Higher Order Thought

  • View
    224

  • Download
    0

Embed Size (px)

Citation preview

  • 8/4/2019 Higher Order Thought

    1/30

    The Southern Journal of Philosophy (2008) Vol.XLVl

    Higher-Order Thought andthe Problem of Radical ConfabulationTimothy LaneNational Chengchi University, TaiwanCaleb LiangNational Taiwan University, Taiwan

    AbstractCurrently, one of th e most influe ntial theori es of consciousness i sRosenthals version of higher-order-thought (HOT). We argue that theHOT theory allows for two distinct interpretations: a one-componentand a two-component view. We further argue that t he two-componentview is more consistent with his effort to promote HOT as anexplanatory theory sui table for application to the empirical sciences.Unfortunately, t he two-component view seems incapable of handling agroup of counterexamples that we refer t o as cases of radical confabu-lation. We begin by introducing the HOT theory and by indicating whywe believe it is open to distinct interpretations. We then proceed toshow that it is incapable of handling cases of radical confabulation.Finally, in t he course of considering various possible responses to ourposition, we show th a t adoption of a disjunctive strategy, one th a twould countenance both one-component and two-component versions,would fail to provide any empirical or explanatory advantage.

    According t o David Rosenthals influential higher-order thought( H O T ) theory of consciousness, what makes a mental stateconscious is that it is accompanied by a suitable higher-order

    Tim L a n e h a s s t u d ie d b r oa d ly , w o r k i ng a t v a r io u s t i m e s i n t h r eed i f fe ren t Ph D programs . Prob lems per ta in ing to m in d a nd consc ious -n e s s h a v e a l w a y s h e l d h i s i n t e r e s t , e s p ec i al ly t h o se p r o b l e m s t h a taf ford the poss ib i l i ty o f in tegrat ing conceptual analys is w i th em pir icalwork . He current ly works in the O f fice o f Research and Development a tNat ional Chengchi Univers i ty in Taiwan .Caleb Lian g is ass is tant professor o f phi losophy a t Nat ional Taiw anUn ivers i t y . H is research in teres t s are in t h e p h i lo s o p hy o f m i n d a n dpercept ion. He recent ly publ ished Phenomenal Character and the My thof the Given (Journal o f Phi losophical Research, 2006).

    69

  • 8/4/2019 Higher Order Thought

    2/30

    Timothy Lane and Caleb Liang

    thought (Rosenthal 1991, 2002, 2004, and 2005). More specifi-cally, Rosenthal (e.g., 2002, 408-11) hypothesizes that mentalstates a re conscious ju st in case they are the objects of occurent,assertoric, seemingly noninferential thoughts t o the effect th atself is in said state. The relevant kinds of higher-order thoughtsneed not themse lves be conscious, bu t to be su it ab le inRosenthals terms, they should not seem t o be mediated,2 not bethose t ha t we a re merely disposed to have, an d not be thosethat are just imagined, wondered, hoped, o r d e ~ i r e d . ~Rosenthal ha s repeatedly made it clear th at his theory isproposed as an explanation of, among other things, phenomenalconsciousness4 (henceforth, p-consciousness). More specifically,his principal goal is to explain what i t is in vi rtue of whichconscious states differ from mental states that arent conscious(2005, 3). He has also emphasized t hat HOTs, ju st like otherkinds of thoughts, can misrepresent. But several authors (e.g.,Byrne 1997; Neander 1998; Rowlands 2001; Seager 1999; andLevine 2001) have expressed doubts about many claims madeon behalf of HOT theory, including the claims that concern p-consciousness and misrepresentation. One worry is that, sincethe HOTs themselves are usually unconscious, how could it bethat they a re able to give rise to there being something-it-is-likefor the subject t o be in a sensory state? A rather hyperbolic wayof expressing this concern has sometimes been called theproblem of the rock. On Goldmans (1993, 366) version: A rockdoes not become conscious when someone has a belief about it.Why should a first-order psychological state become conscioussimply by having a belief about it? A standard response fromthe HOT theorists is that , in order to be a conscious s ta te , th eobject of HOT must be a mental s tate (Rosenthal 2005; Gennaro,2005). Thoughts about rocks dont make them conscious,because rocks arent mental states.Another worry combines the claims made concerning p-consciousness and misrepresentation. This is the suspicion thatif the content of the first-order st at e can be represented withlit tle or no fidelity by HOTs, then the first-order st at e seemsirrelevant to p-consciousness. Yet more vexing, it seems possiblethat HOTs can be the source of p-conscious states, even whenthe first-order st at es a re completely lacking; thus, we a re leftwith the seemingly contradictory claims that first-order sensorystates must be accompanied by HOTs in order for there to bestates that are p-conscious, along with the claim that HOTs canbe sufficient for p-consci~usness.~hat the sensory qualities offirst-order st at es a r e sometimes unnecessary is sometimesreferred t o as the problem of targetless higher-order thought(cf. also Kriegel 2003; Janzen Forthcoming).Rosenthal partially addresses these worries in his discussionof such cases as wine tasting, parafoveal vision, the filling inof blind spots, dental fear, the-subject-as-target, and confabu-

    70

  • 8/4/2019 Higher Order Thought

    3/30

    Higher-Order Thought and the Problem of Radical Confabulationlation. But we believe th at his responses remain inadequate. Inthis paper, we argue that the problem of targetless higher-orderthoughts does pose a serious problem for the HOT theory. Wediffer from other critics, however, in t h a t while t hey focusprincipally on conceptual issues,6 our focus is on explanatoryadequacy, a fail ing that we regard as more damaging toRosenthal because he regards HOTS as posits of an empiricaltheory, and it is a theory th at has been adopted, withmodifications, and promoted by members of the scientific~o m r n u n i t y . ~u r choice of focus is motivated by a desire toengage Rosenthal on that field of endeavor for which his theorywas specially designed. Below, we proceed as follows: section 1briefly describes how Rosenthals theory attempts to explain p-consciousness. Section 2 presents what we call the problem ofradical confabulation, a condition that can be manifest eitherwhen first-order content is completely misrepresented or whenthere simply is no relevant first-order state. We intend to showtha t this is an empirical issue for the HOT theory, an issue thatthe theory-in light of its stated purpose--cannot dismiss lightly.Section 3 considers some possible responses.

    1.Rosenthals version of HOT is intended as an empirical theoryth a t can , among othe r things, explain p-consciousness, thewhat-it-is-l ike aspect of conscious experience.8 First, heemphasizes that not all mental states are conscious; it ispossible for someone to undergo a qualitative or sensory statewithout being conscious of i t (Rosenthal 1997, 2002, 2005). Hereis a recent formulation of the claim:

    To be qualitative, a property need not always occur consciously; itmu st simply be able to occur consciously. Nonmental, physiologicalproperties, by contrast , ar e never conscious. Qualitative propertiesar e potentially conscious, not invariably or essentially conscious.(2005, 177)

    Sensory qualities-understood as th e properties of certainstates th at enable us to discern similarities and differences(e.g., Rosenthal 2005, 202)-as evidenced by subliminal percep-tion, peripheral vision, blindsight, th e cocktail party effect,headaches th at come an d go duri ng the day, and other suchphenomena-seem to be able t o occur without p-consciousness(Rosenthal 2002, 411h9They seem to be mental states, not merephysiological ones, because thei r functional roles parallel th efunctional roles of conscious sensory states . The distinctionapplies to both perception and to bodily sensations. A s for thelatter, jus t as a conscious pain might cause me to adju st myposture, so too would an unconscious pain cause me to adjust

    71

  • 8/4/2019 Higher Order Thought

    4/30

    Timothy Lane and Caleb Liangmy position while I sleep.O Rosenthals preferred pain exampleis of having headaches for extended periods, even thoughintermittent distractions often make us seemingly unaware ofthe pain. Rosenthal would say tha t during those nonconsciousperiods, the headache continues t o play a functional role th atbefits a mental state. As these examples show, sensory qualitiescan be teased apart from what-it-is-like.What makes a sensory s ta t e p-conscious, according t oRosenthal, is that it is accompanied by a suitable HOT. He says:accompanying HOTs do result in there being something it islike for one to be in states with those sensory qualities (2002,413) . A recent elaboration of the same claim is: How Irepresent to myself the sensation I have determines what itslike for me t o have i t. Differences in my HOTs res ul t indifferences in what its like for me t o have my qualitativestates (2005, 187).These sta tem ents seem to show th a t Rosenthals theoryattempts t o explain p-consciousness in terms of two components.First, there must be a first-order mental state. In the case ofperceptual experience, the mental s ta te would be a sensorystate, one th at could perform perceptual functions, even were itnot conscious. Second, there must be a suitable HOT thatmakes the subject conscious of it. By Rosenthals theory, theHOTs cant carry all the weight in explaining p-consciousness.As he says, Strictly speaking, having a HOT cannot of courseresult in a mental states being conscious if that mental statedoes not even exist (1997, 744).This claim seems t o imply thatfor there t o be p-consciousness, the HOTs must target a sensorysta te. When they do so, they make us conscious of ourselves asbeing in certain qual i ta t ive s ta tes , which results in the subjectiveimpression of conscious mental qualities (Rosenthal 2002, 413).We believe this is Rosenthals view, properly understood. Let uscall this account a two-componen t view of p-consciousness.By way of adducing support for this view, Rosenthal assertstha t the re is a str iking connection between what HOTs weare able t o have and what sensory qualities we are able t o beaware o f (2002, 413-14; also see 1997, 745). Reasoning byinference-to-the-best-explanation, he claims th at those HOTsare what enable us to be in conscious states with just thosesensory qualities. If this line of reasoning is correct, it wouldthen seem t o follow that learning new concepts for, say,gustatory and olfactory experiences as when one is learning t obe a wine taster, usually enables us t o be conscious of morefine-grained differences, among sensory qualities.Rosenthal furt her observes th at w hat is tr ue for finelydifferentiated qualities is equally true for crudely individuatedqualities. Taking the sound of an oboe as an example, hehypothesizes th at one can construct a scale of HOTs from thefinely differentiated t o the crudely undifferentiated: we can, for

    72

  • 8/4/2019 Higher Order Thought

    5/30

    Higher-Order Thought and the Problem of Radical Confabulationexample, move from classification of sensations as the sound ofan oboe, to classification as a woodwind, to some yet moregeneric type of sound, and then-yet more crudely undiffer-entiated-to something not even distinctive of sound (perhapsju st indiscr iminat e sensory experience). If we continue toextrapolate beyond th is point, peel away the least differen-tiated HOTs so t o spe ak, then we would be left without anyconscious experience whatsoever (2002, 13-14).The point isthat, without a suitable HOT, all we have would be unconscious,auditory, mental states.An important implication of this apparent two-componentview is tha t i t allows for th e possibility t ha t a HOT can mis-represent th e sensory sta te one is in. And perhaps this shouldnot come as a surpr ise for, af ter all, our HOTs, our occurentasserti ons, par ticu larly when the y seem to be isolated frominferential access to the external world, can go seriously astray.Thus, we can form erroneous HOTs that make it seem, fromthe first-person perspective, as though we are in sensory stateswe arent actually in. One example of this is provided by thecapacity of HOTs to compensate for the low-resolution sensationsof parafoveal vision; HOTs can make those sensations seemclear and focused (see Rosenthal 1997, 744). n such cases, thelow-level resolutions of first-order s ta te s are misrepresented;the first-order sensory state is there, but i t is made t o seemmore distinct and focused tha n it actually is.One among the other examples provided by Rosenthal (2002,415) s the phenomenon sometimes known as den tal fear. Insuch cases, dental patients seem to be experiencing pain evenwhen nerve damage or anesthetic prevents them from being inthe actual sensory state. The usual explanation of th isphenomenon is that a fearful or anxious reaction to vibrationscaused by the drill is what leads one to feel the pain, pain t hatis subjectively indistinguishable from pains felt when the actualsensory s ta te obtains. According to th e HOT theory, wha thappens is that one occurently believes that one is in a painstate. If the patient is later told that, for example, an anestheticha s already been administered, the patient will cease feelingpain. Information thu s provided can lead to inferences t ha talter the subsequent subjective experiences. But thatinformation doesnt change the patients sense of what the priorexperience was like, for conscious experiences are not altered byinferences based upon su bsequent information. Rosenthalclaims that the HOT theory effectively explains this phenome-non: the vibrations sensed by the patient serve as first-orderstates , while the subjective feeling of pain is brought about bythe HOTs having misrepresented those first-order s ta tes in away consistent with the patients beliefs about the commonconsequences of dentistry. This proposed explanation seems tobe compatible with a two-component interpretation of HOT.

    73

  • 8/4/2019 Higher Order Thought

    6/30

    Timothy Lane and Caleb Liang

    But we must hasten to add that, strictly speaking, Rosenthalsaccount is open t o another in terp reta t i~n .~e claims th at notonly can HOTs misrepresent their targets, they may even beabout something th at does not exist a t all (2005, 210). HereRosenthal might be understood as suggesting that even thoughfirst-order sensory states are frequent causal antecedents forHOTs, they are not essential. Wishful thinking, self-deception,confabulation, and other like phenomena can also serve ascausal antecedents and allow for the possibility of HOTs thatare about notional perceptions or sensations, perceptions orsensations th at dont exist. Rosenthal is careful t o emphasizevarious constrain ts on the degree t o which HOTs can divergefrom targets (212),but, in the end, he allows that , there can besomething its like for one t o be in a st at e with particularmental qualities, even if no such state occurs (211).HOTs havethe las t word. Call th is a one-component interpretation of theHOT theory; it will be discussed below, in the final section.Others have criticized Rosenthals position concerning theclaim tha t HOTs can target nonexistents and still bring aboutthe what-it-is-like of p-consciousness. A s Kriegel observes, theclaim th at a person can be under the impression that she is ina conscious state when in reality she is not is, t o say the least,highly counterintuitive (2007, 49). Rosenthal does not denythat his claim is counterintuitive; instead, he treats thisreaction as just a pr et he ~r et ic ~imitation on our intuitions, alimitation tha t is not as troublesome as it might seem. Pre-theoretic experience, after all, is not a good guide to the causalantecedents of consciousness of self as being in a quali tativestate, nor should we expect it to be.We stand with Rosenthal on th e claim th a t pretheoreticintuitions should not be taken as cause for excessive worry.Moreover, as we have indicated above, our concern residesprimarily with explanatory adequacy, not with alleged problemsof conceptual coherence. It is for this reason that we havedevoted special attention t o Rosenthals examples, for it is heretha t explanatory adequacy can be bet ter assessed. We claimthat Rosenthals examples (e.g., parafoveal vision, tasting wine,or listening t o an oboe), a t least those th a t can more clearlybeen seen t o serve an explanatory role, presuppose a two-component view that suffers from explanatory deficiencies.This completes our brief introduction t o Rosenthals accountof how p-consciousness can be explained. Despite its brevity, webelieve it accurately identifies the critical features of hisaccount. In the next two sections, we argue that his view fails t oprovide a satisfactory explanatory account of the phenomenalcharacter of consciousness. Below we will also further considerRosenthals claim that his theory does not, str ictly speaking,require two components.

    74

  • 8/4/2019 Higher Order Thought

    7/30

    Higher-Order Thought and the Problem of Radical Confabulation

    2.In allowing for cases of misrepresentation, perhaps an allowancethat an empirical theory that aspires to appropriate explanatoryscope must make, Rosenthal has created a problem for himself.On the one hand, as mentioned in the previous section, aspects ofhis theory coupled with certain telling examples tend to implythat HOTs cannot, by themselves, account for p-consciousness.But on the other hand, as we will see, in confabulatory cases theHOTs seem, at leas t sometimes, to have a free hand. In th issection, we challenge Rosenthals account of confabulation. Wedistinguish between nonradical and radical cases of confabula-tion and suggest th at his view faces what we call the pr ob l em o fradical confabula t ion . We argue, first, that some of the cases ofconfabulation tha t Rosenthal has discussed-including pa ra-foveal vision and the filling-in of blind spots-are nonradical.For others-for example, de ntal fear-his expla natio ns a reproblematic in ways tha t he seems not to recognize. Second, weargue that there are radical cases of confabulation tha t the two-component view does not adequately explain. Our contention isthat , with regard to p-consciousness, in particular th e what-it-is-like aspect, Rosenthals HOT theory is not empirically adequate.By nonradical cases of confabulation we mean cases that fallunder one (or more) of the following three categories: (a) thosefor which a relevant first-order target s ta te clearly exists, (b)those for which only a less obviously relevant first-order targetstat e exists, and ( c ) those that are perhaps better understood asinstances of epis temic overconfidence. A s for (a), considerparafoveal vision. It might be said that th e HOTs still faithfullyrepresent a f irst-order sensory sta te, but they extrapolateerroneously, exaggerating it s coverage. The same might be saidfor such phenomena as the filling-inof blind spots. No contentalien to the target s tate need be added.

    A s for (b), consider Rosenthals account of d en ta l fear onceaga in. According to th e two-component view, the vibr atio nssensed by t he patient ar e taken to be first-order states, whichare misrepresented by th e HOTs as pain. (The role of anxietyhere, while not intuitively or medically surp ris ing, does notcomfortably fit into the HOT model of misrepresentation.) Thiscase is not as straightforward as ei the r parafoveal vision orunnoticed blind spots. Mis representat ion seems to be doingmore here th an it is in either of the previous cases. Here i t canbe said that the HOT has a target, but the content of tha t targetst at e seems less relevant to wha t is represented. Unlike thetr ea tm en t given to parafoveal vision o r blind spots, the HOTcant simply extrapolate from first-order content; the HOT mustac t i ve ly , c r ea t i ve l y misrepresent. Vibrations cum fear are notpain,15 but the HOT misrepresents them as pain. Or so i t seems,on Rosenthals description.

    75

  • 8/4/2019 Higher Order Thought

    8/30

    Timothy Lane and Caleb LiangA s for ( c ) ,what we call epistemic overconfidence, no suffi-ciently distinct example can be found within Rosenthals work,so we here cite a common textbook case: according to Nisbettand Wilson (1977), when subjects were invited to choose from avariety of panty hose displayed in a row, they preferred those ontheir right. But their preference explanations made no referenceto sp ati al orientation; instead, th e subjects spoke of such(seemingly phenomenal) qualities a s color and texture. Evenafter being told th a t th e pant y hose were identical a nd theywere simply showing a preference fo r those on their right-handsides, subjects tended to defend their confabulations.The reason for treating ( c ) as distinct from (a) and (b) is thatalthou gh subjects may report phenome nal qualities, i t i s

    somewhat less clear that they are actually experiencingphenomenal qualities. Many empirical studies of confabulationdo not strongly o r consistently suggest th a t wha t is confabu-lated is consciously experienced. Subject reports might seemconsistent with a HOT explanation-that is, the HOTs might besaid to misrepresent visual and haptic continuity, leading thesubject to be conscious of discontinuity. But belief constructionof this sort need not imply anything at all about p-conscious-ness, as various clinical and technical definitions of confabulationseldom include specific reference to p-consciousness (HirsteinS o as not to be misunderstood, we should make it clear th a twe agree th at Rosenthal should try to explain misrepresentationas i t might apply t o p-consciousness; after all, as exemplified byconfabulation, it is a common human phenomenon, exhibited inboth mundane, nonpathological cases, and in exotic, pathologicalones. It might jus t be a basic fea tu re of the way we form a ndrepo rt beliefs, a form o f epistemic overconfidence (Hi rst ein2005; Lane 2006). As for th e pa nty hose case, if we p ut th eworries mentioned above aside, the subjects might be reporting

    judgments tha t a re grounded in p-consciousness in a way thatis consistent with wha t Rosenthal says about HOT theory:despite being wrong about th e cause of their behavior, theycould be said t o be conscious of color in virtue of having HOTsth at t arge t certain sensory state s, bu t said st at es seem to beradically misrepresented, for the y a re mak ing distinctionswhere the re a re no distinctions to be made. In th is case, th efirst-order state does not seem to carry any of the content th a tis being represen ted by th e HOT, at least none of the contentthat enables the making of distinctions. S o i t is unlikeparafoveal vision t h a t simply builds upon wh at i s a lreadyavailable t o it . It sees differences where there ar e none to beseen. I t doesnt misrepresent by indicating more of the same;rat her , i t misrep resents by indicating difference despitesensory continuity. As with dental fear, we seem to have a casewherein HOTs can be said t o have targets, but if we presuppose

    2005, 187-203).16

    76

  • 8/4/2019 Higher Order Thought

    9/30

    Higher-Order Thought and the Problem of Radical Confabulation

    th a t t he subjects reports a re reliable an d li teral , thei r p-consciousness includes sensory qualities (i.e., the discontinuitiesin color and texture) that dont allow for the type of nonradicalextrapolation available to the cases considered above.We have suggested that the two-component view, whenapplied t o dental fear and to t he panty hose cases, misrepre-sen ts in ways th at create a constructivist burden for HOTS, aburden not shared by blind spot or parafoveal vision cases. Butwe will not press the point closely here because we believe thatneither represents a substantial challenge for HOT theory. Forexample, dental fear, since the re i s a target sta te, and sincefear and vibration bear significant conceptual similarities topain, a plausible, relatively straightforward defense of HOTcan certainly be managed: the relevant HOT does target first-order sensory states, but these s ta tes are misnamed. Fear cumvibration is unpleasant, as is pain; perhaps patient s are ju stmisnaming their unpleasant experiences in the way thatsubject s who can dis ting uish among colors readily enoughmight incorrectly name them. As for th e Nisbett an d Wilsonpanty hose case, perha ps an a rgum ent could be made th atdespite references t o p-consciousness, there is no evidence thatth e subjects actually had such experiences. Pe rhaps t hei rbehavior can be entirely treated at the doxastic 1 e ~ e l . l ~e arenot fully convinced that such is the case, but we choose to avoidthese complexities fo r now. Instead, we use these constructivistexamples as a bridge or transition to examples that we believeserve as yet more serious challenges t o the explanatoryambitions of HOT.Accordingly, we now turn to radical confabulation, by whichwe mean cases of radical misrepresentation, radical in that (a)there is good reason to say th at the re simply is no target, yet(b) there are robust grounds for claiming that p-consciousnessactually obtains. Rosenthals theory not only allows for mis-representation by extrapolation from target content and bycreative additions to target content, as we have indicated above,it also allows for the possibility of radica l misrepresentation:th at is, cases in which those first-order sensory sta te s simplydont exist (e.g., 2004, 31). However, he does not believe thatthis poses serious problems for his theory. Addressing this issue,Rosenthal writes:

    Suppose my higher-order awareness is of a stat e with property P ,but the ta rget isnt P , but rather Q . We could say that the higher-order awareness misrepresents th e tar get, but we could equallywell say that its an awareness of a s ta te tha t doesnt occur. Themore dramatic the misrepresentation, the grea ter the temptationto say the target i s absent; but its plainly open in any such case tosay ei ther. The two kinds of cases, moreover, should occasion thesame kinds of phenomenological perplexities, if any. A higher-order

    77

  • 8/4/2019 Higher Order Thought

    10/30

    Timothy Lane and Caleb Liangawareness of a P sta te without an y P st at e would be subjectivelythe same whether o r not a Q state occurs. The first-order st at e cancontribute nothing to phenomenology apart from th e way wereconscious of it (2004, 32).*

    Rosenthal seems to be antici pating our distinction, or likedistinctions, between nonradical and radical cases of confabu-lation and downplaying its significance. On one reading of thispassage, it might seem th a t Rosenthal i s trying to distanc ehimself from a two-component interpretation of HOT. But webelieve his attempt to downplay the possibility of targetlessHOTs is significant. Note t hat he writes: We could say that thehigher-order awareness misrepresents th e target, bu t we couldequally well say that its a n awareness of a sta te t ha t doesntoccur. The more dramatic the misrepresentation, the greater thetemptation to say the target is absent, but its p l a i n l y o p e n inany such case t o say either (Rosenthal 2004, 32; emphasisadded). We believe tha t Rosenthal can make such a claim onlybecause he deals with nonradical cases. The radical cases, weargue, are importantly different: they show th at i t is notplainly open . t o say either.lgCase 1.Consider t he case of Antons Syndrome,20 n pa rticula r, th edenial of blindness (Hirstein 2005, 14546). ypically those whosuffer from Antons Syndrome ( A S ) have two brain lesions: onethat affects vision and another th at affects the ability to knowwhether or not they can see. The result is tha t although theirneural mechanism cannot process visual stim ulation , the ybelieve they can, and they can skillfully confabulate when theirfailure t o see things accurately is called to thei r a ttentio n.Confidence in confabulations is not s ha ke n even when th epatients ar e interacting with the environment; naturally theybump into things and encounter other problems, but theyconfabulate anyway, for example, by complaining that t he roomis too dark, or that theyre not wearing their glasses, or t h a ttheyre not familiar with the environment.Here, although we seem to have an example of HOTswithout sensory experience, a defender of HOT might, as withth e case cited above, att em pt to re strict explanation to th edoxastic level. But a yet more s trikin g characteristic of AS isth at patien ts often have visual hallucinations (th at can beeither simple or complex). Typically, hallucinations, as opposedto mere confabulations, are said to have three characteristics:21(a ) they are vivid and immediate, like perceptions; (b) they areexperienced as though external to the person; and ( c ) hey arenot reassessed as imagery, jus t because evidence suggestingthey should be so reassessed is provided. These three charac-

    78

  • 8/4/2019 Higher Order Thought

    11/30

    Higher-Order Thought and the Problem of Radical Confabulationter isti cs might al so be applicable to confabulatory caseswherein relevant first-order ta rge t s ta te s exist; they might,arguably, even be applicable to such cases wherein a lessobviously relevant first-order target st ate exists.22But clearlythey a re not applicable to cases of epistemic overconfidence.And t he problem posed for HOT is t hat (a), (b), and (c) seem toobtain, as clear indicators of p-consciousness, even when th esubject suffers from AS.In one case, for example, a patient with AS not only reportedhallucinations (while suffering from delirium t r e m e n ~ ) , ~ ~e alsobelieved his vision had been restored t o the extent that he couldprovide descriptions of his environment (Swartz and B ru st1984). Unlike those cases for which defenders of HOT might bebetter able t o restrict explanation to doxastic problems, here wehave better evidence that p-consciousness, in the form ofsubstant ial visual experiences actually occurs. Here then weseem to have a case of radical confabulation for which i t isdifficult t o deny that p-consciousness obtains, but for which theHOT seems to be doing all the work. In a word, the HOT seemsto be s u f f i c i e n t for p-consciousness. This vio late s the two-component interpretation of Rosenthals view; it does not seemplainly open to interpretation as target misrepresentation.The problem here is th a t not only are th e HOTs misrepre-senting and confabulating, as in the case of dental fear or in thecase of choosing panty hose, there seem t o be grounds for sayingthere is no mental state to be targeted; yet the subject still seemscapable of p-consciousness. In the cases of A S accompanied byvisual hallucinations, although the patient claims to have visualexperiences, apparently the HOTs c a n n o t b e t a r g e t i n g thesensory states in question.A defender of HOT theory might want t o claim th at here toowe have insufficient evidence for the existence of p-consciousness.Perhaps explanation can once again be handled by reference tointentional states alone. But we would regard such a responseas ad hoc, because the diagnosis of hallucination presupposes p-consciousness, a nd t ha t diagnosis i s made on independent,medical grounds. Notice we are not suggesting th at the reportsof empirical scientists should be taken at face value. But we areclaiming th at t he burden to correct a diagnosis rest s with thecritic who encounters an anomalous case that does not conformto theory expectations. Moreover, HOT is designed t o explainempirical phenomena, s o unless a principled reason can begiven to reject the standard scientific description, the diagnosisand its conceptual implications should stand.What we are claiming is that i f Rosenthals version of HOTis both intended as an explanation of p-consciousness and asallowing for the possibility of radical confabulation, then it flirtswith inconsistency. On the one hand, it seems t o imply thatdoxastic s ta tes alone-even HOTS-are insufficient as a n

    79

  • 8/4/2019 Higher Order Thought

    12/30

    Timothy Lane and Caleb Liangexplanation of consciousness. The other mental sta tes, in th isinstance th e first-order sensory sta tes , seem to be necessary.But on the other hand, the HOT theory also seems anxious toallow for radical confabulation, and thi s makes th e first-ordersensory sta te s unnecessary. We fu rth er contend t ha t radicalcases are not ambiguous in such a way as to support Rosenthalsclaim th at they could plainly be reinterpreted as instances oftarget misrepresentation.This is not a result that Rosenthal could want, for even heseems t o share the widespread intuition th at intentional states,alone, simply dont have the resources for enabling phenomenalconsciousness (1997, 740; 2002, 413). It is being able to formintentional sta tes about certain sensory qualities that resultsin our being able to experience those qua lit ies consciously(2002, 413). But here there are no sensory qualities tha t can beintentionally targeted. At least for the radical cases, it seems tobe inconsistent to allow for the possibility of erroneous HOTSalone giving rise to something-it-is-likeor the subject t o be in asensory state.One immediate response would be t o t reat it as a n excep-tional case, something that is beyond th e reach of the existingversion of th e theory.24But actually this might not be such anexceptional case. There a re other examples in the litera ture onpathologies that suggest that p-consciousness can occur despiteth e absence of the sor t of target t hat is required by the HOTtheory. To cite just two more cases:Case 2.Congenitally deaf patients who suffer from schizophrenia oftenclaim to he ar voices (see Atkinson 2006). Schizophrenia is ascommon among the deaf as i t is among the general population,and approximately half of these patients report hearing voices.The voices are described as being similar to regular speech, int ha t the y can vary along multiple dimensions, includingloudness, pi tch, content, an d complexity. Moreover, voices a reoften personified such t h a t t he pat ien ts can detect accent,gender, and degree of familiarity. It must be admitted that thereare many contentious issues concerning the proper interpre-tation of these da ta ; for example, it i s not altogether clear howwe should understand hallucination descriptions that use signsglossed in English as Heard, or Voices. But at least this is anindication that cases of AS accompanied by hallucinations arenot s o exceptional as they might seem. In some respects, thephenomenon of deaf-hearing might be even more problematicfor Rosenthal because these patients dont believe that they canhear. They claim to hear, despite not believing t ha t they haveth e capacity for hea ring . On the face of it, they seem t o belacking an essential prerequisite for formation of the relevant

    80

  • 8/4/2019 Higher Order Thought

    13/30

    Higher-Order Thought and the Problem of Radical ConfabulationHOT-the belief that they are capable of hearing. Nevertheless,they do, with the onset of schizophrenia, begin t o hear voices.Contrast this with Rosenthals wine tasting example. In th atcase, it is alleged that learning new concepts enables one to beconscious of fine-grained differences among sensory qualities.Concept learning takes a leadership role. The same can be saidconcerning Rosenthals thought-experiment wherein it is allegedth at by systematically stripping away concepts, the sound of anoboe can be reduced to indiscriminate sensory experience. Inthese cases, concepts are enablers. But when the deaf suddenlystart to hear, t he concepts dont lead or enable, at least not inth e way t hat i t seems to be with wine or oboes. When the deafbegin to hear, their concepts seem capable of nothing more thanmerely responding t o phenomena th at abruptly ass ert them-selves onto th e scene. And t hi s is especially odd because thephenomena are in direct conflict with previously held beliefsconcerning what the agent is capable of experiencing.Case3.Ramachandran discusses the cases of phantom limb pati entswho experience involuntary clenching spasms, such tha t nailscan be felt digging into the palm, spasms tha t inflict great painth at the patient cannot relieve (see Ramachandran 1998;Ramachandran and Rogers-Ramachandran 1996). In order toprovide therapeutic assistance to these patients, Ramachandrandesigned a mirror-box, a box that contains a centrally placed,vertical mirror, a box without a top or front. Patients can placetheir existing hands into the box, thereby creating the illusionth at the phantom ha s been resurrected. As a consequence ofth is resurrection, six out of ten pa tien ts claimed they couldnow feel th e phantoms move and four of the ten were able tounclench their the hands, thereby relieving the pain.

    On a t least one description of such cases, patients who dontbelieve in the resurrection of the phantom hand ar e suddenlyable to exercise control over it and eliminate select p-consciousexperience (i.e., the nails digging into th e palm). That is, thereseems to be no intentional target, at least no obvious sensoryst at e available for targeting. And the re is no belief t ha t th ehand ha s been resurrected. Yet p-consciousness is nonethelesst r a n ~ f o r m e d . ~ ~Once again, contrast this with Rosenthals wine and oboeexamples: were occurent, assertoric, noninferential thoughtsplaying a significant explanatory role in the generation of p-consciousness, one might reasonably expect conceptual additionor subtraction to be leading the way. But here th e phenomenon-the illusion that the hand ha s been resurrected-trumps. Atbest, if there are accompanying HOTS, they seem t o just followthe illusion, no matter where i t leads.26

    81

  • 8/4/2019 Higher Order Thought

    14/30

    Timothy Lane and Caleb Liang

    Although it might be possible for the HOT theory to accom-modate such phenomena, still these cases differ strikingly fromRosenthals paradigmatic cases. And those paradigm cases arewhat provide the HOT theory with much of its initial plausi-bility. They suggest a significant explanatory role for HOTs. Inour cases, however, it seems difficult, or a t le as t pointless, toleave explanatory room for HOTs.On our view, in light of th e cases described above, Rosenthalwould seem to be hard pressed to continue to proclaim thatHOT is a promising empirical theory. The point is not t hat goodtheories dont encounter anomalies; they do. The point is tha tHOT ha s not a s yet establish ed any significant empiricalcredentials. Examples like that of wine tasting are suggestive,but suggestive evidence of this sort is a fa r cry from th e sort ofevidence that growing empirical theories require. Such theoriesreq ui re th e evidence provided by confirmed p redictio ns orevidence provided when attempts t o falsify fail. Given thatexis ting evidence is no more th a n weakly suggestive, the seanomalous cases would seem to constitute a serious th reat toall at tem pts to invoke HOTs as pa rt of the explanans for anygiven instance of p-consciousness.In t he nex t sectio n, we consider some fur th er possibleresponses to our criticisms of HOT theory. We argue th a t noneof them succeed in alleviating the worries raised by the problemof radical confabulation.

    3.In this section, the following objections are considered: (1) heconfabulatory cases can be reinterpreted; (2) the counterexamplesexaggerate the relevance of peripheral or input mechanisms; (3)HOTs might be targeting something mental, even if that some-thing is not a first-order mental state; (4) p-consciousness mightfail to occur in the proposed counterexamples; and, (5) Rosenthalcould opt f o r a disjunctive s trat egy , one th a t allows for th epossibility th a t HOTs a re sufficient. We ar gue th a t each ofthese fails.

    (1)Reinterpreting cases of confabulation? A first responseth a t might be at temp ted on behalf of Rosenthal is t o t ry toreinterp ret some cases of confabulation. For example, as fordental fear, a case that we treat as problematic, a defender of theHOT theory might try to argue that i t is significantly analogoust o standard instances of misrepresentation, like the manner inwhich similarity of shape, impaired vision, and distance mightcause one to mistake a cow for a horse. Fear and vibration-orperhaps a philosopher should write fear and trembling-mightbe relevantly analogous to shape and impaired vision.But even if this attempt succeeds, it still leaves us to wonderon just what grounds this can be called an account of p-

    82

  • 8/4/2019 Higher Order Thought

    15/30

    Higher-Order Thought and the Problem of Radical Confabulationconsciousness. For t he less contentious cases, like parafovealvision, at least there is a relevant, substantial, first-order sensorystat e with which to work. Perhaps, as we have suggested above,some similar account could be developed for dental fear: therelevant HOTs in this instance would actually be about fearand vibration, bu t these sensations would then be misnamed.Because fear cum vibration is unpleasant, just as is pain, suchmisnaming might be as common as is the misnaming of similarcolors or sounds. We submit t ha t den tal pa tien ts, generallyspeaking, are not proficient at naming the different types ofexperienced unpleasantness that they undergo, so they tend touse a catchall term like pain.AS accompanied by hallucinations, however, presents a muchmore substantial worry by eliminating even the possibility of ananalogue to shape, impaired vision, misnaming, and so forth. Insome respects it seems th at Rosenthal has not yet put to restthe problem of the rock. Our concern is not that beliefs dontseem t o be the sort of entity that could explain consciousness;actually, we suspect t hat worries about explanatory gaps andhard problems are excessively motivated by the desire to feelthat one understands, and this is an inappropriate expectationt o hold concerning empirical explanations. As Hempel (1965,256-58) once observed in a similar context, the desire to feelthat we understand is just to confuse empathetic familiaritywith theory-based or cognitive understanding.Doxastic states may well have all that is required to causewhat-it-is-like, but once we leave behind paradigmatic caseslike wine tast ing and oboe listening-cases th at imply a two-component view-it is not at all clear th at such states ar e ableto play a n explanatory role. Although at first glance i t mightappear that AS accompanied by hallucinations might lendsupport t o a one-component view, such is not necessarily thecase. One can describe t he phenomenon i n te rm s consistentwith the view that HOTs alone are sufficient for what-it-is-like,but one can ju st as easily describe it in terms compatible withthe view t ha t HOTs a re eit her i rrelev ant or an outcome ofevents over which they yield no control-for example, despiteth e belief t ha t one cant see, one suddenly starts t o see, a neffect t h a t se ts the confabulatory m achine ry of HOTs intomotion. Concerning the relationship between conceptual consis-tency a nd exp lanatory adequacy, we will have more to s aybelow.Par t of th e confusion concerning th e explanato ry role ofHOTs is, we submit, that in his more recent discussions of HOTRosenthal tend s t o minimize talk of causal connection^,^'preferring to use th e term accompany. But if accompany isinterpreted as correlation, something th at is not uncommonfor Rosenthal (e.g., 2002, 413), then i t is not at all clear jus twhat kind of explanation can be forthcoming. Accidental correla-

    83

  • 8/4/2019 Higher Order Thought

    16/30

    Timothy Lane and Caleb Liang

    tions ar e cheap; hence, explanatorily useless. They a re the typeof thi ng th a t sta tisti cs faculty like t o enterta in thei r under-gra du ates with, for example, as with citing inst ances of thecorrelation between stock marke t trends an d womens sk ir tlengths. Unfortunately, Rosenthal has not yet clearly indicatedjust what kind of nonaccidental correlations are involved here,2sand that is wh at we would need were accompany to play anexplanatory role.More substantial interpretations may well be forthcoming,but at least at this stage in th e development of the HOT theorywe have good grounds fo r asking in what sense it could be th a tHOTs accompany when either they wholly misrepresent whatthey are apparently about, or when they target n o n e x i s t e n t ~ . ~ ~What is, at minimum, required is a more systematic account ofaccompany, either when we should expect the two componentsto correlate, and how or when we should expect that they wont,an d what consequences will follow. Rosenthal doesnt intendtha t accompany should apply to cases wherein HOTs in ten-tionally target nonexi~tents,~~ut the worry here, for those of usconcerned with explanation that is applicable t o th e empiricalsciences, is t hat accompanyn can be dismissed too easily, in anad hoc fashion. And if this is the case, then Rosenthals attemptto avoid conceptual issues pertaining t o the hard problems andexplanatory gaps fails because his avoidance is justified by theclaim t h a t he is concerned wit h scientific explanation , bu tscientific explanation cant get off the ground if it allows for thead hoc and relies upon the post hoc.(2) Exaggerated emphasis on peripheral or inpu t mechan-isms? A defender of HOT might also proclaim that our cases ofradical confabulation only seem to challenge HOT because wefocus unduly on sensory inpu t m echanisms. The arg um entmight go that people can be blind o r deaf in different ways.Perhaps it could be said that although they have lost sensorycontact w ith th e world, their in te rn al sensory capabilitiesremain intact. Perhaps i t could be argued tha t these internalsensory mechanisms are mental s tates that can serve asappropriate HOT targets.To take t he cases of AS as examples, perhaps it could besaid that patients are ju st mistaking visual imagery for actualvision. I t might t he n be open t o a defender of HOT theory toclaim th a t visual imagery provides the targ et for a HOT. Butmost clinical neurologists believe th is is not th e case becausethe damaged are as ar e precisely those a rea s th at ar e neededfo r both imagery and actual vision.31 Stil l, clinical cases varygreatly, s o perhaps the general point does need to beaddressed.Rosenth al says, TObe qualitative, a property need notalways occur consciously; it must simply be able to occurconsciously. Nonmental, physiological properties, by contras t,84

  • 8/4/2019 Higher Order Thought

    17/30

    Higher-Order Thought and the Problem of Radical Confabulationare never conscious (2005, 177). Rosenthal defines sensorys ta tes as possessing properties that enable us to makedistinctions and form classifications, s ta te s th a t have t hepotential t o be conscious. A states having qualitative character... is solely a matter of the role that s ta te plays in perceiving(Rosenthal 2205, 13) .The first th ing to observe here is t ha teven if certain nonconscious, neurophysiological states obtainindependently of sensory contact with the world, since they areplaying no perceptual role, it is by no means obvious that thosesta tes can count as mental, something th at would be requiredon Rosenthals view.32Whats more, identifying something asmental because it has the potential t o be conscious seems torisk flirtation with higher-order dispositional views,33views thatRosenthal rejects.34A third thing t o observe is that , according to R ~ s e n t h a l , ~ ~frel eva nt perceptual roles never existed, we would have nogrounds for the ascription of mental qualities. I t would thenseem t o follow that at least for the congenitally blind victims ofA S, we would lack adequate grounds for the ascription ofrelevant, nonconscious, sensory states. Homomorphism theorydoesnt require the persistent function of perceptual roles, but itdoes require that those roles were previously performed.A fourth thing t o observe is that once sensory contact withthe world has been lost, the sense in which i t can be said t hatwe make distinctions and form classifications is less clear.During dream sta tes , for example, when sensory inpu t isblocked, although internal perception proceeds and emotionsmight be enhanced, cognitive functions a re grea tly alter ed(Hobson 2007, 106-7). Perhaps it could still be argued th a tfirst-order sensory distinctions and classifications are still beingmade, but when one hallucinates (as in sensory deprivation) orwhen one dreams, since cognitive functions are altered, it is notclear ju st what role HOTs are able to play in the generation ofconscious sta te s. At th e very leas t we would need fu rthe rdiscussion concerning the nature and role of HOTs when thesenses are deprived or when the agent is asleep. Even worse forthe HOT the orist, if inter nal perception is unimpaired butcognition is distorted, it might be that Rosenthal and otherswould be pushed toward a higher-order p e r c e p t i o n , o r innersense, view of consciousness. And that is a view that Rosenthal(2005, 340) adamantly rejects.(3) Might HOTs still be targeting something mental, even ifthere is no first-order state? After all, on Rosenthals account, astates being conscious consists in ones being conscious of selfas being in that state. Rosenthal writes: Each HOT representsits target st at e a s belonging t o some individual (2005, 347).Perhaps then p-consciousness could still be said t o result fromth e targe ting of something menta l. But in ste ad of a s t a t eserving as the target, a subject would so serve.

    85

  • 8/4/2019 Higher Order Thought

    18/30

    Timothy Lane and Caleb LiangBut even were we to gra nt that either state o r subject couldsuffice, it is not clear tha t the theory can s tand, at least not onRosenthals t erms. Concerning the ce ntr al notion of self hewrites: A minimal concept of self will suffice for reference tooneself; no more is needed t han a concept that allows distin-guishing between oneself and other things. Such a minimalconcept need not specify what sort of thing the self is. Thus itneed not imply th at the self has some special sort of unity, o r isa center of consciousness, or is transparent, o r even tha t i t ha smental propert ies (1997, 741; emphasis added). Rosenthal alsowri tes tha t t he single subject which were conscious of ourconscious sta te s as belonging t o may not actually exist (2005,348). In other words, according to HOT, neither targeted state

    nor subject need actually exist.If there is no target sensory state , nor any self with mentalproperties, then just what is left to accompany th e HOT? Weknow that a rock would be insufficient. But when we considerRosenthals minimalist, even eliminativist, unde rst andin g ofself, it isnt clear th at there is enough substance here to enablep-consciousness, that is, unless the HOT really is doing all thework, unless HOTs really are sufficient for what-it-is-like.Alternatively, he might tr y to incorporate a more robustsense of self. But doubtless such a response would resurrect oldworries about HOT, particularly those raised by Dretske (1995)and others who are concerned that HOT theory excludesprelinguistic infants, feral children, a nd nonhum an anima lsfrom th e community of cre at ure s capable of phenomen alconsciousness. Even if Rosenthal were to bite this bullet, itwould by no means be an easy task t o formulate a theoreticallyfeasible concept of self that has th e resources to play the rolethat he hopes for within HOT theory.(4) Might it be that HOTs are sufficient because p-conscious-ness doesnt occur in the alleged counterexamples? Rosenthalmight concede that for cases of radical confabulation the first-order targ ets ar e indeed missing, while at the same t imedenying t h a t p-consciousness occurs. He could claim t h a tsubjects first -per son repo rts about p-consciousness reflectnothing but epistemic overconfidence.But, as we observed above, the problem here would be thatRosenthal would have t o deny th e neuroscience and clinicalsta ndard s for interpreting pa tie nt behavior (including verbalbehavior) as indicating the actual experience of p-consciousness.This might be possible, but the burden for rejection of standardempirical characterizations re st s with proponents of the newexplanatory theory; in th is in stance , with the proponents ofHOT. Presumably, were t hi s to succeed, it would r equi re asignificant deflationary view of p-consciousness, a view thatsome others have already found appealing.

    86

  • 8/4/2019 Higher Order Thought

    19/30

    Higher-Order Thought and the Problem of Radical ConfabulationEspecially noteworthy in this regard is Dennett (e.g., 1991,369-41 l),who has invested much effort toward deflatingstandard characterizations of p-consciousness. But, despite themany similarities between their two positions, Dennett (e.g.,1991, 362-68) believes intentional states alone (understood asbeliefs, o r interpretations, or as memeplexes) are sufficient toexplain consciousness. He sees no need t o posit two distinctmental levels; even when he turns his attention toconsciousness, he takes beliefs to be the pretheoretical dat a,the quod erat explicatum (2005, 4-45).Were Rosenthal to recas t HOT in th is form, he might bebette r positioned to seek an explanatory scope tha t does notomit cases of radical c~ n f a b u l a t i on . ~ ~implifying somewhat, if

    HOTs carried all of the load, all of the time, then there wouldbe no need t o mark distinctions between radical and nonradical,no need to mar k d istinct ions between mere epistemic over-confidence and cases in which robust p-consciousness actuallyoccurs. However, moving in that direction would mean giving upany role for first-order mental s tates, which would imply thatwe are never really in certain folk-psychological sta tes that wetake ourselves t o be in; it would commit Rosenthal to denial ofa distinction between the appearance and reality of conscious-ness. Obviously, such moves would mark a substantial departurefrom Rosenthals current position (e.g., 2005, 231-56).Still, theadvantage of such moves would be that Rosenthal would not betroubled by differences between ordinary ins tances whereinHOTs make self aware of being in a certain sta te and instancesof radical confabulation. On Dennetts view, there simply is nosubstantial difference-in a sense, all cases are cases of radicalconfabulation because we ar e never in the states th at we takeourselves t o be in.But a move in this direction would also be rife with problems.Recall, for example, the manner in which a simple trick withmirrors can lead to the resurrection of a phantom hand. Whatmight a theory that puts all explanatory weight onto HOTs beable t o say about th is? Pa tients dont report occurent beliefsthat their hands are resurrected. Quite the opposite. Neverthe-less, the therapy works. I t seems th at Rosenthal (or Dennett)would have to posit a nonconscious belief in the handsresurrection, a belief that conflicts with the conscious beliefthat no such thing is possible. We wont pursue this matter anyfurther here, beyond just to say that Rosenthal would owe us aprincipled account of how conflicting HOTs are, or can be,adjudicated. Although according to the HOT theory, most HOTsare nonconscious, stil l the theory does allow for HOTs t o beconscious, when they themselves are targeted by yet higher-order thoughts. Accordingly, what we need is a n account t hatexplains why the nonconscious HOT-which is nothing morethan a theoretical posit-wields more influence than the con-

    87

  • 8/4/2019 Higher Order Thought

    20/30

    Timothy Lane and Caleb Liangscious, repo rtab le HOT-which is someth ing t h a t we havetheory-independent evidence for.One additional, related problem that we will mention, butreserve full development for another time, is what we term th eproblem of conceptual inadequacy. Consider the case of peoplewho suffer from pain asymbolia syndrome: although they claimt o be experiencing pain, they sa y they dont care, th a t theyseem to float above the pain.37Our worry, baldly stated, is thatno societys folk psychology contains the concept of pains thatone can float above. For those people who know nothing ofpain studies, who suddenly find themselves afflicted with painasymbolia, the phenomenon ju st asser ts itself without warning.These people wouldnt have th e conceptual resource s wi thwhich t o form th e relevant HOT. They ar e dramatically unlikepeople who cult ivate a n appreciation of wine, color, or music.The worry then for those who would adopt a move in th isdirection-that is, placing yet grea ter explanatory weight onbeliefs o r on the HOT-is that in cases like pain asymbolia, th ehigher-order thought seems irrelevant, for it lacks the conceptualresources whereby such a distinction could be made. Folkpsychology doesnt dist ingu ish between pain s t ha t we mu stwallow in and pains that we can float above.(5) Finally, one might defend the HOT theory by suggestinga d i s j u n c t i v e strategy. Earlier, in section 1, we noted th a tRosenthals theory is open to a one-component interpretation,according t o which, as Rosenthal says, there can be somethingits like for one to be in a st at e with particular mental qualities,even if no such s ta te occurs (2005, 211). On this interpretation,contrary to the two-component view, the HOTs alone a re takento be sufficient for p-consciousness. A disjunctive strategy wouldallow Rosenthal t o employ a two-component version to accountfor the ordinary cases and nonradical confabulation cases, alongwith a one-component version to account for radical confabula-tion cases. The idea is that , on the one hand, Rosenthal can agreethat in radical cases there ar e no first-order mental s tates, sincethe one-component view allows HOTs to play a sufficient role inexplaining p-consciousness; on the other hand, he can maintainthe two-component view for the rest of the cases.The first problem is that this disjunctive st rategy seems adhoc. Rosenthals (e.g., 2005, 187-88) paradigm cases-forexample, the wine taster who becomes conscious of fine-graineddifferences among sensory qualities by learning new concepts-ar e all bound to th e two-component version. Even the cases ofmisrepresentation th a t he cites, as we have argued above, arebound to the two-component version: recall, typically misrepre-sentations are extrapolations from first-order sensory states. S othese cases do not motivate the strategy.Moreover, the empirical scientists who invoke RosenthalsHOT theory in constructing explanations also work with a two-

    88

  • 8/4/2019 Higher Order Thought

    21/30

    Higher-Order Thought and the Problem of Radical Confabulationcomponent view. Weiskrantz (1997, 71-76; 1997, 203-4), forexample, in proposing explanations of various neuropsycho-logical deficits, invokes HOT and sees it as largely consistentwith his views on conscious experiences, but Weiskrantz doesnot treat HOTs as sufficient; instead, he sees HOTS-or some-thing similar to HOTS-as being only necessary for the sort ofconscious experience t h a t i s lacking in cases of blindsight,unilateral neglect, amnesia, aphasia, and so forth. Likewise forthe research of Dienes and Perner (1999,2001,and 2002) who,despite noting that Rosenthal claims HOTs can be sufficient,only apply his analysis to cases in which first-o rder sta te sactuall y obta in. And Rolls (2001, 245-65) develops a viewdubbed HOLT (higher-order l ingu i s t i c thought), a view tha tdiffers from Rosenthal only in placing g reater emphasis onsyntactic manipulation. But nothing in Rollss discussion orapplication suggests tha t he regards HOLTs as sufficient; hishypothesis is th at consciousness is the st at e which arises byvirtue of having th e ability to think about ones own thoughts,which ha s th e adapt ive advantage of enabling one to correctlong multistep syntactic plans (2001, 258). This is a version ofthe two-component view. He nowhere considers the possibility ofthin king about nonexistent thoughts , nor does he const ructexplanations of th is sor t; instead , his focus is clearly on themanner in which the higher-order and lower-order interact.Rosenthal (2005, 179) also reports on the research of Frithand Fri th (1999), research that includes reference to brainimaging studies of subjects asked to repor t on thei r men ta lstates. F ri th and Frit h record t ha t, despite wide variation inthe nature of the s tates reported on, activity was observed in allthese studies ... along th e border between ro stra1 ant eriorcingulated cortex and medial prefrontal cortex (1999, p. 1693).Subjects were asked to monitor and report on pain, emotions,spontaneous thoughts, actions, and tickling. Rosenthalspeculates th at the brain a rea activated in all these instancesmight be that which subserves HOTs. But even if this is true, itdoes not lend support to the sufficiency claim because the Frithan d Fri th d at a all concern monitoring of actu al first-orderstates.What these empirical cases suggest is th a t a disjunctivest ra te gy would be ad hoc. Paradigm cases-cases used tomotivate serious consideration of HOT that actually offer someexplanatory, predictive leverage-invoke two components. Eventhe cases of nonradical misrepresentation are best explained bya two-component view. Whats more, the empirical scientistswho invoke HOTs in their explanations, or whose research isregarded by Rosenthal as being suggestive of the role that HOTscan play in explanation, are working with a two-component view.The one-component view has not been shown to be empiricallywell motivated; therefore, the disjunctive strategy seems ad hoc.

    89

  • 8/4/2019 Higher Order Thought

    22/30

    Timothy Lane and Caleb LiangA second problem, one int ima tely re lat ed to t he ad hocproblem, is t h a t t he disjunctive st rat egy does not obviouslyprovide any explanatory or predictive advantage. Although onth e surface it may seem t ha t AS accompanied by hallucinationor schizophrenia in the congenitally deaf can be explained byth e one-component version of HOT theory, perhaps reminiscentof the way in which Newtonian theory predicted the existence ofNeptune and Pluto, such an interpretation would be misleading.The one-component version-markedly unlike the case wi thNewton-leaves absolutely no specific expectations concerningwhen targetless HOTs that enable p-consciousness shouldoccur.38Wishful thinking, self-deception, and other standardcases of confabulation-that is, typical causes of misrepresen-

    ta tion (e.g., Rosenthal 2005, 125-26)-provide no reason forexpecting AS accompanied by hallucinations o r schizophrenicvoices in th e congenitally deaf. Most people who ar e blind knowthat they ar e blind and they have no p-conscious visual experi-ences to suggest otherwise; standard forms of wishful thinkingor self-deception provide no relief from their blindness, althoughdelirium tremens might. The same can be said for the congeni-tally deaf: neither wishful thinking nor self-deception help themto hear, only schizophrenia does. Accordingly, no t even ou rexamples can be taken to suggest that HOTs play a sufficientexplanatory role.In other words, the most tha t could be claimed on behalf ofthe disjunctive strategy when applied to our cases is th a t theone-component view is conceptually consistent with them. Butconceptual consistency can be purchased cheaply. The failure toindicate ju st when we should expect wishful th inki ng, self-deception, o r some other factor to give rise to what-it-is-likeindicates that the one-component view does not obviouslycontribute anything substantial to the explanatory status of theHOT theory.

    A third problem is that the disjunctive strategy seems not tobe falsifiable. A virtue, albeit an ironic one, of the two-component view is that it is falsifiable. We say ironic because,if we are correct, it now st an ds a s falsified by our counter-e x a m p l e ~ . ~ ~mployment of the disjunctive strategy, even if adhoc in t he sense of being unmotivated by paradigm cases orempirical research and even if it provides no obvious groundsfor expecting th e phenomena th at occur in the counterexamples,can at leas t contribute to th e development of an explanatorytheory, if it is cast in such a form as to be potentially falsifiable.But the one-component disjunct of th e disjunctive s trategy isnot so cast. What we need, and do not yet have, is a reasonablyclear stat em en t of the conditions under which th e disjunctivestrategy could be empirically challenged.As we have said above, critics (e.g., Kriegel 2007) have foundsome of Rosenthals claims to be counterintuitive, especially his

    90

  • 8/4/2019 Higher Order Thought

    23/30

    Higher-Order Thought and the Problem of Radical Confabulationclaim th a t people can be unde r t he impres sion of being inconscious s ta te s th at they a re not actually in. Importantly,Rosenthal agrees that such claims are counterintuitive. Becausehe cannot use intuition t o justify these claims, he must justifythem in some other way. The way he chooses is justification interms of a theory t ha t he believes t o have sound empirical,explanatory credentials. And thi s is a way th at h e needs inorder to fend off worries pertaining to the alleged hard problemand the explanatory gap. But we a re here showing tha t theHOT theory is not justified in claiming such credentials foritself. Therefore, the HOT theory is confronted with a seriouschallenge.In sum, although it might seem that a disjunctive strategywould enhance the HOT theorys explanatory adequacy, such isnot the case. Explanatory adequacy is not enhanced when theadded option-in th is case, the one-component option-is notempirically motivated by the paradigm cases o r by the way inwhich working scientists actually employ the theory. Ad hocadditions dont help. Second, although the addition might wellbe conceptually consistent with our cases of radical confabu-lation, mere consistency doesnt aid explanation. Third, the two-component view has the virtue of falsifiability; the one-component view, in its present form, does not. Obviously wehave treated just lightly some difficult issues in the philosophyof science, but our intent is simply to show that the disjunctivestrategy is suspect. To remove the shadow of suspicion from thisstrategy, at least if Rosenthal wishes to continue to promote theHOT theory as an explanatory theory for empirical science, hemust accept th e burden of showing tha t the one-componentversion is not ad hoc, that it provides explanatory advantage, orthat it is falsifiable.

    4.We conclude t h a t radica l confabulation pre sen ts a seriousproblem for Rosenthals HOT theory. There seems t o be animportant body of anomalous cases for which the HOT theory-either the two-component version or the disjunctive strategy-fails t o provide any exp lanatory advanta ge. We wish toemphasize th at our inte nt i s not destructive. We sh ar e anadmir ation for the effort t h a t is expended i n rising to thechallenge of consciousness and seeking to devise an empiricaltheory th a t aims to handle even those aspects tha t a re oftentreated as being beyond the reach of science. Our task, as weapproach it here, is to examine the explanatory adequacy of theproposed theory. Our hope is t ha t the discovery of anomaliesleads to fruitful revision, the design of a more adequate theoryof broader scope, enhanced consistency, and greater explanatorypower.40

    91

  • 8/4/2019 Higher Order Thought

    24/30

    Timothy Lane and Caleb Liang

    NotesThe order of authorship was determined arbitrarily; this manuscript iscompletely collaborative.

    In th is paper we focus on Rosenthals version of HOT theory. Forother versions of HOT theories of consciousness, see Rocco Gen naro2004.Th at we a re not conscious of any media ting factors should not betaken to imply that there ar e no mediating factors.Natura l language is nonesse ntial to t he formation of HOTS;hence, prelinguistic infants, feral children, and nonhuman animals a rein principle capable of conscious experience. See Rosenthal 1991, 472-73, and 1997, 741-42.Rosenthal does though make a n impo rtant distinction here: t heclaim is th at some mental sta tes , those th at exhibit th in phenome-nality, have qualita tive properties, even though t he re is nothing t ha tit is l ike for one to have said properties. Mental sta tes t ha t exhibitthick phenomenality, by contra st, are st at es th at have q ualita tiveproperties along with a what-it-is-like for one to have th e experience(e.g., 2005, 190-92). Subliminal perception and the cocktail party effectar e offered as illustr ations of how, empirically, qualit ative st at es canbe independent of what-it-is-like. Below, whenever we use p-conscious-ness, we inte nd t ha t i t be understood in Rosenthals sens e of thickphenomenality.Rosenthal does not endorse thi s characterization of his position;we further explain our position below.We dont intend to belittle th e significance of conceptual issues,bu t they trouble all scientific theories, an d those theories-quantummechanics, evolution by natural selection, and so on-continue togenerate fruitful inquiry in t he domains for which they were designed.Prominent among such investigators have been Rolls (2001, 244-65),Weiskrantz (1997, 71-76), an d Dienes and Perner (e.g., 1999). Wewill ret ur n to consideration of thei r work below.Of st at es th at possess the property of phenomenal consciousness,it is often said th at it is l i k e someth ing to have them (Farrell 1950;Sprigge 1971, 167-68; Nagel 1974). Notational va rian ts of what-it-is-like-to-be include raw feels, qual ia, experience, a nd subjective feel.Many hold t h a t t hi s prope rty can only be defined ostensively, sucht h a t it can only be pointed to i n experience, as when you expla insting by saying how your hands feel when you hi t a fastbal l off th ehandle or th e end of the b at on a cold day. A s a more formalalternative to ostension, Carruthers defines phenomenal consciousnessas event s t h a t we can recognize i n ourselves, non-inferentially, orstraight off, in vir tue of the ways in which they feel to us, or the waysin which they present themselves to us subjectively(2000, 14).To tak e ju st one example among these, pain, a property that istypically regarded as being th e most highly accessible to a subject, it isstri kin g ju st how many investigators of consciousness, rep rese ntingvaried theoretical perspectives and ar ea s of specialization, countenanceta lk of unconscious, unnoticed, unexperienced, unfel t, or sub-clinicalpains (e.g., Car rut her s 2005, 185-86; Chalmers 1996, 17; Dartnall2001; Guttenplan 2000, 28; Searle 1992, 164-67; Tye 1995, 115 and2000, 182;Vertosick 2000,152 and 175;Wilkes 1993, 186).And thoughit is often said th a t th ese a re incoherent or counterintuitive ways of

    92

  • 8/4/2019 Higher Order Thought

    25/30

    Higher-Order Thought and the Problem of Radical Confabulation

    thinki ng about pain , Lycan observes th a t ordinary people quitefrequently speak of pains t ha t go unfelt , without an y sense ofcontradiction (2003, 9 and 13). To convince skeptics, h e ha s evenbegun to compile examples of unfelt pa in from th e popular press.lo This example is frequently cited by Dennett (e.g., 1991, 61; 1996,13 and 95; and 1998, 351).According to Rosenthals homomorphism theory, m en ta lqualities ar e properties of states th at figure in the perceptual function-in g of a part ic ular sensory modali ty , and whose s imilar i t ies an ddifferences are homomorphic to those which hold among th e propertiesperceptible by that modality. Rosenthal claims that the homo-morphism theory provides the theoretical grounds for distin-guishingamong th e conscious, th e nonconscious but menta l, an d t he merelyphysical.l2 Dennetts (1991, 30-31, 337) observations concerning palate

    training for wine tas ters , aural t raining for musicians , and theexperiences of an app ren tice piano tu ne r mi rro r Rosenthals views.Rosenthals and Dennetts theories converge in many respects.Concerning their appa rent differences see Rosenthal 2005, 321-35 andDennett 2000.l 3 In both personal communication an d in recent publications (e.g.,2005, 209-13), Rosenthal ha s explicitly pressed this point.l4 Rosenthal has emphasized this point in personal communication.l 5 Those of us who live in earthq uake-p rone regions ar e perhap smore easily persuaded by th is claim th an those who do not. We havegrown accustomed to not infrequent vibrations accompanied by fear,and t he two combined ar e indeed painless. To those who dont live inearthq uake-p rone regions, perhaps a ride on a roller coaster wouldsuffice to persuade you.l6 The same is tr ue for well-studied cases of th e confabulation ofintent ion (Wegner 2002, 171-86); aft er t he completion of actions,people commonly revise what t hey th ink they intended, an d th er e isno necessary reference to p-consciousness.l7 One of th e explanatory challenges th a t H O T seems, as yet ,unabl e to respond to is th e need to dis t inguish between dis t inc texplananda: beliefs that merely make reference to what-it-is-like an d

    beliefs t ha t hav e a n ac tua l what-it-is-like aspect. More will be sa idabout this below.la Cf. also Rosenthal 2005, 211, where he makes t he same pointconcerning dental fear.In personal communication, Rosenthal has conceded tha t mattersmight not be so plain here. Again, we take th is matter to be significantbecause, among other things, it seems to suggest tha t t he HOT theorycarries more explanatory weight when tar get s exist, i.e., when th er eare two components. And we believe that radical cases of confabulationshow that the two-component view fails.2o Also known as th e Anton-Babinski Syndrome.21 Our discussion of hallucinations and pseudohallucinations in thisand in the next paragraph is based on Davis 2004.22 We dont concede this point; we simply choose not to press ithere.23 The patient in question did recognize some of his hallucinationsas hallucinations. That a hallucination is recognized as such by th esubject does not imply tha t it is inconsistent with t he three criteria for

    93

  • 8/4/2019 Higher Order Thought

    26/30

    Timothy Lane and Caleb Liang

    hallucinations indicated in the text.24 At one point, Rosenthal e nte rta ined th e possibility t ha t targ et-less HOTs would be both ra re and pathological. He ha s since retrea tedfrom th at position (2005, 29). Below we t ry to show tha t he was correct

    to revise the claim tha t such cases ar e rare.25 Yet more difficult cases could easily be cited here, e.g., the case ofcolor-blind color vision among synesthetes who see numbers as tingedwith hues (Ramachandran and Hubbard 2003, 57). But each caseinvolves distinctive complexities, an d our only intention at thi s pointis to suggest tha t ther e is a family of cases that seriously challengethe explanatory efficacy of HOT theory.26 If we a re describing thi s phenomenon accurately, th is case tooseems to lend support to a two-co mpone nt view, albe it a versionwherein HOTs a re passively compliant to t he demands of first-ordersensory states.27 Rosenthal identifies thi s transition of his ideas in several places,e.g., 2005, 56n25.28 Rosenthal does propose an evolutionary account of how it cameto be th at HOTs tend to accurately reflect actual first-order sta tes(2005, 15-16, 218-19, 303-5). He arg ues t ha t HOTs evolved as ameans of at tr ibuting mental s t ates to others , at tr ibutions tha t a re

    based on behavioral observations and other stimuli; this t alen t for th eat t r ibut ion of menta l s ta tes was then turned inward . But th isarg ument doesnt help hi s case her e because it provides no reason toexpect HOTs th at would be involved in enabling th e deaf to hea r orth e blind to see. Evolved capacities can fail to be adaptive, because theenvironment or informational inpu ts change (Buller 2005, 57). But onth e HOT theory we have no reason to believe th at either ha s changedin such a way as to cause radical confabulation.

    29 I t i s wor th men t ion ing th a t the re is a s ignif icant tensionbetween responding to \t he problem of the rock by claiming th at HOTstar get m ental s ta te s while simultaneously claiming th at HOTs cantar get nonexistents. How could targ eti ng a nonexistent give rise towhat-it-is-like when targeting a rock cannot?30 This is a point th a t he ha s emphasized in personal communi-cation.31 Our discussion here is drawn from Churchland 2002, 122-23.32 In o ther words, we ar e saying th at Rosenthal ha s not yetsucceeded in putt ing to rest t he problem of the rock.33 For a d efense of one version of t h e disp osit ion al view, seeCarruthers 2000 and 2005.34 Rosenthal concedes th a t when me ntal sta te s ar e conscious, itseems phenomenologically tha t the re is no HOT present (2005, 111).But h e emphasizes t ha t t he appearance is irrelevant, because HOTsar e posited as th e best explanation of what it is for a mental st at e tobe a conscious state. And th is is yet anoth er reaffirmation t ha t t hecase for-or against-HOT needs to be made in te rm s of explanatory

    adequacy.35 Personal communication.36 It should be noted thou gh th a t Dennetts resi stance to tal k ofmen tal levels might reflect more a difference i n terminology th an insubstance (e.g., Rockwell 1996).37 Here we a re quoting Carruthers (2000, 206).38 The point is not t o deny th e occurrence of improbable events;

    94

  • 8/4/2019 Higher Order Thought

    27/30

    Higher-Order Thought and the Problem of Radical Confabulation

    rather, it is merely to say th a t when offering a n explanation, somereason for expecting th e improbable event should be s tat ed, perhaps,for example, by describing relevant causa l mechanisms (e.g., Railton1993).39 We dont intend to deny Duhemian and other concerns pertainingto f alsification. We app eal to i t here merely as an (admit tedlyimperfect) constraint on the development of explanatory theories.*O We hereby express our heartfelt gratitude to David Rosenthal forhis careful reading of and detailed comments on previous drafts of thispaper. We also wish to express ou r gra titude to anonymous refereeswho motivated us to make apprpriate revisions. Moreover, we aregra tefu l to par t ic ipants in the 2007 LMPS Taipei Conferencesponsored by the Nation al Taiwan Univer sity (April 27-28) and toparti cipan ts in th e Soochow University Conference on AnalyticPhilosophy (June 20-23, 2007) for their man y helpful comments.

    Especially worthy of note a re several constructive suggestions offeredby Richard Fum ert on, E ric Pen g, Wen-fang Wang, an d Huei-Ying(Tony) Cheng. We have not agreed wi th or been persuaded by everycomment or suggestion, but our unders tandin g of th e subject matt erhas been grea tly enriched through these stimulating exchanges.References

    Atkinson, Joanna R. 2006. The perce ptual c harac teris tics of voice-hallucinations in deaf people: Insights into the na tu re of subvocalthought and sensory feedback loops. Sch izophren ia Bu l le t in 32 (4):

    Buller, David J. 2005. A d a p t i n g m i n d s : E v o l u t io n a r y p sy c h o lo g y a n dByrne, Alex. 1997. Some like it hot: Consciousness and higher-orderCarruthers, Peter. 2000. P h e n o m e n a l c o n s ci o u sn e s s: A n a t u r a l i s t i c

    .2005. Consc iousness : Es say s f r o m a h igher -order perspec tive .Chalmers, David J. 1996. T h e c o n sc io us m i n d : I n s ea rc h o f a f u n d a -Churchland, Patricia Smith. 2002. B r a i n - w i s e : S t u d i e s in n e u r o -Dartnall, Terry. 2001. The pain problem. Philosophical Psychology 14Davis, Derek Russell. 2004. Hallucination. In The O x ford companion tot he m i n d , ed. Richard L. Gregory. 2nd ed. Oxford: Oxford UniversityPress.Dennet t, Daniel. 1991. Consciousness explained. Boston: Litt le, Brownand Company..1996. Kin ds of min ds: Toward a n understand ing of consciousness.

    New York: Basic Books.. 1998. B r a i n c h i l d r e n : E s s a y s o n d e s i g n i n g mi n d s . Cambridge,MA: MIT Press..2000. With a little help from my friends. In Dennetts philosophy:A c o m p r e h e n s i v e a s s e s s m e n t , ed. Don Ross, Andrew Brook, andDavid Thompson. Cambridge, MA: MIT Press.

    701-8.the pers is tent ques t for h u m a n nature . Cambridge, MA: MIT Press.thoughts. Philosophical Studies 86:103-29.theory. New York: Cambridge University Press.New York: Oxford University Press.me ntal theory. New York: Oxford University Press.phi losophy. Cambridge, MA: MIT Press.(1): 95-102.

    95

  • 8/4/2019 Higher Order Thought

    28/30

    Timothy Lane and Caleb Liang

    Dennett , Daniel. 2005. S w e e t d r e a m s : P h i lo s o p h ic a l o b s ta c l es t o ascience of consciousness. Cambridge, MA: MIT Press.Dienes, Zoltran, a nd Josef Perner. 1999. A theory of implicit andexplicit knowledge. Behavioral and Brain Sciences 22:735-55..2001 . When knowledge is unconscious because of consciousknowledge and vice versa. In P r oc e ed in g s o f t h e T w e n t y - T h i r dAn nu al Conference of the Cogni t ive Science Socie ty , ed. J. D. Mooreand K. Stenning. Mahwah, NJ: Lawrence Erlbaum Associates.. 2002 . Developmental asp ect s of consciousness: How m uchtheory of mind do you need to be consciously aware? Consciousnessand Cogni t ion 30.Dretske, Fred. 1995. N a t u r a l i z in g t h e m i n d . Cambridge , MA: MITPress.Farrell, B. A. 1950. Experience. M i n d 59:170-98.Frit h, Chris D., and U ta Frith. 1999. Interacting minds-A biological

    Gennaro, Rocco, e d. 2004. H i g h e r - o r d e r t h e o r i e s of c o n s c i o u s n e s s .. 2005. The HOT theory of consciousness: Between a rock and a

    Goldman, Alvin. 1993. Consciousness, folk psychology, an d cognitiveGuttenplan, Samuel. 2000. Minds landscape. Oxford: Blackwell.Heilman, Kenneth M. 1991. Anosognosia: Possible neuropsychologicalmechanisms. In A w a r e n e s s o f d e fi c i t a f t e r b r a i n i n j u r y : C l i n i c a l

    a n d t h e o r e t i c a l i s s u e s , ed. George Pr igatan o, 53-62. New York:Oxford University Press.Hempel, Carl G. 1965. Aspects of scienti fic explana tion. New York: FreePress.Hirstein, William. 2005. Br ai n f ic tion: Sel f -decept ion an d the r iddle o fconfabulation. Cambridge, MA: MIT Press.Hobson, J. Allen. 2007. Normal an d abnormal s ta tes of consciousness.In The Blackwel l com panion to consciousness , ed. Max Velmans andSusan Schneider. Malden, MA: Blackwell.Jan zen , Greg. Forthcoming. Phenomenal chara cter a s implicit self-awareness. Journal o f Consciousness S tud ies .Kriegel, Uriah. 2003. Consciousness as intransitive self-consciousness:Two views a nd a n argu ment. C a n a d i a n J o u r n a l of P h il o so p h y

    . 2007. Philosophical theories of consciousness: ContemporaryWestern perspectives. In Cam br idge handbook o f consc iousness , ed.M. Mosocovitch, E . Thompson, an d P. 0. Zelazo. New York: Cam-bridge University Press.Lane, Tim. 2006. T-Belief. Pap er pre sen ted a t th e Conference onNaturalized Epistemology and Philosophy of Science at SoochowUniversity, May 3 1 J u n e 3.Levine, Joseph . 2001. P u r p l e h a z e : T h e p u z z l e o f c o n s ci o u sn e s s. NewYork: Oxford Press.Lycan, William G. 2003. Dretskes ways of introspecting. In Privilegedaccess: Phi losoph ical a ccounts o f se l f -kn ow ledge , ed. Brie Gertler.New York: Ashgate.Nagel, Thomas. 1974. What is it like to be a bat? Morta l Ques t ions .Cambridge University Press.Neander, K. 1998. The division of phenomen al labor: A problem for

    basis. Science 286:1692-95.Philadelphia: John Benjamins Publishing Company.ha rd place? Journal of Consciousness Studies 12 (2): 3-21.science. Consciousness and Cognit ion 2:364-82.

    33~103-32.

    96

  • 8/4/2019 Higher Order Thought

    29/30

    Higher-Order Thought and the Problem of Radical Confabulation

    representat ional theories of consciousness. In L a n gu a ge , m i n d , a n dontology, ed. James Tomberlin, 411-34. Oxford: Blackwell.Nisbett, R. E., and T. D. Wilson. 1977. Telling more tha n we can know:Verbal reports on mental processes. Psychological Review 84:231-59.Railt on, Peter. 1993. Probability, explanation, a nd information. InE x p l a n a t i o n , ed. David-Hillel Ruben. New York: Oxford UniversityPress.Rama chandran, V. S. 1998. Consciousness and body image: Lessonsfrom phantom limbs, Capgras syndrome and pain asymbolia.P h i l o s o p h ic a l T r a n s a c t i o n s o f t h e R o y a l S o c i e t y of L o n d o nRamachandran, V. S., and Edward M. Hubbard. 2003. Hearing colors,tasting shapes. Scient if ic Am erican 5:53-59.Ramachandran, V. S., and D. Rogers-Ramachandran. 1996. Synesthesiain phantom limbs induced with mirrors. Proceedings o f the Royal

    Society of Londo n 263:377-86.Rockwell, Teed. 1996.Awareness, mental phenomena, and consciousness.Journal of Consciousness Stud ies 3:463-76.Rolls, Edmund. 2001. The bra in and emot ion . New York: Oxford Univer-sity Press.Rosenthal, David M. 1991. Two concepts of consciousness. In T h ena ture o f mind , ed. David Rosenthal. New York: Oxford UniversityPress.. 1997. A theory of consciousness. In The nature o f consciousness ,ed. Ned Block, Owen Flanagan, a nd Guven Guzeldere. Cambridge,MA: MIT Press..2000. Metacognition and higher-order thoughts. Consciousnessand cognit ion 9:231-42..2002. Explaining consciousness. In Philosophy of min d: Classicala n d c o n t e mp o r a r y r e a d i n g s , ed. David J. Chalmers, 406-21. NewYork: Oxford University Press..2004. Varieties of higher-order theory. In Higher-order theorieso f consc iousness , ed. Rocco J. Gennaro, 17-44. Philadelphia: JohnBenjamins Publishing Company.. 2005. Consciousness and m in d. Oxford: Oxford University Press.

    353:1851-59.

    Rowlands, Mark . 2001. Consciousness a nd higher-order tho ughts.Seager, William. 1999. Theories of consciousness: An i n t r o d u c ti o n a n dSearle, Joh n R. 1992. Th e red iscovery o f m in d . Cambridge, MA: MITSprigge, Timothy. 1971. Final causes. Proceedings o f th e Aris to te l ianSwartz, B. E., and J. C. Brust. 1984. Antons Syndrome accompanyingTye, Michael. 1995. Ten prob lem s of consciousness:A representationalTye, Michael. 2000. Con sciousness, color, an d conte nt. Cambridge, MA:Vertosick, Frank T. 2000. W h y w e h u r t : T h e n a t u r a l h i st o ry o f p a i n .Wegner, Daniel M. 2002. Th e i l lusion of conscious wil l . Cambridge, MA:Weiskrantz, Larry. 1997. C o n s c i o u s n e s s l os t a n d f o u n d : A n e u r o -

    Mind and Language 16 (3): 290-310.assessment . New York: Routledge.Press.Society 45:149-70.withdrawal hallucinosis in a blind alcoholic. Neurology 34:969-73.t he or y o f p h e n o me n al m i n d . Cambridge, MA: MIT Press.MIT Press.New York: Harcourt.MIT Press.

    97

  • 8/4/2019 Higher Order Thought

    30/30

    Timothy Lane and Caleb Liangphysiological exploration. Oxford University Press.thought experiments. New York: Oxford University Press.Wilkes, Kathleen V. 1993. R e a l p e o p l e : P e r s on a l i d e n t i t y w i t h o u t