Grazer Philosophische Studien 57 (1999)


Two claims have been prominent in recent discussions of self-consciousness. One is that first-person reference or first-person thinking is irreducible (the Irreducibility Thesis), and the other is that an awareness of self accompanies all conscious states, at least those through which one refers to something. The latter--here termed the Ubiquity Thesis--has long been associated with philosophers like Fichte, Brentano, and Sartre, though each articulated his own version of the claim. More recently, variants have been defended by Dieter Henrich (1970) and Manfred Frank (1991, 1995a, 1995b). In Frank's words:

. . . every mediated reference to something different from consciousness is mediated by immediate self-awareness. This mediation may be called "transcendental" in a weak sense of the term, according to which what is transcendental does not pertain to the objects of which we speak or think but to the preconditions of such speaking or thinking--presuppositions that fundamentally include a subject immediately certain of its self. (Frank 1995b, 49-50)

Like Henrich, Frank finds immediate self-awareness (mental familiarity, Vertrautheit) to be "non-conceptual" in that it requires no mediation of concepts and no identification or classification of an object of thought in terms of a distinguishing set of characteristics (Frank 1995b, 34-35). It is a direct acquaintance with one's own mental acts or subjectivity which, properly speaking, is not an act of reference at all but a "pre-reflective" and "pre-linguistic" state of consciousness (compare Sartre 1957, 41). To establish its ubiquity, Frank appeals to a Dependency Thesis: immediate self-awareness is an original type of consciousness that is presupposed by reference to any sort of object, including indexical references by means of "this," "here," "now," and object uses of "I" (Frank 1995b, 49). He argues for this claim, in turn, by recourse to the characteristics of indexical thinking.

A question arises. That we can be, and often are, immediately aware of our own mental states is clear enough, but why suppose that such awareness is ubiquitous? Are we not sometimes aware of things without being aware of ourselves or our own awareness? If I am engrossed in watching leaves being tossed about in the autumn wind, concentrating upon the opening bars of the Hammerklavier Sonata, or figuring my annual business deductions, am I, by that very fact, aware of myself, my own experiencing, my subjectivity? These experiences are obviously mine and consciousness is constitutive of them, but must they be accompanied by an underlying reflexive state?

Frank 1995b develops an intriguing attempt to answer these questions in the affirmative. But while it wisely bases Ubiquity upon the Dependency Thesis, and motivates the latter by recourse to the characteristics of indexical thinking, it makes a wrong turn in identifying immediate self-awareness as a type of first-person consciousness and articulating it in terms of the Self-Ascription theories advanced by Roderick Chisholm and David Lewis. Here, I intend to avoid the pitfalls of Frank's approach yet exploit its insights in defending a form of the Ubiquity Thesis that draws upon Hector-Neri Castañeda's work concerning indexical reference.

Frank's Argument for the Ubiquity Thesis

Central to Frank's project is a claim that immediate self-awareness does not involve "reference," nor is its content an "object" of an intentional state or a component of a proposition. Accordingly, immediate self-awareness is non-propositional attitude, viz., it is not a state of consciousness whose content is a proposition or proposition-like structure (Frank 1995a, 188; 1995b, 46). If its content were "objectified," then, by the Dependency Thesis, another state of awareness whose content is the first state would be implied. An infinite regress would be inevitable, and most troubling.

To avoid the regress, secure the Dependency Thesis, and derive the non-propositional, non-referential character of immediate self-awareness, Frank provides a three step argument. Step One rehearses familiar grounds for the Irreducibility Thesis by drawing a contrast between the de dicto, de re, and de se modes of ascription, respectively,

(1) The tallest man believes that the tallest man is wise.

(2) There is an x such that x is identical with the tallest man and x is believed by x to be wise.

(3) The tallest man believes that he himself is wise.

Repeating Chisholm's observations about the lack of any equivalence between these three claims (Chisholm 1981, 19), Frank concludes that the de se report (3) cannot be reduced to the de dicto (1) or the de re (2). Assuming that (3) is the proper format for ascribing a first-person attitude, we have reason to accept the Irreducibility Thesis.

Step Two reasons from the failure of the de dicto format to exhaust the de se ascription as evidence for rejecting the claim that first-person thinking is propositional. Frank agrees with Chisholm that because of the "special logical status of the subject of epistemic self-ascription" there cannot be any first-person propositions. This conclusion is further supported by David Lewis's contention that whenever a proposition P is used to account for a belief content it will equally suffice to ascribe the corresponding property of inhabiting a world in which P is true. Lewis argues that this view is superior to the rival propositional account because there are not enough propositions to distinguish among the attitudes (Lewis 1983, 138-139). Consider the amnesiac, Rudolf Lingens, lost in the Stanford Library (Perry 1993, 21-22). Lingens gathers a great deal of knowledge about the details of the library, and even comes to know that the man, Rudolf Lingens, is at a certain place in the main library, say, in aisle five, floor six at a certain time. But none of this lets him know that he himself is in that locale at that time unless he knows that he himself is Rudolf Lingens. Says Frank, following Lewis:

What he is missing is the ability to self-ascribe one or all of the propositions he holds to be true. This ability (the property of self-ascription) simply does not correspond to any propositional knowledge. It follows, therefore, that some knowledge is nonpropositional." (Frank 1995b, 45)

Again, suppose there are two gods who live in the same world but have different attributes; one lives on the tallest mountain and throws down manna; the other lives on the coldest mountain and throws down thunderbolts. Both are omniscient in the sense of knowing every proposition that is true at their world, in particular, each knows that the god that lives on the tallest mountain throws down manna and that the god who lives on the coldest mountain throws down thunderbolts. But neither knows which of the two gods he is, that is, "neither one knows whether he lives on the tallest mountain or on the coldest mountain; nor whether he throws manna or thunderbolts." (Lewis 1980, 139) Frank elaborates on this case;

No enlargement of their propositional knowledge would overcome this strange ignorance; for what they lack is the ability to self-ascribe the property of knowing a specific world. Such knowledge is therefore, not propositional. (1995b, 45, and see 1991, 317-318)

Step Three argues that the self-ascription argued for in Step Two is involved in every act of reference to the world. In one sense, this is a trivial consequence of a self-ascription analysis of all attitudes, and Frank seems receptive to this position (1995b, 44). But he offers an independent consideration; since the subject-use of "I"--as distinct from its object-use in "I am standing by the table"--underlies all demonstrative usage, and since demonstratives like "this," "here," "now" are indispensible to our being-in-the-world, then the subject-use of "I" is presupposed in all our encounters with the world:

It is I myself who provides the "this" with its reference and its meaning. (That is especially true for "here" and "now," indexical expressions that I could not even learn unless I had previously learned to directly self-ascribe--in subject-use--mental states to myself.) For how could "here," "now," or "this" be introduced other than through phrases that already employed the subject-use of "I" ("here" means "the place where I stand," "now" means "the time when I grasped this thought," "this" refers to "the finger with which I press the keys of my computer," and so on)?40 Therefore, self-consciousness--perhaps not in the form of qualia-consciousness, but clearly in the form of attitudes de se--proves to be fundamental to our being-in-the-world. (1995b, 49)(1)

With this, Frank concludes his case for the Dependency Thesis: all cases of reference are dependent upon first-person awareness, hence, upon self-awareness. The Ubiquity Thesis is a direct consequence.

To summarize; Step One is the familiar observation that first-person thinking is irreducible; Step Two argues that such self-awareness is unlike the objective reference typified in the thinking of proposition-like content; and Step Three contends that such self-awareness is presupposed by all other conscious states. In what follows, I will argue that the reasoning within Steps Two and Three are flawed. The grounds offered by Chisholm and Lewis for the Self-Ascription theory are not persuasive, and while the phenomenon of indexicality does support a form of the Dependency Thesis, it is not the subject-uses of "I" which prove fundamental. Nevertheless, certain insights in Frank's argumentation can be salvaged; immediate self-awareness is not propositional and it is presupposed by other referential states. To find an alternative route to these conclusions, let us take a closer look at reference, and at indexical reference in particular.

Thinking Reference

It is crucial to understand that reference is here treated as a psychological state. Only in a derivative sense do terms refer, by conveying or storing what is psychologically or thinkingly referred to. Further, a thinker's direct confrontation with an item in the manifold of consciousness must be distinguished from both what a thinker intends to communicate with a given token (communicational reference) and the denotation associated with the linguistic type in a shared language (semantic reference). These latter forms of reference are dependent upon direct encounters in experience or thought (Castañeda 1989a, chp. 1).

We refer by consciously "picking out" an item in experience and thinking something of it. Reference embodies a relatively high-grade awareness wherein one focuses upon one item to the exclusion of others, distinguishing it from the background material impinging upon us at any moment. Picking out an item is to identify it as being F--for some property or mode F--and not as being G for some contrastive property G. The act thereby embodies a sense of negativity--of what the selected item is not(2)--and presupposes the presence of an experiential manifold, much of which might be unstructured and inarticulate so far as referential content goes. Through reference, structure is discerned, or perhaps, imposed.

From this description it follows that the items picked out--the thinking referents, for example, this paper, Nathaniel's sneezing in the back row, that specific shade of green--can be grasped and thought about. What they are, ontologically speaking, is beyond present scope, but any account must allow that they be accessible to consciousness, This may imply that they are finite in internal complexity and identifiable through finite sets of properties, thus, not massive, infinitely propertied objects like the whole of Mont Blanc "with all its snowfields."(3) Also, judging from the contrastive property of consciousness, thinking referents are never grasped in isolation, but always as involved with other items. Specifically, they are picked out with the purpose of thinking something about them, thus, as components within propositions or other proposition-like contents (interrogatives, intentions, recommendations, etc.). Further, thinking reference requires the mediation of referential modes by whose means the referent is picked out.(4) There can be no reference to an object per se devoid of all qualifications; thinking of the Louvre, for example, we may consider it qua the largest art gallery in France, or qua that huge building over there across the Seine, or, through the non-identifying modes of being a famous museum in Paris or, simply, as that. However we identify a referent, we must acknowledge the following principle:

Principle 1. A person X's thinking reference to an item O is accomplished only via referential modes each of which is a way through which X cognizes (thinks of, conceives, perceives) O and at least one of which is an identifying mode whereby X distinguishes O from everything else.

Let us say that a token designates its referent and expresses relevant referential modes. It need not be assumed that a referent is identified through only one mode in a given act, nor that X's knowledge of an identifying mode M is sufficient for X's thinking reference to any item qua M. While some modes are attributes of external entities, others are aptly described as modes of presentation, that is, manners whereby objects are apprehended. Being this blue pen in my hand, for example, may be a mode through which I refer to a particular pen, but "this," "in my hand," and perhaps "blue" apply to the item only in virtue of its causal relation to myself, an experiencing subject. The entire mode is the means whereby distinct aspects congeal to constitute an "object" for me.

This description has led some to treat modes as ways of cognizing objects, as attributes of the thinking agent or the cognitive act than elements in what is grasped. But the effect of so isolating modes from propositional content is to misconstrue their vital role in both communication and inference. Describing Andrea as "that wealthy singer we met yesterday," for example, I mean to communicate information about her, the person I am talking about, not only about my beliefs or about how I think of Andrea. Similarly, one who accepts the proposition that Cicero was a clever philosopher, and who thinks of the referent "Cicero" qua the most famous Roman Senator, can legitimately infer that some Roman Senator was a clever philosopher, or, from 7 > 3 that some prime number is greater than 3. But this is so only if "Cicero" and "7" contribute implicationally-relevant propositional content. Again, there must be a reason why a conditional like if Cicero had invented the steam engine then some Roman would have been a famous inventor is true whereas if Archimedes had invented the steam engine then some Roman would have been a famous inventor is not. The different modes expressed by the proper names are the telling factors.

The foregoing illustrates a further property of thinking reference, namely, that the referential modes through which a person X refers to O are part of the content of the attitude within which X's reference to O occurs. The modes need not themselves be subjects or predicates within that content; in an uttering "Paris is a large city," a speaker might identify Paris as the capital of France, but it does not follow that the property of being capital of France is a separable component about which one thinks in thinking the proposition that Paris is a large city. The expressed mode is the unconceptualized manner--not explicitly referred to nor predicated--through which one conceptualizes the referent of 'Paris'. Still, it has a presence within conscious content:

Principle 2. In referring to an item via a referential mode M, one is ipso facto aware of M without necessarily referring to M.

In short, every act of thinking reference is accompanied by non-referential awareness of the involved referential modes.(5)

Indexical Reference

According to Castañeda, indexical tokens express an experiential mode of reference whereby we distinguish and locate specific items within the mass of material impinging upon us at any moment. Indexical reference is our most basic means of differentiation, expressed through tokens which reveal our own part in the process by denoting items that are "referred to as items present in experience" (Castañeda 1981, 285-6). If I hear you describe a painting with the sentence "this is beautiful," for instance, I realize not only that beauty is said of what you call "this" but that the latter is an object of your direct encounter. As such, indexical reference is perspectival in reflecting a subject's orientation towards an item or, conversely, a referent's position relative to the place and time of an experience.(6) My indexical tokens express what they do not only because they are anchored in my particular experiences and issue from a unique spatio-temporal vantage point that I happen to occupy, but also because their referents are differently situated relative to that locale. From your perspective, a here might be my there, your she my you, and within my own, a this differs from a that and one this diverges from another. It is perspective that is essential to any experience that provides one's grasp of the whereabouts of indexically-specified items without which particular this's, that's, here's, and beyond's would be denuded of their considerable informative powess.

There are several consequences of this description of indexical reference as rooted in perspectival encounters. First, indexical reference is ephemeral insofar as perspective is constantly changing, rendering the status an entity has relative to a given experiential perspective temporary:

A this quickly turns into a that, and soon enough it is lost to experience and is not even a remote that; a you goes away and is replaced by another . . . Nothing is really an enduring you -- except God perhaps for the abiding mystic. (Castañeda 1989a, 69)

Second, being indexically positioned is an extrinsic feature of an entity; no object in the external world is intrinsically a you, a this or a here; indexical status is possessed only contingently, and always relative to a confronting subject. Third, indexical reference is irreducible since non-indexical mechanisms of reference fail to express the essential involvement with immediate experience that indexicals convey. Nor can the various indexicals be reduced to each other.(7) Fourth, indexical reference is subjective. One person's "I," "this," or "over there" expresses, in part, what is unique to his or her perspective, making it impossible for another to gain a cognitive fix on the very same item in precisely the same way (Castañeda 1990b, 127). No one can entertain the exact indexical contents of anyone else, and even one's own indexical contents must differ over time.(8) Finally, the truth-conditions for indexical sentences must reflect the role of their determining experiences. It will not do to say that the tallest man's statement,

(4) I am wise.

is true if and only if the tallest man is wise. The tallest man might be wise even when he makes no self-reference, even if, by chance, he utters a token of the type 'I am wise'. What must be captured, in addition, is that the tallest man, qua aware of himself in a first-person way, is wise, and it precisely this that we intend to convey with the de se ascription (3) in which the emphatic reflexive 'he himself' is a quasi-indicator, that is, a device for ascribing indexical reference.

Being an indexical expression is always a property of a token. No syntactic type in English, qua type, is indexical for the simple reason that there are both common noun and quasi-indexical uses of the familiar types "I," "you," "this," "now," etc. Similarly, no type is essentially quasi-indexical; even the most apparent quasi-indicators in English, viz., "he himself" and "she herself," can be used in other ways. Indexicals convey not only position relative to access points (here, yonder, now, before, tomorrow), but also, classifications in terms of gender (he, she), personhood or its lack (I, you, it), number (he, we, they), order (the former, the latter), and possession (mine, yours, hers). Each invariably expresses an indexical referential mode, portraying items as involved in experience in some way, either as immediately confronted, e.g., pure demonstratives such as "this" or "there", or as vicariously considered through relation to present experience, say, through the indexical description "the man who sat on that stool yesterday." The temporal demonstratives "now" and "then" not only designate intervals of time, but intervals as related to a present experience, the former to the actual time of that experience and the latter to some preceding interval. Similarly, "this" and "that," "these" and "those," represent items present to consciousness. Indexicals are particularly rich in information value; none is a mere label.

The first-person indexical in (4) expresses a publicly accessible mode that allows others to know how the tallest man thinks of himself. The inclusion of such modes within propositional content accommodates some intuitive implications. For instance, if it is true that today is March 27 then it follows that tomorrow is March 28, but the latter is not implied by the proposition that George's birthday is March 27 even though George's birthday is today. Hence, the indexical term "tomorrow" expresses content that is relevant to logical powers of the propositions expressed by "today is March 27." These are reasons for recognizing publicly accessible indexical modes as part of propositional content.

Yet recourse to publicly accessible modes is not the whole story, since they can never succeed in performing the identifying function of indexicals. With typical uses of the pronouns "I" and "you" we succeed in picking out particular referents, but this is not secured by means of the generic modes associated with the types. The perspectivity ensures that indexical tokens express more than generic modes. Others can imagine that Andrea is thinking of herself as an I--as someone who is aware of herself qua self--without thinking her first-person propositions and, thus, without grasping her particular identifying mode. With her use of "I" she is locating herself within the field of her own consciousness, picking herself out from her present perspective. Generic indexical modes are not identifiers. Again, my demonstrative in "this is beautiful" connotes my particular perspective on an item, say, the Hope diamond pictured in a magazine. But, at the same time, I might use "this" to refer to that very diamond which now appears as a dirty stone before me, subsequently learning, to my own surprise, what I could express by "this is this" (Burge 1977, 355). In short, since indexical modes serve as identifiers, they must be particularized (Castañeda 1989a, 76; 1990c, 303; Searle 1984, 222). Only the tallest man can think of himself in precisely the first-person way, and consequently (3) does not fully report what the tallest man expresses with (4). Only he can grasp the full significance of his "I," just as only I can grasp the full significance of my "this.". A similar perspectivity characterizes what is expressed by every indexical.

Consequently, an "I" token not only designates its producer but, like a "there" or "here" token, expresses the fact that the referent is encountered from a particular perspective. The latter is integral to the particularized mode, and this yields further principle:

Principle 3. Among the particularized indexical modes involved in any act of indexical reference are identifying modes which embody (contain, reflect) the thinker's perspective within that act.

Perspective is only one element of what is expressed by an indexical. The referent of an "I" token is also identified qua encounterer or experiencer or thinking thing,(9) while the referent of a "you" token is identified qua addressee. Again, one does not thereby predicate any of these properties in so doing.

Not all forms of self-reference are first-person indexical. I can refer to something which is the same as myself, e.g., when I refer to the person I see reflected in the mirror, not realizing, perhaps, that this person is myself. Frank cites an experience of Ernst Mach:

Once when he was boarding a bus in Vienna, he saw a man enter at the same time on the other side and was suddenly struck by the thought, "Look at that shabby pedagogue coming on board!"--not realizing that he was referring to himself, becuase he had not noticed that opposite him hung a large mirror. (1995b, 42)

Mach's reference was reflexive, but was not reference to himself qua self. In Castañeda's terminology, non-first-personal self-references exhibit an external reflexivity since they pick out oneself from the outside. The latter is distinct from the internal reflexivity expressed through the first-person indexicals where one is conscious of oneself qua self.(10) In short, first-person reference is only one species of self-awareness.(11)

Frank's Argument Rejected

If this account of indexical reference is correct then there is good reason to be skeptical of Frank's argument above, specifically, of the reasoning within Steps Two and Three. Let me explain in detail.

Self-Ascription is Both Referential and Propositional.

There are good reasons to reject the Chisholm-Lewis approach to first-person attitudes. Chisholm, while rejecting propositional accounts, nonetheless speaks of the attributing subject as the object of direct attribution and of the property as the content. Now I ascribe properties to a good many things in the course of my daily experience, and in so doing I always pick out the items to which I ascribe those properties. To think predicatively is to synthesize distinct factors, each of which enters into cognition, into a single unit. The unit is a posited relationship or proposition. How is it any different when I ascribe something to myself, e.g, that I am currently thinking or that I have blue eyes? Mustn't I still focus on myself and say something about me? Must I not identify--refer to--a logical subject in order to ascribe anything to it? I cannot think otherwise; ascribing a property to oneself, even when we consider ourselves as a thinking subject, is a cognitive synthesis of diverse factors which culminates in the acceptance of a proposition-like content.

Consider how commonly we attribute things to ourselves and affirm our connections to other entities:

(5) I am taller than Bill.

(6) I am currently in Illinois.

(7) I love Myrna immensely, just as much as Richard does. But some time she will make the two of us unhappy.

(8) Harold gave the book back to me.

(9) I believe that Sherry will be thinking about me this evening.

In each of these cases I single out items and affirm something of them, and the first-person pronouns I employ in so doing have all the earmarks of referential expressions. Moreover, in (7) and (9) I am talking about a thinking, feeling subject. There is every indication of a propositional character to what is expressed since I can wonder about their truth values and speculate about their implications. That I am in Illinois, for example, implies that some thinking being is in Illinois, or even that some self-referring being is in Illinois, but the noisiest organism presently in this room is in Illinois need not have this implication even though I may be the noisiest organism presently in this room. This suggests not only that "I am currently in Illinois" expresses a proposition but that "I" introduces implicationally-relevant content. The counting use of "I" in (7) reinforces the conception that it designates a discrete item on a par with the referent of "Richard," and the occurrences of "me" in (8) and (9) serve to identify an object of someone else's endeavor. In each case, self-ascription is referential and propositional.

Sanctioning First-Person Propositions.

Some instances of first-person reference have little to do with ascribing properties to oneself. For example, I can think what I express with sentences (5)-(9) without thereby affirming what I think of. Again, practical thinking is replete with self-reference. If I report my consideration of alternatives with,

(10) I am debating whether to read Hegel's Phenomenology or Spinoza's Ethics next summer.

then I am certainly ascribing to myself a property of deliberating about what to do. But as we probe into the structure of what I am thinking we realize that to deliberate between the options of reading Hegel's Phenomenology or reading Spinoza's Ethics, I must consider my performing these actions. I thereby consider myself as engaged in a certain activities but without attributing those actions to myself in either a believing or intending manner. This phenomenon is common to my anticipation of possible consequences of actions I might take, For instance, if I affirm,

(11) Were I to read Hegel's Phenomenology next summer then I would not be able to finish another paper on pronouns before September.

I am not merely noting an exclusion relation between one action type and another. I am conjecturing about what would happen if I were to read Hegel's Phenomenology, not some necessary or lawlike connection between "reading Hegel's Phenomenology next summer" and "not finishing another paper on pronouns by September." Others might not be similarly incapacitated, for example, I can affirm of my prolific friend George,

(12) It is not the case that if George were to read Hegel's Phenomenology next summer then he would not be able to finish another paper on pronouns by September.

This suggests that my token of "I" in (11) expresses (to me at least) counterfactually or implicationally relevant material, viz., properties that I recognize myself as having, e.g., the fact that reading Hegel fatigues me, or that I have many additional commitments next summer. These are not properties I associate with "George." Perhaps by uttering (11) I am committed to believing that I possess a certain conditional property, but it is does not follow that the occurrences of "I" in (11) serve only to mark a subject to which something is attributed. I can affirm (11) without attributing to myself the property of reading Hegel's Phenomenology next summer, but in order to affirm the conditional I must think a content to which the first token of "I" in (11) makes a crucial contribution.

Chisholm and Lewis dismiss the possibility of accounting for de se ascriptions in terms of propositions too quickly. Chisholm argues that the best account of knowing first-person propositions would require the agent to grasp his or her own haecceity (individual essence), and since it is unlikely anyone can do this then recourse to first-person propositions won't do (Chisholm 1981, chp.3). But his conclusion is premature; the account of indexical and first-person reference given above provides a way of sanctioning the existence of first-person contents without requiring an agent to grasp exotic haecceities. Chisholm simply does not consider all the alternatives and is oblivious to the possibility of a more fine-grained account of propositions, some of which contain subjective modes taking them out of the public realm.

The propositionalist reply to Lewis's arguments is straightforward. Take the Rudolf Lingens case. Frank appears to reason as follows: Since (i) Lingens knows all the (relevant) propositions about Rudolf Lingens, but (ii) Lingens does not know that he himself is Rudolf Lingens, then it follows that (iii) not all knowledge is propositional (Frank 1995b, 45). On the account presented above, however, if Rudolf doesn't initially know where he himself is, then when he comes to adopt the belief that he himself is in aisle five, floor six--perhaps by realizing that he is Rudolf Lingens--he acquires knowledge of a first person proposition, namely, what he might report with "I am in aisle five, floor six of the Stanford Library." This is a distinct proposition from Rudolf Lingens is in aisle five, floor six of the Stanford Library. Why so? Because of the difference in the modes used to pick out the respective thinking referents, that associated with Lingens' "I" on the one hand, and with his "Rudolf Lingens" on the other. Thus, if the first-person proposition which Rudolf Lingens fails to know is a proposition about Rudolf Lingens, then Frank's premise (i) is false and the argument is unsound. If it is not a proposition about Lingens, then the argument fails to be valid--except on the assumption that "I am in aisle five, floor six of the Stanford Library" or "I am Rudolf Lingens" do not express propositions. But this assumption is precisely what I have challenged here.

The solution to gods' dilemma is similar; each god would know which god he himself is were he to come to know a relevant first-person proposition, e.g., that which could be reported by 'I am the god that lives on the coldest mountain and throws down thunderbolts.' If they did not know these propositions then Lewis is either mistaken in describing them as omniscient or begging the question against the propositional account.(12)

The De Se Locution Ascribes an Attitude Towards a Propositional Type.

The subjectivity of indexical reference implies that one who ascribes indexical beliefs to Lingens or to the gods is not in a position to think those very first-person propositions. The latter are subjective. Not sharing the ascribee's perspective, the ascriber has no access to the ascribee's first-person mode and cannot attribute reference via that mode. Instead, the ascriber appeals to the generic first-person mode that we all grasp in understanding the meaning of "I," attributing only the type of proposition with the quasi-indexical locution in (3). Seen in this way, the de se (3) is similar to the de re (2) which also shows only the type of proposition attributed, a type conveyed by the matrix "x is wise." The difference is that the quasi-indicator in (3) is a restricted variable whose semantic contribution is determined partly by its antecedent and partly by the attributed generic first-person mode.(13) The failure of the that-clause of (3) to specify a proposition the tallest man is said to believe does not imply that the tallest man fails to affirm a proposition upon asserting "I am wise." The de se locution is a species of the de re format--and the latter is appropriate whenever ascribers do not intend to, or are unable to, specify the exact propositional content of the ascribee.(14) Along with de te, de nunc, de hic, de hoc attributions, the de se locution presents an irreducible means for ascribing to a thinker an attitude towards a proposition he, the thinker, would express indexically even though the ascriber cannot fully apprehend it. I conclude, that the Self-Ascription theory gives no advantage over the propositional account of first-person attitudes, and that, therefore, Frank has not established that there is a distinct mode of self-consciousness that is non-propositional and non-referential.

First-Person Thinking is not Ubiquitous.

A more direct consideration against the Self-Ascription theory affects both lines of argument that characterized Frank's Step Three. The problem repeats a concern noted at the outset of this essay; why suppose that self-awareness, in the sense of first-person consciousness, is ubiquitous? Frank's premise that demonstrative usage is endemic to our being-in-the-world is plausible enough, but why conclude that this must involve first-person consciousness, thus, that the latter has the epistemological "priority" that he claims?

I have already mentioned that Frank cited Castañeda's 1967 essay, "Indicators and Quasi-Indicators," in support of this priority claim. But Castañeda abandoned this thesis in the 1980s when he became convinced that a fully external consciousness, an "I-less experience," is a more basic level of awareness upon which fully articulated self-consciousness is mounted (see, for example, Castañeda 1987, 440 and 1990b, 136-137). He cited cases of blindsight, visual concentration, and animal consciousness as examples of pure I-less consciousness (see Castañeda 1990b, 132-138)--perhaps Ernst Mach's "Look at that shabby pedagogue coming on board!" provides another. Castañeda also abandoned the attempt to reduce any pure indicator to any other, for example, to explain all indexicals in terms of "I" (1989a, 70-76). To illustrate, no token of "here" can mean exactly "the place where I stand," as Frank suggested, since "here is here" and "here is the place where I stand" differ in modal status. Similarly, "now is now" and "now is the time when I grasped this thought" differ in information value, in which case the meaning of "now" cannot be given by "the time when I grasped this thought."


In sum, I have advanced three consideration against an embrace of the Self-Ascription theory. First, ascription, whether to oneself or an external entity, appears to be a judgmental act which involves both reference and propositional content. Second, there are accounts of first-person propositions that can be sustained against the arguments of Chisholm and Lewis. Third, it is doubtful that self-ascription of the first-person sort accompanies all referential acts. While these considerations do not decisively refute the Self-Ascription theory, they do raise doubts about Frank's understanding of indexical reference in terms of that approach. Consequently, it has not been established that first-person thinking underlies all consciousness, and Frank's attempt to derive the Dependency and Ubiquity theses from the phenomenon of indexicality remains unconvincing.

An Alternative Defense of the Ubiquity Thesis

Despite this criticism, Frank is entirely correct to seek the grounds for the Ubiquity Thesis in indexical thinking, and I will now argue that both Steps Two and Three in Frank's argument contain insights that can be exploited in an alternative argument for the Ubiquity Thesis.

It is the essential perspectivity of consciousness that underlies the irreducibility of indexical thinking. Indexicality is how we locate ourselves within the world, how we pick out and constitute various items within our experience and prepare ourselves for action (Castañeda 1990d, 81-85). But how extensive is it? It certainly permeates perceptual consciousness; visual, tactile, auditory, and olfactory experience is always of content from a perspective. Similarly, perspective is ineliminable from all sensory imagination. Indexical elements creep into seemingly non-perceptual awareness as well; when I think that the tallest man in Bejing speaks Mandarin I make reference to the actual world not only with the description but also with the proper names "Beijing" and "Mandarin," and in each case indexical elements lurk within the relevant modes of presentation. For example, if I identify Beijing in terms of its locale I do so vis-a-vis my present position. I thereby utilize both spatial and temporal indexicals. Again, if my thought of Beijing is qua the capital of China, then indexicality emerges in thinking the referent of "China."(15) Our singular descriptions are replete with indexical elements, and even our use of proper names invokes modes that reflect how their referents are or have been situated within our perspective.

Is all consciousness similarly perspectival? The above reflections suggest that all reference to the actual world of concrete particulars is inescapably so. Whether the abstract thoughts of pure mathematics and similar disciplines are as well is more difficult to determine. Certainly sensory images are common to the symbolism used to facilitate our calculations (though images do not exhaust the relevant mathematical content), and we commonly pick out abstracta indexically, viz., this step function, the sum of that pair, the previously mentioned color. Some might balk at the idea that pure abstracta, untainted by involvement with concrete individuals, can be referred to. Whether their skepticism is justified, I cannot say, but it does seem that all reference to the actual world of spatio-temporal events and objects is inevitably perspectival and, therefore, indexical. Since perspective is embedded within the modes of presentation by which a thinker refers indexically, modes the thinker must be aware of (by Principle 2 above), then awareness of perspective must anchor acts of reference even though neither the awareness nor the perspective are themselves referred to. To this extent, there is an pre-reflective awareness of one's own subjectivity, since subjectivity is nothing more than consciousness from a perspective. But there is no reason to describe such awareness as a type of self-ascription or first-person consciousness. The argument for this can be set forth as follows:

1. Each act of reference to the actual world embodies indexical reference.

2. By Principle 1 and Principle 3, all indexical reference includes a mode of presentation that involves the subject's encountering something from a perspective (a perspectival encounter).

3. If an act of reference includes a mode of presentation that embodies the user's perspectival encounter, then, by Principle 2, the user must be aware of that perspectival encounter, even if the user has not conceptualized or referred to that perspective.

4. Since subjectivity is consciousness from a perspective, then to be aware of one's perspectival encounter is to be aware of one's subjectivity.


5. All referential consciousness presupposes awareness of one's subjectivity.

I submit that this awareness of subjectivity is a viable interpretation of immediate self-awareness. The conclusion of this argument is a form of the Dependency Thesis, and an analogous form of the Ubiquity Thesis is an immediate consequence.

Curiously, this conclusion allows us to embrace Frank's insight that immediate self-awareness is neither a propositional nor a referential state.(16) Not all awareness is characterized by the level of articulation needed for reference; one might not have contrasted one's perspective from any other perspective, one might not have picked it out, nor picked out any "subject" which occupies that perspective. Immediate self-awareness is a pervasive consciousness of one's own perspectival encounter, a phenomenon akin to the presence of background noise within one's experience--literally part of what we are taking in--without being the object of our referential focus or an element in our judgments.(17)

The fundamental refexive encounter that accompanies my conscious states is not first-person-awareness; it is awareness of the conscious activity alone, without any identification as being me, mine, or anything else. It may be nothing more than the presence of the manifold of consciousness from a point of view--and though more than the mere existence of the manifold, it is not the presence of an "I." By recognizing it, the Ubiquity Thesis gains a fresh impetus.(18)


Ameriks, K. 1995. "From Kant to Frank: The Ineliminable Subject," in Ameriks and

Sturma, eds., 217-230.

Ameriks, K. and D. Sturma, eds., The Modern Subject: Conceptions of the Self in Classical

German Philosophy (SUNY, 1995).

Bezuidenhout, Anne, 1996. "Pragmatics and Singular Reference," Mind & Language

Boer, S., and Lycan, W., 1980. "Who Me?" The Philosophical Review 89, 427-466.

Brentano, Franz, 1874. Psychology From An Empirical Standpoint. English edition edited by Linda McAlister (Routledge, 1973).

Burge, T. 1977. "Belief De Re." The Journal of Philosophy 74, 338-362.

Castañeda, H-N., 1977. "Perception, Belief, and the Structure of Physical Objects

and Consciousness," Synthese 35, 285-351.

Castañeda, H-N., 1981. "The Semiotic Profile of Indexical (Experiential) Reference,"

Synthese 49, 275-316. Reprinted as "Indicators: The Semiotics of Experience," in Jacobi and Pape, eds., Thinking and the Structure of the World (de Gruyter), 57-93.

Castañeda, H-N., 1983. "Reply to John Perry: Meaning, Belief, and Reference," in

J. Tomberlin, ed., Agent, Language, and the Structure of the World (Hackett),


Castañeda, H-N., 1987. "Self-Consciousness, Demonstative Reference, and the Self

Ascription View of Believing," in J. Tomberlin, Philosophical Perspectives 1

(Ridgeview Publishing Company), 405-454.

Castañeda, H-N., 1989a. Thinking, Language & Experience (Minneapolis: University of

Minnesota Press).

Castañeda, H-N., 1989b. "Direct Reference, the Semantics of Thinking, and Guise

Theory" in Almog et al, eds., 105-144.

Castañeda, H-N., 1989d. "Persons, Egos, and I's," Psychopathology and Philosophy, in

M. Spitzer et al, eds. (Berlin: Springer-Verlag), 210-234.

Castañeda, H-N., 1990a. "Indexicality: The Transparent Subjective Mechanism for

Encountering a World," Nous 24, 735-750.

Castañeda, H-N., 1990b. "Self-Consciousness, I-Structures, and Physiology," in Philosophy and Psychopathology, M. Spitzer and B. Maher, eds. (Berlin: Springer-Verlag), 118-145.

Castañeda, H-N., 1990c. "Indexical Thinking and Indexical Guises," in Jacobi and Pape,

eds., 299-307.

Castañeda, H-N., 1990d. Intentionality, Modality, and Supervenience, Rotterdamse Filosofische Studies XII

Dennett, D. 1991. Consciousness Explained. (Little, Brown and Company).

Forbes, G., 1987. "Indexicals and Intensionality: A Fregean Perspective," The

Philosophical Review 96, 3-31.

Forbes, G., 1989. "Indexicals" in Handbook of Philosophical Logic, ed., D. Gabbay, vol IV

(D. Reidel), 463-490.

Frank, M., 1991. Selfstbewußtsein und Selbsterkenntnis (Stuttgart: Philipp Reclam).

Frank, M., 1995a. "Is Subjectivity a Non-Thing, an Absurdity [unding]?" in Ameriks and

Sturma, eds., 199-216.

Frank, Manfred. 1995b. "Mental Familiarity and Epistemic Self-Ascription." Common Knowledge (October): 30-50.

Frege, G., 1967. "The Thought: A Logical Inquiry," in P. Strawson, ed., Philosophical

Logic (Oxford), 17-38.

Henrich, D. 1970. "Selbsbewußtsein: Kritische Einletung in eine Theorie," in F. Bubner, ed., Hermeneutik und Dialetik, I, Tübingen, 257-284.

Jacobi, K. and H. Pape, eds. 1990. Thinking and the Structure of the World (Berlin: De Gruyter).

James, W., 1904/1976. Essays in Radical Empiricism, (Harvard University Press, 1976).

Kapitan, T., 1993b. "Keeping a Happy Face on Exportation," Philosophical Studies.

Kapitan, T., 1994. "Exports and Imports: Anaphora in Attitude Ascriptions," in J.

Tomberlin, ed., Philosophical Perspectives 7/8: Philosophy of Logic and Language

(Ridgeview Press).

Kapitan, T., 1996. Review of Fraçois Recanati's Direct Reference, forthcoming in Philosophy and Phenomenological Research.

Kapitan, T., 199-. "On Depicting Indexical Reference," forthcoming in a memorial volume dedicated to Hector-Neri Castañeda, Francesco Orilia and William Rapaport, editors.

Kaplan, D., 1989a. "Demonstratives, An Essay on the Semantics, Logic, Metaphysics,

and Epistemology of Demonstratives and Other Indexicals," in Almog et al, eds.,


McGinn, C., 1983. The Subjective View (Oxford).

McKay, T., 1988. "De Re and De Se Belief," in D. F. Austin, ed., Philosophical Analysis

(Kluwer, 1988), 207-217.

Millikan, R., 1990. "The Myth of the Essential Indexical," Nous 24, 723-734.

Nagel, T., 1986. The View From Nowhere (Oxford).

Perry, J., 1979. "The Problem of the Essential Indexical," Nous 13, 3-22.

Perry, J., 1983. "Castaneda on He and I," in Tomberlin 1983, 15-42.

Perry, J., 1988. "Cognitive Significance and New Theories of Reference," Nous 22, 1-18.

Perry, J., 1990. "Individuals in Informational and Intentional Content," in E. Villanueva,

ed., Information, Semantics and Epistemology (Blackwell), 172-189.

Perry, J., 1993. The Problem of the Essential Indexical and Other

Essays (Oxford University Press).

Recanati, François, 1993. Direct Reference (Blackwell).

Sartre, Jean-Paul, 1957. The Transcendence of the Ego (The Noonday Press). Originally published as La Transcedance de L'Ego: Esquisse d'une description phénomenologique (Recherches Philosophiqes VI 1936/37).

Searle, J., 1983. Intentionality (Cambridge).

Searle, J., 1992. The Rediscovery of the Mind. (MIT Press).

Tomberlin, J. ed., 1983. Agent, Language and the Structure of the World (Hackett).

Whitehead, A. N., 1929. Process and Reality (Macmillan).


1. 1. The superscript in this passage is for a footnote in which Frank cites Castañeda's 1967 paper, "Indicators and Quasi-Indicators" for support of the epistemological priority of first-person reference. For relevant passages, see the reprint of this paper in Castañeda 1989a, 208-210.

2. 2. If Whitehead was correct in describing consciousness as "the feeling of negation" then a felt negativity is a component of all consciousness states. See A. N. Whitehead 1929, Part II, chp. VII, section II, and Part III, chp. II, section IV.

3. 3. This example comes from a well-known letter to Russell in which Frege expressed bewilderment at Russell's insistence that Mont Blanc itself was a propositional constituent (reprinted in Soames and Richard 1988, p. 41).

4. 4. The literature on referential modes or Fregean Sinne is massive. For a recent discussion see Recanati 1993, particularly its contrast between linguistic modes and psychological modes, the former constituting the meaning of a term qua type and the latter being the user's means for cognizing the referent.

5. 5. The graspability and finitude of thinking referents suggests that they cannot be divorced from the modes used to identify them, that is, referential modes are constitutive of thinking referents. In speaking of a referent X qua M, we take the entire unit, the object qua mode, as the logical subject and do not treat it as though it were separable from M. So doing permits the fine-grained distinctions to solve familiar problems of intentional identity, but it raises further questions about reidentifications and correlations among distinct thinking referents. Again, this is a matter for more detailed theorizing. See Castañeda 1989 for a detailed account and a generalized development in Kapitan 1994. Insofar as we speak of the object or event that presumably underlies and accounts for a unity among various manifestations to our consciousness, say, different presentations of one and the same enduring person, we are speaking of a doxastic referent, namely, what we believe to account for the unities or samenesses among thinking referents (Castañeda 1989b, 125).

6. 6. Castañeda emphasized the perspectivity of indexical reference in his early essays (including "Indicators and Quasi-Indicators in 1967 where perspectives are treated in terms of perspectival properties). Subsequently, he spoke more in terms of demonstrative or indexical properties (1989a, 71). That indexical properties "fix" positions in spatio-temporal fields is emphasized in 1977, 320. William James wrote that even the first-person "I" is a "noun of position" and seemed willing to generalize this to all indexicals (James 1904, 86). See also Whitehead 1929, 43, where demonstrative elements are taken to enter into the expression of every proposition.

7. 7. The irreducibility of indexical reference was urged by Castañeda since his earliest papers on the topic in the 1960s while more recent statements can be found in Castañeda 1989a, 70-76. He criticized familiar reductions of indexicals, say, "I", to "this person now speaking," for imputing too much conceptual apparatus to their users, particularly, small children (1989a, 72-5). Again, there are modal difficulties inasmuch as "I am this person now speaking" expresses a contingency truth while "I am I" is necessary.

8. 8. The subjectivity thesis and its reliance upon particularized indexical modes of presentation was articulated in Castaneda 1981, 1989a, 1989b, and 1990a. It is implied by the treatment of indexicals in Searle 1983, 220-230, on similar grounds. Analogous reasoning appears in Frege 1967, 25-6 with respect to first-person reference; "everyone is presented to himself in a particular and primitive way, in which he is presented to no-one else." See also McGinn 1983, 17 and Nagel 1986, chps. 2-4. Boer and Lycan 1980 provides a contrasting view of indexical reference, as do Perry 1979 and 1983, Kaplan 1989, and Millikan 1990, but see Castañeda's responses to these positions, respectively in 1984, 249-256; Tomberlin 1983, 313-328; Castañeda 1990b; and 1990a.

9. 9. Here we must be careful. Frank interprets the Irreducibility and Ubiquity theses as showing the limits of a physicalist description of the world. Recall Hobbes's complaint against Descartes' use of the cogito: all we have a right to conclude from the cogito is that I am a thinking thing, but not that a thinking thing is soul or a mind conceived as distinct substance from any material thing. Thus, while the item referred to in an act of internal self-awareness is aptly characterized as a thinker or experiencer, or, perhaps as an event of experiencing, we are not entitled to automatically conclude that the "I" refers to a non-physical entity. To the contrary, since our experiences are wedded to perspective then they are radically spatial and temporal, and each may be but a wrinkle in the physical field(s) constituting physical reality, abiding fully by physical laws. Whether this implies the adequacy of a physicalistic description depends on what "physicalistic" means. Viewed from the outside, the "physical" categories of locale, extension, motion, mass, force, acceleration, or other field properties seem to leave something out. But whether this indicates that physical science is essentially incomplete, or only that in its current state of development it leaves something out, is another question. Both the Irreducibility and Ubiquity theses, as here interpreted, are neutral on this matter.

10. 10. Castañeda indicated that the particularized I-property is primitive, and that one does not classify oneself as a self, person, etc. in first-person reference (assuming that to classify is to predicate a sortal property or concept). See Castañeda 1990b, 127.

11. 11. Like all reference, self-reference involves contrast, that is, there can be no self-reference without awareness of other items against which one contrasts the self. But if that is so, then self-reference through the first-person pronoun cannot be presupposed by all other sorts of reference. Castañeda emphasized the hierarchies of negativities involved in self-reference in his 1990b.

12. 12. It should be noted that Lewis explicitly accepts the view that propositions are sets of worlds, and says that he "need not quarrel" with the view that the objects of first-person attitudes are propositions in some other sense (Lewis 1983, 135). It is for this reason that his account is insensitive to the possibility of fine-grained distinction between propositional contents. Frank is not sufficiently appreciative of this aspect of Lewis's account.

13. 13. The full account is a bit more complex. Quasi-indicators are restricted variables used to "depict" or produce a facsimile for understanding another's content; they are not referential terms that fully reveal that content. But their semantic contribution is not merely generic indexical modes, otherwise there would be no difference in what one attributes to the tallest man with (3) and to the shortest man in "the shortest man believes that he himself is wise." The quasi-indicator's anaphoric dependence upon its antecedent is also relevant in determining its significance. See my essay "On Depicting Indexical Reference," forthcoming in a memorial volume to Hector-Neri Castañeda, edited by William Rapaport and Francesco Orilia.

14. 14. I follow Searle 1983, Castañeda 1089a, and Richard 1990 in taking the de re/de dicto contrast as a distinction between different kinds of ascription, not between different sorts of attitudes (see Kapitan 1994).

15. 15. I note that nothing depends upon taking tokens of the names "Beijing" and "China" as referring expressions. Even if tokens of "China," say, are viewed as a truncated descriptions such as "the world's most populus country" of "the large country on the east coast of Asia," we still use spatial indexicals or sensory images in fixing their significance.

16. 16. See also, Castañeda 1990b, 132, in which several levels of self-consciousness are distinguished.

17. 17. Frank also claims that immediate self-awareness cannot be explained in an ontology limited to spatio-temporal objects, facts, and propositions. I do not see that the data about the types of self-consciousness treated herein -- internal self-reference, external self-reference, and the pre-reflective awareness of perspectival encounter -- cannot be handled in terms of objects, facts, and propositions, though I think the types of "facts" needed to accommodate experience are better classed as events.

18. 18. The author is indebted to Professors Manfred Frank and James Hart on an earlier version of this paper.