The reach of phenomenal consciousness

Word count 10, 350 (includes references but not footnotes) 11,823 (includes footnotes)

Abstract How far does phenomenal consciousness reach? Answers to this question cluster into two groups. Phenomenal conservatives argue for a restrictive view of phenomenal consciousness, according to which only a narrow range of mental states are phenomenally conscious. Phenomenal liberals have a more permissive view of the range of phenomenal consciousness, holding that most—if not all—types of conscious states enjoy a proprietary phenomenal character. This paper both clarifies the debate between these two positions and examines a number of potential solutions to it. My conclusion is a negative one: the debate resists dissolution. This result is not insignificant, for it calls into question the widespread assumption that we have an adequate grip on the notion of phenomenal consciousness.

2

1. Introduction If the students in your Philosophy of Mind classes are anything like mine, they are likely to be puzzled by the term ‘phenomenal consciousness’ when they first encounter it. What, they might ask, does it mean for a mental state to be phenomenally conscious? Until recently, I assumed that this question had a straightforward answer: phenomenal states are states that there is something it is like for the subject to be in.1 Accounting for ‘what it’s likeness’ might be hard—indeed, it might be very hard—but getting a pre-theoretical grip on the notion sufficient for theory-building is not. So, at any rate, I once thought. I am no longer as sanguine about these matters as I once was. My reasons for worry derive from debates about the reach of phenomenal consciousness. In explicating the notion of what it’s likeness theorists invariably invoke examples. There is, it is said, something that it is like to smell freshly brewed coffee, to see bright sunflowers in bloom, and to hear Big Ben chime. Such examples are fine as far as they go, but they do not go very far. One might want to know whether there is—or could be—something that it is like to (say) understand a proof, wonder whether the Democrats will regain the White House, and decide to go to Ethiopia for a holiday. This is where the trouble begins. Some theorists—I will call them ‘phenomenal conservatives’—deny that there is something distinctive that it is like to be in these mental states. Conservatives hold that phenomenality has a relatively minimal presence in mentality; it extends only as far as our perceptual and affective engagement with the world. Phenomenal liberals, by contrast, hold that

1

Although some have suggested that the ‘what it is like’ locution might function only as a means of identifying

phenomenal consciousness (see Pereboom in press), I take it to provide a definition or explication of the notion of phenomenal consciousness.

3

phenomenality infuses cognition. Unlike conservatives, liberals allow that there is something it is like to understand a proof, wonder whether the Democrats will regain the White House, and decide to go on holiday in Ethiopia. It is hard to over-emphasize the importance of this debate. A theorist’s take on the reach of phenomenal consciousness has the potential to bear on her conception of the functions of phenomenality, the relationship between phenomenality and intentionality, introspection, phenomenal concepts, and perhaps even the deep nature of phenomenality itself. However, my primary concern here is not with these issues but with the implications that this debate has for our claim to know what it is that we are taking about when we talk of phenomenal consciousness. Arguably, the mere existence of this debate undermines our claim to have a determinate—or at least shared—conception of what it’s likeness.

2. Clarifying the contrast Like their political counterparts, phenomenal conservatism and liberalism are clusters of loosely related approaches to consciousness rather than clearly defined positions.2 Michael Tye represents a fairly standard position within the conservative cluster when he writes: Should we include any mental states that are not feelings and experiences [on the list of phenomenally conscious states]? Consider my desire to eat ice cream. Is there not 2

Those sympathetic to some form of liberalism include Chalmers (1996), Flanagen (1992), Goldman (1993),

Horgan & Tienson (2002), Graham et al (2007), Kriegel (2002), Loar (2003), Lycan (2008), Pitt (2004), Siewert (1998), and Strawson (1994). Those sympathetic to some form of conservatism include Braddon-Mitchell & Jackson (2007), Byrne (2001), Clark (2000), Dretske (1995), Jackendoff (1987), Jacob (1998), Langsam (2000), Levine (2001), Lormand (1996), Prinz (2007), Robinson (2005), Shoemaker (1996) and Tye (1995; 2000). Kriegel’s (2007b) terms ‘phenomenological deflationism’ and ‘phenomenological inflationism’ correspond roughly to my distinction between ‘phenomenal conservatism’ and ‘phenomenal liberalism’.

4

something it is like for me to have this desire? If so, is this state not phenomenally conscious? And what about the beliefs that I am a very fine fellow? Or the memory that September 2 is the date on which I first fell in love? … It seems to me not implausible to deal with these cases by arguing that insofar as there is any phenomenal or immediately experienced felt quality to the above states, this is due to their being accompanied by sensations or images or feelings that are the real bearers of the phenomenal character. (Tye 1995: 4). Norton Nelkin also excludes thoughts from the realm of the phenomenal: There are propositional attitudes, and we are sometimes noninferentially conscious about our attitudinal states. But such consciousness does not feel like anything. A propositional attitude and consciousness about that attitude have no phenomenological properties. The mistake is to think they must. There are different sorts of states we call “conscious states,” and the most important sorts are not like having sensations. (Nelkin 1989: 430) More recently, Peter Carruthers has claimed that …thoughts aren’t phenomenally conscious per se. Our thoughts aren’t like anything, in the relevant sense, except to the extent that they might be associated with visual or other images or emotional feelings, which will be phenomenally conscious by virtue of their quasi-sensory status. (Carruthers 2005: 138-9) What unites these theorists is a shared commitment to the claim that phenomenal consciousness is restricted to the sensory, affective and perceptual domains—it does not permeate thought as such. This view is rejected by liberals, such as Galen Strawson:

5

[T]he experience of seeing red and the experience of now seeming to understand this very sentence, and of thinking that nobody could have had different parents … all fall into the vast category of experiential episodes that have a certain qualitative character for those who have them as they have them. (Strawson: 1994: 194). Charles Siewert adopts a similar position: …generally, as we think—whether we are speaking in complete sentences, or fragments, or speaking barely or not at all, silently or aloud—the phenomenal character of our noniconic thought is in continual modulation, which cannot be identified simply with changes in the phenomenal character of either vision or visualization, hearing or auralization, etc. (Siewert 1998: 282; emphasis suppressed) David Pitt also defends a version of liberalism: In addition to arguing that there is something it is like to think a conscious thought, I shall also argue that what it is like to think a conscious thought is distinct from what it is like to be in any other kind of conscious mental state, and that what it is like to think the conscious thought that p is distinct from what it is like to think any other conscious thought… (Pitt 2004: 2) And, finally, Terry Horgan and John Tienson: Quine notwithstanding, it seems plainly false—and false for phenomenological reasons— that there is indeterminacy as to whether one is having a thought that rabbits have tails or whether one is instead having a thought that (say) collections of undetached rabbit parts have tail-subsets. It is false because there is something that it is like to have the occurrent thought that rabbits have tails, and what it is like is different from what it would be like to

6

have the occurrent thought that collections of undetached rabbit parts have tail subsets. (Horgan and Tienson 2002: 522) So, we have a spectrum of views about how far the reach phenomenality extends. Tye occupies a position at (or near) the conservative end of the spectrum. He claims that the features involved in the phenomenal character of visual experience are restricted to properties like “being an edge, being a corner, being square, being red29.” (Tye 1995: 141) On his view, high-level perception—seeing the animal under one’s bed as a tiger; hearing an instrument as a cello; smelling the odour emanating from the fridge as rotten aubergine—does not involve distinctive phenomenal properties. Pitt occupies a position at (or near) the liberal end of the spectrum. On his view, not only are there distinct phenomenal states associated with the various propositional attitudes—judging, wondering, deciding, intending, and so on—but there are also distinct phenomenal states associated with particular propositional states: what it is like to wonder whether today is Monday is distinct from what it is like to wonder whether it is raining in Dubrovnik.3 Conservatives need not be quite as conservative as Tye nor need liberals be as liberal as Pitt. Falling between these ends of the spectrum lie a number of more moderate accounts of the reach of phenomenal consciousness. Proponents of what we might call ‘liberal conservatism’ hold that phenomenality infuses both low-level and high-level perception, even as they deny that it characterizes propositional attitudes. And proponents of what we might call ‘liberal 3

I have assumed here that cognitive states can be understood as relations to propositions as opposed (say) to

objects or states of affairs, but nothing of substance turns on this assumption. I have also presented the debate between conservatives and liberals as a debate about the relationship between intentional and phenomenal states, but nothing much turns on this and the debate could equally be posed in terms of the relationship between phenomenal and intentional events or properties.

7

conservatism’ hold that although the various propositional attitudes possess phenomenal characters particular propositional states do not. According to the conservative liberal, there is something it is like to wonder whether interest rates will go up, but there is no phenomenal difference between wondering whether interest rates will go up and wondering whether a friend’s tenure case will prove successful. The stability of these intermediate positions is something of an open question. For one thing, one might argue that high-level perceptual phenomenality stands and falls with cognitive phenomenality. Arguably, one of the central issues dividing conservatives from liberals is the question of whether phenomenality infuses the conceptual, and on this point those who recognize the phenomenal character of high-level perception may find themselves allied with those who recognize the phenomenal character of propositional states. Each of these theorists, however, might find themselves opposed to the conservative liberal, who holds that phenomenal character attaches to propositional attitudes but not states. I will leave for another occasion the task of sorting out just how much stability there is in the space between the conservative and liberal ends of this spectrum, and turn instead to the question of how the contrast between conservatism and liberalism—as rough and ready as it is—should be construed. We can begin with the question of what it means for a state to have phenomenal character. In what sense do conservatives deny, but liberals assert, that cognitive states have phenomenal characters? To fix ideas, let us suppose that propositional state C has phenomenal character P. Firstly, the liberal should not be committed to the view that any instance of C must be accompanied by an instance of P. Although many liberals are sympathetic to the thought that

8

the relationship between cognitive states and their phenomenal characters is essential, we skew the debate if we saddle liberals with this view. We can see why by noting that conservatives ought not be saddled with the view that perceptual states possess their phenomenal character essentially. All conservatives recognize colour phenomenology, but not all conservatives hold that the phenomenal character of colour experiences is essential to them. Some conservatives allow for the possibility of ‘absent qualia’ and hold that it is (broadly) possible for a creature to perceptually represent an object as yellow without enjoying any phenomenal state; and some conservatives recognize the possibility of ‘inverted qualia’, according to which it is (broadly) possible that the phenomenal character associated with my experience of sunflowers, yolks and yellow crayons could be associated with your experience of broccoli, grass and green crayons (and vice-versa). But if conservatives, qua conservatives, need not be committed to the view that the phenomenal character of experiences of colour is essential to them, then the liberal, qua liberal, need not be committed to the view that the phenomenal character of cognitive states is essential or necessary to them. An alternative conception of what it is for C to possess phenomenal character is for it to be accompanied by some phenomenal state or other. But this conception of liberalism is too weak, for conservatives can—and typically do—hold that conscious thought enjoys some kind of phenomenal character. Thinking about the weather in Paris might be accompanied by images of Notre Dame encased in bright sunlight; wondering whether the Democrats will win the Presidency might involve images of Obama standing on the White House lawn; and deciding to go to Ethiopia for a holiday might bring with it feelings of anticipation. So, how should we understand the contrast between conservatives and liberals, given that

9

conservatives can grant that cognitive states have some form of phenomenal character and liberals need not hold that the phenomenal character of cognitive states is individuative? It seems to me that we can best capture the contrast between conservatives and liberals by invoking a relation of direct association.4 Consider colour experience. Despite disagreement about how the phenomenal character of experiences of red is related to their intentional content, all hands are agreed in holding that the relationship between them is direct, in the sense that it is not mediated by any other intentional state. It is experiences of red themselves that have a ‘what it’s likeness’—their phenomenal character is proprietary. And we can say this without denying that (in principle) such experiences might fail to have this phenomenal character, either because they fail to possess any phenomenal character (as in absent qualia scenarios) or because they possess some other phenomenal character (as in inverted qualia scenarios). Let us say that an intentional state that has proprietary phenomenal character in this way has phenomenal content.5 We this in mind, we can capture the contrast between liberals and conservatives by saying that liberals assert that cognitive states possess phenomenal content whereas conservatives

4

The strategy that I develop here is a fudge, intended to finesse the debate between those (‘intentionalists’ or

‘representationalists’) who hold that phenomenal properties just are representational properties of a certain kind and those (‘phenomenists’) who hold that phenomenal properties merely supervene in some way on representational properties. 5

There is a use of phenomenal content according to which phenomenal content is in some way the ground of

intentionality, see e.g. Kriegel (2007a); Horgan & Tienson (2002) and Graham et al (2007). I am not here committing myself to any such claim. As I use the term, an intentional state has phenomenal content if, and only if, it has proprietary phenomenal character. It does not follow from this that intentional content is somehow dependent on phenomenal character.

10

deny that they do. Conservatives might grant that one cannot enjoy cognitive states without being in some phenomenal state or another, but they deny that the phenomenal states associated with one’s cognitive states are attached to those states as such; instead, they hold that they are associated with one’s cognitive states in virtue of being attached to the perceptual/affective intentional states that are bound up with the subject’s cognitive states. Does every grammatical use of the ‘what it’s like’ locution pick out a state with phenomenal content? Probably not. I suspect that few liberals will want to be committed to the view that there is something distinctive that it is like to be a teenager, to be born into post-industrial Europe, or to live in Boise.6 Instead, liberals will take these phrases to pick out distinctive (and perhaps open-ended) clusters of phenomenal states. What it’s like to be a teenager is to be prone to bouts of moodiness, feelings of ennui, and so on; that is, there is no phenomenal state proprietary to being a teenager. By contrast, liberals hold, reference to what it’s like to understand a proof, wonder whether the Democrats will regain the White, and decide to go on holiday in Ethiopia do pick out distinctive phenomenal properties; what it’s like to decide to go on holiday in Ethiopia may involve a cluster of phenomenal states associated with (say) imagery of various kinds, but is also involves a phenomenal state proprietary to that propositional attitude. So, at any rate, the liberal holds.

6

But perhaps some liberals won’t mind being saddled with such a view. Flanagan quotes Peirce as claiming that

there is a distinctive quale to “every work of art—a distinctive quale to this moment as it is to me—a peculiar quale to every day and every week—a peculiar quale to my whole personal consciousness” (Flanagan 1992: 64). In a somewhat similar vein, William James suggested that “we ought to say a feeling of and, a feeling of if, a feeling of but, and a feeling of by, quite as readily as we say a feeling of blue or a feeling of cold” (James 1890: 245-46).

11

Although the distinction between conservatism and liberalism is a reasonably clear one, the task of identifying whether a certain philosopher might have conservative or liberal leanings is problematized by the fact that the debate is shrouded in multiple layers of terminological ambiguity. (I return to the question of terminology in section 5). Firstly, ‘experience’. Some authors use ‘experience’ and ‘experiential content’ in such a way that it is a definitional truth that (conscious) experiences must be (or be accompanied by) phenomenal states. Block, for example, introduced the term ‘phenomenal consciousness’ by stating that “phenomenal consciousness is experience” and that phenomenally conscious properties are experiential properties (Block 1995: 230). But there are other uses of ‘experience’ and ‘experiential content’—uses that allow for experiential content to be non-phenomenal. Tye, for example, holds that perceptual experience is structured in terms of high-level categories such as , and so on, but denies that such content is phenomenally represented (Tye 2000: 73-76; see also Carruthers 2000; Lyons 2005). Similar problems plague ‘phenomenal consciousness’ and its cognates. In line with what I regard as standard usage, I have taken ‘phenomenal consciousness’ as nothing more than a label for those states that possess what-it’s-likeness; on this usage, it is an open question whether cognitive states might possess phenomenal character. Other authors, however, appear to regard it as a definitional truth that only perceptual (and, sometimes, low-level perceptual) states can possess phenomenal character. Thus, Dretske states that he will use ‘phenomenal’ simply for a sensory mode of awareness (Dretske 1995: 12). Even if Dretske were to allow that there is something it is like for a subject to enjoy cognitive states, his use of ‘phenomenal’ would preclude him describing such states as phenomenal states. What’s more, some of those who reserve ‘phenomenal’ for perceptual states explicitly grant that cognitive states do possess what-it’s-likeness. Jaegwon Kim, for example, denies that

12

cognitive states have phenomenal character, but then goes on to claim that there is something it is like to be in them (Kim 2005: 201). So despite appearances to the contrary, Kim would count as a liberal in my sense of the term. Having shed some light on the contrast between liberals and conservatives I turn to the question of how the debate between them might be resolved.

3. Introspection It is natural to begin with introspection. Surely, one might think, questions about the reach of phenomenality ought to be resolvable by appeal to first-person reflection. Of course, this tempting thought is quickly deflated by the realization that both liberals and conservatives appeal to introspection in support of their cause. Liberals claim that the presence of cognitive phenomenality is introspectively manifest, whilst conservatives profess to be unable to find any trace of it in their experience. This standoff casts a dark cloud over introspection if we assume that are on all fours as far as phenomenal character is concerned: they either both possess cognitive phenomenality or both lack it. But perhaps this assumption should be rejected. Perhaps what it’s like to be a liberal is different—radically different—from what it’s like to be a conservative. It is difficult to rule out this proposal definitively, but I do find it hard to summon much enthusiasm for it. Of course, it would be naïve in the extreme to suppose that there are no significant phenomenal differences between normal human beings. I have no doubt that the phenomenal states I enjoy when tasting wine are radically impoverished relative to those available to my wine-tasting friends, nor do I doubt that the phenomenal states that accompany my partner’s orgasms are quite unlike those that accompany mine. But these judgments are motivated by behavioural differences, and to the best of my knowledge there are no behavioural differences between

13

conservatives and liberals—other, that is, than the fact that conservatives endorse conservatism and liberals endorse liberalism.7 Could the existence of cognitive phenomenality be influenced by top-down factors? Might liberals enjoy cognitive phenomenality because they expect to enjoy it whilst conservatives do not for precisely parallel reasons? It seems doubtful. If cognitive phenomenality were cognitively penetrable in this way then it would be radically unlike any other form of phenomenality with which we are acquainted. It seems safe to assume that either conservatives and liberals alike enjoy cognitive phenomenality or they both lack it. Introspection appears to be leading someone seriously astray. But who? The case for conservatism would be boosted by an account of why liberals (falsely) believe themselves to enjoy cognitive phenomenality, whereas the case for liberalism would be strengthened by an account of why conservatives overlook the phenomenal properties that characterizes their cognitive states. Is it plausible to suppose that conservatives might enjoy cognitive phenomenality of which they are oblivious? There is some reason to suppose that not all phenomenal states are introspectively accessible (Block 2007). One line of evidence for this thesis comes from Sperling’s work on briefly presented visual stimuli, in which subjects are presented with a 3 x 4 matrix of alphanumerical figures for 250 ms (Sperling 1960). Subjects can typically report 7

There might be situations in which it is reasonable to posit phenomenal variation in the face of behavioural

invariance. Martine Nida-Rümelin (1996), for example, has argued that some of us might be spectrum inverted relative to the rest of the population. But there are a number of differences between her proposal and the current proposal, the most obvious of which is that her proposal involves inverted qualia whereas the current proposal involves absent qualia. Phenomenal differences between us might be deeper than we ordinarily think, but it is stretches credulity to suppose that roughly half of us enjoy cognitive phenomenality and half of us do not.

14

only 4.5 of the 12 alphanumeric figures when asked to report as many of the figures as they can, but when a tone is sounded immediately after the matrix is presented, indicating that subjects are able to report a particular row (say, the middle row), they are able to correctly report 3.3 of the 4 figures. A plausible (although by no means uncontroversial) interpretation of these results is that subjects have a rich representation of the matrix that is phenomenally conscious but whose contents are not introspectively accessible (or at least, not reportable). More generally, it does not seem implausible to suppose that phenomenal states falling outside of the subject’s attentional focus might not be introspectively accessible. Does any of this help the liberal explain the conservative’s failure to detect cognitive phenomenality? Not really. Firstly, the debate about how best to interpret Sperling’s results and related phenomena is not about the introspective evidence—there is rough agreement about that—but about whether we are justified in going beyond the introspective evidence in positing phenomenal states. By contrast, the debate between conservatives and liberals is precisely about what introspection reveals. Second, the kinds of factors that arguably prevent phenomenal states from being introspectable in the context of Sperling’s displays, such as high cognitive load and processing bottlenecks, are unable to explain why cognitive phenomenality might be introspectively inaccessible for they simply are not operative in this case. This point is related to a third, namely, that the debate concerning cognitive phenomenality concerns types of phenomenal state types whereas that concerning phenomenality outside of attention concerns phenomenal state tokens. Liberals must explain why conservatives are unable to detect certain types of phenomenal properties, not why they cannot detect particular instantiations of phenomenal properties. As such, the debate over cognitive phenomenology ought to be far easier to resolve via introspection than the debate over phenomenality outside of attention: if there is cognitive phenomenality, then surely all

15

one need do in order to detect it is attend to one’s thoughts in a quiet evening hour. But conservatives have done this and have come up empty-handed. In sum, the kinds of limitations to which introspection is subject seem unable to explain why conservatives might be unable to introspect the phenomenal states that, according to liberals, characterize thought. In defence of their position, liberals might suggest that the phenomenal properties associated with cognition (and high-level perception) are elusive. Unlike (say) the phenomenal properties associated with colours, tastes, smell and itches, those associated with thought are shy, and it takes a certain kind of introspective acuity to detect them. This response cannot be dismissed out of hand. Perhaps some kinds of phenomenal properties are harder to detect than others. It would not be surprising if complex phenomenal properties were harder to detect than simple phenomenal properties, however exactly that distinction is to be drawn. Further, it seems plausible to suppose that more abstract phenomenal properties (such as the property distinctive of visual experience) are harder to detect than less abstract phenomenal properties (such as the property distinctive of experiences of red). But it is not obvious that the phenomenal properties that make up cognitive phenomenology need be particularly complex or abstract (in the relevant sense). Finally, as a matter of dialectical policy liberals cannot hold that cognitive phenomenology is all that elusive—after all, they claim to have introspective access to it! (Perhaps liberals will simply claim that their introspective abilities are more finely honed than are those of their conservative colleagues, but even this claim seems rather implausible.) In short, it seems to me that there is little reason to suppose that conservatives persist in overlooking vast swathes of their phenomenal states—phenomenal states that ought to be open to introspection. For their part, conservatives are committed to the view that liberals think that they enjoy phenomenal states that do not—and perhaps cannot—obtain. Arguably, the most plausible

16

liberal account of the error into which conservatives have fallen is to suppose that they confuse the phenomenal character of cognitive states with that of perceptual states (or clusters thereof). Liberals, so this thought goes, are latching on to phenomenal properties of some kind, it’s merely that they are mistaken in thinking that those properties are directly associated with cognitive states. Let us examine the prospects of this strategy in a somewhat indirect way, by considering a class of arguments that have been offered for liberalism. (These arguments are typically offered in support high-level perceptual phenomenality rather than cognitive phenomenality, but as I suggested in section two, the gap between high-level perceptual phenomenality and cognitive phenomenality is not a large one, and similar arguments can be given for cognitive phenomenology proper.) The arguments in question involve contrasting two scenarios that supposedly differ in phenomenal character but not in low-level perceptual content.8 For example, it is claimed that there is a difference between what it is like to hear the sentence “Il fait froid” when one does not understand French and what it is like to hear the same sentence having learnt French. In a similar vein, it is claimed that what it is like to hear the sentence “visiting relatives can be boring” depends on whether one takes the “boring” to qualify the relatives or the visiting, and that there is a ‘what it’s like’ difference between seeing the young woman/old woman figure as a young woman and seeing her as an old woman. Liberals claim that since these scenarios involve what-it’s-like differences that are not accompanied by difference in low-level perceptual content, phenomenality must extend at least as far as high-level perception, if not into cognition proper.

8

Contrast arguments can be found in Strawson (1994), Goldman (1993), Siewert (1998), Horgan and Tienson

(2002) and Siegel (2006). The method itself is discussed in Kriegel (2007) and Siegel (forthcoming).

17

Conservatives typically respond to contrast scenarios by granting that they involve differences in phenomenal character, but they insist that these differences can be fully accounted for in conservative terms. For example, that phenomenal difference between hearing “Il fait froid” as a monolingual Anglophone and hearing the same sentence having learnt French can be fully accounted for by appeal to changes in how one parses and segments the acoustic stream—pace liberals, there is no need to assume that high-level (semantic) representations possess phenomenal content. Similar accounts can be given of ambiguous expressions and multi-stable figures: the competing perceptual gestalts involve distinctive states of low-level perceptual focus and attention, and these states suffice to account for the phenomenal contrast between the gestalts. Does this response scuttle the liberal’s appeal to contrast cases? Not necessarily, but it does problematize it, for it does suggest that the liberal is not justified in assuming that differences in ‘what it’s likeness’ can be accounted for only by appealing to high-level phenomenal content. Nonetheless, the liberal might be correct to think that contrast cases do involve differences in high-level phenomenality, for it is not implausible to suppose that high-level phenomenality—if there is such a thing—supervenes on low-level phenomenality. The liberal asserts that high-level phenomenality exists, but she need not hold that it is autonomous. The liberal could hold that any difference in high-level phenomenal content must be accompanied by differences in low-level phenomenal content. To take an example of how this might go, consider the phenomenology of causation. One might hold that although visual phenomenology can represent causal relations, the phenomenal representation of causation is nomologically (indeed, perhaps even constitutively) dependent on the experience of spatiotemporal relations, and so in this sense is not autonomous. And it is not implausible to suppose that what holds of the experience of causation might hold more generally.

18

So, the state of play seems to be as follows. Conservatives hold that the appearance of cognitive phenomenality is mere appearance—that what liberals describe as the phenomenal character of cognitive states is actually that of accompanying (low-level) perceptual states. Liberals, for their part, can allow that differences in high-level and cognitive phenomenal properties supervene on low-level phenomenal differences, but they deny that the phenomenal differences that characterize contrast cases are exhausted by low-level differences. Over and above any such low-level differences are phenomenal differences that are proprietary to high-level perception and cognition. In short, liberals insist that they are not guilty of misidentifying low-level phenomenal states. Perhaps such claims are wrong— indeed, if conservatism is right then they must be wrong!—but to accuse liberals of this error is to accuse them of a fairly radical kind of error, an error that it seems implausible to suppose liberals are making. In short, it seems to me that the best account of why liberals are attracted to liberalism, should conservatism be true, is not particularly convincing. Where does this leave us? We have seen that although there are liberal accounts of why conservatives might be attracted to conservatism and conservative accounts of why liberals might be attracted to liberalism, none of these accounts is compelling. What they all have in common is the rather unpalatable consequence that introspection is liable to lead us into serious error. If conservatism is right then introspection is guilty of sins of commission; if liberalism is right then it is guilty of sins of omission: either way, introspection proves itself to be an unreliable witness when it comes to the reach of phenomenal consciousness.9

9

See Schwitzgebel (2008) and Spener (2008) for further reason to think that introspection might not be a

trustworthy—or at least unproblematic—guide to consciousness.

19

It is worthwhile pausing to reflect on the magnitude of this problem. It does not concern some subtle point about (say) how best to describe what it’s like to be about to address a large audience, or how to capture the strange feeling of excitement and fatigue that one has when exploring a new city while jetlagged. Rather, the question at issue is precisely the sort of question that introspection ought to be able to resolve with ease. Unlike the disagreement about whether there is consciousness outside of attention we cannot attribute disagreement about the reach of phenomenality to difficulties in cognitive access to phenomenal content, for the participants in this debate have ample opportunity to screen off the influence of those factors that might impair introspection. Another way in which one might account for the debate about whether or not there is consciousness outside of attention is to suppose that phenomenality comes in degrees, and that different theorists are operating with different thresholds or criteria for how much phenomenality a state must possess in order to qualify as a phenomenal state. But it’s hard to see how appealing to degrees of phenomenality might account for the debate between liberals and conservatives. Even if phenomenality can come in degrees—a proposition whose truth is uncertain—there is no reason to suppose that cognitive states should possess less phenomenality than perceptual or sensory states. Further, most instances of phenomenal disagreement are accompanied by some degree of subjective uncertainty. This is certainly my own situation vis-à-vis the question of whether there is phenomenality outside of attention; I have some inclination to think that there is, but I am not sure whether this inclination amounts to belief, and it certainly does not rise to the level of conviction. I doubt that I am alone in this regard. But it seems to me that precisely the opposite is true of the phenomenal content debate: conservatives tend to be adamant that phenomenal character is restricted to low-level content, whereas liberals tend to be equally

20

certain that it is not. It seems unlikely that the phenomenal content debate will be resolved by appeal to introspection.

4. Knowledge arguments and explanatory gaps Perhaps conceptual analysis can help where introspection has been found wanting. Perhaps there is something in the very notion of what it’s likeness that prevents its extension to cognition.10 One line of thought that is worth considering concerns the connection between phenomenality and appearance. Dennett once remarked that “‘qualia’ is an unfamiliar term for something that could not be more familiar to us: the ways things seem to us” (Dennett 1988: 43). Might there be a case for conservatism in these waters given that cognitive states are not—in the relevant sense—appearance states?11 I think not. Even if phenomenality is in fact tied to perceptual systems there is little reason to regard this thesis as a conceptual truth. Many theorists, conservatives and liberals alike, hold that there are phenomenal states that are not associated with representational content. Examples of states for which such accounts have been defended include bodily sensations, such as pains, and non-directed moods, such as generalized feelings of hedonia. What matters in the present context is not the correctness of non-intentional accounts of these states but

10

Note that arguments of this kind might commit some conservatives to more than they would be happy with,

for some conservatives might want to hold that although phenomenality happens to be restricted to low-level perception this restriction is merely contingent. 11

The ‘in the relevant sense’ is needed in order to rule out beliefs about how the world appears to one, for

although such states are appearance representations in one sense they are not appearance representations in the relevant sense.

21

their coherence, for if it is coherent to suppose that these states are non-representational then it cannot be a conceptual requirement that phenomenal states are the ways things seem or appear to us.12 But perhaps it was naïve to suppose that a straightforward analysis of the notion of ‘what it’s likeness’ might provide any traction on the phenomenal content debate. As far as I know, the phrase ‘what it’s like’ was first used in connection with consciousness by Wittgenstein (1980: §91). It was subsequently picked up by Farrell (1950), Sprigge (1971) and of course Nagel (1974). As best I can tell, the ways in which these authors employ the phrase do not give us any reason to favour one side of this debate over the other. Although Wittgenstein and Farrell focus on the ‘what it’s likeness’ of sensory states, nothing in what they say limits the extension of phenomenal character to sensation as a matter of conceptual necessity. Sprigge (1971) adopts a decidedly liberal conception of the reach of what-it’s-likeness. Nagel, of course, emphasizes our inability to appreciate what it is like for the bat to perceive the world as it does, but questions concerning the bat’s cognitive perspective seem to be no less pressing than those raised by its perceptual perspective. A direct assault on the notion of what it’s likeness seems unlikely to shed much light on the phenomenal content debate, but perhaps progress can be made by examining some of the puzzles with which phenomenal consciousness is most closely associated. Consider Jackson’s expert colour scientist Mary (Jackson 1982). As you will recall, Mary knows everything there 12

Even if we were to restrict phenomenality to appearance properties, we would not be able to secure the kind

of conservatism that many conservatives would like, for high-level perceptual content is as much a matter of appearance as low-level perceptual content is. Monkeys not only appear to one as (say) brown, moving, and furry—they also appear to one as monkeys. There is no sharp division in perceptual processing between lowlevel and high-level content. It’s categories all the way up (or down).

22

is to know about the cognitive neuroscience of colour perception, despite the fact that she has never see red. What happens, Jackson asks, when she first sees red? Jackson’s answer is that Mary learns something new—she learns what it is like to experience red. Since Mary knew everything that there was to know about the physical basis of colour perception, it follows— so Jackson argued—that what it is like to see red cannot be a physical fact.13 Whether or not the knowledge argument is sound it has certainly struck many philosophers of mind—not to mention a few hundred thousand undergraduates—as prima facie plausible. Mary does seem to learn something when she first sees red. Of course, the conclusion was not meant to be restricted to what it is like to see red. As Jackson pointed out, the argument can be deployed for “….taste, hearing, the bodily sensations and generally speaking for the various mental states which are said to have (as it is variously put) raw feels, phenomenal features or qualia” (Jackson 1982: 130). But exactly how far can the argument be extended? Can it be extended to cognitive (or high-level perceptual) states? If not, the conservative might argue, we have good reason to think that such states do not have phenomenal character. Perhaps, the conservative might continue, it is no accident that Jackson’s discussion of what the knowledge argument establishes had a decidedly conservative slant to it: Tell me everything physical there is to tell about what is going on in a living brain, the kind of states, their functional role, their relation to what goes on at other times and in other brains, and so on and so forth, and be I as clever as can be in fitting it all together, you won’t have told me about the hurtfulness of pains, the itchiness of itches, pangs of jealousy, or about the characteristic experience of tasting a lemon, smelling a rose, hearing a loud noise or seeing the sky. (Jackson 1982: 127). 13

It should be noted that Jackson no longer endorses the knowledge argument.

23

It is instructive to put the foregoing argument more formally. 1. It is not possible to run a plausible version of the knowledge argument using cognitive or high-level perceptual states. 2. If cognitive or high-level perceptual states had phenomenal content, then it would be possible to run a plausible version of the knowledge argument using them. 3. So, cognitive and high-level perceptual states do not have phenomenal content. Is it possible to develop a plausible version of the knowledge argument that appeals to highlevel perceptual or cognitive content? Let us leave cognitive content to one side, and focus only on high-level perceptual content. Consider Mary’s brother Harry. Harry is an expert in the cognitive neuroscience of object identification. Despite not having seen a monkey, he knows all there is to know about the neural processes involved in recognizing monkeys. Does Harry learn something new when he first sees a monkey? Perhaps, but it is far from clear that this is so. Whereas we expect Mary to say “Oh this is what it’s like to see red”, we do not expect Harry to say “Oh this is what it is like to see a monkey”.14

14

The reason for this may have something to do with what we might call ‘phenomenal proximity’. In setting up

the original knowledge argument, it seems to be important that we stipulate that Mary has never seen reds of any kind. A version of the knowledge argument in which Mary has seen (say) pink but not purple would not be as potent as Jackson’s original version of the argument. Why is this? Perhaps it is because the ‘phenomenal distance’ between pink and purple is not substantial, and can be traversed from the armchair so to speak. Developing this thought, one might argue that it is difficult to construct a version of the knowledge argument for high-level perceptual content and cognitive content because there will not be sufficient phenomenal distance between Mary’s initial phenomenal states (whatever they happen to be) and the target states with which she is unacquainted. Of course, this is just a sketch of an account; working out the details of this story would take some doing.

24

Perhaps it is possible to construct a liberal version of the knowledge argument—indeed, Goldman invokes knowledge argument intuitions to argue that certain propositional attitudes, such as doubt and disappointment, do have phenomenal character (1993: 24)—but it is undeniable that constructing a plausible version of the argument becomes much harder once one leaves the realm of the sensory. Only the most naïve of philosophy instructors would dare to teach the knowledge argument to undergraduates by replacing Mary with Harry. So, do we have here an argument for conservatism? No, for despite its intuitive plausibility premise (2) is untenable. We can see why by considering experiences of space. Suppose that Mary knows everything that there is to know about the perceptual representation of space but has never perceived a straight line. Does she learn anything when she first sees a straight line? Will Mary say, “Oh this is what it is like to see a line as straight”? Arguably not. The point applies to primary properties in general: it is much harder to construct plausible versions of the knowledge argument around experiences of primary qualities, such as motion, distance, or symmetry, than it is to construct plausible versions of the argument around experiences of secondary qualities, such as colours, flavours, and smells. It is no accident that Mary is colour-blind rather than (say) akinotopsic. Perhaps Jackson’s decision to center the knowledge argument on colour experience was forced on him by the demands of the argument rather than being a concomitant of any commitment to conservatism that he may (or may not) have had. Perhaps it is possible to construct a plausible version of the knowledge argument that appeals only to primary qualities, but it is clear that in order to do so one has to work much harder than Jackson had to work. What follows from this? Should we conclude that representations of primary qualities cannot be phenomenally conscious? Surely not! There is undoubtedly something it is like to see an object as moving, to feel it as straight, or to hear it as located behind one. The kind of über-

25

conservatism that would restrict the scope of phenomenal properties to secondary qualities has no intuitive plausibility whatsoever. (Certain theorists might use “phenomenal character” as a synonym for the perceptual representation of secondary qualities, but this is obviously a different issue.) The fact that high-level and cognitive content resists knowledge argument deployment does not count against liberalism, for knowledge argument intuitions do not function as a litmus test of phenomenality.15 The knowledge argument and its associated intuitions is not the only puzzle associated with phenomenal consciousness—there is also the explanatory gap. Perhaps ‘gap intuitions’ can be used to motivate phenomenal conservatism. The conservative might argue as follows: 1. Every state that enjoys phenomenal content must involve (the appearance of) an explanatory gap. 2. Cognitive states do not involve (the appearance of) an explanatory gap.

15

It also of some interest to note that in introducing the notion of phenomenal consciousness theorists invariably

refer to secondary qualities, such as colours, flavours, sounds and so on. Why is this? One can understand why conservatives avoid high-level perceptual content in introducing phenomenality—after all, they don’t believe in it—but why should they (and liberals too) restrict their attention to secondary qualities and ignore primary qualities? I suspect that the reason for this derives in part from the thought that secondary qualities seem to be non-intentional, or subjective, in a way that primary qualities do not. By focusing on states that appear to lack intentional content, theorists hope to draw the reader’s attention to a feature of mentality—its phenomenal character—that is not merely intentional. (And, of course, some of the appeal of the knowledge argument derives from the assumption that phenomenal character outruns intentional content.) Although there is something to be said for this strategy, it runs the unfortunate risk of encouraging a separatist picture of the mind, according to which there are no deep relations between phenomenal properties and intentional properties.

26

3. So, cognitive states do not enjoy phenomenal content. (Here, too, the conservative might take some comfort from the fact that the proponent of the explanatory gap argument may also have conservative inclinations. To the best of my knowledge Levine does not give an account of the range of phenomenal properties, but the only examples he gives of phenomenal properties are experiences of pain and colour.) Premise (1) seems to be secure. Here, unlike the knowledge argument, there seems to be no asymmetry between primary and secondary qualities. There is as much of an explanatory gap between experiences of movement and the brain areas that subserve them as there is between experiences of colour and the brain areas that subserve them. But premise (2) is false, for there is as much of an explanatory gap—or at least the appearance thereof—between cognitive states and the brain states that subserve them as there is between low-level perceptual states and the brain areas that subserve them. We are as hard-pressed to explain how activity in (say) the prefrontal cortex produces conscious thoughts as we are to explain how activity in V4 produces experiences of red or activity in MT produces experiences of motion. Explanatory gap intuitions no more mark a divide between perceptual states and cognitive states than do knowledge argument intuitions. One might argue—as some have—that to look to the intrinsic features of neural activation for an explanation of phenomenal states is to look in the wrong place (see e.g. Dretske 1995). According to representationalists, neural states subserve particular phenomenal states only because of the properties (objects, events) that they represent. The gap between neural states and phenomenal states is just the gap between representational vehicles and their contents, and that’s the kind of gap that ought to be expected. This is not the place in which to enter into a discussion of the merits of representationalist accounts of phenomenal character.

27

Luckily, no such discussion is required, for the only point that needs making is this: if representationalism defangs explanatory gap intuitions, then it also defangs any explanatory gap based argument for conservatism. I have examined two of the most prominent arguments associated with phenomenal consciousness—Jackson’s knowledge argument and Levine’s explanatory gap argument— with a view to the question of whether the intuitions at play in these arguments might give some support to the conservative cause. The conclusion that we have reached is that they do not. Is this merely a ‘null result’, or does it give some support to the liberal cause? After all, wouldn’t the apparent absence of a priori restrictions on the scope of what it’s likeness provide some support for the liberal cause? Perhaps, but perhaps not. Firstly, the failure of a priori arguments for conservatism does not itself constitute an a priori argument for liberalism. It is one thing to fail to see how phenomenality could be restricted to perceptual states, it is quite another to see that it need not be restricted to perceptual states. (It is one thing to see that a proof for a mathematical postulate fails, it is another to see that the postulate can be proved false.) Secondly, some conservatives hold that the truth of conservatism is a posteriori and cannot be established from the armchair (see e.g. Tye 1995: 141). Conservatives of this ilk hold that although cognitive states do not—and perhaps cannot—possess phenomenal character, there is no conceptual incoherence in supposing that they could. Those who take this line will be unperturbed by the failure of a priori arguments against liberalism.

5. Theories of consciousness? Another source of evidence that could, in principle, have a bearing on the phenomenal content debate derives from accounts of phenomenal consciousness. Unfortunately, we have

28

no such account—or rather, we have too many such accounts, and there is little agreement as to which accounts are most promising. Accounts that favour conservatism are likely to be rejected by liberals precisely because they favour conservatism and vice-versa. Arguably, it is more reasonable to use one’s account of the reach of phenomenal consciousness to constrain one’s account of what it is that differentiates those mental states that are phenomenally conscious from those that are not than to proceed in the opposite direction. Nonetheless, appealing to theories of consciousness might provide us with some traction on this debate. Consider Block’s notion of access consciousness, where a state is access conscious exactly when its content is globally available for verbal report and the rational control of action (Block 1995). According to many, access conscious and phenomenal consciousness are coextensive, perhaps even necessarily so. Most of the debate about the relationship between access consciousness and phenomenal consciousness has focused on whether phenomenal states must be access conscious, but there has been rather less discussion of whether access conscious states must be phenomenally conscious. (From the point of view of the claim that phenomenal consciousness and access consciousness are co-extensive, this kind of dissociation would no less telling than the other form of dissociation.) But if conservatism is correct, not only is it possible for a state to be access conscious without being phenomenally conscious, but in fact such states are ubiquitous, for the description exactly describes conscious thought as the conservative conceives of it. More generally, it appears as though the conservative will not be able to correlate phenomenality with any functional property akin to access consciousness, for any such property is likely to be possessed by low-level and high-level states alike. So, how should conservatives approach phenomenality? It is natural for conservatives to suggest that the distinction between states that possess phenomenal character and those that

29

do not is correlated with distinctions in content. Not surprisingly, this is precisely the approach that conservatives tend to take; Tye and Dretske, for example, argue that phenomenal states must have non-conceptual content (Tye 1995; Dretske 1995). So, generally speaking, reasons to think that phenomenality should be accounted for in terms of distinctions between content would favour conservatism, whereas reasons to think that phenomenality can be accounted for in terms of a state’s availability to systems of (say) theoretical and practical reasoning would favour liberalism. Do we have such reasons? Not as far as I can tell. Again, the appropriate order of analysis is to begin with an account of the reach of phenomenal consciousness and only then to build a theory of consciousness, rather than proceed in the opposite direction. Our brief excursus into theories of consciousness has failed to resolve the debate between conservatives and liberals, but it has revealed additional reasons for taking the debate seriously.

6. A storm in a terminological teacup? In light of the foregoing it is tempting to suppose that the debate between conservatives and liberals might be ‘merely’ terminological.16 Perhaps there are two kinds of what-itslikenesses, one of which maps onto the conservative position and another that maps onto the liberal view. This proposal has a great deal of appeal—for one thing, by taking liberals and conservatives to be talking passed each other we preserve the reliability of introspection—but can it be sustained?

16

As Schwitzgebel has recently remarked, debates about what is introspectively manifest “beggar the

imagination; they plead for reinterpretation as disagreements only in language or theory, not real disagreements about the phenomenology itself” (2008: 258).

30

The thought that there might be two forms of WILness is not a new one. In arguing that the phenomenal extends beyond the sensational Flanagan claimed that the concept of qualia has both a wide and a narrow sense (Flanagan 1992: 67). Although a conservative, Tye allows that there may be “a very broad use of the locution ‘what it’s like’ in ordinary life which concedes a difference in what it is like whenever there is any conscious difference of any sort whatsoever”; however, he goes on to add that this is not his usage (Tye 1996: 302, n. 3). In a similar vein, Georgalis distinguishes a restricted sense of ‘what it’s likeness’ from an unrestricted sense: the former involves only the stimulation of sense organs, whereas the latter includes propositional attitudes and contents (Georgalis 2006: 69). Unfortunately, none of these authors provides any analysis of just what these contrasts amount to. Positing two conceptions of ‘qualia’, ‘what it’s likeness’ or ‘phenomenality’ helps us not one jot unless we have some grip—preferably non-extensional—on how they might differ. I know of only one account that might help. According to Lycan (1996; in press) and Carruthers (1998; 2000), there is a distinction between ‘worldly’ (Carruthers) or ‘first-order’ (Lycan) what-it’s-likeness on the one hand and ‘mental-state’ (Carruthers) or ‘second-order’ (Lycan) what-it’s-likeness on the other.17 Here is how Carruthers presents the contrast: It is possible to draw a distinction between what the world (or the state of the organism’s own body) is like for an organism, and what the organism’s experience of the world (or of its own body) is like for the organism. … it is one thing to say that the world takes on a subjective aspect by being presented to subjects with differing conceptual and

17

A very similar distinction—couched in terms of a contrast between thin phenomenality and thick

phenomenality—can be found in Rosenthal (2002). Byrne (2004) argues, quite persuasively, that Rosenthal’s distinction is basically a terminological variant of the Carruthers-Lycan distinction.

31

discriminatory powers; and it seems quite another thing to say that the subject’s experience of the world also has such a subjective aspect, or that there is something which the subject’s experience is like. (Carruthers 2000: 128). A question: could this contrast dissolve the phenomenal content debate only if it is real, or is it sufficient that one (or more) parties to the debate take there to be a distinction between worldly and mental-state WILness? The answer, I think, is clear: the contrast must be wellfounded, for only then could we explain why conservatives and liberals give such different accounts of the reach of phenomenality. Merely supposing that one can distinguish worldly from mental-state WILNess would fail to explain away the disagreement. Is the distinction between worldly WILness and mental-state WILness well-founded? I have my doubts. Phenomenal properties are properties of the subject of experience—a phenomenal state is a state that there is something it is like for the subject of experience to be in. I do not see anything in this conception of WILness that might support a distinction between two kinds of WILness. Of course, there is a distinction between consciously representing the world and consciously representing one’s own experiences of the world, but this distinction does not bring with it a distinction between two kinds of WILness. Consider a parallel: some intentional states represent non-intentional states and some represent intentional states, but it does not follow that there are two types of intentionality. Similarly, some phenomenal states are associated with first-order states and some are associated with second-order states, but it does not follow that there are two kinds of WILness or phenomenality. But let assume, if only for the sake of argument, that the distinction between worldly and mental-state WILness is secure, and that we can establish either that conservatives are interested in worldly WILness and liberals in mental-state WILness or vice-versa. Would this

32

dissolve the debate? Well, it would certainly help, but it would not itself sever the Gordian knot in which we find ourselves entangled, for the phenomenal content debate is not a tidy one. As I noted in section 2, conservatism and liberalism are not monolithic positions. ‘Liberal conservatives’ hold that high-level perceptual states but not cognitive states possess phenomenal character, whereas ‘conservative liberals’ hold that propositional attitudes but not propositional states themselves posses phenomenal character. Suggesting that the phenomenal content debate arises from conflating worldly WILness with mental-state WILness fails to account for these intermediate positions. Supposing that this problem can be addressed, let us return to the question of how, if at all, the distinction between worldly and mental-state WILness might map on to the distinction between conservatism and liberalism. We can start by putting this question to those who are most keen on the contrast between worldly and mental-state WILness: Carruthers and Lycan. Carruthers seems to pair the liberal view with worldly WILness and the conservative view with mental-state WILness: Of course the world is like something to any perceiver and to any thinker, whether their states are phenomenally conscious or not. For any experience, and any thought, will involve a partial and partially subjective ‘take’ on the elements of perception/thought. What is crucial for phenomenal consciousness, however, is that there should be something that the subject’s own mental states are like, for them….With this distinction in place, there is no reason to believe that non-imagistic thoughts will be like anything. (Carruthers 2005: 139). I suspect that Carruthers holds this view (in part) because he thinks that it is mental-state WILness that is the interesting (important, controversial) notion—it is this form of WILness

33

that deserves the ‘phenomenal consciousness’ moniker.18 But Lycan takes the liberals to be talking about mental-state WILness: “by common consent”, he says, the kind of WILness that proponents of cognitive phenomenology have in mind is higher-order rather than firstorder WILness (2008: 18).19 Further, Lycan holds that liberals are right to think that there is cognitive phenomenality in this higher-order sense. So, Carruthers holds a liberal view of worldly WILness and a conservative view of mental-state WILness, whereas Lycan holds a conservative view of worldly WILness and a liberal view of mental-state WILness!20 It is clear that the distinction between worldly WILness and mental-state WILness does not provide us with any straightforward dissolution of the phenomenal content debate. Is there any reason to limit worldly WILness to perception? Not as far as I can see, for both perceptual and cognitive representations present the world to the subject. What about mentalstate WILness? Would a focus on mental-state WILness lead to conservatism? I do not see 18

He didn’t always hold this view. As Alex Byrne (2004) notes, in an earlier paper Carruthers had allowed that

either sense of what it’s likeness could be given the title of ‘phenomenal consciousness’ (Carruthers 1998: 209). 19

Consider also the following: “Now we can adjudicate the question of whether nonsensory mental states such

as thoughts, desires and intentions involve “qualia” or have “phenomenal character.” In the Lewis-Goodman sense of Q-properties [authorial insertion: equivalent to the first-order sense], they do not, because by stipulation, Q-properties are specifically sensory properties like colors and pitches and smells. In the higherorder sense some non-sensory states do have phenomenology, because … it is sometimes like something, for the subject, to believe that P, to want such-and-such, or to experience doubt and disappointment.” (Lycan in press: 12). 20

This disagreement is echoed in a disagreement about how to understand what it is that other consciousness

theorists are attempting to address. Carruthers (2000) takes theorists such as Block, Tye and Dretske to be giving accounts of mental state (higher-order) what-it’s-likeness, whereas Lycan (Lycan 1999, note 1) takes them to be giving accounts of worldly (first-order) what-it’s-likeness.

34

why it should. If there can be something that it is like for the subject to be in a particular kind of perceptual state why could there not also be something that it is like for the subject to be in a particular kind of cognitive state? Of course, one could stipulate that only experiential states can possess mental-state WILness, but the connection between mental-state WILness and perception must be substantive rather than stipulative if it is to do any work here. The distinction between worldly and mental-state WILness may contribute to a solution to the phenomenal content debate, but it does not itself provide that solution. Perhaps other analyses of ‘what it’s like’ might succeed where this analysis has failed, but I know of no other analyses. Even if some such analysis were to reveal two forms of WILness onto which the conservative and liberal positions might be mapped, further questions would remain about which form of WILness was ‘deepest’, ‘most important’, or ‘most puzzling’. Conservatives would no doubt insist that it is the narrow or restricted notion that poses the deepest challenges to naturalistic accounts of the mind, whereas liberals would presumably hold precisely the opposite view. In short, substantive disagreements between conservatives and liberals may remain even if we were to locate two genuinely distinct notions of ‘what it’s likeness’.21

7. Conclusion It is typically assumed that the notion of phenomenal consciousness is clear enough to be employed for the purposes of theory building; as Byrne puts it, phenomenal character may be

21

As a case in point, consider the contrast between Flanagan and Tye. Flanagan, a liberal, allows that there is a

narrow sense of qualia, but assumes that a theory of consciousness should focus on the broad sense. Tye, a conservative, allows that there is a broad sense of WILness but assumes that a theory of consciousness should concern itself with the narrow notion of WILness.

35

“hard to explain” but it is “easy to understand” (Byrne 2001: 200).22 The burden of this paper has been to suggest that the notion of phenomenal consciousness may be rather harder to understand than we tend to assume. Perhaps it’s not only those approaching philosophy of mind for the first time who ought to be puzzled by the notion. To press this point home, imagine a world—call it “Inverted”—in which debates about the reach of phenomenality concerns only low-level perception. The inhabitants of Inverted agree that there is something distinctive that it is like to wonder whether wombats are monogamous and to recognize a visually-presented object as a tiger, but controversy rages over whether there is something that it is like to see an object as yellow, taste it as peppery, or hear it as moving from one’s right to one’s left. We would be puzzled by such a debate and rightly so, for its mere existence would give us some reason to wonder whether the inhabitants of Inverted had a firm grip on the notion of ‘what it’s likeness’. But if we would be justified in harbouring such suspicions about the Invertians then surely they would be justified in harbouring such suspicions about us. More to the point, conservatives and liberals are entitled to harbour such suspicions about each other. Where does this leave us? Some years ago Bill Lycan recommended that we ‘just say no’ to the phrase ‘what it’s like’; it’s tokening, he claimed, was worse than useless, and ‘sends the struggling mind of even the most talented philosopher into yet another affect-driven tailspin of confusing a welter of distinct issues’ (1996: 77). I have some sympathy with Lycan’s position but I cannot second his motion. For all the difficulties that surround the phrase, it is not clear how else we are to characterize consciousness. Talk of ‘what it’s likeness’ might do a pretty lousy job of demarcating what it is that we care about in trying to understand 22

He does go on to qualify this claim by adding, “[a]t any rate, everyone seems to understand it.” (2001: 200).

36

consciousness, but—for the moment at least—it’s all we have. 23 The question is whether it is enough.

References Block, N. 1995. On a confusion about a function of consciousness, Behavioral and Brain Sciences, 18: 227-87. Block, N. 2007. Consciousness, accessibility, and the mesh between psychology and neuroscience, Behavioral and Brain Sciences, 30: 481-548. Braddon-Mitchell, D. & Jackson, F. 2007. Philosophy of Mind and Cognition (2nd ed.). Oxford: Blackwell. Byrne, A. 2001. Intentionalism defended. The Philosophical Review, 110/2: 199-240. Byrne, A. 2004. What phenomenal consciousness is like. In R. Gennaro (ed.) Higher-Order Theories of Consciousness: An Anthology. John Benjamins, 203-225. Carruthers, P. 1998. Natural theories of consciousness. European Journal of Philosophy, 6: 203-222. Reprinted in P. Carruthers, Consciousness: Essays from a Higher-Order Perspective. Oxford: Oxford University Press, pp. 36-60. Carruthers, P. 2000. Phenomenal Consciousness: A Naturalist Theory. Cambridge: Cambridge University Press.

23

Acknowledgements.

37

Carruthers, P. 2005. Conscious experience versus conscious thought. In Consciousness: Essays from a Higher-Order Perspective. Oxford: Oxford University Press. Chalmers, D. 1996. The Conscious Mind. Oxford: Oxford University Press. Clark, A. 2000. A Theory of Sentience. Oxford: Oxford University Press. Dennett, D. 1988. Quining qualia. In A.J. Marcel and E. Bisiach (ed.) Consciousness in Contemporary Science. Oxford: Oxford University Press. Dretske, F. 1995. Naturalizing the Mind. Cambridge, MA: MIT Press. Farrell, B. 1950. Experience. Mind 49: 170-98. Flanagan, O. 1992. Consciousness Reconsidered. Cambridge, MA: MIT Press. Georgalis, N. 2005. The Primacy of the Subjective. Cambridge, MA: MIT Press. Goldman, A. 1993. The psychology of folk psychology, Behavioural and Brain Sciences, 16: 15-28. Graham, G., Horgan, T. & Tienson, J. 2007. Consciousness and intentionality. In M. Velmans & S. Schneider (eds.) The Blackwell Companion to Consciousness. Oxford. Blackwell. Horgan, T. & Tienson, J. 2002. The intentionality of phenomenology and the phenomenology of intentionality. In David J. Chalmers (ed.) Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Jackendoff, R. 1987. Consciousness and the Computational Mind. Cambridge, MA: MIT. Jackson, F. 1982. Epiphenomenal qualia, Philosophical Quarterly, 32: 127-136. Jacob, P. 1998. What is the phenomenology of thought? Philosophy and Phenomenological Research, 58: 443-448.

38

James, W. 1890. The Principles of Psychology. Harvard: Harvard University Press. Kim, J. 2005. Philosophy of Mind (2nd edition). Boulder CO: Westview Press. Kriegel, U. 2002. Phenomenal content. Erkenntnis, 57: 175-98. Kriegel, U. 2007a. Intentional inexistence and phenomenal intentionality. In J. Hawthorne (ed.) Philosophical Perspectives, Vol. 21: Philosophy of Mind, pp. 307-40. Kriegel, U. 2007b. The phenomenologically manifest. Phenomenology and the Cognitive Sciences, 6: 115-36. Langsam, H. 2000. Experiences, thoughts, and qualia. Philosophical Studies, 99: 269–295. Levine, J. 2001. Purple Haze: The Puzzle of Consciousness. Oxford: Oxford University Press. Loar, B. 2003. Phenomenal intentionality as the basis of mental content. In M. Hahn and B. Ramberg (eds.) Reflections and Replies: Essays on the Philosophy of Tyler Burge. Cambridge, MA: MIT Press. Lormand, E. 1996. Nonphenomenal consciousness. Nous, 30: 242-61. Lycan, W. 1996. Consciousness and Experience. Cambridge, MA: MIT Press. Lycan, W. 1999. A response to Carruthers’ “Natural Theories of Consciousness”, Psyche, 5: http://psyche.cs.monash.edu.au/v5/psyche-5-11-lycan.html. Lycan, W. In Press. Phenomenal intentionalities. American Philosophical Quarterly. Lyons, J. 2005. Perceptual belief and nonexperiential looks. Philosophical Perspectives, 19, Epistemology, 237-56. McGinn, C. 1988. Consciousness and content. Proceedings of the British Academy 74: 21939. Reprinted in The Problem of Consciousness. Oxford: Blackwell, 1991.

39

Nagel, T. 1974. What is it like to be a bat?, Philosophical Review, LXXXIII: 435-50. Nelkin,

N.

1989.

Propositional

attitudes

and

consciousness,

Philosophy

and

Phenomenological Research, 49 (3): 413-30. Nida-Rümelin, M. 1996. Pseudonormal vision, an actual case of qualia inversion? Philosophical Studies, 82: 145-157. Pereboom, D. In press. Consciousness and introspective accuracy. In S. Newlands and L. M. Jorgensen (eds.) Metaphysics and the Good: Themes from the Philosophy of Robert Merrihew Adams. Oxford: Oxford University Press. Pitt, D. 2004. The phenomenology of cognition Or What is it like to think that P? Philosophy and Phenomenological Research, 69: 1-36. Prinz, J. 2007. The intermediate level theory of consciousness. In M. Velmans & S. Schneider (eds.) The Blackwell Companion to Consciousness. Oxford. Robinson, W. S. 2005. Thoughts without distinctive non-imagistic phenomenology. Philosophy and Phenomenological Research, LXX: 534-61. Rosenthal, D. 2002. How many kinds of consciousness? Consciousness and Cognition, 11: 653-665. Schwitzgebel, E. 2008. The unreliability of naïve introspection, The Philosophical Review, 117/2: 245-73. Shoemaker, S. 1996. Self-knowledge and “inner sense”: Lecture III. In S. Shoemaker The First-Person Perspective and Other Essays. Cambridge: Cambridge University Press, 24668.

40

Siegel, S. Forthcoming. How can we discover the contents of experience? Southern Journal of Philosophy. Siewert, C. 1998. The Significance of Consciousness. Princeton, N.J.: Princeton University Press. Spener, M. 2008. Phenomenal adequacy and introspective evidence.

Unpublished

manuscript. Sperling, G. 1960. The information available in brief visual presentations, Psychological Monographs, 74, no. 11. Sprigge, T. 1971. Final causes. Proceedings of the Aristotelian Society (Supplement), 45: 149-70. Strawson, G. 1994. Mental Reality. Cambridge, MA: MIT Press. Tye, M. 1995. Ten Problems of Consciousness. Cambridge, MA: MIT Press. Tye, M. 1996. The function of consciousness, Nous 30/3: 287-305. Tye, M. 2000. Consciousness, Color and Content. Cambridge, MA: MIT Press. Wittgenstein, L. 1980. Remarks on the Philosophy of Psychology, Vol 1. Chicago, IL: University of Chicago Press.