'I' and the First Person Perspective
In any actual utterance the first person singular pronoun 'I' stands for the person who uses it; 'I' refers to the speaker. This simple observation stands in stark contrast to the epistemological and metaphysical claims, advanced on behalf of the first person from Descartes to the present. First person claims, i.e., those formulated with 'I' as the subject of the main verb, seem to have an immediacy, transparency and salience, which cannot be rendered in any other terms. Such claims express the speaker's particular point of view and possess special relevance for her actions. (1) Thus, an intimate connection seems to exist between the use of 'I' and what has been called the first person perspective, construed as the perspective of consciousness. The idea of a first person perspective, a point of view different from and irreducible to any other perspective, seems most striking in the case of sensory states and perceptions. (2) However, it equally pertains to occurrent thoughts, practical reasoning and intentional action.
The 'first person perspective' is so-called after the grammatical form of its typical expression, the first person singular. In English this grammatical form involves the personal pronoun 'I'. How is 'I' related to the first person perspective? I shall argue that the two are not as directly related as some philosophers suppose. First, I discuss the semantics of 'I' as a lexeme of a natural language and trace the special features attributed to 'I' to the fact that with uses of 'I' semantic reference and speaker's reference coincide. Secondly, I show that the "referential guarantees" or the impossibility of "misidentification" attributed to 'I' accrue only to actual uses. Thus, the relation between 'I' and the first person perspective must be explicated within the context of linguistic action. I then argue that 'I' does not function like a proper name, a description or an identifying singular term. Rather, I propose that a use of 'I' indexes a linguistic act with respect to the responsible agent. On this view, the use of 'I' in an utterance does more than express the first person perspective, considered as the perspective of consciousness, since the first person perspective can be expressed by unmarked, impersonal or zero-pronominalization linguistic forms. In conclusion, I illustrate this claim with examples from the Wiener Kreis.
In order to delineate the relation between 'I' and the first person perspective, we must attend to 'I' as a word type of a natural language. 'I' is the first person singular pronoun in the system of English personal pronouns. The personal pronouns compose an internally structured system, which is closely connected to the system of person in verbs. The grammatical category of person depends generally upon the grammaticalization of roles the participant roles of speaker and hearer, and the non-participant roles, the so-called third person. Many languages grammaticalize the category of person by inflecting the main verb. For example, Latin and Spanish have a first person singular form with zero-pronominalization (cogito, pienso) as well as a pronominalized form (ego cogito, je pienso). This shows that it is not necessary to employ a word type corresponding to 'I' to express the first person perspective. Even impersonal linguistic forms like "There is red" can express the first person perspective, a point to which I will return in the last section. Thus, the use of 'I' is not a necessary condition for expressing the first person perspective.
The personal pronouns are essentially deictic or indexical devices (and only secondarily anaphoric). We use them primarily to refer to persons. Though the meaning of a given pronoun remains constant, it can be used literally to refer to different individuals. The actual context of use plays a central role in determining the referent of a pronoun. Indeed, the reference of a personal pronoun is fixed only relative to a specific context of utterance. Independent of a particular use a personal pronoun has no specific referent.
The pronoun 'I' occupies a special position within the system of personal pronouns. First of all, its referent is determined semantically. The meaning of the word type 'I' consists in a rule, specifying how its reference is a function of the context of utterance, namely, who uses it. A token of 'I' stands for whomever produces it. Note that who is using 'I' is a fact about the context of utterance and this fact allows others to identify the referent of 'I' in a concrete situation. Hence, only uses of 'I' can be said to have a specific referent.
In contrast, the reference of other personal pronouns is not determined semantically, but involves additional, pragmatic considerations. The speaker's intention to refer to someone, the so-called 'speaker reference', plays an important role in determining the referent of pronouns like 'we', 'you', 'she', etc. (3) In the case of 'I', however, the meaning of the word type by itself determines the reference as a function of the context of utterance. On any particular occasion of use, 'I' refers to the speaker who employs it solely on the basis of its meaning. As a competent speaker of English, one must use 'I' to refer to oneself and that is a fact about its meaning. (4) In consequence, it is impossible for the semantically determined reference of 'I' in a particular context and the speaker reference to diverge, whereas semantic reference and speaker reference can and often do diverge in the case of other pronouns, leading to reference failure and error. In the case of 'I', however, semantic reference and speaker reference always coincide and they do so solely on the basis of the linguistic meaning of the word type. This often overlooked fact sets 'I' apart from all other pronouns. Moreover, this fact accounts for many of the special features associated with 'I'.
'I' is further distinguished from other personal pronouns by the fact that it is not semantically equivalent to any other expression, e.g. a proper name, a definite description or an indexical description. (5) Although 'I' stands for an individual due to its role in the system of personal pronouns, it would be incorrect to assume it functions as a proper name. A name denotes not only an individual, but the same individual. The referent of 'I', however, changes with each speaker. Furthermore, a proper name can be used by different speakers to refer to the same individual, but 'I' can only be used to denote oneself, even though any speaker can so use it.
'I' does not function like a description either. If 'I' functioned like a definite description, its reference would be tied to its satisfying descriptive predicates like 'the so-and-so'. However, there seems to be no description, not even an indexical one, equivalent to 'I'. Although 'I' stands for the speaker, the expression 'the utterer of these words' is not strictly equivalent to 'I', since I could use the description, 'the utterer of these words', to refer to someone I am quoting or translating.
'I' seems to lack descriptive content entirely. Importantly, there is no need for the speaker to 'know who' he is, i.e. who is uttering 'I', in order to successfully refer by its use. The speaker may have entirely false beliefs about himself or no identifying beliefs at all. None the less, when the speaker utters a sentence containing 'I', he refers to himself. By the use of 'I' one refers to oneself without any further characterization.
A corollary of this point is that 'I' does not function like an identifying singular term. This can be seen by asking whether the speaker identifies herself semantically, i.e. makes identifying reference in using 'I'. If 'I' were like an identifying singular term, it would denote a particular object for the speaker in contradistinction to other objects. It would identify a particular object, e.g. the speaker herself. But, semantically, the word type 'I' does not denote or identify a particular individual at all. It can be used by any speaker in any context. It is its use in an utterance act that provides a referent. Hence the functioning of 'I' is not to be understood on the model of identifying an object.
In sum, the pronoun 'I' does function in a special way: (1) its reference is fixed semantically; (2) in the case of 'I' semantic reference and speaker reference always coincide; (3) 'I' is not semantically equivalent to any name, definite description or indexical description, nor (4) does it function like an identifying singular term. What philosophical implications do these observations have? At first glance, they might seem to entail epistemological and metaphysical theses concerning self-knowledge, self-identity, self-awareness and privileged access. Many philosophers have thought so. Let us review three well-known claims concerning 'I' to see if this is the case. (6)
(I) Uses of 'I' are guaranteed against lack of reference; the referent always exists. The fact that a use of 'I' cannot fail to refer what Anscombe called one of its "referential guarantees", might suggest that 'I' refers to a special sort of entity a Cartesian ego, a metaphysical self or a transcendental subject, and does so in a special way. However, this "guarantee" is much less mysterious than it appears. For it is a rule of competent usage, that 'I' refers to whomever uses it. The fact that there is a speaker is sufficient to guarantee a reference for the use. In this light, the "guarantee" against failure of reference hardly implies substantive epistemological or metaphysical claims.
(II) Uses of 'I' are immune to referential error, i.e. misidentification. As we have seen, it is impossible that the semantically determined reference of 'I' on an occasion of use and the competent speaker's intended reference not coincide. Uses of 'I' are guaranteed against "mistaken" or incorrect reference by virtue of the competent use of the English language system. The speaker will correctly and successfully refer to herself by means of 'I' in English on all occasions other than those on which she acts as a spokeswoman for someone else. It is the performance of a linguistic action by means of a sentence containing 'I', which guarantees the correct referent for 'I'.
(III) The claim that reference in the case of 'I' does not proceed by way of identification might be taken to indicate that a special, unmediated epistemological relation is involved. The attraction of this suggestion is directly proportional to the problems encountered in treating 'I' as an identifying singular term. As Sidney Shoemaker points out, a particular object is identified either by a unique property or a unique relation. (7) In the case of the "use of 'I' as subject", e.g. psychological predicates like "I am in pain", such an identification would require either a prior identification, leading to a regress, or the identification would already presuppose the very identification, which one is trying to achieve. Thus, the lack of descriptive content in the case of 'I' might seem to lend plausibility to the idea of a direct, epistemological relation. However, the inference from the fact that 'I' does not function like an identifying singular term to a special epistemological relation is not compelling. For the special features associated with 'I' can be fully explained on the basis of the pronoun's semantics and consequences of its semantics for actual uses of the pronoun.
Note that the special features adduced accrue only to actual uses of sentences containing 'I', i.e. to utterances in the sense of linguistic actions. Sentence types containing 'I' exhibit such features only potentially, i.e. as potentially spoken or written by a speaker in a particular context. For without an utterance act, there is no specific referent, only a potential role. Actual uses of sentence types are actions. They are actions performed by linguistic means, i.e. with the resources of a language system.
The fact that the special features associated with 'I' pertain only to its use in concrete utterance acts suggests that in order to clarify the relation between 'I' and the first person perspective we must consider the role of 'I' in linguistic action. I take a linguistic action to be a speech act performed by an agent at a particular time and place with the words of a language system. (8) Linguistic actions have at least three characteristics: (1) the meaning of the words uttered plays an indispensable role in performing the action intended by the speaker; (2) a linguistic action is a datable utterance event; and (3) a linguistic action is the action of a particular agent.
What can we say about the use of 'I' in linguistic acts? I want to suggest that the function of 'I' does not lie in its referring to an entity, nor can it be assimilated to purely expressive avowals, although I cannot argue for these claims here. Instead, I want to propose that the function of 'I' lies in its indexing a linguistic act with respect to the agent responsible for the act. This claim concerns first person utterances which exhibit a pronominalized form of the first person, an explicit 'I', as the subject of the main verb. The idea is that in such actions the speaker uses an identifiable lexeme for the first person singular to index her action, thereby signaling that she is the agent. This indexing function of 'I' can be elucidated by considering examples of linguistic acts, which express the first person perspective, but do not use the pronoun 'I'.
Consider the Wiener Kreis's attempts to capture the immediate contents of experience, e.g. Schlick's Konstatierungen and Carnap's Protokollsätze. (9) Despite important differences, these statements were meant to express the pre-theoretical deliverances of experience. They were attempts to capture immediate sensations, die Erlebnisperspektive, which qualifies as the first person perspective on anyone's account. Konstatierungen like "here now red" were to express the phenomenal character of immediate experience. For Schlick a Konstatierung is an indexical utterance event, a linguistic act at a particular time and place, which is why it cannot be written down.
From a semantic standpoint, Konstatierungen are quasi-sentences; they lack particular syntactic and semantic features usually found in complete sentences. First, note that there is no verb. A verb would implicitly involve a agent and a verb of perception verb would involve a subject explicitly, as when Roderick Chisholm reformulates Konstatierungen as 'I sense redly'. Secondly, there is no grammatical subject. Indeed, in Schlick's solipsistic construction, no subject is needed. For there is no need to make the perspective of the agent explicit, as there is only one perspective and it constitutes the world. Note, however, that the indexicals 'here' and 'now' in a Konstatierung cannot be understood except as involving reference to a speaker. In order for the quasi-sentence "here now red" to make sense, it must be construed relative to the agent of the utterance act.
Carnap's Protokollsätze were another attempt to express the phenomenal content of immediate experience. Protocolls like "A experiences red at time t" were to be backed by the content of quasi-sentences like "Here now an experience of red". Such quasi-sentences exhibit the same peculiar semantic features as Schlick's Konstatierungen: no verb and no subject term coupled with the presence of indexicals. As Carnap eventually conceded to Otto Neurath, if Protokollsätze are regarded as describing an observer's immediate experience, they can only be understood solipsistically.
What do these examples reveal about the relation of 'I' to the first person perspective? First, they show that the function of 'I' cannot be explicated within the perspective of a consequent solipsism. On a solipsistic position, there is no 'I', nor is there a need for 'I', because the first person perspective and the world are coextensive. Secondly, an explicit 'I' is not required to express the first person perspective, as the examples illustrate. The first person perspective, considered as the perspective of consciousness, can be expressed by impersonal, zero-pronominalization forms. (10)
Thus, 'I' expresses the first person perspective in an utterance, but it does more. For the first person perspective can be expressed without the use of 'I'. In utterances with an explicit 'I', e.g. "I see red here now", there is an overt indexing of the utterance act with respect to the agent. In indexing a linguistic action by using 'I', the speaker presents herself as the author and agent of the linguistic act. Filling out the details of this proposal must be left to another occasion. However, I believe it points in the right direction, one which locates the relation between 'I' and the first person perspective in the context of action.
(1) As H.-N. Castañeda, "He": A Study in the Logic of Self-Consciousness" Ratio 8 (1966):130-57, and J. Perry, "The Problem of the Essential Indexical" Nous 13 (1979):3-21, have emphasized.
(2) When Thomas Nagel claims that all and only conscious beings have a point of view the 'what-it-is-like' phenomenon, he considers examples of phenomenal consciousness, "What is it like to be a bat?" (1974) in: Mortal Questions Cambridge 1979.
(3) cf. K. Donnellan "Reference and Definite Descriptions", Philosophical Review 75 (1966): 281-304; "Speaker Reference, Descriptions and Anaphora" in: P.A. French, Th. Uehling, H.K. Wettstein (eds.) Contemporary Perspectives in the Philosophy of Language, Minneapolis, MN 1979:28-44.
(4) This is essentially D. Kaplan's point, "Demonstratives" (1977), in: J. Almog, J.Perry and J. Wettstein (eds.), Themes from Kaplan, New York/Oxford 1989.
(5) This has been emphasized by Wittgenstein The Blue Book (1958):107 in: The Blue and Brown Books Oxford 1972, Philosophical Investigations § 410; E. Anscombe, "The First Person" in: S. Guttenplan (ed.) Mind and Language, Oxford 1975; H.-N. Castaneda (1966), (1968) "On the Logic of Attributions of Self-Knowledge to Others" Journal of Philosophy 65 and J. Perry, "Frege on Demonstratives" Philosophical Review 86 (1977): 474-97, (1979).
(6) Anscombe (1975); S. Shoemaker "Self-reference and Self-Awareness" J Phil 65 (1968):555-67; reprinted in Identity, Causality and Mind (1989).
(7) Shoemaker (1968/1989):12
(8) I take linguistic acts to be speech acts in J. L. Austin's sense, How to Do Things with Words Oxford 1962, not in the sense of J. R. Searle, who conflates the locutionary aspect and the illocutionary aspect of a speech act in the case of explicit performatives Speech Acts Cambridge 1969.
(9) M Schlick (1935) "Über Konstatierungen" in: M. Schlick, Philosophische Logik, Frankfurt am Main 1986:230-37; R. Carnap (1928) Der Logische Aufbau der Welt 2. Auflage, Hamburg 1961.
(10) This idea is, of course, not new. In criticizing Descartes' cogito argument, Georg Christopf Lichtenberg suggested that Descartes' premiss should read "there is thinking" in analogy with "there is lightening", cf. Vermischte Schriften, G. C. Lichtenberg und F. Kries (eds.) Göttingen 1800-1803:96.