WAB: "Fragments" | The following contribution is an excerpt from P.M.S. Hacker: "Of knowledge and of knowing that someone is in pain", in: Alois Pichler and Simo Säätelä (eds.): Wittgenstein: The philosopher and his works, Working Papers from the Wittgenstein Archives at the University of Bergen no. 17, Wittgenstein Archives at the University of Bergen (WAB) 2005, pp. 203-235. Publication on WAB's website with kind permission from the author and the editors (2005.3.10).

P.M.S. Hacker: Of knowledge and of knowing that someone is in pain

First person authority: the received explanation

Over a wide range of psychological attributes, a mature speaker seems to enjoy a defeasible form of authority on how things are with him. The received explanation of this is epistemic, and rests upon a cognitive assumption. The speaker’s word is authoritative because when things are thus-and-so with him, then normally he knows that they are. This is held to be because the speaker has direct and privileged access to the contents of his consciousness by means of introspection, conceived as a faculty of inner sense. Like perceptual knowledge, introspective knowledge is held to be direct and non-evidential. Accordingly, the first-person utterances ‘I have a pain’, ‘I believe that p’, ‘I intend to V’ are taken to be descriptions of what is evident to inner sense. Many classical thinkers held such subjective knowledge to be not only immediate, but also infallible and indubitable.

The challenge to the received conceptions came from Wittgenstein. He denied the cognitive assumption, arguing that it cannot be said of me at all (except perhaps as a joke) that I know that I am in pain. For what is that supposed to mean – except perhaps that I am in pain? (PI § 246) If it makes no sense to say that one knows that one is in pain, then the epistemic explanation is a non-starter, since it explains the special authoritative status of a person’s avowal of pain by reference to the putative fact that the subject of pain knows, normally knows, or cannot but know, that he is in pain when he is. It is important to note that Wittgenstein did not mechanically generalize the case of pain across the whole domain of first-person utterances. The case of pain constitutes only one pole of a range of such utterances. Avowals and averrals of belief and intention approximate the other pole, and require independent analysis and grammatical description. Moreover, one must not allow the cases of the emotions and motives, which lie between these poles and where self-deception and lack of self-understanding is common, to overshadow the rest and blind one to the distinctive features to which Wittgenstein drew our attention.

Wittgenstein agreed with the philosophical tradition that being in pain is incompatible with doubting that one is, but pointed out that it is not merely that I do not doubt, I cannot doubt, i.e. doubt is logically excluded. There is no such thing as being in pain and doubting that one is – just as there is no such thing as castling in draughts. The form of words ‘A is in pain, but he doubts whether he is’ is not the expression of a false proposition, but is nonsense – i.e. no sense has been assigned to it. So it is excluded from the language. If someone were to say ‘Perhaps I am in pain, but I rather doubt it’ or ‘Maybe I have a pain, I’m not really sure’, we would not know what he meant. And since doubt is logically excluded, so too is certainty, for certainty presupposes the possibility of doubt.

In place of the received epistemic explanation, Wittgenstein proposed a grammatical elucidation. Rather than explaining why a person’s word carries the kind of weight it normally does when he avows that he is in pain, avers that he thinks this or that, declares his intention to do such-and-such, he sought to describe the grammar of such utterances, the distinctive features of their use, their compatibilities and incompatibilities with other assertions, and the epistemic operators (such as ‘I know’/‘He knows’, ‘I believe’/‘He believes’, ‘I doubt whether’/‘He doubts whether’) which they do or do not accept. Correctly locating such utterances in the web of our concepts, he thought, would obviate the apparent need for philosophical explanations by rendering perspicuous the conceptual structures involved.

Such first-person psychological utterances, Wittgenstein argued, are, in the primitive language games out of which their use arises, essentially expressive, not descriptive. It is, he held, a mistake to construe the characteristic avowal of pain, e.g. ‘It hurts’, ‘I have a pain’ or ‘I have toothache’, the typical utterance of belief, e.g. ‘I believe she is in the garden’, and of intention, e.g. ‘I’m going to V’ or ‘I intend to V’, as descriptions of myself or of my state of mind. On the contrary, they are characteristically expressions or avowals. They are authoritative (to the extent that they are) not because they are assertions of something the agent knows, but because they are manifestations of the agent’s feeling, thinking or intending whatever he feels, thinks or intends. So their truthfulness normally guarantees their truth. Hence, the ‘authority’ they have is, in certain cases, akin to the evidential authority of expressive non-linguistic behaviour. They are logical criteria for the ‘inner’. The cry of pain, in circumstances of injury, is not a sign of pain, which has been discovered in experience to be inductively well-correlated with it. It is a manifestation of pain, and a logical criterion, not an inductive symptom, of pain. Utterances of pain, e.g. the exclamation ‘It hurts’ or the groan ‘I am in pain’, have the same criterial status, for they are acculturated extensions of natural pain-behaviour. (Avowals of occurrent passions, such as ‘I am angry’, ‘I am afraid’, ‘I am so pleased’ approximate this ‘pole of description’, being, in the primitive language-game, extensions of the snarl of anger, the cry of fear, and the exclamation of delight. But the primitive language-game extends to averrals of emotional disposition, and the story there gets more complex and nuanced.) To say that such utterances are acculturated extensions of natural expressive behaviour does not imply that they are just like the primitive behaviour on to which they are, as it were, grafted. On the contrary – unlike the natural behaviour such linguistic behaviour can be truthful or dishonest. What is said by such utterances may be true or false, no less than the third-person counterparts. It stands in logical relations of implication, compatibility or incompatibility with other propositions.

In other cases, for example of thought, belief, expectation, suspicion, etc., the first-person utterance ‘I think (believe, expect, suspect, etc.) that p’ is not grafted on to natural expressive behaviour, but on to forms of linguistic behaviour that have already been mastered, viz. the use of an assertoric sentence ‘p’, which, uttered sincerely, may, in appropriate contexts, express one’s beliefs, opinions, expectations, suspicions, guesses, etc. In such contexts, the utterance of the sentence ‘p’ is a criterion for ascribing to the speaker the belief or opinion, etc. that p. So too is the utterance of the sentence ‘I believe (think) that p’, which is commonly an explicit expression of the belief that p and not an expression of the speaker’s knowledge (or belief) that he believes that p. In avowing that he believes that p, a speaker is also endorsing what he believes. He is answerable, if not to others then to himself, for his beliefs. If he has beliefs for which he can find no reason and which he knows or thinks are contrary to reason, they are not so much beliefs as obsessive thoughts, fantasies and imaginings of which he cannot rid himself. (While avowals of occurrent emotion approximate the case of pain in certain respects, avowals and averrals of emotional disposition approximate the case of belief in other respects.)

Of course, both utterances such as ‘I am in pain’ and utterances of the form ‘I think (expect, suspect, etc.) that palso have a use as statements or reports. ‘I think that p’ is typically an avowal or expression of belief or opinion, but it can also be used as an autobiographical admission, confession or statement. However, the first-person statement that things are thus-and-so with me shares many of the logico-grammatical expressive features of the more primitive utterance from which it grows. So, for example, my statement that I believe that p is nevertheless still an expression of my belief that p, in as much as in stating that I so believe, I am still endorsing the proposition that p (which is why I cannot say ‘I believe that p, but actually it is not the case that p’ or ‘I believe that p, but whether it is the case that p is an open question as far as I am concerned’). As always, generalization is perilous, and different cases must be examined in their own right. In particular, one must not extrapolate from avowals of pain to avowals of belief and related doxastic verbs or to avowals of intention and related volitional verbs, but investigate each case separately. So too, expressions of emotion and statements of motive are separate cases for treatment, which will not be ventured here.

It might well be said that, in view of Wittgenstein’s expressive elucidation of why a person’s avowal of how things are with him carries special weight, the term ‘authority’ is a misnomer. I am not an authority on how things are with me, as I might be an authority on renaissance painting. Rather, in the absence of defeating conditions, my word goes – it is a (defeasible) criterion for others to judge that things are thus with me. The ‘authority’ in question is not cognitive authority, but more akin to verdictive authority (in the case of belief) and executive authority (in the case of decision and intention). So the very term ‘first-person authority’ is misleading. This is, I think, correct. So although I shall use the received term occasionally, this qualification should be born in mind.

Wittgenstein’s arguments are often misunderstood and his conclusion has not won much support.1 My purpose is to elaborate his account and to defend the rejection of the cognitive assumption as an explanation of so-called first-person authority. First, I shall try to elucidate the contour lines of the concept of knowledge and to adumbrate some of its relations to adjacent concepts in its semantic field. This is necessary to evaluate the plausibility of the cognitive assumption and hence of the epistemic explanation. I shall then examine the rather special case of pain (and, by implication, of other sensations and arguably mental images). Belief (and related doxastic predicates) and intention (and related volitional predicates) will be discussed elsewhere.




1A recent examination of his views concludes that ‘the expressivist proposal … is a dead duck …’ (Crispin Wright, ‘Self-knowledge: the Wittgensteinian Legacy’, in Anthony O’Hear ed. Current Issues in Philosophy of Mind, Royal Institute of Philosophy Supplement: 43 (Cambridge University Press, Cambridge, 1998), p. 115). If it seems so, I shall suggest, it is because it has been misunderstood. Part of the misunderstanding stems from an unwarranted extrapolation from the case of avowals of pain to other avowals, e.g. of belief or intention, an extrapolation which Wittgenstein was careful to avoid.