Two Wittgensteins Too Many:

Wittgenstein’s Foundationalism
DANIEL D. HUTTO

If the true is what is grounded, then the ground is not true, nor yet false. (OC 205)

Apparently, Polish logicians count differently from most people. This is because they
operate with different numerical criteria. Disagreements between them and us about
how many objects there are in a room will look not only insuperable but also
unintelligible unless the differences between their system and ours are brought into
the light. Similarly, if we are to put the debate about how many Wittgensteins there
are on a proper footing it will be important to establish the criterion according to
which we are making our count.
Just as we individuate Russells and Putnams by the number of significant changes in
their approaches to philosophy (the former case yielding a legion), so too has the
familiar debate over whether there is one or two Wittgensteins been conducted
against the background of the now controversial claim that there was a radical break
in his thinking. This is regularly presented as a sea change in his views about
language, dividing the work of his initial period which culminated in the Tractatus
Logico-Philosophicus, from the writings of his later period which found their first
mature expression in the Philosophical Investigations and continued along the same
lines until the very end, including the musings of On Certainty. The nature of this
break focuses on the alleged change in his understanding of the way language
operates. Traditionally, it is supposed that he abandoned his purely referential 'theory'
about picturing and its attendant metaphysics in favour of the idea that 'meaning is
use'. Inter alia, this understanding frequently partners the claim that his philosophy
underwent an important shift from realism to anti-realism in the process. There are
two main ways of understanding the nature of this change and, consequently, the
relation between the early and later writings.1 According to some, the later works
develop a new, successor theory that replaces the early one. For others, the change
heralded the introduction of an anti-theoretical dimension to his thinking, isolating his
theoretical ambitions to his early period.2 Although these readings disagree about the
status of the Investigations, they are united in holding that his early approach to
philosophy is defined by theoretical concerns, broadly construed.
This interpretation of the Tractatus has recently been challenged by those who seek
to demonstrate that it cannot be easily reconciled with his claim, made during both
periods, that he was not engaged in any form of philosophical theorising at all. For
example, in a series of important papers, James Conant objects that metaphysical
interpretations cannot accommodate the remark that philosophy 'is not a body of
doctrine' (Conant, 1989, 248, 266; TLP 4.114). On such grounds, both he and Cora
Diamond argue convincingly that any simple doctrinal interpretation fails 'to take
seriously what Wittgenstein says about philosophy itself' (Diamond, 1995, 18). But if
there are no theories or doctrines advanced, then these cannot be appealed to as a
basis for distinguishing the 'early' from the 'late' (or even the ‘latest’) Wittgenstein.
Whatever one makes of this debate, the claim and its counter are clear enough. The
dispute turns on whether or not we can identify, and confirm, the existence of a
radical break, as opposed to a mere expansion or development, in Wittgenstein's
 2

thinking. Presumably, the idea that we ought to recognize a third Wittgenstein rests
on similar claim. But, what is its basis?

Foundationalism and the Third Wittgenstein

In his contribution to this volume, Avrum Stroll makes the assertion that there is ‘a
feature of [Wittgenstein's] later philosophy that occurs only in On Certainty. This is a
unique form of foundationalism that is neither doxastic nor non-doxastic' (Stroll, this
volume, p. 2). He also holds that Wittgenstein’s increased attention to metaphorical
language in explicating this foundationalism is yet another feature that sets it apart
from the rest of his corpus. I raise doubts about appealing to either of these aspects as
a rationale for identifying a third Wittgenstein. I argue that Wittgenstein's
commitment to foundationalism – to the extent we should recognise it at all – and his
concern with the non-literal are not unprecedented; they are present in his earliest
writings.
As a precursor, it is important to clarify what form his foundationalism took, what it
amounts to and what allegedly makes it unique. According to Stroll’s original
formulation, to be a foundationalist one must minimally subscribe to two claims.
First, that there is an asymmetrical relationship of epistemic dependency holding
between some 'primordial' knowledge, F, and the remainder, R. Second, that F itself
depends on nothing. However, this leaves open the question of what slots into F and
R such that, unlike traditional forms, Wittgenstein’s foundationalism is 'a view to the
effect that certitude stands in a foundational relationship to the language game itself'
(Stroll 1994, 141).3 As evidence that Wittgenstein was thus committed, Stroll directs
his readers to over sixty remarks, comprising a tenth of the total remarks in On
Certainty.
According to Stroll Wittgenstein distinguishes what is basic at different times in
'propositional' and 'non-propositional' terms. Within the scope of the former, he
distinguishes 'relative' and 'absolute' forms of certainty. The former involves
particular propositions, which can be doubted in some contexts but not others,
whereas the latter involves propositions that it never makes sense to doubt at all; they
are 'that which stands fast for us' and thus that which funds, supports and makes our
language games possible at all. In this respect, these 'hinge propositions' are not
properly propositions at all. It is this latter idea that Stroll sees as developing to
maturity throughout On Certainty, with section 204 supposedly marking the crucial
turning point. Nevertheless, he acknowledges that it is a matter of consternation for
his reading that Wittgenstein at times vacillates between the propositional and non-
propositional descriptions of what is foundational even beyond this point.
Consequently, it is ultimately the more radical idea – that the foundation of our
everyday practices concerning knowledge and doubt is such that it cannot by itself be
intelligibly brought into question because it is non-propositional – that is meant to set
the insights of On Certainty apart from his early work.
Several authors have found the very suggestion that Wittgenstein was a
foundationalist, at all, untenable on grounds that ought to evoke sympathy (see Li
1999, Richter 2001). Firstly, Stroll’s claim looks immediately suspect given that
Wittgenstein's anti-theoretical emphasis should evoke caution about committing him
to an '-ism' to the extent that this implies that he would have been advancing any kind
of epistemological theory. Yet even in the act of assigning this label, Stroll recognises
that he is running the risk of attributing to Wittgenstein a position or doctrine and
thereby – given his disavowal of the idea that philosophy ought to advance theses –
 3

threatening to 'hoist him by his own petard'. What rescues him from this
uncomfortable fate is that his form of foundationalism is not merely unconventional,
but that it is so in a way that disallows theorising about it or treating it as the product
of a theory. This is a feature of its peculiarly heterogeneous character, given that the
supporting base for knowledge cannot be characterised in the same categorical terms
as that which it supports. Specifically, the thought is that our genuine claims to
knowledge rest on certainty. This manifests itself in our actions, in a way that cannot
be justified or explained. In this regard, his is unlike all traditional forms of epistemic
foundationalism, such as rationalism or empiricism, which treat everyday knowledge
claims as ultimately dependent upon or derived from innate principles or sensory
items that, at least in principle, can be made intelligible by appeal to the very same
categories as that which is derived from them. Hence, it would be unfair to think that
Stroll had not already guarded against the worry that calling Wittgenstein a
foundationalist he was necessarily attributing to him a philosophical 'position'.
A second objection concerns Stroll's formal characterisation of foundationalism and
the way its concomitant imagery encourages us to picture the asymmetric dependence
of our practices on what grounds them. It has been argued that in neatly
distinguishing levels of stratification Stroll misrepresents Wittgenstein’s
understanding of the relationship between what is fundamental and what is not.
Richter tries to capture the force of this worry by noting that, on one hand, nothing
could be a 'chess king' outside a game of chess, yet on the hand, nothing could be a
game of chess without something to play the role of the king. Li makes the same point
in a way that is more directly pertinent to the concerns of On Certainty. Li writes:
I think Stroll is right to say 'that the earth is very old is a proposition that is
presupposed by persons engaging in any historical, anthropological, geological
or etymological inquiry. In that sense the propositions of history, geology and so
on, "depend on" it'. But is it true that 'it [the proposition "The earth is very old"]
will not depend on them [the propositions of history, geology, and so on]'?
Without at least some of the propositions of history, geology, and so on, the
proposition 'The earth is very old' is meaningless. In this sense, the proposition
'The earth is very old' also 'presupposes' ('depends on') the propositions of
history, geology and so on (Li 1999, 12, emphasis added).

In both cases the examples serve to highlight the idea that we cannot neatly separate
or distinguish the base from that which it supports. These sorts of consideration are
meant to lead us to conclude that the two are in an important sense on the same level
and thus one is not more dependent on the other. The crucial point is that both critics
hold that what preconditions our practices cannot be 'logically or temporally prior' to
them (cf. Li 1999, 13; Richter 2001, 355).
It is right to hold that, for Wittgenstein, our form of life is foundational in that it sets
the scope of our various practices and yet can only be characterised by exploring the
full range of practices in which we engage, and in which we might come to engage.
That is to say, it is neither an independent nor a substantial notion (this is something I
will come back to shortly). Yet, the above objection concerns meaning, as Li
expressly stresses, and for this reason, it does not follow that we cannot speak of
levels, with respect to the epistemological issues. It is therefore possible to regard one
level as epistemically foundational to another without overlooking other important
forms of interdependence that must exist between them. Wittgenstein's own
metaphors of the foundation-walls being carried by the whole house and the axis
being fixed by the movement around it, I maintain, are designed to achieve just this
 4

balance and would hardly make sense without it (OC 248, 151). To return to the
example cited, it must be taken as read, as unquestioned, that the earth is very old if
we are even to begin to engage in more specific investigative practices with respect to
the past. Yet the fact that this 'taking as read' can only express itself in and through
these practices does not unseat the idea that this commitment must hold fast for us in
a way that is more fundamental than any of these practices (and not vice versa). Seen
in this light, we can certainly hold on to the idea that there is a kind of bedrock upon
which our practices rest, while at the same time treating this as something intrinsic to
them. There need be no tension between these ideas in the way suggested by Phillips’
remark, which Richter approvingly quotes, that '[t]hey are not the basis on which our
ways of thinking depend (foundationalism), but are basic in our ways of thinking'
(Richter 2001, 356; quoted from Phillips 1995, 123).
Although this much of Stroll’s approach can, I think, be salvaged I fear it is
vulnerable on other fronts. It is rightly complained that, unlike traditional approaches,
there is no sense in which Wittgenstein would have regarded the supervenient forms
of knowledge to be derived or inferred from the subvenient base. If Wittgenstein’s
real innovation was to promote an heterogeneous foundationalism, this much ought to
be obvious, given that there can be no rational relation between what is propositional
and non-propositional (or conceptual and nonconceptual) (cf. Hutto 2000, Ch. 1, esp.
24-26). Nevertheless, Stroll holds that Wittgenstein’s so-called early propositionalist
account of certainty has precisely this feature. He writes: 'insofar as propositions or
even pseudo-propositions or grammatical rules are conceived of as the products of
rational activity, the new view stands in opposition to any such account' (Stroll 1994,
157-8). The contrast between the two accounts he finds in On Certainty is founded on
the idea that the one but not the other regards ordinary knowledge as the 'product of
reasoning and intellection'.
But this is quite wrong. For Wittgenstein there could be nothing to this distinction
precisely because, on inspection, there is nothing genuinely propositional about his
so-called hinge propositions, save their outward appearance. They are propositions in
name only, even calling them such is a misnomer. They do not function as ordinary
propositions at all and hence ought properly to be classified as non-propositional.
Moreover, once it is exposed that they operate like grammatical rules, it becomes
clear that they neither rationally fund other pieces of knowledge nor can be
discovered by means of ratiocination. This is made clear by a host of remarks
concerning rules even in his earlier work Remarks on the Foundation of Mathematics,
which neatly foreshadow those presented in On Certainty. He writes:

…and now I have been told ‘Go on like that’: how do I know what I have to do
next time? – Well, I do it with certainty, I shall also know how I defend what I
do –that is, up to a certain point. If that does not count as a defence then there is
none. (RFM VI 29, emphasis mine; cf. also VI 23-24, 38, 46-47).

…If someone is not a master of a language, I may bring him to mastery of it by

training. Someone who is master of it, I may remind of the kind of training, or I
may describe it; for a particular purpose; thus already using the technique of
language.
To what extent can the function of a rule be described? Someone who is
master of none, I can only train. But how can I explain the nature of a rule to
myself?
The difficult thing here is not, to dig down to the ground; no, it is to recognise
the ground that lies before us as the ground. (RFM VI 31, emphasis mine).
 5

In this light, at best we might regard 'hinge propositions' as expressions of these

sedimentary rules which, to borrow Stroll’s own phrase, we absorb in the course of
customary training such that they become second nature to us. Only in this respect
can they be meaningfully contrasted with our animalistic, instinctive responses: those
that constitute our first nature and which provide the, 'preliminary steps towards
acting according to a rule' (see RFM VI 43).
These considerations expose a more serious objection to Stroll’s characterisation of
Wittgenstein’s foundationalism, concerning the way he tries to understand what is
basic or fundamental to our practices. He identifies 'several candidates for F [what is
absolutely foundational to our language games], and all of them non-intellectual.
Among these are acting, being trained in communal practices, instinct and so on'
(Stroll 1994, 146). He admits that further analysis might reveal them to be at odds but
that Wittgenstein failed to carry out this investigation before his death (Stroll 1994,
158). But this is misguided. Wittgenstein was not prevented by lack of time from
engaging in analysis of that which can only manifest itself in and through our
activities in their different aspects. Nor should these notions be cleanly distinguished
in his thought; it is rather that they all characterise what is basic to our form of life.
Confusion on this score leads to fruitless attempts to separate the various
contributions made to our form of life by the world, our communal practices and our
more animal natures. It is true that Wittgenstein sometimes describes our form of life
in relation to instinctive and natural responses in a way that puts it beyond
justification, as in On Certainty 358-359, which is prominently quoted by Stroll. In
other places, he speaks of it in connection with our more sophisticated practices that
depend on enculturation (cf. PI 19, 23, & p. 174). But it is a mistake to think that
these views are at odds or that they point to two different accounts of our form of life.
Wittgenstein’s seemingly unsystematic remarks on this score serve as an indication
that he thought that our form of life grounds the possibilities for our existence on both
these fronts. Rather, it comprises both in a nested way.
Our natural, animal responses fund, but do not circumscribe, the full range of
possibilities that manifests itself in our sophisticated, cultural and communal
activities. Thus it is wrong to think that they are somehow in competition, as Stroll
appears to when discussing their inter-relations as they appear in the latter part of On
Certainty. Our form of life is inclusive of both our natural reactions and our
education, both of which are inexplicably and intrinsically bound up with the world.
How could we even imagine a separation? The limits of the world are only revealed
through our active engagements with it. Moreover, the world determines the viability
of our various practices and concepts. This is why Wittgenstein says:

…And if things were quite different from what they actually are – if there were
for instance no characteristic expression of pain, of fear, of joy; if rule became
exception and exception rule; or if both became phenomena of roughly equal
frequency – this would make our normal language-games lose their point – the
procedure of putting a lump of cheese on a balance and fixing the price by the
turn of the scale would lose its point if it frequently happened for such lumps to
suddenly grow and shrink for no obvious reason (PI 142).

The full range of possibilities that comprises our form of life is not fixed or definable;
the possibility of change is ever present (cf. OC 63, PI 230). Hence, it is particularly
important to resist a certain relativistic reading that makes it appear as if there might
be many exclusive 'forms' of human life as opposed to a single overarching range of
 6

possibilities (see Garver 1990). Being so, this notion is a gesture at something that
cannot be charted or analysed.
Cleaned up in this way, and with the crucial caveats, proposed above, in place, it
would appear that the attribution of a heterogeneous form of foundationalism is
warranted, however misleading the label might be.4 Having established this we are
now in a position to revisit the question of whether, and to what extent, this
constitutes a unique development of his final period. To make a proper case for this
requires the identification of the putative changes that distinguish the late and the last
Wittgenstein, if we imagine that there is already a way of establishing a clean break
between the early and the late. However, my strategy is to cast doubt on the idea that
there is any ground for thinking that the commitment to foundationalism in On
Certainty, as characterised above, represents anything like a radical break or fissure in
his general approach. For something like this commitment can be identified even in
his earliest writings, although it expresses itself differently there. It is only by
clarifying these matters that we can open the way for a proper characterisation of his
ideas. Consequently I argue that, at least with respect to this issue, it is more
illuminating to focus on what unifies, rather than what divides his thinking.

Tractarian Foundations

Before reviewing Stroll’s reasons for thinking that On Certainty marks a distinct turn
in Wittgenstein’s philosophy I want to consider Genova’s claim that there is a unique
emphasis on action in the late, late Wittgenstein. She holds that his primary metaphor
for our relation to the world, which goes through a staged 'progression from the
Tractatus to On Certainty is marked by a steady advance from thinking through
seeing to acting' (Genova 1995, 65). Undermining this idea not only removes an
independent reason for thinking that On Certainty 'breaks new ground', it also puts us
in a good position to evaluate what I take to be Stroll’s misreading of the Tractatus.
In what follows I concentrate only on the Tractatus for two reasons. The first is
practical – simply a lack of space. The second is strategic for, given transitivity, it is
not possible to make a case for the existence of three Wittgensteins if there is no
demonstrable basis for distinguishing even two of them. If a strong degree of
continuity can be established between the 'early' and 'late' writings, we can regard the
developments in On Certainty as an expansion of his view of the scope of 'grammar',
as it moved in new directions, after being unshackled from his restrictive early vision
of the general form of the proposition. Although this latter observation has merit, it
does not recognise the kind of radical break that warrants shocking talk of a 'third
Wittgenstein'.5 My hope is not just that my discussion will serve as a corrective to
misunderstandings about the Tractatus but that it will also enable us to see more
clearly the continuity of his thought concerning what is inexpressibly fundamental to
our practices. That Wittgenstein was concerned to emphasise action in On Certainty
is not in dispute. The following remark epitomises this:

Our not doubting them all [those facts that stand fast for us] is simply our
manner of judging, therefore of acting. (OC 232, emphasis mine)

The real question is: Why should we think that the identification of judging and
acting is not already present in the Tractatus? Genova’s answer will be familiar to
those who favour a 'representationalist' reading of the Tractarian view of the nature
of thought, of which Malcolm is the most prominent defender. On this reading it is a
 7

special feature of thoughts, as paradigm logical pictures which directly share their
form in common with reality, that they can provide syntactically well-formed
propositional signs with their meaningful links to reality. Apparently if a thought-
picture is formed everything else takes care of itself. Hence:

In the later works, Wittgenstein stresses the dynamic activity of calculating with
pictures, i.e. using them according to rules, rather than the static fact of having a
picture. The activity of thinking rather than the having of thought takes center
stage in the later works (Genova 1995, 65, emphasis in the original).

That Wittgenstein initially viewed thoughts thus is a prima facie plausible way of
making sense of the idea that proposition signs can only be regarded as genuine
bearers of meaning if they are somehow inflated by thoughts and thereby given a
sense. At first blush, the following remarks appear to provide strong evidence of his
commitment to this idea.

The method of projection is to think of the sense of the proposition. (TLP 3.11;
emphasis mine)

Moreover, he also apparently identifies thoughts with logical pictures, the purest form
of a picture.

A logical picture of facts is a thought. (TLP 3)

Ultimately, it appears that the reason for making this identification is that if the
elements of the psychical fact are so related to the elements of pictured fact, this
would explain why propositions, logical pictures, contain their own method of
projection:

So a picture, conceived in this way, also includes the pictorial relationship,

which makes it into a picture. (TLP 2.1512; emphasis added)

Thoughts lend propositional signs content in this way. This interpretation, according
to which language inherits its meaningful properties from underlying thoughts, is
further encouraged by the claim that propositions express thoughts:

In a proposition a thought finds an expression that can be perceived by the

senses. (TLP 3.1)

The upshot of all this seems to be that the projective link between language and the
world goes via a way station of thoughts. The link between words and world is
therefore to be explained by a pre-existing natural projective relation holding
between complete thoughts and specific possible states of affairs in the world.
Accordingly, the objects of thoughts are intrinsically related to the objects of the
world, and this link is transitively conferred on natural language by conventional
means. Hence, a proposition reaches right out to some possible state of affairs
because of the mediating thought that stands behind it. Even if a proposition should
miss its mark, if the world happens not to be as it says, this does not rob the
proposition of its sense. Things might have been as the thought behind it says
precisely because what is thinkable is also possible (TLP 3.02). Being logical
pictures, thoughts can represent what is possibly the case as well as what is actually
the case.
 8

Obviously, such a view need not constitute a retreat to psychologism as the

link between thoughts and reality is not achieved by means of any psychological
processes that could be empirically investigated. It is simply that there is a natural
relation between thought elements and those of the world. Wittgenstein's response to
Russell’s question about the nature of these thought elements is evidence that he had
no interest in these matters, which otherwise ought to loom large. For, in answer to
the query 'Does a thought consist of words?', he replied: 'No, but of psychical
constituents that have the same sort of relation to reality as words. What those
constituents are I don't know' (NB 130). Nor should this admission be surprising,
since speculation on the nature of such constituents ought to have held no interest for
him; it would not be a proper topic for logic. At best the question about their nature
would be one for empirical psychology or, one day, physics. That there must be such
constituents can be asserted without further investigation into their nature. Hence, in
holding that thoughts, language and the world have an internal correlation is not to
make a claim in need of empirical support. He is stating what must be the case if
language is to represent the world.
With the charge of psychologism quashed, there are some obvious attractions to this
reading. Nevertheless, it suffers because it wrongly encourages us to imagine
thoughts as reified entities in a way, pace Genova, that goes precisely against
Wittgenstein’s emphasis on the activity of making sense and applying signs. On this
front, he unequivocally rejects the suspect notions of our being acquainted with
propositions or grasping them as if they had an independent existence. In forming
judgements we are not passively related to pre-formed external items in the mind or
the world, rather in the formation of thoughts we can see the internal relation between
our logical pictures and what they represent. In this way, language becomes related to
the world through being thought out. Thinking, in this sense, is judging, which is an
active process.

What signs fail to express their application shows. What signs slur over, their
application says clearly. (TLP 3.262; emphasis added)

A propositional sign, applied and thought out, is a thought. (TLP 3.5; emphasis
added)

In order to recognize a symbol by its sign we must observe how it is used with
sense. (TLP 3.326; emphasis added)

These applications of which he speaks are bound up with particular acts of

judgement, which involve the making of an assertion as based upon the putting
together of the thought elements to create a psychical fact. It may be useful to think
of the activity of putting together some things – the physical shapes of the bits and
bobs limit how they might go together naturally. Our judgements are limited by the
very chemistry of thoughts, the possibilities of elemental combination. This is why
sense is revealed through our successful applications, whereas nonsense is the product
of malformed attempts at judgement.
This gives rise to questions about what determines the boundaries of sense and
nonsense. In this regard, the existence of logical rules that determine which
combinations are permissible might be thought to underwrite our conventional use of
terms. We find an early expression of this idea in Mounce’s interpretation. He writes:
 9

In order to grasp the logical form of an expression one has to look at the rules for
its use… What is not a matter of convention, however, is how we can use this
mark once we have fixed its meaning by a rule… It is only if a mark is applied
according to a rule which reflects logical form that it has been given a meaning
in the first place. (Mounce 1981, 29-30; emphasis added)

What is right about this is that it underscores Wittgenstein’s insistence that '[a]
proposition is not a blend of words ... A proposition is articulate' (TLP 3.141). To
represent a fact, a proposition cannot simply be a conglomerate of words. For an
individual word to mean anything at all, all the components of a proposition must be
so ordered that they can intelligibly mirror some possible state of affairs. They must
be put together in the right way. Moreover, it is true that in this context he explicitly
talks of the rules of logical syntax. For example, he writes:

What signifies in the symbol is what is common to all those symbols by which it
can be replaced according to the rules of logical syntax. (TLP 3.344)

Of course, as Mounce is well aware, Wittgenstein is not conflating the rules of logical
syntax with the rules of the conventional grammars of different natural languages.
For, unless indefinitely supplemented, the latter always generate ambiguity and
equivocation. Consider propositions (1) and (2) below:

(l) The dog chases the cat.

(2) The house chases the cat.

Propositions (1) and (2) both share the noun-verb-noun syntactic form. Still – (2) is
not a sensible sentence because though 'house' is a noun, the sorts of thing that this
noun designates are in no position to chase cats. This example is exceptionally crude,
but it could easily be replaced with more sophisticated ones. Still, it is enough to
show that what makes sense and what does not is not regulated by the surface syntax
of natural language, rather it is written into the very nature of the things. For
Wittgenstein, the rules of grammar and syntax run deep in that they concern only
intrinsic possibilities, which cannot be violated. They can therefore not be identified
with the rules of various grammars of particular natural languages. This is why '[a]
sign does not determine a logical form unless it is taken together with its logico-
syntactical employment' (TLP 3.327). However, it is also a mistake to think that
Wittgenstein’s talk of rules reveals a commitment to something like the existence of a
Chomskyian universal grammar or Fregean laws of thought. In speaking of rules he is
not advancing any substantive conception of them. They are not something we can, to
use Mounce’s illocution, intelligibly 'grasp'. For this would require that we could, in
contravention to what Wittgenstein expressly denied, give a positive characterisation
of the logical form of the world. Thus logico-syntactic rules for permissible
combinations inherently mirror the logical form of the world and permit us to make
sense of it, but they are only revealed retrospectively and on specific occasions of use.
It is not possible to anticipate them or give them a general characterisation. In no way
can they be specified or codified. These 'rules of use' neither exist as independent
things-in-themselves nor do they occupy a distinct realm, internal or external, that can
be separately investigated or discovered. It must be recognised that Wittgenstein’s
talk of rules here is slippery and it must be compared with what he says elsewhere,
such as:
 10

The rules of logical syntax must go without saying, once we know how each
individual sign signifies. (TLP 3.334; emphasis added)

Indeed, it is on precisely these grounds that Wittgenstein revealed Russell’s theory of

types to be utterly superfluous. He realised that there is no point in charting the rules
for how meanings may be grammatically employed if those rules must be derived
from observing occasion by occasion whether sense has been made or not. At best,
even if we could articulate them, such rules would need to be compiled antecedently
of all particular acts of sense-making.
In the end, to understand Wittgenstein’s position properly we must blend together
certain aspects of the proposals of Malcolm and Mounce, while purging them of
specific errors and of their explanatory concerns. So edited, these readings can be
synthesised to help us to see that his interest was in particular acts of judgement,
which are governed not by an external set of rules, but only internally. This is why the
limits of thought can only be revealed not stated. This is also why he sees individual
acts of making sense along the lines of conducting an experiment:
In a proposition a situation is, as it were, constructed by way of experiment.
(TLP 4.031)

In the process of experimenting, what works and what does not – what makes sense
and what does not – is made manifest. We have no other means of determining this.
Yet even without the existence of a separate set of rules, there are logical limits to
thinking – we cannot think what is illogical (TLP 3.03). That which is illogical is
literally nonsense – it makes no sense. Read in this light it is transparent that from the
very beginning he regarded judgement not as something static but as an activity.
Indeed, this was why he regarded philosophy itself as an activity. This undercuts any
motivation for claiming that Wittgenstein's early emphasis on thought stands over and
against a very late focus on action in his final work. This reading is strengthened by
giving attention to his understanding of the nature and purpose of analysis in a way
that reveals his early commitment to the idea that there was an unsayable foundation
to our particular acts of sense-making. He held that in 'everyday language' the true
form of our thoughts is 'obscured'; nevertheless, determinate thoughts are there,
beneath this clothing, giving meaning to our expressions (TLP 4.002). Since language
obscures the true form of our thought, its propositions are, in principle, capable of
analysis. This requires making transparent the logical form of specific propositions.
For at the level of individual judgements it is possible to reveal how the simple
elements stand proxy, in a particular way, for simple objects of the states of affairs
they depict. In principle, through analysis, the true meaning of our everyday
propositions can be revealed, reflecting the true form of our thoughts. Thus:

In a proposition a thought can be expressed in such a way that elements of the

propositional sign correspond to the objects of thought. (TLP 3.2)

The configuration of the simple signs in the propositional sign corresponds to the
configuration of the objects in the state of affairs. (TLP 3.21)

It is worth considering just how he believed we might prosecute this complete

analysis. At stage one we would need to decompose a seemingly complex proposition
into the elementary ones that comprise it, for each ordinary proposition is in fact,
according to the Tractatus, nothing but the truth functional product of elementary
 11

propositions. Only through analysis would the basic logical form of the elementary
propositions that internally (and truth-functionally) constitute such complex
propositions be revealed. At the end of such a process we encounter elementary
propositions, which show the true form of our thoughts. These are composed of the
names of simple objects, bound together like links in a chain. They mirror basic states
of affairs and constitute the point at which we reach bedrock, where analysis must
come to an end. For, '[n]ames cannot be taken to pieces by definition' (TLP 3.261).
Once we have names that stand proxy for objects there is nothing left to analyse, for
objects in elementary states of affairs are indescribably bound together. We cannot
even begin to imagine what it would be to picture such 'objects' in isolation from
some particular state of affairs or other. At best we could enumerate all the different
states of affairs of which an object might possibly form part. This would be to picture
that which is essential to the full range of possibilities of an object and which could
not be determined in advance. This is why analysing our everyday language would at
best reveal the unspeakable essence of the world.
It is conditional on a proposition’s having sense, whether it is true or false, that only
its most basic components name simple objects. Ordinary propositions and the names
within them are clearly not elementary in this respect since the very existence of their
component objects can be sensibly put into question without loss of sense. Real
names, in contrast, stand directly for objects. For this reason they are indefinable.
Nevertheless, they are the firm, fixed ground upon which language gets its feet. If it
were possible to define them, if they could be further analysed, they would not be
simple and language would have no secure basis. Thus, although his analytic
procedure is clearly reminiscent of Russell’s, it would be a mistake to think that their
ambitions were similar. Wittgenstein did not hope, even in theory, to discover the
basic metaphysical structure and components of the world. He did not regard
philosophy as the most general science. However, this is how he is frequently
misread.6 As Pears writes:

First, he had argued dogmatically and a priori that all language and thought must
have an atomic structure imposed on it by the world. Then when he failed to find
this structure on the surface or on any level of analysis that he actually explored,
he still insisted that it must be discoverable on some deeper level of analysis.
That gave his early theory of language a curiously ambiguous status. (Pears
1988, 206)

As Pears sees it Wittgenstein's 'whole way of doing philosophy is modelled on the

methods of science' (1988, 203). However, if he was interested in such pseudo-
scientific theorising, it ought to have been an embarrassment to him that he never
succeeded in providing any intelligible examples of elementary propositions and
simple objects. Yet he never thought such items were going to come into view as a
result of analysis. It is precisely because objects cannot be understood apart from their
possibilities of combination that it follows that the substance or essence of the world
cannot be described. It is no accident then that he provides no examples of elementary
propositions. The focus of his concern, as I argued above, was rather to clarify
propositions used on particular occasions. We can see this by considering the
frequently discussed example of the watch lying on the table that appears in the
Notebooks, not in the Tractatus. He employs it to show that in clarifying precisely
what we mean on a given occasion we must clearly and unambiguously delineate
what must be the case if our proposition were to be true. In this case it would entail
getting clear about the many, many facts of how things must stand with all the
 12

constituent parts of the watch and the table and all their inter-relations, including their
position in space and time, etc. Thus:

I tell someone ‘The watch is lying on the table’ and now he says: ‘Yes, but if the
watch were in such-and-such a position would you still say it was lying on the
table’? And I should become uncertain. This shows I did not know what I meant
by ‘lying’ in general. ... In order to show that I did know what I meant, I should
say: ‘I know what I mean, I mean just this’ pointing to the appropriate complex
with my finger. And in this complex I do actually have the two objects in a
relation. (NB 70; emphasis added)

The point is that this kind of demonstration serves just as well as the would-be
products of analysis that would specify what we meant in terms of an indefinitely
complicated series of propositions. Ultimately, either would serve to clarify what we
wanted to express by a given proposition on a particular occasion in a way that
nobody would raise concerns on grounds of ambiguity (NB, 19.9.14, cf. Ostrow 2002,
57, and Friedlander 2001). Here we find Wittgenstein going completely against the
idea of analysis, determining, once and for all, a common set of essential but general
properties that comprise the intrinsic nature of objects. He is explicit about this:

It is impossible, however, to assert by means of propositions that such internal

properties and relations obtain: rather this makes itself manifest in the
propositions that represent the relevant state of affairs and are concerned with the
relevant objects. (TLP 4.122)

As he sees it, since we can only make sense of individual propositions on particular
occasions of use, it is not possible to capture what is essential to them in a way that
can be articulated at all. All such attempts, which constitute the bulk of traditional
metaphysics, are doomed to failure from the outset. He sees the implications for the
ends, limits and method in philosophy completely otherwise.

A particular mode of signifying may be unimportant but it is always important

that it is a possible mode of signifying. And that is generally so in philosophy:
again and again the individual case turns out to be unimportant, but the
possibility of each individual case discloses something about the essence of the
world (TLP 3.3421; emphasis added)

Rather than analysis resulting in a set of intelligible propositions that describe the
essence of the world – whether ineffable or otherwise – it functions to clarify our
thoughts on particular occasions. Its ultimate achievement, if successful, is
elucidation. Thus:

The meanings of primitive signs can be explained by means of elucidations.

Elucidations are propositions that contain the primitive signs. So they can only
be understood if the meanings of those signs are already known. (TLP 3.263,
emphasis mine)

Yet again, the point is that the activity of making sense (or failing to do so) and our
evaluation of this cannot be settled in advance because there is no general schema that
permits it. There is no a priori order of things upon which such determinations could
be based. This is why, in line with his views on the nature of judgement, philosophy
is necessarily an activity. But this is also entirely consistent with the idea that
language is the totality of propositions that express thoughts.
 13

With this in hand, it is possible to re-assess Pears’ best evidence for thinking that
Wittgenstein was originally committed to a form of pseudo-scientific analysis.7 It
stems from a conversation recorded by Waismann, in which he said of the Tractatus:
There is another, much more dangerous error, that permeates my whole book –
namely, that conception that there are questions for which answers would be
discovered at a later date. Although one does not have an answer, one thinks that
one has the method by which an answer can be found. I for one thought that it
was the task of logical analysis to discover the elementary propositions. I wrote
that we are unable to specify the form of elementary propositions, and that was
quite correct too. It was clear to me that there are no hypotheses here and that
one cannot proceed with these questions the way Carnap did. But I did think that
one would be able to specify the elementary propositions later (WVC pp. 182-
183)

Yet there is a way of reading these remarks that is in line with the idea that the
elucidation of elementary propositions was meant to provide clarification of what is
meant on particular occasions of use as opposed to an attempt to provide any general
results. Even though he talks of an early commitment to the idea of a postponed
discovery, this alone hardly constitutes a reason to think he was initially inclined to
think of philosophy as having a Russellian scientific task. Indeed, even in the quoted
fragment recorded by Waismann he openly rejects, yet again, the idea that analysis
could result in revealing a fixed set of such touchstone propositions, detailing what is
both general and necessary. Rather, what he thought lay hidden within our ordinary
propositions was the deep logical form that is the foundation of all sense-making; but
although this could be revealed it could not be articulated. Nor should it surprise us
that the very need to engage in analysis with respect to ordinary propositions is put
into question if:

In fact, all the propositions of our everyday language, just as they stand, are in
perfect logical order. (TLP 5.5562)

I suggest that this is what he later came to recognise: that there is no need to elucidate
what we mean in ordinary cases in which we are already making sense, when
language is functioning properly. This is what lies behind his rejection of the idea that
there is anything to be discovered in philosophy. For in his discussion with Waismann
he goes on to say: 'What I want to oppose here is the false idea that we could hit upon
something that today we do not see, that we can discover something entirely new.
That is a mistake. In truth, we already have everything; in fact it is present, so that we
do not have to wait for anything. We move within the realm of something already
there, the grammar of our accustomed language' (WVC 182-183). That he could not
see this at the time of writing the Tractatus was, in part, because his vision was
clouded by the idea that all language served a uniform representational function. Yet,
this was something he came to revise in precisely the way that enabled him to make
sense of the saying/showing distinction in more appropriate, metaphorical terms as
his understanding of what constitutes 'ordinary speech' broadened, in just the way
Stroll denies. Stroll also tries to establish that only in the later writings did he
abandon the idea that language had a limit. However, in doing so, Stroll conflates the
idea that language has a single, general function, e.g. to represent, with the idea of it
being in some way limited. Only the first view was held in the Tractatus, not the
latter. Yet, if we recognise that he changed his view with respect to the former there is
no bar to thinking that the later emphasis on metaphorical language was his new,
 14

'down to earth', way of expressing what he had earlier relegated to the mystical (see
Lippitt and Hutto, 1998, 274-275).
Crucially, we can see from this that even in the Tractatus he was committed to the
idea of what might be awkwardly labelled 'heterogeneous foundationalism'. For in
both periods he held that at the bottom of our ordinary practices is a foundation that
admits no justification or intelligible description. This undercuts Stroll’s primary
ground for identifying a radical break between the early and the late period. For, like
Pears, he too identifies the tack taken in the Tractatus with a kind of scientism and
claims that 'the major difference lies in his disavowal of an Ideal Language for
dealing with philosophical problems. All sorts of important consequences flow form
that decision. One of those is his rejection of the contrast between what can be said
and what can be shown' (Stroll, this volume, p. 20, emphasis added).

Conclusion

In deploying the notions of logical form and form of life, Wittgenstein was always
clear that despite occupying this fundamental position, neither could be deployed for
the purpose of making philosophical pronouncements about 'what there is' or 'what
there must be'. Both remain outside the scope of the explicable, sayable or articulable,
and in no circumstances could the limits they embody be stated or positively charted.
Just as there was no way of charting the limits of logic independently of particular
acts of sense-making, there is no understanding of our form of life from some
theoretical 'hide' outside the activities and practices that it makes possible. There is no
getting behind, beneath or above our form of life for the philosophical purpose of
providing explanation or justification. Just as logical form could not be represented,
therefore precluding the possibility of determining the limits of what we might
encounter in the world in advance of inquiry, so too our form of life cannot be
explicated and thereby provide a boundary, since it too is the ground of all inquiry, of
all saying and doing. Thus we cannot anticipate the limits of logical form or grammar
– we can only explore them from within. Wittgenstein's central insight, early and late,
is that there can be no transcendental setting of limits to sense in advance, or once and
for all. It is for this reason that philosophers go wrong in attempting to theorise and
give explanations of topics relating to these fundamental matters; instead, they must
merely describe and be vigilant of transgressions of sense, occasion by occasion. We
can see a direct reference to Wittgenstein's early view, in a variant of the famous
remark in the Investigations in which we are told that forms of life are given. He
writes:

Instead of the unanalysable, specific, indefinable: the fact that we act in such-
and-such ways, e.g. punish certain actions, establish the state of affairs thus and
so, give orders, render accounts, describe colours, take an interest in others’
feelings. What has to be accepted, the given – it might be said – are facts of
living [forms of life]. (RPP I, 630; PI p. 226)

Against the backdrop of these considerations, although it is right to think of this

transition as marking a shift from a focus on 'reality' to a greater attention to 'human
practices', there is a real danger in understanding what this embodies in too simplistic
terms and in a way that obscures the continuity in his overall method. Certainly, it is
 15

not a shift from realism to idealism as some commentators hold (see Hutto 1996). We
can better understand the nature of this shift by reminding ourselves that Wittgenstein
initially held that linguistic activity was to be understood in terms of particular
instances of modelling reality. This is what lay behind his view of the general form of
the proposition and what drove his early views concerning the truth-conditional and
determinate character of all sense-making. In his later period a more liberal, multi-
functional understanding of grammar replaced the idea that the combinatorial
possibilities of objects underpin the very substance of the world, and formed the basis
of a different vision of what underwrites essence.
Seen in this light, it would be a mistake to regard On Certainty as marking a unique
turn in Wittgenstein’s thinking, if this is meant to be constituted by a hitherto absent
focus on the importance of our fundamental activities. To the extent that it is right to
say he espoused foundationalism at all, it can be found even in his earliest writings.

NOTES

1
Sometimes, as in the case of Pears, theoretical and doctrinal readings are combined to form
an unholy alliance.
2
Thus, as Pears sees it, the important thing to keep in mind, 'is that he is moving away from
theorizing and towards plain description of the phenomenon of language' (Pears 1988b, 218).
3
Although Stroll tends to employ the definite article when talking of language games it is
clear from what he writes that these are multifarious.
4
Although I believe that attributing what seems to be a position to Wittgenstein is to court
confusion, he made a similar comparison himself when he wrote, 'So I am trying to say
something that sounds like pragmatism. Here I am thwarted by a kind of Weltanschauung'
(OC 422).
5
There is not enough space in this essay to properly explore and evaluate Stroll’s claim that
even in the Investigations Wittgenstein was still primarily concerned with the literal (see
Stroll, this volume, , p. 17), though I admit to an initial skepticism.
6
As Pears remarks, 'The theory advanced in the Tractatus belongs to a type that has had a
long vogue. The general mark of the type is the idea that the world imposes a fixed structure
on our thought' (Pears 1988, 206).
7
Similarly and by appeal to the very same evidence, Schulte maintains that, despite himself
and what he wrote in Tractatus, Wittgenstein did not take enough heed of his own
methodological scruples and failed to sufficiently distinguish the nature of the problems of
science from those of philosophy (Schulte 1992, 41-42). But this would have been a quite
incredible oversight given the amount of attention that he gives to contrasting the two
domains.