Myth, Memory and Misrecognition in Sellars' \"Empiricism and the Philosophy of Mind\"

$Page 1: Myth, Memory and Misrecognition in Sellars' \"Empiricism and the Philosophy of Mind\"$
REBECCA KUKLA

MYTH, MEMORY AND MISRECOGNITION IN SELLARS’“EMPIRICISM AND THE PHILOSOPHY OF MIND”

(Received in revised form 20 May 1998)

The Laws are not of yesterday or today, but ever-lasting. Though where they come from, none of uscan tell. Theyare.

Antigoneln. 500, quoted by Hegel,Phenomenologyof Spirit, §437.

William James, in order to shew that thought ispossible without speech, quotes the recollection ofa deaf-mute, Mr. Ballard, who wrote that in hisearly youth, even before he could speak, he hadthoughts about God and the world. – What couldhe have meant? . . . And why does this question –which otherwise seemed not to exist – raise its headhere? Do I want to say that the writer’s memorydeceives him? – I don’t even know if I shouldsay that. These recollections are a queer memoryphenomenon – and I do not know what conclusionsone can draw from them about the past of the manwho recounts them.

Wittgenstein,Philosophical Investigations, §342.

INTRODUCTION

In increasing numbers, philosophers are coming to read Sellars’“Empiricism and the Philosophy of Mind” (1997, hereafter EPM)as having dealt the definitive death blow to the idea that innerstates with epistemic authority could have this authorityimmedi-ately. EPM purportedly proves that instead, such states necessarilyshow up already embedded within a web of inferentially articulatedconceptual knowledge, and that in order for this to be possible,

Philosophical Studies101: 161–211, 2000.© 2000Kluwer Academic Publishers. Printed in the Netherlands.

162 REBECCA KUKLA

the epistemic subject must be a negotiator of anormativespace inwhich standards of justification and correctness are already recog-nized. This normative space is the so-called ‘space of reasons’,which is organized by relations of epistemic authority. Since allepistemic authority is mediated by and legitimized within such aspace, there can be no self-authorizing epistemic states whose legit-imacy is immediately given. Robert Brandom, Richard Rorty andJohn McDowell lead the way among authors who have tried toreveal this move as a central thrust of Sellars’ work, and as the moralof Sellars’ famous attack on the ‘Myth of the Given’.1

Now philosophers of all stripes, from both sides of the Atlantic,have long wondered how we could possible get inducted into norm-ative space in the first place. Various paradoxes and problems haveaccompanied attempts to explain how authority first gets off theground and comes to bind us, or how a distance between ‘is’ and‘ought’ opens up for us. Philosophers as diverse as Hegel, Louis,Althusser and McDowell have argued that all efforts to tell a storyof our transformation from merely natural beings driven around bycausal forces, into negotiators of normative space capable of recog-nizing and wielding authority (whether this is epistemic or socialauthority), have been failures; these stories either fail to get genuinenormativity into their picture, or they beg the question and illegit-imately smuggle it in from the start. In contrast, these authors arguethat in order for authority to get a grip upon us, we need to recognizeit as always alreadyhaving been there, and likewise we need tounderstand ourselves as always already having been inhabitants ofnormative space.2

EPM is centrally concerned with the origins of epistemicauthority and the problem of our induction into normative spaces.But I will try to show in this paper that Sellars rejects the tradi-tional strategies for engaging with this problem, neither opting tooffer natural histories of these origins, nor resting content withclaiming that we literally inhabit a world always already laden withnorms.3 I argue here that Sellars is best read as attempting a solu-tion of a radically different form, in which he explains the originof authoritymythically, where the philosophical function of thesemyths is irreducibly distinct from the function of literal or pseudo-literal descriptions of how things are. In this paper I will attempt

MYTH, MEMORY AND MISRECOGNITION 163

to show that Sellars’ mythical explanations in EPM employ a veryspecific and rhetorically complex methodology, and likewise that wewill not be in a position to critically assess the paper’s argumentsunless we give careful attention to its overall textual structure and tothe nature of the mythical explanations it employs.

EPM opens with the Myth of the Given, and closes with theMyth of Jones. Sellars claims that the argumentative strategy ofhis paper is to “use a myth to kill a myth” (EPM §63), andindeed, he insistently and consistently uses the language of mythand fiction throughout the text. We thus have an immediate philo-sophical question on our hands: how can fiction be used to establisha non-fictional result such as an account of epistemic authority?More specifically, how can a myth kill a myth, in a way thatmeets the standards of philosophical argumentation? Sellars givesno direct methodological analysis of myth in EPM, and it is notclear to me that he understood all the details of how his own rhet-orical strategies functioned. But we do have plenty of evidence forhis ongoing concern with the peculiarities of origin stories, theircomplex relationship to and distinction from literal stories, and theirphilosophical and justificatory functioning. This concern becomesa central theme, for instance, in bothScience and Metaphysics(Sellars, 1968), and “Philosophy and the Scientific Image of Man”(Sellars, 1963, hereafter PSIM). It is thus remarkable that next to nocommentators concern themselves with the nature and methodolo-gical function of myth-telling in Sellars’ work. Most commentatorsunabashedly and without comment eliminate the mythical languageof EPM when analyzing its arguments.4 Those who do take note ofSellars’ myth-telling countenance only very simple understandingsof myth (such as that a myth is a mere fiction, a ‘nonexistent object’,or an armchair historical hypothesis),5 all of which foreclose fromthe start the possibility of an interesting epistemological role forsuch myth-telling in philosophy.6 Having failed to take seriouslythe need for a philosophical analysis of myth in reading EPM, it isnot surprising that such commentators conclude that “to be told [asSellars tells us] that the story of Jones and the Ryleans is more like a‘social contract theory’ or a ‘Platonic myth’ [than a literal history] isscarcely illuminating” (Echelbarger, 1974, p. 240), and that as longas we take Sellars at his word and treat his myths as myths, their

164 REBECCA KUKLA

epistemological role in EPM will remain a mystery (Ibid., p. 241).In contrast, I will try to show that it isonly when we treat Sellars’myths as myths that their place in EPM canceaseto be a mystery,and that we can read and assess the arguments of EPM justly.

1. MYTHS AND THE CONSTITUTION OF AUTHORITY

EPM is structured around three myths, each of which, in a differentway, seeks to tell a story about the origin of the epistemic authorityof our experiences. I will briefly recap these three myths, althoughtheir details will not matter until later sections. The first is the Mythof the Given, which seeks to ground our observational knowledgein immediate, self-authorizing epistemic states, most paradigmatic-ally but not necessarily sense-data (such as the state of something’slooking green to me now). Sellars seeks to reveal the incoherenceof this myth, by showing how the epistemic authority of anythingwe try to identify as ‘immediate’ experience or knowledge will infact depend upon its location within our already inferentially artic-ulated web of authoritative conceptual knowledge, which meansthat it cannot in fact serve as an origin of epistemic authority.This revelation occurs through the telling of a second story, that ofJohn the Necktie Salesman, which Sellars does not actually call amyth (though as we will see it has some mythical elements). Thissecond myth problematizes the Myth of the Given, by showing thatthe authority of John’s reports (and experiences) that a thinglooksgreenactually derives from the prior authority of his reports (andexperiences) ofseeing thata thing is green, which in turn dependsupon his having learned how to recognized the (essentially norm-ative) standard conditionsunder which his experiences can countas providing him with legitimate observational knowledge. Finally,Sellars tells the Myth of Jones the Rylean Genius. Here we receivethe myth of the origin of authoritative inner states, not in a singleknower, but in ‘history’, via a story of their ‘discovery’ whereinJones posits them as part of a theory designed to explain variousfeatures of our overt public behavior. Instead of locating the originof the epistemic authority of specific inner episodes in other, found-ational episodes (the Myth of the Given) or in prior observationalknowledge (the Myth of John), the Myth of Jones seeks to locate


this general type of epistemic authority within a theoretical project.This myth will ‘kill’ the Myth of the Given, not only by reinforcingits incoherence, but in revealing it, in good Hegelian fashion, to bea mistaken momentwithin the Myth of Jones.

There are many understandings of myth, and of what distin-guishes myths from other types of fictions. For the purpose of thispaper, my working definition of myth will be a narrative that is rhet-orically set in thepast, and which is used by its tellers as part ofan attempt tomake sense of the present, in some way other than bypurporting to provide a literal, chronological history. This hopefullyintuitively plausible definition is general enough to remain neutralbetween many of the theories of myth but forward by narrativetheorists and rhetoricians. Notice that as defined, a myth need notturn out to befalse. Notice also that it is built into my definition thatmyths aretold, and that they are told for a specificpurposeand thushave a project to accomplish and certain standards of success up towhich they must live. Within the general category of the mythical,there is a certain special type of myth that is of specific interestfor the purposes of reading EPM. These are myths thatlegitimizeauthority, by telling a story about theorigin of that authority. Likeall myths, these make sense of the present via a fiction about thepast, but the particular features of a present state that they seek tomake intelligible are normative features, namely the ability thingsor people have to make authoritative claims of various sorts.7

Now making sense of something’s authoritative status is not thesame kind of project as making sense of, for example, its empir-ical properties. Something is authoritative only if it is binding, andmakes a claim on the subject of its authority. Furthermore, for it togenuinely bind or make a claim, its authority must be legitimate.There can be no such thing as real yet illegitimate authority, sincesuch ‘authority’ would not in fact bind us; the closest there could beto such a thing would be coercive force which makes no normativeclaims upon us. Thus, recognizing authority is inseparable from atleast implicitly recognizing that this authority isalready legitimate– to recognize authority is to recognize the claim it makes uponus. In fact, it seems thatmaking sense ofsomething’s authoritativestatus can mean nothing other than providing the legitimizationwhich grounds it and makes it binding. Thus myths of the origin of

166 REBECCA KUKLA

authority are legitimization myths, and they succeed exactly to theextent that, in their telling, they do legitimize authority and bringus to recognize the claims we can make and that can be made uponus. Furthermore, according to Sellars authority needs a subject –it must make a claimon someone or something, and in order forthis to happen, that someone must be capable of, as Sellars willput it, ‘in some sense recognizing’ the claims of that authority.8

We cannot make claims on inanimate objects, for instance, becausethose claims don’t matter to them and hence can’t count as claims.But, to come to recognize authorityis to come to recognize its legit-imacy. Thus for Sellars, nothing can truly be said to have authority atall until the legitimacy of its claim is of a sort that can in principle berecognized by those upon whom it makes a claim. Therefore, mythsand other justificatory stories whose purpose is to make sense ofthe legitimacy of authority may, if they work, not onlyrevealbutconstitutethat authority. Making sense of authority and constitutingauthority are essentially linked projects, in a way that making senseof empirical phenomena and constituting empirical phenomena donot seem to be. Therefore, to the extent that they are successful,these legitimization myths have a special performative status: in theact of beingtold they help constitute what they seek to illuminate.

All of this only becomes relevant to reading Sellars in virtue ofthis recasting the problem of epistemic foundations as a problemfirst and foremost about the nature and source of epistemic authority,and hence as a normative problem. But once he has shown that beingan epistemic agent depends essentially upon being a negotiator ofnormative space (specifically the space of reasons), then questionsabout the source and nature of our epistemic standing become oneand the same as questions about the legitimization of epistemicauthority. My claim is that Sellars answers these questions by telling– and showing how we already tell –constitutive myths, which donot merelyrevealan already-present legitimate basis for epistemicauthority, butperformthe legitimization through being told. At thesame time, his myth-telling disables the performative function ofthe Myth of the Given by revealing how it fails to serve as a mythlegitimizing authority.

But how can a mythical, non-literal story legitimize authority? Iam going to turn my attention away from Sellars for a while, and


take a detour through some social theory, where we can see mythworking in much the same way I will claim it works for Sellars,in a different yet (I hope to reveal) importantly analogous domain.Just as we can ask how epistemic states come to be placed withinthe space of reasons, we can also ask how subjects or persons cometo be placed within the space of social norms. Now an epistemicstate is an authoritative state, and hence a myth which constitutesits authority will also help constitute its existence as an epistemicstate. Plausibly, we can say something similar for people: to be asubject is, in part, to be bound by and a negotiator of norms, andin fact to be a specific person is partly a matter of being placed ina specific way within the space of norms. By drawing heavily onthe account developed by Louis Althusser in his essay “Ideologyand Ideological State Apparatuses” (1971),9 we can start to see howmyth might function to induct us into or place us within normativespace, and thus help constitute our status as subjects.

Althusser argues that subjects get inducted into normative spacethrough an odd process ofmisrecognition, wherein other authorit-ative subjects (mis)recognize a (potential) subject asalreadyboundby norms and capable of negotiating normative space, and in turn thesubject (mis)recognizes herself asproperly identifiedin that recog-nition, and asalready having beenthe subject she is recognized asbeing. This misrecognition plays a constitutive role; it ‘produces’or ‘recruits’ subjects by placing them within normative space andbinding them in the appropriate ways, and hence it serves to bringabout its own correctness.

The paradigmatic example of this process at work, for Althusser,is the event of ‘hailing’ or ‘interpellation’: that is, the moment whenI (who am already a subject, and hence wield some authority) greetsomeone else or identify him by name (“Hi, Richard!”). The hailhas the following peculiar structure. It presents itself as a recogni-tion of an already-given fact, namely that Richard, a specific personalready bound by specific norms, is there to be hailed. But at thesame time it functions as ademandthat Richard be that person,which involves his acknowledging the bindingness of these normsupon him. In turn, Richard, in recognizing that the hail isreallyaimed athim – that he really is the self being acknowledged, andhence that its demand is binding – accepts the authority of the

168 REBECCA KUKLA

relevant norms. Althusser’s claims is that it is though respondingto this demand, and hence acknowledging the appropriateness ofits recognition, that Richard becomes the (normatively defined)self that he is. He recognizes this hail as reflecting already-truefacts about who he is and what norms legitimately bind him. Butit is through recognizing my hail, and the many other hails thatconstantly acknowledge him and demand recognition back from him(for the act of hailing must be very broadly construed here to includeall sorts of social acknowledgments both literal and metaphorical)that he in fact becomes this self subject to these norms at all. Inconcrete terms, if I say “Hi, Richard”, I am both acknowledgingRichard’s presence and demanding that he respond appropriately tomy greeting. In responding, he lives up to my demand, but in doingso also recognizes that it was really him who I was greeting. Further-more, for him to recognize this is for him to acknowledge that mygreeting legitimately binds him to responding, and to upholdingthe norms of greeting behavior. The reason why Richard in factrecognizes himself as hailed, and responds appropriately, is thatmy greeting, in demanding a response from him, also demands thathe take himself as being that person. The demanding nature of thehail is often accompanied and indicated by an easily accessible andrecognizable psychological impact; to see this, we need only thinkabout how uncomfortable we feel when we ignore a greeting anddecline to respond appropriately.10

On this account, because being a person involves having aspecific normative status, becoming one involves the repetition ofa mythical structure whereby we become bound in the proper wayby being recognized and recognizing ourselves asalreadyhavingbeen bound. If we do not do this, the claims made upon us, such asthe demand implicit in a hail, will not get a grip on us. Recognizingthe hail involves recognizing not just its presence but its legitimacy,which is inseparable from taking it as really aimed atme, and henceas already making a real claim on me in virtue of having identifiedme correctly. The recognition of legitimacy and the constitution oflegitimacy (and likewise, to the extent that being a certain person is anormative status, the recognition of who we are and the constitutionof who we are) occur together. My hail, which on the surface has theform of an acknowledgment of a given fact that Richard is present,


actually functions as a demand that he behave in ways that wouldbe appropriate if he were this person. In recognizing this demandand responding ‘as though’ it reflected an already-given truth byacting ‘like himself’, Richard thereby does just the kind of thing thatmakes it so that he is this person, which in turn makes the demandlegitimate, bringing about the appropriateness of the hail and thecorrectness of its identification. Thus this process of recognition, onboth our parts, is in some sense always a process of misrecognition,whereby we mask our demand that a fact become true in the formof an acknowledgment that it already was true. This gap betweensurface structure and performative force is in fact at the center ofAlthusser’s definition of ideology. Ideology is typically understoodas the masking of a demand or an interested purpose as a already-given fact, and Althusser adds to this analysis the idea that ideology,in perpetrating this misrecognition, actually helps constitute socialfacts.

I have not gone into this story in nearly enough detail to count ashaving argued that it provides a convincing theory of the constitu-tion of subjectivity. My aim here, instead, is merely to make clearthe peculiar structure ofconstitutive misrecognitionthat Althusser ispurporting to reveal. Interpellation employs myth in demanding thatwe recognize facts as havingalready in the pastbeen true, while notbeing able to literally assert these facts about the past. The hail treatsus as though we werealways alreadythe person it recognizes us asbeing. This type of appeal to the always-already has an intrinsicallymythical structure, since it asserts at once our inability to locate anorigin (it was ‘always’ true) and at the same time the fact that suchan origin must have occurred (it is ‘already’ true). Together theseterms cite the past in order to tell us something about how we aresupposed to understand and treat the present, rather than telling ushow we are to literally reconstruct this past. The reason why suchmythical (mis)recognitions can play constitutive roles is that thefacts they recognize arenormativerather than merely empirical facts(if there are such things). If they succeed in binding us to certainstandards of appropriateness, they can thereby bring about their ownlegitimacy through being performed, even if they do so by basingtheir demands upon mythical facts treated as though they were

170 REBECCA KUKLA

‘always-already’ true. Slavoj Zizek gives us the following flowerybut vivid example of how misrecognition can establish authority:

‘Being-a-king’ is an effect of the network of social relations between a ‘king’ andhis ‘subjects’, but – and here is the fetishistic misrecognition – to the participantsof this social bond, their relationship appears necessarily in an inverse form: theythink that they are subjects giving the king royal treatment because the king isalready in himself, outside the relation to his subjects, a king; as if the determin-ation of ‘being-a-king’ were a ‘natural’ property of the person of a king (Zizek1989, pp. 24–25).

I will need to do quite a bit of work in order to make apparenthow this structure of constitutive misrecognition is relevant tothe founding ofepistemicauthority in the way I think it is forSellars. In PSIM, however, he discusses the complex role of recog-nition (although not yet constitutivemisrecognition) in identifyingnormatively bound subjects. This text is the one where Sellars ismost explicit about his concern with the methodological differencesbetween establishing empirical and normative facts. In the course ofdiscussing ascriptions of subjectivity, Sellars points out, as I havebeen, that ‘recognition’ can in such cases function as a normativelyweighted demand instead of merely a descriptive reflection of thefacts:

To say that a person desired to do A, thought it his duty to do B but was forcedto do C, is not todescribehim as one might describe a scientific specimen. Onedoes, indeed describe him but one does something more. . . The irreducibility ofthe personal is the irreducibility of ‘ought’ to ‘is’. . . The most fundamental prin-ciples of a community, which define what is ‘correct’ or ‘incorrect’, ‘right’ or‘wrong’, ‘done’ or ‘not done’, are the most general commonintentionsof thatcommunity with respect to the behavior of members of the group. It follows thatto recognizea featherless biped or dolphin or Martian as a person requires thatone think thoughts of the form, ‘We (one) shall do (or abstain from doing) actionsof kind A in circumstances of kind C’.To think thoughts of this kind is not toclassify or explain, but to rehearse an intention(PSIM 39, some of the italics inthe original).

Thus recognitions of someone’s status as a subject, for Sellars,govern and bind behavior rather than just describing how thingsalready are, although they employ the grammar of reportinglanguage (‘Richard is a person’; ‘Richard wants to fly toAlbuquerque’) and thus importantly misportray their own function.We can ignore the troublingly simplistic aspects of this passage (its


one-dimensional account of communal norms, and the risk it facesof psychologizing the nature of normative demands and commit-ments by employing the language of intentions), and see that Sellars,like Althusser, is pointing out that ‘reports’ on normative facts donot merely reflect fixed truths, but demand from us a set of activecommitments, behaviors and plans. Functioning as an authoritativedemand is the ‘something extra’ that a recognition of a personasa persondoes, which prevents such a recognition from ever beingreduced back into the descriptive language of science and makes itimpossible to give a naturalized story about what it is to identifysomeone as a person. In order for such a recognition to becorrectit must thus belegitimate, and for it to be legitimate we must infact be able to recognize the authoritative claims it commits us to asbinding; hence this recognition plays a constitutive and performativerather than merely a reporting role.

Even once we acknowledge that recognition of normativelyinflected facts look like mere descriptions but function asconstitutive demands, it remains to be seen why and when thesedemands must take the form of mythical misrecognitions of thesefacts as always-already the case. This argument will emerge overthe course of the remaining sections. In the meantime we can turn atlast to the issue of epistemic authority and to the text of EPM itself,beginning with an examination of Sellars’ first and most famousmyth.

2. THE IMPOSSIBLE ORIGIN: THE MYTH OF THE GIVEN

The Myth of the Given begins from the fact that we do in facthave observational knowledge – we make authoritative claims basedon experience. This fact raises an already familiar question: whatenables us to be authoritative epistemic agents of this sort, or,equivalently, what is the source of the legitimacy of these claims?Our observations can only be authoritative to the extent that thereare good reasons to think that they live up to the tribunal of theworld. The Myth of the Given tells a story about the origin of thisauthority: there are some experiences – typically sensory experi-ences – that are directlyself-authorizing, because theyimmediatelypresent the facts they claim to report (such as the fact that some-

172 REBECCA KUKLA

thing ‘looks green’), rather than deriving these facts inferentially.11

Thus, the myth continues, our normal, everyday epistemic authorityconcerning empirical claims derives ultimately from its beginningsin these immediately authoritative given experiences.

In light of the historical fact that it has been notoriously diffi-cult for philosophers to isolate and identify these pieces of non-inferential knowledge, to abstract them from their heavily inference-laden and mediated epistemic context, we have someprima faciereason to suspect that Sellars’ portrayal of these experiences asrecovered through myth is appropriate. It is at least clear, for a start,that at the level of everyday inferences which are or can be madeavailable to consciousness, we do notbeginby noticing the given,and then build up our knowledge inferentially from there. Rather,the self-authorizing states portrayed in the Myth of the Given firstbecome the kinds of things we can turn our attention to when weare in the midst of the theoretical project of telling a retrospectivestory about how we came to have the authority to make our everydayclaims. Even the proponents of the Myth of the Given ought toadmit that whether successful or not, the strength of their accountdoes not lie in its serving as a literal reconstruction of our epistemichistory, wherein we report upon our actual past inferential processes,or discover and examine actual immediate, self-authorizing states;even if the Myth turns out to be literally true, its theoretical plausib-ility lies elsewhere than in its relationship to observable facts aboutinference. Instead, we posit the given as an origin of epistemicauthority for justificatory purpose – we ‘retrieve’ the given in orderto show how wecould havereasoned, and thereby to give authorityto the epistemic position in which we actually find ourselves. Hencethe mythicality of the Given is not necessarily a problem, in lightof how it seeks to play a role in our understanding of epistemicauthority. The question we should be asking is not whether thegiven gets recognized first in myth, but whether this myth servesits purpose.

Sellars rightly attacks the coherence and success of the mythas opposed to its literality. He argues that such self-authorizingexamples of immediate knowledge are impossible, and hence thatthis origin myth fails to ground authority. This argument will notbe complete until the end of EPM, but the problematic status of the


given is set forth in a dilemma he puts forth right near the beginningof the paper. Either our immediate experiences make a claim aboutthe world or they do not. If not, then they are not binding in anyparticular way, and they cannot be said to have epistemic authority,in which case they cannot serve as a ground for inference or a sourceof the legitimacy of other claims. If they do make a claim, thenthey must claim something specific, in contrast to something else.But this seems to require the conceptual resources that enable us todiscern and sort (for otherwise, we can note, the content of our givenexperiences would be a mere ‘this’, which we are unable to distin-guish from any other mute ‘this’), and this in turn brings our wholeconceptual and inferential machinery into play as a precondition forour experience making a claim, thereby destroying the immediacyof its epistemic status (EPM §§6–7).

Rephrased somewhat, the difficulty is this: For our experiencesto make a claim that things are some specific way, there must bestandards of truth to which we can hold them. They must haveenough content so that there could be some way for things to besuch that their claims would not be true. This means that epistemicauthority requires the meaningful possibility of a gap between howthings are and how things are presented in our claims. In turn, whatmakes experiences authoritative in the face of the gap is their claimto living up to the standards which close this gap. The problem isthat epistemic status is a kind of normative status, as it must be sinceit essentially carries with it a form of authority. Sellars, furthermore,is committed to the following form of anti-reductionism:

The idea that epistemic facts can be analyzed without remainder – even ’in prin-ciple’ – into non-epistemic facts, whether phenomenal or behavioral, public orprivate, with no matter how lavish a sprinkling of subjunctives and hypotheticalsis, I believe, a radical mistake – a mistake of a piece with the so-called ‘naturalisticfallacy’ in ethics (EPM §5).

The normative standards that an epistemic state must live up to arethose which open up and then define how to bridge a gap betweenhow things are and how things might be. The standards therefore, inorder to get a grip, take us beyond what is immediately presented inthe state itself, and this state therefore ceases to be understandable asimmediately self-authorizing. Later, Sellars will fill out this critiqueby showing that the standards that mediate the claims of experi-

174 REBECCA KUKLA

ence and make it possible for them to be contentful must be alreadygrasped by an epistemic agent, who negotiates an articulated norm-ative space, before her experiences can have epistemic authority ofany sort, immediate or otherwise. But even at this early stage, thefact that an experience must make a specific claim before it can haveanything to be authoritativeaboutalready appears inconsistent withthe immediate authority of the experience. The given cannot serveits assigned role as the origin of epistemic authority, for in order forit to make an authoritative claim at all it must already have a locationin a normative space which is partially external to it.

We are now in a position to see that the Myth of the Given notonly is told from the perspective of hindsight (since the tellers of itare in fact trading in inferentially articulated, mediated claims), butthat it must beso told, whether or not it turns out to be a successfuljustificatory story. If we begin only with the epistemic materialprovided by given experience, we would not be able to articulatethe content of our knowledge, but could at best indicate it through amute ostension. It is only retroactively that I candescribethe contentof experience as something’s ‘looking green’ – in whatever sense Iknow this immediately according to proponents of the myth, I couldnot be applying concepts such as ‘looking’ and ‘green’ to it at thetime, since this would take it out of the realm of the immediate. Buttherefore we could not, from the perspective of this epistemic originin the given, find the resources to articulate the grounds for a movefrom the given to inferential, mediated knowledge. We manage toarticulate how the given grounds other knowledge only by movingbackwards, without leaving behind the resources available to usfrom our current perspective as epistemic agents within a mediatedspace of reasons. Our everyday observational concepts, accordingto the myth, get their grip from their grounding in given experience.But these concepts are needed to make sense of the legitimizingstory of this grounding, and hence this story seems to predate itsown possibility if we take it literally.

Overlooking the necessarily retrospective character of the appealto the given is what seduces us into the following mistake:

When we picture a child . . . learning hisfirst language,we of course, locate thelanguage learned in a structured logical space in which we are at home.. . . Butthough it iswewho are familiar with this logical space, we run the danger, if we


are not careful, of picturing the language learner as havingab initio some degreeof awareness – ‘pre-analytic’, limited are fragmentary though it may be – of thissame logical space (EPM §30).

Here we misrecognize the epistemic beginner as already possessingthe tools we use to tell the retroactive story of how we come tobe language-users. Of course, using tools we havenow to makesense of our past in ways we could not have done at the time is notnecessarily problematic. We face difficulties when we get confusedand believe that we can literally ascribe our sense-making justific-ations and resources to early knower. So, we can tell a story abouthow we came to be able to make claims that things were green onthe basis of things looking green to us, as a piece of backwardsjustification. But we cannot make the mistake of thinking that wecould, originally, begin with the immediate knowledge that some-thing looks green andreason to ourselvesthat therefore it mustbegreen. The backwards-looking justification thus serves as a legitim-izing myth rather than as a literal history; sense-making stories thatcan only be told retroactively become non-literal stories when theyare storiesaboutour past sense-making activities. However, this isa special problem for the Myth of the Given, which it will not befor other myths, for this myth’s specific premise is that the givenis immediately self-authorizing. But if this authority can show uponly in hindsight, then it seems to depend upon the mediation of ourprocess of myth-telling and recovery, rather than ‘adhering’ to theexperiences themselves. That is, if the given receives its authorityretrospectively, it cannot have it immediately, and hence the mythfails.

On my reading, we can take very seriously Sellars’ claim inEPM §1 that he is setting out to critique ‘Hegelian immediacy’,as well as his description of EPM as a ‘Hegelian meditation’. AsI have presented it, the structure and critique of the Myth of theGiven are nearly identical with Hegel’s presentation and under-mining of the knowledge of ‘sense-certainty’ (immediately known,self-authenticating takings-in of sensory experience) in Chapter Iof the Phenomenology of Spirit.12 Hegel also tells the story ofsense-certainly as a retroactive myth presented from our positionas mediated knowers. Sellars and Hegel both begin by attempting to‘remember’ the given, in order to show how the features which give

176 REBECCA KUKLA

it its place in the myth (its epistemic authority, its non-inferentialcontent) depend upon the very forms of mediate perceptual know-ledge that it was purported to fund. In order to show the impossib-ility of this origin. Hegel asks us to try to begin from the perspectiveof sense-certainty itself. What we discover in doing so is that fromthis perspective, we actually have no contentful knowledge at all,but only the empty generality of the ‘this’. The knowledge of sense-certainty, or the content of immediate experience, shows up onlyas retrieved from within a more sophisticated epistemic perspective,from which we already have available to us the conceptual resourcesof mediated perception. This undermines the epistemic authority ofsense-certainty, not because that authority can only be articulatedretroactively – for Hegel, this isalways the case for any form ofknowing – but because in addition, sense-certainty is specificallythat form of knowing that presents its authority as by definitionimmediate and self-legitimizing, and hence the fact that the legit-imacy of this authority is mediated by this retrospective storydestroys that authority altogether.

I have argued the given is mythical because it shows up withina retroactive story intended to serve a legitimizing function, andnot, as some commentators think, because it is a ‘false or wrongidea’.13 We tell the story for the purpose of legitimating our obser-vational knowledge, just as we tell the story of the social contract,not because we think there was one in history, but because we hopeto thereby give legitimacy to the authoritative claims we now make.Unfortunately, the Myth of the Given fails at this legitimizing task,because its starting point builds in the presupposition of the verysort of authority it seeks to ground.14 Sellars undermines the myth,not by showing its mythicality or its factual inaccuracy (neither ofwhich need be a topic of contention between him and the proponentof the given), but by showing thatas a mythit turns out to beincoherent and incapable of serving its own legitimizing function.Sellars’ detailed argument undercutting the success of the myth isnot yet complete; to finish the argument he will tell a second myth,the myth of John the Necktie Salesman.


3. HOW JOHN LEARNED TO LOOK

In Sellars’ ‘historical fiction’ about John the Necktie Salesman,John struggles to grasp that things might look green under certainnon-standard conditions such as bad lighting, yet not actually begreen. The moral of the story is that “the concept oflooking green. . . presupposes the concept ofbeing green”, and in turn that theconcept of being green makes essential appeal to the standard orproper conditions of observation, which are the conditions underwhich the way things look give authoritative information about theway things are (EPM §18). If this is right, then what is requiredin order that our experiences be able to make authoritative claims,in the way they failed to do within the Myth of the Given, isthat we already be conceptual and inferential negotiators of thespace of reasons, capable of recognizing authority and being heldto standards. Hence these experiences cannot serve as the origin ofepistemic authority. I want to show in this section how the rhetoricalstructure of Sellars’ telling of the Myth of John is essential ratherthan incidental to the philosophical point he is using it to make.The Myth of John undercuts the Myth of the Given, not merelyby giving an alternative story in which there is no given, but bygiving a temporal narrative that reveals the impossibility of the givenshowing up in the originary position in which it necessarily showsup within the original myth.

John’s ability to report authoritatively on colors is called intoquestion, in Sellars’ story, when he reveals that he is unable todistinguish between ties that look green under nonstandard lightingand ties that actually are green. He learns to make the distinctionbetween experiences of seeing that something is green and exper-iences of something’s looking green by learning the observationalconditions under which his experiences are authoritative. Once hehas learned to assess the authority of his own experiences in thisway, he can claim that he sees that various things are green. Hecan also now claim that things merely look green, but this becomesa contentful claim in virtue of his now understanding the properconditions under which he shouldwithhold from asserting that thislook makes an authoritative claim about how things in the worldstand (EPM §16). Thus we see that John could not in fact haveexperiences with epistemic authority until he was deeply enough

178 REBECCA KUKLA

embedded in the space of reasons to recognize and understand thenormative requirement that his reports live up to the standards thatguarantee that they live up to the tribunal of the world. To putthis another way, John must be able to recognize the authority thatobjectshave as the arbiters of the legitimacy of his claims, beforethese claims can themselves be authoritative. His claims must, asSellars puts it, be ‘wrung from him by the object’, so that ‘Nature. . . puts [him] to the question’ (Ibid.). The epistemic authority of hiswon experiences is not immediate, but derivative upon the authorityof things and his ability to recognize and be bound by this authority.To do this, he must be able to meaningfully posit a gap betweenthings and experiences, and he must understand enough about howepistemic authority works and when it becomes binding to knowwhen he can claim that this gap is closed. All of this, of course, takeshim far beyond the immediate and self-authorizing experientialknowledge posited by the Myth of the Given.

Sellars never calls the story of John a myth, which is noteworthygiven his insistent use of the language of myth during this telling ofboth the Myth of the Given and the Myth of Jones. This fiction qual-ifies as a myth given the broad definition I offered. Like the othertwo myths, it is a reconstruction of a fictional past, which purportsto reveal the nature of the epistemic authority that we now enjoy.However, it differs from the other myths in a couple of importantways. Rather than seeking tolegitimizeauthority, this myth aims toshow how such authority canfail. Since John is capable of engagingin normal conversations and activities when the story begins, he isclearly already a negotiator of an elaborate inferentially articulatedconceptual structure, and he can already recognize what it is forauthority to be legitimate and binding. The story shows where hisabilities break down rather than how they begin. Furthermore, thisstory does not rhetorically mask itself as a description about howthe past actually was. It presents itself only as a helpful fiction thatpoints us to an insight about how we know, rather than governing orgrounding our claims. Hence it is not playing a constitutive role inlegitimizing authority. These differences may or may not have beenbehind Sellars’ choice of different language to describe this fiction.

Let us look at how John’s story can be told from a couple ofdistinct perspectives. Once John has successfully learned how to


look, and can report appropriately on both how things look and howhe observes them to be, it seems that he can now look back andrememberhow things were before he had learned al this. He willlikely make statements about his past such as “When things lookedgreen to me back then, I thought I knew that they were green, but Iwas wrong”. In a sense, this is obviously a reasonable description onhis part. His past experiences will clearly show up in his memory asexperiences of things looking green. This might not take the formof his being able to make determinate ‘green images’ present tohis mind, but he will be able to use his memory to make reason-able claims about which of the past things he saw ‘looked green’.He also now knows that what he ‘experienced’ did not count asknowledge because he was not holding those lookings to standardsthat would justify according them authority concerning how thingswere. On the other hand, his memory is making available to hima piece of information that was not available at the time, namelythat something looked green. In fact he is necessarily misrecog-nizing the past when he talks about ‘when things looked green’ tohim, since the whole thrust of Sellars’ argument was that nothingcould meaningfully look green until we had the ability to know thatsomething was green. So although John’s early experiences reallydo show up in his memory as cases of something looking green tohim, these experiences show uponly in memory (and in the nextsection, we will see that it turns out to benecessarythat John be ableto engage in the project of remembering his learning processes andearly experiences). So we seem to be caught, paradoxically, sayingthat John can have a legitimate memory of a fact that was not true atthe time. Yet John’s memory cannot possibly reach into the past andchange any of the empirical facts of the situation.

How can this be? The trick is that while John’s memory is notchanging or embellishing the empirical facts, it is creating newnormative facts about the past that were not true at the time.15

Specifically, in memory he can recognize the epistemic role thathis experiences played or failed to play as evidence that some-thing was in fact green. Although our memory cannot change theempirical or natural facts about the past, time works importantlydifferently in the case of normative facts: it is possible for some-thing to become justified, or to come to play a legitimate role in

180 REBECCA KUKLA

authoritative knowledge, only after the fact, once we place it withinthe space of reasons through memory. Thus John’s memory canplay a constitutiverole in his past experienceshaving been casesof experiences that something looked green. This works because forSellars, something’s having been or being a particular determinateexperienceis a normatively fact, which essentially involves thatexperience having certain authoritative and legitimizing relations toother facts – we must avoid slipping back into the idea that there issome determinate character that ‘the experience’ has independent ofthe role it plays in structuring the space of reasons and determiningthe epistemic standing of the experiencer. In turn, this means thatthere is nothing paradoxical about saying that John’s memory, byactively telling a story which places his ‘past experiences’ in thespace of reasons, therebymakesthem the determinate experiencesthey (correctly) show up as being during the process of remem-bering. John’s memory is not merely reflecting (or failing to reflect)already-determinate facts, but retroactively constituting those facts,by telling a story about their epistemic standing. The active processof ‘remembering’ the story about the limits and possibilities of theepistemic authority of his experiencesis the process which createsand demarcates that authority, and thus helps constitute the exper-iences themselves. Thus we can say both that John’s memory thatsomething looked green to him is correct, and that this was not soat the time. His memory ‘misrecognizes’ the past, but this is an oldsense of misrecognition since there are no past facts with which hismemory conflicts, and in fact, through constitutive misrecognition,his memories turn out to recognize the facts properly.

Let us now switch perspectives, and instead of telling the storyfrom the perspective of hindsight, we will see what happens whenwe tell it as a forward-looking, chronological tale, beginning withhow John sees things before he learns to look. When his reportthat he sees that a tie is green is challenged, he cannot take thischallenge as a demand that he back down and only claim that the tielooksgreen, since he does not yet have the conceptual resources torecognize such a demand. In fact, he can tell no story at all aboutwhat is happening to him, for the events that he will later rememberas occurring are not ones that he now has the tools to understand orarticulate. As Sellars puts it, once he recognizes that his authority is


being somehow challenged, all John can say is “I don’t know whatto say!” (EPM §15).

Luckily, all is not lost for John, for he has plenty of other concep-tual and inferential abilities, and so he can at least realize that hisauthority is being challenged and that he had better come up with alegitimization of his claims or figure out why this cannot be done.As he begins to learn the logic of looks, therefore, he will have theresources needed to articulate what he is learning, and he can notethat he is learning to make a new kind of claim, and that he nowrealizes he had no warrant for the old claims. But at each step hecan narrate his own story only retrospectively, for the events becomedeterminate only in memory. Along the way, John’s learning willhave to proceed via a series of constitutive misrecognitions. Forinstances, when his peers ‘point out’ to him that the tie he thinksis green only looks green to him, this is again not literally true atthe time, since as they tell him this he is not yet capable of eitherthinking the tie is green or having the tie look green to him. Inpointing this out they misrecognize the situation, but they do so insuch a way that will help make it the case that later, in memory,John can see that this was the right thing to say after all, and hewill ‘remember’ that in fact he was making the mistake they wereaccusing him of making. His peers could be careful to ‘point out’only that the tie looks green rather than claiming that it looks greento him; but even in this case, in order to teach him, they must ‘pointout’ facts to him which, at the time they are pointed out, he doesnot have the resources to recognize through this ostension. We willbe troubled by this only if we insist on thinking it appropriate toask whether John’s peers’ accusations are literally true or literallyfalse. What we must see is that these accusations are authoritativedemands upon how Johnshouldunderstand what is happening. Tothe extent that the accusations serve a descriptive function, we needto say that they start out false and bring about their own correctness.But their essential work is to play a performative role, whereby theyattempt to govern and bind John’s epistemic behavior, rather than atraditional descriptive role. Thus the question of their literal truth orfalsity poses an inappropriate and inapplicable dilemma.

In order for Sellars to make his argument work, he cannotdistance himself from the temporal complications and the rhet-

182 REBECCA KUKLA

orical distinctions between story-telling perspectives that he hasintroduced. We see these complexities in the following pivotalpassage:

While [the conclusion of the story of John] does not imply that one must haveconcepts before one has them, it does imply that one can have the concept ofgreen only by having a whole battery of concepts of which it is one element.It implies that while the process of acquiring the concept green may – indeeddoes – involve a long history of acquiringpiecemealhabits of response to variousobjects in various circumstances, there is an important sense in which one hasnoconcept pertaining to the observable properties of physical objects in Space andTime unless one has them all – and, indeed, as we shall see, a great deal morebesides (EPM §19).

Here Sellars points to the need to tell at least two incommensur-able stories about the origins of conceptual knowledge. One the onehand, Sellars is committed to making room for the fact that as faras literal, forward-looking history is concerned, we must come tobe negotiators of normative space by evolving to display patternsof behavior of gradually increasing complexity. There cannot be anexplicable single moment at which our induction into the space ofreasons and our ability to make authoritative claims begin. For aswe have seen, we can only do this once we recognize the way thatthings make authoritative claims on us, and unless we are alreadynegotiators of normative space, such authority could never show upfor us in a way that could enable us to recognize it. Hence, barringour induction into normative space via a sudden and inexplicablesupernatural burst of spectacular insight, we need to say that in thisstory of gradual complexity, no moment will show up as the momentat which normativity begins. Yet, it is the case that at the beginningof this story we were not authoritative knowers, and at the end ofit we are doing things that count as making claims and wieldingepistemic authority, even while no elements of the story will explainthis transition.

At the same time, when we tell the story of the origin of authorityfrom the perspective of hindsight, from our positions within thespace of reasons, our induction into this space will always turnout to predate any moment that we try to place at the beginningof this induction. We must always understand ourselves retroact-ively as already in this space when we explain how we know whatwe know, in terms of our having made various mistakes, recog-


nized various facts and the like in the past. In other words, oncewe are knowers, we must always remember ourselves as alreadymembers of the space of reasons, insofar as we use our memory toexplain our epistemic status. We need to be able to ‘remember’ thissecond story, even though it is based on misrecognition, for other-wise we cannot legitimize our claims to epistemic authority. This isthe ‘important (retrospective) sense’ in which one has no conceptsuntil one has lots of them. If our ability to make claims sprungout of a pre-conceptual, pre-authoritative past, then our authoritywould be ungrounded and would cease to exist. This past needs tobe reconstituted in memory so as to provide this conceptually richgrounding; Sellars will make explicit the necessity of reconstructingour epistemic history through memory in EPM §§36–38.

Both these stories are ‘right’, at least in the senses that they areappropriate to tell and that they do not conflict with the facts. Neithercan be sensibly told from the perspective of the other, and in neitherone is there an explanatory story of the beginning of epistemicauthority. We should not be surprised that Sellars takes this doublestory-telling approach, since he attempts just such a combination ofincommensurable stories in PSIM (though the rhetorical structureof the stories is less complex in that paper). In the next section,I will give fuller attention to whether or not Sellars can get awaywith this use of mythical misrecognition and constitutive memoryin the context of a philosophically rigorous analysis of the nature ofexperience and epistemic authority.

4. MEMORY AND MISRECOGNITION

Sellars comes as close as her ever will to addressing the methodo-logical complexities of EPM over the course of a crucial and verydense series of paragraphs right in the middle of the essay, placedafter the story of John but before the Myth of Jones. I will devotethis section to a close reading of this series (§§35–38). This is thesection of EPM that is the most crucial for me, both in the sensethat on my account, these turn out to be the passages doing the mostphilosophical work for Sellars, and in the sense that these passagesprovide the most direct and rigorous support for my overall readingof the essay. In order to see how this section forms the linchpin of

184 REBECCA KUKLA

EPM, we need to attend to the rhetorical details of Sellars’ argumenthere with extreme care.

§35 presents ‘two hurdles’ which stand in the way of under-standing epistemic authority merely in terms of the ability to makereports that correlate reliably with the facts. This reliablist positionis in essence a deflationary response to the problem of explainingthe special role of authority in distinguishing epistemic claims fromother sorts of reliable indicators. To take this tact is just to put to oneside the various insights about the role of normativity in knowledgethat Sellars has been developing in EPM, and to insist instead thatepistemic authority is reducible to a type of reliable empirical gener-alization. For Sellars, an epistemic report that P is different fromother reliable indicators that P precisely because it reports P in virtueof beingboundby the claims of the world andsubjectedto epistemicnorms. The reliablist neither can nor admits the need to distinguishbetween these two types of case. The first hurdle thus “concerns theauthoritywhich, as [Sellars has] emphasized, a sentence token musthave in order that it may be said to express knowledge” (EPM §35).This hurdle really just restates the disagreement between Sellars andhis interlocutor. Granting for the moment his opponent’s right toignore his insights in the name of upholding the deflationary line,Sellars moves onto the ‘second hurdle’, which he says is ‘decisive’,and in fact shows the impossibility of ignoring the ‘first hurdle’:

For we have seen that to be the expression of knowledge, a report must not onlyhaveauthority, this authority mustin some sensebe recognized by the personwhose report it is (Ibid.).

The import of saying that the authority must be ‘in some senserecognized’ by the knower, I believe, is to return us to the insightthat authority is essentiallylegitimateauthority, and that legitimateauthority isbinding. We need not explicitly recognize what legitim-izes our claims at each moment, but we must be capable of makingsense of and responding to a demand that we justify ourselves. Thereliablist account of authority has no room for the fact that authoritydisappears and becomes mere coercion or coincidence when it doesnot hold us to something in virtue of being justifiable. Legitimacyand justification are, in turn essentially epistemic notions. For usto be bound by authority is for the legitimacy of that authority tohave a hold over us, and that can only be the case if we at least


implicitly understand and are in a position to justify that legitimacy.Thus to know something is to know (in some sense) why that pieceof knowledge is authoritative, which,contra the reliablist, dependsupon our recognizing the normative shape of the space of reasonsin a way which requires more from us than our ability to respondreliably.

Sellars’ inclusion of the qualifier ‘in some sense’ in the abovepassage serves an important purpose in leaving open the possibilitythat this recognition of authority need not be traditional recognitionof a literal, given fact, but might instead be a constitutive recogni-tion with the performative force of a demand. Sellars finally makeshis use of this type of move explicit in §36), where he begins tocomplicate the temporal picture:

Now it might be thought that there is something obviously absurd in the ideathat before a token uttered by, say, Jones could be the expression of observationalknowledge, Jones would have to know that overt verbal episodes of this kind arereliable indicators . . . This requires an abandonment of the traditional empiricistidea that observational knowledge “stands on its own feet” [or is self-authorizing]. . . it runs counter to the idea that we come to know general facts of this form onlyafterwe have come to know by observation a number of particular facts.

And it might be thought that there is an obvious regress in the view we areexamining [namely that all observational knowledge presupposes observationalknowledge] . . . This charge, however, rests on . . . amistaken conception of whatone is saying of Jones when one says that he knows that p . . . . The essential pointis that in characterizing an episode or a state as that of knowing, we are not givingan empirical description of that episode or state; we are placing it in the logicalspace of reasons, of justifying and being able to justify what one says(EPM §36,my italics in the second paragraph).

The first half of this quotation reiterates Sellars’ rejection of theproblematic account of observational authority employed by theMyth of the Given. The second paragraph gets to the heart of theissue (while making explicit the different status Sellars assigns toclaims about normative and empirical facts). As we have seen,observational knowledge, because it is knowledge, must be authorit-ative, and hence justifiable, and this justification seems to depend onother observational knowledge, putting Sellars at risk of potentiallyfatal regress. In the last section, it was this pressure that led meto conclude that once we are knowers, we must always rememberourselves as already having occupied the space of reasons. But froma literal, forward-looking perspective this is not the case.

186 REBECCA KUKLA

Sellars tries to alleviate this crucial tension in two steps, one ofwhich shows up here, and the next in §37. Here, we get the claim thatwhen werecognizethat a state (such as a past observation) countsat knowledge, we should not assume that this recognition takes theform of a reflection of a given fact. Our recognition, or ‘characteriza-tion’, does not function as a description at all, but rather a ‘placing inthe logical space of reasons’, or in other words, in normative space.Our characterization is thus ademandthat the purported case ofknowledge live up to certain standards and be bound by the claimsit makes. Furthermore, to recognize itas knowledgeis specifically,among other things, to recognize it asjustified, so for something tobe an example of knowledge it must explicitly play a specific norm-ative role in epistemic space. Hence, like the Althusserian hail, ourrecognition of its being knowledge may look like a description thatreflects fixed facts, but it is actually a performative act that makescertain demands and acknowledges certain commitments.

But how does this reinterpretation of what is involved in recog-nizing something as a case of knowledge help us out of the apparentregress? Once we see that recognition of facts about epistemicauthority takes a performative rather than a descriptive form, wecan argue that, as we saw in the case of John, this recognitioncan play a constitutive role in determining the normative facts thatare being recognized. This would not be a possible solution if therecognition in question were of regular empirical facts, rather thanstates defined by their normative status. But among the facts thatwe can constitute through recognition are ‘remembered’ facts aboutour own past epistemic states. This can potentially get us out of theregress, because our memories of ‘past’ observational knowledge– knowledge that justifies the epistemic authority of our currentsates – can be constituted through this memory, rather than needingto have literally and determinately existed in the past, in a waythat we could have recognized at the time. It isvery important tonotice that when Sellars says that authority must ‘in some sense’be recognized by the knower, he by no means insists that thisrecognition need occursimultaneouslywith the relevant authorit-ative claim. The recognition may – and sometimes mayonly –occur in memory, in which cast the claim becomes authoritativeonly retroactively. Sellars makes use of this move in §37, where


he distinguishes between what is known as it is happening and whatis known through memory. He presents this argument as though itwere a direct consequence of the argument from §36:

Thus, all that the view I am defending requires is that no tokening by Snowof‘This is green’ is to count at ‘expressing observational knowledge’ unless it isalso correct to say of S that henow knows the appropriate fact of the formXis reliable symptom of Y . . .And while the correctness of this statement aboutJones [that he has the relevant observational knowledge] requires that Jones couldnow cite prior particular facts as evidence for the idea that these utterancesarereliable indicators, it requires only that it is correct to say that Jonesnowknows,thusremembers, that these particular factsdid obtain. It does not require that it becorrect to say that at the time these facts did obtain hethen knewthem to obtain.And the regress disappears (EPM §37).

We need to sort through this crucial paragraph very carefully,taking our cue from Sellars’ consistent italicizing of the temporalterms and noting their complicated employment here. From §36, welearned that for me to know that Jones knows that ‘this is green’is for me to recognize his (normative) place within the space ofreasons, so that I take his claim to be authoritative, legitimate andbinding. Similarly, for Jones to know ofhimselfthat he knows that‘this is green’ would be for him to recognize the authority of hisown reports, and this involves his placing himself and his claimswithin space of reasons (which, we learned in §35, he must ‘insome sense’ do if he is to have the observational knowledge at all).For him to grant this authority to his own reports is for him totake himself as legitimately bound by them. But Jones legitimizeshis own authority, Sellars now tells us, through his knowledge ormemorythat in the past, his observational reports have correspondedto the facts under the right conditions and thus that they could betaken as reliable – he must remember that when he said somethingwas green, and said it under appropriate observational conditions,the thing was indeed green. In other words, Jones must recognizehimself asalreadya reliable reporter in order to give authority to thispresent report. For this to be possible, Jones must remember that thefacts that he reported actually occurred. But, Sellars argues, he neednot rememberknowingthese facts at the time; he must simply knownow that they obtainedthen. This is what ends the regress, becausealthough observational knowledge always depends on observationalknowledge, it need not always depend onpastobservational know-

188 REBECCA KUKLA

ledge. Instead we can ‘discover’ through memory that we werealready reliable reporters (though not knowers), and hence that weare justified in according authority to our present reports.

For this to be possible, Jones must be able to “remember” thatthe facts he reported upon reliably in the past actually occurred. Butnow Jones seems to have an analogous problem to the one Johnfaced when he remembered how things looked to him in the past:the relevant facts could not have been available to Jones at the time,since what gets us out of the regress is the assumption that in thepast he was not yet an epistemic agent of the right sort to knowthe facts. Thus the facts that need to show up in memory could nothave shown up in the past, which at first makes it seem impossiblethat they could be available to memory.16 We see the same problemin Brandom’s gloss on §37 in the study guide, where he gives theexample of a six year old who has moved into the space of reasons,and hence can look and ‘see’ that his former, four year old self hadthe right reliable responsive dispositions.

We have here a very similar structure of constitutive misrecogni-tion to that which we saw during my reading of the Myth of John:Jones, through the process of remembering, is giving a determinatecharacter to what did not have this character at the time, namelypast experiences of how things are. Whereas John reconstructedhow things looked, Jones reconstructs how he saw that things were.Both, however, are epistemic states which can be reconstituted inmemory because they are defined by their normative status. Solvingany potential paradoxes embedded in this move is more crucial herethan it was in John’s case, since then we were merely speculatingabout how John would look back upon what happened to him ashe learned to look, should he choose to do so, whereas we see nowthat Jonesmustbe able to remember these reliable correlations if heis to count as understanding why his reports now are justified, andhence if he is to count as a knower at all. Jones’ authority apparentlyrisks resting on a circle, despite Sellars’ attempt to end the regress.The reason that Jones can now reconstruct the past through memory,remembering facts that were not available to him at the time, isthat he is now a full-fledged, authoritative negotiator of the space ofreasons. But this authority is only legitimate in virtue of his ability to(mis)recognize the past. So just as we must deal with the “paradox


that . . . man couldn’t be man until he encountered himself”, and thusthat he seems to necessarily predate his own existence (PSIM 6),we seemingly must also confront the fact that authority can’t beauthoritative until it recognizes itself. Sellars is apparently troubledby the temporal complexities of this part of his argument, and awareof how critical they are to its success, since in a reprinted versionof EPM he returns to this point and is moved to add a somewhatredundant footnote to §37 that reads, “My thought was that one canhave direct (non-inferential) knowledge of a past fact which one didnot or even (as in the case envisaged)couldnot conceptualize at thetime”.17

This apparent circle would undermine Jones’ authority if theonly way his memory could function correctly were for it to givea literal reflection of past events. But we can alleviate the problemby reminding ourselves that the act of remembering, like the actrecognizing, can function as a constitutive demand. Jones insistsupon reconstructing his epistemic history in the right way through anactive process of remembering: his memories areachieved productsof this process rather than found contents of his mind. This ispossible because what his memory demands is not a change in ora creation of empirical facts, but rather a certain normative stancetowards ‘past’ experiences and reports; he demands of his memoriesthat they live up to the claims of the present.

There is an important difference between the demands beingmade by Jones’ memory and by John’s, despite their structuralsimilarities. John’s memory constituted new facts, albeit normativefacts, about the past, such as that ties had looked green to him.Jones, on the other hand, ‘remembers’ what reports he made, andwhat the facts were when he made these reports. These are empir-ical facts, and Jones had better be right about them – his memorycan’t change these sorts of facts retroactively, and his status as anauthoritative knower depends on his memory of such facts beingcorrect. What Jonesconstitutively‘remembers’ are not the empiricalfacts, but rather his own past observations of the facts upon whichhe was reporting. These observations are, of course, the way thefacts became available to him. But in turn this seems to raise a newproblem: if Jones’ memory is giving determinacy to experiencesthat did not have them at the time, isn’t he at risk of doing so in

190 REBECCA KUKLA

a way that leads him to incorrect conclusions about what the factwere? However, we need not worry unduly about Jones makinghis memories determinate in a way that conflicts with how thingsactually were, for there are all sorts of ways for him to check theempirical, publicly accessible facts about the past. Sellars remindsus at EPM §38 that “empirical knowledge . . . is rational not becauseit has a foundation, but because it is self-correcting enterprise”. Thushe can make sure that the facts he remembers were true.

The project his memory must undertake is the recovery andreconstruction of the fact that his reports werebound by andresponsive to the claims ofthe world, and here we reenter thedomain of normative rather than empirical facts. This will always,in a sense, be a misrecognition, since he cannot be truly bound bythe claims of things until he is an authoritative knower. However,again, his memory here is making a demand rather than describing,and the demand is that he understand his past responses as bound inthis way. Since what this requires is that he treat them as having acertain normative status, rather than that he misrecognize the empir-ical facts about what actually happened, there does not seem to beany inconsistency here. But this works only if we see that when hetakes himself to already have been bound by the facts in groundinghis present epistemic authority, his memory is functioningmyth-ically rather than literally. Our need to recognize and legitimizeourselves as authoritative knowers requires us to ‘remember’ ourpast responsesas if we were always already bound by things. Thismemory is neither a true description nor a false description of thepast, but an ongoingproject of mythical legitimization carried outthrough insistently treating the past as necessitating certain demandsand commitments.

The forward-looking, literal story of how Jones became able tomake authoritative observational reports will look different from themythical story that he must treat, in retrospect, as a telling of hisown past. Father down in §37, Sellars claims.

Thus while Jones’ ability to give inductive reasonstoday is built on a longhistory of acquiring and manifesting verbal habits in perceptual situations, and,in particular, the occurrence of verbal episodes, e.g. “This is green”, which aresuperficially like those which are later properly said to express observationalknowledge, it does not require that any episode be characterizable as expressingknowledge. (At this point, the reader should reread Section 19 above.)


Sellars here reminds us of the forward-looking historical storythat he committed himself to earlier. Jones gets to the point ofbeing able to be an authoritative knower, capable of ‘remembering’the past in the right way, in virtue of some gradual process ofdeveloping increasingly complex behaviors. But his chronologicalhistory, which gives the conditions under which he becomes aknower, will at no pointexplain his induction into the space ofreasons and his corresponding acquisition of epistemic authority,because this explanation would necessarily take the form of alegit-imization of his authority, and this legitimization, we have justbeen seeing, can be constructed only mythically and retrospectively.Hence it is a good thing that as we are telling the chronological story,we need not characterize its episodes as expressing knowledge, fortheir epistemic status will not show up within this story. The chro-nological story of increasing complexity and Jones’ ‘remembered’story will be different, not with respect to the empirical facts, butwith respect to the authoritative and normative status of the episodeswithin the story. In neither story will an origin of authority show up,for authority is such that by the time it shows up, it logically mustalready have been there.

Now that we see how one of these stories functions as a mythicallegitimization, which, while not contradicting the facts, works bydemanding that we treat the past in a certain way, we need not bebothered by the gap between the stories. Not all stories serve thesame methodological purposes, nor is there only one single relation-ship that a story can bear to literal reality in order for it to be agoodstory which successfully serves its purpose. Thus Sellars has theright to accept the legitimacy of two stories with two different meth-odological structures at once. Sellars falls short of making this wayof reconciling the two stories explicit. In order to address lingeringconcerns we might have about the tension between the stories, herefers us back to §19. But this is just the already cited sectionwhere he first laid out the two stories and their seeming tension,and then referred us forward to the argument he is in the processof giving. This frustratingly recreates the very mythical structurehe is revealing – his support for his current argument is a pastargument which he claimed at the time would only become determ-inate and complete later! In order to complete his argument, Sellars

192 REBECCA KUKLA

should have made it clear not only that recognition of normativefacts involves placing them within the space of reasons rather thanjust describing them (this was the moral of §36), but also how thisenables our recognition of the past to take the form of constitutivemisrecognition.

Sellars’ story about the grounding of epistemic authority involvesthe constitution of this authority through mythical misrecognitionof the past. But memory, in Sellars’ account, can at least poten-tially serve its legitimizing purposes better here than it did in thecontext of the Myth of the Given. The legitimizing myths we areexamining now do not misrepresent any of the facts, nor do theytell an inconsistent narrative. The Myth of the Given failed becauseit assigned an inconsistent role to supposedly originary epistemicstates. It was part of the Myth of the Given that the knowledge thatgrounded authority wasimmediateandself-authorizing, hence oncewe revealed that this knowledge only got its authority derivatively,and was mediated by our retrospective story-telling, it ceased to playits legitimizing role. In contrast, Sellars needs to give epistemicauthority to past states, but there is no reason why that authorityshouldn’t be derived retroactively from future retellings.

The key to understanding Sellars’ methodology, I have beenarguing, is to see that his mythical retellings serve a regulative ratherthan a reporting role. As long as our demands upon the past donot require us to actually contradict the facts – or as long as wecan continue to ask, as Sellars will later in EPM, “couldn’t it havehappened in this way?” – then these demands can work as legitim-izations at the same time, which constitute authority through myth.It is only once we understand the demanding, constitutive functionof mythical memory that we can properly make sense of the rhetoricand methodology of Sellars’ final and most complex myth, the Mythof Jones.

5. USING A MYTH TO KILL A MYTH

Sellars introduces the Myth of Jones §48, in order to show howwe can have inner episodes with epistemic standing and yet denythat these inner episodes are given immediately. This myth is there-fore intended to replace the Myth of the Given. We have seen that


knowers such as John and Jones ‘remember’ such inner episodeswhen legitimizing their own status as authoritative reporters. ThusSellars must show that such authoritative episodes still make sensein the face of the demise of the Myth of the Given. Constitutivemisrecognitions of the sort performed by John and Jones are allwell and good, but they cannot, in the end, provide us with falsefacts or appeal to impossible entities if they are to play their legit-imizing role successfully. We must make sure that the demands thatthese knowers make on how to understand their own experiences arecoherent ones. The last part of EPM comprises Sellars’ explanationof the existence of authoritative inner states by way of a legitimizingorigin myth. This is the only legitimizing myth in EPM which hetells in his own voice, rather than in the voice of his interlocutors (inthe case of the Myth of the Given) or in the voice of an imaginaryknower whose epistemic activities Sellars is examining from theoutside (such as John).

That the coda of EPM takes a mythical form is by no meanshidden by Sellars. His proclaimed method is “to answer these andother questions I have been raising by making a myth of my own”(EPM §48). He insistently and consistently uses the language ofmyth, fiction and temporal narrative throughout the rest of the essay,referring to the ‘denouement’ of his story, the ‘final chapter of ourhistorical novel’, and the ‘hero’ of the story. When we put togetherthis language with the argument of §§35–38, which reveals theineliminable role of myth and constitutive memory in ‘discovering’epistemically authoritative states, I believe it becomes clear that wemust attempt to understand the structure and methodology of theend of the paper (§§48–63) as specifically mythical. Furthermore,the Myth of Jones is a myth within a myth. Sellars tells a story, inthe course of which Jones the Genius uses the mythical methodsthat we have so far explored to ‘discover’ inner states and placethem appropriately within the space of reasons. It will thus turn outto be important to distinguish Sellars’ voice from Jones’ as we readthis part of EPM, and not to collapse them together as some readersdo.18

The Myth of Jones tells the story of ‘our Rylean ancestors’ –an explicitly fictional society whose language is purely behavior-istic, containing no references in inner or private episodes. Jones

194 REBECCA KUKLA

is the ‘Rylean genius’ who ‘discovers’ inner states by developinga succession of two theories, one which posits inner-speech-likethoughts, and one which posits impressions or inner experiences.Sellars argues that inner states or episodes can enter our conceptualspace as theoretical entities and still bereal, just as theoretic-ally introduced scientific entities can be real (as Sellars believesthey normally are). Jones can therefore come to count as havinggenuinely discovered an inner realm.

Jones’ first theory is that utterances are the culmination of aprocess beginning in inner thought episodes, and he models thesethoughts upon the very linguistic reports which they purport toexplain (EMP §56). At the end of his discussion of this theory,Sellars makes a complex move that will, once dissected, reveal themythical structure of both Jones’ story and Sellars’:

For once our fictitious ancestor, Jones, has developed the theory that overt verbalbehavior is the expression of thoughts, and taught his compatriots to make useof the theory in interpreting each other’s behavior, it is but a short step to theuse of this language in self-description. . . . And not it turns out –need it have?– that [others] can be trained to give reasonably reliable self-descriptions, usingthe language of the theory, without having to observe [their] overt behavior. . . .Our ancestors begin to speak of the privileged access each of us has to his ownthoughts.What began as a language with a purely theoretical use has gained areporting role(EPM §59, first italics mine).

Thus Jones tells a story about verbal behavior expressingthoughts, and Sellars tells a story about how this story comes toenable us to use Jones’ theoretical language toreport our ownthoughts. If Jones’ theory can come to support and enable thesereports, then we must assume that it is at least implicitly part ofthat the theory that the thoughts were already the kind of thing thatcould, in principle, be noticed and reported upon without referenceto the evidence of overt behavior, using (what would later come to becalled) ‘introspection’ instead. However we must also assume thatpeople were not noticing their thoughts through introspection beforeJones formulated his theory, for otherwise he would not have to positthem as theoretical entities, as he does in Sellars’ myth, but couldinstead cite them as direct evidence supporting his theory. Thus histheory must help constitute the inner accessibility of thoughts, whilealso ‘discovering’ these thoughts, which show up within the theoryas the kinds of things that were already there awaiting discovery.


That is, Jones’ theoretical language takes the forms of a report onthe inner, and eventually it does get used in this reporting role, butit must take this form even before we learn to introspect.

Now it seems consistent to claim that Jones might have developedhis theory, and upheld this theory because of its coherence andexplanatory value, without its ever turning our that we are able toknow our thoughts introspectively (as, one might plausibly claim,Freud developed a theory of unconscious thoughts and their relationto behavior). So how are we to read Sellars’ odd interjected question– need it have turned our that we can report our thoughts in this way?The answer is ‘yes’, but not because this result is demanded by thelogic of Jones’ story. Rather,Sellars’ myth must have this ending,simply because we can in fact do this! It is our current recognition ofepistemic authority that necessitates reconstructing a past in whichintrospectively accessible thought will be possible. It must be a partof Sellars’ myth of Jones that Jones’ theory turns out to be authorit-ative, and hence that the thoughts turn out to have been recognizableas already there so that we could learn to report on them. But atthe same time, thoughts modeled on utterances makeclaims andhence wield epistemic authority, and we know that such authority,for Sellars, must be recognized in order to get off the ground. Hencethere could not, literally, have been thoughts there to report beforeJones ‘discovered’ them. Thus if the story, in the end, must cometo the conclusion that Jones discovered real thoughts that are nowaccessible to us, then we must say that Jones’ theoretical activityplayed a constitutive role in the existence of the inner, even as thetheory formally misrecognizes thoughts as the kinds of things thatwere already there to be discovered.

The same goes for Jones’ second theory, which is that impres-sions are the inner end results of the impingement of physicalobjects and processes on the body (EPM §60). We have seen thatthese impressions only become determinate, reportable experiencesonce we are already able to distinguish between a look and a seeing-that. But we can only make this distinction if we have already been‘noticing’ our impressions and how they may diverge, under inap-propriate observational conditions, from the world (which is whythe impressions cannot function as the ‘given’). When Jones’ theoryposits these impressions, and thus sets the ground for our learning

196 REBECCA KUKLA

how to notice them, his theorizing is therefore again playing aconstitutive role. For as Sellars puts it in EMP §45,

Now we recognize that instead of coming to have a concept of something becausewe have noticed that sort of thing, to have the ability to notice a sort of thingis already to have the concept of that sort of thing, and cannot account for it(Sellars’ italics).

Yet the theory describes these impressionsas if they were alreadythere to be discovered – its constitutive role does not show up withinthe theory itself. In order to justify the legitimacy of the realistlanguage the theory will employ, in the face of these complications,Sellars gives us an analogy:

Notice that while our ‘ancestors’ came to notice impressions, and the languageof impressions embodies a ‘discovery’ that there are such things, the language ofimpressions was no more tailored to fitantecedentnoticings of these entities thatthe language of molecules was tailored to fit antecedent noticings of molecules(EPM §62).

By using the molecule analogy, Sellars makes this conclusion soundless radical and less problematic than it in fact is. As realists, we caneasily make sense of the idea that molecules existed determinatelyprior to our having the epistemic tools required for observing them,and thus that we could genuinely discover them by theoreticallypositing them even though we could not discover them on the basisof observation. But Sellars is here glossing over his own results from§§35–38, which reveal a crucial disanalogy. Impressions, unlikemolecules, are themselves entities within epistemic space, definableonly in relation to other epistemic notions, and this means theycannot exist independently of being ‘in some sense recognized’;in particular, they cannot exist prior to being posited by the theorywhich enable them to be ‘noticed’. Thus our ‘discovery’ of impres-sions, unlike our discovery of molecules, must take the form of aconstitutivemisrecognition, in that it presents itself as if it wereuncovering things that already existed. Sellars’ scare quotes around‘discovery’ in the above passage therefore seem quite appropriate,and they belie his attempt to make his myth turn out sounding likea traditional realist’s story. Within Sellars’ myth, impressions arediscovered as ‘always already there’, but they cannot have literallyand determinately existed prior to their discovery. This does not


mean that impressions aren’t real after all, but it does mean thattheir temporal status and the conditions of their existence are morecomplex, on Sellars’ account, than he makes them sound here.

Once Jones develops his dual theory, he has a couple of possib-ilities available to him. He may misrecognize his own theory ashaving literally uncovered already-existing states. He thereby

mislocates the truth of these conceptions, and . . . confuses his own creative enrich-ment of the framework of empirical knowledge, with an analysis of knowledge asit was. He construes asdata the particulars and arrays of particulars which hehas come to be able to observe, and believes them to be antecedent objects ofknowledge which have somehow been in the framework from the beginning. It isin the very act oftakingthat he speaks of thegiven(EPM §62).

What we have here is the myth of the birth of the Myth of the Given,which turns out to be what Jones gets if he fails to notice his ownconstitutive role in producing the structures of epistemic authoritythat he ‘discovers’. Alternatively, he can take a more sophisticatedattitude towards his own theoretical activities, and recognize that his‘discoveries’, which present themselves rhetorically as descriptionsfor the purposes of the explaining and justifying our observationalknowledge, are in fact demands upon how we are to explain andunderstand experience. Our inner states must serve as the tribunalthat grounds the authority of our claims about these states, and thusthe theory must treat themas if they literally predate these claims.To avoid the Myth of the Given, Jones must realize that his is not astandard empirical theory giving a literal description of how things(already) are, but a myth that plays a performative and constitutiverole in legitimizing epistemic authority. Sellars’ own myth does notmake this sufficiently clear, since it presents Jones as giving a theoryin a much more traditional sense.19

But now we are able to see the sense in which the Myth of Jones‘kills’ the Myth of the Given. By revealing the temporal complex-ities of Jones’ theory, and seeing that it enable the ‘discovery’ ofpotentially noticeable entities of a funny sort that could not benoticed in advance of this discovery, we see that these entities cannotpossess immediate epistemic authority, both because our access tothem is mediated by the relevant theoretical understandings, andbecause any epistemic claims they make must be derivative uponother authoritative claims that we must already be able to recog-

198 REBECCA KUKLA

nize and make. We can also see now that the illusion of givennessis produced by the mythical structure of the Jones’ theory itself,which must ‘recognize’ the inner, retrospectively, as already determ-inate and accessible. If we do not notice the retrospective statusof this recognition, we can easily think that such an inner, onceuncovered, is the perfect thing to place explanatorily at the originof our authoritative epistemic processes. Whereas Sellars’ earlierarguments in EPM already demonstrated the Myth of the Given tobe riddled with incoherence, the Myth of Jones kills the Myth ofthe Given bysubsumingit and changing its status and performativerole. Jones’ myth, taken properly, does not place accessible innerstates at the origin of authority nor accord any of them immediate,self-authorizing status. The original Myth of the Given now showsup, within Sellars’ myth, as Jones’ misconceived literalization andtemporal rearrangement of his own process of theoretical discovery.

Sellars ends EPM with a question:

I have used a myth to kill a myth – the Myth of the Given. But is my myth really amyth? Or does the reader not recognize Jones as Man himself in the middle of hisjourney from the grunts and groans of the cave to the subtle and polydimensionaldiscourse of the drawing room . . . (§63).

If we put together the fact that (a) the myth of Jones is historicallyimplausible and Sellars has presented no evidence at all for its literaltruth, (b) he has used the language of fiction and myth throughoutthis part of the essay, and (c) he puts this passage in question form, Ibelieve that we ought not to read Sellars as making a literal sugges-tion here. In virtue of (a), Echelbarger (1974) points out that therole of the Myth of Jones in EPM would be incomprehensible, andany conclusions it leads us to so weakly supported as to be useless,if we read the argumentative force of the myth as depending inany way upon its literal or semi-literal truth. In virtue of (b) and(c), we have good evidence that Sellars is trying to do somethingmore interesting with the Myth of Jones, rhetorically speaking, thanpursuing implausible armchair historical hypotheses. We owe it tohim as readers to try to find a subtler reading of the question he askshere.

Furthermore, an alternative reading is available. Byaskingus ifwe recognize ourselves in Jones, he asks us to attempt to understandour own past in this way. If we can succeed at this imaginative task,


which involves making sense of the idea that our ability to observeour own inner statescouldbe the result of theoretical activities likeJones’, then this myth can serve its legitimizing purpose. Even ifthe story is false, we can work to maintain a mythical stance forthe purpose of our practices of justification, and (mis)recognize ourpast in this way so as to ground our claims about the inner. On thisreading, Sellars is exactly right to claim, as he does twice inScienceand Metaphysics, that the Myth of Jones is a rational reconstructionplaying a role analogous to that played by social contract myths(1968, pp. 71, 155). We use each myth to legitimize (and perhaps inminor ways correct) the claims of authority as we now understandthem. The usefulness of the myth does not depend upon its truth oreven its plausibility, but upon its function as amethod for regulating,governing and legitimizingour position in a space of norms. Likethe social contract myth, this myth serves a normative rather thana descriptive purpose, although it presents itself in the guise of apossible descriptive history. We can see that Sellars is in fact issuingan imperative in §54, where he moves from claiming that wecantell stories true to the principles of methodological behaviorism toclaiming that weshoulddo so. In order for us to be able to followthe mythical imperative, the myth must express a coherent possiblehistory – a requirement that I shall return to in a moment. Butwhile its possibility is crucial, its plausibility and its probability arecompletely beside the point. The question form that Sellars employsnow makes sense: if the myth asks us to misrecognize our past forcertain purposes of legitimization, then we do not want to eitherassert its literality (which would just be a mistake) nor to deny it(which would undercut its ability to serve its pragmatic purpose).Rather, we must maintain a stance like the one suggested by Sellars’question, wherein we activelykeep openthe possibility that the mythis true, but withoutassertingits truth.

The Myth of Jones tells the story of the discovery of the inner,and within this myth, Jones tells the story of how the inner wasalready there to be discovered. Sellars’ myth does not commititself to the literal truth of Jones’ myth, but on the contrary, under-mines the possibility of taking it literally, even while Jones’ story-telling plays a crucial role in Sellars’ story. Within Sellars’ myth,public, intersubjective claims fictionally and conceptually predate

200 REBECCA KUKLA

private, internal thoughts and experiences. But Jones’ myth ‘recog-nizes’ these internal things as predating public utterances much asmolecules do. Which is the ‘true’ order of dependence? To ask thisis to misunderstand the methodological status of both myths. Just as“there is a logical dimension in which other empirical propositionsrest on observation reports, [and] there is another logical dimensionin which the latter rest on the former” (EPM §38), here too we mustsee that there is no single answer to questions of logical depend-ence, outside of the context of a particular story serving a particularmethodological purpose.

My reading of Sellars’ methodology at the end of EPM is import-antly different from Brandom’s reading in his study guide to theessay. Brandom notes that the Myth of Jones is ‘explicitly putforward as a myth’, which on his reading, means that:

Sellars is not claiming that things actually happened this way, that we really hadRylean ancestors, or owe our concepts to a primitive genius (never mind onecalled ‘Jones’). Sellars’ pragmatism dictates that issues of conceptual priority betranslated into questions of the relative autonomy of different strata of language –that is, into questions concerning what language games can be played independ-ently of and antecedently to which others. Telling an as-if historical, develop-mental story is a way of exhibiting those relations of conceptual dependency andpresupposition. (Brandom, 1997, p. 170).

While Brandom is right to point out that this is an ‘as-if’ story, hehas translated the legitimizing role of myth into a merely exhibitiveor descriptive role. The performative force of the myth is lost here,as is its role in constituting the very relations of conceptual depend-ency that Brandom sees it as exhibiting. These methodologicalfeatures of the myth are similarly lost on Echelbarger’s reading,in his article “Sellars on Thinking and the Myth of the Given”(1974), the only article I have found that makes an explicit issueof the nature and function of myth in Sellars. Echelbarger considersthree possible readings of the mythical status of the Myth of Jones:It might function as an armchair, speculative history, or a parableabout the psychological development of children, or a roundaboutway to make an abstract claim about the logical priority of speechover thought (p. 234). The first and second options render the mythuseless, since it is either far too unsupported to do any epistemolo-gical work or patently false. The third possible reading once againhas the shortcoming of missing the performative role of the myth,


and it also suggests that there is only one answer to this question oflogical priority, which, I have argued, is not the case for Sellars.

The Myth of Jones can only succeed in its legitimizing functionif it is in fact a coherent possibility (though again, it need not behistorically plausible). This, in turn, depends upon its being in factpossible for there to be a theory such as the one Jones supposedlydevelops, which systematically links inner and outer states. Sellarscertainly doesn’t give us this theory, or even point in the directionof where to look for it. We know that the theory takes the form ofa model, but all we know is what the source and target domain ofthe model is, without knowing anything at all about the theoreticallinks between the domains. As long as we have no such links, thetheory will be empirically empty, and as Echelbarger points out,Jones’ theoretical entities will “predict only the phenomena theywere introduced to explain” (1974, p. 238). Until we have at leastsome minimal hints about how the elaboration of the theory mightproceed, we cannot assess whether such a theory could plausibly bedeveloped.

Since Sellars gives none of these hints, he cannot be said to havecompleted his task of showing that the Myth of Jones provides acoherent possibility for how inner episodesmight haveentered ourepistemic space by first being introduced as theoretical posits. Tothis extent, his myth cannot count as serving its legitimizing func-tion. For the myth demands that we understand the past as if ithappened a certain way, but not only do we not have a complete ideaof what such an understanding would come to, we don’t yet know ifthis type of understanding is even possible. If we are not sure sucha theory can exist, we cannot use the Myth of Jones to completelyground the authority of our claims about the inner. If Sellars hadintended the Myth of Jones to be taken literally, this objection wouldbe fatal: we cannot authorize our claims via their history if there areno hints of any historical details that would tell us how the legitim-izing story would go. Since Sellars’ approach is instead mythical,the problem remains troubling but perhaps not yet fatal. In the faceof the demand that, for purposes of legitimization, we actas if thisstory had occurred, it seems we can meet this demand up to a point.However, without any details about Jones’ theory, we just don’t

202 REBECCA KUKLA

know how to go about fully meeting this demand, nor whether doingso is ultimately a coherent possibility.

6. CONCLUDING REFLECTIONS

I have tried to show that the guiding problem of EPM is the problemof legitimizing epistemic authority, and that for Sellars, addressingthis problem requires searching for ways to narrate the origin ofthis authority. EPM is primarily concerned with how the spaceof reasons, or dimensions of this space,open upfor us. Further-more, I have argued that Sellars’ solution to this problem takesan essentially mythical form, and that his account of our inductioninto normative space relies upon our constitutively misrecognizingour epistemic history. If we look carefully at Sellars’ arguments,as I have read them here, we will notice that his turn to a myth-ical foundation for the normative is necessitated specifically by hiscommitment, made explicit at EPM §35, to the idea that genuineepistemic authority must be ‘in some sense recognized’ by theknower. Throughout my sections 3, 4 and 5, which comprise theheart of my reading to EPM, it was the demand that werecognizeour own epistemic authority that in each case led to the need forconstitutive misrecognition through memory. We saw that in orderto recognize our own authority, we must tell a story about the pastwhich legitimizes that authority, for to recognize it as authorityis to recognize it as legitimate. But the peculiarities of myth andconstitutive misrecognition enter in because we must rememberhaving always already been able to recognize authority if the pastis to be usable in our legitimizing stories in the right way.

Sellars argues, as we have seen, that a non-mythical, forward-looking story of gradually evolving complexity is insufficient toexplain normativity. From the perspective provided by such a story,normativity will never begin, because at no point can a claim showup and be recognized as authoritative and binding if until that pointwe had not yet been recognizers of authoritative claims. In PSIMSellars reminds us that we must avoid an account that leads us intothe following trap:

To be able to think is to be able to measure one’s thoughts by [normative] stand-ards of correctness . . . In this sense a diversified conceptual framework is a whole


which, however sketchy, is prior to its past, and cannot be constructed as a comingtogether of parts which are already conceptual in character. The conclusion isdifficult to avoid that the transition from pre-conceptual patterns of behavior toconceptual thinking was a holistic one, a jump in level of awareness which isirreducibly new, a jump which was the coming into being of man.(6)

This suspect conclusion – that we enter normative space in asudden burst of ungrounded understanding – is incoherent within aSellarsian framework. If there were a moment at which conceptualthinking began and the space of reasons opened up, this momenthave to be a beginning of authority not grounded in previousauthority. This is in fact a temporally indexed version of the Myth ofthe Given, since it must treat this first moment of authority as self-authorizing. On the other hand, a story which explains our behaviorbut not the source of our epistemic authority will not do as a storyabout our past for the purposes of justifying the authority we havenow. For a story that does not show us how we became authoritativemight causally explainbut cannotlegitimizeour being authoritativenow, and it is this legitimization story that we need in order to beable to recognize our own authorityas authoritative. Sellars pointsout in PSIM that conceptual thought always emerges within thespace of conceptual thought (or the space of reasons), and hence wecannot explain it by an appeal to an origin in the non-normative,because for conceptual thought to think itself out of conceptualspace is precisely for it to lose its legitimate authority and epistemicstanding (pp. 16–17). Hence as long as we remain committed toa story about authority that makes some essential reference to therecognition of that authority, it seems we must follow Sellars inrejecting forward-looking natural histories purporting to explainnormativity, and turn instead to mythical stories.

It is possible that we can sidestep the whole problem of recog-nition of authority depending upon constitutivemisrecognition ofauthority by adopting the approach articulated by McDowell inhis influentialMind and World(1994). I will not try ot arbitratebetween McDowell and Sellars here; instead I wish to indicate howMcDowell provides us with a genuine alternative to the Sellarsianaccount I have been developing, through beginning with a rejectionof Sellars’ central assumption that authority depends upon recogni-tion. McDowell agrees with Sellars that no natural history revealing

204 REBECCA KUKLA

an origin of normativity is possible. But unlike Sellars, this does notcompel him to search for a retrospective, constitutive origin myth.Instead, he claims that there could be no good narrative of the originof our induction into normative space of any sort, because we arealways already within it, though we may need to learn to recognizeits features. Like Sellars, he thinks that when we seek to understandepistemic authority, we must understand ourselves as always alreadyhaving been bound by norms, but for him this is literally true ratherthan a mythical constitutive misrecognition. McDowell is thereforecommitted to claiming that authority needs no subject at all, and canbe binding even if it is not taken as binding by anyone. The worldmakes claims even if those claims are not recognized, and thus whenwe seek to legitimize our states as epistemically authoritative, wecan appeal to how wereally werebound in the past even if we didnot know it.

For both Sellars and McDowell, our epistemic claims can haveauthority only if they are in turn subject to the authority ofthings;the world binds us and serves as the tribunal of our claims. But forSellars, the world does not have authority over us in this way unlesswe in some sensegive it this authority. We must take on the commit-ment of making claims that are true to the world and hold ourselvesto this standard before the world can make a binding claim on us.Objects maydemand of usthat we make certain claims (these claimsare ‘wrung from us by the object’ – EPM §16), but they can do soonly in virtue of our being already committed to recognizing thesedemands and taking them as binding. Sellars thus finds himself ina circle, wherein epistemic authority depends upon recognition ofthe authority of things, and recognition of the authority of thingsdepends upon our already being negotiators of the space of reasonswho are capable of this recognition. By telling a retrospective myth,we can get this circle off the ground by constitutively misrecog-nizing ourselves in memory as already within the space of reasons,bound by the authority of things and committed to respecting thisauthority. For McDowell, on the other hand, things have authoritywhether or not we are polite enough to give it to them, and comingto be bound by this already-present authority involves coming torecognize the claims things already make. If McDowell’s position isa tenable one, then the need for a mythical grounding of authority


will never arise. The disagreement between Sellars and McDowellcomes down to their disagreement over the claim I made back inSection 1, namely that authority requires a subject that it makesclaims upon, and that one can be such a subject only if one canin some sense recognize the legitimacy of those claims.

Sellars’ story about how normativity and authority fit into thenatural world is, as we have been seeing, exceptionally complex. Hedoes make use of a story about the natural history of normativity– or at least, he makes use of his commitment to the in-principleexistence of such a story, though he never makes any attempt to giveit even in its broad outlines. This story gets him out of the paradoxof claiming that in a literal, chronological sense, we must have lotsof our concepts before we can have any of them. But epistemic andother normative concepts will not show up within that story, exceptin the minimal sense that by the end of it we turn out to be doingthings that count as making claims, being bound by norms, and soforth. Instead, the stories that he uses to explain normative facts takea mythical, retrospective form and involve constitutive misrecogni-tion. These stories by no means count as naturalizing stories in anytraditional sense, since they of necessity treat normative space asalready open for the purposes of explanation, and since they retro-spectively play a constitutive role in creating the types of facts theyexplain.

Still, commentators often read Sellars as a naturalist, who, asRorty puts it in the Introduction to EPM, “makes it possible to under-stand mind as gradually entering the universe by and through thegradual development of language, as part of a naturalistically explic-able evolutionary process, rather than seeing language as an outwardmanifestation of something inward and mysterious” (1997, p. 7).20

While it is certainly right that Sellars doesn’t understand languageas an outward manifestation of something inner and mysterious, Ido not see how we can read EPM as enabling us tounderstandmind as gradually and naturalistically entering the universe. Wecannot point to any place in EPM where Sellars does anything morethan assert that a naturalistic history of mind must in principle bepossible; he does not argue for this, but merely takes it as a piece ofcommon sense which he wishes to show does not conflict with theother account he is giving. The text is focused on his mythical story

206 REBECCA KUKLA

about the emergence of the inner, which is incommensurable withthe natural history (as we know both from reading EPM carefully,and from reading PSIM where he asserts the incommensurability ofthe two stories). The mythical story does not describe the gradualemergence of mind, but rather begins in the present, and looks backretrospectively to show how the inner gets recognized as alwaysalready there. Sellars, it seems to me, is not trying to make possiblethe type of naturalistic understanding that Rorty describes at all, butis rather attempting to give an explanation of a radically differentsort, which at best manages not to conflict with the naturalistic,non-explanatory facts of the matter.

It may seem that I am claiming that far from telling a naturalisticstory, Sellars is completely casting aside the tenets of naturalism thatareau courantin contemporary philosophy, and which he elsewhereprofesses to uphold. Whether this is fair depends on what sort ofnaturalism we have in mind. Manning (1997) distinguishes betweenthree types of naturalistic commitments. ‘Reductive naturalists’insist that “only physical entities and processes and the explanatorymethods suitable to these are naturalistically permissible” (Ibid.).Clearly Sellars is not engaged in a reductive naturalistic project.He disavows the possibility of explaining epistemic facts in thesereductive terms right at the outset, as we saw (EPM §5). ‘Scient-istic naturalists’, on the other hand, “see traditional philosophicalquestions as demanding natural scientific answers or no answersat all” (Manning 1997). The demand for a naturalistic account ofsomething, on this view, “is just the demand for a theory fromscience itself” (Ibid.). We cannot really tell whether a given positionis consistent with scientistic naturalism until we have a story aboutwhat counts as a legitimate scientific explanation or theory. But itseems that we would be hard-pressed to count Sellars as a scientisticnaturalist, given that for him, there are genuine (normative) facts thatcanonly be explained mythically, and which are in part constitutedby our process of ‘remembering’ their mythical history. One wouldthink that explanation via mythical constitutive misrecognition hasgot to count as non-scientific explanation if anything does, unlessby ‘scientific’ we simply mean ‘good’ (in which case the jury is stillout). On the other hand, it turned out that Jones’ scientific theor-izing, in the Myth of Jones, itself made use of mythical constitutive


misrecognition. So we must either say that Jones does not count asdoing science in the story, or we must countenance the possibilitythat Sellars’ own uses of myth might be acceptable to the scientisticnaturalist. Finally, a ‘respectful naturalism’ is one whose “imper-ative . . . above all is to ensure that our philosophical theories do notcontain elements of what, from the point of view of natural science,is beyond explanation, [or] supernatural” (Ibid.). It might appear asthough Sellars has violated even respectful naturalism, in claimingthat mythic memory can constitute determinate facts about the pastthat were not true at the time. But as we have seen, memory cannotchange or reconstruct the empirical facts, but only the normativefacts, which it does by understanding and locating past events withina normative space organized by relations of authority, demand,commitment and the like. Epistemic states are standings in the spaceof reasons, and I have tried to show that there is nothing mysteriousor supernatural about our ability to relocate past events within thatspace through memory. Furthermore, Sellars goes out of his wayto argue that this process is consistent with a natural history thataccounts for all of the empirical facts without remainder. He thusappears to be fully committed to the spirit of respectful naturalism.And although I certainly won’t defend this claim here, one mightreasonably suspect that respectful naturalism is the only kind worthphilosophically fighting for.

My reading of EPM has aimed to find a place for myth in philo-sophy. The mythical, I have tried to show, can play a role which isneither merely metaphorical, nor a substitute for genuine argument,but is rather full-bloodedly epistemological, and (when employedproperly), a genuine form of philosophical methodology. A carefulreading of EPM reveals that Sellars could not have substituted literalor pseudo-literal stories for his myths, for the force of his argumentsdraws crucially upon their mythical structure. The specific functionof the mythical in EPM shows up at the level of the particular mythshe uses and discusses (such as the Myth of the Given, the Myth ofJones, and John’s mythic reconstruction of his epistemic history).But it also shows up at the level of the text as a whole, whichis structured by the complex and sometimes constitutive relationsthat his myths bear to each other, as they nest within, undermine,supplant and draw upon on another.21

208 REBECCA KUKLA

ACKNOWLEDGMENT

I would like to thank the National Endowment for the Humanitiesfor supporting my participation in their summer institute on Back-ground Practices and Intelligibility (Santa Cruz, CA 1997), duringand with the help of which I formulated many of the ideas for thispaper. I am grateful for the help and insights of Tim Adamson, DanGold, Katharine Loevy, Amy Lund, John Lysaker, and Joe Rouse.Mark Lance gave me extremely helpful comments on a penul-timate draft. I am especially indebted to Richard Manning, whosupplied nearly-endless conversation, comments, criticism, supportand inspiration, and without whom this paper quite literally wouldnot exist.

NOTES

1 See Brandom (1994, 1997), McDowell (1994) and Rorty (1997). The import-ance of recognizing this theme in Sellars’ work appears to have been a mainmotivating factor in the recent reprinting of EPM in the form of a self-containedbook. All these commentators also recognize that this is a move which had alreadybeen made in different forms by earlier figures such as Kant, Hegel and Wittgen-stein.2 See Hegel’sPhenomenology of Spirit(1977), Althusser’s “Ideology and Ideo-logical State Apparatuses” (1971), and McDowell’sMind and World(1994) forthree very different versions of this line of argument.3 This latter strategy is the one adopted by McDowell in his recent bookMindand World(1994). See section 6.4 See for example Garfield (1989) and Echelbarger (1981).5 See Robinson (1975) and Echelbarger (1974).6 For instance, Robinson (1975) sets out to show that the Myth of the Given isnot a myth at all, but rather a ‘legend’. Unfortunately, Robinson’s initial definitionof myth is not promising. The first line of his paper tells us, “To say of an objectthat it is myth is to imply that the object does not exist” (p. 63). The idea thatan objectas opposed to a narrative could be a myth already seems misguided.Myths (such as the Myth of Oedipus) may be about specific objects, but surely itis only in sloppy parlance that we call an object a myth. Equally problematic isRobinson’s equation of mythicality with simple non-existence, or, in a narrativeterms, its fictionality. This starting point blinds him to any philosophical use fora myth other than telling it in order to show its falsity. In particular, it rendersincomprehensible Sellars’ attention to developing his positive myth, the Myth ofJones, along with his claim in EPM that he is using a myth to kill a myth – surely


one can’t use something non-existent to kill something nonexistent, nor is it evenclear what such a killing would involve.7 Mythical legitimations of authority, from the social contract to the originalposition, have been time-honored tools for philosophers. Of course, various philo-sophers may use these as mythical tools in only a very minimal sense; for instancewhen Rawls uses the original position in the context of devising a test for justice,he does not offer us a full-blown narrative of the past with either literal or non-literal status. But he is still asking us to imagine what we would decide if wewere not alreadypositioned within society and aware of this position. Hence histest still works mythically, in my sense, since it asks us to place ourselves in anon-literal (in this case counterfactual) past in order to make sense of normativefeatures of the present.8 Though others, such as Brandom (1997, pp. 157–159), disagree that this recog-nition plays any necessary role in being bound.9 See especially his discussion of interpellation, misrecognition, and the produc-tion of the subject (pp. 167–177).10 There is a question to be asked here concerning how we couldfirst recog-nize somethingas a demand, before we are in any way negotiators of normativespace. Similarly, we may wonder at what point our urge to respond ‘appropriately’becomes indicative of our genuine recognition of a normative claim, rather than amere impulse or habit devoid of normative import. On the difficulty of answeringsuch questions about the beginnings of normativity, see later sections of this paper.11 Sellars points out that some proponents of the myth take the given to be partic-ulars rather than facts. But as he argues, this makes their epistemic status and theirrole in inference incomprehensible.12 Hegel (1977), pp. 58–66.13 For example Robinson, as we saw, equates the mythical with the nonexistent,and takes Sellars, therefore, as arguing simply that there is no given (1975, p. 83).14 See Althusser (1982) and Kukla (1995) for arguments that at least Rousseau’sversion of the social contract myth also turns out to fail its legitimizing purposes,for analogous reasons.15 Richard Manning has accused me, or Sellars, of relying upon an untenabledistinction here. After all, empirical facts, for Sellars, are in a sense always norm-ative ‘all the way down’, in the sense that they can make determinate, authoritativeclaims on us only when we take them as normatively related to other facts. Thiswas in fact a moral of the last section. So how can he rely on a distinction betweennormative and empirical facts? A normative fact, for our purposes, is one whichis at least partiallyaboutthe structure of normative space itself. As will becomeclearer in the next section, to state a normative fact is not merely to describe a stateof affairs, but rather to make a set ofdemands. Standard examples like ‘killing iswrong’ are obvious cases; more interestingly, ‘p is evidence’ is a normative factin a way that ‘p is green’ is not, even if we grant that the claim ‘p is green’ is onlymeaningful as placed within normative space. To say that ‘p is evidence’ is notjust to report on a state of affairs, but to make a claim about what sort of authority

210 REBECCA KUKLA

p has. To say that ‘p is green’ commits us to various inferences and other claims,and generally embeds us in the economy of epistemic authority, but it does notmake a claim about this normative melee.16 Sosa (1997) points out this same tension in §37, though towards quite differentends.17 Quoted by Sosa (1997) in his note 10.18 See, for instance, the interchangeability of Sellars’ voice and Jones’ in thereadings of the Myth of Jones given by Brandom (1997, especially pp. 180–181)and Garfield (1989, p. 22).19 Although in Science and Metaphysics, Sellars admits rather cryptically thatJones is giving something which is both importantly like and importantly distinctfrom a regular empirical theory. In addition, Echelbarger (1974, p. 238) convin-cingly argues that there are various reasons to be suspicious of taking Jones, as heis presented in EPM, to be engaged in the process of traditional theory-building.20 Brandom makes almost the same claim in his study guide to EPM, for instanceat 1997, p. 130.21 In comparison to past readings of EPM, this reading brings Sellars’ meth-odology far closer to Hegel’s, particularly the Hegel of thePhenomenology ofSpirit.

REFERENCES

Althusser, L. (1971): ‘Ideology and Ideological State Apparatuses’, inLenin andPhilosophy and Other Essays, trans. Ben Brewster, New York: Monthly ReviewPress.

Althusser, L. (1982):Montesquieu, Rousseau, Marx: Politics and History, trans.Ben Brewster, London: Verso.

Brandom, R. (1994):Making it Explicit: Reasoning, Representing, and DiscursiveCommitment, Cambridge, MA: Harvard University Press.

Brandom, R. (1997): ‘Study Guide’, in Wilfred Sellars (ed.),Empiricism and thePhilosophy of Mind, Cambridge, MA: Harvard University Press.

Echelbarger, C. (1974): ‘Sellars on Thinking and the Myth of the Given’,Philosophical Studies25, 231–246.

Echelbarger, C. (1981): ‘An Alleged Legend’,Philosophical Studies39, 227–246.Garfield, J. (1989): ‘The Myth of Jones and the Mirror of Nature: Reflections on

Introspection’,Philosophy and Phenomenological Research50, 1–23.Hegel, G.W.F. (1977):Phenomenology of Spirit, trans. A.V. Miller, Oxford:

Clarendon Press.Kukla, R. (1995):Conformity, Creativity, and the Social Constitution of the

Subject, Ph.D. Dissertation, Department of Philosophy, Pittsburgh, PA: Univer-sity of Pittsburgh.


Manning, R. (1997): ‘Function and Intentionality’, presentation delivered to theDepartment of Philosophy, Albuquerque, NM: University of New Mexcio,September 1997.

McDowell, J.H. (1994):Mind and World, Cambridge, MA: Harvard UniversityPress.

Robinson, W.S. (1975): ‘The Legend of the Given’, in H.-N. Castañeda (ed.),Action, Knowledge and Reality, Indianapolis, IN: Bobbs-Merrill.

Rorty, R. (1997): ‘Introduction’, in Wilfred Sellars (ed.),Empiricism and thePhilosophy of Mind, Cambridge, MA: Harvard University Press.

Sosa, E. (1997): ‘Mythology of the Given’,History of Philosophy Quarterly14,275–287.

Sellars, W. (1963): ‘Philosophy and the Scientific Image of Man’, inScience,Perception and Reality, Atascadero, CA: Ridgeview Publishing Co.

Sellars, W. (1968):Science and Metaphysics, London: Routledge and Kegan Paul.Sellars, W. (1997) [1956]: ‘Empiricism and the Philosophy of Mind’, Cambridge,

MA: Harvard University Press.Wittgenstein, L. (1953):Philosophical Investigations, trans. G.E.M. Anscombe,

New York: Macmillan Publishing Co.Zizek, S. (1989):The Sublime Object of Ideology, London: Verso.

Department of PhilosophyCarleton University1125 Colonel By DriveOttawa, ON K1S 5B6Canada


Myth, Memory and Misrecognition in Sellars' \"Empiricism and the Philosophy of Mind\"

Documents