Identity Theories of Truth and the Tractatus

Peter M. Sullivan

Abstract
The paper is concerned with the idea that the world is the totality
of facts, not of things – with what is involved in thinking of the
world in that way, and why one might do so. It approaches this issue
through a comparison between Wittgenstein’s Tractatus and the iden-
tity theory of truth proposed by Hornsby and McDowell.The paper’s
positive conclusion is that there is a genuine affinity between these
two. A negative contention is that the modern identity theory is vul-
nerable to a complaint of idealism that the Tractatus can deflect.
Philosophical Investigations 28:1 January 2005
ISSN 0190-0536

Identity Theories of Truth and the Tractatus

Peter M. Sullivan, University of Stirling

§1 Introduction

This is, as they say, part of a research project. It is one of two papers
on the opening statement of Wittgenstein’s Tractatus, that the world
is the totality of facts. The first paper (Sullivan 2000) concentrated
on the idea of totality and asked whether there can be one totality of
facts. This paper addresses the idea that the world is a totality of facts
rather than of things: its concern is with what it is to conceive of
the world in that way, and why one might do so. (In later stages of
the project I hope to reach the third sentence of the book.)
For a long time the conception of the world as the totality of facts
was associated with the correspondence theory of truth. Armstrong
(1973), for instance, took himself to be in agreement with Wittgen-
stein in portraying the world as a totality of correspondents or truth-
makers for beliefs.The association was upheld as much by opponents
of correspondence as by its advocates. For instance, Strawson chose to
express his opposition to any form of correspondence theory by
reversing Wittgenstein’s slogan:“The world”, he insisted,“is the total-
ity of things, not of facts” (Strawson 1950: 198 n.).
But things changed.The opening words of the Tractatus came to be
used almost as a motto of an influential line of thought that is funda-
mentally opposed to correspondence conceptions. One leading expres-
sion of it is the following passage from McDowell’s Mind and World.
. . . there is no ontological gap between the sort of thing one can
mean, or generally the sort of thing one can think, and the sort of
thing that can be the case. So since the world is everything that is
the case . . . , there is no gap between thought, as such, and the
world.
But to say that there is no gap between thought, as such, and
the world, is just to dress up a truism in high-flown language. All
the point comes to is that one can think, for instance, that spring
has begun, and that very same thing, that spring has begun, can be
the case. (McDowell 1994: 27)
© Blackwell Publishing Ltd. 2005, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA
02148, USA.
44 Philosophical Investigations
McDowell concludes: “the world is made up of the sort of thing
one can think” (1994: 27–8); and the sort of thing one thinks, if one
thinks truly, is a fact. So McDowell’s thought seems to be the same
as, or at any rate very much like, the one that Jennifer Hornsby pre-
sents in “the simple statement that true thinkables [or propositions]
are the same as facts”, a statement which encapsulates what she calls
“the identity theory of truth”. That theory, she says, “is worth con-
sidering to the extent that correspondence theories are worth avoid-
ing” (Hornsby 1997: 1–2).
Of the many questions raised by this curious turn of events I want
to discuss just one: namely, whether McDowell and Hornsby have any
right to claim the Wittgensteinian formulation on behalf of their
identity theory. Is there any genuine affinity between the conception
of truth they advance and Wittgenstein’s thought in the Tractatus?
My answer will be a qualified ‘Yes’. The ‘Yes’ part of the answer
is the most important, and concerns motivations. The qualification
concerns an aspect of the Tractatus that the modern identity theo-
rists want nothing to do with, namely, its notion of analysis. To sep-
arate the positive answer from the qualification I need to separate
two ways in which thinking of the world as the totality of facts
might seem to involve idealism.
While facts, as both Wittgenstein and the modern identity theo-
rists think of them, are certainly not made by mind, there can seem
point in saying that they are made for mind: facts are structured
exactly so as to be thinkable, and in that way conform to the form
of thought. So conceiving the world as the totality of facts might
seem to involve, as the only possible explanation of that conformity,
a Kantian variety of idealism (cf. Moore 1997: 118).The positive part
of my answer is that both Wittgenstein and the identity theorists
recognize that it does not. More than that, they have in common
that they advance that conception precisely to avoid the genuinely
idealist consequences of the superficially more realist-sounding
alternative. In another shared slogan, their aim is to allow that
thought “reaches right up to [reality]” (TLP 2.1511; cf. PI 95), rather
than stopping short at something less, reality-as-we-conceive-it.
The qualification concerns a charge of idealism more like
Johnson’s stone-kicking challenge to Berkeley. A totality of facts
strikes many people’s intuitions as not thick or solid or brute enough
to make a world. There is, I think, something right in that reaction.
At the end I’ll suggest that, whereas the Tractatus can accommodate
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 45
or defuse what is right in it through the project of analysis, the reac-
tion stands as an objection to the modern identity theory.
That, then, is the destination. I should give a sketch of the route
I’ll take to it. My starting point in §2 will be a brief review of Straw-
son’s famous paper “Truth”, in which he aligned the conception of
the world-as-facts with the correspondence theory, and rejected
them both. I’ll claim that his arguments do not make the alignment
essential. Instead, his central arguments are best seen as setting an
agenda for an acceptable version of the world-as-fact conception. I’ll
then turn in §3 to Hornsby’s advocacy of the identity theory, rec-
ommending that we see it as pursuing that agenda. §4 will give
reasons for my positive conclusion, that the same agenda and the
same motivations are at work in the Tractatus. The last section (§5)
will then explain the qualification, that what Hornsby and McDow-
ell share with the Tractatus is not so easily separated from other, less
attractive parts of the scheme.

§2 Strawson against the Correspondence Theory

“If ”, Strawson says, “we read ‘world’ (a sadly corrupted word) as

‘heavens and earth’, talk of facts . . . as ‘included in’ or ‘parts of ’ the
world is, obviously, metaphorical” (1950: 198 n.). With ‘world’ so
read, that is clearly right. ‘Heavens and earth’ denotes a vast spatio-
temporal complex, whose parts are located and dated, and hence not
facts. But why should we read ‘world’ in that way?
One obvious reason not to is that not everything we speak or
think about has a place in heavens and earth. (Numbers don’t, for
instance.) An independent and more relevant reason is given by
Russell’s well-known observation, that a catalogue merely of things
– of those things that are, uncontroversially or non-metaphorically,
included in heavens and earth – is not enough to fill one centrally
important role that philosophy has assigned to the notion of world:
the role of what thought answers to, or what thoughts are measured
against to be assessed as true or false. “The world”, Russell remarked,
“is not described by merely naming all the objects in it” (1922: 12).
If ‘heavens and earth’ is not up to that task, then we should under-
stand the notion of world less restrictively.
Strawson had powerful reasons for not locating facts “in the
world” (1950: 195), even in the face of Russell’s observation. But I
© Blackwell Publishing Ltd. 2005
46 Philosophical Investigations
think those reasons fall short of compelling his understanding of
‘world’. Instead, the problems and confusions he points to can be
taken as markers for how a conception of world-as-fact ought to be
developed.
Strawson’s first and most fundamental objection to counting facts
parts of the world is that it fosters an illusion of explanation. The
illusion is a matter of imagining that we have bridged a divide
between thought and reality when all we have actually done is to
shuffle around amongst truisms that remain squarely on the nearer
side of it.This illusion comes about by fudging between two incom-
patible pulls on the notion of a fact: on the one hand, statements
and facts are supposed to have a peculiarly intimate connection with
each other, facts being exactly what it takes to make statements true;
on the other, facts are supposed distanced from statements, as belong-
ing to the world we describe rather than its description. To counter
the illusion Strawson insists that to talk of a state of affairs’ being
represented is just to allude in other words to a statement’s being
made; likewise, for a fact to be expressed just is for a statement to
be true. ‘Statement’, ‘fact’, and ‘truth’ are internally related terms
which collectively signal the occurrence of informative or ‘fact-
stating’ discourse. They belong to that discourse. They do not
comment on it, and they cannot explain it. To locate facts in the
world (or to imagine them located there) encourages us to mistake
truisms recording the internal order of that discourse – such truisms
as that a statement is true if what it expresses is a fact – for a sub-
stantial account of the relation between that discourse and some-
thing external to it that gives it its point. A correspondence theory
of truth, Strawson claims, inevitably but hopelessly aspires to be just
such an account (1950: 200–201).
Strawson’s second complaint is still broader. Facts being “made
for” statements (1950: 197), to count them constituents of reality is
to “model the world on the word” (1950: 190). The complaint was
already a standard one when Strawson wrote, and Austin, his imme-
diate target, had tried to avoid it (in part1) through a similarly stan-
dard idea, that “statements fit the facts always more or less loosely,

1. This is no doubt unfair to Austin. What was standard was the attempt to avoid a
threatened “collapse” (O’Connor 1975: 48) or “coallescence” (Armstrong 1973: 113)
of proposition and fact by the notion of a loose-fitting garb. What is original in
Austin is the idea that the fit is sensitive to context and purpose; for development
of this theme in Austin, see Travis (2000), especially chs. 6 & 10.

© Blackwell Publishing Ltd. 2005

 Peter M. Sullivan 47
in different ways on different occasions for different intents and pur-
poses” (Austin 1950: 130). Strawson’s simple reply aims to close off
that defence. He asks: “What could fit more perfectly the fact that
it is raining than the statement that it is raining?” (Strawson 1950:
197).
Suppose, as I think, that there is something right in both points.
Still, Strawson gives no hint as to how Russell’s observation might
be respected along with them. Without that, there is an obvious
danger that the pseudo-explanatory theorizing he condemns will not
be halted, but instead will just be redirected to questions of how the
internally related pair of statement and fact connect with the world
of objects. The danger is apparent in one of Strawson’s most com-
pelling formulations of his first complaint.
And it is, indeed, very strange that people have so often proceeded
by saying ‘Well, we’re pretty clear what a statement is, aren’t we?
Now let us settle the further question, viz. What it is for a statement
to be true’. (1950: 200)

Compare that with the following: ‘Well, we know what it is to

describe an object, don’t we? Now let us address the further ques-
tion, what it is for a description to fit the object it describes’. The
second is surely ‘strange’ in just the same way as the first. But by
placing the division of internal from external relations where he does
– statement and fact internally related on one side, these two exter-
nally related to objects on the other – Strawson invites the second
strange question in the very same breath as he closes off the first.
That is an indication that Strawson’s fundamental point will not
reach its proper conclusion until it is reconciled with Russell’s
observation.
Suppose, then, that we take Russell’s observation as given: a cat-
alogue of facts, and not merely of things, is needed to distinguish
this world from any other, or how things are from how they might
have been; in whatever sense that implies, facts belong to the world.
Strawson’s first complaint tells directly not against that placing, but
against any imagined separation of the internally related statement-
fact pairing. So the point can be adapted to Russell’s observation as
the insistence that truths, or true propositions, belong to the world
along with facts.
That shift inevitably opens up another distance for theorizing to
fill, between ‘discourse’, or language use, and the proposition-fact
© Blackwell Publishing Ltd. 2005
48 Philosophical Investigations
pair. But that is as it should be. The empirical sciences of language
shouldn’t be left with nothing to do. The relevant point is just that
the kind of pseudo-theorizing Strawson condemns patently has
nothing to contribute to bridging the new gap. (Theorizing that
helped itself to any of his internally related notions – as our
ordinary, non-theoretical characterizations of language use always
do – would be already on the proposition-fact side of the new
distance.)
Strawson condemned the correspondence theory as offering a
pseudo-explanation of fact-stating discourse, but he held out the pos-
sibility of genuine explanation. Anyone attempting such an explana-
tion would, he said, have to “stand back from language and talk about
the different ways in which utterances are related to the world
(though he must get beyond ‘correspondence of statement and fact’
if his talk is to be fruitful)” (1950: 201). The kind of ‘standing back’
he had in mind was, I believe, a Kantian transcendental explanation:
nothing less offers the chance of a non-circular grounding of the
fundamental categories to be explained. But however things stand
with that speculation, it is plain that the shift recommended here
closes the opening Strawson left. His strictures on genuine explana-
tion, translated into the new setting, would require it to ‘stand back
from propositions and talk about the different ways they are related
to the world’. If propositions already belong with facts to the world,
this makes no sense at all. So part of the agenda for an identity con-
ception set by Strawson’s arguments would be to acknowledge that
there will be no explanation to put in place of the correspondence
theory.
An identity theory’s response to Strawson’s second complaint, that
a world of facts is objectionably ‘mind-shaped’, is bound to be indi-
rect. An identity theory cannot deny that the world has the shape
of thought – that it has propositional structure built in. Its only alter-
native is to deny that there is anything objectionably idealist in
acknowledging that. It must insist that to display propositional struc-
ture need not mean to have been structured in conformity to pro-
positional thought. The traditional, and I think only, defence of that
insistence is to disarm supposed alternatives by comparison with
which a propositionally structured world seems less than fully real.
We get one indication of how that defence should run in noting
how Strawson’s second complaint also falls short of its proper con-
clusion. Objecting to Austin’s claim that statements fit the facts
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 49
loosely he does not altogether reject of the idea of a loose-fitting
garb. His complaint is only that it is things, not facts, that wear the
garb. Things, but not facts, can be variously described, for various
purposes, from various perspectives, and with varying degrees of
adequacy. While there could be no looseness in the internal relation
of statement and fact, things are merely externally related to dis-
course; as external relata, they do not have the form of thought
built in.
That, at any rate, is the picture suggested by perhaps the most
memorable and striking passage of Strawson’s article:
If you prise the statements off the world, you prise the facts off
it too; but the world would be none the poorer. (You don’t also
prise off the world what the statements are about – for this you
would need a different kind of lever.) (1950: 197)

This is a picture in which objects provide just the kind of invidious,

external comparison that makes the identity theory’s conception of
the world appear idealist. But its separation of facts from the objects
they concern opens the picture to complaint from any number of
directions. In the present context it might be natural to ground the
the complaint on Wittgenstein’s echo of Frege’s context principle,
according to which the notion of an object just is the notion of
something referred to in the expression of a proposition (TLP 3.3).
A nicely ad hominem variant of that would be to ask the author of
The Bounds of Sense whether the “System of all Principles” might be
‘prised off the world’ while leaving behind objects, the general
concept of which those principles elaborate – the categories being
“concepts of an object in general” (CPR B128). But the most imme-
diately telling point is the one we have already met: that the picture
invites precisely the kind of pseudo-theorizing that Strawson’s first
complaint rejects.
This consideration of Strawson’s article has suggested a two-stage
agenda for an identity theory. It should first accept Russell’s obser-
vation, that the world is, as one might say, not less than the totality
of facts. The second and more delicate part of the conception to
make out will hold that the world is not more than that, or to borrow
from Wittgenstein, that there is nothing “outside the facts” (cf. NB
51). Along the way we’ve noted two further points about this agenda.
First, that to avoid the charge of idealism it will be important that
objects are not left over as something “outside the facts”. And sec-
© Blackwell Publishing Ltd. 2005
50 Philosophical Investigations
ondly, that adopting this agenda will involve renouncing completely,
and not merely transposing, the kind of explanatory ambitions that
Strawson associates with the correspondence theory.

§3 Hornsby against the Correspondence Theory

I suggest that we read advocates of the modern identity theory as,

in part, pursuing that agenda, and in this section will illustrate the
suggestion by sketching Hornsby’s criticism of the correspondence
theory. While her criticism at first seems directed against kinds of
correspondence theory somewhat different from Strawson’s target,
the overall shape of the complaint is similar. She presents the corre-
spondence theory as a product of misguided ambitions characteris-
tic of an outlook we are better off leaving behind: “philosophers
formuations are”, she says, “apt to create an outlook that is forsworn
when an identity theory displaces a correspondence theory”
(Hornsby 1997: 9). Closer parallels will emerge when we follow
through the complaint in a little more detail.
Central to Hornsby’s complaint is a contrast between the explana-
tory modesty of her own identity theory and the correspondence
theorist’s characteristic
willingness to reconstruct thinkables from posited entities of a
different sort, entities which make things true . . . [the correspond-
ents] are supposed to be items which we can specify independ-
ently of an account of thinkables, items which may confer truth
upon a thinkable. When they are introduced, however, we cannot
hold onto the truism that inspires the identity theory. The fact
. . . that autumn has begun, if it were [such a thing as a corre-
spondence theory claims], would not be the same as what I think
when I think truly that autumn has begun. (1997: 8)
The willingness mentioned here can be seen as the result of a piece
of reasoning. Correspondence is a general theory, but involves spe-
cific commitments: there must be an answer to the question, for a
given true proposition, what it corresponds with; for instance, there
must be something about the way things are, or about what there
is, that makes true the proposition that the Eiger is steep. Now the
proposition that the Eiger is steep is itself a specification of a way
things are, which seems exactly to meet that bill. But correspondence
implies difference, a difference that will be invisible if the corre-
spondent is specified in terms too close to those of the original
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 51
proposition, for instance, as ‘the steepness of the Eiger’ or ‘the Eiger’s
being steep’. So to preserve the difference his theory requires a cor-
respondence theorist will naturally resort to other terms. For
instance, he might speak of the ratio of the Eiger’s height to its base
area being relatively large.
One can see, then, why a correspondence theorist might be
willing, and might even feel compelled, to specify a correspondent
for a given proposition differently from the proposition itself. But it
doesn’t emerge from that train of thought what is philosophically
suspect about that willingness. Specifying something differently is
getting a different angle on it, and that often seems a good thing to
do. Hornsby, though, clearly has in mind a much more radical depar-
ture from one’s first view of things.
[I]t will be distinctive of correspondence theorists to seek items
located outside the realm of thinkables, and outside the realm of
ordinary objects of reference, but related, some of them, to whole
thinkables. The idea is widespread and it takes various guises. In
the Russell of An Inquiry into Meaning and Truth . . . the basic cor-
respondents are percepts . . . In the Quine of Philosophy of Logic,
the correspondents are cosmic distributions of particles. (1997: 7)
That the ratio of the Eiger’s height to its area is relatively large is
something one can think. So that specification, on the face of it, isn’t
of something “outside the realm of thinkables”.The same would hold
of any other specification one might come up with. So what might
it be to find anything of the kind that the correspondence theorist
is here said to seek?
Hornsby’s examples give a grammatical clue. To specify a propo-
sition is to propound it, and propounding is done in a sentence. Col-
lections of percepts or particle distributions seem rather the kind of
thing that one might refer to by a singular term. Something “outside
the realm of thinkables” would then be something lacking proposi-
tional articulation reflected in a sentence expressing it. An example,
so far, would be a chair. A chair is something one can refer to, hence
think about, but not something one can think. A correspondence the-
orist’s correspondents are to be in that respect like chairs, but unlike
a chair in being a correlate of a whole sentence rather than a ref-
erential part of one. If a correspondence theorist were committed to
specifying such a thing his position would surely be hopeless, since
the very notion of it imposes inconsistent demands on the form the
specification is to take.
© Blackwell Publishing Ltd. 2005
52 Philosophical Investigations
To this point Hornsby’s argument parallels one formulation of
Strawson’s first objection to the correspondence theory, that it over-
rides “the complete difference of type” (Strawson 1950: 195)
between fact and thing, or between stating and referring. When the
notion of a fact is pulled towards ‘discourse’, a fact is thought of as
what a statement expresses; when it is pulled towards the world, it
becomes what a statement is about (1950: 194–5). Correcting that
confusion needs only the grammatical reminder that nothing less
than a statement will do to specify what makes a statement true. But
that is just Russell’s observation in linguistic dress. So correcting the
confusion and respecting the observation yields only the first stage
of the agenda, that the world is nothing less than the totality of facts.
We have to introduce a further theme to explain how Hornsby
addresses the second stage of the agenda.
Hornsby’s argument involves, not just the notion of a thinkable,
but that of a realm of thinkables. Like McDowell she opposes any
position on which a notion of normative significance – justification
in McDowell’s case, truth in ours – rests on the obtaining of a rela-
tion between something within and something outside that realm.
It seems that a correspondence theorist could be pushed towards
such a position only by a radical generalization of the line of thought
that prompts him to specify the correspondent for a proposition dif-
ferently from the way it is specified in the proposition itself. Applied
locally – one proposition at a time, as it were – that seems only to
push one around one’s repertoire of description to find a new con-
ceptual angle. But those local commitments derive from a general
theory of what it is for anything one might think to be made true
by a corresponding reality. Allowing the same impetus to operate at
a global level yields the idea that the whole system of our thoughts
is true if it corresponds to reality – but now, a reality the only ade-
quate specification of which must be different from anything we
might think or say. The ‘other angle’ called for by this global appli-
cation of the correspondence theorist’s thought is thus a perspective
that we, trivially, cannot occupy. The result is not that we cannot
describe the reality that makes our propositions true. That we do so
is a starting point. But the mere idea of that other angle suggests
that any employment of our concepts yields only a description from
this angle, and so a description only of reality as it appears to a mind
equipped with our concepts. If one were then tempted to gesture
towards how reality might present itself to that other angle, the gram-
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 53
matical inarticulacy we noted in Hornsby’s examples would seem
suited to the ambition: lack of propositional structure in the speci-
fication would register that what one aims to specify is a reality not
conditioned by our conceptual perspective.2 Relations between our
thoughts and such a reality are necessarily inscrutable to us. So
resting normative notions on the obtaining of such relations presents
the traditional choice, between scepticism and an idealism that
preserves knowledge claims by restricting their pretensions to
appearances.
Those, I think, are the threatened consequences that lead Hornsby
to reject the ambitions of the correspondence theory to rest with
the explanatory modesty of the identity theory’s truisms. A corre-
spondent for the whole scheme of our thoughts would be an exter-
nal object of invidious comparison, a comparison that reduces what
we can think to mere appearance. To reject the theory that calls for
that correspondent is to accept that the world is no more than the
totality of facts. It completes the second stage of the agenda.

§4 Parallel Thoughts in the Tractatus

Supposing what I’ve said is at least partly right about what moti-
vates the modern identity theory, and about the sense it gives to its
Tractarian motto, how do things stand with my opening question,
whether there is any genuine affinity between this recent line of
thought and the Tractatus? That question divides, I think, into two.
First, does the shared claim also have the sense in the Tractatus that
the world is ‘no more than’ the totality of facts? In this section I will
argue that the answer to that is ‘Yes’. That answer gives rise to a
second question, that I will consider in the next and final section:
namely, whether its having that sense is separable from other Trac-
tarian commitments that proponents of the modern identity theory
would certainly not be willing to endorse. My answer to the second
question will be ‘No’.
Part of the correspondence package that Hornsby rejects is any
“reconstruction of thinkables from something else” (1997: 10).
McDowell likewise has long rejected any ambition to explain
meaning ‘as from outside’. ‘From outside’ means from a perspective

2. To avoid confusion: I said it would register that intention, not realize it.

© Blackwell Publishing Ltd. 2005

54 Philosophical Investigations
other than the standpoint that our concepts provide, and therefore
a perspective whose bare possibility threatens to reduce to mere
appearances the facts that present themselves to our standpoint. Jus-
tifying a positive answer to our first question about the Tractatus thus
has an easy and a hard part. To show that it too rejects any ‘side-on’
perspective is easy. To argue that its rejection of that perspective
serves to avoid an idealist reduction of the facts presented to our
standpoint is harder. I’ll do the easy thing first, before sketching how
one might go about doing the harder one.
Central to the picture theory is that the form of a picture con-
stitutes its standpoint on the reality it represents, a standpoint that
cannot be transcended to become an object of representation in turn.
A picture represents its object from without. (Its standpoint is its
form of representation.) That is how a picture represents its object
truly or falsely.
A picture cannot, however, place itself outside its form of
representation.
(TLP 2.173–4)

So far that might seem to leave open the possibility of stepping aside
from the standpoint of any particular proposition or region of
thought by shifting to another. But the unity implied by the gener-
ality of logic ensures that there is a single such standpoint:
Propositions can represent the whole of reality, but they cannot
represent what they must have in common with reality in order
to be able to represent it – logical form.
In order to be able to represent logical form, we should have
to be able to position ourselves with propositions outside logic,
that is, outside the world.
(TLP 4.12)

Because of the all-embracingness of the logical perspective, the idea

of a side-ways view on it is just the emptily impossible idea of extra-
logical thought (TLP 3.03). That is all that needs to be said, I think,
to answer the easy part of the question.
To answer the hard part would be to show how in the Tractatus
the impossibility of a side-on view counters a threatened idealist
reduction of the facts on which my language provides a standpoint.
That is to say, it would be to explain how that idea contributes to
Wittgenstein’s treatment of solipsism, to show that the impetus of
solipsism evaporates to leave a kind of realism. I have attempted that
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 55
elsewhere,3 so know that it is not just lack of space that prevents me
succeeding fully in it here. But any persuasive answer would have to
make central Wittgenstein’s metaphor of the visual field.
5.6331 For the form of the visual field is surely not like this:

Eye

The centrally important fact about this metaphor is that it is intro-

duced by Wittgenstein’s idealist interlocutor, and then immediately
subjected to criticism.Wittgenstein points out that as a representation
of the visual field, rather than a third person’s representation of a
region of the world falling with the range of the eye, everything about
the drawing is wrong. In the first place,“you really do not see the eye”
(TLP 5.633): the eye can have no place in a representation of the
field, but must provide the perspective of the representation. But then
in immediate consequence of that, the rest of the drawing is wrong
too: from the perspective of the eye, the field has no boundary.
Wittgenstein’s intention in the passage is to suggest that an anal-
ogous representation of the relation of the subject to its world makes
the same mistakes. That representation, Wittgenstein holds, may be
prompted by a genuine insight, that the notions of thought, and of
what it is for there to be a world, are internally related. That, again,
is the upshot of Strawson’s first point, shifted, as we recommended,
to suit Russell’s observation. It yields the first stage of the agenda,
that the world is not less than the totality of facts. But the insight
lends itself, as McDowell remarked, to high-flown expression. In
McDowell its high-flown form is that “there is no gap between
thought, as such, and the world” (1994: 27). In Wittgenstein it
becomes: “The world is my world” (TLP 5.62). However natural,
that way of voicing the insight makes the same mistakes as the
picture. It makes the first mistake, of placing the subject. And in con-
sequence of that it is forced into the second mistake of drawing a
boundary around the subject’s world. That the second mistake is a
mistake is, outside of the metaphor, no longer a phenomenological
point. Drawing a boundary around the world is a mistake just
because nothing so bounded can be everything that is the case, or
3. Sullivan 1996; this and the following paragraph draw on §II of that paper.

© Blackwell Publishing Ltd. 2005

56 Philosophical Investigations
claim for itself the title of the world.Whatever is representable belongs
to the world. We therefore have no choice but to award that title,
no longer to the region circumscribed in the representation, but to
all that presents itself to the external perspective of the representa-
tion in which the subject and its ‘world’ are placed. And in doing
that we necessarily reduce the subject’s ‘world’ to appearance, and
allow the principles that structure it only a conditioned necessity.
Kantian idealism rests with that retreat, counting the restricted objec-
tivity it secures as at any rate better than none. Wittgenstein diag-
noses it instead as an imposition of an external representation that
distorts the very thing it was invoked to capture, the internal con-
nection between thought and world. Once that is recognized as a
distortion one has no further use for the external perspective, and
the threat of idealist reduction it carried then simply lapses. As
before, abandoning the impetus to that external perspective is accept-
ing that the world is no more than the totality of facts. It yields the
second stage of the common agenda.

§5 A Significant Point of Divergence

The answer to our first question, then, is that recent advocates of

the identity theory share much more than a turn of phrase with the
Tractatus. They give to its opening characterization of the world a
sense informed by a similar vision of the metaphysical alternatives,
and their endorsement of it derives in part from the same pressures.
But only in part. And that raises the second question, whether
accepting that the world is ‘no more than’ the totality of facts is
separable from other Tractarian commitments which Hornsby and
McDowell would not recognize.
The question is best approached by noting a division I have so
far suppressed between three historical advocates of the identity
theory: Moore, Russell, and Frege.4 For Moore and Russell the con-
4. The idea of an identity between true propositions and facts has had a discontin-
uous career. It was subscribed to, at least for a while, and in one form or another,
by the three founding figures of analytical philosophy. It was advanced by Moore in
“The nature of judgement” (1899), and retracted in Some Main Problems of Philoso-
phy (1953, written 1910–11). Russell accepted it in The Principles of Mathematics
(1903), argued most explicitly for it in his 1904 articles critical of Meinong (reprinted
in Essays in Analysis, 1973), first expressed reservations about it in “On the nature of
truth” (1906), and rejected it in the 1910 essay “On the nature of truth and false-

© Blackwell Publishing Ltd. 2005

 Peter M. Sullivan 57
stituents of propositions, and so of facts, include those things a propo-
sition is about. For Frege the constituents of a proposition (or
‘thought’) are senses, while what the proposition is about are the ref-
erents determined by those senses.
The opposition between these conceptions was a running theme
of the correspondence between Frege and Russell in 1903–4 (Frege
1980: 130–70). Frege’s central argument in those letters is now very

hood” (in Philosophical Essays). Frege’s case is somewhat different: he never rejected
the identity claim, but it is perhaps unclear how early he would have subscribed to
it. His only explicit endorsement of the claim comes in the 1918 essay “Thoughts”,
where in answer to the question “What is a fact?” he says: “A fact is a thought that
is true” (Frege 1984: 368).
Recent interest in the identity claim derives from Richard Cartwright’s paper “A
neglected theory of truth” (1987). The role of the claim as a connection between
Bradley and Moore is explored in papers by Tom Baldwin (1991) and Stewart Can-
dlish (1989, 1995), and it seems to be Candlish who first spoke of an “identity theory
of truth” (1989: 338). Whilst these authors’ interest was primarily historical, Hornsby
and Dodd advocate versions of such a theory in its own right (Dodd and Hornsby
1992), while disagreeing over which version to advocate: see Dodd (1999) and
Hornsby (1999) in response.
A line of thought common to the three historical advocates of the identity claim
runs as follows. One who believes truly believes things to be how they are. What he
believes is that things are thus-and-so, and things are thus-and-so. Otherwise put,
what is believed is how things are. And if what is believed were not how things are,
how could the belief be true? There can, then, be no distinction between what is
truly believed and what is so. In particular, there can be no such distinction between
a truth, and that in virtue of which it is a truth, as is presupposed in a correspon-
dence theory.
In Russell this line of thought appears as a compressed objection to the idea that
“a true proposition expresses fact”:
This at once raises the problem: What is a fact? And the difficulty of this
problem lies in this, that a fact appears to be merely a true proposition,
so that what seem a significant assertion becomes a tautology. (1904: 75)
In Moore (1902) the presentation is more developed and more clearly focused on a
presupposition of correspondence theories:
It is commonly supposed that the truth of a proposition consists in some
relation which it bears to reality, and falsehood in the absence of this rela-
tion. The relation in question is generally called a ‘correspondence’ or
‘agreement’; and it seems to be generally conceived as one of partial sim-
ilarity to something else, and hence it is essential to the theory that a truth
should differ in some specific way from the reality, in relation to which
its truth is to consist. It is the impossibility of finding any such difference
between a truth and the reality to which it is supposed to correspond
which refutes the theory.
Finally, and succinctly, in Frege:
A correspondence can only be perfect if the corresponding things coin-
cide and so are just not different things. (1984: 352–3)

© Blackwell Publishing Ltd. 2005

58 Philosophical Investigations
familiar: the things it is about cannot be the constituents of a propo-
sition, because there are significant differences between propositions
which do not differ in what they are about; expressed linguistically,
substitution of co-designative terms in a sentence can alter the
proposition expressed (1980: 164). Russell’s opposed argument is
much less familiar, and seems to turn on the status of the relation
that Frege must presume between the sense figuring in a proposi-
tion and the thing it is about. He wrote:

I believe that in spite of all its snow fields Mont Blanc itself is a
component part of what is actually asserted in the proposition
‘Mont Blanc is more than 4000 metres high’.We do not assert the
thought, for this is a private psychological matter: we assert the
object of the thought, and this is, to my mind, a certain complex
(an objective proposition, one might say) in which Mont Blanc is
itself a component part. If we do not admit this, we get the con-
clusion that we know nothing at all about Mont Blanc. (Russell
to Frege, 12.12.04, in Frege 1980: 169)

Why Russell should think that conclusion follows is not exactly

obvious, but comparison with the argument that was to appear six
months later in “On denoting” suggests a kind of regress argument.
Grasp of the sense cannot make possible knowledge of the thing,
Russell is saying, without an appreciation of the relation between
the sense and the thing. But knowledge of that relation can be of
help only if the thing can figure in its own right in what is then
known. If it cannot so figure, then the problem of connecting knowl-
edge with the thing is just reiterated. But if it can, then we might
just as well accept that it can likewise figure in the original
proposition.
Suppose – because we cannot here assess them – that there is
force in both arguments. The position they jointly dictate is, barring
terminological shifts, the position of the Tractatus. Objects figure in
their own right in propositions: “The name means the object. The
object is its meaning.” (TLP 3.203) But grasp of these objects meets
Frege’s constraint on the grasp of senses: it is impossible that one
should grasp the same object in significantly different ways (TLP
4.243). Objects, in consequence, must be simple (TLP 2.02).
Acquaintance with them is an all or nothing matter: in knowing an
object one knows all there is to its being that object (TLP 2.01231).
Objects are not the kind of things on which there could be another
© Blackwell Publishing Ltd. 2005
 Peter M. Sullivan 59
angle. (So, though it carries some danger, there would be point in
saying that, if Wittgenstein’s objects were to be described in Frege’s
terms, they should be counted more as minimal elements of sense
than of reference.)
But then how can such objects constitute an external world? Part
of what is involved in externality is what common sense talks of as
the cussedness of things, their having other sides responsible for often
unexpected and sometimes unwanted consequences. External things
do not reveal themselves completely to a single viewpoint, and our
grasp of them will be in various ways incomplete.
The answer to the question I just posed, of how Wittgenstein’s
simple objects could compose an external world, is that they don’t.
The features of externality I mentioned are not neglected in the Trac-
tatus. They are, though, located at a different level of analysis. It is
complexes that constitute the external world, and it is complexes that
display the multi-facetedness and contingency of external things.
Taking different angles on a complex involves the recognition of dif-
ferent facts (TLP 5.5423), and realizing the common bearing of those
different takes is recognizing still further facts. The effect of the
scheme of analysis in the Tractatus is in that way to remove a source
of tension between a conception of the world as the totality of facts
and a legitimate sense in which it is also a world of things.
Now compare with this the conception of our modern advocates
of the identity claim, McDowell and Hornsby. The true propositions
with which they identify facts are Fregean thoughts.These thoughts,
and their constituents, are cognitive relata that completely determine
a mind’s cognitive relation to them.5 The standard they are required
to meet is: same item–same grasp; different grasp–different item
grasped. So they, too, are not the kind of thing on which there can
be different takes or different angles. So again, to identify the world
with the totality of such facts so far fails to accommodate the con-
tingency and multi-facetedness of things. In contrast to the Tractatus,
however, ordinary objects provide no anchor for these characteristic
features of externality at another level of analysis. Instead, ordinary
objects seem to have dropped out of the frame altogether.

5. This is the theoretical essence of Frege’s notion of sense: Frege’s ‘thoughts’ are to
be objects of thought which have built into them every logically or epistemologi-
cally relevant similarity and difference between thoughts with those objects; or, as
one might tendentiously express the point, they are ‘thoughts with their under-
standing built in’. Cf. McDowell (1994: 180).

© Blackwell Publishing Ltd. 2005

60 Philosophical Investigations
This is, I realize, an extraordinary charge to bring against a leading
advocate of object-dependent senses, as McDowell is.6 But then the
charge is precisely that he has made an extraordinary mistake. It is
not, however, an inexplicable mistake. We saw, in connection with
Strawson, that to cast objects in the role of something ‘outside the
facts’ would make for the kind of invidious comparison that grounds
the threat of a Kantian version of idealism. That threat is something
McDowell is conscious of and keen to avoid. So, lacking Wittgen-
stein’s way of locating objects ‘inside the facts’, he responds to the
invidious comparison in a cruder and historically more common
way: by cutting off the offending limb. But what then remains is a
picture in which the ordinary features of externality have no anchor
at all. It is one that positively invites the Johnsonian, stone-kicking
charge of idealism.7
Essentially the same point can be made in connection with
Hornsby’s advocacy of the identity theory. Her identification of the
world with the totality of facts aims to avoid the idealist conse-
quences of treating the world as a common focus between our own
and an impossibly external perspective. I agreed, and took the Trac-
tatus to agree, that to do that is a mistake. But again, it is not a gra-
tuitous or inexplicable mistake. It appeared rather in our discussion
of Hornsby to be an excessive generalization of a good thought,
indeed a thought that one might regard as non-negotiable: that
reality does serve as the common focus of those perspectives that are
possible. In one way or another, that is, reality must include the kind
of thing on which different angles are possible. The Tractatus has an

6. McDowell anticipates an objection that the “drift” of this account is “idealistic”

in providing, as he puts it, for “an alignment of mind with the realm of sense, not
with the realm of reference” (1994: 179); and he replies, in effect, that on the right
view of sense an alignment with senses is an alignment with references. That is, he
presumes that the objection must be motivated by a wholly descriptive or specifi-
catory conception of the relation of sense to reference, and so takes it to be ruled
out by the account of object-dependent senses he shares with Evans (McDowell
1977, 1984; Evans 1982). Even if McDowell’s diagnosis of the objection was right,
I think this response is too local to be effective: the descriptive conception of sense
is, after all, not universally wrong. But more important than that is that McDowell
misconstrues the objection, which has to do with his characterization of reality rather
than any question of ‘alignment’.
7. The invitation is obvious in one of McDowell’s metaphorical summaries of his
view, that “thought and reality meet in the realm of sense” (1984: 180). If that is
right, then to identify the world with a totality drawn from that realm is to iden-
tify the world with its meeting point with mind.

© Blackwell Publishing Ltd. 2005

 Peter M. Sullivan 61
unexpected way of acknowledging that. So far as I can see, modern
proponents of the identity theory have no way at all.8

References
Armstrong, D. M. (1973) Belief, Truth and Knowledge, Cambridge:
Cambridge University Press.
Austin, J. L. (1950) ‘Truth’, Proceedings of the Aristotelian Society, SV
XXIV. Reprinted in his Philosophical Papers, Oxford: Clarendon
Press, 1979.
Baldwin, Tom (1991) ‘The identity theory of truth’, Mind 100.
Candlish, Stewart (1989) ‘The truth about F. H. Bradley’, Mind
98.
(1995) ‘Resurrecting the identity theory of truth’, Bradley Studies 1.
Cartwright, Richard (1987) ‘A neglected theory of truth’, in his
Philosophical Essays, Cambridge, Ma.: MIT Press.
Dodd, Julian (1999) ‘Hornsby on the identity theory of truth’, Pro-
ceedings of the Aristotelian Society XCIX.
Dodd, Julian and Hornsby, Jennifer (1992) ‘The identity theory of
truth: reply to Baldwin’, Mind 101.
Evans, Gareth (1982) Varieties of Reference, Oxford: Clarendon Press.
Frege, Gottlob (1980) Philosophical and Mathematical Correspondence,
ed. B. F. McGuinness, Oxford: Blackwell.
(1984) Collected Papers on Mathematics, Logic and Philosophy, ed. B.
F. McGuinness, Oxford: Blackwell.
Hornsby, Jennifer (1997) ‘Truth: the identity theory’, Proceedings of
the Aristotelian Society XCVII.
(1999) ‘The facts in question: a response to Dodd and to Can-
dlish’, Proceedings of the Aristotelian Society XCIX.
Kant, I. (1929) Critique of Pure Reason, trans. N. Kemp Smith,
London: Macmillan. Referred to as ‘CPR’; page references in
standard A and B format.
McDowell, John (1977) ‘On the sense and reference of a proper
name’, Mind 86.

8. Ancestors of this paper were given to seminars at the universities of Michigan

(Ann Arbor), Illinois (Champaign-Urbana), Birmingham, and at a conference on the
Tractatus hosted by the University of Utrecht. Thanks to all who commented on
those occasions, and to Jennifer Hornsby and Ian Proops for comments on a draft.
One round of revision occurred during research leave funded by the Arts and
Humanities Research Board, whose support I gratefully acknowledge.

© Blackwell Publishing Ltd. 2005

62 Philosophical Investigations
(1984) ‘De re senses’, Philosophical Quarterly 34.
(1994) Mind and World, Cambridge, Ma.: Harvard University Press.
Moore, A. W. (1997) Points of View, Oxford: Clarendon Press.
Moore, G. E. (1899) ‘The nature of judgement’, Mind 8.
(1902) ‘Truth’, in J. M. Baldwin, ed., Dictionary of Philosophy and
Psychology, vol. 2, London: Macmillan.
(1953) Some Main Problems in Philosophy, London:Allen and Unwin.
O’Connor, D. J. (1975) The Correspondence Theory of Truth, London:
Hutchinson.
Russell, Bertrand (1903) The Principles of Mathematics, Cambridge:
Cambridge University Press. (2nd ed. 1937, London: Allen and
Unwin.)
(1904) ‘Meinong’s theory of complexes and assumptions’, Mind
13. Reprinted in Russell (1973).
(1905) ‘On Denoting’, Mind 14. Reprinted in Russell (1973).
(1906) ‘The nature of truth’, Proceedings of the Aristotelian Society VII.
(1910) ‘On the nature of truth and falsehood’, in Philosophical
Essays, London: Allen and Unwin.
(1922) ‘Introduction’ to Wittgenstein (1922).
(1973) Essays in Analysis, ed. D. P Lackey, London:Allen and Unwin.
Strawson, P. F. (1950) ‘Truth’, Proceedings of the Aristotelian Society, SV
XXIV. Reprinted in his Logico-Linguistic Papers, London: Methuen,
1971.
(1966) The Bounds of Sense, London: Methuen.
Sullivan, Peter M. (1996) ‘The “truth” in solipsism, and Wittgenstein’s
rejection of the a priori’, European Journal of Philosophy 4.
(2000) ‘The totality of facts’, Proceedings of the Aristotelian Society C.
Travis, Charles (2000) Unshadowed Thought, Cambridge, Ma.: Harvard
University Press.
Wittgenstein, Ludwig (1922) Tractatus Logico-Philosophicus, London:
Kegan Paul. Referred to as ‘TLP’.
(1953) Philosophical Investigations, Oxford: Blackwell. Referred to
as ‘PI’.
(1979) Notebooks 1914–1916, Oxford: Blackwell. Referred to as
‘NB’.

Department of Philosophy
University of Stirling
Stirling
FK9 4LA
© Blackwell Publishing Ltd. 2005