'A Defense of First-Order Representationalist Theories of Mental-State Consciousness' by Robert W. Lurz

	Peter Carruthers' book Language, Thought and Consciousness may be purchased from Amazon.Com

A Defense of First-Order Representationalist Theories of Mental-State Consciousness

Robert W. Lurz
Indian River Community College
Ft. Pierce, Florida
U.S.A.

rlurz@ircc.net

PSYCHE, 6(1), February 2000
http://psyche.cs.monash.edu.au/v6/psyche-6-01-lurz.html

KEYWORDS: First-order and higher-order representationalism, mental-state consciousness, phenomenal consciousness, animal consciousness.

COMMENTARY ON: Carruthers, P. (1998). Natural theories of consciousness. European Journal of Philosophy, 6 (2), 203-222. Precis available at: (http://psyche.cs.monash.edu.au/v4/psyche-4-03-carruthers.html

ABSTRACT: Recently, Peter Carruthers has advanced the debate over first-order representationalist theories (FORs) and higher-order representationalist theories (HORs) of consciousness by offering two innovative arguments in support of dispositionalist HORs. In this article, I offer a limited defense of actualist FORs by showing that Carruthers' two arguments either beg the question against such accounts, equivocate on important concepts, or rest on suspect claims about our abilities to attribute phenomenal consciousness to animals. In addition, I argue that dispositionalist HORs face an apparent counterexample, one which Carruthers does not consider.

1. Introduction

There has been considerable debate recently over which kind of representationalist theory of mind -- higher-order representationalist (HOR) theories or first-order representationalist (FOR) theories -- offers the best explanation of mental-state consciousness.<1> Both theories attempt to explain the nature of mental-state consciousness in part by appealing to a mental state's representational properties. But HOR theories also appeal to a mental state's causal/functional relation to higher-order cognitive states -- that is, cognitive states whose contents are about the mental state in question; whereas, FOR theories appeal to a mental state's causal/functional relation to cognitive states of only the first-order -- that is, cognitive states whose contents are not about the mental state in question. Each theory, of course, has its defenders and its detractors. Recently, however, Peter Carruthers (1998a) has advanced the debate by offering two innovative arguments designed to show that HOR theories like his own are superior to FOR theories like Michael Tye's (1995), Fred Dretske's (1995), and Robert Kirk's (1994). My intention in this paper is to give a limited defense of FOR theories by replying to Carruthers' two main arguments against Tye's theory and by offering an argument of my own which reveals, I believe, a fundamental problem with Carruthers' HOR theory, a problem which Carruthers does not seem to consider.

2. Three Explananda

Before I begin my defense, some groundwork needs to be laid. It is generally agreed that an acceptable theory of mental-state consciousness must explain the following three features of conscious mental states:

The aware/unaware distinction: We are consciously aware of some of our mental states but not others. Consider the familiar case of the distracted driver who, as a result of her preoccupation with her thoughts about work, is completely unaware of what she is doing or seeing on the road yet succeeds in navigating around a double-parked vehicle.<2> The driver certainly sees the double-parked vehicle -- she would have crashed her car otherwise -- but she is not consciously aware of seeing it. Or consider the following case due to Ned Block (1995): while engaged in an intense conversation, you suddenly become aware that outside the window there is -- and has been for some time -- a pneumatic drill digging up the street. You realize that you have been hearing the drill for the entire time that you have been conversing, although you just recently became consciously aware of hearing it. An acceptable theory of mental-state consciousness must explain the difference between being consciously aware of one's mental states and not being consciously aware of them.<3>

The phenomenal/nonphenomenal distinction: Some mental states have phenomenal properties -- that is, the property of there being something that it is like to have -- and others do not. There is something that it is like, for instance, to have a sharp pain in your back or to taste a fine claret; but there may be nothing that it is like -- no distinctive "feel" that is -- to think about Fermat's last conjecture or to remember your twenty-first birthday. It is true, of course, that your thoughts about Fermat's last conjecture may cause you to feel perplexed, or that your memory of your twenty-first birthday may cause you to feel nostalgic, and these feelings surely have phenomenal properties; but your thoughts and memories are, in and of themselves, phenomenally empty.

The phenomenal/nonphenomenal distinction also applies to perceptual states in some rare cases. The best example of this is blindsight, an unusual visual deficit in humans and animals that is caused by brain damage in the striate cortex. Human blindsight subjects sincerely declare that they have no visual sensation in certain "blind" areas (scotoma) of their visual field; however, if they are asked to guess at the position of a light source, or the orientation of a line, or even to reach out and grasp an object placed in their scotoma, they generally succeed at accomplishing these tasks.<4> Blindsight subjects, then, seem to have perceptual states of which there is nothing that it is like for them to have. An acceptable theory of mental-state consciousness must explain why some mental states have phenomenal properties and others do not.

The phenomenal character distinction: What it is like to experience a pain in one's back is different from what it is like to experience a tickle in one's back; what it is like to see something as red is different from what it is like to see something as green. Although, all of these experiences have phenomenal properties, each has its own determinate phenomenal property or character. One's pain, for example, has its own unique phenomenal character that is different from the phenomenal character that one's tickle has; and one's visual experience of seeing something red has it own unique phenomenal character that is different from the phenomenal character that one's visual experience of seeing something green has. An acceptable theory of mental-state consciousness needs to explain why some experiences have the particular phenomenal characters that they have.

3. Tye's FOR Theory

According to Tye's (1995) FOR theory, the phenomenal character of an experience is one and the same as its phenomenal content -- that is, content that is abstract, nonconceptual, and appropriately poised to make a direct impact on a subject's belief-forming mechanisms (pp. 143-144). This theory explains the phenomenal character distinction by showing that differences in phenomenal character (e.g., the phenomenal difference between seeing something as red and seeing something as green) are simply differences in phenomenal contents (e.g., the differences between one's visual experience representing something as red and one's visual experience representing something as green). The theory also has the resources to explain the phenomenal/nonphenomenal distinction: propositional attitudes and the nonphenomenal perceptual states of blindsight subjects do not have phenomenal properties because they either do not have phenomenal contents (in the case of propositional attitudes) or they do not make a direct impact on a subject's belief-forming mechanisms (in the case of blindsight states). Finally, Tye's theory explains the aware/unaware distinction by appealing to activated higher-order thoughts. The explanation, roughly, is that mental states of which we are (not) consciously aware are those about which we (do not) have an activated higher-order thought. Although this appeal to higher-order thought makes Tye's account a nonunitary FOR theory, it does not prevent his account from being in competition with Carruthers'; for the latter theory claims that only a unitary account in terms of (potential) higher-order thoughts can explain the explananda above.<5>

4. Worldly-Subjectivity/Mental-State-Subjectivity Argument

Carruthers offers two main arguments that are designed to show that FOR theories like Tye's are explanatorily inferior to HOR theories like his own. The first of these arguments rests on the distinction that Carruthers draws between what the world is like for a subject of experience (i.e., worldly-subjectivity) and what the experience of the world is like for the subject of experience (i.e., mental-state-subjectivity or what I call "phenomenal character"). Carruthers maintains that FOR theories can explain worldly-subjectivity, but they cannot explain mental-state-subjectivity (i.e., phenomenal character); HOR theories, however, can explain both. So, Carruthers concludes, FOR theories fail to account for an important feature of mental-state consciousness which HOR theories do not.

This argument has one main problem: it rests on a distinction that is either harmless or question-begging. To see this, consider that Carruthers' worldly-subjectivity/mental-state subjectivity distinction can be taken as stating either (i) that there is a difference in the concepts expressed by "worldly-subjectivity" and "mental-state-subjectivity" or (ii) that there is a difference in the properties picked out by "worldly-subjectivity" and "mental-state-subjectivity." If Carruthers takes his distinction as stating (i), then the distinction and the argument of which it is a premise are harmless; for Tye, like Carruthers, is not in the business of giving a conceptual analysis of phenomenal character, but of giving a substantive theory of its nature. Phenomenal character, according to Tye, is phenomenal content in the same way that water is H20 and heat is mean molecular motion (p. 184). Furthermore, phenomenal content just is the content of experience; and the content of an experience just is the way the world is presented to the subject in experience; and the way the world is presented to a subject in experience just is worldly-subjectivity. By the transitivity of identity, then, Tye's view can be expressed as stating that the property of mental-state subjectivity (i.e., the property of phenomenal character) just is the property of worldly-subjectivity (i.e., the property of phenomenal content). So, if Carruthers takes his distinction as stating (ii), then he clearly begs the question against Tye's account, since Tye's account denies that there is such a distinction.

5. Aware/Unaware Argument

Carruthers' second argument attempts to show that Tye's FOR theory faces three intractable problems in accounting for the aware/unaware distinction. These three problems I call "the counter-intuition problem", "the incoherency problem", and "the dilemma problem". Since I have already replied to Carruthers' dilemma problem in Lurz (forthcoming), I shall examine only the first two problems here.

5.1 The Counter-Intuition Problem & The Animal Case

As we noted above, Tye's account is nonunitary, and, as such, it entails that a subject may have an experience with phenomenal properties (as a result of the experience being poised to make a direct impact on the cognitive system) even though he or she is not consciously aware of having the experience (as a result of failing to have a higher-order thought about the experience). Carruthers finds this possibility "highly counter-intuitive" and suggests that it ought to be resisted by any acceptable theory of mental-state consciousness.

It should come as no surprise that defenders of FOR theories, like myself, do not find this possibility counter-intuitive at all; rather, we find it quite plausible. Its plausibility is derived from four main sources: ordinary-language cases, memory cases, physiological cases, and animal cases. However, given the limitations on the nature of this paper and Carruthers' unique view on animal experiences, I shall discuss only the latter case.<6>

It is highly likely that animals have experiences: they see, hear, taste, smell, and feel pain. It is also highly likely that there is something that it is like for them to have experiences. However, it is highly unlikely, as Carruthers points out, that animals are capable of forming higher-order thoughts about their own experiences.<7> The experiences of animals, then, strongly suggests that phenomenal states can exist without higher-order thoughts.

Carruthers claims, however, that we have no grounds for attributing phenomenal states to animals and that our confidence in thinking that animals have such states can be easily explained away. Let's examine this second claim first. Carruthers maintains that the reason we think that animals have phenomenal states is that we naturally "project" our own phenomenal states onto them in the process of attributing experiences to them:

[W]hen we ascribe an experience to the cat we quite naturally (almost habitually) try to form a first-person representation of its content, trying to imagine what it might be like "from the inside." But when we do this what we do, of course, is imagine a conscious experience -- what we do, in effect, is represent one of our own experiences, which will bring its distinctive phenomenology with it. (Carruthers, 1998b)

The idea here seems to be that whenever we attribute an experience to an animal, we do so on the basis of an imagined experience of our own, which, of course, we imagine as having phenomenal properties; and, as a result of what we imagine, we come to think that the animal has an experience with phenomenal properties too.

I am not convinced, however, that this is how things work. For instance, I highly doubt that I have ever experienced anything remotely like what a bat experiences as it uses echolocation to navigate through the night sky in search of bugs. I realize that it would be futile for me to try to imagine "from the inside" what it is like for the bat to do these things; for I know that anything that I could imagine would be inaccurate. So I do not try to imagine the bat's experiences. Nevertheless, this failure of mine does not stop me from strongly believing that there is something that it is like for the bat. Or consider a dog that smells the scent of another animal on the ground. Again, I highly doubt that I have ever experienced anything like what the dog experiences when it smells the scent of another animal, and so I do not try to imagine what it is like for it. Nevertheless, I strongly believe that there is something that it is like. Carruthers' "projectionist" account, however, would seem to predict that I would not have these strong convictions in the absence of the mental imaginations. So Carruthers' account does not appear to give an adequate explanation for why we intuitively believe that animals have phenomenal states.

One might think that what explains our strong conviction that animals have phenomenal states is simply that we have good reasons to believe that they do. Carruthers, however, claims that we do not have any such reasons, that there is nothing that animals have or can do that entitles us to attribute to them phenomenal states as opposed to nonphenomenal states. But this is surely wrong. It is well known that the brains and nervous systems of many animals are very similar in structure and function to those of adult humans; and since the structure and function of human brains and nervous systems support phenomenal states, it is plausible to suppose that the brains and nervous systems of some animals support phenomenal states too. Moreover, many animals behave in ways that entitle us to attribute phenomenal states as opposed to nonphenomenal states to them. When my dog runs across the lawn in hot pursuit of the neighbor's cat, for instance, it is implausible to account for his behavior in terms of his having a nonphenomenal visual experience. For such experiences, as we saw in the case of blindsight above, do not usually elicit purposeful and intelligent behavior. My dog's behavior, however, is quite purposeful and intelligent: it is aimed at catching the neighbor's cat, and it is appropriately modified in light of perceived changes in the environment. We seem to have, then, a good reason to believe that there is something that it is like for my dog to see the cat in the yard. Phenomenal states generally cause purposeful and intelligent behavior whereas nonphenomenal states do not, as evidenced by the difference in the behaviors of blindsight subjects and normal-sighted subjects. Since many animals engage in purposeful and intelligent behaviors upon experiencing an object in the world, we have some reason to believe that their experiences are phenomenal as opposed to nonphenomenal.

In reply, Carruthers might maintain that although my dog's behavior is not explainable in terms of nonphenomenal states akin to those of ordinary blindsight subjects, it is explainable in terms of nonphenomenal states akin to those of super-duper-blindsight subjects".<8> The perceptual states of super-duper-blindsight dogs, for example, are functionally identical with those of normal-sighted dogs, except that they lack phenomenal properties. Super-duper-blindsight states have the same sort of direct impact on the cognitive system of super-duper-blindsight dogs as normal perceptual states have on the cognitive system of normal-sighted dogs. Furthermore, the information contained in the nonphenomenal states of super-duper-blindsight dogs is just as rich as the information contained in the perceptual states of normal-sighted dogs. As a result, super-duper-blindsight dogs can make the same visual discriminations and perceptual judgments as normal-sighted dogs. There is, then, no functional or behavioral difference between normal-sighted dogs and super-duper-blindsight dogs. But since the latter type of dog lacks phenomenal states, there is no need to explain either type of dog's behavior in terms of phenomenal states -- we can simply appeal to super-duper-blindsight states. So, the argument goes, there is no reason to attribute phenomenal states to dogs or any other animal: we can simply appeal to super-duper-nonphenomenal states.

The problem with this argument, however, is that it begs the question against Tye's FOR theory. For it assumes that super-duper-blindsight states (or other super-duper-nonphenomenal perceptual states) are possible, and Tye's FOR theory denies this. Super-duper-blindsight states are perceptual states that are poised to make a direct impact on a subject's cognitive system but lack phenomenal properties. But Tye's account claims that perceptual states with phenomenal properties just are perceptual states that are poised to make a direct impact on a subject's cognitive system. According to Tye's account, then, there are no super-duper-blindsight states (or any other super-duper-nonphenomenal perceptual states); for any states that functioned like a super-duper-blindsight states would not be "blind" -- they would have phenomenal properties. So the above reply cannot be used against FOR theories like Tye's to show that we have no reason to attribute phenomenal states to animals. However, since we do have reasons to make such attributions, it is intuitively plausible to suppose that there can be phenomenal states without higher-order thoughts about them.<9>

5.2 The Incoherency Problem

Carruthers maintains that Tye's theory is not only counter-intuitive, but incoherent. His argument runs as follows: according to Tye's view, there can be something that it is like for a subject to have an experience even though the subject is not having a (actual or potential) higher-order thought about the experience; but for there to be something that it is like for a subject to have an experience, the experience "must be one which is available to the subject of that experience -- and that means being a target (actual or potential) of a suitable [higher-order thought]." So Tye's view seems to entail that there can be something that it is like for a subject to have an experience even though there is no experience available to the subject about which there could be something that it is like, and this certainly seems incoherent.

As elegant as this argument is, it does not work: it commits a fallacy of equivocation. There are two distinguishable senses of the phrase "available to the subject." To say that an experience is available to the subject, for example, may mean that the experience is available to the subject's higher-order thoughts, or it may mean that the experience is available to the subject's first-order thoughts. Furthermore, the reason that there is this distinction is that there is a distinction between two different notions of a subject. A subject (in this context) may be something that has a mind -- that is, something that has first-order mental states which guide its actions -- or a subject may be something that has a self -- that is, something that has the ability to be consciously aware of its own mental states and to unify and organize them as a result.<10> That is, a subject may be a creature with first-order mental states, or a subject may be a creature with higher-order mental states. Most animals have minds but not selves; and, hence, they are subjects in the first-order sense but not in the higher-order sense. Normal adult human beings, on the other hand, are subjects in both senses.

With this distinction in mind, let's return to the main premise in Carruthers' argument -- that an experience must be available to a subject in order for there to be something that it is like for the subject to have the experience. We are now able to see that this premise can be given either a first-order reading or a higher-order reading, depending on one's reading of the phrase "available to the subject". If the premise is given the first-order reading, then the conclusion that follows is simply that an experience must be available to a subject's first-order mental states in order for there to be something that it is like for the subject to have the experience, and this is perfectly in line with what FOR theories claim. However, if the premise is given the higher-order reading, then FOR theorists will naturally object; for on this reading, the premise simply denies that FOR theories are correct -- that is, it simply asserts that an experience must be available to a subject's higher-order mental states in order for there to be something that it is like for the subject to have the experience. Either way, Carruthers argument does not succeed in showing that FOR theories like Tye's are incoherent.

6. A Problem For Carruthers' HOR Theory

As I briefly mentioned above (footnote 5), Carruthers' HOR theory offers a dispositionalist account of the aware/unaware distinction: a mental state of which we are (not) consciously aware is one which is (not) available to our higher-order thoughts.<11> So the distracted driver and the blindsight subjects, on this account, are not consciously aware of their visual experiences precisely because their experiences are not available to their higher-order thoughts.

The problem with this account, however, is that there appear to be cases in which a subject's mental state is available to higher-order thoughts -- in one clear sense of this word -- even without conscious awareness of having the mental state. The sense of "available" that I have in mind here is roughly this: a subject's mental state is available to her or his higher-order thoughts just in case there is nothing in the mind or brain that would prevent formation of a higher-order thought about the experience. This sense of "available" offers a clear explanation for why the experiences of the distracted driver and the blindsight subjects are (as Carruthers maintains) unavailable to these subjects' higher-order thoughts. The explanation is simply that in each of these cases, there is something in the subject's mind or brain that would have prevented formation of a higher-order thought about the experience had that been attempted. In the distracted-driver case, there is the driver's preoccupation with her thoughts about work; and in the blindsight case, there is the damage to the subject's striate cortex.

So on the above sense of "available", the experiences of the distracted driver and the blindsight subjects are not available to higher-order thoughts. But there appear to be cases in which an experience is available (in the above sense) to a subject's higher-order thoughts but the subject is not consciously aware of having the experience. For example, consider a case that is very similar to the distracted-driver case described above. In this case, like in the original, the driver is unaware of seeing the double-parked vehicle around which she successfully navigates. However, the reason that she is unaware of seeing the vehicle is not that she is preoccupied with her thoughts about work or that she suffers from some sort of brain damage; rather, the reason is that she has driven this particular route and has navigated around this particular double-parked vehicle many times. As a result, she is simply "going through the motions" and no longer pays any attention to what she is seeing or doing on the road. But now we seem to have a case where a subject is unaware of her visual experience even though the experience is available to her higher-order thoughts. For, in this case, there is nothing in either the driver's mind or in her brain that would have prevented her from forming a higher-order thought about her visual experience if she had attempted to do so -- there is no distraction, no preoccupation, no neurological pathology -- and yet she is not consciously aware of her visual experience. In order for Carruthers' dispositionalist account to be acceptable, a distinction must be drawn between a mental state's being available to a subject's higher-order thoughts, although the subject is not availing himself of the mental state by forming a higher-order thought about it, and a mental state's being unavailable to a subject's higher-order thought. The above sense of "available" offers a straightforward way to draw this distinction. However, as we have just seen, this way of drawing the distinction leads to a counter-example to Carruthers' dispositionalist account.<12>

An actualist account of the aware/unaware distinction does not have the above problem, however. For on the actualist account, the distracted driver is not consciously aware of her visual experience in both the original and the modified cases because in both cases she lacks an activated higher-order thought about her experience. Actualist FOR theories, then, are free from the above counter-example and free from the problems that Carruthers claims they have. Of course, this freedom is not enough to prove that actualist FOR theories are preferable to HOR theories like Carruthers', but it does offer some presumptive evidence in their favor.

Notes

<1>Mental-state consciousness is the type of consciousness that mental states possess. It is to be contrasted with creature-consciousness, a type of consciousness that creatures possess when they are awake and/or aware of items in their environment.

<2>This case is due to Armstrong (1968).

<3>It should be noted that the aware/unaware distinction applies to propositional attitudes as well (see Carruthers, 1996, pp. 137-138).

<4>It should be noted that blindsight subjects succeed in these tasks only when they are given a very limited number of options (usually two) from which to guess.

<5>For a rough characterization of Carruthers' HOR account, simply replace "belief-forming mechanisms" and "activated higher-order thoughts" in my description of Tye's account with "higher-order thoughts" and "potential higher-order thoughts" respectively. The result, of course, is not a precise rendering of Carruthers' account, but it does offer a good sketch of the substantial differences and similarities between the two theories.

<6>For use of the ordinary-language case see Tye (1995, p. 6); for use of the memory case see Tye (1995, p. 6) and Martin (1992, p. 750); and for use of the physiological case see Dretske (1993).

<7>I argue in Lurz (forthcoming) that there is no good reason to think that animals cannot have certain kinds of higher-order thoughts about their experiences. However, these higher-order thoughts are certainly not possessed by very many animals. Which animals possess them will only be answered by further empirical investigation.

<8>I borrow this terminology and idea from Block (1995), although he uses them to show something different from what I use them to show here.

<9>This intuitive plausibility also renders Carruthers' explanation of the phenomenal/nonphenomenal distinction implausible. For Carruthers' explanation assumes that mental states with phenomenal properties have these properties in virtue of the causal/functional role they play among external stimuli and higher-order thoughts. But the phenomenal states of animals, it is reasonable to believe, do not play this causal/functional role.

<10>McGinn (1982, p. 104) makes a similar distinction.

<11>Carruthers also expresses his account in terms of a special short-term memory store, C, whose function is to hold those mental states that are available to higher-order thoughts. According to this way of expressing it, Carruthers' account states that a mental state of which we are (not) consciously aware is one which is (not) in C.

<12>David Rosenthal (1993) offers an objection that is similar to mine here. He maintains that a conscious experience cannot be a disposition of any kind since conscious experiences are actual events and dispositions are potential events. Carruthers (1996) replies by maintaining that on his dispositionalist account "there is something categorical taking place in me whenever I have a conscious experience ... since the perceptual information is actually in the short-term memory store C, which is defined by its relation to higher-order thought" (p. 172). But this reply begs the question. For as we noted (see footnote 11), the "relation" to higher-order thought that defines C is the dispositional relation of availability to higher-order thoughts. So to say that an experience is in C is just to say that it is available to higher-order thoughts. Rosenthal's point, then, stands: one cannot define conscious experience in terms of a disposition of any kind, including the disposition of availability to higher-order thoughts. I think that Rosenthal's point here is right, and I take my modified distracted-driver case to give independent support for it.

References

Armstrong, D. (1968). A materialist theory of mind. London: Routledge.

Block, N. (1995). A confusion about a function of consciousness. Behavioral and Brain Sciences, 18, 227-247.

Carruthers, P. (1998a). Natural theories of consciousness. European Journal of Philosophy, 6(2), 203-222.

Carruthers, P. (1998b). Animal Subjectivity. PSYCHE, 4(3). http://psyche.cs.monash.edu.au/v4/psyche-4-03-carruthers.html.

Carruthers, P. (1996). Language, thought and consciousness. Cambridge: Cambridge University Press.

Dretske, F. (1995). Naturalizing the mind. Cambridge, MA: MIT Press.

Dretske, F. (1993). Conscious experience. Mind, 102, 268-283.

Kirk, R. (1994). Raw feeling. Oxford: Clarendon.

Lurz, R. (forthcoming). Animal consciousness. Journal of Philosophical Research.

Martin, M.G.F. (1992). Perception, concepts, and memory. The Philosophical Review, 4, 745-763.

McGinn, C. (1982). The character of mind. Oxford: Oxford University Press.

Rosenthal, D. (1993). Thinking that one thinks. In M. Davies and G. Humphreys (Eds.), Consciousness. (pp.207-211). Oxford: Blackwell.

Tye, M. (1995). Ten problems of consciousness. Cambridge, MA: MIT Press.