Transparency in Self-Reference

TRANSPARENCY IN SELF-REFERENCE (Jan Müller) My talk concerns an apparent tension in Gareth Evans’ account of self-reference and the role it plays in knowledge we have of ourselves as rational subjects of our attitudes and actions. On the one hand, Evans holds that the First Person Pronoun, just as other demonstrative expressions, can indeed possess a Fregean Sense: In ‘I’-thoughts, the First Person Pronoun directly relates to its object in such a way that its sense depends on the mode of representation of the object in thought. This mode of representation can very well be conceived of as incommunicable, which would entail that ‘I’-thoughts might be essentially private in that they cannot be shared between different thinkers. But if that is correct: how, on the other hand, can ‘I’-thoughts be understood as “objective” (as indeed they must be, if they are to be Thoughts at all)?1 I will first sketch the main points in Evans’ characterization of ‘I’-thoughts. I will then discuss a critique put forward by José Luis Bermúdez, that can be interpreted as a showcase example of an interpretation that takes Evans to develop a bottom-up account of demonstrative and self-reference. I do not believe that is accurate, and will in conclusion present a few indications to this effect that draw on the role of transparency in self-knowledge.

(1) Self-reference is, as Evans puts it, “the essence of self-consciousness […], that is to say, thinking, by a subject of judgements, about himself, and hence, necessarily, about a subject of judgements” (VR, 213). Self-consciousness consists in a subject having an actual ‘I’-thought, a thought that is exclusively articulated in judgments using the first person pronoun, and which in turn is based on the right informational input; and if it does, then the sense of the demonstrative expressions involved is object-dependend. “In self-conscious thought”, Evans says, “the subject must think of an object in a way that permits it to be characterized as the 1

Now, it has been argued that these problems dissolve if one were to focus on the peculiar ‘channels’ of information-acquisition and –distribution that facilitate knowledge of oneself; for, if an empirical account of this information-processing apparatus could be given, the question of the objectivity of a person’s ‘I’-Thoughts could be sidestepped. But such a sturdy naturalization of the model would miss that in self-reference, a person represents herself in the “indirect reflexive”, not herself as other.

1

subject of that very thought” (VR 213). It need not, however, necessarily think of itself “as the author of this very thought” (VR, 213) – “if, indeed, such a thing is intelligible”, as Evans writes. Depending on how we are to understand this “author”-metaphor, it might well be intelligible; for Evans, however, the crucial point is this: If self-conscious-thought is objectdependent, and if we are to avoid the pitfalls of Cartesian substance-dualism, then to think of oneself at all is, immediately, to think of an object in the world, that is an object whose existence is represented as being independent of the thinker and, indeed, independent of its being an object of thought. This in turn means that to have a thought of oneself is think of an embodied subject. The reason for this is that “our thinking about ourselves conforms to the Generality Constraint. And this means that one’s Idea of oneself must also comprise, over and above the information-link and the action-link, a knowledge of what it would it would be” for something that conforms to the fundamental identification of a person to satisfy the predicates one wants to couple with one’s own fundamental Idea of oneself (VR, 209). So the term “Idea” plays a double role here: he first designates fundamental types of thinking an object (the “fundamental level of thought”). An ‘I’-idea in this sense describes what it is for someone in general to relate to himself as himself, or to have oneself being given to onself in a way that differs from the mode of givenness of other objects or persons. One “has a fundamental Idea of an object if one thinks of it [the object] as the possessor of the fundamental ground of difference which it in fact possesses. (Such an idea constitutes, by definition, distinguishing knowledge of the object, since the object is differentiated from all other objects by this fact)” (VR, 107). Secondly: to actually have a fundamental idea of oneself is to think of oneself as an instance of this Idea, which implies the ability to “to know what is involved in locating oneself in a spatio-temporal map of the world […, for] no one can be credited with an ‘objective’ model of the world if he does not grasp that he is modelling the world he is in – that he has a location somewhere in the model, as do the things that he can see.” (VR, 211 a. 212). It belongs to the form of thought of oneself to think essentially of an embodied subject – an object that not only occupies a place in space in time alongside other, but does so selfconsciously. Only in this way it is possible that a judgement about oneself actually conveys an ‘I’-thought. In this way self-reference, in effect, puts demonstrative reference into practice – namely by shedding light on the way it is based upon an account of ourselves as the subjects of such 2

referring acts. So the notion of “space” rests upon a practice of identifying spacial relations; and this includes relations to oneself as sharing the space that, in a certain way, one helps to establish. Hence, to fully understand spatial demonstration, one has to give an account of what it is to find oneself in space. To have an idea of oneself, Evans says, involves the same two elements that are active in spatial demonstratives: “an element involving sensitivity of thoughts to certain information, and an element involving the way in which thoughts are manifested in action” (VR 207); those two elements, however, receive a new status in selfconscious thought. Self-conscious thought, although sharing “the same general character” (VR, 205) with other forms of demonstrative thought, is nevertheless sui generis.

(2) Self-knowledge, as opposed to knowledge of some person that per accidens is identical with myself, consists in thinking ‘I’-thoughts – that is, thoughts in which the sense of the referring expression, the actual ‘I’-Idea depends solely upon the subject’s way of gaining knowledge about it. Such thought, as all demonstrative identification, identifies its object directly. It is “identification-free”: it is presupposed for any descriptive account of the object, but does not in turn require an descriptive identification-component. Such thought immediately singles out the object it is about, and it does so in virtue of the ways the subject can gain knowledge of it, not in virtue of an additional identifying judgement. It should of course be noted that this freedom from identification and the subsequent immunity to error through misidentification are not designed to immunize arbitrary knowledge claims as if by a feat of “epistemological magic” (Grush 2007, 611f.).2 But Evans is not concerned with factual (or “empirical”) instances of knowledge-acquisition, but with the form of such knowledge in general. His Fregean Framework hence allows him to make a grammatical distinction between arbitrary mental episodes that might look like judgements – but are not judgements, for they lack the appropriate information-link between subject and object, and hence are for all intents and purposes not judgements about something, but mere judgement-attempts3; in order for a 2

It is precisely this line of reasoning that led Elizabeth Anscombe to her – often misinterpreted – claim that “I” does not refer – for it doesn’t, if one models reference solely in a Russelian framework. If one does not, then, of course, “I” refers – but the grammar of the term “reference” has been changed – which was just what Anscombe intended. 3

Cf. Grush 2007. The difference between a judgement-attempt and a judgement is that only in the latter the referring expression actually homes in onto its referent by means of the subject’s disposition to register certain

3

judgement to be intelligible at all, it has to be conceived of as enabled by something other than the judging subject alone4. Now, Evans is understandably cautious about the notion of identification-free immediacy. But the question whether a factual, singular utterance (be it outspoken or just “within one’s mind”) actually is a judgement is, it could be said, a practical question; that is, it is a question the answer to which would call for additional judgement about the conduct of the subject, both in the first and third person perspective, etc.5 Evans only shows, by way of conceptual analysis, what is already in place when the distinction between an attempted and a successful judgement is intelligible: “someone who understands a term as referring to himself must be disposed to regard, as relevant to the truth or falsity of certain utterances involving that term, the occurrence of certain experiences which he is in a position immediately to recognize” (VR, 233). And it is now that the privileged access to experiential information that pertains also to other demonstrative references and the according thoughts is not only immediately used, but can also be justified: Knowledge of this object there is founded upon the fact that I see it, or: that I can judge that I am would be reliably disposed to see it, were the conditions right, etc.6 Now, Evans is understandably cautious about the proposal that knowledge should be justified by reference to the knowledge’s subject and its mental activities. After all, his conception of object-dependence is essentially designed to prevent such idealistic notions of pure constitution. Of course, this begs the question which status can be assigned to the information link – that is, the privileged relation one has to the object of one’s thought; for it is this relation which essentially makes it conceivable that a subject can refer to the object independent of prior identification, or that demonstrative thought presents the object itself.

perceptual information; only in actual judgements is articulate a thought about the object whose content contains the mode of presentation. This does by no means imply that demonstrative judgement is susceptible to errors through misidentification; it only shows that naturally judgement attempts can fail. However, to understand what it means for a judgement attempt to fail, one already employs the idea of a successful judgement, that is: the idea of a judgement’s freedom from identification and its subsequent immunity against misidentification is already at work. 4

Being disposed to answer to such information, or being susceptible to it, is of course a necessary condition for judgement; however, receiving information und thus being in certain informational states does not cause the judgements that take them as their material: “the judgement's being a judgement with a certain content can be regarded as constituted by its being a response to that state“, rather than „the state [...] be regarded as constituted by dispositions to make certain judgements“ (VR, 229). 5

Evans makes this point regarding the solipsist’s account of the sense of „I“, pointing out that what is wrong with the solipsist is not a theoretical argument, but rather his attitude towards his relation to the world. 6

„A subject does not need to have information actually available to him in any of the relevant ways in order to know that there is just one object to which he is thus dispositionally related” (VR, 216 Fn. 21).

4

The notion of the transparency of our mental states, then, fulfils a twofold purpose. First, it acknowledges that the idea of object-dependent thought demands that we either assume a Cartesian thinking substance, or that we accept that the very idea of one’s own mental states and activities has to be essentially linked to the objective relation to the world. As the famous passage reads: “in making a self-ascription of belief, one's eyes are, so to speak, or occasionally literally, directed outward — upon the world” (VR, 225). The second function is that Transparency finalizes Evans’ externalist take on the Kantian idea that intuition, the content of experience, necessarily conforms to certain forms. This includes subscribing to the methodological thesis that in order to find out the subject’s contribution to this form it is necessary to start with her involvement with the perceived content. In perception, the contact with one’s object – if, indeed, there is such contact – is seemingly immediate. This immediacy changes, however, when referring to one’s own mental states and attitudes, for we do relate to our mental states via our engagement with the world. What changes in this shift of perspective is not the sighted object, but rather the mode of the object’s representation. To borrow from another language-game: the immediacy of the object’s givenness turns out to be already mediated by thought, just as an immediate occurence of coloured light might turn out to be mediated (and, in fact, facilitated) by a glass prism which is transparent to the naked eye, and would be invisible, were it not for the light refraction that indicates its effect. This change in the “mode of givenness” and the subject’s awareness to it indicates what we would call “self-conscious thought”; and this change of the mode of givenness affects the model of direct demonstrative that Evans was concerned with before. Object-dependent demonstrative reference is the basis of, and in turn becomes fully intelligible only through, self-reference. This is indicated by Evans’ metaphorical introduction of a “procedure” to “gain knowledge of one of his own mental states“ (VR, 225). Of course, one does not infer from a certain statement, p, that ergo “I believe p”7. Upon asserting p, it is characteristic for a rational being 7

Campbell (2005) holds that there is indeed inference involved. He writes: "This is a transition you are entitled to make, from a conscious state you enjoy to a judgment. If a thinker comes to judge, by this means, that he sees that that desk is covered with papers, his judgment can thereby be knowledge. [...] This is the starting point of a general model of self-ascriptive knowledge of one’s own perceptual states. Because the thinker sees that p, he moves, rationally, to the judgment I see that p, and thereby gains knowledge that he so sees" (Campbell 2005, 222). But, as Boyle points out, this would be absurd; for, „even if it is true that P, this by itself has no tendency to show that I believe it“ (Boyle 2011, 230). It does show it if I presuppose what alledgedly was shown in the inference: That I hold p to be true. Self-

5

that he is able to justify this assertion; it may be demanded of him that he present his reasons. And one does not “produce” these reasons by “producing”, e.g., a certain believe – one upon reflection gives one’s reasons (the term I borrow from Matthew Boyle’s recent work).8 To have an Idea of oneself means just that: to be able to point out this immediate connection between states of affairs in the world and our mental states and attitudes that are transparent to them. “If there is to be a division between the mental and the physical, it is a division which is spanned by the Ideas we have of ourselves” (VR, 213) – that is: it is by reflecting upon our judgements and the reasons we can give for them that we can draw an intelligible distinction and henceforth competently navigate between “the physical” and “the mental”.9 The model is this: The subject’s being in a certain informational state makes it possible for his rational capacities to respond to this state; and upon reflection on this “mental” activity (which might as well be manifest in overt action) he can relate to himself by ascribing attitudes, volitions, intentions, and even by metaphorically “working out” in which state he is or ought to be in.

(3)

knowledge is formally inherent in knowledge of the world; reflection is just the name of the rational capacity we exercise constantly to make it explicit and articulate to ourselves. – An analogue argument holds for the immediacy of „experiential evidence“: Just as „seeing” does not mean: observing based on evidence. For something to count as “evidence”, it has to play a certain role in a procedure that at least involves identifying judgement and inference, namely the judgement that something counts as evidence for something. Seeing something as something, on the other hand, is just that: one episode of the exercise of the rational capacity to see something (this, thing); cf. Cassam 2009, 9. 8

The notion of reflection is put forward – at least in the analytic philosophy of mind – by Richard Moran (2001), and recently defended by Matthew Boyle (2011, 237): “suppose I knowingly accept P, and understand that what I accept as true just is what I believe. Then I have everything I need to knowingly judge: I believe P. All I need do is reflect. “ 9

It is for this reason that Evans can, as he does, omit further inquiry into the action-link in self-knowledge; for in effect he already has pointed out how our actions, knowledge of what we are doing, is immediately available to us. Just as we know from our state of believing by, upon reflection, noticing our disposition to hold certain assertions to be true, that is, to be assertions based on which we would justify further beliefs etc. – just as that we know of our actions as being actions for a reason, or as being rational action. Evans, in other words, makes it clear that the distinction between bodily and mental activities is not a distinction between classes, but rather a distinction between aspects of activities that we can ascribe to ourselves and to one another. And such ascribing judgements, he argues, always depend on experiential information of the world, acquired through the proper channels – channels such that the identity of the object perceived rests solely upon the ways in which it is presented to the subject. Cf. VR, 224, Fn. 34: Evans neglects “our knowledge of our own actions”, where the same consonance would have to be explained: between the perspective upon the agent himself (“the subject of desires, thoughts, and intentions”), and the perspective upon an object that, by moving, changes and moves other things. This is precisely what Anscombe had in mind by alluding to the nature of practical knowledge as “cause of what it understands”; where Evans, forced by the language-game on “singular reference” he is participating in, stresses the duplicity of perspectives, Anscombe conversely stresses the requirement to ultimately understand their unity.

6

Of course, the overall model rests upon this notion of “being in an informational state”. Its function is to support the idea that self-knowledge is objective, that is: to provide for the possibility that self-conscious judgement conveys a Fregean ‘I’-thought. This, of course, entails Frege’s dictum that “everyone is presented to himself in a particular and primitive way in which he is presented to no one else" (cit. UD, 312). Evans takes this idea to be an indication that his interpretation of the object-dependence of demonstrative thought in general, and of self-identification in reference in particular, is sound.10 This characterization however does seem to imply that since only I am given to myself in the special way that pertains to my ‘I’-Idea, only I can have an ‘I’-thought that depends solely on the unique way I am given to me; conversely, it seems as if ‘I’-thoughts cannot be shared with other thinkers.11 And this seems to clash violently with the idea that communication is, as Evans writes, “essentially a mode of the transmission of knowledge” (VR, 310; italics JM). Such is the objection José Luis Bermúdez raises; for, if there is an “information component of an ‘I’-idea [as] a set of information links that can hold only between a person and himself”, then “thoughts containing such an “I”-idea are not shareable […, since] [m]y sensitivity to information about myself that is immune to error through misidentification is not something that you can share. Hence, building such sensitivity into the sense of the first-person pronoun will have the result that “I”-thoughts are incommunicable.” (Bermúdez 2005, 185f.) But uncommunicable thoughts, Bemúdez reasons, would stand against the very idea of knowledge, and immediately give rise to scepticism. Let me briefly discuss the solution Bermúdez proposes, because it is exemplary for worries that stem from a certain reading of Evans, that I will return to. – Bermúdez proposes to strip the information component of reference of the crucial role it plays in Evans’ account, and assign this role to the spatial component, that is: the ability to locate oneself in space. For, Bermúdez argues, the “requirement here is […] that one knows what is involved in locating oneself in public space. If one understands that one is such a person, then one can understand the truth condition for any sentence of the form “I = ∂” where “∂” is a name or description picking out a particular person." (Bermúdez 2005, 191) 10

If “everyone is presented to himself in a primitive way”, then obviously “Frege did not hold that all ways of thinking of objects must involve thinking of those objects as uniquely satisfying some description" (UD, 312). 11

Evans argues that this poses no problem, since Frege only took shareability as sure indication, not a necessary condition for objectivity: It “is the inference from shareability to objectivity which is of paramount importance to Frege, rather than shareability itself. Since an unshareable thought can be perfectly objective […] there is no clash between what Frege says about 'I'-thoughts and this […] aspect of his philosophy" (UD, 313).

7

But what this proposal confuses is the difference between thinking of oneself* in the indirect reflexive and thinking of oneself-as-another. For a judgement of the form [∂ = I] expresses different thoughts if we take ∂ to designate a Fundamental Idea of a Person (any person that can either be characterized descriptively, or pointed out demonstratively) on the one hand – or if we take ∂ to designate an exemplification of my fundamental Idea of myself, that is: my singular ‘I’-thought. To attach as little value to the information component in self-reference as Bermúdez does amounts to the suggestion that referring to oneself is merely to identify oneself as a certain kind of object. But as Evans points out: Even though it is part of one’s self-conception to be able to refer to oneself as an object, what is essentially for selfknowledge is that the mode of givenness of that object comes, so to speak, in a special flavour. In self-conscious thought, the object is represented as subject. And this, Evans holds, leads back directly to the crucial role of experiential information: “The idea that there is an objective world and the idea that the subject is somewhere cannot be separated, and where he is is given by what he can perceive” (VR, 222).

(4) Let me try to bring out the problem with Bermúdez’ account in different way. Bermúdez understands the grammar of “subject” to be kind-referring and to a certain extent co-extensive with the descriptive term “person”. But, to borrow from Peter Geach, the term “subject” has an attributive role: it indicates a mode and manner of representation. So Bermúdez’ attempt to secure objectivity by means of securing communicative verifyability is dearly bought by abandoning a central point of Evans’ neo-fregean account. For although the objectivity and truth-value of a thought is independent of its actually being thought by any individual thinker, its sense actually does depend on actually being thought; for it is only in actually being entertained that an ‘I’-thought rests solely upon the special way in which a certain rational subject is presented to itself. The actuality of being thought of course cannot, for want of objectivity, be an essential feature of demonstrative thought; however, in the case of demonstrative thought of myself as a rational subject (or in self-knowledge), it is nevertheless necessary to understand the ‘I’-thought as being actually thought; for it is the activity of this thinker which determines the sense of the referring expression (the representation of the object as subject). And this only comes out when not only the ability to locate oneself among other objects, but also the ability to articulate the attitudes and activities (or “mental states”) 8

which are transparent to locating objects, is exercised. For this exercise of rational capacities, Evans says on various occasions, relies on a certain “background” that is already indicated in the generality constraint, videlicet „an appreciation of the fact that the kinds of evidence which he is prepared to recognize as relevant to the ascription of the predicate to others bear also upon the truth of his claim, and a willingness to recognize, as relevant to the ascription of the predicate to others, evidence of their having executed the same procedure […]. Without this background, we might say, we secure no genuine 'I think' […] to accompany his thought (‘p’): the 'I think' which accompanies all his thoughts is purely formal.” (VR, 226)12 There is no gap between the particular and primitive way in which every subject is presented to herself, and the fact that this is possible precisely because the subject is able to conceive of his thought as exemplifying a general ‘I’-Idea that also figures in other’s thought and speech. It is bridged by the “background” (that Evans only hints at) of a shared practice of interrelated judgements and ascriptions; and it is within this practice that the adequacy of ascriptions to others is weighed and judged. From this perspective, it becomes trivial that actual ‘I’-thoughts cannot be shared; for what individuates thoughts is their sense, the way in which they relate to their object. But from this it does not follow that their objectivity cannot be accounted for.13 Although I cannot share your ‘I’-thought, for the first person pronoun would identify the wrong object, I can very well understand what it is for an ‘I’-thought to be actual; and I can very well understand that in your self-referential ‘I’-thought, you relate to the object of your thought in just the same way as I would do. Bermúdez thinks that if we cannot share actual ‘I’-toughts, we can never judge with certainty whether a factual utterance or judgement-attempt ever amounted to an articulation of Thought, that is: whether they conveyed knowledge. Hence, he designs a method to testing ascriptions third-personally by reference to a factively presupposed spatial-

12

Evans mentions this in passing, and seemingly only to ward off the idea that the self-ascription of belief (and other mental states) could be modelled after the idea of perception; so to speak, a kind of second-order perception (of perceptive states); cf. also: “The subject who genuinely has this capacity for self-knowledge must understand the content of his judgement ‘It seems to me as though p’, and his understanding of it must determine it to have a content different from that of the judgement '’Possibly p‘ [etc. ...]. This requires a background of a sort [...]. [W]e may have here an interpretation of Kant's remark about the transcendental ‚I think‘ which accompanies all our perceptions. Without the background, we have at most a formal ’I think‘“ (VR, 228). 13

On the contrary: If the idea of self-reference, that is: of true and actual self-reference, presupposes “a willingness to recognize” the rational conduct of others as essentially relevant to one’s own self-ascriptions, then, we might translate, self-consciousness essentially is located in a shared space of reason, a rational practice (I take this to be exciting only regarding the interpretation of Evans’ thought). And this in turn means that, of course, our actual ‘I’-thoughts conform to the generality constraint.

9

temporal order.14 But such a third-personal method of testing would only be needed if otherwise the success of our communicational efforts were generally dubious, so that such a test would be an essential part of understanding what it means to be a subject, or that other minds exist. But this, Evans holds, isn’t the case. We already share a form of thought; and this sharedness is exemplified in every actual ‘I’-thought as a condition for the possibility of its structure15. That for every singular occurrence of a judgement-attempt the possibility of failure remains comprises just as little sceptical horror as does the insight that we do not share one another’s actual ‘I’-thoughts. Both cases simply present us with the practical problem to coordinate our perspectives upon one another and the world; they “cannot be transformed into a substantial account of what the subject means by ‘I’” (VR, 234).

(5) Bermúdez argued that because informational states16 are non-conceptual and strictly private, it follows that ‘I’-thoughts are uncommunicable and therefore unshareable. This shows that Bermúdez understands the term “informational system” and its relatives factively – as a description of a factual information processing system that figures in a bottom-up account: Starting from a description of the informational system delivered from the third-person perspective, and leading up to an account of self-knowledge delivered first-personally. This narration then creates a gap between a descriptive, third-personal, and an expressive (articulatory, or reflexive) first-personal account.17 Of course, in Evans’ account these two perspectives figure prominently; they are, however, not separated, but rather form a tension that is mediated by the fundamental Idea of a Person. If, however, one reads Evans’ proposal as comprising a bottom-up approach, then the promise 14

His proposal is similar to the proposal of the Solipsist that Evans discusses: The solipsist reacts to the problem Bermúdez raises by purporting that at least for himself, he is quite cabaple of explaining the meaning of “I” (namely, “By ‘I’ I mean the person such that, when he is in pain, something frightful is to be expected”, VR 234f.); Bermúdez in turn tries to reunite a crowd of solipsists by giving them at least an external method of testing third-personal ascriptions. Evans, however, is not worried by the Solipsist’s demand for such methods. 15

“The point is that 'I think' (or 'it seems to me’) acquires structure (“XI thinks' or 'it seems to XI', with ‚I’ in the argument-place) only when it is related to (at least possible) other exemplifications of the same predicate” (VR, 228 Fn.) 16

upon whose deliveries any self-referring judgement is based, and from whose deliveries the object-dependence of the resulting thought stems… 17

We can see now that Bermúdez does not try to bridge this gap; he merely notices that Evans seems to set out from a descriptive starting-point while restricting epistemic access to it, and proposes to swap this descriptive starting point with one that is “public” from the beginning.

10

of mediation can never be held: Such readers are bound to take seriously, on the one hand, Evans evaluation of singular demonstrative thought’s conceptual form; the clou of this approach of course being that this conceptual form is essentially shared and manifested in a rational practice of normative evaluation, that is a practice which constantly and consistently judges factual judgement-tokens according to normal or paradigm cases.18 On the other hand, such readers are liable to challenge just this reliance upon a normal practice by stretching the boundaries of what might seemingly be counted as conforming to a paradigmatic normal case; and it can be shown that Evans’ model falls short, for example, in providing a subject with the possibility to realize its mistakes when surgically wired to a wrong source of information. This inevitably leads to the conclusion that the problem lies with the very idea of a singular, immediate source of unverifiable information itself19 – an idea that Lucy O’Brien quite rightly associates with “an observational model” of self-reference (O’Brien 1995, 246). But it is just this characterization of Evans’ account “observational” that begs the question. It indicates two ideas: First, the idea that self-conscious knowledge of our activities, attitudes and states, is based upon an observation. Of course, one cannot observer one’s own mental states directly, for they are transparent against the world – but one learns about them by “observing” occurrences and states of affairs in the world. Secondly, “observational” indicates the opinion that observation is our first and foremost, most natural access to the world. This world, in turn, is modelled after realist and empiricist intuitions: a de facto impersonal, public spatial and temporal order, for which our participation in it, although necessary for us, is purely incidental20. But to suppose this was Evans’ view would move him too close to the Russellian descriptivism he explicitly refuses. For, it meant to insinuate that a descriptive approach to the world is most natural, wereas demonstrative reference needed special clarification. But Evans on the contrary tries to elucidate that direct demonstrative reference is the basic approach, for it is in this relation to the world that our judgements about it have sense that depends on the object’s presenting themselves to us, and us being susceptible to them.21 18

Cf. VR, 257.

19

„A suggestion to the effect that the subject identifies itself as that thing which is the source of most of the information it has received, is hopeless. For one thing, it involves a conceptual element that Evans does not countenance as being involved in normal uses of "I". Further, this identification cannot underwrite first person reference because it already involves the subject thinking about itself, in a first personal way, as the receiver of information. (O’Brien 1995, 245) 20

It adds neither to its existence nor objectivity.

21

By this, Evans basically inverts Russell – where Russell introduces a notion of “acquaintance” to help with the sceptic problems emerging from his account of descriptive reference, Evans shows how such descriptions are even possible only in the light of our acquaintance with the world.

11

(6) So I would like to conclude that this rather suggests a top-down reading of the information component’s function in object-dependent thought. Let me briefly sketch this. Evans writes: The “informational states which a subject acquires through perception are nonconceptual, or non-conceptualized. Judgements based upon such states necessarily involve conceptualization: in moving from a perceptual experience to a judgement about the world […], one will be exercising basic conceptual skills” (VR, 227), only to add that “this formulation (in terms of moving from an experience to a judgement) must not be allowed to obscure the general picture” (ibid.). Evans tries to navigate between two equally absurd ideas: First, that this “movement” is solely the subject’s doing, for a Thought is object-dependent; and secondly that immediate perception could somehow “cause” or “bring about” a judgement; for judging is essentially a rational act. The danger of obscurity, I think, arises from this: Evans accomplishes his navigational feat by apparently narrating the process “of conceptualization or judgement” which “takes the subject from his being in one kind of informational state […] to his being in another”, and, at the same time, the evolution of conceptual from non-conceptual content. But we saw that if we take this narration literally, it only serves to accentuate the gap between the two ideas. Taken literally, the narration is an account of the problem, not its solution. But Evans shows that the direction of fit of this narrative is reversed. That is why he can omit from giving criteria to decide whether a factual “cognitive state” satisfies the conditions for “knowledge”: Reflection, he says, “is a way of producing in himself, and giving expression to, a cognitive state whose content is systematically dependent upon the content of the informational state, and the systematic dependence is a basis for him to claim knowledge of the informational state” (VR, 228), but, he adds, this basis for a claim can itself not be object of consideration: “in no sense has that state become an object to him” (ibid.). Were this passage about a singular knowledge claim it would follow that one could claim knowledge of one’s mental states without being able to produce empirically verifiable, descriptively accessible evidence. But Evans doesn’t talk about factual “cognitive states”. He analyses the form in which knowledge is überhaupt 12

intelligible, and formulates a clause that secures that knowledge cannot be conceived of as constituted solely by the factive consciousness that possesses it, although it plays an essential role in it. Hence, it belongs to the grammar of the term “cognitive state” that it rests upon a precursor which made it possible and was itself not, in turn, somehow “constituted” by this cognitive state. To understand this concept, or to be familiar with the grammar of “cognition”, means to understand how it is generally conceivable that a judgement hits its mark. In this, the idea of an “informational state” functions as a logical condition for judgement in general. Such a state can be reconstructed as having enabled and, in retrospect, even caused a judgement. But it cannot be identified without reference to the judgement it developed into22 – for it is nothing but the precursor of this or that actual judgement.23 This top-down interpretation of course entails two consequences. First, it would imply that the relatively simpler varieties of demonstrative reference Evans analyses at first become fully comprehensible only in the light of self-conscious thought. While at the former stages a rollback into the framework of Russellian descriptivism is at least possible, in self-reference the whole framework has finally shifted into a decidedly Kantian gear. It is only then that our reliance upon the experiential relation to the world becomes intelligible, for only by virtue of the ascription of transparent mental states that can only be accounted for as exemplifications of our conceptually structured relation with the world can be understood how this reliance is immediately and permanently at work. (Which, in turn, conforming to the generality constraint, indicates a general practice of self-conscious judgement and action.)

22

One could, of course, refer to “informational states” that are individuated by means of empiricial criteria. But in doing so, one would have switched the grammar of the term “informational state”, using a homonymy, and have entered a descriptive language game. I take the aforementioned to be an indication that Evans, albeit struggling with the empiricist language game of his tradition, tried to develop a top-down account of the role of experiential states in demonstrative and self-reference. 23

It cannot, or not in any systematically substantial way, be understood as something else than “the state that did provide the material for this or that actual judgement”, for it cannot even be conceived of independently of the judgement it enabled. “For an internal state to be so regarded, it must have appropriate connection with behaviour – it must have a certain motive force upon the actions of the subject” (VR, 226); and: “any informational state in which the subject has information about the world is ipso facto a state in which he has information about himself” (VR, 230): Both characterisations are grammatical remarks on the idea of such a state, not a criterial distinction – for pain of misusing the metaphor of “force” in the first, and to confuse “being in a informational state” with “having an information” in the second instance. – This is crucial: in the same way as a factual state that could be the fulfilment of this condition would be immediate for any subject and couldn’t be subject-matter of his interest without changing its essence – just in the same something like such states would be generally out of bounds for human cognition. This isn’t an empirical assertion; if it were, then Bermúdez’ critique would be most appropriate indeed. It is a grammatical assertion.

13