PHILOSOPHICALPSYCHOLOGY , VOL. 15, NO. 4, 2002

Whenconsciousness matters: acritical review ofDaniel Wegner’s The illusionof consciouswill

EDDY NAHMIAS

ABSTRACT In The illusion ofconscious will ,DanielWegner offers an exciting, informative,and potentiallythreatening treatise on the of action.I offerseveral interpretations of the thesis thatconscious willis anillusion. Theone Wegner seems to suggestis “modularepiphenomenalism” : conscious experience ofwill is produced by abrainsystem distinct fromthe system that produces action;it interpretsour behavior but does not,as it seemsto us, cause it.I arguethat the evidence Wegnerpresents to support this theory, though fascinating, is inconclusive and, in any case,he has notshown that conscious willdoes notplay a crucialcausal role in planning,forming intentions, etc. This theory’s potentialblow to our self-conception turns out to be a glancingone.

1.Introduction: a blowto our self-image Freud claimed thathis theorymarked thethird ofthe “ majorblows at thehands of science”to “the na¨õ ve self-love ofmen.”First, Copernicus moved us fromthe center ofthe universe, thenDarwin knocked us fromthe pinnacle ofcreation, and Žnally, Freud showedus that“ theego is noteven masterof its ownhouse, butmust content itselfwith scanty ofwhat is going on unconsciously in its mind” (1989/1920, p.353). But at least Freud offered some rolefor the conscious will, especially if, throughpsychoanalysis, weunveil theunconscious machinations that move us. Daniel Wegner might beseen as arepresentativeof a fourthscientiŽ c blowto ourself-image, apost-Freudian knockoutblow that says ourconscious will plays no rolein causing ouractions butis, instead, an illusion. His The illusionof conscious will is an exciting, informative, entertaining, and potentially terrifying treatiseon the psychology ofaction. Intheend, Wegner seemsuncertain about howthreatening he means his view tobe. He wants toexposeas illusions whatwe take tobetrue: “The factis, it seemsto each of us thatwe have conscious will. Itseems we have selves. Itseems we have minds. Itseems we are agents. Itseems we cause whatwe do” (p. 342). But healso wants tosuggest thathis explanation ofhowthese illusions are

Eddy Nahmias,Department of Philosophy, Florida State University, Tallahassee,FL 32306–1500, USA,email: [email protected] ISSN 0951-5089/print/ISSN 1465-394X/online/02/040527–15 Ó 2002Taylor &FrancisLtd DOI:10.1080/ 0951508021000042049 528 E. NAHMIAS produced by thebrain allows theman importantplace in ourlives: “Afterall, the apparent mental causation approachdoes not throw out conscious will entirely; it instead explains howthe experience comesto be … Ultimately, ourexperience of conscious will may have morein uence on ourmoral lives than doesthe truth of our behavior causation”(pp. 334, 341). Idon’t think hecan have it bothways. Ifwe accepted his theoryas true, it would threatenour conception of ourselves as freeand responsible (as Wegner recognizes in various places—e.g. in his contrastbetween the positive illusion of controland a“resignation todeterminism” ; pp. 332–333). In any case, Iwill argue thathis thesisis overstated. Wegner has gathered atremendousamount ofinterest- ing empirical workand presentedit in an engaging manner. Despitehis arsenal of empirical Žndings, however,he has notshown that our conscious will is an illusion—at least notin thestrong sensethat says ourconscious experience ofwilling ouractions plays no importantcausal rolein howwe act. Weaker interpretationsof Wegner’s thesis, moreaccurate in light oftheresearch he presents,suggest thatthis blowto our self-image is aglancing one.

2.Interpretations of Theillusion of conscious will Wegner summarizes his main thesisin several differentways, and it can bedifŽ cult todiscern exactly whathe is trying tosay and, hence, howthreatening his thesisis toour sense of self and freewill. Iwill describethree ways onemight interpretthe claim thatconscious will is an illusion. Wegner hints at theŽ rsttwo but argues most clearly fora third interpretation. Iwill argue thatthe evidence hemarshals in supportof this interpretation is unconvincing. Wegner writes,“ The experience ofconsciously willing an action is nota direct indication thatthe conscious thoughthas caused theaction. Conscious will, viewed in thisway, may bean extraordinary illusion indeed”(p. 2). Whatmight thismean?

2.1.Dualism is false? One way tointerpretthe claim thatconscious will is an illusion is thatthe prevalent commonsenseview ofthe mind— what Ryle called the“ OfŽcial Doctrine”— is wrong: thereis nota mental forcethat causes bodily behaviors, no ghostly soul that movesthe machine. So, in thequotation above, themistake would involve our Cartesian beliefthat “ conscious thought”certainly doesn’t seem like aphysical entity butit certainly does seemlike it movesus toact. The illusion would beanalogous to ourold beliefthat the sun revolves around theearth because it certainly seems like weare notmoving and thesun is. Itwould beodd ifdualism is theillusion Wegner is describing, since thisclaim would beuncontroversial tomany ofhis readers, certainly tomostpsychologists and philosophers.Furthermore, though neuroscientiŽc evidence continues towhittle down any rolefor a non-physical soul, thisis notthe sort of evidence Wegner discusses. Yetsometimes it seemshe is trying tomake acase against dualism when heemphasizes thatthe way mental causation appears tous is like Cartesian interac- WHENCONSCIOUSNESS MATTERS 529 tionism—“We intended todo it, sowe did it”(p. 27)—but the way actions are actually caused involves mechanisms about whichwe are unaware. Wegner writes, “peoplehave at hand tworadically differentsystems of explanation, onefor minds and onefor everything else”(p. 21). Weapply thementalistic explanations, with theirreference to beliefs, desires, purposes,and intentions, toour own and others’ behaviors: “weŽ nd it enormously seductive tothink ofourselves as having minds” (p. 26). But, Wegner suggests, such explanations are mistaken in thesense that “ the real causal sequenceunderlying human behavior involves amassively complicated setof mechanisms …[and] themind can’t everknow itself well enough tobe able tosay whatthe causes ofits actions are”(pp. 27–28). Again, thisclaim is uncontroversial toanyone whothinks weare notconscious oftheneural mechanisms in ourbrains. And theclaim thatwe are notconscious of thecauses ofour behavior becomestautological whenhe asserts, “The real causes ofhuman action are unconscious”(p. 97). What is controversial is howto cash out theseclaims. IfWegner is notjust denying theuntenable dualism suggested byour ordinary language ofmental causation, perhapshe is suggesting thatconscious processessimply cannot cause physical behavior.

2.2.Epiphenomenalism is true? Perhapsthe way tointerpret the illusion, then, is thatall conscious mental states simply do nothave any causal inuence on whatwe do; theyare epiphenomenal. The roleof consciousness is illusory, then, because despitehow it seems,the existence ofconscious will has no effectson howwe behave. Itis aside effect,like Huxley’s example ofthe steam whistle that is caused by thetrain’ s engine butdoes notmove the train (cited on p. 317). Wegner suggests thatthe conscious will is caused by the“ engine”of the brain (the“ empirical will”) butdoes not itself move us toact: “Whateverempirical will thereis rumbling along in theengine room—an actual relation betweenthought and action—might in factbe inscrutable tothe driver ofthe machine (themind)” (p. 28). Perhapsconscious will could even be strippedoff of the mind’ s activity withoutthis affecting whatwe do. This strong view is related tothe issues philosophersworry about whenthey discuss theproblem of zombies, absent qualia, and ofcourse, mental causation. One worryis that, if the cause ofbehavior can besufŽ ciently explained by low-level (“unconscious”) mecha- nisms, whatrole is leftfor conscious processes? In fact, Idon’t think Wegner means toraise thismetaphysical worry, despitehis suggestions thatmechanistic explanations are real and mentalistic explanations are illusory. Though his bibliography includes an impressive number ofreferences to worksof philosophy, hedoes not really engage withthe philosophical questions about mental causation. Instead, hedenies thatconscious will is entirely inefŽcacious: “theexperience ofconscious will thatis createdin thisway [by unconscious mechanisms] need notbe amereepiphenomenon. Ratherthan aghost in themachine, theexperience ofwill is afeeling thathelps us toappreciate and rememberour authorship ofthe things ourminds and bodies do”(p. ix; seeChapter 9). Nonetheless,he clearly has in mind somesort of epiphenomenalism: “Why, if 530 E. NAHMIAS thisexperience ofwill is notthe cause ofaction, would weeven go tothetrouble of having it? What good is an epiphenomenon?”(p. 318). So whatdoes he have in mind?

2.3.Modular epiphenomenalism is true The third way tointerpret the claim thatconscious will is an illusion putsan empirical faceon themetaphysical claim ofepiphenomenalism. Itsays thatthe conscious will is explicable in termsof brain processesbut that these brain processes are not“ in theloop” when it comesto producing actions. Itis not thatconscious mental statesin general are epiphenomenal butthat speciŽ cally thosethoughts and intentions weexperience justbefore actions as thecause ofthose actions do notin factcause ouractions. Furthermore,the conscious experience ofwill is notentirely impotent—it “helpsus appreciate and rememberwhat we are doing”(p. 325), and it is “theperson’ s guide tohis orher own moral responsibility foraction” (p. 341). (In offeringthese roles for conscious will, Wegner thinks hecan dampen theblow ofhis thesisand even provide an explanation offreewill— but see discussion below.) Wegner thussuggests whatI will call “modular epiphenomenalism.”He de- scribesconscious will as amodule orsystem in thebrain thatfunctions tointerpret behavior in termsof mental statessuch as beliefs, desires, intentions, and plans. But thismodule, responsible forproducing theexperience ofwill, is separatefrom the modules orsystemsthat actually cause behavior, whathe calls the“ empirical will.” Mostof the evidence discussed in thebook is used byWegner todemonstrate that theexperience ofwill and theempirical will “comeapart oftenenough tomake one wonderwhether they may beproduced by separatesystems in themind” (p. 11). The “interpretivemodule” is epiphenomenal, then, in thesense that its activity is independent fromand occurstoo late toaffectthe brain systemsthat cause behavior. Itsfunction is toobserve and theorizeabout behavior, notproduce it. So, modular epiphenomenalism is an empirical claim about thetiming ofand connections betweenevents in thebrain: sluggish conscious processesare tooslow tooccur before behavior is generated, and theyare on aone-way exit fromthe behavior-production system: “The brain, in turn, showsevidence thatthe motor structuresunderlying action are distinct fromthe structures that allow theexperi- enceof will. The experience ofwill may bemanufactured by theinterconnected operationof multiple brain systems, and thesedo notseem to be the same as the systemsthat yield action”(p. 49) [1]. The brain systemsinvolved in producing conscious intentions are modules (or, whendiagrammed, boxes) whichreceive input about whatbehaviors are being produced butwhich themselves produce no output relevant tothebehavior theyrepresent. Wegner’ s central diagram capturesthis idea. Wegner citesthree sources of evidence tosupport the anatomical and temporal claims ofthis modular theory.First, hediscusses psychologists, such as Alan Leslie and Simon Baron-Cohen, whoargue that“ theoryof mind” (our ability tounder- stand mental statesand theirrelations tobehavior) may bea specialized module (p. 24). This view oftheory of mind is controversial, buteven iftrue, it would not showthat theory of mind always involves retrospectiveinterpretation, never prospec- WHENCONSCIOUSNESS MATTERS 531

FIG.1:The experienceof consciouswill arises when apersoninfers an apparent causalpath fromthought to action.The actualcasual paths arenot presentin the person’s consciousness.The thought iscaused by unconsciousmental events, and the actionis caused by unconsciousmental events, and these unconscious mentalevents may alsobe linked to eachother directlyor through yet other mentalor brainprocesses. The willis experienced as aresultof what isapparent, though, not what isreal. Source:Wegner(2002, p. 68). tive inuence. Many actions involving, forinstance, social coordination, deception, and self-controlrequire theoryof mind (as suggested bythedifŽ culty theypose for autistics, young children, and animals whichsupposedly lack theoryof mind). There is no reason tothink thismodule (if it is amodule) represents“ theendpoint ofa very elaborateinference system underlying apparent mental causation”(p. 325) ratherthan an importantcomponent in planning and executing sometypes of action. Wegner’s second sourceof evidence formodular epiphenomenalism is drawn fromBenjamin Libet’s experimentson therole of conscious will in voluntary behavior. Thesewell-known experimentssuggest thatunconscious neural activity (thereadiness potential) precedessubjects’ conscious awareness oftheir desire to, say, movea Žnger. Wegner uses theseresults tosuggest thatthe experience of conscious will might noteven bea link in the“ causal chain leading toaction” ; instead, “it might justbe a looseend— one of those things, like theaction, thatis caused by priorbrain and mental events”(p. 55). The image, whichrecurs through- outthe book, is thatunconscious mental activity A is acommoncause forboth behavior B and theexperience ofwilling theaction C: A is the real cause of B, and C is aseparateeffect that comes too late toin uence B.(Wegner also notesthat theremay betwoseparate unconscious causes A1 for B and A2 for C and that A1 and

A2 may ormay notbe connected— see Figure 1.) 532 E. NAHMIAS

But Libethimself suggests thatour awareness ofvolition occursin timeto veto theaction— so conscious will is notcausally cutoff from action production. More importantconsiderations involve commoncritiques ofLibet’ s work.First of all, Libet’s discussion neglects therole of earlier conscious intentions—for instance, the subjects’accepting theinstructions and deciding tospontaneously lift aŽnger later. Libet’s results, unless generalized as Wegner proposes,do notchallenge thecausal roleof the “ action plan”(or pre-intention) formedwhen subjects agree todo what is asked ofthem. Second, thereis thedifŽ culty ofpicking outthe phenomenology and timing of“ willing,”and oftrying totime it precisely. Try spontaneously lifting your Žnger several times(and pay attention tothe experience). When Iask people todo this, theyreport quite differentfeelings: somefeel themselves initiate the movement,some feel the urge comeupon them,and somefeel the desire tomove only after theymove. The language Libetuses in various places (“urge,”“desire,” “decision,”“intention” ) is as diverse as thephenomenology. In any case, as commentatorshave oftenobserved (see Libet, 1985), anyone whois nota dualist should notbe surprised thatbrain activity precedesconscious awareness. The question is whether,as Wegner suggests, only unconscious processescause action and all brain processes“ underlying”conscious will (touse Wegner’s vague language ofthe metaphysical relation) are epiphenomenal. Libet’s results do notshow this. Finally, Wegner argues thatthe modular view is supportedby experiments using directelectrical stimulation todifferent parts of thecortex. Some stimulations producebehaviors experienced as unwilled—“I didn’t do that. You did”(p. 45)— while othersproduce behavior thatis accompanied by an experience ofwill (so subjectsthen invent areason whythey performed the stimulated behavior). Wegner concludes: “thebrain structurethat provides theexperience ofwill is separate from thebrain sourceof action” (p. 47). This interpretationneglects thepossibility that differentstimulations may simply hitthe complicated “causal chain”in various places—after all, you can hitmy knee witha mallet and I’ll moveit without experiencing themovement as willed. In any case, it should notbe surprising that thebrain’ s systemsfor of action and production ofaction are not identical. Weneed toperceive actions withoutproducing themwhen, forinstance, weobserve others acting, and weneed toproduce actions withoutperceiving what weare doing whenwe act very quickly (e.g., hitting afastball). But tosay the systemsare notidentical is notto say theyare unconnected [2].

3.Exceptions as rules? Wegner seemsto realize thetenuous nature ofthesupport these research programs provide forhis conclusions. Thus, heoften uses tentative language todescribe his conclusions: “This, in turn, suggests theinteresting possibility thatconscious will is an add-on, an experience thathas its ownorigins and consequences. The experience of will may not be very Žrmly connectedto the processes that produce action, in that whatevercreates the experience ofwill may function in away thatis only loosely coupled withthe mechanisms thatyield action”(p. 47, myitalics). Thereis nothing wrong withscientists being cautious in theirinterpretations of empirical work—it WHENCONSCIOUSNESS MATTERS 533 should becommended. But Wegner’s use ofsuch language can itselfbe interpreted in twoways— as an indication ofhis caution oras an indication thatthe work he citescan only supportthe claim that sometimes theexperience ofwill can occur withoutthe relevant action and vice versa. In fact, mostof the experiments and experiences Wegner discusses throughout therest of the book are bestseen as supporting thisinterpretation— that there are various exceptionsto the rule thatour conscious experiences ofour actions corre- spond withthose actions. But thefact that there are theseexceptions does not show that, in the normal cases ofcorrespondence,conscious will is causally irrelevant. Just because consciousness ofwilling an action can beseparated fromthe action doesnot mean thatwhen they are notseparated, consciousness is notcausally relevant (just because A can occurwithout B doesnot mean thatwhen A does occur with B, A is nota cause of B) [3]. Wegner assertsthat the exceptions may in factbe the rule, buthe offers little supportfor that claim beyond thehypothesis of modular epiphenomenalism itself [4]. Forinstance, afterdiscussing automatisms—behaviors an agent performswith- outexperiencing conscious will, such as moving aOuija board orautomatic writing—Wegner concludes, “Ifconscious will is illusory, automatisms are somehow the‘ real thing’, fundamental mechanisms ofmind thatare leftover once the illusion has beenstripped away. Ratherthan conscious will being therule and automatism theexception, theopposite may betrue: automatism is therule, and theillusion of conscious will is theexception” (p. 143). His idea is thatautomatisms are produced by “thesame kinds ofprocesses that create intentional action”but the interpretive process,an add-on foractions experienced as intentional, is bypassed orobstructed in someway soit doesnot produce the illusion ofconscious control(p. 130). But anotherinterpretation is thatautomatisms are not produced by thesame kinds of processesthat create intentional action precisely because thecausal roleof conscious intention has beenbypassed. Itseemsthat neuroscientiŽ c data would bemost useful totest between these hypotheses, but Wegner offersno such evidence (beyond that discussed above, whichdoes not directly apply toautomatisms). Wegner argues in Chapter 3thatthe interpretive process that creates the experience ofconscious will worksaccording tothe theory of “ apparent mental causation.”We experience ourconscious thoughtsand intentions as causing our actions whenthey Ž tthreecriteria which, heclaims, drawing on Hume, “peopleuse in theperception of causality moregenerally” (p. 64): (1) priority—theconscious thoughtsand intentions (hereafter,just “ thoughts”) mustprecede the behavior they represent,usually by 1–5 seconds; (2) consistency —thethoughts must be consistent withthe actions (this will oftenbe apparent because theyare semantically related); (3) exclusivity —webelieve ourthoughts cause ouractions tothe extent that we do notperceive other factors that might belikely causes ofthe actions [5]. Mostof Wegner’ s examples focuson cases in whichone of these three principles breaks down sothat we do notexperience ourselves as causing behavior thatwe are in factcausing (illusion ofnon-control) and cases in which, because the threeprinciples are met,we experience ourselves as causing behavior thatwe are not 534 E. NAHMIAS in factcausing (illusion ofcontrol). Forinstance, in the“ ISpy”study, by creating theexperience ofconsistency, exclusivity, and especially priority, Wegner could induce subjectsto think theycontrolled acomputermouse when they did not (pp. 74–78). Conversely, peoplefeel out of control when they perceive somealternative cause oftheir behavior (thus losing theexperience ofexclusivity), such as internal impulses orexternal agents (as in hypnosis orschizophrenics’ hearing voices). Wegner citesthe classic automatisms (Ouija boards, automatic writing, etc.) as primeexamples ofvoluntary actions—though theuse of“ voluntary” seemsquestionable— which are notexperienced as consciously willed, usually because theprinciple ofconsistency breaks down. Wedo notunderstand howour speciŽc movementscould becaused by ourthoughts since wedo notconsciously intend thosemovements. Sometimes our actions are even produced by trying to resistthem, as in ironic processesin whichtrying not tothink about ordo something can increase thelikelihood ofthinking about ordoing it (Wegner is known forhis workon ironic processes,summarized in White bearsand other unwanted thoughts , 1989). Otherexamples Wegner discusses in depth(in Chapter4) include dissociative experiences, ideomotoraction, dowsing, and thechevreul pendulum, eachof which involves agents’producing movementswithout feeling like theyare. InChapter8 he analyzes hypnosis in asimilar way: theempirical will succumbs tothesuggestions of thehypnotist while theconscious will, “associated speciŽcally withan executive or controlling module ofmind,” is bypassed (p. 301). The hypnotized agent doesnot feelout of control but he does not feel he is consciously controlling his actions either.In Chapters6– 7 Wegner discusses theconverse cases in whichour interpre- tive module goes awry and attributesconscious intentions and controlwhere there is none. This so-called “action projection”includes thecase ofClever Hans (and otherexamples ofanthropomorphizing animals and objects),the tragic case of facilitated communication [6], and jointventures in whichpeople may projecttheir intentions ontothose of the group. Italso includes thecreation of“ virtual agents,” such as ghosts, spirits, gods, and alternate personalities, by projectingpurposeful intentions ontopatterns of events orbehavior. Wegner ends thechapter by postulat- ing thatthese exceptional cases may representthe rule forhow individuals createa sense ofself: “Weachieve thefact of having aperspectiveand being aconscious agent by appreciating thegeneral idea ofagents overall and thenby constructing a virtual agent in whichwe can reside”(p. 269). Ishould pointout that Wegner describesall ofthese cases withlucid proseand charming wit. Mostof the examples are inherently interesting since theyare, well, weird, butWegner also slips in lotsof jokes and puns tokeep the reader entertained (thebad puns are betterthan no puns at all). Icertainly recommendthe bookas an informative and fascinating journey throughsome of psychology’ s puzzles—and as Iexplain in my conclusion, Irecommendit as atypeof book that needs tobe reckoned with. But returning tohis central thesis, Ithink Wegner has misinterpreted, orat least exaggerated, theramiŽ cations ofthese psychological puzzles. WHENCONSCIOUSNESS MATTERS 535

4.When consciousness matters When all is said and done, Wegner has offeredno evidence orarguments against this proposal: certain brain processeshave theproperty of being consciously represented tothe agent as themental stateswe describe as beliefs, desires, intentions, and actions (forinstance, mybrain is currently going throughprocesses which I experi- enceas something like “Ithink thisproposal makes sense,”“Type outthe words ‘thisproposal makes sense’,” and soon). Howit is thatthese brain processeshave theseexperiential propertiesis currently amystery (that is thehard problemof consciousness). But ifthese processes did nothave theirrepresentational properties, thenthey would nothave thecausal powersthey have (forinstance, theywould not allow meto represent various ways ofcarrying outmy intentions and hencewould notallow meas much exibility toinhibit oradjust my actions) [7]. Hence, my conscious experiences have importantcausal inuences on my actions. Granted, Ido notrepresent the brain processesthemselves and so, in some sense, myexperience ofconscious will is illusory (perhaps in thesame sensethat my experience oftables as solid is illusory). But my experience is notillusory in the sense ofbeing epiphenomenal oreven inessential formost of my actions. And even though theconscious experiences may involve brain systemsdistinct fromthe systemsthat produce some behavior, consciousness is nota separatemodule, abox whichreceives input butproduces no outputrelevant tobehavior. Furthermore, thereis no reason tosuggest thatthese brain processeshave theircausal powersonly in virtue oftheirmechanistic, neural propertiesrather than theirpsychological/ func- tional/representational properties—no morethan thereis reason tosuggest thatthe neural properties“ loseout” their causal powersto thequantum-level propertiesthey are identical to(or supervene on), orthat genes do nothave causal properties because only DNAmolecules do [8]. Finally, sometimes Ibehave in ways thatlook (but don’t feel) similar topurpose- ful actions butwithout engaging thebrain processesthat have theserepresentational properties(e.g. hypnosis and automatisms). And sometimesmy behaviors, orthose Iobservein otheragents, are sosimilar tobehaviors thatare usually accompanied by conscious intentions thatI (unconsciously) attributeconscious controlto them (e.g. anthropomorphism).That purposefulbehaviors and conscious awareness of theircauses can comeapart should notbe too surprising, butwithout the “ rule”that theyusually do not comeapart, theexceptions would notbe intriguing. Indeed, I think Wegner’s intriguing bookhas in factoffered some evidence that supports the proposalI have justdescribed. But even ifWegner’ s modular epiphenomenalism is moreaccurate than my proposal, hedoesnot give conscious experience its properdue. Perhapsto make his thesismore exciting by making it morethreatening, heseemsto push it towardsthe moreexpansive epiphenomenalism thatallows no importantrole for consciousness will in producing actions. He allows theexperience ofwill aretrospectiverole— it connectsus toour past actions— but he allows it no prospectiverole. This is a mistake. Thereare several ways thatconscious representationsof future actions seem 536 E. NAHMIAS essential tothe way weperform them. One involves inhibition. Formany typesof action, in orderto inhibit doing them,it is crucial tobe able torepresent them. Consider examples ofintentional deceptionthat require inhibiting an action (say, an eye gaze towardsa hidden object)to avoid providing information toa target (theory ofmind may also benecessary forsuch deception). Libet’s interpretationof his experimentssuggests conscious awareness ofourvolitions allows us toinhibit them (what hecalls “vetopower” ). Wegner himselfwrites, “ Conscious processesare more exible and strategic”than unconscious ones(p. 57). This is in partbecause they allow us torepresentand comparevarious actions and chosewhich of themsatisŽ es certain goals. And ironically, theironic processesWegner studies suggest that conscious volition is causally relevant toaction since, in theseinstances, trying to achieve acertain goal (e.g., toinhibit thoughtsabout whitebears) often causes the goal tobe frustrated. Theseexceptions, however,should notbe confused withthe rule thatconscious intentions toachieve agoal generally facilitate actions that achieve thegoal. Indeed, themost obvious way thatconscious will plays acausal rolein behavior involves developing general goals orplans forbehavior— that is, theactivity involved in deliberating, deciding, forming priorintentions, and carrying outextended actions (see, forinstance, Bratman, 1987; Searle, 2002). Itis strange thatWegner discusses someof these points when he overviews thephilosophy ofaction early in thebook (pp. 18–20), only toignore themwhen he advances his main thesis. His modular epiphenomenalism seemsmeant toapply tothe conscious intentions we experience as occurring only justbefore we act, and mostof his examples do not address planning, deliberation, orextended actions. But Ithink theseare activities in whichwe feel our conscious thoughtsand intentions as causally inuencing our actions even ifthey occur much earlier than theactions (and even ifwe do notthen experience an immediately priorconscious thought). Forinstance, whenwe decide todrive home,to proposemarriage, toreconcile witha friend, tostudy, orto water the plants, weform general “action plans” withoutŽ lling in mostof the detailed behavior thatwill constitutethe actions. Indeed, alittle phenomenological investigation seemsto suggest thatwe do not consciously will mostof our speciŽ c movements,except in cases wherethe actions are being learned, are difŽcult, orrequire inhibition, etc.The “quick actions,”such as typing, thatWegner discusses (pp. 56–59) todemonstrate that consciousness is tooslow to occurbefore such actions seeminstead tobolsterthe idea thatconscious will simply doesits jobearlier— for instance, whenwe decide whatto typeor thatwe will try tohit the fastball or, in theLibet experiment, thatwe will moveour Ž nger at someunspeciŽ ed time. Wegner instead concludes that“ consciousness popsin and outof the picture and doesn’t really seemto do anything”(p. 59). Furthermore,what we feel most responsible forare theplans and policies we make forour own lives, thedecisions wemake afterconscious deliberation about whatsort of motivations wewant tomove us, whattypes of actions wewant to perform,what sort of person we want tobe. Conversely, wecan also feelresponsible foractions weperform without conscious will. Consider an athletewho is playing “unconsciously”— shemay behyper-conscious ofwhat is going on around herbut WHENCONSCIOUSNESS MATTERS 537 sheis notconsciously willing herspeciŽ c actions. Nonetheless,she feels responsible because shehas trained precisely sothatshe can actwithout consciously willing her actions [9]. Musicians and actorsaim forthe same goal. Indeed, weall aim tobe able, in many typesof situations, toactwithout the feeling ofwill, butin accord with theplans wehave consciously formed.And wefeel free and responsible forthe actions thataccord withthese plans, regardless ofwhetherwe consciously will each of them. Wegner ends his bookwith his ownsolution tothe question offree will. Basically, hesuggests thatby explaining howthe brain gives rise tothe experience (theillusion) ofconscious will, wecan explain whywe feelin controlof someof our actions and feelmorally responsible forthem. This, ofcourse, is no solution at all, since no onewants theexperience ofconscious will, self-control, and personal responsibility tobe illusory, tobe mere feelings thatdo notcorrespond to reality. What wewant is thatour conscious selves have areal inuence on whatwe do, on whatwe become.We want (prospectively) toacton ourreasons, notŽ rstto act and then(retrospectively) torationalize whywe acted. Wewant toknow what we’ re doing and (at least sometimes)why we are doing it. Ithink wecan have thissort of freewill, and it is mostevident in ourconscious, sometimesself-re ective, deliber- ations and plans (see, forinstance, Bratman, 1984; Nahmias, 2001; Searle, 2002). Thereis areason wedon’ t think thecorrelation between a lifetimewithout stepping onasidewalk crack and amothernever breaking herback is areal causal correlation. Even if weare toosuperstitious totest the counterfactual, wecannot imagine howthe two events are physically connectedin theright way. But the correlationbetween conscious intentions, plans, and thoughtsand theactions they representare notso physically isolated. Even if, as Wegner suggests, theyare produced by distinct systemsin thebrain— aclaim Ihave disputed—nothing in the brain is completelyisolated. Itwould bea miracle ifthe system that deliberates, decides, and formsgeneral action plans did notcausally inuence thesystem that producesbodily movements.Or perhaps it would justbe epiphenomenalism— the strong kind thatsays no conscious processdoes anything. But since Wegner does notdiscuss whathis modular epiphenomenalist thesishas tosay about thesetypes ofconscious processes,he leaves thereader wondering howthreatening theillusion ofconscious will is meant tobe [10].

5.Conclusion: whyphilosophers should respondto Wegner’s work (and work likeit) Wegner claims, “Philosophershave given us plenty of‘isms’to use in describing the positionsthat can betaken on thisquestion [offree will and determinism], mean- while notreally answering it in asatisfying way”(p. iv); theyhave “generated a literature thatis immense, rich, and shocking in its inconclusiveness”(p. 26). We philosopherswould respond thatwe have notsolved thisproblem because it is not easy! In fact, someclaim ourlimited human brains will never beable tounderstand howour brains are conscious orhow it is thatwe could have freewill (e.g., Noam 538 E. NAHMIAS

Chomsky’s and Colin McGinn’s mysterianism and Petervan Inwagen’s recentviews on themystery of metaphysical freedom—positions I do not agree with). But thereis adanger in theway philosophershave dealt withthe problem of freewill. Wehave generated volumes in thedeadlocked debatesbetween compati- bilists and incompatibilists, withoutappreciating thatthese two camps generally agree about almost everything necessary forfree will— everything butwhether inde- terminism ofsome sort is essential. Wegenerally agree, forinstance, thatfree will requiresthat our conscious deliberations make adifferencein whatwe do, thatwe have theability toact according toour re ectively considered desires and reasons, and thatwe have conscious controlover some of our actions as weperform them. So, whensomeone— like Wegner, whoaptly recognizes theessential connection betweenthe problems of consciousness and freewill— suggests thatwe don’ t even have theseabilities, weshould collectively take notice. Weshould take noticebecause otherpeople will. Formost philosophers the problemof may beabout whetherit is compatiblewith the metaphysical thesisof determinism. But therest of the world does not feel as threatenedby metaphysical vagaries (which may beunveriŽ able). Forthem, the problem of free will, whenit goes beyond thesociopolitical problemsof freedom, is usually about whetherfree will is compatiblewith the way science tells us weare (and Ihasten to add that, despiteconfusion about theissue, scientiŽc explanation is notequivalent todeterministic explanation). What makes peopleworry about freewill is atypeof reductionist view disseminated throughpopular science books,like RobertWright’ s The moralanimal and MattRidley’ s genome,and throughubiquitous media reports about genes forpersonality traits, brain areas thatcontrol deliberation and emotion, and evolutionary storiesabout oursexual, addictive, and aggressive instincts [11]. Theseviews representthe scientiŽ c revolution thatnovelist Thomas Wolfesummed up forthousands ofDuke graduates at the2002 Commencement: Man’s view ofhimself is changing very rapidly. Within thenext twenty years Ithink you’ll feelits effects.After Darwinism, afterFreudianism, there’s nowa tremendousin uence ofneuroscience …IfI may reduce withterrible reductiveness thebottom line ofneuroscience, it is thatlet’ s notkid ourselves, we’re all concatenations ofmolecules containing DNA, hard wired into achemical analogue computerknown as thehuman brain, whichas softwarehas acertain genetic code. And your idea thatyou have asoul oreven aself, muchless freewill, is justan illusion. The idea thatthe self and freewill are illusions may never beable toforce its way into ourcollective consciousness (though thiswill likely depend onwhatthat idea is supposedto mean). But Wolfe’s public address is buoyed by bookslike Wegner’s, despiteWegner’ s concluding attemptsto put a happy faceon his thesis. And such views matter.The degree towhich people feel they can controltheir actions and be responsible forthem varies acrosscultures, eras, and individuals. Informationfrom thesciences (and philosophy) can inuence howthese feelings play out, as demon- stratedby debates overlegal responsibility—from the insanity defenseto the Twinkie defense—and political philosophies—from conservatives’ view thatindivid- WHENCONSCIOUSNESS MATTERS 539 uals are generally responsible fortheir position in life toliberals’ view thatsociety bearsmuch responsibility forindividuals’ adversities. In theface of these important issues, philosophersshould notsequester their debatesabout freewill toissues like metaphysical determinism, whichmay be relevant butwhose relevance is oftenfar from clear. Weshould engage withbooks like Wegner’s because it is partof our job to examine thephilosophical and ethical implications scientists draw fromtheir experimental results and tomake sure they are notinferring morethan theyshould— indeed, abooklike Wegner’s helpsremind aphilosopherhow important our distinctions and debates(e.g., about themind– body problem)can be. Atthe same time, weshould also learn about theempirical factssuch bookspresent so that our own theories are notout of touchwith the way things are. Ihave suggested thatWegner’ s The illusionof conscious will is successful in presenting ahostof empirical factsthat inform us about theway wethink and act (though mostlyin marginal situations). Weshould pay attention tothese facts. But thebook is notsuccessful in presenting adecisive challenge tothe folk intuition at theheart of philosophical conceptionsof free will, thatour conscious experiences of ourdeliberations, planning, intentions, and actions oftenplay an essential rolein whatwe do. Wereit otherwise,then this fourth scientiŽ c revolution would truly be aknockoutblow to whatFreud called our“ na ¨õ ve self-love”— or, as JerryFodor put it, in thecontext of aclosely related debate: “Ifit isn’t literally truethat my wanting is causally responsible formy reaching …and mybelieving is causally responsible for mysaying …thenpractically everything Ibelieve about anything is false and it’s the end ofthe world” (1990, p. 196). Wephilosophers should keepour guards up against any blowthat would bethe end ofthe world [12].

Notes [1]Notice in this quotation (andothers citedbelow) a phenomenoncommon throughout Wegner’s book (andmany scientiŽc books that discussthe mind–brain relationship): the words usedto describethe relationship—“underlying,” “allow,” “manufactured by,” “yield” — arevague and sometimescontradictory. Questions about the metaphysicalrelationship between mental and physicalproperties (e.g. supervenience, identity, causation, correspondence, etc.) are avoided. [2]Interesting research on “ mirrorneurons” has foundneurons in primates’ premotor cortex that activate both when the subjectmoves in particular ways and when the subjectperceives similar movementsin others, suggesting that at leastin this brainarea, representation of actionpro- ductionand actionproduction are not cutoff fromeach other, perhaps facilitatingimitation (see Gallese et al., 2002). [3]This sort of fallaciousreasoning sometimes appears indiscussions of phenomenalike blindsight when peopleargue that, because(some) visual stimuli can bediscriminated without conscious awareness, consciousnessis not inprinciple necessary to visualperception. Or morecomprehen- sively:because it seems possible to performcomplex behaviors likeours without consciousness (e.g.robots), our consciousprocesses are not causallynecessary for ourbehavior. [4]However, another sourceof evidenceWegner takes to support the rulethat “consciousness doesn’t know how consciousmental processes work” (p. 67) comes from social psychology researchthat suggests we cannot correctlyintrospect on the causesof ourmental states or behaviors—e.g. generally we misattributecausation to charactertraits rather than situational factors. InNahmias (forthcoming),I argue that the implicationsof such researchhave been 540 E. NAHMIAS

exaggerated, often because,like Wegner, the theorists interpretabnormal cases as representingthe norm. [5]Wegner’ s useof Humeis suspect (seepp. 13–14, 64, 66); he suggests that Humeshowed us that ourperception of the constant conjunctionbetween our conscious intentions and ouractions is not evidenceof arealcausal relation, but Wegnerdoes not apply Hume’s skepticismto the low-levelmechanistic processes dubbed the “realcauses of human action”(p. 97). Also, a consistentuse of Humewould allow that, evenif conscious will and actionare not constantly conjoined,certain instances are;the exceptionsmight, as Isuggest, be exceptions.Note alsothat Humeanaccounts of causation arecontroversial. [6]Facilitated communication involves facilitators holding their hands overthose of adisabledperson to helphim or her type messages ona keyboard. Despitetheir belief that they arenot activein composingthe messages (theirlack of consciouswill), it has beenshown that they areresponsible for the contentof the messages. [7]I presentthis counterfactualin terms of physical(or natural) necessity: I makeno claim about consciousnessbeing metaphysically (much less, logically) necessary for certainbehaviors. Forthe purposes of understandingWegner’ s empiricalclaims, this type of necessityis the relevantone. [8]This analogy betweendifferent examples of reductionismis, of course,too simple,but the point isthat Wegner’s continualclaim that onlyunconscious mechanisms are real causes requires an argument hedoes not provide. [9]Olympic soccer star Tiffany Milbrettexplained her game-tying goal duringstoppage timeof the 2000gold-medal match inthis way: “Ihad nothoughts. Mybody just didwhat itdid.” Also, see the Norman–Shallice (1986) model of actionwhich posits two complementaryprocesses in controllingand guidingaction, one that dealswith novelor difŽ cult actions that requireconscious attention and onethat dealswith simple,well-learned, and habitual skillsthat donot require attention. The latterinvolve “ motor schemata”which cancarry out quitecomplex movements, though consciousnessof the behavior stillallows the agent to becomeaware of any deviationsand correctfor them (aswhen amusicianhits awrong note). [10]The closestWegner comes to discussingthe roleof consciousnessin deliberationand planningis when hementions controlled mental processes and actions(p. 97), where he says, “The abilityto know what onewill do, and particularlyto communicatethis to others verbally,would seem to bean important human asset,”but healso claims that “evenif [theseactions] occur along with an experienceof controlor conscious will, this experienceis not adirectindication of theirreal causalin uence.” He refers to aforthcoming articlewhere he will spell out this argument,“ Who isthe controllerof controlledprocesses?” My guess ishe will end up sinkinginto the stronger formof epiphenomenalism. [11]See Chapter 17,“ Blamingthe victim,”of Themoral animal (Pantheon,1994), and Chapter 22, “Freewill,” of genome (Fourth Estate, 2000).Earlier books that sparked these amesinclude RichardDawkins’ TheselŽ sh gene and E.O.Wilson’ s On human nature .Needlessto say, eachof these books has beenmore widely read and discussedthan any recentphilosophical work onfree will. [12]Thank you to AlMele, Brandon Towl, and CherylKopec Nahmias for helpfulcomments. Fundingsupport froman FSU FYAPgrant.

References

BRATMAN,M.E.(1987). Intentions, plans, and practicalreason .Cambridge,MA: Press. FODOR, J. (1990). Atheoryof contentand otheressays .Cambridge,MA: MITPress. FREUD,S.(1989/1920). Introductory lectureson psycho-analysis .New York:Norton. GALLESE, V., PEIRFRANSCESCO, F., KOHLER, E. & FOGASSI,L.(2002).The eyes,the hand, and the mind: behavioral and neurophysiologicalaspects of socialcognition. In M. B EKOFF, C. ALLEN & G. BURGHARDT (Eds) Thecognitive animal .Cambridge,MA: MITPress. WHENCONSCIOUSNESS MATTERS 541

LIBET,B.(1985).Unconscious cerebral initiative and the roleof consciouswill in voluntary action. Behavioraland BrainSciences , 8, 529–566. NAHMIAS,E.(2001).Free will and the knowledgecondition. Unpublished dissertation. NAHMIAS,E.(forthcoming).Free will and the threat of socialpsychology. NORMAN, D.A. & SHALLICE,T.(1986).Attention to action:willed and automaticcontrol of behavior. In R.J. DAVIDSON, G.E. SCHWARTZ & D. SHAPIRO (Eds) Consciousness and self-regulation:advances in researchand theory , 4, 1–18. SEARLE, J. (2002). Rationalityin action .Cambridge,MA: MITPress. WEGNER,D.M.(1989). Whitebears and otherunwanted thoughts .New York:Viking Penguin. WEGNER,D.M.(2002). Theillusion of conscious will .Cambridge,MA: MITPress.