by things S, according to the
Nyāya mainstream.
brahman: the “Absolute” or “God,” the central concern of Vedānta philosophy.
cakraka: “circularity,” the epistemic defeater of a thesis that proposes an impos-
sible circular chain of causation or qualification; circular reasoning.
darśana: a “seeing,” a perspective on the truth, a philosophy; the discipline of
philosophy understood as holistic and concerned with the most important
truths and values (sometimes contrasted with “critical inquiry,” anvīkṣā).
dharma: (1) property; (2) right religious and moral practice.
dharmin: property-bearer.
dhvani: sound, tone; “suggestion” as a third power of words, or mode of mean-
ing, śakti, in addition to reference and figurative meaning; for example, the
statement, “The village is in the Gaṅgā,” whose figurative sense is equivalent
to “The village is on the bank of the Gaṅgā,” is said by the third power to sug-
gest (by association with the sacred river) a meditational atmosphere.
doṣa: “fault,” “epistemic defect,” veridicality-undermining causal condition
responsible for the generation of wrong views, according to Nyāya.
 Sanskrit Glossary 165
dravya: substance, one of seven ontic primitives or categories (padārtha) of main-
stream Nyāya-Vaiśeṣika; nine fundamental types of substance are said to be
earth, water, fi re (or the “fiery element,” tejas), air (these are the four atomic
elements), ether (the fi fth “material” element, which is non-atomic and the
medium of sound), time, space, manas (the “mind” or “internal organ”), and
self (ātman).
dṛṣṭānta: a pervasion-supportive “example,” a basis for induction, typically a
locus known to exhibit both the prover and the probandum (i.e., an instance of
sapakṣa, q.v., that exhibits the prover), e.g., a kitchen hearth where both smoke
and fi re are known to have occurred; providing an example suggests that proper
inductive procedure has been followed (cf., anvaya-vyatireka) and that the “per-
vasion” or vyāpti which a proffered inference presumes is properly evidenced.
eka-vṛtti-vedyatā: the doctrine of a qualificandum being known simultaneously
as qualified by a qualifier, such that any verbalizable cognition will have three
types of objecthood or intentionality, viṣayatā: a qualifier portion, a qualifican-
dum portion, and a portion concerning the qualificative relation.
gauravatva: “heaviness,” theoretic cumbersomeness (alleged comparatively); a
form of tarka, “suppositional reasoning.”
gavaya: a rare kind of wild buffalo.
guṇa: (1) “excellence,” “epistemic excellence,” a veridicality-indicating causal con-
dition, according to Nyāya; (2) quality; one of seven fundamental categories
(padārtha) in traditional Nyāya-Vaiśeṣika ontology.
guru: teacher; Prabhākara, the Mīmāṃsaka philosopher.
hetu: inferential mark, “prover,” sādhana (q.v.).
hetvābhāsa: false or misleading prover or inferential mark, “fallacy of the prover”
or simply “fallacy.”
indriya-sannikarṣa: (= indriya-pratyāsatti) connection between an operative
sense organ and an object perceived; see also ALAUKIKA- SANNIKAR Ṣ A and JÑĀNA-
LAK Ṣ A Ṇ A- SANNIKAR Ṣ A.
īśvara: God, “Lord”; an individual self or ātman responsible for the creation or
material arrangement of the world but not for atoms, which are eternal and
uncreated, nor karma, which is an independent force (adṛṣṭa), nor other selves,
such as human selves, which are eternal but also subject to spiritual ignorance
and other weaknesses whereas God is not—according to Vātsyāyana and most
but not all later Nyāya philosophers.
jalpa: debating for victory where (in contrast with vāda) using tricky arguments is
all right so long as one is not caught.
jāti: universal, natural kind, a property occurring in more than a single instance
or locus.
jāti-bādhaka: a “blocker,” defeater, or counterconsideration against taking a prop-
erty to be a true universal, jāti, generating an “infi nite regress,” for example,
which blocks cowhood-hood.
jñāna: “cognition” (see the chart, p. 29); both presentational awareness, anubhava,
and remembering, smaraṇa; an episodic psychological quality or occurrent
knowledge, according to Nyāya-Vaiśeṣika, that has an object or objecthood,
viṣayatā, and that guides voluntary action.
jñāna-lakṣaṇa-sannikarṣa: (jñāna-lakṣaṇa-pratyāsatti) cognitively mediated con-
nection with a cognized object that enables perception of something not in direct
connection with the operative sense organ; for example, the recognition, “This
is that Devadatta,” is perceptual with the “thatness” provided—“retrieved”
(upanīta)—by memory; see also INDRIYA-SANNIKAR Ṣ A and UPANĪTA.
jyeṣṭha-pramāṇa: perception as the fundamental (“eldest”) pramāṇa; according to
Nyāya, non-perceptual knowledge sources are each said to depend in some way
on perception.
166 Sanskrit Glossary
karaṇa: “trigger,” “proximate instrumental cause,” a necessary condition that on
being met an effect regularly comes about.
kāraṇa: “cause,” a necessary and regulative condition for an effect, by default of
the subtype called “instrumental cause,” nimitta-kāraṇa but including causes
called “inherent,” samavāyin, and “emergent,” asamavāyin; see also SĀMAGRĪ,
“sufficient causality.”
karman: (1) action, motion; (2) psychological dispositions to act in a certain man-
ner accrued through previous actions; habits, karma; conceived metaphysically
as an environmental influence (adṛṣṭa, “Unseen Force”) affecting future hap-
piness or suffering (as in having good luck or bad), in particular the nature of
one’s body in one’s next reincarnation.
kevala-anvayin: (1) “universally positive,” a property with no negative range,
nowhere, that is, where it is not; (2) “only positive,” an inference whose induc-
tive support consists only of instances of co-presence between the prover (H)
and probandum (S), with no known instances of co-absence, i.e., things S and H
but not things not-S and not-H.
kevala-vyatirekin: “only negative,” an inference whose inductive support consists
only of instances of co-absence between the prover (H) and probandum (S), with
no known instances of co-presence, i.e., things not-S and not-H but not things
S and H.
kṣaṇa: “point-instant,” the smallest increment of time, too small to be perceptible,
according to Nyāya; a temporal atom, so to say, according to Yogācāra.
lāghavatva: “lightness,” parsimony, theoretic simplicity; a form of tarka, “suppo-
sitional reasoning.”
lakṣaṇa: characterization, defi nition.
lakṣaṇā: figurative speech; a second power or śakti of words in addition to denotation,
abhidhā.
lakṣya: that which is indicated in indirect, figurative speech, the beloved’s face, for
instance, indicated by reference to the moon, the śakya, direct referent.
manas: the mind or internal organ, the sense-mind, the sense organ operative in
the perception of psychological properties such as pleasure and pain and the
conduit of sensory information to the perceiving self, according to Nyāya and
other classical schools.
māyā: illusion; cosmic illusion; the nature of the entire phenomenal display accord-
ing to Advaita Vedānta.
mīmāṃsā: (1) science of interpretation; (2) a philosophic school, see M Īm Ā ṂsĀ.
mithyā-jñāna: false cognition; delusion.
mukti: (mokṣa) “liberation” from rebirth, salvation, the summum bonum accord-
ing to several schools.
nigraha-sthāna: “grounds for rebuke” in a debate; informal fallacy (several are
identified in the Nyāya-sūtra).
nimitta-kāraṇa: instrumental cause, necessary condition, such as a potter, potter’s
wheel, and so on, for the production of a pot.
nirākāra: “without form,” said of cognition by Naiyāyikas who see cognitions as
given shape or form by their objects—their intentionality (viṣayatā) object-de-
termined—a view contrasting with Buddhist sākāra-vāda.
nirṇaya: reflective knowledge; fi nal ascertainment as a result of philosophical
investigation or nyāya.
nirvikalpaka: “indeterminate”; indeterminate perception grasps a qualifier “in the
raw,” not as qualifying a qualificandum; indeterminate perception is not apper-
ceptible and not directly verbalizable, according to later Nyāya.
niścaya: certainty, warrant, sufficient epistemic confidence to act unhesitatingly.
niṣkampa-pravṛtti: voluntary action taken unhesitatingly.
nyāya: (1) the philosophic school, see NyĀya; (2) philosophical examination and
argument; (3) maxim or principle.
 Sanskrit Glossary 167
padārtha: category, primitive, “type of thing to which words refer”; there are seven
fundamental categories according to mainstream Nyāya-Vaiśeṣika, (1) sub-
stance, (2) quality, (3) motion, (4) universal, (5) fi nal particularizer, (6) inher-
ence, and (7) absence.
pakṣa: “inferential subject,” a locus where a prover is known or presumed to reside
(e.g., the mountain in the stock example of inference from smoke to fi re).
pakṣa-dharmatā: the rule and certification condition that the prover has to be
known as qualifying the inferential subject.
parāmarśa: “reflection” or “consideration” necessary to the generation of infer-
ential awareness; a cognitive putting together, so to say, of the evidence of the
prover as qualifying the subject with that of the prover’s being pervaded by the
probandum.
param-parā: chain relation, as with the indirect, figurative meaning that works
from designation of a fi rst item, the moon, śakya, to a second item ultimately
indicated and meant, the beloved’s face, lakṣya.
para-prāmāṇya: cognition as “certified through another,” “extrinsic justification,”
usually by means of inference; Nyāya’s doctrine of certification as requiring infer-
ence or apperceptive source identification to establish the truth of a target cogni-
tion; an epistemological thesis opposed to self-certification (sva-prāmāṇya).
parārthānumāna: “inference for others”; a formal proof embedding a valid infer-
ence, to be best expressed, according to Nyāya (with a few dissenters), in a five-
member argument understood as a single sentence.
paribhāṣika: technical (as in “technical terms”), defi nitional, stipulational.
prakāra: as a Nyāya technical term, “predication content,” more generally, “way,”
a way something appears; e.g., a cognition Fa has F-hood as its “predication
content,” as the way something, a, is appearing, provoking, in appropriate cir-
cumstances, the cognizer to say, “It’s F,” or “It’s an F,” or the like; a qualifier
(viśeṣaṇa) as cognized.
prakaraṇa: (1) textual and extra-textual “context” of a statement; (2) a section of
a philosophic or scientific treatise.
prakṛti: “nature” in Sāṃkhya philosophy, the uniform ground of all phenomena,
of everything material as presented in experience, all effects being transforma-
tions of a common material cause in which they in a sense pre-exist.
pramā: knowledge, veridical cognition, veridical “presentational experience,”
anubhava; unreflective knowledge.
pramāṇa: “knowledge source,” means to pramā; according to Nyāya, there are
four, perception, inference, analogy, and testimony.
pramāṇa-ja: knowledge-source-produced; certified as knowledge.
pramāṇa-samplava: coalescence of one or more knowledge sources in making
known an object or fact.
prāmāṇya: (1) knowledge, the condition of being generated by a knowledge source,
i.e., a true cognition’s being generated by a genuine knowledge source, pramāṇa;
(2) justification, truth-grounded evidence.
prāmāṇya-ābhāsa: pseudo-justification, the appearance of having knowledge for
good reasons without actually knowing the truth.
prameya: object of knowledge; the knowable.
prāpya-kāri: a doctrine concerning a sense organ such as touch which is said to
“work by reaching” its object through a direct contact at the object’s own locus;
contrasting with a-prāpya-kāri, “not reaching its object,” said of a sense organ
such as hearing in relation to an object that is not grasped through contact at
the object’s own locus.
prasaṅga: dialectical difficulty, predicament drawn out by tarka, “suppositional
reasoning,” such as contradiction or infi nite regress; the equivalent of tarka in
Buddhist usages.
pratibandhaka: preventer, blocker.
168 Sanskrit Glossary
pratijñā: “proposition”; the fi rst member of a formal “inference for others,”
parārthānumāna, the “proposition to be proved.”
pratikūla-tarka: reasoning unfavorable to a position, such as prompting an infi-
nite regress or having a consequence known to be false; reasoning “against the
grain.”
pratiyogin: (1) in Navya Nyāya, the countercorrelate or relatum of a dyadic rela-
tion, i.e., the second term of a relation, the anuyogin being the fi rst (e.g., the
pratiyogin of “objecthood,” viṣayatā, is cognition and the object of cognition
is the anuyogin; the pratiyogin of “subjecthood,” viṣayitā, is the object of cog-
nition and the cognition is the anuyogin); (2) counterpositive of an absence,
“absentee” (with an absence of a pot on the floor, the pot is the pratiyogin, the
floor the anuyogin).
pratyāsatti: see INDRIYA- SANNIKAR Ṣ A.
pratyakṣa: (1) perception or perceptual knowledge, the veridical result of perception
as a knowledge source; (2) the source itself, the “knowledge source,” pramāṇa,
that generates bits of perceptual knowledge.
pravṛtti: voluntary action.
pravṛtti-nimitta: grounds for a particular activity of speech, for use of a particular
word.
pūrva-pakṣa: a prima facie position, the opponent’s position; a portion of a text
devoted to exploring views and arguments not accepted by the author who will
express his own views in an upcoming and correlate siddhānta (q.v).
pūrva-pakṣin: an opponent or fi rst-level interlocutor in relation to a siddhāntin
(q.v.).
ṛṣi: ancient sage; instrument of Vedic revelation, according to some Nyāya
philosophers.
rūḍhi: convention; the principal way the meaning or denotation relation between
word and object is set up according to Nyāya and some other schools (not
Mīmāṃsā).
śabda: (1) “testimony” as a knowledge source according to Nyāya and some other
classical schools; (2) “sound,” the quality grasped by the organ of hearing; (3)
revelation, śruti (q.v.).
śābda-bodha: knowledge from trustfully understood words, veridical cognition
generated by testimony.
sādhana: “prover,” hetu, in relation to a probandum term within a knowledge-pro-
ducing inference (e.g., smokiness, in an inference to fieriness at a certain spot).
sādhya: “probandum,” the property to be proved by a prover in an inference (e.g.,
fieriness in an inference from smokiness).
sādṛśya: similarity.
sahakārin: auxiliary cause, a necessary condition that is not in itself sufficient for
an effect nor its trigger.
sākāra-vāda: a Buddhist doctrine about cognition “taking form internally,” not
depending on external things, as opposed to Nyāya’s view that cognition is in
itself “without shape or form,” nirākāra (q.v.).
sakhaṇḍa upādhi: a surplus property that is ontologically composite and analyz-
able, e.g., “blue-cow-hood,” which is a composite of the quality, blue, and the
universal, cowhood.
śakti: “power” of words to connect with objects; the meaning relation; there are two
such powers according to Nyāya, “reference,” abhidhā, and indirect, “figurative
meaning,” lakṣaṇā, not three, as according to some, adding “suggestion,” dhvani.
śakya: the intermediate item or “referent” in indirect, figurative speech, the moon,
for instance, indicating the lakṣya, the beloved’s face, the ultimately indicated
or meant.
samādhi: yogic trance, the ability to shut off mental fluctuations, concentration.
 Sanskrit Glossary 169
sāmagrī: collection of causal factors sufficient to produce an effect.
sāmānya: universal or jāti; “common characteristic.”
sāmānyato dṛṣṭa: a type of inference, discussed in the Nyāya-sūtra commentar-
ies on NyS 1.1.5 and elsewhere, an inference proceeding “on the basis of a
general principle holding for all things experienced (as of such a kind)” to an
inferred analogue in a case at issue; e.g., “Pleasure and the like have a substra-
tum, since they are qualities, like color” (the key step, according to Vātsyāyana,
to inferential knowledge of the self, which is to be identified with the substratum
proved).
samavāya: “inherence”; ontic glue, according to Nyāya and other realist schools,
relating certain types of things to their loci, such as true universals, jāti; e.g.,
cowness inheres in cows.
samavāyi-kāraṇa: “inherent cause,” a substratum in relation to a superstratum, e.g.,
a property-bearer in relation to its properties and parts in relation to a whole.
sannikarṣa: see INDRIYA- SANNIKAR Ṣ A.
saṃśaya: doubt, uncertainty.
saṃskāra: memory-impression, mental disposition; impetus.
saṃyoga: contact, conjunction.
sapakṣa: things or loci where the probandum is known to occur, e.g., a kitchen
hearth, etc., with respect to an inference from smoke to fi re.
śāstra: science or craft; a scientific textbook.
sat-pratipakṣa: the “counterinference” fallacy and defeater of a putative inference, where
a first thesis to be proved (Sa) is countered by an opposite thesis (~Sa), each with a
prover counterbalanced by another prover equally well-supported by correlations.
savikalpaka: “determinate”; determinate cognition, which is a “propositional”
cognition, verbalizable cognition, cognition of an entity as qualified by a quali-
fier, according to Gaṅgeśa and New Nyāya, paradigmatically a cognition of a
qualificandum (viśeṣya) qualified by a qualifier (viśeṣaṇa) or property, thus hav-
ing objecthood or intentionality (viṣayatā) with the structure, a-as-F; opposed
to nirvikalpaka.
siddhānta: (1) the “right view” answering or displacing a correlate “prima facie
view” or pūrva-pakṣa; an author’s “own position” and that of his school; (2) a
portion of a text devoted to elaborating an author’s own views and arguments in
relation to a previous pūrva-pakṣa of contending views and arguments.
siddhāntin: the “proponent of the right view” answering a correlative pūrva-pakṣin
(q.v.); the siddhāntin is normally the author himself speaking for himself and his
school.
siddha-sādhana: the fallacy of trying to prove what is already known or generally
admitted.
smaraṇa: “remembering,” an occurrent cognition, jñāna, distinguished from
anubhava, “presentational experience,” by depending in its correctness on
the presentational experience responsible for formation in a self (or mind) of a
saṃskāra memory-disposition making it and other rememberings possible.
smṛti: (1) memory; (2) second-order of sacred texts in epistemological contrast with
a fi rst order, śruti.
sphoṭa-vāda: theory of sentential holism proposed by Bhartṛhari; the sentence,
which is the basic semantic unit (not the word), is understood “in a flash” or
“burst,” sphoṭa.
śruti: the Veda, the highest order of “revelation,” according to popular Hinduism.
sthiti-sthāpaka: the dispositional property of elasticity, e.g., the disposition of a
stretched rubber band to assume a shorter length.
sūtra: “thread”; an aphorism or summary statement meant to express succinctly a
position or argument which is to be explained and fleshed out by commentary
whether in written form or orally by a teacher or expert.
170 Sanskrit Glossary
sūtra-kāra: sūtra-maker; e.g., Gautama as the author of the sūtras of the Nyāya-
sūtra.
sva-bhāva: (1) self-nature, independent existence; (2) according to Buddhist logic, a
type of prover that works by a relation of identity; see also ANTAR-VYĀPTI.
sva-lakṣaṇa: particular, individual, the unique, “that which is its own mark”; an
important Buddhist ontological concept.
sva-prāmāṇya: cognition as “self-authenticating,” a Mīmāṃsaka and Vedāntic
view of cognitive confi rmation contrasting with the Naiyāyika view that a cog-
nition is certified “through another,” para as opposed to sva; see also PARA-
PRĀMĀ Ṇ YA.
svārthānumāna: “inference for oneself,” natural inference, contrasting with formal
“inference for others,” parārthānumāna (q.v.).
sva-rūpa-sambandha: self-linking relation, e.g., a rope (tying itself, so to say, to a
tree as well as a goat).
svayam-prakāśamāna: “irreflexively self-illuminating”; an Upanishadic doctrine of
the nature of self-consciousness championed by, among others, the Prābhākara
Mīmāṃsaka.
tarka: “suppositional reasoning”; according to Nyāya and other classical schools,
usually a hypothetical argument revealing a fallacy or defeater with respect to a
position ~p opposed to another position p that is to be accepted so long as p has
independent evidence in its favor.
tarka-ābhāsa: apparent or fallacious instance of suppositional reasoning, tarka.
tātparya: “intention,” what a speaker or author means to say.
trairūpya-hetu: doctrine of the “threefold inferential mark” due to the Buddhist
philosopher, Dignāga: see p. 57 in this book.
udāharaṇa: “application,” the fourth member of a formal “inference for others,”
parārthānumāna, the assertion that the inferential subject falls under the gen-
eral rule expressing a “pervasion,” vyāpti.
upādhi: in Nyāya, (1) the inferential upādhi, the inferential “undercutter,” a spe-
cial condition cognition of which undercuts an inference, blocking or prevent-
ing the occurrence of genuine inference through bringing about reasonable
suspicion of deviation from a rule or pervasion (vyāpti); standardly defi ned
as something that pervades the probandum of a targeted inference while fail-
ing to pervade its prover (all things S, the probandum, are U, the undercutter,
but some things H, the prover, are not U); (2) the ontological upādhi, a “sur-
plus property,” an “imposed property” contrasting with a genuine universal,
sāmānya or jāti.
upalakṣaṇa: (1) indirect indication; (2) indicator; a feature presented, or referred
to, to indicate something else; a differentiator that determines reference having
little or nothing to do with what the referendum is, i.e., not one of its genuine
qualifiers, e.g., hovering crows mentioned to pick out Devadatta’s house.
upamāna: “analogy”; analogical acquisition of vocabulary, a pramāṇa according
to Nyāya and some other classical schools.
upamiti: “analogical knowledge,” e.g., a subject S knowing from a previous ana-
logical statement that the word ‘gavaya’ is used to designate the sort of animal
that S is currently perceiving, and thus knowing (and possibly saying in the
speech act), “That’s what the word means.”
upanīta: “retrieved,” “revived”; memory-impressions, saṃskāra, can be provoked
to project previous perceptual content or objecthood into a current presenta-
tional experience, as happens in perceptual illusion and in several further types
of sensory cognition according to Nyāya.
vāda: “discourse,” enquiry undertaken to discover the truth on a topic or issue of
concern.
vipakṣa: loci where a probandum is known to be absent, e.g., a lake with respect to
an inference from smokiness to fieriness at a spot.
 Sanskrit Glossary 171
virodha: “incompatibility” or “contradiction”; a fallacy and defeater of a putative
inference where the prover proves not the probandum but its opposite.
viṣaya: cognized object.
viṣayatā: “objecthood,” intentionality, one side of the relation between a cognition
and its object, with the other as viṣayitā, according to New Nyāya.
viṣayitā: “subjecthood,” intentionality, one side of the relation between a cognition
and its object, with the other as viṣayatā, according to New Nyāya.
viśeṣa: (1) distinction, difference, particular; (2) individualizer, particularizer, or
numeralizer, one of seven ontic primitives or categories according to mainstream
Nyāya-Vaiśeṣika.
viśeṣaṇa: qualifier; qualities (e.g., blue), universals (e.g., cowhood), motions, along
with unclassified properties (upādhi) all count as qualifiers that qualify a quali-
ficandum, viśeṣya, according to later Nyāya.
viśeṣaṇatā: “being-a-qualifier,” “qualifierhood”; the relation (sannikarṣa) between
an absence and a sense organ that is operative in perception of an absence; for
example, it is in virtue of being a qualifier of a floor that a pot-absence there can
be known perceptually.
viśeṣya: the qualificandum as determined by a cognition of an entity as qualified;
the thing cognized, the bearer of qualifiers (= dharmin).
viśiṣṭa: an entity as qualified; a qualificandum together with one or more of its
qualifiers as known by a qualificative cognition (viśiṣṭa-jñāna, “cognition of an
entity as qualified”).
viśvāsa: trust; belief.
vitaṇḍā: refutational debate by captious argument with no express concern to
establish a thesis of one’s own.
vyabhicāra: “deviation” of the occurrence or extension of a putative prover prop-
erty from the occurrence or extension of its probandum such that there is no
pervasion (vyāpti) of the prover by the probandum.
vyākaraṇa-śāstra: science of grammar.
vyāghātatā: incompatibility, opposition; contradiction as a defeater of a thesis as
shown by tarka.
vyañjaka: indicator, sign, manifester, e.g., a dewlap for a cow.
vyāpāra: “operation,” “employment in causal operation,” which is said to be
required of “triggers” (karaṇa) in relation to effects; e.g., the activity of an axe
being used to fell a tree is the “causal employment” of the axe with respect to
the tree’s being cut down.
vyāpti: “pervasion,” a relation that grounds inference, a factual relation, according
to Nyāya and some other classical schools, such that everything exhibiting a
prover (H) also exhibits a probandum (S), e.g., wherever smokiness, there fieri-
ness: (x) (Hx → Sx).
vyāpya-vṛtti: “locus-pervading,” as the universal cowhood of every part of Bessie
the cow.
vyatireka: negative correlations (things not-S and not-H) entailed by a natural per-
vasion, vyāpti, constituting negative evidence for a pervasion; see also ANVAYA-
VYATIREKA.
vyavahāra: common experience and linguistic practice, what people commonly do
and say; taken by Naiyāyikas and other classical philosophers as presumptively
authoritative; convention, agreement.
yathārtha: matching or corresponding to an object, cognition “as something is (or
was)”; correctness or truth of a remembering.
yaugika-pratyakṣa: yogic perception, special perceptual capacities presumed to be
brought about by yoga practices.
yoga: (1) practices of self-discipline; (2) etymological synthesis (“union”) or deriva-
tion as establishing meaning, e.g., ‘swimmer’ as composed of the verb ‘swim’
and the agential suffi x ‘-er’ to mean a person who swims.
172 Sanskrit Glossary
yoga-ruḍhi: meaning established both by convention and etymology, e.g., ‘catcher’
as composed of the verb ‘catch’ and the agential suffi x ‘-er’ to mean not only
a person who catches but, by convention, a particular position on a baseball
team.
yogyatā: semantic fitness, a condition or “excellence” (guṇa) governing the intel-
ligibility of a statement and the acquisition of knowledge through testimony,
according to Naiyāyikas and others; e.g., “He is watering the plants with fi re,”
is unintelligible in a non-metaphoric sense, because the semantics of “watering”
precludes fi re as an instrument in the action.
Texts and Translations

References are normally to translations, and to the fi rst listed when there
are more than one. The exception is reference to the Nyāya-sūtra (NyS) and
core commentaries (NySBh and NyV) which is normally to the Sanskrit of
the Tarkatirtha et al. edition.

Ānandavardhana. Dhvanyāloka. Tr. Daniel H. H. Ingalls, J. Masson, and M.

Patwardhan. 1990. The Dhvanyāloka of Ānandavardhana with the Locana of
Abhinavagupta. Harvard Oriental Series 49. Cambridge: Harvard University
Press.
Annambhaṭṭa. Tarka-saṃgraha. With the auto-commentary, Tarka-saṃgraha-
dīpikā. Ed. and tr. Gopinath Bhattacharya. 1976. Tarkasaṃgrahadīpikā on
Tarkasaṃgraha by Annaṃbhaṭṭa. Calcutta: Progressive Publishers.
Appaya Dīkṣita. Vṛttivārttika. Ed and tr. Edwin Gerow. 2001. The Vṛttivārttika
or Commentary on the Functions of Words of Appaya Dīkṣita. American Ori-
ental Society Series 87. New Haven.
Asaṅga. Mahāyāna-saṃgraha. Ed. and tr. Étienne Lamotte. 1973. La Somme du
Grand Véhicule d’Asaṅga. 2 vols. Louvain-la-Neuve: Université de Louvain,
Institut Orientaliste.
Bhartṛhari. Vākyapadīya, Chapter 1. Tr. Joseph Ouseparampil. 2005. Bhartṛhari’s
Vākyapadīya Kaṇḍa I. Pune: Indian Institute of Indology.
Bhāsarvajña. Nyāya-bhūṣaṇa (auto-commentary on Nyāya-sāra). Ed. Swami
Yogindrananda. 1968. Varanasi: Satdarshana Prakashan Pratisthanam.
. Nyāya-sāra. With the editor’s Sanskrit commentary. Ed. Vasudevi Suri.
2005. Delhi: Sanjay Prakashan.
Caraka. Caraka-saṃhitā. Ed. and tr. Priyavat Sharma. 1981. 2 vols. Varanasi:
Choukhambha Orientalia.
Dharmakīrti. Hetu-bindu. Tr. Pradeep P. Gokhale. 1997. Hetubindu of Dharmakīrti
(A Point on Probans). Delhi: Sri Satguru Publications.
. Nyāya-bindu. Ed. F. I. Shcherbatskoi. With the Nyāya-bindu-ṭīkā com-
mentary of Dharmottara. 1918. Bibliotecha Buddhica 7. Reprint. Delhi: Motilal
Banarsidass. (1) Tr. Alex Wayman. In Alex Wayman, ed. 1999. A Millennium of
Buddhist Logic, vol. 1. Delhi: Motilal Banarsidass. (2) Tr. (including Dharmot-
tara’s commentary) Th. Stcherbatsky (F. I. Shcherbatskoi). Buddhist Logic, vol.
2. Reprint. New York: Dover.
. Pramāṇa-vārttika. Ed. Dvarikadas Sastri. 1968. With the commentary of
Manorathanandin. Varanasi: Bauddha Bharati.
. Pramāṇa-viniścaya, pratyakṣa chapter. Tr. (from the Tibetan) Tilmann
Vetter. 1966. Sitzungsberichte der Österreichischen Akademie der Wissen-
schaften 250, Band 3. Vienna: Böhlaus.
174 Texts and Translations
. Vāda-nyāya. Tr. Pradeep P. Gokhale. 1993. Vādanyāya of Dharmakīrti:
The Logic of Debate. Delhi: Sri Satguru Publications.
Dharmarāja Adhvarin. Vedānta-paribhāṣā. Tr. S. S. Suryanarayana Sastri. 1971.
Madras: Adyar Library and Research Centre.
Dharmottara. Nyāya-bindu-ṭīkā. See under Dharmakīrti.
Dignāga. Pramāṇa-samuccaya. Tr. Masaaki Hattori. 1968. Dignāga, On Percep-
tion: An Annotated Translation of the First Chapter of the Pramāṇasamuccaya.
Cambridge: Harvard University Press.
Gadādhara. Śakti-vāda-vicāra. Ed. and tr. Gerdi Gerscheimer. 1996. La Théorie
de la signification chez Gadādhara. 2 vols. Paris: Collège de France: Edition-
Diffusion de Boccard.
. Viṣayatā-vāda. Tr. Sibajiban Bhattacharyya. 1990. Gadādhara’s Theory of
Objectivity, vol. 2. New Delhi: Indian Council of Philosophical Research.
. Vyutpatti-vāda. Tr. V. P. Bhatta (fi rst and second kāraka). 1990. Delhi: East-
ern Book Linkers. Tr. Achyutananda Dash (the abhedānvayabodhaprakaraṇam).
2003. Delhi: Pratibha Prakashan.
Gaṅgeśa. TCM. Tattva-cintā-maṇi, perception chapter. (1) Ed. N. S. Ramanuja
Tatacharya. 1972. With the Prakāśa commentary by Rucidatta Miśra and a
sub-commentary by Rāmakṛṣṇādhvarin. Kendriya Sanskrit Vidyapeetha Series
20. Tirupati. (2) Ed. Kamakhyanath Tarkavagish. 1884–1901 (reprint 1991).
With the Māthurī commentary of Mathurānātha. Calcutta: Asiatic Society.
Translation: S. H. Phillips and N. S. Ramanuja Tatacharya. 2004. Episte-
mology of Perception. New York: American Institute of Buddhist Studies.
TCM, inference chapter. (1) Ed. N. S. Ramanuja Tatacharya. 1982 and 1999.
With the Prakāśa commentary by Rucidatta Miśra and a sub-commentary by
Dharmarājādhvarin. 2 vols. Kendriya Sanskrit Vidyapeetha Series 33. Tiru-
pati. (2) Ed. Kamakhyanath Tarkavagish. 1884–1901 (reprint 1991). With
the commentary of Mathurānātha. 2 vols. Calcutta: Asiatic Society. Trans-
lation (piecemeal, by several scholars): Mrinalkanti Gangopadhyay. 1975.
“Gaṅgeśa on Vyāptigraha: The Means for the Ascertainment of Invariable
Concomitance.” Journal of Indian Philosophy 3: 167–208. C. Goekoop. 1967.
The Logic of Invariable Concomitance in the Tattvacintāmaṇi. Dordrecht:
D. Reidel. S. H. Phillips and N. S. Ramanuja Tatacharya. 2002. Gaṅgeśa on
the Upādhi, the “Inferential Undercutting Condition.” New Delhi: Indian
Council of Philosophical Research. S. H. Phillips. 2009. “From the Tattva-
cintā-maṇi by Gaṅgeśa, the kevala-vyatireki-prakaraṇam, ‘Negative-Only
Inference,’ Annotated Translation and Commentary.” In Piotr Balcerowitz,
ed., Logic and Belief in Indian Philosophy, Warsaw Indological Studies, Vol-
ume 3: 435–505. Delhi: Motilal Banarsidass. John Vattanky. 1984. Gangeśa’s
Philosophy of God. Madras: Adyar. TCM, analogy chapter. (1) Ed. Gauri-
nath Sastri. 1983. With the Pragalbhī commentary by Pragalbhācārya. Vara-
nasi: Sampurnanand Sanskrit University. (2) Ed. Kamakhanath Tarkavagish.
With the Dīpanī commentary by Kṛṣṇakānta (reprint 1990). Calcutta: Asiatic
Society. TCM, testimony chapter. (1) Ed. and tr. V.P. Bhatta. 2005. 2 vols.
Delhi: Eastern Book Linkers. (2) Ed. Kamakhyanath Tarkavagish. 1884–1901
(reprint 1990). With the commentaries of Mathurānātha and Jayadeva. 2 vols.
Calcutta: Asiatic Society.
Gautama. NyS. Nyāya-sūtra. (1) Ed. A. M. Tarkatirtha, Taranatha Nyayatar-
katirtha, and H. K. Tarkatirtha. 1936–1944 (reprint 1985). Nyāyadarśanam
(Nyāya-sūtra with four commentaries, the Nyāya-sūtra-bhāṣya of Vātsyāyana
(NySBh), the Nyāya-vārttika of Uddyotakara (NyV), the Nyāya-vārttika-
tātpārya-ṭīkā of Vācaspati Miśra, and the Vṛtti of Viśvanātha). Calcutta San-
skrit Series 18. (2) Ed. Anantalal Thakur. 1997. Gautamīyanyāyadarśana with
Bhāṣya of Vātsyāyana. New Delhi: Indian Council of Philosophical Research.
 Texts and Translations 175
Translation: (1) Ganganatha Jha. 1919. With Vātsyāyana’s and Uddyotakara’s
commentaries. The Nyāyasūtras of Gautama with the Bhāṣya of Vātsyāyana
and the Vārttika of Uddyotakara. 4 vols. Reprint. Delhi: Motilal Banarsi-
dass. (2) Mrinalkanti Gangopadhyay. 1982. With Vātsyāyana’s commentary.
Nyāya-Sūtra with Vātsyāyana’s Commentary. Calcutta: Indian Studies. (3)
Michel Angot. 2009. Le Nyāya-sūtra de Gautama Akṣapāda, le Nyāya-Bhāṣya
d’Akṣapāda Pakṣilasvāmin: l’Art de Conduire la Pensée en Inde Ancienne; Édi-
tion, Traduction et Présentation. Paris: Les Belles Lettres.
Hemacandra. Pramāṇa-mīmāṃsā. Ed. Nagin J. Shah and tr. Satkari Mookerjee
with Nathmal Tatia. 2002. Hemacandra’s Pramāṇamīmāṃsā (A Critique of
Organ of Knowledge): A Work on Jaina Logic. Ahmedabad: International Cen-
tre for Jaina Studies, Gujarat Vidyapith.
Jagadīśa. Tarkāmṛta. Ed. and tr. Baliram Shukla. 1997. Pune: Nimitta Prakashan.
Jaimini. Mīmāṃsā-sūtra. Tr. Mohan Lal Sandal. 2 vols. Reprint (1980). Delhi:
Motilal Banarsidass.
Jayanta Bhaṭṭa. Nyāya-mañjarī. Ed. Surya Narayana Sukla. 1971. Kashi Sanskrit
Series 106. Varanasi: Chowkhamba Sanskrit Series Office. Translation: J.
V. Bhattacharyya. 1978. Jayanta Bhaṭṭa’s Nyāya-mañjarī. Delhi: Motilal
Banarsidass.
Jayarāśi. Tattvopaplava-simha. Tr. Eli Franco. 1987. Perception, Knowledge and
Disbelief: A Study of Jayarāśi’s Scepticism. Stuttgart: Franz Steiner.
Jhalakikar, Bhimacarya, compiler. 1978. Nyāya-kośa (Dictionary of Technical
Terms of Indian Philosophy). Ed. and rev. Vasudev Shastri Abhyankar. Poona:
Bhandarkar Oriental Research Institute.
Kaṇāda. Vaiśeṣika-sūtra. Ed. (with the commentary of Candrānanda) Muni Sri
Jambuvijayaji. 1961. Gaekwad Oriental Series 136. Baroda: Oriental Institute.
Tr. Nandalal Sinha. 1911. The Sacred Books of the Hindus, vol. VI.
Keśava Miśra. Tarka-bhāṣā. Ed. Devadatta Ramkrishna Bhandarkar and Pandit
Kedarnath. 1937. Tarkabhāṣā of Keśava Miśra. Bombay Sanskrit and Prakrit
Series 84.
Kevalananda Sarasvati, ed. Mīmāṃsākośa. 7 vols. 1874–1888, (reprint) 1952–
1966. Wai: Prajna Pathashala Mandal.
Kṛṣṇakānta. See under Gaṅgeśa, TCM analogy chapter.
Kumārila. SV. Śloka-vārttika (commentary on part of the Mīmāṃsā-sūtra). Ed.
Dvarikadas Sastri. 1978. Varanasi: Tara Publications. Translation, the percep-
tion chapter: John Taber. 2005. A Hindu Critique of Buddhist Epistemology:
The “Determination of Perception” Chapter of Kumārila Bhaṭṭa’s Śloka-
vārttikā. London: Routledge. Other chapters: Ganganatha Jha. Slokavartika.
1900, 1908. Reprint. Delhi: Sri Satguru Publications.
. Tantra-vārttika (commentary on part of the Mīmāṃsā-sūtra). Tr. Gan-
ganatha Jha. 1924. 2 pts. Bibliotheca Indica 161. Asiatic Society of Bengal.
Calcutta.
Laugākṣi Bhāskara. Tarka-kaumudi. Ed. and tr. M. N. Dvivedi. 1886. Bombay
Sanskrit Series 32.
Mādhava. Sarva-darśana-saṃgraha. Ed. Iswarachandr Vidyasaga. 1858. Bibliotheca
Indica 63 and 142. Asiatic Society of Bengal. Calcutta. Reprint. Osnabrück:
Biblio Verlag, 1981. Tr. E. B. Cowell and A. E. Gough. Compendium of All Phi-
losophies: Review of the Different Systems of Hindu Philosophy by Madhava
Acharya. London: Kegan Paul, 1892. Reprint. Delhi: Cosmo Publications.
Mahābhārata. Ed. Ramchandrashastri Kinjawadekar (with a commentary by
Nīlakaṇṭha). 6 vols. Reprint. New Delhi: Oriental Books. Tr. (in part) J. A. B.
van Buitenen. 1973–1978. 3 vols. Chicago: University of Chicago Press.
Mammaṭa Bhaṭṭa. Kāvya-prakāśa. Ed. and tr. Ganganatha Jha. Reprint 2006.
Delhi: Bharatiya Vidya Prakashan.
176 Texts and Translations
Maṇikaṇṭha Miśra. Nyāya-ratna. Ed. V. Subrahmanya Sastri and V. Krishna-
macharya. 1953. Nyāyaratna of Maṇikaṇṭha Miśra. Madras: Madras Govern-
ment Oriental Series.
Mathurānātha. Māthurī (commentary on Gaṅgeśa’s TCM). See under Gaṅgeśa.
Mīmāṃsākośa. See Kevalananda Sarasvati.
Mokṣākaragupta. Tarka-bhāṣā. Tr. Yuichi Kajiyama. 1998. An Introduction
to Buddhist Philosophy: An Annotated Translation of the Tarkabhāsā of
Mokṣākaragupta. Vienna: Arbeitskreis für Tibetische und Buddhistische Stu-
dien, Universität Wien.
Nāgārjuna. Mūla-madhyamaka-kārikā. Ed. J. W. de Jong. 1977. Madras: Adyar.
Translation (from the Tibetan): Jay Garfield. 1995. The Fundamental Wis-
dom of the Middle Way: Nāgārjuna’s Mūla-madhyamaka-kārikā. New York:
Oxford University Press.
. Vigraha-vyāvārttinī. Ed. and tr. Kamaleshwar Bhattacharya. 1978. The
Dialectical Method of Nāgārjuna. Delhi: Motilal Banarsidass.
Nīlakaṇṭha. See Annambhaṭṭa.
NyS. See Gautama.
Nyāya-kośa. See Jhalakikar.
Patañjali. Yoga-sūtra. (1) Ed. Jibananda Vidyasagara. 3rd. ed. 1940. With the com-
mentaries of Vyāsa and Vācaspati Miśra. Calcutta. (2) First chapter, ed. (in
a critical edition) Philipp André Maas. 2006. Samādhipāda. Aachen: Shaker.
Translation: Stephen Phillips. 2009. Yoga, Karma, and Rebirth: A Brief History
and Philosophy (Appendix C). New York: Columbia University Press.
PDS. See Praśastapāda.
Pragalbhācārya. See under Gaṅgeśa, TCM analogy chapter.
Praśastapāda. PDS. Padārtha-dharma-saṃgraha. Ed. (with a Hindi translation)
Durgadhara Jha. 1977. Padārthadharmasaṃgraha of Praśastapāda. With the
Nyāyakandalī commentary of Śrīdhara. Varanasi: Sampurnanand Sanskrit Vish-
vavidyalaya. Translation: Ganganatha Jha. 1982 (reprint). Padārthadharmasaṃgraha
of Praśastapāda. Varanasi: Chaukhambha Orientalia.
Raghunātha Śiromaṇi. Padārtha-tattva-nirūpaṇam. Tr. Karl H. Potter. 1957. The
Padārthatattvanirūpaṇam of Raghunātha Śiromaṇi. Cambridge: Harvard-
Yenching Institute.
Ratnakīrti.Kṣaṇa-bhaṅga-siddhiḥ.Ed.AnantalalThakur.1975.Ratnakīrtibandhāvaliḥ.
Patna: Kashi Prasad Jayaswal Research Institute. Anvayātmikā portion tr. Joel
Feldman and Stephen Phillips. Proof of Momentariness by Positive Correlation.
(Forthcoming) New York: American Institute of Buddhist Studies. Vyatirekātmikā
portion tr. A. C. Senape McDermott. 1969. An Eleventh-Century Buddhist Logic
of ‘Exists’. Dordrecht: D. Reidel.
Śabara. Mīmāṃsā-sūtra-bhāṣya. Tr. Ganganatha Jha. Shabara-bhasya. 1933–1936.
Gaekwad’s Oriental Series 66, 70, and 73. Baroda: Oriental Institute.
Śalikanātha Miśra. Prakaraṇa-pañcikā. Ed. A. Subrahmanya Sastri. 1961. Banaras
Hindu University Darśana Series 4. Varanasi. (Partially) translated: K. T. Pan-
durangi. 2004. New Delhi: Indian Council of Philosophical Research.
Śaṅkara. Brahma-sūtra-bhāṣya. Ed. J. L. Shastri. With the Bhāmatī commentary
by Vācaspati Misra. Delhi: Motilal Banarsidass. Translation: Georg Thibaut.
The Vedānta Sūtras: With Commentary by Śaṅkara. New York: Dover.
Śivāditya. Sapta-padārthi. Ed. J. S. Jetly. 1963. Ahmedabad: Lalbhai Dalpatbhai
Bharatiya Sanskriti Vidya Mandira.
Śrīharṣa. Khaṇḍana-khaṇḍa-khādya. Ed. Navikanta Jha. 1970. Kashi Sanskrit
Series 197. Varanasi: Chowkhamba, 1970. Translation: Ganganatha Jha. 1986
(reprint). The Khaṇḍana-khaṇḍa-khādya of Śrīharṣa (“Sweetmeats of Refuta-
tion”). Delhi: Sri Satguru Publications. See also Granoff 1978.
TCM. See Gaṅgeśa.
 Texts and Translations 177
Udayana. Ātma-tattva-viveka. Ed. Vindhyesvariprasada Dvivedin and Lakshmana
Sastri Dravida. 1986 (reprint). Bibliotheca Indica 170. Calcutta: The Asiatic
Society. Translation: N. S. Dravid. 1995. Ātmatattvaviveka by Udayanācārya.
Shimla: Indian Institute of Advanced Study.
. Kiraṇāvalī. Ed. Siva Chandra Sarvvabhouma. 1911 (reprint 1989).
Kiraṇāvalī of Udayana. With the commentary of Vardhamāna. Calcutta: Asi-
atic Society. Translation: see Tachikawa 1981.
. Nyāya-kusumāñjali. Ed. Mahaprabhulal Goswami. 1972. Mithila Insti-
tite Ancient Texts Series 23. Darbhanga: Mithila Research Institute. Tr. N. S.
Dravid, Nyaya-kusumanjali of Udayanacarya. New Delhi: Indian Council of
Philosophical Research, 1996.
. Nyāya-vārttika-tātparya-pariśuddhi. Ed. Anantalal Thakur. 1996. New
Delhi: Indian Council of Philosophical Research.
Uddyotakara. NyV. Nyāya-vārttika. (1) See under Gautama. (2) Ed. Anantalal
Thakur. 1997. Nyāyabhāṣyavārttikam. New Delhi: Indian Council of Philo-
sophical Research.
Vācaspati Miśra. NyVTatp. Nyāya-vārttika-tātparya-ṭīkā (“Notes on [Uddyotaka-
ra’s] Intention in his Nyāya-vārttika Commentary”). (1) See under Gautama. (2)
Ed. Anantalal Thakur. 1996. Nyāyavārttikatātparyaṭīkā of Vāchaspatimiśra.
New Delhi: Indian Council of Philosophical Research.
. Bhāmatī (commentary on Śaṅkara’s Brahma-sūtra-bhāṣya). See under
Śaṅkara. Translation: (fi rst four sūtras) S. S. Suryanarayana Sastri and C. Kun-
han Raja. 1933. The Bhāmatī: Catussūtrī. Madras: Theosophical Publishing
House.
. Tattva-kaumudi. Ed. and tr. Ganganatha Jha. 1896. Reprint 2008. Dillī:
Bhāratīya Buka Kāraporeśana.
. Tattva-vaiśāradī (commentary on the Yoga-sūtra). See under Patañjali.
Vallabha. Nyāya-līlāvatī. Ed. Dhundhiraj Sastri. 1927–1934. With the Nyāya-
līlāvatī-prakāśa commentary of Vardhamāna. Chowkhamba Sanskrit Series
407. Benares.
Vasubandhu. Viṃśatika. Tr. Stefan Anacker. 1984. Seven Works of Vasubandhu.
Delhi: Motilal Banarsidass.
Vātsyāyana. NySBh. Nyāya-sūtra-bhāṣya. (1) See Tarkatirtha et al. editors under
Gautama. (2) See Thakur ed. under Gautama.
Viśvanātha. Bhāṣā-pariccheda (also called Kārikāvalī). Ed. C. Shankar Ram Shas-
tri. With Muktāvalī autocommentary and five sub-commentaries. 1988. Delhi:
Chaukhamba Sanskrit Pratishthan. Translation: (1) Swami Madhavananda.
Bhāṣā-pariccheda with Siddhānta-Muktāvalī. 3rd ed. 1977. Calcutta: Advaita
Ashrama. (2) John Vattanky. Nyāya Philosophy of Language. 1995. New Delhi:
Sri Satguru Publications.
Bibliography

Alston, William. 1989. Epistemic Justification: Essays in the Theory of Knowl-

edge. Ithaca, New York: Cornell University Press.
. 1991. Perceiving God: The Epistemology of Religious Experience. Ithaca,
New York: Cornell University Press.
Armstrong, David M. 1973. Belief, Truth and Knowledge. Cambridge: Cambridge
University Press.
Arnold, Dan. 2006. “On Semantics and Samketa: Thoughts on a Neglected Problem
with Buddhist Apoha Doctrine.” Journal of Indian Philosophy 34: 415–478.
Ayer, A. J. 1946. Language, Truth and Logic. 2nd ed. New York: Dover.
Bagchi, Sitansusekhar. 1953. Inductive Reasoning: A Study of tarka and Its Role
in Indian Logic. Calcutta: Munishchandra Sinha.
Bandopadhyay, Nandita. 1977. The Concept of Logical Fallacies. Calcutta: San-
skrit Pustak Bhandar.
. 1988. “The Concept of Contradiction in Indian Logic and Epistemology.”
Journal of Indian Philosophy 16: 225–246.
Belzile, Jean-François. 2009. Vaincre et Convaincre: Une Dialectique Indienne de
la Certitude (IIIe-VIIe s.), son Éthique et sa Comparaison avec la Dialectique
Grecque. Quebec: Les Presses de l’Université Laval.
Bhatt, Govardhan P. 1989. The Basic Ways of Knowing. 2nd ed. Delhi: Motilal
Banarsidass.
Bhattacharya, Dinesh Chandra. 1958. History of Navya-nyāya in Mithilā.
Darbhanga: Mithila Institute.
Bhattacharya, Gopinath. See Annambhaṭṭa.
Bhattacharyya, Sibajiban. 1987. “Some Principles and Concepts of Nyāya Logic
and Ontology.” In his Doubt, Belief and Knowledge, 201–244. New Delhi:
Indian Council of Philosophical Research.
. 1990a. Gadādhara’s Theory of Objectivity, vol. 1. New Delhi: Indian
Council of Philosophical Research.
. 1990b. “Some Features of the Technical Language of Navya-Nyāya.” Phi-
losophy East and West 40.2 (April): 129–149.
Bijlert, Vittorio A. van. 1989. Epistemology and Spiritual Authority: The Devel-
opment of Epistemology and Logic in the Old Nyāya and the Buddhist
School of Epistemology, with an Annotated Translation of Dharmakīrti’s
Pramāṇavārttika II (Pramān ̣asiddhi) vv. 1–7. Vienna: Arbeitskreis für Tibet-
ische und Buddhistische Studien Universität Wien.
Bilimoria, Purushottama. 1985. “Jñāna and Pramā.” Journal of Indian Philosophy
13: 73–102.
. 1988. Śabdapramāṇa: Word and Knowledge. Dordrecht: Kluwer.
BonJour, Laurence. 2003. “A Version of Internalist Foundationalism.” In Laurence
BonJour and Ernest Sosa, eds. Epistemic Justification. Oxford: Blackwell.
180 Bibliography
Bulcke, Camille. 1947. The Theism of Nyāya-Vaiśeṣika: Its Origin and Early
Development. Delhi: Motilal Banarsidass.
Burge, Tylor. 1998. “Computer Proof, Apriori Knowledge, and Other Minds.”
Philosophical Perspectives 12: 1–37.
Cassam, Quassim. 2007. The Possibility of Knowledge. Oxford: Clarenden.
Chadha, Monima. 2004. “Perceiving Particulars-as-Such Is Incoherent: A Reply to
Mark Siderits.” Philosophy East and West 54.3: 382–389.
Chakrabarti, Arindam. 1994. “Telling as Letting Know.” In Chakrabarti and Mat-
ilal, eds. 1994.
. 1997. Denying Existence: The Logic, Epistemology and Pragmatics of
Negative Existentials and Fictional Discourse. New York: Springer.
. 2000. “Against Immaculate Perception: Seven Reasons for Eliminating nir-
vikalpaka Perception from Nyāya.” Philosophy East and West 50.1: 1–8.
. 2006. “Knowledge from Trusted Tellings and its Preventers.” In
Śabdapramāṇa in Indian Philosophy. Ed. Majulika Ghosh and Bhaswati Bhat-
tacharya Chakrabarti. New Delhi: Northern Book Centre.
Chakrabarti, Arindam, and B. K. Matilal, eds. 1994. Knowing from Words: West-
ern and Indian Philosophical Analysis of Understanding and Testimony. Dor-
drecht: Kluwar.
Chakrabarti, Kisor. 1975. “The Nyāya Theory of Universals.” Journal of Indian
Philosophy 3: 363–382.
. 1984. “Some Remarks on Indian Theories of Truth,” Journal of Indian
Philosophy 12: 339–355.
. 1995. Defi nition and Induction. Honolulu: University of Hawaii Press.
. 1999. Classical Indian Philosophy of Mind: The Nyāya Dualist Tradition.
Albany: State University of New York Press.
Chatterjee, S. C. 1939. The Nyāya Theory of Knowledge: A Critical Study of Some
Problems in Logic and Metaphysics. Calcutta: Calcutta University Press.
Chemparathy, George. 1972. An Indian Rational Theology. Vienna: Indologische
Institut der Universität Wien.
Chisholm, Roderick M. 1942. “The Problem of the Speckled Hen.” Mind 51.
. 1977. Theory of Knowledge. 2nd ed. Englewood Cliffs, New Jersey:
Prentice-Hall.
Crane, Tim. 2011. “The Problem of Perception.” The Stanford Encyclopedia of
Philosophy (Spring 2011 Edition), Edward N. Zalta (ed.), URL = <http://plato.
stanford.edu/archives/spr2011/entries/perception-problem/>.
Dasti, Matthew, and Stephen Phillips. 2010. “Pramāṇa Are Factive—A Response
to Jonardon Ganeri.” Philosophy Easy and West 60.4: 535–540.
Davis, Lawrence. 1981. “Tarka in the Nyāya Theory of Inference.” Journal of
Indian Philosophy 9: 227–257.
Dravid, N. S. 1996. “The Nyāya-Vaiśeṣika Explanation of Illusion.” Journal of
Indian Philosophy 24: 37–48.
Dravid, Raja Ram. 1972. The Problem of Universals in Indian Philosophy. Delhi:
Motilal Banarsidass.
Dreyfus, Georges B. J. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and
Its Tibetan Interpretations. Albany: State University of New York Press.
Ducasse, C. J. 1951. Nature, Mind, and Death. La Salle, Illinois: Open Court.
Dummett, Michael, 1978. “Frege’s Distinction between Sense and Reference.” In
his Truth and Other Essays. London: Duckworth.
. 1994. “Testimony and Memory.” In Chakrabarti and Matilal, eds. 1994.
Dunne, John D. 2004. Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom
Publications.
Feldman, Joel. 2005. “Vasubandhu’s Illusion Argument and the Parasitism of Illu-
sion upon Veridical Experience.” Philosophy East and West 55.4: 529–541.
 Bibliography 181
Feldman, Richard. 1985. “Reliability and Justification.” The Monist 68.2:
159–174.
Firth, Roderick. 1965. “Sense Data and the Percept Theory.” In John Troyer, ed.
1998. In Defense of Radical Empiricism: Essays and Lectures by Roderick
Firth. Lanham, Maryland: Rowman & Littlefield Publishers.
Fish, William. 2010. Philosophy of Perception. London: Routledge.
Franco, Eli. 2004. “A Note on Nāgārjuna and the Naiyāyikas.” In Shoun Hino
and Toshihiro Wada, eds. Three Mountains and Seven Rivers: Prof. Musashi
Tachikawa’s Felicitation Volume. Delhi: Motilal Banarsidass.
. ed. 2009. In collaboration with Dagmar Eigner. Yogic Perception, Medita-
tion and Altered States of Consciousness. Sitzungsberichte der Österreichischen
Akademie der Wissenschaften, Philosophische-Historische Klasse, 794 Band.
Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Ganeri, Jonardon. 1996. “Numbers as Properties of Objects: Frege and the Nyāya.”
In Epistemology, Meaning and Metaphysics after Matilal. Ed. A. Chakrabarti.
Studies in Humanities and Social Sciences 3.2: 111–121.
. 1999. Semantic Powers: Meaning and the Means of Knowing in Classical
India. Oxford: Clarendon Press.
. 2001. Philosophy in Classical India: The Proper Work of Reason. London:
Routledge.
. ed. 2001b. Logic in India: A Reader. Richmond, Surrey: Curzon.
. 2003. “Ancient Indian Logic as a Theory of Case-Based Reasoning.” Jour-
nal of Indian Philosophy 31: 33–45.
. 2011. The Lost Age of Reason: Philosophy in Early Modern India 1450–
1700. Oxford: Oxford University Press.
Ganguly, Krishna Chakravorty. 1993. A Bibliography of Nyāya Philosophy. Cal-
cutta: National Library.
Gerow, Edwin. 1971. A Glossary of Indian Figures of Speech. The Hague:
Mouton.
. 2001. See Appaya Dīkṣita.
Gettier, Edmund. 1963. “Is Justified True Belief Knowledge.” Analysis 23:
121–123.
Gillon, Bendon S. 1997. “Negative Facts and Knowledge of Negative Facts.” In
Purushottama Bilimoria and J. N. Mohanty, eds. Suffering and Beyond: Essays
in Honor of Bimal K. Matilal. New Delhi: Oxford University Press.
Ginet, Carl. 1975. Knowledge, Perception, and Memory. Dordrecht: D. Reidel.
Gokhale, Pradeep P. 1992. Inference and Fallacies Discussed in Ancient Indian
Logic, with Special Reference to Nyaya and Buddhism. Delhi: Sri Satguru
Publications.
Goldman, Alvin. 1967. “A Causal Theory of Knowledge.” Journal of Philosophy
64.12: 355–372. Reprinted in George S. Pappas and Marshall Swain, eds. 1978.
. 1999. Knowledge in a Social World. Oxford: Oxford University Press.
. 2002. Pathways to Knowledge: Private and Public. Oxford: Oxford Uni-
versity Press.
Gradinarov, P. I. 1990. Phenomenology and Indian Philosophy: Studies in
Nyāya-Vaiśeṣika Transcendental Logic and Atomism. Delhi: Ajanta Books
International.
Graham, Peter. 2006. “Liberal Fundamentalism and Its Rivals.” In J. Lackey and
E. Sosa, eds.
Granoff, Phyllis. 1978. Philosophy and Argument in Late Vedānta (with a transla-
tion of part of Śrīharṣa’s Khaṇḍana-khaṇḍa-khādya). Dordrecht: D. Reidel.
Grice, H. Paul. 1961. “The Causal Theory of Perception.” Reprinted in Robert J.
Swartz, ed. 1965. Perceiving, Sensing, and Knowing. Berkeley: University of
California Press.
182 Bibliography
. 1975. “Logic and Conversation.” In Syntax and Semantics, vol. 3. Ed.
Peter Cole and Jerry L. Morgan. New York: Academic Press.
Gupta, Bina. 1998. The Disinterested Witness: A Fragment of Advaita Vedānta
Phenomenology. Evanston, Illinois: Northwestern University Press.
Halbfass, Wilhelm. 1980. “Karma, apūrva, and ‘Natural’ Causes.” In Wendy Doni-
ger O’Flaherty, ed. Karma and Rebirth in Classical Indian Traditions. Berkeley:
University of California Press, 268–302.
. 1992. On Being and What There Is: Classical Vaiśeṣika and the History of
Indian Ontology. Albany: State University of New York Press.
Hamblin, C. L. 1970. Fallacies. London: Methuen.
Harmon, Gilbert. 1968. “Knowledge, Inference, and Explanation.” American
Philosophical Quarterly 5: 164–173.
Hayes, Richard P. 1987. Dignāga on the Interpretation of Signs. Dordrecht:
Kluwar.
Hempel, Carl. 1965. Aspects of Scientific Explanation. New York: Macmillan.
Herzberger, Radhika. 1986. Bhartṛhari and the Buddhists. Dordrecht: D. Reidel.
. 1986. “Apoha and Śiṃśapāvṛkṣa.” In B. K. Matilal and R. D. Evans, eds.
Hume, David. 1748. An Enquiry Concerning Human Understanding. Ed. Tom L.
Beauchamp. 2000. Oxford: Clarendon.
Ingalalli, R. I. 1988. Śabda Pramāṇa: An Epistemological Analysis. Delhi: Sri Sat-
guru Publications.
. 1989. Meaning and Knowledge: An Interpretation of Indian and Contem-
porary Epistemological Concepts. Delhi: Sri Satguru Publications.
Ingalls, Daniel H. H. 1951. Materials for the Study of Navya-Nyāya Logic. Cam-
bridge: Harvard University Press.
Jacob, Colonel G. A. 1925. A Second Handful of Popular Maxims Current in San-
skrit Literature. Bombay: Nirnaya Sagar Press.
Jha, Ganganatha. 1964. Pūrva-Mīmāṃsā in Its Sources. Varanasi: Benaras Hindu
University Press.
. 1978. The Prābhākara School of Pūrva Mīmāṃsā. Delhi: Motilal
Banarsidass.
Jha, V. N. 1986. Studies in Language, Logic and Epistemology. Delhi: Pratibha
Prakasan.
Junankar, N. S. 1978. Gautama: the Nyāya Philosophy. Delhi: Motilal
Banarsidass.
Kahrs, Eivind. 1998. Indian Semantic Analysis. Cambridge: Cambridge University
Press.
Kaviraj, Gopinath. The History and Bibliography of Nyāya-Vaiśeṣika Literature.
Ed. Gaurinath Sastri. 1982 (reprint). Varanasi: Saraswati Bhavana Studies.
Kelly, Thomas. 2005. “The Epistemic Significance of Disagreement.” In Oxford
Studies in Epistemology, vol. 1. Ed. Tamar Szabo Gendler and John Haw-
thorne. Oxford: Clarendon Press.
Kleine, Peter. 1971. “A Proposed Defi nition of Propositional Knowledge.” Journal
of Philosophy 68: 471–482.
Krishna, Daya. 2004a. “Nyāya: Realist or Idealist: Is the Debate Ended, the Argu-
ment Concluded.” In Daya Krishna, ed. Discussion and Debate in Indian
Philosophy: Vedānta, Mīmāṃsā and Nyāya. New Delhi: Indian Council of
Philosophical Research.
. 2004b. The Nyāyasūtras: A New Commentary on an Old Text. Delhi: Sri
Satguru Publications.
Kumar, Shiv. 1980. Upamāna in Indian Philosophy. Delhi: Eastern Book Linkers.
Kuznetsova, Irina, Jonardon Ganeri, and Chakravarthi Ram-Prasad, eds. Forth-
coming. Self: Hindu and Buddhist Ideas in Dialogue.
 Bibliography 183
Lackey, J., and E. Sosa, eds. 2006. The Epistemology of Testimony. Oxford:
Oxford University Press.
Lehrer, Keith. 1990. Theory of Knowledge. London: Routledge.
Lehrer, Keith, and Thomas D. Paxson, Jr. 1969. “Knowledge: Undefeated Justified
True Belief.” Journal of Philosophy 66.8: 225–237. Reprinted in Pappas and
Swain, eds. 1978.
Lindtner, Christian. 1986. Nagarjuniana. Delhi: Motilal Banarsidass.
Matilal, B. K. 1968. The Navya-Nyāya Doctrine of Negation. Cambridge: Har-
vard University Press.
. 1986. Perception: An Essay on Classical Indian Theories of Knowledge.
Oxford: Oxford University Press.
. 1998. The Character of Logic in India. Ed. Jonardon Ganeri and Heera-
man Tiwari. Albany: State University of New York Press.
. 2002. Mind, Language and World: The Collected Essays of Bimal Krishna
Matilal. Ed. Jonardon Ganeri. New Delhi: Oxford University Press.
Matilal, B. K., and R. D. Evans, eds. 1986. Buddhist Logic and Epistemology.
Dordrecht: Kluwar.
McCrea, Larry. 2009. “‘Just Like Us, Just Like Now’: The Tactical Implications of
the Mīmāṃsā Rejection of Yogic Perception.” In Franco, ed. 2009: 55–70.
McDermott, A. C. Senape. See Ratnakīrti.
Meurath, Annette. 1996. Untersuchungen zur Kompositionsgeschichte der
Nyāyasūtras. Würzberg: Echter Verlag.
Mill, John Stuart. 1843. A System of Logic, Ratiocinative and Inductive. Ed. J. M.
Robson, 1973. Collected Works of John Stuart Mill, vol. 7. Toronto: University
of Toronto Press.
Mohany, J. N. 1992. Reason and Tradition in Indian Thought. Oxford:
Clarenden.
. 1994. “Is There an Irreducible Mode of Word-Generated Knowledge?” In
Chakrabarti and Matilal, eds. 1994.
. 2000. Classical Indian Philosophy. Lanham, Maryland: Rowman and
Littlefield.
Mondal, Pradyot Kumar. 1982. “Some Aspects of Perception in Old Nyāya.” Jour-
nal of Indian Philosophy 10: 357–376.
Mookerji, Satkari. 1975. The Buddhist Philosophy of Universal Flux: An Expo-
sition of the Philosophy of Critical Realism as Expounded by the School of
Dignāga. Delhi: Motilal Banarsidass.
Mukhopadhyay, P. K. 1984. Indian Realism. Calcutta: K. P. Bagchi.
Narain, Harsh. 1976. Evolution of the Nyāya-Vaiśeṣika Categoriology. Varanasi:
Bharati Prakashan.
Narayanan, T. K. 1992. Nyāyasāra of Bhāsarvajña: A Critical Study. New Delhi:
Mittal Publications.
Nozick, Robert. 1981. Philosophical Explanations. Cambridge, Massachusetts:
Harvard University Press.
Oetke, Claus. 1996. “Ancient Indian Logic as a Theory of Non-Monotonic Rea-
soning.” Journal of Indian Philosophy 24: 447–539.
. 2004. “The Role of the Example in Ancient Indian Logic.” In Shoryu Kat-
sura and Ernst Steinkellner, eds. The Role of the Example (dṛṣṭānta) in Classical
Indian Logic. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien
Universität Wien.
Pappas, George S., and Marshall Swain, eds. 1978. Essays on Knowledge and Jus-
tification. Ithaca, New York: Cornell University Press.
Patil, Parimal G. 2009. Against a Hindu God: Buddhist Philosophy of Religion in
India. New York: Columbia University Press.
184 Bibliography
Perrett, Roy. 1984. “The Problem of Induction in Indian Philosophy.” Philosophy
East and West 34: 161–174.
. 1985. “A Note on the Navya-Nyāya Account of Number.” Journal of
Indian Philosophy 13: 227–234.
. 1999. “Is Whatever Exists Knowable and Nameable.” Philosophy East and
West 49: 410–414.
Phillips, Stephen. 1995. Classical Indian Metaphysics: Refutations of Realism
and the Emergence of “New Logic.” Chicago: Open Court. Delhi: Motilal
Banarsidass.
. 2001. “There’s Nothing Wrong with Raw Perception.” Philosophy East
and West 51.1: 104–13.
. 2004. “Perceiving Particulars Blindly: Remarks on a Nyāya-Buddhist Con-
troversy.” Philosophy East and West 54.3: 389–403.
. 2005. “Self as Locus/Substratum (adhikaraṇa) of Psychological Conti-
nuities and Discontinuities.” American Philosophical Association Newsletter
05.1: 4–8.
. 2009. Yoga, Karma, and Rebirth: A Brief History and Philosophy. New
York: Columbia University Press.
Pingree, David. 1981. Jyotiḥśāstra: Astral and Mathematical Literature. Wies-
baden: Otto Harrassowitz.
Pollock, John. 1986 Contemporary Theories of Knowledge. Towota, New Jersey:
Rowman & Littlefield.
Potter, Karl H. 1961. “Ontology of Concrete Connectors.” Journal of Philosophy
58.3.
, ed. Encyclopedia of Indian Philosophies, vol. 1. Bibliography. http://fac-
ulty.washington.edu/kpotter/.
, ed. 1977. Encyclopedia of Indian Philosophies, vol. 2. Nyāya-Vaiśeṣika up
to Gaṅgeśa. Delhi: Motilal Banarsidass.
Potter, Karl H., and Sibajiban Bhattacharyya, eds. 1993. Encyclopedia of Indian
Philosophies, vol. 6. Indian Philosophical Analysis: Nyāya-Vaiśeṣika from
Gaṅgeśa to Raghunātha Śiromaṇi. Delhi: Motilal Banarsidass.
Prasad, Rajendra. 2002. Dharmakīrti’s Theory of Inference. New Delhi: Oxford
University Press.
Preisendanz, Karin. 1989. “On ātmendriyamanorthasannikarṣa and the Nyāya-
Vaiśeṣika Theory of Vision.” Berliner Indologische Studien 4–5: 141–213.
. 1994. Studien zu Nyāyasūtra III.1 mit den Nyāyatattvāloka Vācaspati
Miśras II. 2 vols. Alt- und Neuindische Studien 46. Stuttgart.
. 2005. “The Production of Philosophical Literature in South Asia during
the Pre-Colonial Period (15th to 18th Centuries): The Case of the Nyāyasūtra
Commentarial Tradition.” Journal of Indian Philosophy 33: 55–94.
Prets, Ernst. 2001. “Futile and False Rejoinders, Sophistical Arguments and Early
Indian Logic.” Journal of Indian Philosophy 29: 545–558.
Price, H. H. 1969. Belief. London: Allen & Unwin.
Putnam, Hilary. 1982. Reason, Truth, and History. Cambridge: Cambridge Uni-
versity Press.
Quine, W. V., and J. S. Ullian. 1978. The Web of Belief. 2nd ed. New York: Ran-
dom House.
Raja, K. Kunjunni. 1969. Indian Theories of Meaning. 2nd ed. Madras: Adyar.
Ramanuja Tatacharya, N. S. 1992. Pratyakṣatattvacintāmaṇi Vimarśa: A Compre-
hensive and Evaluative Study of the Pratyakṣakhaṇḍa of the Tattvacintāmaṇi.
(In Sanskrit.) Tirupati: Rashtriya Sanskrit Vidyapeetha.
. 2005. Śābdabodhamīmāṃsā: An Inquiry into Indian Theories of Verbal
Cognition. Pt. 1: “The Sentence and Its Significance.” (In Sanskrit.) New Delhi:
Institut Français de Pondichèry; Rashtriya Sanskrit Sansthan.
 Bibliography 185
Ram-Prasad, Chakravarthi. 2002. Advaita Metaphysics and Epistemology. Lon-
don: RoutledgeCurzon.
. Forthcoming. “Self and Memory: Personal Identity and Unified Conscious-
ness in Comparative Perspective.” In Irina Kuznetsova, Jonardon Ganeri, and
Chakravarthi Ram-Prasad, eds.
Rao, Srinivasa. 1998. Perceptual Error: The Indian Theories. Honolulu: University
of Hawaii Press.
Ruegg, David Seyfort. 1986. “Does the Madhyamika Have a Thesis and Philo-
sophical Position.” In Matilal and Evans.
Russell, Bertrand. 1912. The Problems of Philosophy. London: Williams and
Norgate.
Saha, Sukharanjan. 1991. Meaning, Truth and Predication: A Reconstruction
of Nyaya Semantics. Calcutta: Jadavpur University and K. P. Bagchi and
Company.
. 2000. Epistemology in Pracina and Navya Nyaya. Jadavpur Studies in
Philosophy. Kolkata: Jadavpur University.
Sastri, Gaurinath. 1959. The Philosophy of Word and Meaning: Some Indian
Approaches with Special Reference to the Philosophy of Bhartṛhari. Calcutta:
Sanskrit College of Calcutta.
Schayer, Stanislaw. 1933. “Studies in Indian Logic.” In Ganeri 2001b.
Schweizer, Paul. 1994. “Momentary Consciousness and Buddhist Epistemology.”
Journal of Indian Philosophy 22.
Shastri, D. N. 1964. Critique of Indian Realism: A Study of the Conflict Between the
Nyāya-Vaiśeṣika and Buddhist Dignāga School. Delhi: Motilal Banarsidass.
Shaw, Jayasankar Lal. 1978. “The Nyāya on Existence, Knowability and Name-
ability.” Journal of Indian Philosophy 5: 255–266.
Shope, Robert K. 1983. The Analysis of Knowing: A Decade of Research. Princeton:
Princeton University Press.
Siderits, Mark. 1991. Indian Philosophy of Language. Dordrecht: Kluwar.
. 2004. “Perceiving Particulars: A Buddhist Defense.” Philosophy East and
West, 54(3): 367–382.
Solomon, Ester. 1976. Indian Dialectics. 2 vols. Ahmedabad: Gujarat Vidya
Sabha.
Sosa, Ernest. 1964. “The Analysis of ‘knowledge that p’.” Analysis 25: 1–8.
. 1980. “The Raft and the Pyramid: Coherence versus Foundations in the
Theory of Knowledge.” Midwest Studies in Philosophy, vol. 5. Studies in Episte-
mology. Ed. P. French et al. Minneapolis: University of Minnesota Press. Repub-
lished in Sosa 1991.
. 1991. Knowledge in Perspective. Cambridge: Cambridge University Press.
. 2003. “Beyond Internal Foundations to External Virtues.” In Laurence
BonJour and Ernest Sosa, eds. Epistemic Justification. Oxford: Blackwell.
Staal, J. F. 1962. “Negation and the Law of Contradiction in Indian Thought.”
Bulletin of the School of Oriental and African Studies 25.
. 1973. “The Concept of Pakṣa in Indian Logic.” Journal of Indian Philoso-
phy 2.2 (reprinted in Ganeri 2001b).
Steinkellner, Ernst. 1991. “On the Interpretation of the svabhāvahetu in
Dharmakīrti’s Vādanyāya.” In E. Steinkellner, ed. Studies in the Buddhist Epis-
temological Tradition. Proceedings of the Second International Dharmakīrti
Conference, Vienna, 1989. Wien: Verlag der Österreichischen Akademie der
Wissenschaften.
Strawson, P. F. 1994. “Knowing from Words.” In A. Chakrabarti and B. K. Mati-
lal, eds.
Swain, Marshall. 1974. “Epistemic Defeasibility.” The American Philosophical
Quarterly 11.1: 15–25. Reprinted in Pappas and Swain, eds. 1978.
186 Bibliography
Taber, John. 2009. “Yoga and Our Epistemic Predicament.” In Franco, ed:
71–92.
. Forthcoming. “Uddyotakara’s Defense of a Self.” In Irina Kuznetsova, Jon-
ardon Ganeri, and Chakravarthi Ram-Prasad, eds.
Tachikawa, Musashi. 1981. The Structure of the World in Udayana’s Realism
(including editions and translations of Udayana’s Lakṣaṇāvalī and a portion of
his Kiraṇāvalī). Dordrecht: D. Reidel.
Tillemans, Tom J. F. 1999. Scripture, Logic, Language: Essays on Dharmakīrti
and His Tibetan Successors. Boston: Wisdom Publications.
Tuske, Joerg. 2008. “Teaching by Example: An Interpretation of the Role of
Upamāna in Early Nyāya Philosophy.” Asian Philosophy 18.1: 1–15.
Verpoorten, Jean-Marie. 1987. Mīmāṃsā Literature. Wiesbaden: Otto
Harrassowitz.
Wada, Toshihiro. 1990. Invariable Concomitance in Navya Nyāya. Delhi: Sri Sat-
guru Publications.
. 2000. “The Origin of Navya-nyāya and Its Place within the History of
Indian Logic.” In S. Hing and T. Wada, ed. Three Mountains and Seven Riv-
ers: Prof. Musashi Tachikawa’s Felicitation Volume, 439–462. Delhi: Motilal
Banarsidass.
. 2007. The Analytical Method of Navya-Nyāya. Groningen: Egbert
Forsten.
Wittgenstein, Ludwig. 1953. Philosophical Investigations. Tr. G. E. M. Anscomb.
New York: Macmillan.
. 1969. On Certainty. Ed. G. E. M. Anscombe and G. H. von Wright.
Oxford: Basil Blackwell.
Index

A 130, stock example 75, see also

abduction 51 statements, analogical
ābhāsa (the apparent, pseudo-) 10, 11, Ānandavardhana 95, 153n55, 153n58,
41, 42, 101, 135n22, 137n13, 159
162 Annambhaṭṭa 92, 94, 134n13, 151n27,
abhāva, see absence 152n48, 153n52, 153n53,
abhidhā, see meaning, reference 153n56, 153n57, 159
abhihita-anvaya-vāda (theory of the anubhava (presentational experience)
connection of the referents) 20, 21, 28, 29, 33, 37, 42, 64,
88–89, 162 98, 102, 138n25, 163, 165
Abhinava Gupta 95 anumiti, see inferential knowledge
absence (abhāva, negative fact) 24, anupalabdhi (non-cognition, non-
46, 47–48, 61–62, 66, 141n32, perception) 30, 48, 64, 163,
162, 168, knowledge of 40, 155n9, 163
64, 142n41, 162, 168, see also anvaya (connectedness of meaning) 89,
anupalabdhi 92, 123, 125, 152n27, 152n46,
action (voluntary, pravṛtti) 15, 17, 19, 163
56, 76, 98, 108, 134, knowl- anvaya-vyatireka (positive and negative
edge and 6, 11, 15–16, 17, 32, correlations) 23, 32, 54, 58–59,
142n39, 165, 166 62, 66, 163, 171, see also gener-
adṛṣṭa (Unseen Force) 141–142n33, alization, inductive
162, 165, 166 anvita-abhidhāna-vāda (theory of ref-
Advaita Vedānta 159, 160, 161 erence of the connected) 88, 163
aesthetics, classical Indian (alaṃkāra- anyathā-khyāti, see under illusion
śāstra) 76, 93, 152n45, 162 Appaya Dīkṣita 152n45, 153n54
āharya-buddhi 29, 97, 98 apperception (anuvyavasāya) 20,
ākāṅkṣā (syntactic expectation) 55, 88, 22–23, 44, 48–50, 85–86,
90, 162 135n21, 137n11
ākāśa (ether) 27, 46, 78, 121, 153n51, āpta (expert and trustworthy testifier)
155n18, 162, 164 83, 84, 94, 99, 123, 150n4,
ālaya-vijñāna (storehouse conscious- 150n9, 154n7, 164
ness) 134n13, 162 Armstrong, David 97, 153n2
analogical knowledge (upamiti) 29, 75, artha-kriyā 143n12
78, 116, 130, 155n9, 170, defi ni- arthāpatti (circumstantial implication,
tion of 131–132, 155n9, 156n15 postulation) 24, 30, 63, 77, 117,
analogy (upamāna) 46, 64, 74–81, 119, 146n41, 155n11, 164
103, 107, 112ff, 118, 119, 120ff, Asaṅga 134n13, 141n23, 161
124, 126ff, 129ff, 138n27, āsatti (proper proximity) 90, 164
149n20, 170, defi nition 107, astrology 102
188 Index
ātman, see self certainty (niścaya) 56, 126, 127,
atoms 23, 44, 52, 70, 138n16, 142n34 152n38, 166
attention 45, 49–50, 141n30 certification 5, 11–12, 13, 14, 15, 16,
Ayer, A.J. 17 17–32, 47, 48, 50, 99, 101, 167,
certifiability 21–22, 164, condi-
B tions of 12, 44, 45, 55, 57, 62,
Bagchi, Sitansusekhar 139n32, 63, 83, 84, 85, 86, 150n7, 167,
139n33, 139n34 see also guṇa and doṣa, extrin-
Bandopadhyay, Nandita 147n62 sic 134n8, 167, objective 13, see
belief 5, 6–7, 49, 134n11, perceptual also under justification, pseudo-
38, see also statements, observa- (prāmāṇya-ābhāsa) 21, 42, 100,
tion, see also cognition 167, self-certification 18, 22,
Bhartṛhari 35, 88, 89, 140n8, 159 138n15, 170, source identifica-
Bhāsarvajña 4, 30, 134n12, 135n15, tion 11, 14, 15, 17, 21, 23, 24,
139n1, 148n7, 159 34, 47, 48–49, 97
Bhāṭṭas (followers of Kumārila) 24, Chakrabarti, Arindam 73, 84, 90,
30, 40, 48, 77, 88, 89, 108–109, 139n1, 139n4, 140n14, 142n41,
116–118, 152n29, 155n9, 159, 142n42, 150n3, 150n9, 152n36,
162, 163, 164 154n6
Bhattacharya, Gopinath 95, 151n27, Chakrabarti, Kisor 136n8, 138n15,
153n52 138n26, 141n29, 143n4,
Bhattacharyya, Sibajiban 91, 134n14, 144n30, 145n39, 147n63
139n4, 142n42, 143n4, 144n23, Chemparathy, George 148n65
145n38, 151n23, 152n43, 154n10 Chisholm, Roderick 138n17
bhūyo-darśana (wide experience) 54, circumstantial implication, see
148n4, 164 arthāpatti
BonJour, Laurence 50, 98–99, 135n24, cognition (jñāna) 5ff, 42, 61, 140n15,
142n43, 153n2, 153n3 141n30, 162, 165, as verbaliz-
Bradley problem 67 able 38, chart of types 29, inten-
Brahman (the Absolute) 72, 159, 164 tionality of, see intentionality,
Buddhist logic 54, 170 non-veridical 8, see also illusion,
Buddhist philosophy 30, 36, 37, 39, recurrent 25, 28, 145n39
45, 47, 60, 61, 67, 73, 77, 166, cognizedness 41
168, apoha theory 89, 138n19, coherentism 14, 15, 19
152n29, see also Yogācāra, sva- concepts 25, 34, 35, 37, 38, 42, 50, 62
lakṣaṇa conjunction 108
Bulke, Camille 146n51 contradiction (self-contradiction), 31,
Burge, Tyler 102, 103, 154n11 three kinds 32, 136n8, see also
defeaters
C conventions (rūḍhi) 75–77, 93, 149n14,
Caraka 142n34, 159 168, 172
Cartesian epistemology 15, 19 counting 102
Cārvāka 30, 56, 144n18, 154n7
Cassam, Quassim 102, 103, 154n11, D
154n12 Davidson, Donald 145n38
category (padārtha) 28, 46, 61, 111, debate 71–73, 165, types 72
165, 167 defeasibility (bādhaka-abhāva) 18, 21,
causal fallacy 96 82, 164, see also fallibilism
causality (kāraṇatā) 45, 64–65, defeaters, epistemic (bādhaka) 18, 22,
145n37, 145n38, 166, 168, 56, 82, 99, 106, 136n27, 162,
causes as reasons 96, inherent 164, 165
(samavāya) 65, 145n37, 169, suf- defi nitions 66
ficient (sāmagrī) 45, 169, trigger Descartes 17, see also Cartesian
(karaṇa) 65, 145n37, 166 epistemology
 Index 189
desire 54, 67–68, 127 Feldman, Richard 138n18
destruction 71, see also absence fictions (imaginations) 28, 29, 62, see
dharma (righteousness) 19, 150n3, 164 also āharya-buddhi
Dharmakīrti 3, 42–43, 47–48, 51, 53, foundationalism 14, 15, 19, 20–23, 50,
146n41, 159, 161 136, see also under knowledge
Dharmottara 53, 143n8, 143n10 Franco, Eli 136n4
dhvani, see under meaning Frege, G. 78, 88, 97
Dignāga 42, 51, 53, 54, 74, 86,
trairūpya-hetu (threefold test of G
the prover) 57, 161, 170 Gadādhara 140n9, 145n31, 151n23,
disjunctivism 10 152n32, 159
dispositions, mental (saṃskāra, Ganeri, Jonardon 56, 57, 58, 59,
memory dispositions) 7, 8, 36, 133n3, 137n14, 139n30,
37, 39, 46, 91, 134n13, 163, 139n32, 143n5, 144n20,
164, 169, 170 144n22, 149n14, 154n10
distinctness (bheda) 110 Gaṅgeśa 2, 6, 20, 25, 26, 27, 32, 35,
doṣa, see guṇa and doṣa 38, 43, 47, 53, 56, 57, 59–60,
doubt 2, 12, 15, 17, 18, 20, 21, 22, 64, 66, 67–68, 69, 70–71, 73,
23, 29, 59, 82, 90, 135n22, 77, 78, 79, 80–81, 82, 84–85,
136n8,142n39, 169 89, 92, 95, 135n22, 138n27,
doxastic ascent 136–137n8 140n9, 140n15, 146n42,
Dravid, Raja Ram 138n26, 145n39 147n54, 149n12, 149n18,
Dreyfus, Georges 141n22 149n19, 149n21, 150n9, 151–
Ducasse, C.J. 141n29 152n28, 155n5, 157n28, 159,
Dummett, Michael 135n22, 148n9 169, Tattva-cintā-maṇi (TCM)
Dunne, John 141n22, 143n11 2, 105–106, 134n7, 134n9,
135n19, 135n25, 136n26,
E 136n8, 137n11, 138n20,
earthen, the 46, 66, 140n9, 164 138n23, 138n28, 140n11,
effort 7, 43–44 140n13, 141n25, 141n28,
eka-vṛtti-vedyatā 26, 77, 134n12, 141n30, 141n32, 142n33,
149n17, 165 142n39, 142n41, 142n42,
elenchus 15, 154n7 143n13, 144n25, 144n29,
empiricism 3, 44, 103 145n32, 145n35, 145n40,
endurance 40 146n46, 146n53, 148n2, 148n3,
epistemology 9, 52, analytic 50, 96–103 148n4, 148n6, 149n22, 149n24,
ether, see ākāśa 149n25, 150n1, 150n8, 150n10,
etymology (grammatical derivation) 150n11, 151n12, 151n15,
76–77, 79, 93, 130, 149n20, 152n35, 152n37, 152n39,
156n19, 171, 172 152n42, 152n46, 152n47,
experience, presentational, see 153n49, 153n52, TCM analogy
anubhava chapter 105–132
externalism, epistemic 5, 13–15, Ganguly, Krishna 133n3
98–99, new evil-demon problem Gautama 3, 4, 18, 37, 45, 48, 51, 72,
101, 136n28 75, 77, 83, 92, 93, 157n33,
extrapolation 54, 56–57 159, as Akṣapāda 133n2,
Nyāya-sūtra (NyS) 2, 3–4,
F 18, 25, 42, 44, 51, 53, 54, 72,
fallacies 51, 54, 71–73, 166, 169, 171, 73, 76, 83, 87, 133n1, 133n2,
see also defeaters 133n4, 133n5, 134n15, 135n22,
fallibilism 20–21, see also defeasibility 139n1, 139n2, 140n10, 140n12,
falsity, see under veridicality 140n16, 140n20, 142n36,
familiarity (abhyāsa-daśā-āpanna) 47, 142n40, 147n59, 148n3, 148n8,
138n15, 142n39, 162 148n10, 148n11, 150n2, 150n4,
190 Index
150n5, 150n6, 151n13, 151n21, smaraṇa, Dharmakīrti’s types
152n44, 157n28, defi nition of 64, for oneself (svārtha) 53–56,
perception 34 67, 170, for others (parārtha)
generality problem 24–30, 44–48, 101 53–56, 67, 167, 168, 170,
generalization, inductive (induc- negative-only (kevala-vyatireka)
tion) 51, 58–60, 64, see also 62, 65–68, 103, 117, 119, 129,
anvaya-vyatireka and inferential 145n41, 157n28, 166, positive-
example only (kevala-anvaya) 62, 166,
Gerow, Edwin 152n43, 153n54 pseudo- 100, see also fallacies,
Gettier, Edmund 135n23, see also sāmānyato dṛṣṭa (by general
under knowledge principle) 126, 127, 157n28, 169,
Ginet, Carl 135n24, 154n6 self-certifying 138n15, terms of
God, see īśvara 54, Uddyotakara’s types 62
Gokhale, Pradeep 147n62 inferential example (dṛṣṭānta) 54, 70,
Goldman, Alvin 102, 154n11 75, 112, 144n15, 165
Graham, Peter 87, 151n20 inferential knowledge (anumiti) 29, 33,
grammar 78, 171, see also ākāṅkṣā 57, 58, 75, 127, 163
grammarian literature 78, 160 inferential probandum (sādhya) 52ff,
Grice, H. Paul 90, 152n33 144n15, 168, familiar (prasid-
guṇa and doṣa (epistemic merits and dha) 53, 66, 68
flaws) 12, 14, 23–24, 28, 42, 47, inferential prover (hetu, sādhana) 9,
49, 98. 100, 164, 165, 172, see 51ff, 144n15, 165, 168, pseudo-
also certification (hetvābhāsa) 165, see also
fallacies
H inferential reflection (parāmarśa) 22,
Halbfass, Wilhelm 141n33 55–56, 57, 58, 75, 120, 144n25,
Hamblin, C. L. 147n59 167
Harman, Gilbert 154n7 inferential subject (pakṣa) 52ff, 65–66,
Hempel, Carl 146n43 68, 112, 143n13, 144n15,
Hinduism 83 145n32, 167
homonyms 89 inferential undercutting condition, see
Hume, David 45, 86, 151n17 upādhi
infi nitism 136n6
I Ingalls, Daniel 143n4, 143n11, 144n14,
illusion (pseudo-perception, pratyakṣa- 144n29
ābhāsa) 9, 28, 36–37, 41–44, inherence 61, 145n37, 169
49, 170, anyathā khyāti (presen- intentionality (viṣayatā, objecthood)
tation otherwise) 8, 38, 41, 49, 6, 7, 8–9, 26, 33, 41, 48, 61,
61, 134–135n15, 163, concept 151n2, 159, 165, 168, 171
as parasitic on that of veridical internalism, epistemic 5, 13–16, 19
experience 42, 135n22, 140n10 intersubjectivity 8
imagination, see āharya-buddhi irony 153n54
India 2 īśvara (God, Lord) 52, 70, 71, 75,
indirect indication (upalakṣaṇa) 78–80, 93, inference to 68–70, 130,
121, 122, 149n22, 149n25, 170 146n52, 147n64, 148n6,
individuals 38–39, 77, 92, 170, 171, 149n14, 165
see also qualificanda, sva-
lakṣaṇa J
induction, see generalization, inductive Jacob, Colonel 140n10
inference (anumāna) 10, 51–73, 74, Jagadīśa 140n9
95, 103, 116, 119, 162, and Jagannātha 95
testimony 85–87, certifica- Jaina philosophy 30, 74, 159, 162
tion conditions 55–56, see Jayanta Bhaṭṭa 4, 47, 106, 120–121,
also pakṣa-dharmatā, vyāpti- 137n12, 151n14, 159
 Index 191
Jha, Ganganatha 151n22 152n34, 152n41, 152n45,
jñāna, see cognition, see also 153n50
knowledge Kuznetsova, Irina 140n18
Junankar, N. S. 140n17
justification (epistemic) 2, 5, 15, L
regress 17–19, 50, objective Lakṣmīnṛsiṃha 153n52
7, 11, 13, 97, 101, 135n16, Laugākṣi 6, 134n7
135n24, 136n27, 167, see also law, natural 65
certification Lehrer, Keith 96, 97, 153n1, 153n2
logic, deductive 53, in India 51–52,
K inductive 66, rule of transposi-
karma (karman) 155n7, 166 tion 145n35, see also Buddhist
Kelly, Thomas 154n7 logic, inference
Kleine, Peter 135n24
knowledge 4–13, 27, 29, 64, 98, 167, M
a priori 64, 102–103, defi nition Mādhava 133n5, 138n28, 141n24,
134n7, foundations 11, Gettier 144n18, 160
cases 12–13, 100–101, 135n24, Mādhyamika 42, 135n22, 160
135n25, 136n27, 154n7, knowl- Mahābhārata 72, 147n57
edge of 12, nirṇaya (reflective manas (mind, sense-mind) 27, 45,
knowledge, knowledge certified) 150n6, 165, 166
5, 12, 106, 135, 166, see also Maṇikaṇṭha Miśra 47, 54, 58, 74,
siddhānta, non-occurrent 9, 144n26, 148n4
pramā (unreflective knowledge) Mathurānātha 26, 27, 28, 136n26,
5, 6, 10, 12, 13, 26–27, 29, 98, 138n24, 149n19, 149n20, 160
99, 167, prāmāṇya 6, 26, 134n8, Matilal, B. K. 57, 72, 137n10, 138n15,
reflective and unreflective 2ff, 139n5, 140n8, 140n19, 143n4,
86, 96–97, 98, social dimension 143n5, 143n10, 144n23,
30, 56, 99, 154n7, see also ana- 144n24, 145n33, 145n34,
logical knowledge, inferential 147n58, 147n62
knowledge, perceptual knowl- māyā (cosmic illusion) 61, 166
edge, testimonial knowledge McCrea, Larry 153n51
knowledge sources (pramāṇa) 3, 5, McDermott, Charlene 143n5, 144n23
9–11, 13, 17, 27, 81, 97, 99, 103, meaning 103, 168, connotation
123, 126, 131, 138n27, as meth- 79, 80, dhvani (suggestion)
ods 15, convergence (pramāṇa- 91, 94–95, 153n58, 159,
samplava) 147–148n64, 164, figurative (lakṣaṇā) 80,
153–154n5, 167, factivity 89, 90 –93, 94, 112, 116,
9–10, identification, see under 123–124, 125, 152n43,
certification, pseudo- 7, 137n13, 156n22, 166, dead metaphor
reducibility 30 76, injunctive theory 150n3,
Krishna, Daya 137n14 learning 75, 76–78, 149n20,
Kṛṣṇakānta Vidyāvāgīśa 106, 155n3, 149n21, 157n26, 157n58,
155n7, 155n11, 157n32, 158n35 reference (abhidhā) 76, 77, 79,
Kumar, Shiv 148n1, 148n7 87, 91, 121, 122, 129–130,
Kumārila 33, 35, 36, 134–135n15, 162, sense 78–9, 91, sentence
139n1, 139n5, 140n6, 76, 87–91, 151–152n28,
140n15,140n18, 142n41, 152n29, see also statements,
144–145n31, 147n64, 149n12, word 76, 78, 87, 152n28, see
151n24, 152n30, 153n51, also yogyatā
155n3, 155n10, 160, 161, see memory (smṛti, remembering,
also Bhāṭṭas smaraṇa) 7–8, 28, 29, 86,
Kunjunni Raja 93, 149n12, 149n13, 103, 165, 169, 171, correct
149n15, 149n19, 152n31, (yathārtha, as something is
192 Index
or was) 5, 6, 28, 171, see also pain 46, 146n54
dispositions, mental pakṣa, see inferential subject
mental dispositions, see dispositions, pakṣa-dharmatā (prover as qualifying
mental an inferential subject) 55, 57,
metaphor, see meaning, figurative 167
metaphysics 3 parāmarśa, see inferential reflection
Meurath, Annette 133n2 parsimony (theoretic simplicity,
Mill, J. S. 79, 80, 134n13, 149n23 lāghavatva, heaviness, gaura-
Mīmāṃsā 3, 4, 10, 18, 45, 60–61, vatva) 9, 31, 32, 165, 166
63, 77, 78, 80, 83, 88, 93, particulars (viśeṣa, particular, indi-
94, 96, 119, 134n8, 146n41, vidualizer, viśeṣya), see qualifi-
147–148n64, 150n3, 152n45, canda, see also sva-lakṣaṇa
153n51, 160, 166, 170 Peirce, C. S. 51
Mīmāṃsā-sūtra 150n3, 159 perception (pratyakṣa, the perceptual
mind, see manas process) 3, 10 14, 20, 29, 33–50,
Mohanty, J. N. 85, 135n18, 150n7, 86, 103, 116, 165, 167, concept-
151n13, 151n15, 151n16 laden (savikalpaka, determinate)
Mokṣākaragupta 146n48 and concept-free (nirvikalpaka,
momentariness 10, 67, 145n32 indeterminate) 25, 33–34, 35–38,
mukti (mokṣa, liberation) 19, 70–71, 50, 139n1, 139n4, 142n42166,
146–147n54, 147n64, 166 169, cross-sensory 36–37,
extraordinary sensory connection
N 46–47, 64, 162, Nyāya’s causal
Nāgārjuna 18, 42, 72, 136n3, 136n4, theory of 44, pseudo- 10, 41–44,
147n61, 160 110, see also illusion, sensory
Nageśa Bhaṭṭa 149n12 connections 45–48, 141n30, 165,
names 87, 88 types of 101, the word ‛pratyakṣa’
Narayanan, T. K. 148n7 9, 33, yogic 47
Nīlakaṇṭha 95 perceptual knowledge (pratyakṣa) 10,
nirṇaya, see under knowledge 13, 15, 27, 29, 33ff, 44, 45, 47,
Nozick, Robert 97, 99, 135n24, 145, 168, see also apperception,
139n31, 153n4, 154n6, 154n7 recognition
number 102 Perrett, Roy 137n14, 154n10
nyāya (critical reasoning, philosophic pervasion (vyāpti) 23, 54, 58, 59, 60,
method) 2, 14, 133n1, 133n5 144n15, antar-vyāpti (internal
Nyāya (the philosophic school) 1ff, 10, pervasion) 63, 145n41, 163, 171
96, 160, 166, and the Veda 83, pervasion, knowledge of 148n4,
93–94, history of the school see generalization, inductive,
1–4, mind-body connection ontology of 60–65, terms of 52,
70, New (navya) 2, 4, 20, 34, 53, 54
43, 54, 64, 75, 84, 87, 92, 106, Phillips, Stephen 138n26, 139n30,
133n3, 134n14, 160, 161, 171, 139n32, 139n35, 140n11,
Old (prācīna) 1, 4, 118–121, 159, 140n18, 141n26, 142n38,
rational theology 4, theism 93–94 145n33, 147n55, 152n28
Nyāya-kośa 134n7, 140n9 philosophy 12, 164, classical Indian 1
Nyāya-sūtra, see under Gautama Pingree, David 154n9
pleasure 70, 71, 114, 146–147n54,
O 155n6, 155n7
Oetke, Claus 51, 52, 142n1, 143n4, Pollock, John 13, 100, 135n16,
144n17 135n24, 136n27, 154n7
ontology 3, 46, 61, 77, 89, 102 Potter, Karl 133n3, 142n35, 143n4,
152n28
P Prabhākara (the individual) 149n12,
padārtha, see category 160, 165
 Index 193
Prābhākaras (followers of Prabhākara) reasoning, critical, see nyāya
30, 36, 43, 80, 84–85, 88, 89, reasoning, suppositional, see tarka
106, 107–111, 135n15, 135n25, reasons 96–97
142n39, 152n29, 154n2, 160, recognition 35, 38–41, 110–111, 113
161, 163, 170, New 114–116, reference (abhidhā), see under
155n5, view of illusion 39, meaning
43–44, 49 Reid, Thomas 87
Pragalbhācārya 106, 155n4 reliabilism 24, 25, 98–99
pragmatics 77 Russell, Bertrand 17
pramā, see under knowledge
pramāṇa, see knowledge source S
Prasad, Rajendra 143n4, 143n8, Saha, Sukharanjan 135n25
143n10 Śālīkanātha Miśra 155n5, 161
Praśastapāda (and PDS) 28, 47, 53, Sāṃkhya 10, 142n34, 161, 167
134n13, 161, PDS 137n14, saṃskāra, see disposition, mental
138n19, 142n37, 145n39, 160 Śaṅkara 141n27, 161
pravṛtti-nimitta (grounds for activ- Sanskrit 2, 87–89
ity of speech) 87, 122–123, Sanskrit literature 2
126–132, 168 Śāntarakṣita 3
Preisendanz, Karin 141–142n33 sapakṣa (where a probandum is known
Price, H. H. 86, 151n18 to occur) 67, 74, 165, 169
pronouns 130–131, 149n22 Sastri, Gaurinath 106
properties 26ff, psychological 6ff, 20, Schayer, Stanislav 57, 143n4, 143n5
67, 68, sensory 35ff, 40–41, self (ātman) 35, 69, 72, 140n15,
universal (all-qualifying) 52, 140n18, 141n30, 164, 165,
62, 166, see also dispositional inference to 65, 66–68, 127,
properties (saṃskāra), guṇa and inference to 147n64, 157n29,
doṣa, qualifiers, universals 169, perception of 147n64
propositions 8 sentence, see under statements, see also
pūrva-pakṣa (prima facie position or under meaning
text devoted thereto) 1, 105– shape (ākṛti) 87, 162
106, 167 Shaw, J. L. 137n14
Putnam, Hilary 136n28 Shcherbatskoi, F. I. 143n8
siddhānta (text answering a previous
Q pūrva-pakṣa) 1, 105–106
qualificanda 38–39, 61ff, 118, 171, the siddhānta (the established view, certi-
thick particular 64 fied knowledge) 5, 11, 19, 22,
qualification (ontological) 143n5 31, 32, 133–134n5, 169
qualifiers (viśeṣaṇa) 38, 39, 41, 44, 49, Siderits, Mark 88 151n25, 151n28,
50, 61ff, 80, 87, 118, 141n32, 152n29
149n25, 150n3, 171, see also similarity (sādṛśya) 3, 75, 79, 107–116,
properties 118–119, 124, 148n2, 152n45,
Quinn, Philip 154n6 importance to inference 148n3,
defi nitions of 108–112, dis-
R similarity 115, 117–118, 121,
Raghunātha 4, 25, 47, 61, 102, knowledge of and inference 74,
138n21, 138n25, 140n9, ontology of 80–81
145n36, 154n10, 161 skepticism 17–18, 32, against inference
Ramanuja Tatacharya, N. S. 136n27, 56
152n32 Socrates 154n7
rasa (aesthetic experience) 95 Solomon, Esther 147n56, 147n57
Ratnakīrti 3, 145n32 Sosa, Ernest 136n6, 138n17, 140n7
ravens paradox 66 154n7
realism 3, 33, 36, 44, 103, 155n8 sound 156n18
194 Index
speaker’s intention (tātparya) 84, Thomism 75
85, 89, 92, 93–95, 124, 125, touch 46
150n11, 156n24, 170 trust 6, 12, 22, 134n11
sphoṭa-vāda (sentential holism) 88, 89, truth, see veridicality
169
Śrīharṣa 31, 56, 135n25, 141n26, U
144n19, 161 Udayana 4, 28, 31, 63, 64, 69, 73, 75,
Staal, J. F. 143n4, 143n5, 143n6, 94, 137n13, 138n26, 138n28,
144n23 139n31, 139n34, 144n29,
statements 12, 22, 55, 82, 83–85, and 147n54, 148n65, 149–150n25,
facts 87–91, analogical 38, 75, 160, 161, Ātma-tattva-viveka
78, 79, 120–121, 122–123, 140n18
124, 127–129, 156n15, 156n21, Uddyotakara 4, 40, 45, 46, 48, 51,
156n23, 157n33, observation 52, 55, 56, 57, 58, 62, 135n20,
18, 34, 61, sentential unity 137n10, 140n117, 142–143n2,
88–89, see also testimony 144n16, 147n64, 152n40, 161
Strawson, P. F. 86, 151n19 universals (jāti, sāmānya, class
subjectivism 36 character, general characteris-
substance (dravya) 46, 61, 165, 167 tic) 25–26, 27, 63, 64, 77, 79,
suffering 70, 71 80, 108, 110, 111, 113, 118,
supervenience 111, 113 144–145n31, 145n39, 148n4,
sva-lakṣaṇa (pure particular, that 154n2, 155n11, 155n8, 157n33,
which is its own mark) 108, 162, 165, 167, 169, universal-
139n1, 170 blocker 28
upādhi (inferential undercutting condi-
T tion) 58–60, 69, 73, dubious
Taber, John 135n20, 150n2 59–60, 144n28, 170
tarka (suppositional reasoning) 11, upādhi (quasi-universal, accidental,
15, 19, 21, 23, 28, 29, 30–32, mind-imposed property) 25–26,
58, 62, 65 70, 73, 97, 100. 101, 28, 63, 102, 170, akhaṇḍa
123, 127, 129, 139n29, 139n30, (unanalyzable accident) 63–64,
148n4, 162, 163, 164, 165, 166, 162, sakhaṇḍa (analyzable acci-
168, 170, anukūla (reasoning dent), 64, 168
favorable to one’s own position) upalakṣaṇa, see indirect indication
139n31, 163 Upanishads 71–72, 147n55, 161, 170
Tarkavagish, Kamakhanath 106
tātparya, see speaker’s intention V
testimonial knowledge (śābda-bodha) Vācaspati Miśra 4, 33, 42, 43, 68, 69,
29, 82, 87, 88, 150n7, 168 70, 73, 83–84, 86, 87, 134n14,
testimony (śabda) 3, 10, 12, 64, 78, 138n15, 139n1, 141n27,
82–95, 99, 103, 116, 149n20, 146n47, 146n49, 146n50,
168, apparent testimony 146n51, 146n52, 147n60,
(śabda-ābhāsa) 10, 85, certifica- 151n16, 161
tion conditions 83–86, 150n7, vāda (discourse) 72, 170
150n9, the expert, see āpta, not Vaiśeṣika 4, 10, 30, 46, 60–61, 77, 85,
inference 85–87, principle of 102, 134n13, 160, 161
charity 82, see also speaker’s variegatedness (citratā) 109
intention, statements Vasubandhu 134n13, 161