Grammar II

JVG/ LV/ UTN 2011

THEORY: LESSON 3
THREE DIFFERENT CONCEPTIONS OF LANGUAGE by Aldo Blanco

What is language? Three conceptions

Nearly every linguist would agree that linguistics is the scientific study of language. But what is
language, and what is science, are controversial matters.

Let us consider three different conceptions of language as an object of study. For the first two,
language is external to the mind; for the third conception, language in internal to the mind.

1) Language as performance: a nominalist view

To very many people, language is a type of human behaviour, or the result of that behaviour:
utterances or marks on paper. In one word, performance, the actual use of language, what we
actually say, or write, at a given moment: an acoustic or graphic object. Linguistic activity or
behaviour, or its products: spoken utterances and written texts. From this point of view, language
involves four uses: speaking, writing, listening, and reading. It involves a medium or substance:
phonic or graphic.

Instances of the products of performance are an utterance, a text, a corpus, a sample – the data of
linguistics. The utterance is a bit of performance. Its nature is concrete, physical, acoustic. It takes
place in time and space. For example: “Come here”, “Come here”, “Come here”, would be three
utterances but one sentence. Three uses of the same sentence, three tokens of the same type.

Utterances are never exactly alike, although the difference may be very slight. If we cannot perceive
any differences, it is because we do not possess a fine enough measuring device. An utterance is a
fleeting event. It is produced and it is gone immediately.

Linguistic performance is like the performance of a symphony. It is also called speech, parole.

What we say may be acceptable or not. It is acceptable if it is understood as evoking a meaning, even
if it is not grammatically correct, e.g. How do you do?, Lo fueron. Idioms are acceptable, but they are
not grammatical (in the technical sense of the word).

Performance involves certain features which are not linguistic:

a) Differences in sex, age, state of mind, state of health, etc.

b) Mistakes, repetitions, pauses, hesitation points, breaks, etc.

Linguistic performance is concerned with psychological or physiological limitations of human

beings: memory limitations, for instance. It is directly observable. It can be recorded on a tape or on
paper.

All the linguistic performance of a child of three or four may be recorded on tape: everything that
the child has said since he was born. A long recording. In this there will be utterances that the child
has never heard, for instance: goed, me ponió.

Performance is extremely complicated. It is the outcome of the interaction of several systems

(cognitive, physiological, physical, etc.). Grammar is only one of these. Performance is “impure” in
the sense that it incorporates elements derived from faculties other than the language faculty, such
as:

Page 1 of 13
 a) the conceptual system
b) pragmatic competence
c) common-sense understanding: knowledge, beliefs, expectations, assumptions,
conventions, evaluation, judgement, about:
i) the nature and behaviour of objects,
ii) their place in a system of “natural kinds”,
iii) the properties that determine the categorisation of objects and the analysis of
events.
d) the place and role of people in a social world
e) the structure of human action
f) will and choice
g) artistic creation and expression
h) play, etc. etc.

Man’s behaviour is determined by the interaction of numerous internal systems operating under
conditions of great variety and complexity. Knowledge of language as a system of practical abilities
to do such-and-such is an ill-defined, obscure object of study. The range of facts potentially relevant
is vast and unknown. It is a mistake to think of knowledge of language as a collection of learned
practices to respond and make utterances used for particular ends.

We had better study the mental structures that enter into the exercise of capacities to do such-and-
such, exploring the properties of these mental structures and separating out the various strands
distinguishing grammar from pragmatic competence, distinguishing the computational structure
from conceptual systems of a different sort. Performance is derived from grammars. Performance is
an epi-phenomenon, a by-product. The use of language is derivative. Grammars have a real existence
in the brain. But there is nothing in the real world corresponding to language. A natural language,
e.g. French, is not a well-defined linguistic concept. It has crucial socio-political dimensions. In
colloquial usage, we say that German is one language and Dutch another, but some dialects of
German are more similar to Dutch dialects than to other, more remote dialects of German.We say
that Chinese is a language with many dialects (Pekingese, Cantonese) and that French, Italian and
Spanish are different languages. But the diversity of the Chinese “dialects” is roughly comparable to
that of the Romance languages. A linguist knowing nothing of political boundaries or institutions
would not distinguish “language” and “dialect” as we do in normal discourse. Nor would he have
clear alternative concepts to propose, with anything like the same function.

Furthermore, even within the more restricted “languages” there may be considerable diversity. Two
dialects of what we call a single language may be mutually incomprehensible. A single individual
will generally command diverse modes of speech, in part associated with varying social conditions
of discourse. No clear principles are known that determine the range and character of possible
variation for a particular individual. Indeed, there is little reason to believe that such principles
exist.

There are two aspects of performance that we must keep apart:

a) performance as processing and production of utterances, and

b) performance as linguistic variety.

Processing and production refer to what we do when we hear and speak a sentence. Studies of
linguistic processing are quite advanced today and they relate to linguistic theory in a very
interesting way. They can be used to support or to reject certain theoretical assumptions.

Page 2 of 13
The study of linguistic variety and context of situation has been going on for about two or three
decades (socio-linguistics, ethnography of speech, dialectology, etc.). Language varies along many
dimensions:

a) spatial or geographical (regional dialects)

b) temporal (Shakespearean English)
c) social (social dialects)
d) situational (registers), etc.

It is necessary to abstract the linguistic components of the utterance in order to get at language
itself. It is necessary to leave out the non-linguistic elements present in the use of language. This
process of abstraction is also called idealisation, and the opposite is called manifestation, or
realisation, or exponence. It is necessary to get at the sentence. The sentence is an abstraction from
utterances, so to say. The sentence above is an abstract object, ‘Come here!’ which can be
manifested in actual speech or writing many times, and each time it is a different acoustic object,
but always an illustration or exponent of the same sentence.

(Horrocks p 8) All grammars abstract away from non-grammatical phenomena. The effects of
drunkenness, nervousness or a bad cold on the speech of native speakers. Also abstract away from
variety (the very real heterogeneity of actual speech communities).

2) Language as an infinite set of sentences: a realist or Platonist view

Abstracting away from the use of language, we could conceive of it as being made up of an infinite
set of sentences: potential, possible sentences, all of them.

In this view, the sentence is a unit of language: an abstract object. The sentence as such is not
located anywhere: it is a hypothetical construct. A tool of the linguist. A concept. A scientific
construct, like the phoneme, morpheme, word, noun, etc. or like the concepts atom, chair, etc. or in
fact, number, set, proposition.

A sentence is like a symphony. Linguistic performance, like the performance of a symphony. A

sentence is like message which may be transmitted many times. A sentence is a type, of which the
various instances are the tokens. A sentence is form. The utterance is substance.

Saussure defined language (la langue) as a system of signs. Hjelmslev defined language as a system
of relations.

The sentence has three levels of structure: semantic, syntactic, phonological, because we need three
kinds of rules to generate them.

The sentence is constructed in the mind by means of the rules of the grammar, and then manifested
or spoken. But sentences do not exist in the mind in the form of a list. When I speak, when I make an
utterance, I do not choose a sentence from a list, or stock, in the mind. I have to build it up. That is
why the sentence is abstract, and not mental. The rules are mental or psychological.

The relation between the sentence and the utterance (likewise between language and performance)
is one of abstraction or idealisation in one direction, and manifestation, realisation or exponence in
the other direction.

The sentence is abstracted from the utterance by leaving out all the non-linguistic features or
elements in it. We have to impose some structure on the utterance at the time we analyse. It is a
process of scientific research: imposing structure and discovering structure at the same time. We
have to generalise from the use of language, from the utterances.

Page 3 of 13
This is the realist or Platonic view. Language exists in “the world of ideas”. It is like a mathematical
object. This is Katz’ view.

The problem is the impossibility to unify linguistics with psychology. If language is outside the
mind, it doesn’t seem to be part of the mental equipment of man. If so, it couldn’t be a psychological
(hence biological) capacity. Obviously it is.

3) Language as competence: a conceptualist view

The third conception is that language resides in the mind. In other words, it is located in the brain
but we don’t know how it is realised there. What we do know, or at least can safely hypothesise, is
that there are no sentences in the mind. Language does not exist as a list of sentences. Sentences
have to be constructed by the mind by means of principles which are applied to words. Language is
thus a set of principles, or very general rules, which make up the sentences and other expressions.
Language is competence, the capacity possessed by human beings to build up expressions.
Language is a type of knowledge, a mental object. This object cannot of course be seen or observed.
It has to be studied in the way we study other objects which are unobservable, through an
examination of its consequences: the utterances produced by people. Language is a cognitive
structure, a part of the brain, a part of the body. Linguistics is a branch of biology. This position is
called conceptualism, the cognitive approach.

Language is in the brain and it can’t get out of there. There is no language outside the mind.
Whatever is printed or spoken is a manifestation of language, but not language itself. If an utterance
is printed on paper, it can be burnt. The only way to burn language is to set light to a human body. If
all the books were burnt, language would still exist in the mind of man.

Furthermore, language is a possession of the individual. It is in the brain of every individual that is
alive. There is no language outside the mind. There are no linguistic communities. In fact, there are
no languages as such, nor dialects –only idiolects, as many languages as there are people. What we
usually call a language is an abstraction from the similarities of the languages of individuals. My
language and your language are similar enough to make that abstraction. That is why we can
communicate: through the similarities in our languages, i.e. idiolects.

Language is a kind of knowledge: linguistic competence, a product of human intelligence, a system

of knowledge. This knowledge is also called “grammar” and it is represented in the brain-mind
somehow.

Chomsky used to think that these cognitive structures consisted of a finite set of rules (that
generated the sentences of a language). Now he thinks that these rules can be expressed in a much
more general way; they can be formulated as principles. These principles have a mental or
psychological nature: they are in the mind. These principles represent the knowledge of the native
speaker of his language: the cognitive state or structure attained. They constitute the creative aspect
of language use.

Language is like a code, in terms of which we can construct messages (the sentences of the code).
When we encode a message and transmit it, we have an utterance, which is then received and
decoded (or processed) back into grammar.

The encoding of a message (the construction of a sentence) is called the process of generation or
synthesis, and the decoding of the message is called analysis. These notions are the relationship
between grammar and language.

When we go from the sentence to the grammar, we analyse the sentence. Analysing a sentence
means describing its structure, providing the rules (or principles) which are behind its construction.

Page 4 of 13
Generating a sentence means going from the rules of the grammar towards the language: using the
rules in order to build it up.

A grammar is a theory about a particular language: its description.

It is necessary to distinguish two related concepts: the intuitive grammar and the scientific
grammar. The latter is the study of the former. Intuitive grammar is the speaker’s knowledge: his
competence, his language. Scientific grammar is the description of that knowledge made by the
linguist.

Intuitive grammar, language, is not observable. It has to be studied through hypotheses or

assumptions that the linguist sets up which are meant to account for the speaker’s ability. The
hypotheses are the scientific grammar.

Whenever an object of study is not observable, we have to examine the results of it: the observable
manifestations of this object, in the case of language, the utterances spoken by people. From them
we generalise and set up sentences, which are then studied in order to construct the grammar for
that language.

The rules and principles of a grammar can generate sentences which speakers wouldn’t use, or can’t
use, because of non-linguistic limitations that they have, such as memory limitations. An example of
sentences which are grammatical in that they are generated by the rules, but which are not
manifested are:

The horse – the man – my sister … married – bought – died.

The butcher’s daughter’s boyfriend’s brother (regressive structure)

El hermano del novio de la hija del carnicero de la esquina (progressive structure)

Chomsky, 1981, p 3: The theory of Universal Grammar must meet two obvious conditions. On the
one hand it must be compatible with the diversity of existing (indeed, possible) grammars. At the
same time, UG must be sufficiently constrained and restrictive in the options it permits so as to
account for the fact that each of these grammars develops in the mind on the basis of quite limited
evidence.

Linguistic theory requires restrictiveness (i.e. selection from the narrowest set of possible
grammars). A maximally restrictive theory best fits the facts of language acquisition: speed,
uniformity, little stimulus, no intelligence. We ought to severely restrict the class of grammars from
which the child, selects his own. The most restrictive theory is best. A theory that allows anything
(everything) explains nothing. A theory of grammar has to be restrictive, constrained. (p 157).

Summing up, Nominalism holds that language is just a name for something that does not really
exist. Realism maintains that language is an abstract object. Conceptualism claims that language is
a conceptual or cognitive structure in the brain, a part of the human body, very much like our leg or
nose.

Three conceptions of the nature of language

(3) (2) (1)

Page 5 of 13
Competence Sentences Performance

A particular grammar An infinite set of Acoustic object

Rules and principles sentences
Knowledge of language Phonic (or graphic)
An abstract object substance or medium
Competence
No physical Text, data, corpus
Mental object realisation ABSTRACTIO
N or ‘Come here, come here,
In the mind/brain ANALYSIS No location IDEALIZATIO come here.’
N
A code ------------- Three utterances, one
A particular grammar ----- --------------- sentence
------
Internal to he mind Non-linguistic features:
sex, age, mood, state of
Not observable health, etc.
Cognitive structure Mistakes, repetitions.

Use, behaviour, activity.

SYNTHESIS Speech
or MANIFESTAT
ION or
GENERATIO REALIZATION
N
------------------
----------------
---
--

The faculty of language and Universal Grammar.

The faculty of language is our linguistic capacity at birth, our uniquely human inheritance. It is
present in the mind of every child and it is going to develop in one way or another depending on
external factors: linguistic exposure. If you talk English to the baby, he is going to develop
something very much like English, but always with some differences from the input. If you talk
Spanish to him, he is going to develop something very much like Spanish, but with some differences.
Depending on the input, the faculty of language will unfold along a certain path. This development
or growth of a particular language is technically called parameter setting. Universal grammar is
structured in such a way that given a certain input, it develops along a certain path. Given a different
input, it develops along a different path.

Universal grammar is the innate language faculty, represented in some still unknown way in the
brain. It is a species property, a capacity specific to the human species. It is the truly distinctive and
most remarkable characteristic of the species. It is made up of principles and parameters and when
the parameters have been set, it develops into a ‘particular’ language. Chomsky calls this cognitive
state the initial stage or stage zero.

Once the parameters have been set, and a lexicon has been acquired (though much of it is also
universal), the child reaches the so-called steady stage. He now knows most of the syntax of his
language, while the vocabulary continues developing throughout his life in a very different
(cumulative) way. The development of the syntax is thus explosive. It is over when the values of the
parameters have been selected, at a very early age indeed: four or five years.

Page 6 of 13
Examples of parameters are:

1) The subject may be obligatory or optional. It is obligatory in English and optional in Spanish.
2) The position of the subject may be before the verb or after the verb.
3) The presence or absence of clitic pronouns, such as le, lo, se in Spanish.
4) The subject of the infinitive may occur in English in certain structures (I want John to come) and
in Spanish in other kinds of structures (Al salir el sol).

(5) (4) (3)

Linguistic experience The faculty A particular grammar
of language
Primary data Innate language The parameters have
Input faculty been set
What the child hears A species property A lexicon has been
-------------- Natural endowment -------------- acquired (learnt)
-- Principles and -- The values of the
parameters parameters have been
UG selected

5) Linguistic experience
The performance of others. A normal child acquires knowledge of a language on relatively slight
exposure and without specific training. This experience is extremely fragmentary and
impoverished; scattered and restricted evidence. And yet we are able to learn so much.

Growth of language from the initial state (Si) to the steady state (Ss) through
interaction with linguistic experience (= exposure)

Si ------------------------------------------- Ss

How are languages “learnt”? Language development (learning, acquisition, growth, maturation)
Languages are not learnt nor are they acquired. You cannot learn or acquire something that you
have already. Language is a cognitive capacity that we inherit from our parents. It is present in
every human being at birth and it grows or develops according to a genetic programme, the same
programme that guides the growth of the body. The best word to describe this process is
maturation. Language matures like the rest of the organism. We may continue talking about
language learning, or acquisition, but now we are using these words with new senses. Language
development is part of our human endowment. It comes in the genes together with much else.

The problem of acquisition: poverty of the stimulus (evidence)

The problem of under-determination of the theory by the data
Every single human being develops knowledge of a particular language, English, French, Swahili,
with the only exception of pathological cases. This is strong evidence that the innate faculty of
language is a rich system, tightly structured, which with only a minimum of exposure will develop
into a particular language. It is also strong evidence that exposure determines which particular
language the innate state will develop into. The process is uniform across the species and virtually
without exceptions.

Furthermore, children of 3 or 4 years often say words that they have not heard before (goed, ponió)
and our normal use of language is creative, that is most of the sentences we speak and hear are
created anew. They are not quotations or repetitions of things we have heard before. We do not talk
by heart.

Page 7 of 13
How much exposure is necessary for the initial state to become fixed, a native language? Apparently
very little, just enough for the child to set the parameters of his innate capacity. That is why
language development is so uniform and so swift. Because languages are not really learnt or
acquired from outside. Linguistic experience is just a trigger which sets the innate mechanism of
linguistic growth into action, allowing the innate faculty to make the choices it must on the
parameters it consists of according to the principles which it is made up of.

The stimulus, the evidence, the primary linguistic data, is said to be poor in relation to the richly
structured initial state. The theory of grammar is badly under-determined by the data, that is why it
has taken so long to hypothesise its principles.

Similarities and differences among languages

There are several reasons why language differences are so striking and so it is difficult to believe
that there is really only one human language. The following may be valid reasons:

1) In every day life we are more interested in differences than in similarities. John falls in love with
Mary because she has blue eyes. Scientists however abstract from irrelevant differences in their
attempt to formulate general laws or principles.
2) We have to wrestle with differences when we wish to speak ‘another language’. We even believe
that the essence of language is the word and its meaning. It is not. The essence of language is the
syntax, the computational system which allows me to utter long sentences with sometimes quite
involves structures and my listener follows me seemingly without much effort. But a speaker of
another language cannot.

If we regarded similarities among languages with the same interest and concern (as generative
grammarians do) with which we regarded differences, the picture of linguistic variety that would
emerge would be quite different from the one presented by socio-linguists.

What is innate and what is learnt?

The performance systems do not seem to vary from individual to individual. We all seem to use the
mechanism of perception and articulation, interpretation and expression in very much the same
way.

Differences among languages seem to lie in the language system, more precisely:

a) in the lexicon, the association between sounds and concepts is arbitrary,

b) in the inflexional and derivational morphology (not nouns and verbs),
c) in the sound system.
Variation lies then in a restricted area of the lexicon and in the phonological component. This
leads Chomsky (1995: 170) to claim that ‘there is only one computational system and one
lexicon, apart from this limited kind of variety.”

The reason why most variation is in the phonological component is that variation is found in the
primary linguistic data that the child hears (Lightfoot 1991).

Language: its knowledge, origin and use

There are three things that we want to study about language:

a) its knowledge, i.e. knowledge of a particular language (its structure), (specify the
parametric options)
b) its origin or development, i.e. our acquisition of a particular language (the result of the
interaction of UG and linguistic experience),
c) its use, i.e. processes of understanding and production, (pragmatics, discourse).

Page 8 of 13
The use of language is a human right
Every human being possesses his own language and is free to make use of it as he pleases. This idea
offers a new focus on the descriptive versus prescriptive dichotomy. I say what I like, in the way I
like. I choose my lexical items and the grammatical construction that is going to put them together.
And nobody can take away that freedom from me. The right to self-expression is like the right to
walk the way I like, or to think as I do. I am writing this on May 3, the International Day of Freedom
of the Press. Today, the newspapers are full of articles on this human right. It is the same as the
freedom to use language the way we like. It is also called Freedom of Speech. It is a human right and
an individual freedom. We are free to express the ideas we believe in and also to use the language of
our choice.

2
What is science?
Science is the study of objects and phenomena, or aspects, of the world, that seeks theoretical
understanding expressed in the form of general principles. These principles explain the causes of
the phenomena under study.

This intellectual pursuit has been very successful in the natural science (physics, chemistry, biology,
astronomy, the earth sciences, etc.) because nature lends itself to be simplified and experimented
on as autonomous systems. Social and cultural studies are far too complex to be submitted to
rigorous analysis and explanatory principles. They often deal with problems which results from the
interaction of many systems which are very hard to disentangle.

Generative grammar, the study of language as a natural object from an internalist perspective,
claims to be part of natural science. It enquires into the structure of language and seeks to formulate
the principles that underlie the generation of sentences and their structural descriptions.

The study of the language faculty is often called Universal Grammar, and the study of a particular
language is called a particular grammar.

The formulation of principles was beyond the description of linguistic forms and the classification of
their constituent elements. Traditional grammar and structural grammar both had as their goals the
presentation of the forms of language, especially the taxonomy of irregularities. They did not
inquire into why the structures differed from one another. Traditional grammar did not change
much since it was originated over 20 centuries ago. It did not change precisely because it restricted
its descriptions to observable features. It was a structured version of primary linguistic data. It did
not develop any theories about the forms that language may assume in the mind of its speakers.
Little syntax.

Two kinds of science: natural and socio-cultural

There are two views with respect to the nature of science: one view imposing a weaker claim on it,
the other restricting a scientific study to very stringent conditions.

The weaker view is that science is any serious, objective and systematic study of a given subject.
This view is very comprehensive and it is prepared to consider very many disciplines as scientific:
the study of translation, pedagogy, socio-linguistics, the study of behaviour (human as well as
animal), economics, etc.

The stronger view regards a scientific study as a theoretical enquiry into certain aspects of nature. It
restricts the term science to chemistry, physics, biology, astronomy, geology and a few other fields.
Generative grammar (theoretical linguistics) is a scientific study in spite of the fact that it cannot be
unified with the basic, central and successful natural sciences yet, perhaps never.

Page 9 of 13
This classification of ‘science’ into soft and hard is not merely a terminological question. Chomsky
uses the term ‘naturalistic enquiry’ to refer to the study of language as a part of the natural world.

Language is part of the mental phenomena, on a par with chemical, electrical or optical phenomena.
The study of language is very much like the study of complex molecules, electric fields or human
vision. Science develops concepts and postulates the existence of certain objects in the natural
world. These objects may not be material in the sense of ‘physical contact’. Magnetism, fields of
energy and other concepts of contemporary physics do not seem to refer to any proposed notion of
matter. Science builds intelligible explanatory theories and tries to unify them.

Two kinds of adequacy: conceptual and empirical

The two driving forces that bring about progress in science are conceptual and empirical adequacy.
The latter is attained when more data are covered by the theory. A grammar gets better when it
describes a larger and larger number of structures. The former adequacy is reached when the
principles formulated are general in scope and explanatory in quality.

Generative grammar has been in constant search of general principles and empirical coverage. What
was known about the faculty of language in 1957 was not much more than the linguistic universals
hypothesised then (substantial such as noun and verb, and formal such as phrase-structure and
transformational rules). Universal grammar has expanded dramatically in the forty years that
ensued since its inception. In fact the recent changes in the theory were the result of realising that
Universal Grammar had become too rich and should be minimised.

Two shifts in focus

Generative grammar has produced several models in its short history of slightly over forty years,
but there have been two changes which may be considered as the most important ones because of
the great advances in theory that they brought about. Chomsky often calls them ‘two shifts in focus’.
One is the change in the subject-matter of the enquiry. It used to be the analysis of sentences. Since
its inception generative grammar set up a different object of study: I-language, that is the system
that generates sentences. The difference between analysis and synthesis (or generation) is that the
description now has to meet a stringent requirement of explicitness (formality). There cannot be
vague descriptions or inconsistent analyses.

The other change came about by 1980 when the current model was devised, the theory of Principles
and Parameters. Gradually though, grammatical studies moved from the search for rules and the
description of particular constructions to the search for filters, conditions, principles. The end-
result is a minimalist theory with very few and very general principles. This trend gives generative
grammar a deductive character. The aim is that English and Swahili should be deduced from
particular ways of setting the parameters. The shift from rules to principles is a move from
description to explanation.

A model as a deductive system

A model or theory consists of primitive terms and axioms. On the basis of primitive terms we define
derived terms and on the basis of axioms we define theorems.

The primitive terms of a linguistic theory are the lexical features (phonological,
structural/formal/grammatical/morphological and semantic) and the values on parameters. On the
basis of these, we define grammatical categories (lexical and functional).

The axioms are the general principles that constrain derivations, e.g. economy, last resort, etc. As
research progresses, we should be able to deduce particular languages from the setting (or fixing) of
parameters of variation.

Page 10 of 13
Leading ideas and modes of execution
As a linguist works on the description of a language and on the explanatory principles underlying
the description, he follows leading ideas concerning the structure of language and tries various
modes of execution (hypotheses on the application of leading ideas).

Abstraction or idealisation of the data

As an I-language is a property of each individual speaker, there will always be differences between
I-languages. No matter how striking the similarities may be, there will always be some differences.
Language varies from speaker to speaker. But for purposes of explanation, the linguist abstracts
from these differences. The highest abstraction is achieved in the study of universal grammar.

It is always possible to describe differences, which is what students of variation do. The interesting
thing however is to relate them to parameters of variation.

The leading research strategy in linguistics has been to abstract general principles restricting
derivations from the apparent variation observed on the surface of linguistic expressions.

Idealisation to instant learning

It takes a child around five or six years to develop an I-language. This process follows an inner-
directed program of growth, just like the rest of the body. Again there may be differences among
children during the course of language development but for theoretical purposes, the linguist
abstracts from the stages of language growth and idealises the process to instant learning. At this
stage of research, he wants to find out what parameters are set and how.

Language is a nearly perfect system

By perfect is meant as simple as possible, minimal, economical, not redundant, symmetrical, natural,
with as few elements in the representations and as few steps in the derivations as possible. The
concepts and structure of the theory should be minimal and necessary. The theory should keep to
bare essentials, excluding what is superfluous. These features account for ease of acquisition. E.g.
structural case should be assigned in a single way.

The hypothesis is that language was perfect when it first appeared in the human brain, but it had to
adapt itself to the requirements of the external and extraneous sound and concept systems. The
process of adaptation brought about changes in language design which now make it less perfect
than it used to be. So language exhibits imperfections, deviations, departures from perfection. But
the changes have been minimal.

Other systems that process information do not have these properties: logic and computer
languages. What changes did language have to undergo to adapt itself to the external performance
systems? What are the departures from perfection?

a) linearity (a restriction on structure), imposed by interacting systems, e.g. the sound system
b) movement, imposed by the need to check features for convergence.

3
The components of language
The cognitive system and the performance systems
The lexicon and the computational system
The five operations: selection, merge, move, deletion and spell-out

The components of language

In his latest publications, Chomsky (1995a: 12, 1995b: 12, 1993a: 47) claims that ‘there is evidence
that the language faculty has at least two different components: a “cognitive system” that stores
information in some manner, and performance systems that make use of this information’. The

Page 11 of 13
performance systems are just two: an input receptive system and an output production system. But
there is also a common body of information that these two systems have access to. That is the
cognitive system mentioned in the first place, which is thus separate from the performance systems.
Then Chomsky (1995: 13) adds that ‘the performance systems may well be specialised for
language.’
Through linguistic experience the faculty of language that every normal child possesses at birth
develops into a particular language, a stat of the cognitive system of the innate language faculty, an
I-language, where I stands for internal, individual, and intensional.

This I-language, the cognitive system, consists of a lexicon and a computational system. The lexicon
is a list of items, each one a set of three kinds of features or properties: phonetic, grammatical and
semantic. The computational system consists of a number of five operations: selection, merge,
move, delete and spell-out. It is a generative procedure that determines an infinite class of linguistic
expressions, each a collection of instructions for the performance systems. The procedure
constructs pairs (P, L) of representations that are interpreted at the two interfaces: A-P and C-I.

A linguistic expression is at least a pair (P, L) meeting the conditions of FI and under minimalist
assumptions, at most, such a pair (meaning no other levels).

The computational system may be invariant, and so do the performance systems seem to be.
Variation seems to lie only in the properties in the formal features which have no interpretation at
the interface. Or even restricted to the properties of morphemes, a subpart of the lexicon. Then
there is only one human language.

‘Learning’ a particular language means fixing the lexical options on the basis of primary linguistic
data. The linguist now has to deduce particular languages and dialects by fixing the options within
the finite lexical variety allowed.

An overriding principle is economy. It operates at several places in the theory. Economy in

representations means few symbols. Economy in derivations means few and short steps. Economy
in the general theory means few levels in the architecture of the theory.

Noam Chomsky:

(1995a) Language and Nature. Mind 104, 1-61. (1. Language as a natural object. 2. Language from

an internalist perspective.)

(1995b) The Minimalist Program. MIT. Pp. 420. It includes the following chapters:

1) The Theory of Principles and Parameters. With Howard Lasnik. Written in 1991. Pp. 13-
127.
2) Some Notes on Economy of Derivation and Representation. Written in 1988. Pp. 129-
166.
3) A Minimalist Program for Linguistic Theory. Written in 1992. Pp. 167-217.
4) Categories and Transformations. Written in 1994. Pp. 219-394.

(1994a) Bare Phrase Structure. In G. Webelhuth (ed.) Government and Binding Theory and

the Minimalist Program. Blackwell. Pp. 383-439

(1994c) Naturalism and dualism in the study of language and mind. International Journal

of Philosophical Studies 2, 181-209.

Page 12 of 13
(1993a) Language and Thought. Wakefield, IR & London: Moyer Bell. Pp. 96.

(1993b) A minimalist programme for linguistic theory. In Hale & Keyser (1993) The View

from Building 20. MIT. Pp. 1-52

(1992a) Explaining Language Use. Philosophical Topics 20, 205-231.

(1992b) Language and Interpretation: Philosophical Reflections and Empirical Enquiry. In

J. Earman (ed.) Inference, explanation and other philosophical frustrations.

Berkeley: University of California Press.

(1991a) Linguistics and Adjacent Fields: A Personal View. (Prospects for the study of

language and mind) Pp 3-25. In A. Kasher (ed.) The Chomskyan Turn. Blackwell.

Pp. 3-25.

(1991b) Linguistics and cognitive science: problems and mysteries. In A. Kasher (1991)

Pp. 26-53.

(1990) On formalisation and formal linguistics. Natural Language and Linguistic Theory 8:

143-147.

Page 13 of 13