1. The Blue Book Passage

In considering Wiseman’s case for her interpretation, it will help to have the Blue Book passage before us. I quote it in full.

Wiseman highlights two key passages in support of her contention that the Blue Book passage does not endorse a distinction between uses of ‘I’ with respect to whether they involve the recognition of a person and whether they are vulnerable to error through misidentification. I consider them in turn. I shall argue, against Wiseman’s reading, that there is nothing in what Wittgenstein says to suggest that he does not think there is an important distinction to be drawn in those terms between uses of ‘I’.

The first passage comes near the beginning of Wittgenstein’s discussion. Having said that there are ‘two different cases’ in the use of ‘I’, and given his two lists of examples, Wittgenstein says this:

One ‘can’ point to the difference between the two cases like that. But, Wiseman says, that invites the question, ‘but ought one? What is the pointing for? What purpose is it to serve?’ (Wiseman Reference Wiseman2019: 668). And, she claims,

Against the standard reading, she offers this alternative suggestion:

What should we make of these points?

First, it is not true that Wittgenstein’s way of pointing to the difference between the as-object and as-subject uses of ‘I’ ‘says nothing about the second category – only the first’. Nor is it true that the idea that the as-subject use does not involve the recognition of a particular person ‘is supplied by’ the standard reading, rather than by Wittgenstein himself. Wittgenstein introduces the points about recognition of a person and the possibility of error in connection with the use of ‘I’ as object. But he immediately goes on to make the contrasting points about the use of ‘I’ as subject: ‘there is no question of recognizing a person when I say I have toothache’; and ‘when in this case no error is possible, it is because the move we might be inclined to think of as an error … is no move of the game at all’ (Wittgenstein Reference Wittgenstein1969a: 67). So the application of the points about recognizing a person and the possibility of error to the as-subject case is not an artefact of the standard interpretation. It is present in Wittgenstein’s text. Of course, Wiseman might still be right that his point is to warn against drawing the distinction in these terms rather than to recommend it. But we should be clear about what Wittgenstein actually says, however we go on to interpret it.

Second, it is certainly important to be sensitive to the distinction between remarks that express Wittgenstein’s own views and remarks that appear in some other voice. However, if something appears in the text merely as a possible view, and something else is offered as a corrective, we should expect to be able to identify which is which; we should be able to tell where the presentation of a tempting but misleading idea ends and where Wittgenstein’s corrective begins. But as far as I can see, there is no such transition in the Blue Book passage. There are many points in the Blue Book at which, having articulated something that we are inclined or tempted to say, Wittgenstein goes on to say explicitly that it involves a misunderstanding, a delusion, a confusion, or something similar. But there is no such statement in the current passage.Footnote ³ Nor is there any indication that Wittgenstein does not endorse the points he describes about the recognition of a person and the possibility of error. Thus, having introduced the idea that when ‘I’ is used as object there is ‘the possibility of an error’, he goes on to rephrase that idea in his own preferred way: ‘or, as I should rather put it: the possibility of an error has been provided for’. That characteristically Wittgensteinian formulation seems a clear indication that he is endorsing the idea.Footnote ⁴ Similarly, his analogy with the kinds of error that are allowed for in ‘a pin game’ (he means ‘pinball’), seems to be proposed entirely in his own voice. And there is nothing to suggest that he regards the broken-arm case or the bump-on-the-forehead case as anything other than straightforward examples of a kind of error that he does indeed take to be possible.

Turning to the as-subject use of ‘I’, Wittgenstein says that, if I say I have toothache, ‘To ask “are you sure that it’s you who have pains?” would be nonsensical’. That point, again, is one that he seems happy to endorse. It recalls a remark from much earlier in the Blue Book:

That earlier remark is certainly an expression of Wittgenstein’s own view; he thinks the question would indeed be absurd. But it would not be absurd if we held the view of expectation that Russell advances in The Analysis of Mind (Russell Reference Russell1921). Wittgenstein takes that to be a decisive reason for rejecting Russell’s view. Similarly in the present case. Wittgenstein himself thinks it would be ‘nonsensical’ to ask someone who says they have toothache ‘are you sure it’s you who have pains?’.Footnote ⁵ And there is no reason to suppose that he does not also endorse the preceding remark – ‘there is no question of recognizing a person when I say I have toothache’ – or the remark that comes next: ‘when in this case no error is possible it is because the move we might be inclined to think of as an error … is no move of the game at all’. So far as I can see, then, there is nothing in what we have seen so far to suggest that Wittgenstein is not talking in his own voice when he characterizes the difference between the as-object and as-subject uses of ‘I’ in the way he does.

However, Wiseman argues that the next lines of the Blue Book passage give us good reason to challenge the standard reading. (This is the second of the ‘key passages’ I mentioned at the start of this section and referred to in footnote 3.) Wittgenstein says:

Wiseman comments:

There are two points here: one about the ‘clear red flag’ of the way that Wittgenstein introduces the ‘way of stating our idea’; the other about its modal form. I will take them in reverse order.

It is true that the later Wittgenstein thinks that talk of impossibility must be handled with care. But he does not think such talk automatically ‘signals an error’. Error arises only if we try to conceive of what is really a logical or grammatical impossibility on the model of physical or empirical impossibility. In the Blue Book, he explains the point by contrasting the physical impossibility of feeling pain in someone else’s tooth with the logical impossibility of feeling someone else’s toothache (Wittgenstein Reference Wittgenstein1969a: 49ff.). It is, he thinks, ‘conceivable that I feel pain in a tooth in another man’s mouth’. But the way in which human beings are actually constructed means that, as things are, ‘we can’t have ... pains in another person’s tooth’. The impossibility in that case is physical impossibility; Wittgenstein compares it to the impossibility of an iron nail’s scratching glass. The impossibility of feeling another person’s toothache, he thinks, is quite different. What makes a toothache S’s toothache is the fact that it is S who feels it. So any toothache that I feel will be my toothache, by virtue of the fact that it is me who feels it. The reason why it is impossible for me to feel your toothache, then, is not that I am physically (or metaphysically) prevented from doing so, as if by an ‘insurmountable barrier’ that lies between me and your toothache (Wittgenstein Reference Wittgenstein1969a: 55–56). It is, rather, that nothing would count as my feeling your toothache; the possibility of feeling another person’s toothache ‘has not been provided for’. Here, the impossibility is logical or grammatical. Wittgenstein does not object to the claim that it is impossible for one person to feel another person’s pain. What he objects to is only the tendency, when doing philosophy, to think of this impossibility on the model of the physical impossibility of feeling pain in someone else’s tooth.

We should understand the modal talk in the Blue Book passage about ‘I’ in the same way. When I have toothache and judge ‘I have toothache’, it really is impossible for me to be right that someone is in pain but wrong that it is me; as Wittgenstein puts it, the possibility of an error has not been provided for. But we must not misunderstand the nature of this impossibility. It is not as if, when I judge that I have toothache, there is in principle room for me to be wrong about who it is who has toothache, but that my capacity to distinguish my own toothache from other people’s is so acute as somehow to make it impossible for me to make such a mistake. Rather, when I feel toothache and judge that I have toothache, there is nothing that would count as making that sort of error. But as long as we avoid misconstruing the kind of possibility involved, Wittgenstein is perfectly happy with the idea that, when I say ‘I have toothache’, ‘it is as impossible that … I should have mistaken another person for myself, as it is to moan with pain by mistake, having mistaken someone else for me.’ There is, pace Wiseman, nothing intrinsically misleading or non-Wittgensteinian about the modal form of that claim.Footnote ⁶

Nonetheless, I agree with Wiseman that, when Wittgenstein introduces the formulation we have been discussing with the words ‘this way of stating our idea suggests itself’, he is indicating that he does not fully endorse every aspect of what comes next. But what he objects to is not the idea that, when I judge ‘I have toothache’, it is impossible to have mistaken another person for myself. Nor is it the idea that there is an instructive parallel between moaning with pain and saying, ‘I’m in pain’.Footnote ⁷ What Wittgenstein objects to comes at the beginning of the suggested ‘way of stating our idea’, with the phrase ‘it is as impossible that in making the statement “I have toothache” …’ (my emphasis). That is what he picks up in the remark that comes next: ‘To say, “I am in pain” is no more a statement about a particular person than moaning is.’ Wittgenstein clearly thinks there is something wrong with the idea that someone who says ‘I am in pain’ is making a statement about a particular person. There are several points he may plausibly have in mind. First, that ‘I am in pain’ is typically not a statement but, e.g., an expression.Footnote ⁸ Second, that ‘I am in pain’ is not a statement (or whatever) about a particular person but, e.g., a statement (etc.) by a particular person. Third, that ‘I am in pain’ is not a statement (or whatever) about a particular person but a statement (or whatever) about myself.Footnote ⁹ But none of that distances Wittgenstein from the idea that when ‘I’ is used as subject, there is no question of recognizing a person, and therefore no possibility of being wrong about which person has the property that is being ascribed.

To sum up. There is no indication in the Blue Book passage that Wittgenstein is not endorsing the distinction he describes: between a use of ‘I’ that involves the recognition of a person and is vulnerable to the possibility of error through misidentification and a use of ‘I’ that involves no recognition of a person and is immune to errors of that kind.

2. Solipsism and the dialectical role of the Blue Book passage

I turn now to the second of Wiseman’s objections to the standard reading of the Blue Book passage: that it misrepresents the point of Wittgenstein’s discussion because it ignores the context in which it occurs, as part of an extended discussion of ‘personal experience’ and the seductive attractions of solipsism. In that context, she thinks, the point of the passage is to diagnose and reject the temptation to think that the word ‘I’ has two different uses, ‘depending on whether we are talking about RW or the subject [of] experience which “inhabits” RW’ (Wiseman Reference Wiseman2019: 673). It is not, as on the standard reading, to endorse a distinction between two different uses of ‘I’.

It is certainly true that most discussions of the Blue Book passage have ignored its dialectical role and its context in the discussion of solipsism. When we do pay attention to the wider context, however, we see that the standard reading fits the context very well. It makes complete sense of the role of the as-object/as-subject distinction in the passage as a whole. In particular, it provides a very natural reading of Wittgenstein’s point that ‘the idea that the real I lives in my body is connected with the peculiar grammar of the word “I”, and the misunderstandings this grammar is liable to give rise to’ (Wittgenstein Reference Wittgenstein1969a: 66).

The passage about the use of ‘I’ as subject and as object comes in a discussion that aims to diagnose and cure us of the attractions of solipsism. The solipsist, here, is someone who wants to explain or articulate the difference between the case where she is in pain, say, and the case where someone else is in pain. She starts with an expression of the standard epistemic problem of other minds:

That thought quickly leads to the statement of a conceptual problem about other minds:

That drives the solipsist to think that if I ascribe pains and so forth to other people – as we do in ordinary language – the word ‘pain’ must mean something different when applied to other people from what it means when I apply it to myself. So there is something wrong, or at best misleading, about ordinary language, which uses the same word ‘pain’ in both first-person and third-person ascriptions. As the discussion proceeds, Wittgenstein works through a series of ways in which someone who feels the pull of this solipsistic line of thought may try to articulate their supposed insight: ‘only my own experience is real’ (Wittgenstein Reference Wittgenstein1969a: 46), ‘only my pain is real pain’ (Wittgenstein Reference Wittgenstein1969a: 57), ‘only I really see’ (Wittgenstein Reference Wittgenstein1969a: 60), ‘I am the vessel of life’ (Wittgenstein Reference Wittgenstein1969a: 65), and more. But what is the solipsist trying to say when they make such claims? In particular, what do they mean by ‘I’?

Wittgenstein says:

But what makes the solipsist think that there is a way for me to use ‘I’ to talk about something other than the human person, NN? Wittgenstein says that,

And then comes the famous passage that we have been discussing: ‘There are two different cases in the use of the word “I” (or “my”) which I might call the “use as object” and “the use as subject” …’.

In that context, it is very natural to understand Wittgenstein’s point like this. There really are two cases in the use of ‘I’: the use as object, which involves the recognition of a person, and the use as subject, which does not. In distinguishing between these two cases in the way that he does, he is describing features of the ordinary use of ‘I’: part of ‘the peculiar grammar of the word “I”’. But we are easily misled by that ‘peculiar grammar’. For the fact that I can use ‘I’ in a way that doesn’t require the recognition of a particular person can tempt me to think that, in those cases, what I am talking about is not a human person with bodily characteristics but something else: ‘the real I [that] lives in my body’. And two pages later, Wittgenstein articulates exactly that diagnosis of the temptation to think that ‘the real I lives in my body’:

On this reading, Wittgenstein agrees that, ‘in the cases in which “I” is used as subject, we don’t use it because we recognize a particular person by his bodily characteristics’. He treats that as an important feature of the use of ‘I’ in these cases; he does not, as Wiseman suggests, introduce that idea merely in order to warn us against it. He cautions us against being misled by this feature of the use of ‘I’ as subject into the solipsist’s mistake of thinking that such uses of ‘I’ refer to an immaterial entity that inhabits the body. But, he thinks, we do not have to be misled. If we keep a clear head, we can acknowledge that, when ‘I’ is used as subject, we do not identify ourselves by our bodily characteristics, whilst recognizing that what we are talking or thinking about in these cases is an embodied human being.

Read in this way, Wittgenstein’s discussion of the two cases in the use of ‘I’ can be seen to follow a pattern that we find in many places in the Blue Book. He draws attention to some feature of our language. He shows how that feature of grammar can mislead us when we are reflecting philosophically: how it can tempt us into confusions or misunderstandings. Then he shows us how to avoid being misled: how to resist the confusions. Here is an example from another part of the Blue Book:

In this case, the grammatical feature to which Wittgenstein draws attention is the use of the past tense in contexts such as ‘I meant the man who won the battle of Austerlitz’; he calls this ‘a peculiar form of expression of ordinary language’ (Wittgenstein Reference Wittgenstein1969a: 38). We are easily misled by that feature of ordinary use into thinking that, if it is true that someone meant the winner of the battle of Austerlitz when they said ‘Napoleon was crowned in 1804’, there must have been something going on in their mind at the time that made it true. But, he urges, we should resist the temptation ‘to think of meaning or thinking as a peculiar mental activity’ (Wittgenstein Reference Wittgenstein1969a: 39), and adopt the broadly dispositional understanding of meaning or intending that he develops in the Blue Book and elsewhere. The passage about the use of ‘I’ has the same structure.

3. Three interpretative issues

In this section, I address three issues that are highlighted by the contrast between Wiseman’s interpretation and my own. The first two concern further objections to the standard reading. The third poses a challenge for Wiseman’s reading.

First, as Wiseman points out, the standard reading of the Blue Book passage faces a puzzle about the lists of examples Wittgenstein offers of the two cases in the use of ‘I’. His examples of uses of ‘I’ as object all involve physical self-ascriptions: ‘My arm is broken’, ‘I have grown six inches’, ‘I have a bump on my forehead’ and so on. His examples of uses of ‘I’ as subject all involve self-ascriptions of ‘narrowly psychical’ mental properties: ‘I see so-and-so’, ‘I hear so-and-so’, ‘I think that it will rain’ and so on.Footnote ¹⁰ But, on the standard reading, the use of ‘I’ as subject is not confined to its use in self-ascriptions of mental properties. For there are also self-ascriptions of physical properties in which ‘I’ or ‘my’ is used as subject: for example, the judgement, ‘My legs are crossed’, made in the normal case where I can tell that my legs are crossed without needing to look and see how they are arranged; or the judgement, ‘The wind is blowing my hair about’, made when I feel the wind blowing through my hair, without needing to look at my reflection in a mirror. In cases like that, there is no recognition of a person involved and accordingly no possibility of error through misidentification. So the challenge for the standard reading is this. If the use of ‘I’ as subject, as Wittgenstein conceives it, extends to some self-ascriptions of physical properties as well as to self-ascriptions of mental properties, why does his list of examples of the use of ‘I’ as subject include no uses of ‘I’ in physical self-ascriptions?

My response is this. As we have seen, Wittgenstein’s reason for introducing the distinction between as-subject and as-object uses of ‘I’ is to show why, when we reflect philosophically, we are easily tempted to think that the real I lives in the body. But that temptation only arises when we use ‘I’ as subject in mental self-ascriptions; there is no such temptation when ‘I’ is used as subject in a physical self-ascription. When I judge that my legs are crossed, the property I am ascribing to myself – the property of having crossed legs – is a property of an embodied person; an immaterial ego could not have crossed legs. So when I make that judgement in the normal case, using ‘I’ or ‘my’ as subject, there is no temptation to think that ‘my’ ‘refer[s] to something bodiless’; it is evident that the subject of the ascription is an embodied human being, even though no recognition of a person by their bodily characteristics is involved. The reason why Wittgenstein’s discussion focuses on the use of ‘I’ as subject in mental self-ascriptions, then, is that those are the uses that give rise to the temptation to think that the real I lives in the body. For all that, the use of ‘I’ as subject, understood as Wittgenstein describes it, extends to its use in some physical self-ascriptions. But those cases are not relevant to his immediate diagnostic aims.Footnote ¹¹

Second, Wiseman thinks it is a mistake to read Wittgenstein as holding that ‘I’ has two different uses, one that involves the recognition of a particular person and another that does not. His point, she says, is that there are two different cases in the use of ‘I’: the case where ‘I’ is ‘used with a predicate expression that describes a physical state or process’, and the case where ‘the predicate expression is a description of personal experiences’. But ‘what Wittgenstein insists upon’, she says, ‘is that these two cases are two cases of the (single) use of “I”, where “I” is an expression that is used by a particular person to say, among other things, how things appear to her’ (Wiseman Reference Wiseman2019: 673–674). The implication is that the standard reading, by contrast, must deny that the use of ‘I’ as subject and the use of ‘I’ as object are ‘two cases of the (single) use of “I”’: that it must treat the as-subject and as-object uses of ‘I’ as distinct, self-standing practices.Footnote ¹² But there is no reason at all for the advocate of the standard reading to take that view. On the contrary, she can insist, with P. F. Strawson and many others, that ‘the first personal pronoun refers, on each occasion of its use, to whoever then uses it’.Footnote ¹³ Whether ‘I’ is used as subject or as object, its use is governed by the same reference rule. And in both cases it is used by, and refers to, the same kind of thing: a human person. In that respect, the use of ‘I’ as subject and the use of ‘I’ as object are parts of a single use: a single, unified practice of self-ascribing mental and physical properties, sometimes in ways that are vulnerable to errors of misidentification, sometimes in ways that are not; a practice that is in turn part of a unified practice of ascribing such properties to others as well as oneself.

There is also a textual point to make. It is true that Wittgenstein introduces the distinction between the use of ‘I’ as object and its use as subject by saying, ‘There are two different cases in the use of the word “I” (or “my”)’ (my emphasis). But he immediately goes on to talk about ‘Examples of the first kind of use’ and ‘examples of the second kind [of use]’ (my emphasis). So Wittgenstein’s text gives us no reason to prefer talking about two different cases in the use of ‘I’ rather than two different uses of ‘I’; he himself seems to use the two expressions interchangeably.

Third, as I have said, Wiseman thinks it is wrong to ascribe to Wittgenstein the view that there are two uses of ‘I’, one that involves the recognition of a particular person and another that does not. But it is hard to see how we can avoid recognizing the distinction that the standard reading of Wittgenstein draws in those terms. So a further challenge for her view is to say how we should draw that distinction. I have said that, on the standard view, the distinction between the use of ‘I’ as object and the use of ‘I’ as subject cuts across the distinction between self-ascriptions of physical properties and self-ascriptions of experience and other mental properties. When I judge that my legs are crossed, for instance, in the normal case I can simply feel that they are crossed; I do not need to look and see how they are arranged. But there are other situations, e.g. when my legs are anaesthetized, where I do need to look and see in order to tell whether or not my legs are crossed. We could say, using Anscombe’s terminology, that in the first case I know without observation that my legs are crossed; in the second case, my knowledge about the position of my legs is based on observation (Anscombe Reference Anscombe1963: §8). That distinction, between observational and non-observational knowledge about oneself, plays a crucial role in Anscombe’s discussion of the first-person pronoun, ‘I’, which is a reference point in Wiseman’s discussion.Footnote ¹⁴

It is agreed on all sides, then, that for a judgement like ‘My legs are crossed’, there are these two different cases. But then there must be something to say about what the difference is. And it is hard to see what is wrong with saying that, in the case where my legs are anaesthetized and I need to look and see how they are arranged, the judgement ‘My legs are crossed’ involves recognizing the person whom I see to have crossed legs as myself; whereas, in the normal case, no such recognition of a person is required. Similarly, it is hard to see what is wrong with saying that, in that respect, these are two different uses of ‘my’. Of course there are similarities between these two uses of ‘my’ as well. In both cases, as we have said, ‘I’ or ‘my’ refers to the person who produced it; and both uses of ‘I’ or ‘my’ are parts of a single practice of ascribing mental and physical properties to oneself and others. But that does not remove the difference. If it is a mistake to mark the difference in terms of whether or not a use of ‘I’ or ‘my’ involves the recognition of a person, and whether the judgement is vulnerable to error through misidentification, why exactly is it a mistake? And how else should we describe the difference?

4. A ‘complete reorientation of our philosophical thinking about self-consciousness’?

Wiseman contends that the standard reading of the Blue Book passage fails to appreciate ‘the complete reorientation of our philosophical thinking about the self and self-consciousness’ that is implied by Wittgenstein’s work. In particular, as we have seen, she thinks that the standard reading approaches self-consciousness in the wrong way, by understanding it ‘in terms of a special epistemological capacity’ (Wiseman Reference Wiseman2019: 676). The correct, and properly Wittgensteinian, approach to understanding self-consciousness is instead to ‘describe the form of life of creatures with self-consciousness’ (Wiseman Reference Wiseman2019: 677). But why might it be objectionable to think that self-consciousness involves the possession of a distinctively first-person way or ways of knowing about ourselves? Does Wittgenstein give us good reason not to think in that way? And does the project of ‘describ[ing] our shared form of life through a study of [our] linguistic practices’ (Wiseman Reference Wiseman2019: 676) – specifically, the practice of using the word ‘I’ – provide a genuine alternative to the kind of appeal to distinctive, first-person epistemic capacities involved in the standard reading?Footnote ¹⁵

We have said that whether or not a judgement of the form ‘I am φ’ is immune to error through misidentification depends on how it is made. Where the judgement ‘My legs are crossed’ or ‘The wind is blowing my hair about’ is made on the basis of observation, it is vulnerable to error through misidentification. Where it is made in the normal, distinctively first-personal way, it is immune to error of that kind.

Now suppose we gloss ‘judgement made in a distinctively first-personal way’ as ‘judgement made on a distinctively first-person basis’.Footnote ¹⁶ Then we can see why the idea that self-consciousness involves the possession of a distinctive epistemological capacity (or capacities) might be thought to be at odds with Wittgenstein’s approach. For it is a characteristic feature of Wittgenstein’s discussions to insist that in the normal case a person’s judgements about what she is feeling, what she believes, how her limbs are arranged, and so on are not based on anything at all. I say, or judge, that I have toothache, that I believe it will rain, that my mental image is red, that my legs are crossed, that if I had been asked what someone should put after ‘1000’ I would have said ‘1002’, and so on. But those judgements are groundless. As Wittgenstein puts it in connection with the last of those examples:

So the challenge for someone who thinks, as I do, that the standard reading of the Blue Book passage is perfectly compatible with Wittgenstein’s approach to self-consciousness is this: to characterize our distinctive, first-person ways of knowing about ourselves, which are operative in judgements in which ‘I’ is used as subject, without appealing to the idea that self-consciousness involves access to a special source of evidence, that judging the colour of one’s own mental image involves applying introspective criteria for the redness of an image, that one judges the positions of one’s limbs on the basis of sensations, and so on.Footnote ¹⁷ But that challenge is easily met.

When I know that someone else is in pain, or has crossed legs, etc., I know it by observing what she says and does. By contrast, as we have seen, in the normal case my knowledge that I am in pain, or that my legs are crossed, etc., is not based on observing my body or behaviour. Now suppose that that negative claim is all we can say about how, in the normal case, one knows that one is in pain, or that one’s legs are crossed. It will still be true that self-consciousness involves a special epistemic capacity: a capacity to know what we are feeling, how our bodies are arranged, and so on without depending on the kind of evidence and observation we deploy in the third-person case. That minimal characterization of the special epistemic capacity is all that is required by the standard reading of the Blue Book passage; and it is all that is required by the standard view of immunity to error through misidentification.Footnote ¹⁸

But is this really an epistemic capacity? An epistemic capacity is a capacity to know something. And someone might object that, in the standard case, judgements like ‘I have toothache’ or ‘I hope it will rain’ do not express knowledge at all. For, they will say, the concept of knowledge is tied to the concepts of evidence, giving grounds, satisfying oneself, the possibility of making a mistake, and so on. And we have agreed that when I say or think ‘I have toothache’, ‘I hope it will rain’ etc., my judgements are typically groundless, that there is normally no question of satisfying myself that I have toothache or that I hope it will rain, and so on. So, the objection goes, I cannot be said to know that I have toothache, that I hope it will rain, etc.; my relation to my own sensations and mental states is non-epistemic.

That objection reflects an important strand in Wittgenstein’s own thinking about knowledge.Footnote ¹⁹ In line with that thinking, he famously remarks that ‘it can’t be said of me at all … that I know I’m in pain’ (Wittgenstein Reference Wittgenstein, Hacker and Schulte2009a: §246). He is similarly suspicious of the claim that I know what I want, wish, believe, feel etc. (Wittgenstein Reference Wittgenstein and Wittgenstein2009b: §309). How should we respond to this objection?

Even if we were tempted by the argument that the groundlessness of judgements like ‘I have toothache’ or ‘I hope it will rain’ shows that such judgements do not express knowledge, that argument would have much less plausibility for judgements about the position and movement of one’s body and limbs. In the normal case, I do not need to satisfy myself that my legs are crossed before judging that they are; nor do I make that judgement on the basis of evidence. But that is no reason for denying that I know that my legs are crossed. (Anscombe agrees; she is happy to talk about non-observational knowledge of the position and movement of our limbs.) And if we accept that our capacity to make reliable judgements about the position and movement of our limbs is a capacity to know their position and movement, there seems no objection to saying that that aspect of self-consciousness involves a distinctive epistemic capacity. The same is true for our awareness of our own sensations, attitudes, etc.. After all, my beliefs that I have toothache, or that I hope it will rain, are normally true. And they count as knowledge on standard conceptions of knowledge. For instance, such beliefs are safe; in similar situations where I believe that I have toothache or that I hope it will rain, those beliefs would not easily be false.Footnote ²⁰ It may in normal circumstances be pointless or uninformative to say that I know that I have toothache or that I know that I hope it will rain. But that does not make it untrue.Footnote ²¹ I suggest that there is no good reason to deny that a person’s capacity to make reliable, groundless judgements about her own sensations and attitudes is a capacity to know what she feels, hopes, believes, intends etc.; self-consciousness involves distinctive epistemic capacities. Accepting that point does not require us to deny that our relation to our sensations and mental states is different in many other ways from our relations to our bodies and to other physical objects. Nor does it require us to think that, in the normal case, self-ascriptions of sensations, attitudes and so on depend on evidence or observation.

I conclude with a brief comment on Wiseman’s proposal that the right way to achieve a philosophical understanding of self-consciousness is by ‘describ[ing] our shared form of life through a study of [our] linguistic practices’ (Wiseman Reference Wiseman2019: 676). What would such a study look like? Anscombe, focusing as we have noted on self-ascriptions ‘relating to actions, postures, movements, and intentions’ (Anscombe Reference Anscombe and Guttenplan1975: 35), says that what distinguishes self-conscious subjects like us from beings that lack self-consciousness, and who cannot use ‘I’ as we do, is the ability to make unmediated reports of their actions, postures, etc., which are not based on observation. We can say the same thing about self-ascriptions of experiences, attitudes, and so on; self-conscious subjects can make unmediated reports of their experiences, beliefs, etc., which are not based on observation. Remarks like these are the beginnings of a description of the capacity that underlies our use of ‘I’; or, more colourfully, the beginnings of a description of the form of life of creatures with self-consciousness. But, I suggest, giving such a description is not an alternative to describing the distinctive epistemic capacities that are characteristics of creatures with self-consciousness; it is another way of describing those very capacities. For as I have argued, the capacity to make such judgements about my body and limbs, and about my experiences and attitudes, is an epistemic capacity.

Similarly, Anscombe’s talk of our capacity to make ‘unmediated’ reports of actions, postures, and so on suggests a contrast between those unmediated reports and reports of actions, postures, etc. that are mediated by observation of a person or subject. But that is essentially the same as the contrast that the standard reading of the Blue Book passage draws between judgements in which ‘I’ is used as subject, where there is no recognition of a person, and judgements in which ‘I’ is used as object, which do involve the recognition of a person. Here too, Anscombe’s description of our capacity to use ‘I’ is not an alternative to the standard reading’s way of thinking of the two uses of ‘I’ (or, if we prefer, the two cases in the use of ‘I’). It is another way of making the same point.

Wiseman’s comments about the new approach to understanding self-consciousness that she finds in Anscombe and in Wittgenstein are programmatic. In order properly to assess the proposal, we would need to know much more about how the new approach will be developed in practice. For the reasons I have given, however, my suspicion is that the new proposal is not in tension with the standard reading of the Blue Book passage as I have presented it. It is simply a different way of presenting the same facts.Footnote ²²