1 The Use of ‘I’ Sentences
The different theses and theories that are concerned with the topic ‘I’ are motivated by different research interests and deal with different aspects of the subject. As I mention in the Introduction, one of the core issues concerns the concept of ‘self’. How are we to explain this concept? By transforming the reflexive pronoun into a noun, John Locke aimed to designate a particular kind of entity that is distinguished by only being accessible via consciousness. He identified consciousness as an awareness of one’s own mental states and actions. The Cartesian underpinning of Locke’s notion of self can be avoided by reverting to the reflexive pronoun and considering its use in connection with first-person singular pronoun. After all, selves are beings who refer to their own states or conditions by using this pronoun; and reflexive self-awareness cannot be expressed in language without this kind of self-reference. Thus, I will explain the concept of self by examining the use of ‘I’ as it occurs in sentences. There is an obvious objection to this approach, since it assumes that one has to know a language in order to have a concept of self. This assumption is not made in traditional philosophy, and the different theories considered so far also reject it explicitly. So, why make such an assumption at all?
To avoid misunderstandings, I want to point out that my approach does not presuppose the view that only beings that actually know a language are to be considered as relevant for the discussion of the issues raised in the Introduction. Below, I present some arguments in favor of such a claim; but for the time being, this should be understood only in a methodological sense. My analysis of the ability to use sentences that contain first-person pronouns identifies paradigmatic cases of different activities of human beings who know a language. This should be central in the discussion of the topic ‘I’. Even if one believes, as Tyler Burge does, that nonhuman higher animals have “lower level egocentric sensitivities or modes of reference”, one still has to distinguish such sensitivities from the ability to articulate a “full first-person concept” that we do not ascribe to nonhuman animals or small children. The possession of such a concept is manifested paradigmatically by the ability to use various kinds of sentences that contain the first-person singular pronouns. This linguistic competence reveals different cognitive faculties that involve reflexive reference. I explore these abilities by examining their linguistic manifestations. This does not mean that “self and consciousness would emerge after language”, as Damasio puts it. I am not concerned with genetic priorities, nor do I postulate a “verbal self”, in contrast to Damasio’s favored “nonverbal self”, whatever that may be. What matters here is to understand what the possession of a self-concept amounts to; and this will be clarified by considering sentences containing the word ‘I’.
In this chapter I argue that a philosophical understanding of the use of such sentences cannot be confined to semantic issues but has to take into account epistemic issues as well. Ludwig Wittgenstein addressed this topic by distinguishing different uses of ‘I’ in terms of immunity from error through misidentification; but there are other peculiarities in the use of ‘I’ sentences. A complete account of the use of such sentences has to focus on how the attitude of the first-person point of view is invoked. I will start with semantic considerations concerning the pronoun ‘I’ and then account for the other features of its use.
In contemporary philosophy of language there is growing interest in the semantics and pragmatics of indexicals, which include demonstratives, personal pronouns and temporal adverbs. What does the pronoun ‘I’ refer to? The obvious answer is that it refers to the person uttering it. Consequently, Hans Reichenbach coined the term token-reflexive. He claims: “The word ‘I’ … means the same as the person who utters this token …”. To understand this statement one must grasp the distinction between words and their utterances or uses of them as tokens. The same word can be used by different speakers or by the same speaker at different times, in writing or speech. The word is an abstract entity, a type; its uses are concrete occurrences called tokens. According to Reichenbach, it is peculiar of indexicals in general that they refer to something that has to do with their different uses: “The words under consideration are words which refer to the corresponding token used in an individual act of speech, or writing; they may therefore be called token-reflexive words.” This claim can be misleading: indexicals such as ‘here’ or ‘I’ when uttered, do not refer to the utterances made by a speaker at a certain place; they refer, to a particular place or a particular person. What Reichenbach means, is rather that these words refer to something which can be determined only by taking into account the circumstances of their use. Words can only be uttered by people at a certain time and in a certain place. These circumstances constitute the context of the utterance, and words such as ‘I’, ‘here’ and ‘now’ refer to different elements of that context.
dp n="37" folio="29" ? Reichenbach says: “It is easily seen that all these words can be defined in terms of the phrase ‘this token’.” However, the meaning of ‘here’, for example, cannot be defined that way. What he has in mind is that what the word refers to in a given context can be determined by using ‘this token’. The word ‘here’, uttered by me here and now, i.e., “this token”, refers to the place where it is uttered by me. Therefore Reichenbach does not talk about the meaning of a word, as given by a definition, but rather about its referent. Concerning the first-person singular pronoun, he states that “the word ‘I’ means the same as ‘the person who makes this utterance’ …”. He considers this to be a rule for determining the referent of the word in a given context of its utterance.
As David Kaplan points out, it is not the referent of the word (which can be uttered by many speakers at the same time) but rather the referent of its utterance or of the token which is determined by this rule: “‘I’ refers to the speaker or writer of the relevant occurrence of the word ‘I’, that is, the agent of the context.” Each use of a word occurs within a particular context, and for every context there is a particular speaker or writer, “the agent of the context”, as Kaplan says. This rule gives the meaning of the word ‘I’, understood as a rule for determining the referent of its tokens, called character by Kaplan, and allows one to determine the referent of its use in a given context. By knowing the meaning of an indexical expression such as ‘I’, ‘here’ or ‘now’, one knows how to determine the referent of its tokens. This does not imply that one knows what is referred to. That knowledge requires information about the context of a use, and there can be a remarkable difference between the knowledge of a speaker and that of a listener, as Kaplan’s following example makes clear: “I may twice use ‘here’ on separate occasions and not recognize that the place is the same, or hear twice ‘I’ and not know if the content is the same. What I do know is this: if it was the same person speaking, then the content was the same.” It is strange that Kaplan moves from the point of view of the speaker to that of a listener to show that understanding the meaning of an indexical expression (character) may not be sufficient to know what it refers to in a given context of use (content). He does not explain the change in point of view, nor does he explain why he illustrates this point by considering different indexicals.
dp n="38" folio="30" ? As far as my use of ‘here’ is concerned, I know what the term refers to in a given context of its use: the place where I am now. I may be completely wrong about where ‘here’ is: the definite descriptions I give may not refer to the place where I am. But this does not matter for my knowledge that I am here now, however uninformative this knowledge may be; it does not require that I recognize the place where I am now as the same place where I was once before, for example. For a listener, things can function very differently; the listener may guess or may just not know where I am on hearing my utterance of ‘here’. This does not imply a lack of knowledge of the linguistic meaning of the word; it probably shows a simple lack of complete knowledge of the context of my utterance, and hence the inability to identify the place where I am.
As far as my utterance of ‘I’ is concerned, Kaplan considers only the listener’s point of view in order to establish his general claim that understanding the linguistic meaning of an indexical expression does not guarantee knowledge of its referent. However, what is true for listeners does not hold for the speakers; speakers cannot know that they have uttered the word ‘I’ without knowing that it refers to themselves. Hence Kaplan’s general claim that understanding the linguistic meaning of indexicals does not provide eo ipso knowledge of what their uses refer to in a given context, is true only for the listener. While, with regard to my utterance of ‘here’, he is postulating a condition which is too strong, he does not consider my use of ‘I’ from the perspective of the speaker. Kaplan’s account of the referential mechanism of indexicals fails to fit the peculiar epistemic situation of the speaker.
This applies in particular to the use of ‘I’ from the perspective of the speaker. Instances of such use are conscious references by speakers to themselves: they refer to themselves and know that they do so. Not every reference to oneself is a conscious self-reference. One can refer to oneself without knowing that one does so. I may be angry, for example, at the person who most recently entered the room and did not shut the door, without knowing that I am angry at myself, since I had not realized that it was me! Lucy O’Brien calls this a de facto reflexive reference represented by the schema: ‘A refers to B when A = B’. Such self-reference does not require that, in referring to B, A knows that A and B are one and the same person; but it does not rule it out, either. Furthermore, O’Brien draws a distinction between “systematically reflexive reference” and “fully self-conscious self-reference”: While the former is represented by the schema ‘A refers to A’, she defines the latter as: ‘A refers self-consciously to A’. The former has something to do with the linguistic meaning of the expression used by the speaker or thinker. A reference is called systematic when it is determined solely by the linguistic meaning of the expression; which does not hold either for “de facto-reference” or for “self-conscious self-reference”.
The use of the word ‘I’ from the perspective of the speaker, not considered by Kaplan in the scenario mentioned, is a conscious self-reference. Ernst Tugendhat has suggested that the “rule of use” for this word is that “by means of this expression each speaker designates himself”. He believes that this rule provides a “definitional explanation of the word”. Speakers can designate themselves, however, without using the word ‘I’; e.g., by using their names or definite descriptions. Thus, Tugendhat’s “rule of use” states only a sufficient condition for designating oneself. When using names or definite descriptions, there is the possibility that speakers refer to themselves without knowing it. As far as the use of ‘I’ is concerned, no such possibility exists: its use is distinguished by the fact that it is a conscious self-reference. However, Tugendhat’s rule has to be explained in a certain way. What does a speaker refer to, if he or she knowingly refers to himself or herself? Of course, to himself or herself. However, this trivial answer needs clarification. The reflexive pronouns ‘herself ’ and ‘himself ’ have to be understood so that they account for the peculiar character of conscious self-reference. Elizabeth Anscombe speaks therefore of an “indirect reflexive pronoun”: “For what is in question is not the ordinary reflexive pronoun, but a peculiar reflexive, which has to be explained in terms of ‘I’.” What she has in mind can be explained in the following way: one can talk about or refer to oneself without knowing it, as in O’Brien’s de facto reflexive reference. The use of ‘I’, however, is distinguished by the fact that the speaker knows that he or she is speaking of himself or herself. Therefore, Anscombe claims the referent of the reflexive pronoun ‘oneself’ can be determined only by the speaker’s use of ‘I’. She objects to Tugendhat’s proposal and points out that his explanation is circular. This charge is justified, since what the reflexive pronoun ‘oneself’ refers to can be determined by the speaker only by means of using ‘I’. This is true of the speaker, not of the listener.
O’Brien disagrees with this charge of circularity, but I think that she misses Anscombe’s point. O’Brien writes: “The general point seems to be that ‘I’ cannot be understood merely as a device of reflexive self-reference because reflexive self-reference is possible in the absence of genuine first-person self-conscious self-reference.” But Anscombe is not concerned with the difference between reflexive self-reference and self-conscious self-reference, and does not ask for the additional conditions that must be met by reflexive self-reference in order for it to become self-conscious self-reference. Anscombe deals only with the latter: the proposal that “‘I’ is the word that each one uses when he knowingly and intentionally speaks of himself.” This proposal is circular, because the speaker can only determine what is referred to by the reflexive pronoun by means of using ‘I’. O’Brien rejects the idea of an epistemically transparent “indirect reflexive pronoun”, but she does not show that the rule of use of ‘I’ as a device of self-conscious self-reference can be stated without the use of this pronoun. While Anscombe does not want to give an explanation of conscious self-reference by criticizing the rule of use of ‘I’ proposed by Tugendhat, O’Brien does indeed aim at just such an explanation. Her explanation has two parts: the semantic claim that the word ‘I’ is a device for designating the person who uses it; and the epistemic claim that the speaker knows they are using the word this way when they use it. By understanding the meaning of ‘I’ – that is, the rule of its use – I know that when people use it they refer to themselves. As Kaplan shows in his scenario, it is possible that this knowledge is not yet sufficient to identify who is referred to in a given context of its use, and it is certainly not sufficient for knowing that one refers to oneself. Therefore, O’Brien claims that one has also to know that one uses the word according to its rule of use. Thus, conscious self-reference is explained by the knowledge a speaker has when using the word ‘I’. But how does the speaker know about that use without knowing that they themselves are using it in that ...