Intentionality inherently de se and the Problem of Self-Consciousness

Dec 3, 2021
[ PDF / print ]

Abstract

In this paper I present what Kristina Musholt, in her book Thinking of Oneself—From Non conceptual Content to the Concept of the Self (MIT Press, 2015), calls the problem of self-consciousness. This is an old problem, reaching back from Shoemaker and analytic philosophy on the one side, Henrich and the Heidelberg School on the other, through Sartre and phenomenology to Fichte. My goal is to identify the assumption generating it and to argue that this assumption is false. I will end by sketching, in admittedly very hand-waving fashion, what this means for the notion of self-consciousness. My argument presupposes an understanding of the central thesis of my research—what I call the thesis of intentionality inherently de se —so I begin by elaborating the four claims constitutive of it.
This is a slightly revised version of a paper held on October 26th, 2021, in the Research Seminar of the Department of Philosophy at the University of New South Wales.

Introduction

In her book Thinking of Oneself—From Nonconceptual Content to the Concept of the Self (MIT Press, 2015) Kristina Musholt outlines what she calls the problem of self-consciousness. This is in fact an old problem, reaching back from Shoemaker and analytic philosophy on the one side, Henrich and the Heidelberg School on the other, through Sartre and phenomenology to Fichte. My goal is both to identify the assumption generating it; and to argue that this assumption can and must be rejected. The latter will involve making some brief and insufficiently grounded remarks about how to think better about self-consciousness—what Musholt calls thinking of oneself, the thinking of ‘I’-thoughts or explicit self-representation. In order to understand all this, however, one must first understand what I call the thesis of intentionality inherently de se.

Section I: The Thesis of Intentionality inherently de se

The thesis consists of four claims. The first is the following:

1. There are beliefs and other forms of intentionality only perspicuously characterisable by sentences of the form “S believes that he (or she or indeed it) is F,” “S desires that he (or she or indeed it) be F,” and indeed “S sees something F standing in relation R to it,” etc.

In these sentences the third-person pronoun refers to S through an ineliminable anaphoric link to the subject-term of an intensional context.

The use of such sentences must be distinguished from that of sentences such as “S believes that a is F” and “S believes that the (one and only) G is F,” even when the proper name ‘a’ and the definite description ‘the (one and only) G’ are co-referential with the singular term ‘S’. Moreover, the use intended here must be distinguished from another use of these same sentences since sometimes a sentence of the form “S believes that it is F” can be used to characterise a form of intentionality not de se but de re. In the latter use, the pronoun ‘it’ still refers anaphorically to S but no more is meant than that S believes of S that it is F. Evidently, this anaphoric link is eliminable. One may replace it without loss by an anaphoric link to the object-term in the slightly different intensional context “S believes of S,” thereby generating a sentence more perspicuously expressive of what is meant, precisely the sentence “S believes of S that it is F.”1 In order to remove such ambiguities, I will borrow some symbolism from Castañeda. When sentences of the relevant form characterise intentional states or experiences de se, the relevant third-person pronoun shall be furnished with an asterisk. Forms of intentionality de se are now only perspicuously characterisable by sentences of the form “S believes that it* is F”, “S desires that it* be F”, “S sees something F standing in relation R to it*,” and so on.2

The first claim in effect maintains that a sentence of this form, containing as it does the asterisked third-person pronoun, characterises the relevant form of intentionality as it ‘really is’. That is, the sentence articulates not merely what the form of intentionality is objectively about—what items are in the so-called context of evaluation and what property or relation is true of these items in order for content to be true—but also the causal role3 which ultimately distinguishes the form of intentionality for what it is.4 As one may also put it, the characterisation captures not merely what the form of intentionality is objectively about but also how it is about it.

The allusion to causal role yields the second claim constitutive of the thesis of intentionality inherently de se, a claim which motivates the first:

2. Forms of intentionality inherently de se are essential to the generation or motivation of purposive behaviour, hence must be mentioned in any complete everyday explanation of it.5

As one might also put it, intentional states and experiences de se mediate between intentionality as such and purposive behaviour. Intentionality not de se only takes effect through intentionality that is.

This second claim is justified by the following consideration: when two agents A and B recognise that A is about to be attacked by a bear, they behave differently: A rolls up into a ball while B runs to get help—see Perry 1977, p. 494. Yet both agents may have all the same beliefs and desires with contents anyone can have as contents of belief and desire. For example, A and B may both believe that bears are an endangered species and may both desire that bears be protected. Indeed, A and B presumably both believe that A is about to be attacked by a bear and presumably both desire that A not be thus attacked. Let us call such sharable contents and the forms of intentionality of which they are contents impersonal. So A and B have all the same impersonal intentionality. Yet A and B behave differently.6 They must therefore differ in respect of beliefs and desires that are not impersonal—beliefs and desires with contents personal in the sense that only their subject can believe or desire them. Clearly, such personal contents and forms of intentionality are precisely ones de se: A’s belief that it* is about to be attacked by a bear and A’s desire that it* not be thus attacked; correspondingly, B’s belief that the person it* is observing is about to be attacked by a bear and B’s desire that this person distinct from it* not be thus attacked.

Claim 2 is widely endorsed.7 It is, however, widely understood in a manner I reject. In order to appreciate what I am getting at here, we must consider a third claim, one which constitutes a first divergence from the mainstream:

3. In no strong sense does intentionality de se implicate or entail the use, or the capacity for the use, of the first-person pronoun.

There is certainly a weak sense in which intentionality de se implicates the first-person pronoun: If S believes that it* is F, then, should S be capable of using the first-person pronoun at all, S must use it in order to characterise the content of its belief perspicuously. In other words, S must assert the sentence “I believe that I am F.” The vast majority, however, believe, or are committed to believing, that forms of intentionality de se are indexical, indeed first-personal, in a sense stronger than this.8

This is shown by how the vast majority understand claim 2. It is invariably understood as asserting that indexical, indeed first-personally indexical rather than, say, third-personally anaphoric belief and desire is essential to agency even though the examples used to ground the claim only show, as any impartial consideration of them reveals, that personal belief and desire, however it is characterised, is essential. And so intentionality de se is invariably said to be indexical, indeed first-personal even though the phrase ‘de se’ itself is third-personally anaphoric. Moreover, the content of intentionality de se is regularly articulated by citing a sentence in the first-person (“I am F”) rather than by using a sentence in the third-person (“it* is F”).

This preference for characterisation as indexical and first-personal is never explained. Nor is any strong sense in which intentionality de se is distinctively indexical, indeed first-personal, ever clearly identified.9 The nearest one comes to either explanation or identification is Castañeda’s widely endorsed doctrine of quasi-indication. According to Castañeda, the asterisked third-person pronoun has two roles: firstly, the semantic role of referring to the subject of the form of intentionality it helps to characterise. Secondly, however, it has the pragmatic role of indicating that this subject is using, or is at least able to use, the first-person pronoun in order to self-ascribe this form of intentionality de se and its content. This second role is quasi-indication, which Castañeda explicitly characterises as intimation of an actual use, or at least an actual capacity for the use, of the first-person pronoun.10 Even Perry, who appreciates the dangers implicit here, endorses this: he claims that we ordinarily take the pronoun ‘he’ used as ‘he*’

to be what … Castañeda calls [a] quasi-indicator … . Used as a quasi-indicator, “he” performs two functions. First, like a pronoun, it picks up the reference to [the subject] X. But it also tells us how [X] thinks of X. “He” tells us X thinks of X as himself. He thinks of himself in the way that we think of ourselves when we use the word “I.”” (Perry 1980a, p.538/p. 51)

Note Perry’s caution: he does not say that the pronoun ‘he’ tells us that X thinks of himself in the way X thinks of himself when X uses the word ‘I’. Room is left for X’s being unable to use the word ‘I’ or any linguistic equivalent.

All in all, what is meant when one speaks of indexical or first-person belief and the like is very unclear. But if the third claim is right, then this problem simply disappears. For if it is right, then no pragmatic function of quasi-indication attaches to the asterisked third-person pronoun. Indeed, another way of putting the claim is that the pronoun is not a quasi-indicator.11 With this, we come to the fourth and most important claim, one which entails that it is not a quasi-indicator:

4. Intentionality de se does not eo ipso implicate or entail self-consciousness12

In order to understand this claim, we must start with the notion of self-consciousness itself. Musholt defines self-consciousness as

the ability to think “I”-thoughts, where “I”-thoughts are such that they nonaccidentally refer to the subject entertaining them. That is to say, [i.e., this alone entails that] in having an “I”-thought the subject cannot but be aware that the thought refers to herself. (Musholt 2015, p. 63)

By non-accidental reference Musholt means the character of any form of intentionality de se as being about its subject in the latter’s capacity as its subject: when a subject S believes that it* is F, its belief is about S in virtue of S’s being the subject of the belief.13 This is then said to constitute a form of intentionality as an ‘I’-thought, that is, a thought of which “the subject cannot but be aware that [it] refers to herself.”

Note, however, that the latter claim does not follow. Non-accidental reference to the subject of intentionality does not alone entail that “… the subject cannot but be aware that the thought refers to herself.” In the first instance, non-accidental reference connotes the way in which, in the sentence “S believes that it* is F,” the asterisked third-person pronoun refers to the referent of the subject-term ‘S’. One may, however, extend the notion to S’s belief itself. But when thus extended, the notion entails merely that in recognising the belief, that is, in identifying it for the belief it is, a subject S◆ cannot but be aware that the belief refers to the subject of belief, namely, S. Evidently, S◆ need not be S itself. Non-accidental reference does not entail that in having the belief, S is aware that the belief refers to itself, or rather, to itself*.

Our intuitions tell us, of course, that “in having an “I”-thought the subject cannot but be aware that the thought refers to herself.” (Musholt 2015, p. 63; italics added) But this only shows that non-accidental reference is not sufficient for being an ‘I’-thought, whatever the latter might be precisely. Yet the notion of non-accidental reference is sufficient for distinguishing S’s belief that it* is F from S’s belief merely that S is F. S’s belief that S is F does not possess non-accidental reference because it is not about S, hence does not refer to S, in virtue of S’s being the subject of it. The belief is about S, hence refers to S, merely accidentally.

We have here a crucial conflation or equivocation. Musholt effectively identifies intentionality de se with self-consciousness: simply to possess any form of intentionality de se is always already for its subject to know, or at least in some sense to be aware, that it* possesses it.14 Nor is Musholt alone in this. Jaszczolt describes it as a fact that “(t)he essence of [the] de se is self-awareness, the epistemic concept of one’s conscious access to oneself”15 (Jaszczolt 2013, p. 82)—actual self-awareness.16 Castañeda declares that the asterisked third-person pronoun is a pronoun of self-consciousness.17 Use of this pronoun in the attribution of a form of intentionality de se does not simply refer anaphorically to the subject; it also marks the fact that the subject is aware of this form of intentionality as its*. For this reason, Castañeda describes the pronoun as intimating that “ONE is conscious of ONEself qua oneSELF,” or again, “that one thinks of ONEself as oneSELF.” (Castañeda 1999b, p. 256 and p. 264) We have already seen Perry to endorse this.18 But so, too, does Manfred Frank, member of the so-called Heidelberg School19—see Frank 2015, p. 129. They all believe that the pronoun intimates that its referent is actually thinking, or is at least actually able to think, of itself as itself,20 which they all understand to be self-consciousness. The fourth claim denies all this.

In order to see why it is right to do so, let us briefly examine how and why one is led to understand the phrase ‘as itself’ in this fashion. One is thus led when one follows Castañeda and takes the third-person reflexive pronoun ‘itself’ to be a possible instance of the schematic letter ‘Z’ in the general schema “X thinks of Y as Z.”21 When S asserts that the Morning Star is a planet, S refers to Venus as the Morning Star. Here, the phrase ‘as the Morning Star’ intimates that S is specifying the property of being identical with the Morning Star as necessary and sufficient for being what S’s assertion is about. This is then extended, in the first instance to belief, but ultimately to intentionality in general: when S believes that the Morning Star is a planet, or desires that the Morning Star be a planet, S thinks of Venus as the morning star. So analogously, being identical with the Morning Star is necessary and sufficient for being what S’s belief or desire is about; the aboutness of S’s belief or desire consists in possession of the biconditional property of being about something x just in case x is the Morning Star.

But S’s thinking of S as itself cannot be a thinking of something Y as Z in any sense like this. The property of being identical with itself is necessary and sufficient for being about anything, hence necessary and sufficient for being about nothing.22 So if there is any sense at all to talk of S’s thinking of itself as itself, it cannot derive from the general schema. What could this other sense be? At this point, one just shifts to some variation on a theme from Frege, who postulates an original way in which each subject is given to itself precisely as itself.23 For Castañeda the asterisked third-person pronoun marks the presence of a so-called ‘I’-guise.24 The later Perry speaks of a self-notion, Musholt of a self-concept.25 All these moves, however different amongst themselves, are ways of appropriating the generic Fregean move. A subject’s having one or the other of these things as an element in the content of belief, or indeed any form of intentionality at all, is declared to constitute its thinking of itself as itself. The phrase ‘as itself’ is thus stipulatively explained, indeed defined, in terms of what is effectively a first-personal element in content.

Of course, some acknowledge the Fregean move to be problematic. As we shall shortly see in a little more detail, Perry and Musholt recognise that it cannot be made for all forms of intentionality whose perspicuous characterisation requires the asterisked third-person pronoun. Both in the case of perceptual experience and, according to Perry at least, lowest-level perceptual belief; and in the case of the intentionality of non- or pre-linguistic beings: it is implausible to regard the pronoun as marking the presence of a first-personal element in content—of its referent’s ‘I’-guise, self-notion or self-concept. Moreover, some recognise that an account of the move is necessary. Thus, Perry attempts to explain what a self-notion is—lest one suspect that talk of it merely rebadges the notion of thinking of oneself as oneself.

It is not possible here to investigate these diverse ways of conceiving a distinctively first-personal element in content. Fortunately, it is also not necessary since the relevant point has already been made: in all cases non-accidental reference, that is, the kind of reference constitutive of the de se, does not alone entail self-consciousness qua thinking of an ‘I’-thought, whatever that might be. The fourth claim may therefore be put as follows: the asterisked third-person pronoun is not to be understood as itself indicating that its referent is thinking of itself as itself—not even in those higher-level cases to which Perry and Musholt restrict its ostensible indicatory function. In effect, the fourth claim denies that Castañeda’s ‘he*’ is, as this latter maintains, a pronoun of self-consciousness.26 And this is a very good thing because, as we shall see, the assumption that the pronoun itself intimates, even in merely some cases, that its referent is thinking of itself as itself, hence is self-conscious, engenders what Musholt calls the problem of self-consciousness.

We must, however, work our way slowly towards understanding and diagnosing the problem. Consider this perceptual experience:

The small pug sees a large pit bull bearing aggressively down on it*

For the very same reasons as in the case of agent A who believed that it* was about to be attacked by a bear, the asterisked third-person pronoun is also essential here. Yet as Perry and Musholt recognise, it would be implausible to maintain that in thus perceiving, the small pug is thinking of itself as itself. This has nothing to do with the fact that the pug is not a Kantian Vernunftwesen. The point is true of perceptual experience in general: when S perceives a tree standing in front of it*, then S, whoever and whatever S might be, does not think of itself as itself.

Perry and Musholt respond to this by denying that in the case of perceptual experience the asterisked third-person pronoun reflects an element in content at all (since this, they believe, could only entail that the referent of the pronoun is thinking of itself as itself). In this low-level case, content is, as they sometimes put it, self-less. That is, content involves no element which the pronoun would articulate. Rather, content is, to put things roughly,27 a propositional function, e.g., ‘A tree standing in front of (x)’, or, for current purposes equivalently, a property,28 that, namely, of a tree’s standing in front of something.29 In perceptual experience, what secures the character of non-accidental reference to the subject, hence character as de se, is simply psychological mode, that is, the simple fact of being perceptual experience.30

This application of a propositionally functional or property-theoretic conception of content to the most basic forms of perceptual experience encounters specific difficulties.31 The assumption, however, that at least in higher-order cases the asterisked third-person pronoun alone indicates that its referent is thinking of itself as itself, hence intimates some such thing as a self-concept, generates a general difficulty independent of whether one endorses Perry’s and Musholt’s account of low-level forms of intentionality. This is precisely the problem of self-consciousness. An analysis of the problem shows that in order to account for self-consciousness precisely as a thinking of oneself as oneself in the only coherent sense the phrase ‘as oneself’ can have, the asterisked third-person pronoun must not be understood as itself indicating such a thinking, which would require whatever element in content it marks to be some such first-personal element as a self-concept. Precisely in order to account for self-consciousness, one must not regard the asterisked third-person pronoun as even sometimes a pronoun of self-consciousness. This constitutes the fundamental justification for claim 4, hence also indirectly for claim 3.

Section II: The Problem of Self-Consciousness and Its Ground

Musholt writes that self-consciousness, that is, the thinking of ‘I’-thoughts, cannot be explicated in terms of the traditional subject-object model according to which “the subject takes itself as the object of reflection or (inner) perception.” (Musholt 2015, p. 3) So to subscribe to the subject-object model is to regard self-consciousness as a subject’s taking itself as an object either of reflection or, more specifically, of (inner) perception. Inner perception, whatever this is precisely, is therefore a particular kind of reflection32 and self-consciousness cannot be a form of reflection of this kind because it cannot be a form of reflection at all. “We cannot,” claims Musholt, “explain … how self-consciousness is possible … as long as we conceive of self-consciousness as a relation between a subject and an object” (Musholt 2015, p. 4)—as any kind of intentional relation between a subject and an object.

It seems, then, that the subject-object model makes three distinguishable moves: First it makes the general assumption that self-consciousness is some kind of reflection. It then makes the more specific assumption that this form of intentionality is reflection in no respect essentially different from the way ordinary perceptual experience is reflection: self-consciousness is about what it is about in no way essentially different from how perceptual experience is about what it is about. Lastly, it assumes that the first assumption entails the second: self-consciousness, because it is reflection, can only be reflection in the way perceptual experience is. To construe self-consciousness as any kind of intentionality at all in whose perspicuous expression of content there is reference to its subject is always already to assimilate it to perceptual experience.

This linking of the first assumption to the second is even clearer in Musholt’s source for her characterisation of the model: the Heidelberg School. Manfred Frank writes,

Self-consciousness cannot be understood as >objectual< consciousness33 [>gegenständliches< Bewusstsein], such as we have of external objects. Perceptual consciousness is objectual in this sense … . Such objectual consciousness we also call—in contrast to self-consciousness—»intentional« or »representational« because it directs itself towards something which is usually not it itself. (Frank 2015; my translation and italics added)34

The decisive move lies in the italicised sentence. Here, Frank contrasts self-consciousness with, hence distinguishes it from, anything intentional or representational on the ground that the latter can only ever be directed at something contingently identical with itself. Tacitly, Frank has assumed that not just the specific notion of perceptual consciousness, i.e., perceptual experience, but the very notion of intentionality and representation as such, exclude non-accidental reference.

In order to understand why Frank and Musholt make these claims, we must examine their account of what is wrong with the subject-object model. Why can we not explain “… how self-consciousness is possible …” if “… we conceive of self-consciousness as a relation between a subject and an object”? Because, writes Musholt,

a model that conceives of self-consciousness in terms of a relation between a subject and an object leads into a regress. … (F)or a subject to take itself as an object of its self-reflection, it needs to perform an identity judgement. In other words, to know that the object of my self-reflection is in fact myself, I need to self-identify. But [if self-consciousness is a self-reflection, as the subject-object model maintains,] this act of self-recognition in turn relies on another act of self-recognition, and so [on] ad infinitum. Alternatively, to stop the regress, the model must assume that I already possess self-consciousness, in which case the model is circular. (Musholt 2015, p. 4)

Musholt is citing the classic argument of the Heidelberg School, an argument understood by the latter both as an interpretation of Fichte and as a critique of the subject-object model in general.35 Note that in it there is no mention of perceptual experience. The argument has as its target the attempt to assimilate self-consciousness to any consciousness of something at all. It thus targets the first assumption constitutive of the subject-object model.

But what exactly is the argument telling us? That, on pain of either vicious circularity or infinite regress, self-consciousness cannot involve any kind of content in which an element occurs such that the linguistic articulation of this element constitutes a reference to an ‘object’, i.e., to a referent? Surely not! Musholt herself describes self-consciousness, the thinking of an ‘I’-thought, as explicit self-representation, which involves a so-called self-concept. And a self-concept is precisely an element in content whose linguistic articulation consists in reference to a referent, namely, the subject—in the case of characterisation by another, reference by means of the asterisked third-person pronoun, in the case of characterisation by the subject itself, reference by means of the first-person pronoun.

This obscurity results from imprecision, imprecision we can identify by independently showing what the argument is getting at when it claims that self-consciousness cannot be a form of reflection in the specific way perceptual experience is. Importantly, this independent demonstration has a bonus: it permits us to identify a sense in which self-consciousness might very well be, pace Frank and Musholt, an intentional relation of a subject to an ‘object’, i.e., a referent. Even more importantly, it permits us to identify this sense as already implicit in perceptual experience.

To model self-consciousness on perceptual experience is to construe it as about something in the way in which perceptual experience is about something. So how is perceptual experience about what it is about? Let us assume that a subject S sees a man, or perhaps the next door neighbour, or perhaps even John Perry skulking in the garden—this latter being the garden of S’s house, or perhaps simply the house in which S currently finds itself. We may certainly say that in these three cases, S’s perceptual experience is about such and such an individual. And it is about this individual as both (a) satisfying a certain property or propositional function; and (b) satisfying a certain relational property or function which implicates S. These are (a) the properties of being a man, being the next door neighbour or being John Perry; and (b) the relation of skulking in the garden of a certain house, perhaps one belonging to S, perhaps one in which S now happens to find itself. In general, the aboutness of perceptual experience is, at least typically, complex36: to perceive is typically to perceive something as an F, or as the F or as A, standing in some possibly complex relation R to the perceiver.

The relation R must, however, be characterised more precisely since otherwise one will fail to capture the causal role of perceptual experience. In particular, one must advert to the asterisked third-person pronoun: S perceives a man, or the next door neighbour, or John Perry skulking in the garden of its* house, or perhaps simply the house in which it* currently is. In general, for S to perceive something is for S to perceive it as an F, or as the F, or again as A, standing, however, in relation R to it*.37 This already explains why self-consciousness cannot be anything like perceptual experience: S can perceive something as an F, or as the F, or again as A, standing standing in relation R to it* without this latter’s being identical with S. Such difference is indeed the rule. So if self-consciousness were a form of ‘inner’ perception anything genuinely like ‘outer’ perception, then it would have to be a consciousness of something only ever accidentally identical with the subject whose self-consciousness it is. Self-consciousness would not possess non-accidental reference, that is, would not be about its subject simply in virtue of this latter’s being its subject.38 So self-consciousness would not be self-consciousness—in effect, a reductio ad absurdum.39

The reductio reveals, however, more: S’s perceptual experience, while certainly about a man, or about the next door neighbour or about John Perry, is not about the garden, the house or indeed S itself. Yet perspicuous characterisation of it involves reference to all three.40 We must distinguish, in a way the subject-object model and indeed the argument against it do not, between aboutness and reference.41 This obviously permits us to say that S’s perceptual experience is about one thing only, the individual skulking in the garden. Crucially, in so doing, it also permits us to say that although perspicuous characterisation of S’s perceptual experience involves reference to S,42 in no sense is S also perceiving itself!43

Just this points to how one might conceive self-consciousness, whatever it is precisely, “as a relation between a subject and an object” (Musholt 2015, p. 4), indeed, as a consciousness of something in a perfectly acceptable sense of the term. Notice that in her presentation of the classic argument Musholt speaks of “… a subject [taking] itself as an object of its self-reflection.” This can only be the engendering of aboutness, not reference. So the classic argument actually only shows that self-consciousness cannot be about the subject in the manner of perceptual experience. It does not rule out that in self-consciousness—more precisely, in the perspicuous linguistic characterisation of self-consciousness—there is not only reference to the subject, but reference in the same sense in which in the perspicuous linguistic characterisation of perceptual experience there is reference to the subject.

Indeed, the argument does not in principle rule out that what accounts for this reference in both perceptual experience and self-consciousness is (a) an element in content; and (b) the same kind of element. More precisely, it does not rule this out provided one rejects a certain presupposition, one which turns an argument showing merely that self-consciousness is not about anything in the manner of perceptual experience into an argument that self-consciousness cannot be “a relation between a subject and an object” in any sense at all. This presupposition forces the conclusion that either one takes self-consciousness as inexplicably primitive—one can only say such things as that it is the thinking of oneself as oneself, the thinking of an ‘I’-thought, etc.—; or one attempts to explicate it in terms of perceptual experience. The presupposition does this because it denies one the means of explicating self-consciousness. Failure to distinguish between aboutness and reference then intervenes to make perceptual experience look like a viable or perhaps rather the only remaining means.44 So this presupposition explains why the subject-object model assumes that to construe self-consciousness as any kind of “relation between a subject and an object” at all is to construe it as ‘inner’ perception. That is, it explains why the model ties its first assumption to the second.

What then is this presupposition? As just shown, one can appeal to an element in content in order to explain the necessity of the asterisked third-person pronoun in the characterisation of perceptual experience without absurdly maintaining that in perceiving the perceiver perceives itself. One need not, therefore, embrace Perry’s and Musholt’s propositionally functional or property-theoretic explanation of the necessity, at least not for this reason. But this is only true if the element is not some such first-personal item as a self-concept since were it this, a perceiving subject would be thinking of itself as itself.45 Or, to put the same point in another way, it is only true if the necessity of the asterisked third-person pronoun does not ground, even merely in some cases, in its referent’s thinking of itself as itself, hence if the pronoun does not itself`, even in merely some cases, indicate this and so is not ever, pace Castañeda, subject to a pragmatic rule of attributing indexical reference.46 In short, all this is only true if the fourth claim of the thesis of intentionality inherently de se is true. The presupposition is thus the negation of claim 4: the presumption that intentionality de se is eo ipso a matter of its subject thinking of itself as itself.

Musholt presumes this. This is shown by her tacit assumption that the only element in content the asterisked third-person pronoun could possibly express is some such first-personal item as a self-concept. Of course, at this point, one might object that only if this element is some such first-personal item can it play the role it needs to play in an account of self-consciousness, in which case it cannot also be implicated in perceptual experience. To which I reply, “Not so!” There is no need to think that in the case of self-consciousness the asterisked third-person pronoun does anything other than what it does in the case of perceptual experience. In both cases it simply expresses an element in content ensuring non-accidental reference. There is only inability to do the same thing in both cases if one regards the pronoun as intimating that its referent is thinking of itself as itself. Everything about the classic argument against the subject-object model, from obscurity through to claiming that self-consciousness is not an intentional relation while appealing to self-concepts and the like, grounds in this presumption—a presumption which it is the point of the fourth claim to deny.

So in order to understand what self-consciousness is—what is meant by talk of ‘I’-thoughts or of one’s thinking of oneself as oneself—, we must understand the asterisked third-person pronoun as doing no more or less than what it does, at least arguably, in the case of perceptual experience: expressing an element in content ensuring non-accidental reference to the subject without, however, itself entailing that the subject is thinking about itself as itself. This character of being about oneself as oneself, clearly in some sense an undeniable aspect of self-consciousness, must come from elsewhere. How might this be so?

Conclusion: Thinking Better about Thinking about Oneself*

Self-consciousness is, Musholt tells us, the ability to think ‘I’-thoughts. It is one’s thinking of oneself as oneself. These kinds of claim suggest that self-consciousness is primarily either something occurrent or the capacity for such occurrence. Let us now break with the insouciance of tradition and ask what occurrence. We need an example to which we can phenomenologically resonate. So let us recall what happens to Perry in the supermarket:

Finally it dawned on me. I was the shopper I was trying to catch. (Perry 1979, p. 27/p. 3)

One can really only understand this dawning as a matter of the sentence “I am the shopper I am trying to catch!” running assertorically through Perry’s head, in an act of what Musholt calls inner speech.47 Note how this example shows inner speech to be neither necessarily voluntary nor necessarily fatuous: the situation wrings Perry’s insight from him and it is an insight. The content presenting itself to Perry in this dawning, namely, that he* is the shopper he* is trying to catch, makes sense of his immediately past behaviour. It explains the latter’s futility, thereby pointing to what he needs to do now. In so doing, it gives him the belief he needs in order to solve the problem of someone’s making a mess, the belief, namely, that he* is making a mess. Thereupon he re-arranges the items in his shopping trolley.

So what is inner speech? Here I can only propose a bald hypothesis: it is a matter of a subject’s imagining itself* speaking, typically, asserting. Explicitly self-conscious thinking of the kind illustrated by Perry’s dawning is therefore a subject’s imagining itself* asserting some first-personal sentence. In general, a subject S thinks the ‘I’-thought “I am F” just in case S imagines itself* asserting that it* is F. Note that this does reconstruct at least one good sense in which one can think of oneself as oneself: in general, to assert a sentence is to know that one* is asserting such and such about such and such. So to assert a first-personal sentence is to know that one* is asserting such and such about oneself*.

Such inner speech is clearly the thinking of an ‘I’-thought in one good sense of the term. It is thus at least a form of self-consciousness. I would maintain, however, that it is the only form; it just is self-consciousness. But I cannot argue for this here. So let me just note a crucial feature of the account: the phrase ‘as oneself’ is explained in terms of the asterisked third-person reflexive pronoun and not the other way around. The account thus presupposes that the asterisked third-person pronoun does not itself entail that S is thinking of itself as itself, in other words, that it does not simply abbreviate, as Castañeda maintains, the phrase ‘it (as) itself’. Rather, the character of self-consciousness as a thinking about oneself as oneself derives from elsewhere, namely, from the character of the whole occurrence as an imagining oneself* asserting a first-personal sentence. The account thus positively demands the truth of the fourth claim made by the thesis of intentionality inherently de se.

The claim that first-personal inner speech is not simply one but the form of self-consciousness creates a problem. What about so-called pre-reflexive self-awareness? The Heidelberg School and, following them, Musholt take the classic argument against the subject-object model to show that underpinning self-consciousness is some pre-, hence non-reflexive, pre-, hence non-intentional form of self-awareness. We need not ask whether inference to a non-reflexive form of reflection, or again, to a non-intentional form of intentionality is not simply inference to an oxymoron.48 For the inference is itself illicit, something more evident in Musholt’s case since she, it seems, takes the classic argument directly to entail that self-consciousness presupposes some such primordial form of self-awareness in which there is no element in content referring to the subject of the awareness.49

This is surely not right. The classic argument claims to show that one may not conceive self-consciousness, on pain of circularity or regress, as a relation of a subject to an object. Let us ignore how unclear the argument is since there is one thing it cannot be: a demonstration of the conditions under which self-consciousness is possible. How could it be? The argument is simply a refutation. It takes us back to the drawing board and no further.

In the case of the Heidelberg School the matter is less straightforward because more obscure. The Heidelberg School also seems to regard the classic argument as intimating the need for some primordial form of self-awareness underpinning explicit self-consciousness. Yet it also makes another move, one not clearly distinguished from this dubious one. For it argues that some primordial form of self-awareness is a condition of the possibility, not of self-consciousness but of consciousness. By consciousness the Heidelberg School means consciousness of an object, i.e., any form of intentionality possessed consciously. So its appeal to a primordial form of self-awareness is motivated at least also by the idea that one needs such self-awareness in order to account for the character of certain intentional states and experiences as conscious. With this, the Heidelberg School shows itself to be one more camp in a long tradition of maintaining that consciousness entails self-consciousness,50 alongside the so-called higher-order (HOT)51 and same-order (SOT) theories of consciousness.52 In fact, the Heidelberg School belongs to the phenomenological camp of Sartre and the early Heidegger, latterly represented by Dan Zahavi.53

But why do so many think that the character of certain intentional states and experiences as conscious must be construed as some kind of awareness of them had by their subject as its*? I suspect that this is because one is assuming, firstly, that the asterisked third-person pronoun must be used in any account of this character as conscious; and secondly, that the pronoun intimates that its referent is thinking of itself as itself. First one recognises that character as conscious consists in there being something it is like for a subject S to be in a conscious intentional state or experience, more precisely, something it is like for it* to be in this state or experience. Then the second assumption kicks in, prompting a hunt for some plausible sense in which this character implicates a thinking of oneself as oneself. But there is no such sense; the second assumption is just false.

Let me end with a hand-waving general remark. The account of self-consciousness sketched here might seem deflationary and in a way it is. In another way, however, it is not, this for two reasons: firstly, it entails that self-consciousness presupposes language and the first person in particular. Secondly, its appeal to imagination (Einbildungskraft) suggests that self-consciousness has an anschaulichen character ultimately grounded in perceptual experience. To this extent, it in no way dislodges self-consciousness from that “highest point to which one must attach all use of the understanding, indeed all logic and, along with it, transcendental philosophy.”54

References

Anscombe, G. (1975) “The First Person,” in Mind and Language, Wolfson College Lectures 1974, edited by Samuel Guttenplan, Oxford University Press, Oxford, pp. 45-65

Brentano, F. (1874/1973) Psychologie vom Empirischen Standpunkt, Vol. 2, Meiner Verlag, Hamburg

Cappelen, H., and Dever, J. (2013) The Inessential Indexical, Oxford University Press, Oxford

Castañeda, H-N. (1966) “‘He’—A Study in the Logic of Self-Consciousness,” in Castañeda 1999a, pp. 35-60

Castañeda, H-N. (1967a) “Indicators and Quasi-Indicators,” in Castañeda 1999a, pp. 61-88

Castañeda, H-N. (1987b) “The Self and the I-Guises, Empirical and Transcendental,” in Castañeda 1999a, pp. 180-203

Castañeda, H-N. (1989) Thinking, Language, and Experience, University of Minnesota Press, Minneapolis

Castañeda, H-N. (1998) “She*: Pragmatically Imparted or Semantically Encoded?,” in Orilia and Rapaport 1998, pp. 217-234

Castañeda, H-N. (1999a) The Phenomeno-Logic of the I: Essays on Self-Consciousness, edited by James G. Hart and Tomis Kapitan, Indiana University Press, first edition

Castañeda, H-N. (1999b) “I-Structures and the Reflexivity of Self-Consciousness,” in Castañeda 1999a, pp. 251-292

Chisholm, R. (1981) The First Person, University of Minnesota Press, Minneapolis

Christensen, C. (2007a) “What are the Categories in Sein und Zeit? Brandom on Heidegger on Zuhandenheit,” in European Journal of Philosophy, Vol. 15, No. 2, pp.159-185

Christensen, C. (2007b) “Nichts Neues unter der Sonne: Bewußtsein und Selbstbewußtsein bei Paul Natorp,” in Kant-Studien, Vol. 98, No. 3, pp. 372-398

Feit, N. (2008) Belief about the Self: A Defense of the Property Theory of Content, Oxford University Press, Oxford

Feit, N., and Capone, A., eds., (2013) Attitudes de se: Linguistics, Epistemology, Metaphysics, CSLI Publications, Stanford

Frank, M. (2015) Präreflexives Selbstbewusstsein—Vier Vorlesungen, Reclam, Stuttgart

Frege, G. (1918) “Der Gedanke,” in Frege 1976, pp. 30-53; translated in Mind, Vol. 65, No. 259, 1956, pp. 289-311

Frege, G. (1976) Logische Untersuchungen, edited and introduced by Günter Patzig, second, extended impression, Vandenhoeck & Ruprecht, Göttingen

Haddock, A. (2016) “Self-Identification and Its Significance,” unpublished paper held in April 5th, 2016, in the seminar of the Philosophical Society of the School of Philosophy at the Australian National University

Heidegger, M. (1979) Sein und Zeit, 15.te Auflage, Max Niemeyer Verlag, Tübingen

Heidegger, M. (1987) Zollikoner Seminare: Protokolle–Gespräche–Briefe, edited by Medard Boss, Vittorio Klosterman, Frankfurt am Main

Henrich, D. (1967) Fichtes Ursprüngliche Einsicht, Vittorio Klostermann, Frankfurt am Main (First Edition 1966)

Jaszczolt, K. (2013) “Contextualism and Minimalism on de se Belief Ascription,” in Feit and Capone 2013, pp. 69-104

Kant, I. (1781-87/1976) Kritik der reinen Vernunft, in Immaneul Kant—Werke, Bde. III und IV, edited by W. Weisschedel, Suhrkamp Verlag, Frankfurt am Main

Kaplan, D. (1989) “Demonstratives—An Essay on the Semantics, Logic, Metaphysics, and Epistemology of Demonstratives and Other Indexicals,” in Themes from Kaplan, edited by J. P. Almog, Oxford University Press, Oxford, pp. 481-614

Kriegel, U. (2003) “Consciousness as Intransitive Self-Consciousness: Two Views and an Argument,” in Canadian Journal of Philosophy, Vol. 33, No.1, pp.103-132

Lewis, D. (1979) “Attitudes De Dicto and De Se,” in The Philosophical Review, Vol. 88, No. 4, pp. 513-543

Loar, B. (1976) “The Semantics of Singular Terms,” in Philosophical Studies, Vol. 30, No. 6, pp. 353-377

Liu, JeeLoo, and Perry, J. (2012) Consciousness and the Self—New Essays, Cambridge University Press, Cambridge

McGinn, C. (1983) The Subjective View: Secondary Qualities and Indexical Thought, Oxford University Press, Oxford

Musholt, K. (2015) Thinking about Oneself—From Nonconceptual Content to the Concept of the Self, The MIT Press, Cambridge, Massachusetts

Orilia, F., and Rapaport, W., eds., (1998) Thought, Language, and Ontology—Essays in Memory of Hector-Neri Castañeda, Kluwer Academic Publishers, Dordrecht

O’Rourke, M., and Washington, C. (2007) Situating Semantics: Essays on the Philosophy of John Perry, The MIT Press, Cambridge, Massachusetts

Perry, J. (1977) “Frege on Demonstratives,” in The Philosophical Review, Vol. 86, pp. 474-497; also in Perry 2000, pp. 1-26

Perry, J. (1979) “The Problem of the Essential Indexical,” in Noûs, Vol. 13, No. 1, pp. 3-21; also in Perry 2000, pp. 27-44

Perry, J. (1980) “Belief and Acceptance,” in Midwest Studies in Philosophy, Vol. 5, pp. 533-542; also in Perry 2000, pp. 45-56

Perry, J. (1986) “Thought without Representation,” in Perry 2000, pp. 171-188

Perry, J. (1990) “Self-Notions,” in Logos: Philosophical Issues in Christian Perspective, Vol. 11, pp. 17-31; also available at http://john.jperry.net/cv/1990c.pdf

Perry, J. (2000) The Problem of the Essential Indexical and Other Essays, expanded edition, CSLI Publications, Stanford, California

Price, H. (2007) “The Effective Indexical,” unpublished ms., Centre for Time, University of Sydney, http://philsci-archive.pitt.edu/4487/

Recanati, F. (2007) Perspectival Thought: A Plea for (Moderate) Relativism, Oxford University Press, Oxford

Rosenthal, D. (2012) “Awareness and Identification of the Self,” in Liu and Perry, pp. 22-50

Spencer, C. (2007) “Is there a Problem of the Essential Indexical?,” in O’Rourke and Washington 2007, pp. 179-197

Zahavi, D. (2004) “Back to Brentano?,” in Journal of Consciousness Studies, Vol. 11, No. 10-11, pp. 66-87

Zahavi, D. (2007) “Subjectivity and the First-Person Perspective,” in Southern Journal of Philosophy, Vol. 45, pp. 66-84

Zahavi, D. (2009) “Is the Self a Social Construct?,” in Inquiry, Vol. 52, No. 6, pp. 551-573

Notes

  1. Incidentally, since in this sentence reference to that of which S believes that it is F is distinct from reference to that which believes, the sentence could have been formulated rather differently: “S believes of S that the latter is F.” This is actually clearer in that it explicitly indicates to which occurrence of the singular term ‘S’ the anaphoric link extends. For this reason, it is also clearer philosophically: it better indicates what is going on within the belief itself. The possibility, indeed the preferability, of using the term ‘the latter’ instead of the pronoun ‘it’ shows that when the pronoun is used, it is not functioning as a variable. Rather, the pronoun is to be understood as borrowing the referent (but crucially not the reference!) of some singular term in the relevant linguistic context, here, of some singular term in the one sentence. And as thus borrowing its referent, it is, as the preferability of the term ‘the latter’ shows, functioning as a parasitic singular term and not at all as a variable. The assumption that the pronoun ‘it’ is functioning as a variable, hence that the nominal clause ‘that it is F’ is somehow articulating or expressing a propositional or predicate function, is a long-standing and tendentious prejudice reaching back at least to Kaplan 1969. ↩

  2. This first claim is widely (if not universally) endorsed but all who endorse it do not do so in the same way. In particular, not all endorse it as radically as I do. Thus, Loar 1976, Lewis 1979, Chisholm 1981, Recanati 2007 and Feit 2008 claim that while sentences of the form “S believes that it* is F” are ineliminable, the forms of intentionality they characterise distinctive, they also claim that these sentences are slightly misleading. For according to them whenever a sentence of this form is true, the subject S is not standing in a relation of belief to some content traditionally called a proposition but is in fact self-ascribing the propositional function ‘x is F’ or again, for current purposes equivalently, the property of being F.
    Lewis, Chisholm and Feit go on, of course, to extend this propositionally functional or property-theoretic move to all sentences, even those of the form “S believes that a is F”; others reserve it solely for forms of intentionality inherently de se. Thus, Perry and Musholt maintain that at least some sentences in which the asterisked third-person pronoun occurs may be interpreted as the self-ascribing of propositional functions or properties. According to Perry, the simplest forms of perceptual experience and belief, such as one might characterise with the sentences “S perceives a tree standing in front of it*” or “S believes that a tree is standing in front of it*,” are such self-ascribings—see Perry 1986, p. 179. Musholt agrees, at least as far as perceptual experience is concerned. By contrast, I, along with Frege and Castañeda, reject the propositionally functional or property-theoretic interpretation of intentionality de se and the sentences which characterise it entirely. ↩

  3. In the everyday hence ‘Humean’ sense of causal role, of course, not in whatever, say, neurophysiological sense there might be. ↩

  4. This is to assume, at least for intentional states and experiences, the so-called Eleatic Principle. That is, it is to assume, at least for intentional phenomena, that “anything has being that is so constituted as to possess any sort of power either to affect anything else to to be affected, in however small degree, by the most insignificant agent, though it be only once.” (Plato, The Sophist, 247e) What a particular intentional state or experience objectively is, in and of itself, is determined by its content. And what its content objectively is, in and of itself, entails, indeed grounds—only together, of course, with the kind of intentional state or experience whose content it is—a particular causal potential or role. ↩

  5. At least insofar as the agent is only able to effect change in the world on the basis of perceptual experience and through a spatiotemporally located body. According to Cappelen and Dever, a being is conceivable, perhaps some kind of supremely potent if not necessarily omnipotent divinity, who can simply will states of affairs into existence—see Cappelen and Dever 2013, p. 37. It would, for example, simply will that there be world peace and there would be world peace. Such a being would not need to be embodied or perhaps even spatiotemporally located (although it could be both, of course). Cappelen and Dever maintain that such a being would not require intentionality de se.

    Whether such a being really is possible is perhaps arguable. More importantly in the current context, why should we think that its capacity simply to will states of affairs into existence entailed no need for intentionality de se? Would not it have to believe at t that it* could bring about world peace then* and desire at t that it* bring about world peace then*? In particular, would it not have to be able to know that it* has brought about world peace then*? The italicized pronoun ‘it*’ and the italicized adverb of time ‘then*’ are intended here to be understood as ineliminable, this in virtue of their contributing to a perspicuous characterisation of what the relevant forms of intentionality inherently are. The qualification made here concerning perceptually-based knowledge of the world and spatiotemporally located, embodied existence in the world is made simply in order not to get bogged down in speculative theology. ↩

  6. And given, too, that intentionality alone is sufficient for purposive behaviour, an assumption disputed by Cappelen and Dever in their critique of claim 2—see Cappelen and Dever 2013, esp. Ch. 3. ↩

  7. Most notably by John Perry and David Lewis in their well-known papers from 1979. But since the publication of these two papers numerous authors have endorsed their conclusions more or less without question, e.g., McGinn 1983, Price 2007, Spencer 2007 and many more. Cappelen and Dever detail how widespread the conviction is—see Cappelen and Dever 2013, Ch. 1, § 3. ↩

  8. Note that one may not argue that Lewis does not believe this, for Lewis never makes clear what his position is. Moreover, in general, adherents of the property-theoretic or propositionally functional conception of content speak just as freely and exclusively of first-person and indexical thought, which they typically characterise just as readily oratio recta rather oratio obliqua, as anyone else. So at most one may argue only that Lewis and other adherents of this conception need not believe it. ↩

  9. To his credit, Recanati both clearly recognises this commitment and makes an attempt to spell this strong sense out—see Recanati 2007, pp. 13-14. But what Recanati characterises as a notion of ‘indexical’ content—of what he calls mental indexicality—is in fact the notion of a content whose truth or falsity is relative to different times and locations. There is nothing indexical about the latter kind of content. This confusion goes hand in hand with another: the claim that because something is true or false relative to such and such, it is a propositional function, i.e., something incomplete, rather than a proposition, which is complete. ↩

  10. Castañeda describes the reference accomplished by an asterisked third-person pronoun such as his artificial pronoun ‘he*’ as “a … vicarious presentational way of referring to objects not necessarily present to us through representation of their presence to others.” (Castañeda 1989, p. 5; italics added) Following Goodman, he says, he will “… call indicators the personal and demonstrative pronouns and adverbs like ‘this,’ ‘that,’ ‘I,’ ‘you,’ ‘here,’ ‘there,’ and ‘now,’… when they are used purely referentially either to single out an item present in the speaker’s current experience or to pinpoint a self that is a relatum in the cognitive relation evinced by the speaker. I call quasi-indicators the expressions which in oratio obliqua represent uses, perhaps only implicit, of indicators, i.e., uses which are ascribed to some person or persons by means of a cognitive [or in some other sense psychological] or linguistic verb. For example, suppose that at a certain place p and time t, A says to B: (i) “I am going to kill you here now.” Suppose that C reports this by asserting: (2) “A said to B at p at t that he was going to kill him there then.” Sentence (2) contains the quasi-indicators ‘he,’ ‘him,’ ‘there,’ and ‘then,’ which represent uses, ascribed to A, of the indicators ‘I,’ ‘you,’ ‘here,’ and ‘now,’ respectively. … (T)he very same marks or noises ‘I,’ ‘you,’ ‘here,’ ‘there,’ ‘he,’ ‘this,’ function as either indicators, quasi-indicators, or something else. It is a mere accident of grammar that the same physical objects are used in different logical roles. The underlying rationale is this: indicators are a primary means of referring to particulars, but the references made with them are personal and ephemeral; quasi-indicators are the derivative means of making an indexical reference both interpersonal and enduring, yet preserving it intact.” (Castañeda 1967a, pp. 61-62) ↩

  11. This is, of course, not to deny the phenomenon which leads Castañeda to speak of quasi-indication; rather, it is a call to understand the phenomenon properly. Castañeda’s sentence “The Editor of Soul believes that he himself, i.e., he*, is a millionaire” certainly does indicates that the Editor of Soul is actually able to use the first-personal sentence “I am a millionaire” in a self-conscious assertion to the effect that he* is a millionaire. But it does so not (just) because it contains the expression ‘he himself’, i.e., ‘he*’, but because it (also) contains the expressions ‘the Editor of Soul’ and ‘… is a millionaire’: only self-conscious beings can be editors and only self-conscious beings can, at least typically, be millionaires. So while the apposition of ‘himself’ to the third-person pronoun ‘he’ does mean that anyone who asserts this sentence literally intends to indicate both that the Editor of Soul is able to wield the first-person pronoun in an assertion of the content of his belief; and that the Editor of Soul is self-consciously aware of being a millionaire and indeed self-conscious in general: this is simply a matter of emphasising or drawing attention to something already implicit in the whole sentence “The Editor of Soul believes that he is a millionaire.”

    This, incidentally, intimates how one must reply to the claim that the asterisked third-person pronoun need not be understood as functioning as a quasi-indicator in all cases of its use and indeed that Castañeda, Perry and everyone else have never understood it as thus functioning in all cases. For in what cases is it functioning as a quasi-indicator? One can only answer that it is thus functioning in all and only those cases of its use in which the whole sentence, in virtue of its meaning and in particular, in virtue of the referent of the term to which it is anaphorically linked indicates that this latter either actually is or actually could wield the first-person pronoun. Then, however, its function of quasi-indication is redundant; there is no such function or, as Castañeda terms it, pragmatic rule—see Castañeda 1998, p. 232. ↩

  12. This might seem counterintuitive: surely if, to use Castañeda’s example, the Editor of Soul believes that he* is a millionaire, then he actually does know how to assert, “I am a millionaire,” and he actually is self-consciously aware, not the least of the fact that he* is a millionaire. Yes, he does and yes, he is—but not simply because his belief is de se but also because it involves the property of being a millionaire, something one can only be, at least paradigmatically, if one is self-conscious and can wield the first-person pronoun. See in this connection note 11. ↩

  13. Or in virtue of S’s belonging to the so-called context of belief. ↩

  14. In other words, Musholt believes that to grasp a thought de se in the sense of having it is to grasp it in the sense of knowing it. This points to a crucial ambiguity in Frege’s notion of das Fassen eines Gedankens. ↩

  15. That the Editor of Soul believes that he* is a millionaire does, of course, implicate actual self-consciousness and (arguably therefore) also actual ability to wield the first-person pronoun—in, for example, an assertion of the sentence, “I am a millionaire.” But one can and, I argue, must construe this as a consequence of what the sentence means in total; it is not an accomplishment of the asterisked third-person pronoun alone. See once again in this connection note 11. ↩

  16. And ultimately so, too, does Anscombe—see Anscombe 1975, p. 47. For according to her Castañeda’s ‘he’ presupposes ‘I’ and the latter sets itself apart from her curious expression ‘A’ precisely through the character of its use as expressing self-consciousness. Lewis’s thought-experiment of the two gods suggests that he, too, at least as a matter of brute fact, associates the *de se with actual self-consciousness. But the matter is not clear. ↩

  17. See Castañeda 1967a and 1966 respectively. ↩

  18. ““He” tells us X thinks of X as himself.” (Perry 1980a, p. 538/p. 51) Perry, however, also appreciates that this cannot apply in the case of non- or pre-linguistic beings. He initially tries to resolve this problem by regarding intentionality de se as only entailing a (capacity for the) use of the first-person pronoun, hence self-consciousness, in the weak sense: If S believes that it* is F, then, if S is in general capable of self-ascribing its intentionality and if S wishes to characterise its* belief perspicuously, then S must assert and will indeed be disposed to assert, “I believe that I am F.” (Note that the assertion at issue may be ‘mental’, that is, the kind of thing nicely exemplified by that event of dawning constitutes Perry’s recognising, “I am the shopper I am trying to catch!”)

    Unfortunately, this dispositional account—Perry later denies that it is an analysis (Perry 1980, p. 55)—is too weak: it presupposes, hence cannot be used to explicate, the accomplishment of the asterisked third-person pronoun since this must be used in any explication of the first-person pronoun. This point is simply an application of the general idea that the dispositional character of a phenomenon is grounded in its structural, hence non-dispositional constitution. Perry eventually recognises this and introduces, on the one hand, self-notions in order to deal with what Musholt calls explicit self-representation; and, on the other, borrows from the propositionally functional or property-theoretic view of content in order to deal with perceptual experience and also, it seems, low-level perceptual belief. In this way, too, he presumably hopes to accommodate non-linguistic animals and pre-linguistic children. ↩

  19. Whose other members include Dieter Henrich and Ulrich Pothast. The founding text of the School, in which the argument is first presented, both as an interpretation of Fichte and as a critique of the subject-object model in general, is Henrich 1967. ↩

  20. See Castañeda 1999b, p. 264f. This intimates, incidentally, how Castañeda might explain the fact that when the pronoun ‘he’ is used as ‘he’, it often occurs in conjunction with the matching third-person *reflexive pronoun, in the complex pronominal phrase ‘he himself’: the reflexive pronoun does not merely perform the function of underscoring that the form of intentionality at issue is de se; it also functions as an abbreviation of the phrase ‘he in his capacity simply as himself’. Indeed, Castañeda might claim that it only thus emphasises because it thus abbreviates. ↩

  21. Castañeda is quite explicit about this: see Castañeda 1999b, p. 258. ↩

  22. Needless to say, the propositional function ‘x is identical with S’ will not do; it articulates what S is thinking of when S believes merely that S is F. ↩

  23. See Frege 1918, p. 39. Castañeda reveals very nicely what this move really involves: “Self-consciousness,” he writes, “is executed in episodes of thinking about oneself qua oneself. The contents thought in such episodes are expressed in natural language in utterances of sentences containing (at least apparently) singular-referring uses of the first-person pronoun. The internal reflexivity of self-consciousness is the appearing of the thinker to HIMself as himSELF, that is, as an I. Self-consciousness is I-consciousness. Not because we identify I-consciousness with the use of the pronoun ‘I’; but because uses of ‘I’-sentences reveal the speaker’s thinking I-contents, and, consequently, his having a brain I-representation. A thinking episode is not an event of uttering. It is embodied in—indeed, in some appropriate sense of ‘sameness’, a thinking episode is the same as—an event or process in the thinker’s brain, or thinking box, whatever this may be.” (Castañeda 1999b, p. 257) So the appearing of a thinker to ITself as itSELF is original, that in which the first-person pronoun, precisely as a first-person pronoun which one must distinguish from, e.g., Anscombe’s letter ‘A’ (Anscombe 1971, pp. 49-50), is grounded and not conversely. ↩

  24. See, e.g., Castañeda 1987b. ↩

  25. See, e.g., Perry 1990 and Musholt 2015, esp. pp. 123-124. ↩

  26. See Castañeda 1966, pp. 36-41. ↩

  27. Musholt puts things in this rough fashion, that is, without making clear that in thus speaking, one is no longer using the old notion of content according to which content is what is expressed by a nominal clause—see Musholt 2015, p. 100. Perry comes close to this in Perry 1986, p. 178. ↩

  28. Evidently, this brings them at least close to the views of Loar, Lewis, Recanati and Feit—see Loar 1975, Lewis 1979, Recanati 2007 and Feit 2008. These views cannot, of course, be discussed here. ↩

  29. Musholt even suggests that it could be simply the property of standing in front of a tree—see Musholt 2015, p. 80. ↩

  30. Perry’s and Musholt’s differential treatment of the asterisked third-person pronoun as marking an element in content in some cases, viz., cases of what Musholt calls explicit self-representation; and not doing so in other cases, namely, in those ‘self-less’ forms of perceptual experience, leads them to say that the latter cases provide information concerning the perceiving subject but not about it. At the same time, they do not deny that we, in our characterisation of such experience, must use the asterisked third-person pronoun. In this sense, the character even of such low-level intentionality as involving the non-accidental reference distinctive of the de se is preserved. ↩

  31. Thus, one might ask whether the application of the idea to certain forms of intentionality de se and not to others does not make the account of the asterisked third-person pronoun unacceptably equivocal—for how does its use in the characterisation of those forms in which the subject is “thinking of itself as itself” relate to its use in the characterisation of those other forms in which the subject is not? Perhaps more importantly, how could the psychological mode of perceptual experience ever combine, in whatever way, with a propositional function or property to constitute a form of intentionality which requires for its ordinarily if not philosophically perspicuous characterisation the asterisked third-person pronoun?

    Perry attempts to explain this by appeal to what he calls epistemic and pragmatic methods, Musholt by a similar appeal to know-how. Indeed, according to Musholt, in the case of such intentionality, content is know-how—see Musholt 2015, pp. 37-42, p. 61 and p. 68. This is surely unsatisfactory: one’s knowing how to deal with what one objectively sees presupposes, hence cannot be identified with, the perceptual experience which mobilises one’s know-how. In general, the functional or causal role of perceptual experience is precisely to select between which of one’s various know-hows it is appropriate to mobilise, given what one perceives—in which case perceptual experience cannot just be a know-how or any collection thereof. Nor does it help to appeal, as Musholt does, to Gibsonian notions of affordance. There is no reason in principle why one could not regard perceptual experience as both inherently and irreducibly a form of intentionality in which various entities are quite literally represented, possibly even conceptually; and as affording (or inhibiting) this or that. Indeed, this is how, at least according to Heidegger, the very first thinker explicitly to link perceptual experience to affordance, viz., Augustine, conceived the matter: Augustine speaks of perceptual experiences as having adversa and prospera as their contents. See Christensen 2007a.

    Yet one specific difficulty deserves some mention since it is indirectly relevant to the so-called problem of self-consciousness. What goes for the pronoun must also go for the propositional function or property: if the former does not reflect an element of content but merely refers for us, then neither does the latter reflect an element of content but merely how we, using our concepts, describe S’s perceptual experience and its content. Perry uses his example of the Z-landers and their assertions and beliefs about the weather to show that it is possible to assert or believe propositional functions. But Perry’s own example shows that we only recognise them as believing the propositional function ‘Raining at (t, l)’ (or the corresponding relation of raining at a time and location) because we have always already recognised them as wielding their own more specific propositional function ‘Raining-where-one*-is at (t)’ (or property of raining-where-one*-is at a time), i.e., their own more specific concept of raining-where-one*-is at a time—wielding it in the asserting or believing of a perfectly complete, hence to this extent fully propositional content. In short, relative to their own conceptual framework, the contents of Z-landers’ assertions and beliefs are not at all propositional functions but propositions—propositions, moreover, absolutely true or false, hence not those relativised propositions with which Perry occasionally confuses, others, e.g., Recanati 2007, systematically conflate, propositional functions. (The contents of Z-landers’ assertions and beliefs about the weather are absolutely, not merely relatively true or false.) The upshot of this is that insofar as a propositional function or property can be a content of assertion or belief at all—and Perry does speak of “taking the propositional content of [Z-landers’] beliefs to be propositional functions, rather than complete propositions … ” (Perry 1986, p. 179)—, this can only be so in a sense which entails that either these subjects are not aware of their contents as their contents or that these subjects are only derivatively aware of them, this because they have always already recognised that full propositional content from which the propositional function has been abstracted. ↩

  32. So by her use of the word ‘or’ Musholt appears to mean genuine disjunction, not mere apposition. ↩

  33. Frank is using the term ‘consciousness’ in a very non-analytic, German way: it connotes any intentional state or experience with the property of being consciously possessed or undergone. ↩

  34. Elsewhere Frank writes, “Intentionality commits us to a clear and actualised subject-object distinction.” (Frank 2015, p. 72) This only possesses the non-trivial meaning Frank clearly intends it to have if it is understood as maintaining that something is intentional just in case if linguistic articulation of its content involves reference to something x, then x is not necessarily identical with its subject. ↩

  35. In fact, this kind of argument has been deployed by a number of other thinkers since Fichte, e.g., Herbart and Natorp. See in this connection Christensen 2007b. ↩

  36. Whether it is not merely typically but in fact always complex in the sense indicated is not relevant for current purposes. ↩

  37. This is entirely consistent with the view, held by both Musholt and Perry, that in the case of perceptual experience what explains the necessity of the asterisked third-person pronoun is not some element in the content of perceptual experience (since according to them in this case there is no such element). ↩

  38. Or, as one also might put, in virtue of this latter’s occurring in the context of intentionality. ↩

  39. This reductio is certainly implicit in the classic argument of the Heidelberg School. ↩

  40. Analogously, the phrase ‘the Prime Minister of Australia’ possesses reference both to a certain individual and to a certain country but it is only about the former, as is demonstrated by its use in the sentence “The Prime Minister of Australia is a marketing man.” ↩

  41. As both Musholt and the Heidelberg School fail to do. ↩

  42. Note that nothing in this rough-and-ready account of what is going on in perceptual experience entails that the reference accomplished by either the asterisked third-person pronoun (in other-attribution) or indeed the first-person pronoun (in self-attribution) reflects an element in the content of the experience. Thus far, the account is neutral on this score. This neutrality is crucial because Musholt believes that in the case of perceptual experience the necessity of either pronoun does not reflect an element of content. Precisely in the case of low-level perceptual experience, she argues, what makes either pronoun necessary is not content but mode, i.e., the fact that the form of intentionality at issue is the perceptual experience it is. ↩

  43. This remains true even when S sees itself* in a mirror. For there is more to this intentional experience than mere perception: S sees a person in a mirror before it* whom it* takes to be itself*. S’s taking the person seen to be itself* is so to speak added on, as is shown by the fact that S could in fact be seeing an exact look-alike in the mirror, i.e., someone who is not in fact S. ↩

  44. Which of course it is not since by resorting to it one cannot capture non-accidental reference. In fact, the claim that self-consciousness construed as ‘inner’ perception is either circular or regressive is only a convoluted way of putting this point: either one question-beggingly presupposes non-accidental reference or chases endlessly after it through recursive application of the subject-object model to itself. ↩

  45. It would be wrong to argue that short of taking this element to be a self-concept or some such one cannot talk of elements in content at all. By an element in content is not meant a constituent in any literal sense, but rather a moment in, or aspect of, content. In general, what belongs in content is a function of when content is or is not true. One may now say that a singular term marks an element in content just in case the sentence on the right hand side of a biconditional specifying truth conditions ineliminably contains the term as a genuinely referring one. This is as true of the asterisked third-person pronoun as it is of any other singular term. ↩

  46. See Castañeda 1998, p. 232. It is the whole point of the thesis of intentionality inherently de se to deny that there is any general function or rule attaching to the pronoun which would distinguish its use in sentences of the form “S believes that it* is F” from its use in such sentences as “Jack and Jill went up the hill and then they came tumbling down.” The asterisk attaching to the pronoun merely disambiguates its specific semantic character, this by making its distinctive grammatical character explicit. That is, it tells us that what the pronoun is anaphorically linked to is the subject-term of an intensional context in which it occurs.

    One could, of course, regard the asterisk as combining with the third-person pronoun to constitute a distinctive grammatical, hence syntactical category for which there is a rule determining correct and incorrect use. This rule would be, however, merely a grammatical or syntactical one. Semantically, the asterisked third-person pronoun is a singular term alongside proper names and referentially used definite descriptions, on the one hand; and indexical expressions on the other. If there is no problem regarding as singular terms expressions which shift their reference from one non-linguistic context to another, there should in principle be no problem in similarly regarding expressions which shift their reference from one linguistic context to another. (If there were such a rule, it would, of course, no more entail that the asterisked third-person pronoun meant ‘the subject of this very form of intentionality’ than does the corresponding rule for the first-person pronoun entail that the first-person pronoun means ‘the speaker of this very utterance’. Such token-reflexive singular terms are parasitically dependent on the asterisked third-person pronoun in the case of other-attribution; and on the first-person pronoun in the case of self-attribution.) ↩

  47. See Musholt 2015, p. 137 and p. 142. ↩

  48. At this point, Toulmin’s non-tauroid cows spring to mind. ↩

  49. See Musholt 2015, § 1.3, pp. 5-7. She calls this primordial awareness self-related information. ↩

  50. This is, I believe, what the later Heidegger means and is objecting to when he writes, „Where does [the focus on] consciousness begin in philosophy? With Descartes. Every consciousness of something is at the same time self-consciousness, whereby the self that is conscious of an object is not necessarily conscious of itself.” (Heidegger 1987, p. 284; my translation) The later Heidegger thus clarifies his position, thereby rejecting his earlier, implicitly Sartrean one. Note that this clarification is not only consistent with, it is in fact the upshot of, everything said in his early work Sein und Zeit. ↩

  51. See, e.g., Rosenthal 2012. ↩

  52. Whose first representative is Brentano—see Brentano 1874/1973. The SOT is currently represented by Uriah Kriegel—see, e.g., Kriegel 2003. ↩

  53. See Zahavi 2004, 2007 and 2009. ↩

  54. Kant 1781-87/1976, Anmerkung zu B 134, my translation. ↩