This is the final, pre-publication draft. Please cite only from published paper in Hinge Epistemology, special
issue of the International Journal for the Study of Skepticism, 6 (2016), 97-119.
The Animal in Epistemology:
Wittgenstein's Enactivist Solution to the Problem of Regress
Danièle Moyal-Sharrock
University of Hertfordshire
d.moyal-sharrock@herts.ac.uk
Abstract: In this paper I briefly summarize the nature of Wittgenstein's 'hinge certainties',
showing how they radically differ from traditional basic beliefs in their being nonepistemic,
grammatical, nonpropositional and enacted. I claim that it is these very features that enable
hinge certainties to put a logical stop to justification, and thereby solve the regress problem of
basic beliefs. This is a ground-breaking achievement – worthy of calling On Certainty
Wittgenstein's 'third masterpiece'. As I go along, I question some differing interpretations and
respond to some objections from fellow-readers of On Certainty: Duncan Pritchard, Michael
Williams and Crispin Wright.
Keywords: Wittgenstein, On Certainty, epistemology, regress problem, basic beliefs, hinge
certainty, knowledge, Enactivism.
*********
The first, and only time, I heard Crispin Wright speak on On Certainty was at Kirchberg in
2003. At the Q&A session, I asked him: 'What have you done with the animal in On
Certainty'? And his reply was: 'There is no animal in epistemology'. Well, Wright might want
to keep the animal out of epistemology, but Wittgenstein certainly didn't, and I think he was
right:
I want to conceive [this certainty] as something that lies beyond being justified or
unjustified; as it were, as something animal. (OC 359)
Wittgenstein's revolutionary insight in On Certainty is that what philosophers have
traditionally called 'basic beliefs' – those beliefs that all knowledge must ultimately be based
on – cannot, on pain of infinite regress, themselves be propositional beliefs. They are really
animal or unreflective ways of acting which, once formulated (e.g. by philosophers), look like
propositional beliefs. It is this misleading appearance that leads philosophers to believe that at
the foundation of thought is yet more thought. For, though they often resemble empirical
conclusions, basic certainties (or 'hinge certainties' or 'hinges' – as I shall also call them
following Wittgenstein's hinge metaphor [OC 341]) constitute the ungrounded,
nonpropositional underpinning of knowledge, not its object. In thus situating the foundation
of knowledge in nonreflective certainties that manifest themselves as ways of acting,
Wittgenstein has found the place where justification comes to an end, and solved the regress
problem of basic beliefs – and, in passing, shown the logical impossibility of radical or global
scepticism. I believe that this is a groundbreaking achievement for philosophy – worthy of
calling On Certainty Wittgenstein's 'third masterpiece'. In this paper, I briefly flesh out the
2
account of On Certainty that I have just summarized while considering some objections and
differing interpretations from fellow-readers of On Certainty.
1. Certainty versus knowledge
Wittgenstein's last notes were posthumously entitled On Certainty. They constitute his
attempt, prompted by G. E. Moore's 'Proof of an External World', to understand the nature of
our basic assurance – our assurance about such things as 'Human beings are born and die',
'The earth has existed long before I was born', 'I am standing here', 'I have a body', 'Here is a
hand'. Wittgenstein agrees that the objects of Moore's assurance are those of our most
unquestionable beliefs, but disagrees that the certainty in question is of an epistemic nature;
he believes this assurance to be of a more foundational breed than knowing:
When I say 'how do I know?' I do not mean that I have the least doubt of it.
What we have here is a foundation for all my action. But it seems to me that it is
wrongly expressed by the words 'I know'. (OC 414)
Wittgenstein does not take this certainty to be a knowing because he adheres to the standard
view of knowledge as justified true belief: 'One says 'I know' when one is ready to give
compelling grounds. 'I know' relates to a possibility of demonstrating the truth' (OC 243);
'Whether I know something depends on whether the evidence backs me up or contradicts me'
(OC 504; see also OC 91). Basic certainty, as he sees it, has no truck with truth or
justification: 'If the true is what is grounded, then the ground is not true, nor yet false' (OC
205). To say that basic certainties are indubitable is not to say that they have been proved true
beyond doubt, but stems from their not being susceptible of confirmation or falsification at
all; from their being logically impervious to doubt. At some point, justification and doubt lose
their sense; there where the spade turns is the ungrounded ground, where 'justification comes
to an end' (OC 192).
2. Conceptual, not contextual
Our basic certainties are ungrounded; we do not hold them because we have compelling
grounds for them (OC 243); nor are they objects of subjective or psychological conviction,
but of logical conviction (OC 494). That is, doubting them would be tantamount to having
lost the bounds of sense: 'If someone said to me that he doubted whether he had a body I
should take him to be a half-wit' (OC 257); 'I cannot doubt this proposition without giving up
all judgment' (OC 494); 'There are cases where doubt is unreasonable, but others where it
seems logically impossible' (OC 454)
Nor is it possible to be mistaken about a hinge certainty: if I believe that I am sitting
in my room when I am not or that my biological parents are lions, it isn't my possibly being
mistaken that would be under investigation, but my knowledge of English or, more sadly, my
sanity. A mistake results from negligence, fatigue or ignorance; we cannot say of someone
who believes that she was never born that she is 'mistaken':
3
In certain circumstances a man cannot make a mistake. … If Moore were to
pronounce the opposite of those propositions which he declares certain, we should
not just not share his opinion: we should regard him as demented. (OC 155)
In thus logically closing the door to doubt and mistake as regards our basic certainties,
Wittgenstein closes the door to the contextualism Michael Williams attributes to him (1991,
26). For Wittgenstein, it isn't, as Williams suggests, that the justificatory process need not
actually occur (though grounds must be produced on demand), or that it need not be self-
conscious (2001, 35), but that basic certainty is conceptually groundless, groundless by
nature: 'I want to conceive it as something that lies beyond being justified or unjustified' (OC
359). If our certainty stems, or could stem, from justification, it is not a hinge certainty (OC
94).
On Williams' neo-Humean reading, Wittgenstein believes sceptical doubt to have no
bearing in the pragmatic air of ordinary life, but to be legitimate and serious in the context of
philosophical reflection. However, as we've just seen, there can be for Wittgenstein no
context in which our hinge certainties can be doubted or justified, for their indubitability is
conceptual, not contextual. It isn't a hinge that can be doubted in some contexts and not in
others; but a hinge can never be doubted whereas the doppelgänger of a hinge (that is, a twin
sentence which expresses an empirical or an epistemic proposition) can be, and this misleads
us into thinking that the hinge itself can, in some contexts, be doubted
1
. 'The world has
existed for more than five minutes': this sentence may well constitute a falsifiable or
empirical proposition in a sci-fi movie, but it cannot be a falsifiable proposition, and therefore
(on Wittgenstein's view) a proposition tout court, for a human being in any context
2
. For
Wittgenstein, the Cartesian demon is never a plausible threat; he has no more grip in the
philosophical study than he does in our ordinary life.
What may give the impression of contextualism is the difference Wittgenstein marks
in On Certainty between the use of 'I know...' in ordinary life and its use in philosophical
discourse:
1
Note that the notion of doppelgänger helps us to consider the use of identical sentences in different contexts
without any sense or meaning being intrinsically attached to these sentences. As Pritchard rightly notes: '... there
is nothing in Wittgenstein's argument ... which licenses the kind of contextualism that Williams proposes'
(2011a, 16).
2
So it isn't that the same proposition becomes unfalsifiable in one context; but that qua hinge the sentence does
not express a proposition at all (it expresses, as we shall see, a rule of grammar or rule of a language-game (OC
95) or bound of sense). The same goes for personal hinges (for a taxonomy of hinges listing them as linguistic,
personal, local and universal, see Moyal-Sharrock (2007)): my speaking French is a personal hinge, a certainty
which is for me neither falsifiable nor verifiable; the fact that someone may say of me that I speak French
(because they've heard me speak it) and that this is a falsifiable proposition for them (I could have been aping a
French speaker) does not make it a falsifiable proposition for me. Or again, what is for me a hinge certainty
(e.g., 'I am in this room') can constitute a verifiable fact for someone else (e.g., who is not in this room or who is
blind), but this does not make my hinge certainty into a verifiable empirical proposition. There is a first-third
person asymmetry here that we must be wary of. Relatedly, to take hinge certainties to be unfalsifiable empirical
propositions (Hacker (1996, 217); see also Glock in this issue), or judgments that have a normative role (Coliva,
2010, 80; and in this issue) is, I suggest, to attribute to hinges uses that belong to their Doppelgänger. Moreover,
the idea of unfalsifiable empirical propositions, at best, takes us back to the notion of 'necessary propositions'
that Wittgenstein took great pain to reform (see, e.g. MWL 61-65).
4
What I am aiming at is also found in the difference between the casual
observation 'I know that that’s a...', as it might be used in ordinary life, and the
same utterance when a philosopher makes it. (OC 406)
But Wittgenstein suggests we treat these knowledge claims differently not because he thinks
we know our basic certainties in ordinary life and not in the philosopher's study, but because
Moore's being a philosopher ought to constrain him to use 'I know...' with technical precision
– that is, exclusively in cases of justified true belief – whereas this cannot be demanded of the
ordinary person: we cannot and should not expect her to use 'I know...' only when it is
'justified true belief' she means by it. Wittgenstein refuses to admonish or correct our
ordinary use of language, but the philosopher must be made accountable:
For when Moore says 'I know that that’s a...' I want to reply 'you don’t know
anything!' – and yet I would not say that to anyone who was speaking without
philosophical intention. (OC 407; original emphasis)
So that, for Wittgenstein, we are entitled to say: 'I know...' in ordinary life where a
philosopher in the study is not thus entitled, but this does not imply that we do in fact know in
ordinary life.
It isn't, as Williams claims, that sceptical doubts are unnatural doubts (1991, 2), and
therefore sustainable only in the artificial or unnatural conditions of philosophical reflexion,
but that they are not doubts at all. In On Certainty, Wittgenstein elucidates the concept of
doubt in two important ways: he shows that universal doubt is impossible and that not
everything that has the appearance of doubt is doubt:
If someone said that he doubted the existence of his hands, kept looking at
them from all sides, tried to make sure it wasn't 'all done by mirrors', etc., we
should not be sure whether we ought to call that doubting. We might describe his
way of behaving as like the behaviour of doubt, but his game would not be ours.
(OC 255)
In some cases, what looks like doubt is only doubt-behaviour. Of course, where doubt has no
rational motivation or justification, it may have (pathological) causes (OC 74), but normal
doubt must have reasons. It isn't enough to say or imagine we doubt: genuine doubt, like
suspicion, must have grounds (OC 322, 458). If Williams thinks sceptical doubt possible, it
may be because – like Hume, Moore and most philosophers since Descartes – he takes the
mere articulation of doubt for doubt.
Wittgenstein's recognition that the sceptic's doubt is only doubt-behaviour is spurred
by his realization that it is hinged on the very certainties it dismisses. For, were she not
hinged on some certainties, the sceptic could not even formulate her doubt:
If I wanted to doubt whether or not this was my hand, how could I avoid
doubting whether the word 'hand' has any meaning? So that is something I seem
to know after all. (OC 369) ['know' is in italics]
But more correctly: The fact that I use the word 'hand' and all the other words
in my sentence without a second thought, indeed that I should stand before the
abyss if I wanted so much as to try doubting their meanings – shews that absence
5
of doubt belongs to the essence of the language-game, that the question 'How do I
know...' drags out the language-game, or else does away with it. (OC 370)
The possibility of meaningful thought (language) demands certainty: this means that the
sceptic's scepticism depends on certainty for its very conception or formulation. At 'the
foundation of all operating with thoughts (with language)' (OC 401) is an essential certainty,
a certainty endorsed every time a doubt (towards it) is formulated. Its being essential to our
making sense means that this certainty underpins all our questions and doubts (OC 341),
including the sceptic's universal doubt, thereby invalidating it. It seems to me that what we
have here is a knock-down objection to universal scepticism.
Although Hume may be seen to have progressed from Descartes when he admits that
sceptical doubt is not sustainable in ordinary life, it takes Wittgenstein to recognize that
universal doubt is not sustainable at all, inside the study or out – and this, not for pragmatic
but for conceptual reasons: 'A doubt that doubted everything would not be a doubt' (OC 450);
'If you tried to doubt everything you would not get as far as doubting anything. The game of
doubting itself presupposes certainty' (OC 115). Wittgenstein demystifies sceptical doubt; he
shows that the sceptic – and any philosopher who, in the study, takes universal doubt
seriously – are only under, what he calls, an illusion of doubt (OC 19).
In On Certainty, then, Wittgenstein takes Moore to task for confusing knowledge with
the nonepistemic brand of conviction that logically underlies it, and he drives a categorial
wedge between them: '"Knowledge" and "certainty" belong to different categories' (OC 308).
In doing this, he breaks with the traditional presupposition in epistemology that we know our
basic beliefs. But in spite of Wittgenstein's categorial distinction, some commentators insist
on seeing the certainty which underpins knowledge as itself a knowing. As Williams puts it:
'Knowledge ... emerges out of prior knowledge' (2001, 176), which makes our 'bedrock
certainties' unavoidably propositional for, he asks, how could our basic beliefs not be
propositional if they are to generate our nonbasic beliefs (ibid., 97)
3
?
Yet the clear message of On Certainty is precisely that knowledge does not have to be
at the basis of knowledge. For Wittgenstein, underpinning knowledge are not default justified
propositions that must be susceptible of justification on demand but nonpropositional
certainties – certainties 'in action' or ways of acting – that can be verbally rendered for
heuristic purposes and whose conceptual analysis uncovers their function as unjustifiable
rules of grammar. So that basic certainties stand to nonbasic beliefs, not as propositional
beliefs stand to other propositional beliefs, but as rules of grammar stand to propositional
beliefs
4
. Hence the absence of propositionality as regards them. Knowledge need not emerge
3
Williams: 'However basic knowledge is understood, it must be capable of standing in logical relations to
whatever judgements rest on it. For example, it must be capable of being consistent or inconsistent with them.
But this means that even basic knowledge must involve propositional content' (2001, 97).
4
This is the logical relationship between hinges and non-hinge propositions that Williams demands: that
between rules and propositions. For most epistemologists, to ask for the ground of a belief is to ask for yet
another belief, for only beliefs (and other intentional states) can stand in logical relation to other beliefs. But this
is to overlook the fact that grammatical rules can stand in logical relation to beliefs – as indeed they must: as the
necessary enablers of sense. I cannot come to the belief that 'It is indeed a hand I see on this blurry photograph'
unless I am 'hinged' on the grammatical rule that 'This is what we call "a hand"'. In the same way that I cannot
investigate the age of the earth unless I am 'hinged' on such grammatical rules as 'The earth exists'; 'The earth
has existed for a long time'; 'We have a concept of time, and it includes measurement of age in years' etc.