1. Introduction
Evans is very clear about what his theory of demonstrative thought comprises.Footnote 1 A statement he iterates time and again in the Varieties of Reference is that his account is made up of two main constituents: an information component, and an action component. Here is one of many such statements:Footnote 2
Section 6.3 contains the materials for a functional characterization of ‘here’-thoughts, with two components: a component involving the relation between ‘here’-thoughts and perceptual intake, and a component involving the relation between ‘here’-thoughts and action. We can see the information-link which is a necessary condition for possession of a ‘this’-Idea (6.2) as constituting a component in a corresponding functional characterization of ‘this’-thoughts.Footnote 3
Roughly the same words are iterated in a later chapter, where the concern is first-person thoughts:
… the Idea which one has of oneself involves the same kinds of elements as we discerned in the case of, say, ‘here' (6.3): an element involving sensitivity of thoughts to certain information, and an element involving the way in which thoughts are manifested in action.Footnote 4
As these quotes – and many other similar passages I could have easily quoted – make abundantly clear, Evans is unequivocal about what he takes his theory of first-person thoughts to be.
Given the emphasis Evans places on the double feature of his theory, it comes as a surprise that these warnings have gone largely unheeded, or what is more likely, simply ignored by his critics. What has transpired, in the years following the issuance of Evans' major work, is that one of the components – the information element – has received all the attention, and is treated as the sole centerpiece of his theory. As for the other component, it simply does not exist.
Whatever the exact reasons may be, this one-sided reading of Evans has cast his theory in a very unfavorable light. Reduced to a theory with an information component and a couple of supporting elements, Evans' theory turns out to be just another one of those perceptual model theories of self-reference and self-knowledge, and as such is easily subject to the standard arguments against such theories. The theory can't explain for example why self-knowledge is not always observational knowledge.Footnote 5 It has trouble explaining what separates ‘I’-thoughts from other demonstrative thoughts. Worst of all, the theory can't even take a lame punch from an ancient Cartesian war-horse of the kind exemplified by Anscombe's sensory deprivation argument against a perceptual/informational conception of ‘I’-thought.
Would the critical study of Evans and the line of research he helped usher in have taken a different path if his warnings had been heeded? The idea is tantalizing, but it is not the objective of this paper to engage in such speculations. The question I want to pursue is a rather different one. I want to know what Evans' theory would look like if we take his words literally and put back into it the forgotten ‘action component'. What impact would this have on our view of Evans' theory?
All this of course hinges on what ‘the action component’ is. Does Evans' theory really have such a component? And if it does, what's its content? What job is it doing in the theory? Answering these questions is the task of this paper. But here is a quick look-ahead: I'm going to show that not only does Evans' theory feature a second, and equally critical component, but that at the core of this second element is a conception of intentional action epitomized in his memorable one-liner, ‘I do not move myself; I myself move', which Evans utilizes in his account of ‘I’-thoughts. With this component restored, the charge of Evans' theory as a perceptual model of such thoughts falls apart, and the theory turns out to have enough merit to project itself as a legitimate contender for a plausible account of ‘I’-thought.
The structure of the paper is as follows. In section 2 I examine a widely accepted assessment of Evans' account of demonstrative thought – one that is the result of ignoring Evans' advice about the twin feature of his theory. Section 3 is an attempt to piece together Evans' ‘action component'. Section 4 offers a picture of what Evans' theory looks like with the missing action component put back in. I end the paper with some brief remarks on the contemporary relevance of Evans' theory thus restored.
2. Evans As the Information Component
The account of ‘I’-thought Evans offers in the VR is quite complex. If we take his words for it, it comprises two core components. But he is also known for saying things that seem to suggest otherwise (see VR, 209, line 6 ff.). Compounding the matter is the fact that in addition to ‘components’, the account also features several other elements, such as the so-called ‘Generality Constraint’, ‘Russell’s Principle', the notion of demonstrative concept as a ‘practical skill’, the notion of a fundamental level of thought, the notion of an ‘identification-free thought’, and the idea of ‘special ways’ the thinker of a demonstrative thought has for gaining information pertaining to the object of her thought. Figuring out what these pieces really come down to and how they relate to the centerpieces and to one another is a daunting task.
One popular interpretation of Evans characterizes his theory as one underwritten by two claims – or more precisely, a principal claim plus one largely tacit assumption. The claim is of course the so-called ‘information component’, which states that all demonstrative thoughts are information-based thoughts. This core component is taken to be buttressed by several minor pieces (some of which we mentioned in the previous paragraph) which function essentially as constraining or strengthening conditions on the centerpiece. The tacit assumption, on the other hand, is a conception of perception according to which perception is essentially a receptive faculty of the mind, much as volition is an agential faculty of the person. In what follows I'll refer to any reading of Evans' account of ‘I’-thought – when pared down to its bare bones – that takes it to comprise these two notions as ‘Evans as the information component’ reading (or ‘EAIC’ for short).Footnote 6
Viewed from the EAIC perspective, Evans' theory comes out a breed of the ‘perceptual model’ theory. The connection is supposed to be straightforward: If an ‘I’-thought is an information-based thought and information is always gained through the exercise of a receptive faculty, then any information-based theory of ‘I’-thoughts is a perceptual model of such thoughts.
One of the early statements of this view appears in Lucy O'Brien's ‘Evans on Self-Identification’, which focuses exclusively on Evans' information component in assessing his theory.Footnote 7 What she says about that component is revealing. Here are two quotes from that article:
This is what Evans aims to do in The Varieties of Reference. He offers an account of the information channels that are in operation, that will, he claims, underwrite first person reference. ‘I’ is treated along the lines of other demonstratives. Unlike Descartes, Evans does not find it acceptable to treat first person reference like some sort of internal demonstrative underwritten by introspection, but rather treats it as securing demonstrative reference to an external object: a person, a human being on the basis of information we have about that object.Footnote 8
Evans … claims that my kinaesthetic knowledge, my sensations of the position of my limbs, my knowledge of my environment all contribute to the ability I have to conceive of myself as an object in the world with a certain orientation, standing in certain relations to others. And importantly, he claims that it is this knowledge which serves to distinguish me as an object in the world from all other objects in the world, and thus enables me to self-refer.Footnote 9
O'Brien's understanding of perception as a receptive faculty of the mind and her reading of Evans based on this take of the notion is most clearly seen in her supposed contrast between Descartes' approach to self-reference and Evans'. The former, according to her, takes self-reference to be underwritten by ‘introspection’, whereas the latter takes it to be grounded in ‘the basis of information’, obtained from ‘an external object: a person, a human being’.
O'Brien is not alone in adopting the EAIC reading of Evans. The same interpretive stance has also been adopted or endorsed by many other writers. In a survey piece on names and definite descriptions, Jason Stanley expresses agreement with O'Brien regarding the latter's criticism of Evans on ‘I’-thought. As Stanley puts it, demonstrative thoughts such as ‘this’-thoughts are thoughts one would express with the use of ‘perceptual demonstratives’ (i.e., terms such as ‘this’, ‘that’, ‘this color’, ‘that shape’, etc.). Such thoughts, Stanley maintains, are significantly different from ‘here’-thoughts and ‘I’-thoughts. The difference renders dubious attempts such as Evans' to give a ‘unified account’ of ‘I’-thoughts and other demonstrative thoughts on perceptual ground:Footnote 10
The case is more difficult in the case of “I” and “here”, for their reference is “guaranteed”. See Evans (1982, chapters 6 and 7), for an attempt to fit an account of these words into a model more closely paralleling perceptual demonstratives than seems, prima facie, to be possible, and see Lucy O'Brien (1995) for a recent critique of Evans's account.Footnote 11
In an article written several years before O'Brien's and Stanley's, Andy Hamilton expresses a similar dissatisfaction with Evans' account.Footnote 12 To be sure, Hamilton's chief concern in that article is not Evans but rather the various responses to Anscombe's notorious claim that ‘I’ is not a referring term. Hamilton groups philosophers who have responded to Anscombe's claim into two groups: those who conclude that some version of Cartesian theory of self-reference must be correct, and those who opt for what may be called ‘the bodily view’ – the idea that self-reference is reference to a bodily being, a human organism for example. But Hamilton's EAIC reading of Evans is hard to miss. The following remark of Hamilton's bears this out: ‘More radically, one may reject the Fregean model, as Gareth Evans does in assimilating self-reference to demonstrative reference.’Footnote 13 The similarity between Hamilton and Stanley is striking. In both cases, the underlying assumption is the same: demonstrative reference, epitomized in demonstrative reference to perceptual objects, is different from self-reference, the two are treated as similar cases by Evans, and this way of handling the two cases makes Evans' theory ‘radical’ (Hamilton) and ‘difficult’ to maintain (Stanley).
Hamilton's complaint against Evans actually goes further. Evans, according to Hamilton, simply turned a blind eye toward the blatantly obvious problems of the perceptual model theory of ‘I’-thoughts: ‘[Anscombe’s] article shows how easy it is to succumb to Cartesianism; but it does not consequently show that the right way to escape its dangers must be to adopt Gareth Evans' pose of sublime indifference.'Footnote 14
The view these authors share is that, by taking the information component (and the special ways we have of gaining such information) as central to his theory, Evans is essentially committing himself to a perceptual model of ‘I’-thoughts. But the perceptual model, his critics claim, is the wrong kind of theory for such thoughts, and probably for ‘here’-thoughts as well.
The standard arguments against the perceptual model of first person thoughts are familiar enough; so I won't rehearse them here. Earlier I quoted Hamilton's characterization of Evans' stance on the issue as a ‘pose of sublime indifference’. A casual remark no doubt, but it speaks volumes. It represents a widely shared assessment of Evans' theory – that it has a vulnerability all too clear to everyone but Evans.
To the critics, the verdict is clear: a perceptual model of ‘I’-thought is simply the wrong type of theory for such thoughts, and since Evans' theory is a variant of the perceptual model, it is destined to fail.
3. ‘The Action Component’
3.1 Outline of an Alternative Account
But the EAIC reading is not forced upon us. The simplest, and most-straightforward reason for rejecting this reading is its disregard of the plain words of the author it criticizes. I think that if an author says his theory comprises thus-and-so, then we should simply heed what he says and take his words for it unless there are strong reasons against doing so. In addition, the EAIC reading renders Evans' theory unnecessarily crude and naive, hardly the type of stuff one would expect from a philosopher of Evans' caliber. But more importantly, there is readily available an alternative reading that comports well with both the language and the spirit of Evans' work.
The alternative reading I'm going to delineate goes something like this: Evans' theory of first-person thoughts is centered around two critical elements, an information component and an action component. The job of the information component is to account for information-based ‘I’-thoughts – that is, thoughts which feature the concept of ‘I’ as object (e.g. what the ‘I’ in ‘My arm is broken’ expresses). The job of the action component is to account for genuine ‘I’-thoughts – first-person thoughts expressive of self-consciousness (the ‘I’ in ‘I myself move’), and the action component is made up of two strains of thought. The first is the thesis ‘I manifest self-conscious thoughts … in acting.’Footnote 15 The second is an conception of intentional action that underwrites the first one. On this view, an action is what it is because it is something I (decide to) do, not something with a particular mental genealogy. In contrast to her knowledge of a physical event with a particular mental history, the self knowledge an agent expresses, in the form of a genuine ‘I’-thought, of what she (decides/intends) to do, is not perceptual or observational knowledge; it is self-knowledge immune to two types of errors. First, it is self-knowledge immune to the type of errors to which ‘I’-thoughts are commonly taken to be immune, that is, failure to refer, and referring to the wrong object. Second, and more importantly for what concerns us here, it is a claim not refuted solely on grounds of observational facts about the agent (e.g. an onlooker's remark ‘But your arm is not up’ does not refute my announcement ‘I’m raising my arm' when my arm fails to go up but not for lack of trying). Having the action component as an integral element thus prevents Evans' theory from collapsing into a mere variant of the perceptual model of self-conscious thought and self-knowledge. It is in this sense that the action component is a centerpiece of his theory.
This, then, is the gist of the alternative account; the rest of the paper is devoted to substantiating it. For ease of reference, I'll call this view the ‘Evans with the Action Component’ reading (or EWAC for short).
3.2 Action a Distinct Notion in Evans
Evans addresses a wide range of action- and behavior-related issues. Three of these in particular feature prominently in his discussion. There is, in the first place, the discussion of issues concerning full-blown intentional action, evidenced in passages like the following:
It is true that I manifest self-conscious thought, like ‘here’-thought, in action; but I manifest it, not in knowing which object to act upon, but in acting. (I do not move myself; I myself move). Equally, I do not merely have knowledge of myself, as I might have knowledge of a place: I have knowledge of myself as someone who has knowledge and who makes judgements, including those judgements I make about myself.Footnote 16
Then there is Evans' wide-ranging discussion of behavior and behavior-related issues. Here are two memorable passages, one from the VR, and the other from the article ‘Molyneux's Question’:Footnote 17
Egocentric spatial terms are the terms in which the content of our spatial experiences would be formulated, and those in which our immediate behavioural plans would be expressed. This duality is no coincidence: an egocentric space can exist only for an animal in which a complex network of connections exists between perceptual input and behavioural output.Footnote 18
… spatial concepts applicable upon the basis of one mode of perception must generalize to the other. There is only one behavioral space.Footnote 19
In addition to action and behavior, Evans also devotes a significant amount of his discussion to a capacity he calls ‘practical skill’ ('practical capacity', ‘practical abilities’, see VR, page 236, para. 4 for a typical statement).
Philosophers have had surprisingly little to say about this aspect of Evans' thought. That Evans might have a view on intentional action seems to be an idea never crossed the mind of his critics. But even the part of his discussion concerning behavior and practical skill has failed to attract serious attention, and this is borne out by the fact that most writings on Evans merely reiterate his own words on the topic.
One writer who did try to go beyond restating Evans on the matter is John Campbell. Let it be clear that Campbell's critique of Evans on this score takes place in the context of addressing ‘Molyneux’s Question';Footnote 20 Campbell's concern is not squarely the topic of ‘I’-thought. Nonetheless, his discussion highlights a limit of the popular reading of Evans in that it too fails to recognize an important feature of Evans' thought concerning action.
Evans wrote two pieces on Molyneux's Question. In the earlier of the two, ‘Things Without the Mind’, Evans maintains that shape concepts gain their meaning in the context of a naive folk physics. In the later writing, entitled ‘Molyneux’s Question',Footnote 21 Evans seems to have changed his view on the issue. He now holds that shape concepts are the result of the unifying effect of the behavioral space on modal-specific information in egocentric coding: ‘…spatial concepts applicable upon the basis of one mode of perception must generalize to the other. There is only one behavioral space.’ (390)
Campbell is sympathetic to the earlier view but expresses serious doubts about the revised view. I'll focus on Campbell's examination of the later paper – actually, on just one particular point he makes about this second writing of Evans.Footnote 22 This is the suggestion that we can take Evans' notion of behavior (as in ‘behavioral space’, ‘behavioral plans’, ‘behavioral output’) as the equivalent of J. J. Gibson's notion of affordance.
… what are the implications of supposing that egocentric spatial content is actually constituted by its implications for behavior? The Natural way to implement that idea would be to suppose that identifying the egocentric location of an object is identifying something like an affordance of the object, in the sense of Gibson (1979).Footnote 23
Campbell's suggestion can be challenged, but for the purposes of this paper I'm going to accept it. Affordance would seem to be a good candidate for Evans' notion of behavior – and his notion of practical skill as well. By Evans' own lights, the content of perceptual information is non-conceptual in the sense of being in egocentric terms ('to the left', ‘to the right’, ‘above’, ‘in front of’, etc.); thus the behavioral outputs that feed on information of this type are best interpreted as a type of practical skills (e.g. homing in on an object of interest, stalking a prey, shaping one's hand for the gripping of an object) rather than mere reflex (e.g. eye-blinking, and the patellar reflex) or full-blown intentional action, which requires exercise of conceptual thought.
The problem, however, is whether affordance is all Evans talks about when it comes to action and behavior, and I think the answer is ‘No’. Talk of ’practical skill' or ‘affordance’ is always talk that proceeds against certain assumptions about the background of such talk. A normal adult human has practical skills in the sense of affordance, but this is also true of many non-human animals (think of the coyote that has learned to adapt to a 21st century urban environment). The exercise of practical skill can land the human in the court of law, but not when the coyote does so. In other words, ‘skills’ take on different significance when the background against which they are exercised and subsequently looked at is different. An organism like you and me is something capable of putting skills to use for a particular purpose; but when a coyote exercises its skills, all there is to it is just that, the exercise of an (adaptively advantageous) skill. The same thing can be said about ‘behavior’.
So what assumption or assumptions are operative in Evans when he talks about ‘practical skill’ and ‘behavior’ as these feature in his account of first-person thoughts? The answer is found in this much quoted passage:
we arrive at conscious perceptual experience when sensory input is not only connected to behavioural dispositions in the way I have been describing – perhaps in some phylogenetically more ancient part of the brain – but also serves as the input to a thinking, concept-applying, and reasoning system; so that the subject's thoughts, plans, and deliberations are also systematically dependent on the informational properties of the input. When there is such a farther link, we can say that the person, rather than just some part of his brain, receives and possesses the information.Footnote 24
So sensory input that carries behavioral implications can be input to not just one, but two very different systems. It can be input to a ‘thinking, concept-applying, and reasoning system’, and it can be input to an organism that lacks such a system but nonetheless has affordance – a system of inherited or learned skills in handling objects and navigating the environment. To Evans the difference is a crucial one. It is, in his own words, what separates ‘a person’ from ‘some part of his brain’. ‘Thoughts, plans, and deliberations’ are of persons, ‘behavioral dispositions’, on the other hand, pertain to only a sub-part of a person.
If you examine closely some of the best known passages where Evans takes himself to be addressing the ‘behavioral’ implications of perceptual information, you'll see that the distinction we've been looking at is also at work, and well-kept. Consider the passage where Evans quotes Charles Taylor with apparent approval:
… having heard the sound directionally, a person can immediately say to himself ‘It's coming from over there' (pointing with what is in fact his right hand), and may then reflect as an afterthought and that's the hand I write with.' As Charles Taylor writes: “Our perceptual field has an orientational structure, a foreground and a background, an up and down … This orientational structure marks our field as essentially that of an embodied agent. It is not just that the field's perspective centres on where I am bodily – this by itself doesn't show that I am essentially agent. But take the up-down directionality of the field. What is it based on? Up and down are not simply related to my body – up is not just where my head is and down where my feet are. For I can be lying down, or bending over, or upside down; and in all these cases ‘up’ in my field is not the direction of my head. Nor are up and down defined by certain paradigm objects in the field, such as the earth or sky: the earth can slope for instance … Rather, up and down are related to how one would move and act in the field”.Footnote 25
Where Taylor uses ‘the body’, ‘the head’, etc., Evans speaks of ‘some part of his brain’; where Taylor speaks of ‘embodied agent’, Evans talks of ‘persons’. Both writers share the view that the ‘orientational structure’ in the content of perceptual content – the ‘foreground and background’, ‘up and down’ – has different behavioral significance depending on whether it is a person or a mere sub-part of it that is at issue.
To sum up: Evans is clearly operating with a notion of action distinct from his use of the notion of behavior. While the latter fits naturally with the notion of information and is the focus of writers on Evans, the former plays a more critical role in his theory but has not received the attention it deserves.
3.3 ‘I Do Not Move Myself; I Myself Move'
The behavioral output of the ‘thinking, concept-applying, and reasoning system’ is what we call intentional action, and intentional action is part and parcel of Evans' theory. However, saying that in the current context is like drawing a blank check. There is a plethora of theories of intentional action many of which are squarely opposed to one another; and if we can't determine to which of the warring camps Evans belongs, simply insisting that he operates with a notion of intentional action is not very helpful. We need to know Evans' exact stance on the issue.
Here is a way to go about this. One of the major issues that divide philosophers of action is the issue concerning the nature of practical knowledge. By ‘practical knowledge’ I mean the agent's knowledge of what is it that she intends to be doing, or has done intentionally.Footnote 26 If we can find out where Evans stands, express or implied, on the issue, we'll have a pretty good chance of figuring out what conception of action is operative in his theory of ‘I’-thought. Needless to say this is an unorthodox way of grouping theories of action, but then we're not here to cut the terrain at its ‘natural joints’.
Viewed from this vantage point, theories of intentional action fall quite neatly into two large classes. The first class comprises theories which hold that practical knowledge is observational knowledge, acquired through the exercise of either an inner, or introspective faculty (e.g. ‘I intend/have an intention to do so and so, because I can introspectively see such and such.’), or the outer senses (‘I'm doing/have intentionally done so and so, because I can/could see/hear such and such.'). On theories of this type, the agent knows what she intends to do/to be doing, or has done intentionally, because she sees, or is in a position to see, the goings-on inside her, or around her. Let us adopt the name ‘RT’ (short for ‘receptive theories of practical knowledge’) for this type of theories. In contrast, the second type of theories – call them ‘NRT’ (short for ‘non-receptive theories’) – hold that an agent has knowledge of her intention and intentional action that she does because it is something that comes around as a result of her decision, not something she acquires with the help of observational or introspective capacities. Because intentions and actions are what an agent decides to have or do, not something she learns about through the exercise of her receptive cognitive faculties, the capacity to perform an intentional action is not a receptive faculty, and the knowledge the capacity delivers is not receptive knowledge.
There is a good amount of evidence, both textual and substantive, which supports reading Evans as operating with an NRT conception of action. A good place to start is his criticism of Thomas Nagel.Footnote 27 Evans diagnoses the problem with Nagel's claim about the unbridgeable gap between ‘the subjective’ and ‘the objective’ as ‘looking for the impact in the wrong place’. To Evans, ‘subjective’ or ‘objective’, the knowledge the thinker has of herself in both cases is mere contemplative knowledge, not practical knowledge; it is knowledge of oneself as an onlooker, not as an agent, or ‘a subject of thought and action’ as Evans puts it.Footnote 28 Nagel gets wrong about the nature of ‘I’-thoughts because his subjective vs. objective dichotomy merely operates in the confines of contemplative knowledge, which is but a particular species of observational knowledge (albeit acquired via a different route, through the use of introspection, intuition, the inner eye, or what have you).
Evans' criticism of Nagel is revealing. Unfortunately for our purposes, a mere acknowledgement of the existence of a link between ‘I’-thought and action is not enough. What we need is something – a claim, or a suggestion, made by Evans himself – as to the specific nature of the relation between the two. And this we find in the opening pages of the chapter on ‘I’-thoughts, particularly in the following passage:
… the Idea which one has of oneself involves the same kinds of elements as we discerned in the case of, say, ‘here' (6.3) … Of course, it must be recognized immediately that there are crucial differences…And an even more important difference lies in the fact that the essence of ‘I' is self-reference. This means that `I'-thoughts are thoughts in which a subject of thought and action is thinking about himself – i.e. about a subject of thought and action. It is true that I manifest self-conscious thought, like ‘here’-thought, in action; but I manifest it, not in knowing which object to act upon, but in acting. (I do not move myself; I myself move.). Equally, I do not merely have knowledge of myself, as I might have knowledge of a place: I have knowledge of myself as someone who has knowledge and who makes judgements, including those judgements I make about myself.Footnote 29
The importance of this passage for a proper assessment of Evans' theory of ‘I’-thought couldn't be overemphasized. The passage is immediately preceded by a footnote in which Evans announces his strategy going in the chapter: the focus here is going to be the information component, not the action component, because ‘[t]his element [the action component] in an account of ‘I’ is stressed in much recent work', whereas the information component hasn't received its fair share of attention.Footnote 30 The purpose of these terse remarks is apparently to give the gist of the author's view on an issue that is going to be shelved from this point on for the sake of the better pursuit of another. This would explain why what we have here is a few summary pages on the action component rather than a full scale treatment.
As unfortunate as the strategy is, Evans leaves us enough in these pages to piece together the conception of action he shelved for the sake of the exposition of the information component. Of the many remarks on the issue we find in this part of the VR I'm going to focus on three pithy statements, all found in the passage quoted earlier, ‘I manifest it [self-conscious thought], not in knowing which object to act upon, but in acting’, ‘I do not move myself, I myself move’, and ‘The essence of ‘I' is self-reference’.
The first statement – the most crucial of the three for my purposes – needs little exegesis, it contains two key ideas. The first is the idea that self-knowledge is not ‘know-which’ or discriminative knowledge, or roughly for the purposes at hand, knowledge I have of myself as the object upon which my action directs. The second is the idea that action is key to self-consciousness (‘I manifest self-conscious thought … in acting'). The first statement thus provides a strong reason for taking Evans to be operating with an NRT conception of practical knowledge, and it is also evidence that Evans finds a link between action and the phenomenon of self-consciousness.
Getting to the real meaning of the second statement will take a bit of unpacking. An RT theory of action is typically a causal theory.Footnote 31 Actions on these theories are a sub-class of the more general category of events. What sets actions apart from other types of events is that they're events with a particular mental genealogy. Practical knowledge on such theories is thus just causal knowledge – observational or introspective knowledge of what's going/went on inside the head of the subject leading up to the eventuality of a physical event. In this framework, ‘I did such and such’ is loose talk; the truth is, belief and desire, and whatever they in turn cause, do their job within the confines of the agent's skin, the resultant event is what it is because of this mental ancestry.
In contrast, NRT's typically hold that actions are sui generis; they are what we do, not what we – or rather, our mind as one link in a chain of causal events – cause to happen. Practical knowledge on this view is not knowledge of what causal events transpired leading up to the happening of an event out there; it is, as Anscombe puts it, ‘the cause of what it understands’.Footnote 32 I know for example that I am doing so-and-so and not something else, because the doing of it is what I intended, or had decided upon (and I've not changed my mind). In this framework, talk of an agent's ‘doing so and so’ is no loose talk, and it is no loose talk exactly because the tie between a decision to do a certain thing and the thing's being what it is is a tie by virtue of reason, not by law of nature.
Evans' rejection of the claim that practical knowledge is a type of discriminative knowledge and his emphatic statement, ‘I do not move myself; I myself move’, should be viewed against the background of this basic divide between the two types of theories of intentional action. In particular, the point of his second statement is, as I read him, to highlight two different approaches to actions which are expressive of self-consciousness. One can take actions which are expressive of self-consciousness to be actions one's mind qua belief-desire complex causes or makes happen, which is the approach a theory of ‘I’-thought incorporating this conception of action would take and which is what the EAIC reading takes Evans to be expounding. In contrast, a theory of ‘I’-thought based on the second approach cuts out the causal intermediary; it takes the thinker of ‘I’-thought (the ‘I’ in ‘I move’) and the agent of self-consciousness-expressing intention/action to be two aspects of the same thing. The superiority of the second approach is easy to see, for while the causal approach is capable of giving us an account of the object (e.g. the ‘myself’ in ‘I cut myself’) upon which our actions direct, it leaves the agent unaccounted for, and subsequently, is incapable of explaining why ‘I’-thoughts expressive of self-consciousness exhibit the three types of immunity to errors they do (see §3.1). The second theory – to which Evans is arguably committed in view of the two statements discussed – has no such problem.
There has been a growing consensus in contemporary philosophy, that the exercise of an agential capacity, and not that of a receptive faculty, is what gives us self-consciousness. Exactly because thoughts which express knowledge of the mental genealogy of an action and what it has caused to happen are not thoughts expressive of the exercise of such a capacity, they can't be thoughts the having of which manifests self-consciousness. This is the insight that inspired the Rödl-McDowell distinction between self as other and self as self (agent),Footnote 33 and Anscombe's distinction between practical knowledge, which she calls ‘the cause of what it understands’, and observational knowledge,Footnote 34 which has something other than practical reason itself as its cause. And I'm claiming that the same insight is what is behind Evans' pithy one liner, ‘I do not move myself, I myself move.’
Evans' third statement, the thesis that the essence of ‘I’ is self-reference, should be read as part of the same package with the thoughts expressed in the first two. To self-refer is to perform a speech act with the use of some bits of language. But this is not just any kind of event or happening; it is an intentional action, of the type one has practical knowledge of, immune to the three types of errors, and ‘manifesting self-conscious thought’. Furthermore, given Evans' well-known hostility toward Cartesianism, it's pretty straightforward that the agent of the speech act can only be an embodied being, ‘a subject of thought and action’,Footnote 35 and not a Cartesian ego, or a center for the processing of mere factual information. A final thing to note is Evans' dismissal of a view he attributes to Strawson and Geach – the idea that the essence of ‘I’ is its role in linguistic communication. To Evans, self-reference is reference to self as a subject of thought and action, and this role of ‘I’ is not lost even when it's not used for the purposes of communication. If self-conscious thoughts are thoughts which manifest in acting, as Evans claims, communication cannot be what is responsible for the having of self-conscious thoughts, for to claim that would be like claiming that communication is the reason for which we have the capacity to perform intentional action.
This view of self-reference avoids the pitfalls of the view which is the chief target of Anscombe's criticism in ‘The First Person’; it also enables Evans to steer clear of Cartesianism on the issue. The view also readily brings to mind another familiar name, John McDowell. In an article written a number of years after he assisted with the editing and the publication of the VR, McDowell delineates a view according to which self-consciousness is a phenomenon associated with bodily agency, or ‘physical intervention’ as he calls it. Some bit of the physical world – a physical aspect of me, say the vocal chord – becomes something for me to be only when physical intervention upon it is possible for me (e.g. I can make it vibrate through intention). Something is me only if bodily agency on my part is involved. This is why, says McDowell, the ‘A’ in the mouth of Anscombe's imaginary ‘A’-users is not the equivalent of our first-person pronoun ‘I’. The ‘A’ user's use of ‘A’ requires observation, in contrast to our ‘I’ in ‘I’-thoughts expressive of self-consciousness which does not. To quote McDowell:Footnote 36
When we try to think ourselves into the “A”-practice, a certain singled out bodily thing appears at best as an instrument for the will of a self that can only dubiously find a place in our imaginings, a self that recedes inward to the point of vanishing. With bodily agency in the picture, that bodily thing becomes something one can identify as oneself; now there is something plainly present in the world for oneself to be. And it is not clear that we can understand the specific agency exercised in referring … except by beginning with referring as an element in speaking, a particular mode of physical intervention in the world on the part of the bodily agent that one can identify as oneself. Surely this cannot be less fundamental, in our understanding of self-consciousness, than providing a distinction between how things are and how things are experienced as being.Footnote 37
Evans would have agreed. The first of his three statements would have made it clear that the use of ‘A’ in the mouth of a competent ‘A’-user presupposes ‘knowing which object to act upon’ – the speaker would have to have first looked at the letter written on the inside of her wrist, which is the bane of any suggested candidate for knowledge embodying self-consciousness. On grounds of his second statement and his staunch anti-Cartesianism, Evans would have agreed with McDowell on self-reference as a form of embodied agency, and on the latter's reason for dismissing even a cognitively generously enriched ‘A’-user as one capable of self-reference.Footnote 38
4. Evans With the Action Component
I began this paper with quotes from Evans stating his theory's having two components. We have by now a good grasp of both the nature and the content of Evans' ‘action component’. It is time we address the architectural issue brought about by our EWAC reading, ‘What’s the overall picture of the theory with the missing piece restored?'
At some suitable level of generality, Evans' main philosophical undertaking can be viewed as centered on answering one single question, ‘How does perception make thought possible?’ The question is not new; it's one that has exercised all the great thinkers of the empiricist tradition. Evans' chief contribution to the contemporary thinking of the question, all parties agree, is his cognitive science inspired revamping of the notion of perception. The notion of perception featured in traditional empiricist epistemology is the idea of an epistemic faculty whose job is to deliver to the mind mental building blocks – impressions and ideas, sense-data, and what have you – out of which the mind is supposedly to construct its version of reality. Following the ascending cognitive science of his day, Evans swept aside this dated conception, and replaced it with one that takes perception to be an information processing capacity whose rightful home is an organism that relies on it for survival – for navigating the environment, avoiding predators, and securing prey and mates.
This new, cognitive science-inspired notion of perception, when used as a component in an account of demonstrative thoughts, brings about at least two problems. The first I call ‘the coding issue’, and the second ‘content preservation’. Roughly: If perceptual information picked up by the sense modalities is content encoded in a non-conceptual type of representation (as Evans claims), and thoughts are content encoded in another, how is the transformation accomplished? This is the gist of the code-switching issue. The content preservation issue is related but different: if there are such things as information-based thoughts, there must be content continuity – some bit of information comes in at the sensory end, journeys all the way and ends up in a thought at the other end, and somehow manages to stay the same. For example, your ear picks up a sound as coming from ‘the left’, and at some point you end up entertaining a thought featuring the concept of ‘a noise coming from the direction of our camp site’. ‘Noise from the camp site' is not the same as ‘noise from the left’, but presumably the same representational content underlies both. How do we – or our brain – manages to pull this off?
Some parts of Evans' theory are clearly designed to handle the two issues (and other related issues perhaps). His notion of a practical skill, and the idea of a content-unifying ‘behavioral space’ are two such components.Footnote 39 The question that interests us here though is whether the action component is regarded by Evans himself as an element of this part of his theory.
There is little textual evidence which would suggest that this is how he views the job of the action component. On the contrary, given the amount of conceptual as well as empirical resources available to him, Evans has no need to view the action component this way, but ample reasons for rejecting it. The conceptual resource I'm talking about here is of course the distinction we already covered in the last section, between a concept of self based on, or expressive of discriminative knowledge, and exemplified by ‘My arm is broken’, and a concept of self as an embodied agent, as expressed by the ‘I' in ‘I myself move’ and the ‘self’ in ‘the essence of ‘I' is self-reference’, as discussed in the previous section. The empirical resource is the common observation that agents do usually stand in informational contact with, and benefit from having information (and unfortunately misinformation in rare cases) about their own bodies when they succeed in executing what they had intended to do.
Given what is available to him, there is little that could prevent Evans from having it both ways. Evans can claim that there are ‘I’-thoughts that are indeed a type of information-based thoughtsFootnote 40 which are a piece with ‘this’-thoughts and which his information component is designed to account for. At the same time, he can insist that to account for the genuine form of ‘I’-thoughts, that is, thoughts which are expressions of self-consciousness, something akin to his action component – in the way I have interpreted it – is required. In other words, there is a clear division of labor for the two components.
This, I suggest, should be regarded as the overall structure of Evans' theory of the first-person. Note how nicely the two components work in tandem in explaining what normally goes on with a cognitive agent capable of genuine first-person thoughts. Under normal circumstances, the intentions or intended actions are mostly if not always carried out with the agent's having the benefit of being in information contact with her body. For example, one decides to straighten one's leg, and one's leg – not someone else's, nor one which is one's own but a straight one – does end up getting straightened, thanks to one's having true information-based ‘I’-thoughts about the whereabouts of one's limbs. As familiar as this may be with most of us most of the time, the having of such benefit is not only subject to cognitive luck, it is also not the reason a use of ‘I’ counts as an expression of self-consciousness when it does. An ‘I’-thought is an expression of self-consciousness because it is the result of the exercise of an embodied agential capacity.
This reading of Evans obviates the criticism leveled at him by philosophers such as O'Brien and Stanley. It is true that there could be cases where one's information-based ‘I’-thoughts have their sources in other individuals; it might happen for example that I'm having the proprioception of a leg's being bent, when really this is due to my being hooked up to someone else's body without my knowledge. Cases such as these have been used to show the inadequacy of Evans' theory as an account of ‘I’-thought, since, the critics point out, our concept of first-person is not open to such mishaps. The point is a valid one, but is irrelevant as a criticism of Evans. With the help of the action component, Evans can easily answer his critics: per the action component, I am what I (intend to) do. If the information delivered by the proprioception I'm currently undergoing does not originate from a body I can identify as me through my ability to command it in acting, it is not of me the agent, but only of me the subject of my information-based ‘I’-thoughts.Footnote 41
5. Concluding Remarks
The widely accepted view of Evans' theory of first-person thought identifies the core of the theory to be Evans' notion of identification-free ways of gaining information about the object of such thought. We've grown so accustomed to this view that we are no longer able to see the crudeness of the Evans it portrays. But without the action component, and having only the information element to hang on to, Evans' theory looked incredibly naive and crude; it exhibited an attitude toward its problems that can only be characterized as ‘sublime indifference’.
But we do not have to accept this assessment of Evans. The popular interpretation is based on the neglect of a key component of Evans' theory. If what I've argued in this paper is on the right track, there is an alternative reading available – one which incorporates a credible line of thinking about the nature of intentional action and is capable of warding off the criticisms coming from the EAIC reading.Footnote 42
