Borges's Menard 2013 Text&Work

Text and Work: The Case of Menard, PetrKotat’ko, ed. (Prague: Univcrzita Karlova v

Praze, 2013). – DRAFT

Borges’ Menard: His Work and HisIntentions

James Hamilton

Abstract: Authors make assumptions about the fitness andadaptability of their “materials” – such things asmultiplicity or simplicity of sentence structures,techniques for making action appear simple or complex,techniques for achieving depth or shallowness of character,detailed or spare descriptive language, and so on – to theiraims. But, in this paper I show that this reasonable notionof what is involved in authorial intentions is not only atodds with, but is actually obscured by, conceptions ofauthorial achievement that must be presumed in the standarddiscussions of the ontology of art that have employedBorges’s story.Keywords: ontology of art; works of art; adaptability offit.

In her 1988 essay, ‘Losing Your Concepts’, CoraDiamond describes a way in which the adoption ofa particular philosophical theory can renderconcepts that have been necessary unavailable forthe adequate description and analysis ofimportant human concerns. Diamond focusesattention on the way deflationary theories oftruth make it impossible to articulate any sense

James Hamilton

in which truth itself may be said to be valuable.And she assembles powerful examples of people whohave claimed that truth is valuable in order toshow the importance of such claims and theimpotence of deflationary theories of truth whenfaced with the challenge of making sense of them.

In this essay I maintain that something likethis kind of loss has occurred with respect tothe use made of Jorge Luis Borges’ story, ‘PierreMenard, Author of the Quixote’, in philosophy ofart. Here, however, it is not a particular theorythat renders important concepts unavailable tous. Instead, I maintain, the concept of authorialintention – so prominent in Borges’ story –disappears as a robust concept of intention whenthe story is taken to raise questions in theontology of art. It is the focus on thosequestions that renders a robust concept ofauthorial intention unavailable to us.

To make good on this position, I undertakeseveral tasks. First I sketch the alternativepositions on the ontology of artworks that havebeen thought to be supported or challenged by theseveral chapters and part of another of DonQuixote that Borges’ narrator says Menardproduced. Second, I extract from the ontologydebate a position on Menard’s intentions that isconsistent with the positions taken in thedebate. Third, in contrast, I sketch an accountof authorial intentions derived from attending to

160

Borges’ Menard: His Work and His Intentions

details in Borges’ story. I also provide groundsfor thinking that Borges’ implied account ofauthorial intentions is largely correct. And Ishow how we should use this account to explainwhat makes Menard’s authorial intentionsquixotic. Fourth, I develop an objection to thecentral claim of this paper, an objection thatasserts that we must settle the ontologicalquestions even if we adopt the conception ofauthorial intentions I have extracted fromBorges’ story. Finally, I show that any way wehave of stating that objection clearly, it fallsout that the conception of authorial intentionsextracted from Borges’ story loses its robustnessand gets distorted in such a way that it nolonger describes anything like what authorsactually do.

1. The ontology debate, and views of Menard’sachievement

The stage is set for the use of Borges’ story indiscussions of the ontology of works of art byNelson Goodman’s claim that

A literary work…is not the compliance class of a textbut the text or script itself. All and onlyinscriptions and utterances of the text are instancesof the work; and identification of the work frominstance to instance is ensured by the fact that the

161

James Hamilton

text is a character in a notational scheme – in avocabulary of syntactically disjoint and differentiatedsymbols. Even replacement of a character in a text byanother synonymous character […] yields a differentwork. (Goodman 1968: 209)

This claim, that work identity is text identity,is what Borges’ story has sometimes been taken tochallenge.

The challenge has been made in at leastthree, equally unsuccessful, ways. The firstchallenge, at least historically, is found in anessay by Anthony Savile (Savile 1971) whereinSavile uses an account of the story to contestNelson Goodman’s view that notational orsyntactical identity of a text alone issufficient to establish the identity of aliterary work.1 Key to Savile’s account of thestory is the claim that Menard has composed astory with a different meaning than that of thestory composed by Cervantes and that, therefore,although the texts are notationally identical,Menard has produced a different work ofliterature (Savile 1971: 21-22).

1 There are two previous sightings of the Borges story.Zemach 1968 discusses the story in connection with offeringan account of time sensitive properties of works of art.Agassi 1970 also mentions the story, but in a quitedifferent context with quite different implications. Thanksto Gary Iseminger for pointing me to this latter sighting.

162


Goodman has a brief and fairly obviousdefense to the thought underlying Savile’scontention there are two works here. That asingle text can have more than one interpretationis, of course, no evidence for claiming there ismore than one literary work at hand. Indeed,Goodman and Catherine Elgin offered preciselythis defense in 1988. They observe that, even ifwe construe Menard as having the ambition tocreate a text notationally identical withCervantes’ and even if that ambition wereachievable, what we should then conclude is that

what Menard wrote is simply another inscription of thetext. Indeed [...] if infinitely many monkeys [...]produce[d] a replica of the text [...] that replica[...] would be as much an instance of the work DonQuixote as Cervantes’ manuscript, Menard’s manuscript,and each copy of the book that has ever been or will beprinted. (Elgin – Goodman 1988: 62)

A second attempt to use the story to challengeGoodman is found in Arthur Danto’s ‘Artworks andReal Things’ (Danto 1973) to assist him inmotivating the problem that was to become thecornerstone of his later book, TheTransfiguration of the Commonplace (Danto 1981).The book’s title suggests the contents of boththe article and the book well enough: in themDanto asks us to consider how it can come aboutthat commonplace objects become works of art; and

163

James Hamilton

he provides an answer, namely, that commonplaceobjects become works of art, if they do, whenthey are transfigured by being taken as being‘about something’ or, in other words, by being‘interpreted’ in terms of a theory of art.

Of course, the very idea that commonplaceobjects can just become works of art, in anymanner or by any means at all, needs somemotivating. It is not enough that Duchamp puts abottle rack forward in a gallery or museum, evenwith a title attached. Nor will it suffice thatmany members of the museum going public, galleryowners, art-lovers, and purchasers of art acclaimthe bottle rack as an ‘artwork’. For, as B. R.Tilghman (1982: 293) puts it, “there is no demandgrounded either in logic or in sensibility thatobliges us to the same”. If philosophy is tosettle this – and I am not sure it can – what isneeded is an argument or a compelling andincontrovertible case, something that does appealto logic or analysis of sensibility. Duchamp’s‘Readymades’ do not fit the bill preciselybecause they are the controversial objects thatarguments or appeals to cases are supposed tohelp us come to resolve.

Rather than a compelling case, Danto offersan argument. He reasons that, if such objectscould become works of art, the change in statuscould not be in virtue of any changes in theirperceptible properties (for the simple reason

164


that there are no such changes). Moreover,whatever could bring about the change in status,it must be something that is involved,essentially, in anything’s being a work of art.That is, Danto reasons that the identification ofsomething - when it is a work of art- as the workof art it is, cannot be the result of aninference from a description of only itsperceptible properties. Correlatively, Danto(1973: 5-7) thinks it must be possible for thereto be different works of art that areperceptually indiscernible from each other.

What Danto is thought to have done here isshow that identification of works of art demandsspecification of not only intrinsic but alsoextrinsic, specifically relational, properties ofthe work. This view may be true. But Danto hasnot shown it by his use of Borges’ story. AsTilghman points out, “the only ground,” Tilghman(1982: 298) writes, “for saying that Borges hascontributed to literary ontology by discoveringthat [historical and critical matters bearing oninterpretation] are constitutive [of the identityof a work] is the claim that Menard has reallywritten a new work.” But, the grounds Dantooffers for the latter claim just are thedifferent relations between Menard and Cervantesand their two (albeit notationally identical)texts, the relations between the texts and whatwould be their place in literary history were

165

James Hamilton

they different works, and so on. In fact, asTilghman shows, Danto’s connection betweenconstitution and individuation appears to rest ona circular argument or none at all, “to come as apackage”.

David Lewis (1978) provides a third route tochallenging the Goodman position. Lewis maintainsthat Borges’ story illustrates that where thereare “different acts of storytelling” there are“different fictions” because a fiction is “astory told by a storyteller on a particularoccasion” (Lewis 1978: 39). However, as GeorgeBailey has observed, what bearing Lewis’s claimshave on issues about the ontology of artworksdepends on whether we construe the expression‘different fictions’ to mean ‘different works’.If not, then the fact - if it is one - thatMenard has written a different fiction from thefiction written (on another occasion) byCervantes in no way counts against there beingjust one (version of) the Quixote. But, if‘different fictions’ does mean ‘different works’,Lewis is faced with the serious difficulty ofdenying, implausibly, that literary works canhave multiples. Lewis, Bailey (1990: 349) notes,“tries to avoid this by noting that producing acopy is not an act of storytelling [...] [but

166


then, and even more implausibly] by hisdefinition copies are not instances of fictions.”2

2. Menard’s intentions as understood in theontology debate

Although the discussion so far seems to favorGoodman’s stance in the ontology debate, that isfar from my purpose in this essay. We can take astep towards the actual purpose by noting that,in the middle of this debate about work identityin art has been an assumption about the nature ofMenard’s intentions in producing the chapters ofthe Quixote that he manages.

In Flint Schier’s discussion of the story wefind a rehearsal of the claims first stated bySavile and later implied by Danto. But what headds is revealing: he characterizes Menard’sintention as that of “creat[ing] a work that willbe the same, word for word, as Cervantes’ DonQuixote.” And, he concludes, “If Menard hadsucceeded in his mad ambition...his work wouldnot be Don Quixote by Cervantes, but rather theemanation of a peculiar symbolist poet of theearly twentieth century” (Schier 1986: 28).2 I am grateful to Bailey for sending me a copy of thisessay. Without it, I would also have missed two notable‘sightings’ of reference to the Borges story, those ofSavile and Schier.

167

James Hamilton

Pretty obviously, Schier’s conclusion is offeredas a criticism of Goodman, suggesting that Menardhas, indeed, produced a notationally identicaltext that is a new work of literature. But theimportance of Schier’s contribution lieselsewhere; for he is the first in the debate totry to articulate the content of Menard’sintention. This characterization has become, withfew exceptions, the standard story about Menard’sintentions among parties to the debate.3 If onethinks Menard realized his intentions, sodescribed, one sides against a generallyGoodmanian ontology; if one thinks Menard had notrealized his intentions, so described, onesupports a roughly Goodmanian line on ontology.

Christopher Janaway (1992) offers an enrichedaccount of Menard’s intentions and examines howdoing so helps sort things out ontologically.4

Janaway asks us to imagine the case of apsychological experiment in which two people –let us call them A and B – are given slips ofpaper. A is told to write down the name of acolor or arrange a set of words into a string andB is told to write down what she thinks A willwrite down (Janaway 1992: 75). What is there tocount against our saying that both A and B are3 Even Elgin and Goodman accept it. They seek only toseparate the question of Menard’s achievement from thequestion of his intention (Elgin – Goodman 1988: 61).4 His essay is a reply to Wreen 1990.

168


composing even if, on the bizarre chance itshould happen that what is written down on thetwo slips of paper coincides exactly?

We might be inclined to say this is a case inwhich either both or neither is composing. Thatis, if writing down a color word or arranging aset of words into a string counts as compositionat all (or not), it should count (or not) boththe case in which there is no guiding idea and inthe case in which there is a guiding idea, namelythat of writing down what the other person willwrite down. But in fact the latter seems to havegreater claim to be a case of composition thanthe former.

Neither of these descriptions of Menard’sintentions squares easily with Borges’ story. (Iwill argue later that they are also inadequate asaccounts of authorial intention, considered moregenerally.) In this regard notice first howBorges sets out Menard’s project: “[Menard] didnot want to compose another Quixote – which iseasy – but the Quixote itself”. Nor did he intendto copy it out; “his admirable intention was toproduce a few pages which would, word for wordand line for line, coincide those of Miguel deCervantes” (cf. Borges 1964: 39). Now notice thatthere are two parts to this intention. Menardintends a particular product, namely, that thereshould be some pages that coincide with those ofDon Quixote. But Menard also intends to compose

169

James Hamilton

the Quixote: on the one hand this was not to be acopy, not to be plagiarized; but, on the other,the text was not to coincide entirelyaccidentally with the Quixote. The product wasintended to be the result of the intendedactivity.

Schier’s account of Menard’s intentionsclearly does not measure up, at least when takenas an account of Menard’s intentions consistentwith Borges’ story. This is both because it failsto conform to the to the first part of Borges’characterization (in that it makes Menard aim atmore than that at which Borges has Menard aim)and because it disregards the second part, theintention to compose, altogether.

Janaway’s account of Menard’s intentionsmight be thought to fare better because itappears to give us a handle on why what Menard isdoing could be described as ‘composing’. But, onreflection, this will not do. Somethingcritically important is missing from Janaway’scase. As Borges tells it, the idea is not thatMenard (1) intends to write down some words and(2) hopes the resulting text coincides with somestretches of Don Quixote. One way to characterizewhat Janaway has omitted is this: there must bean intended link between the writing down and thenature of the text produced in order for us tohave a case of ‘composing’. This is part of whatI meant when I earlier wrote that the product is

170


intended to be the result of the intendedactivity. Another way of putting this point5 isthis: Menard may have peculiar literaryambitions, and it may be that some of hisliterary ambitions are quixotic; but they areliterary ambitions. The conjunction of (1) and(2) do not express a literary ambition. At bestthey can be used to set up a ‘thought experiment’of a kind that is a staple of ontologicaldiscussion.

In contrast, Borges’ statement of Menard’sambition – “to compose the Quixote” – seemspeculiarly telling. Yet, what can he have meant?

3. Authorial intention as implied in Borges’story

3.1. Constraints on accounts of authorial intentionsBefore trying to formulate Menard’s intentionsmore explicitly, I believe it would do well toconsider a pair of constraints on any formulationwe might provide.

5 At any rate, I think these had better be close to the samepoint. More precisely we may expect that an analysis of whatit is to compose a literary work and an analysis of what itis to put literary ambitions into motion will turn out to bethe same analysis.

171

James Hamilton

The first constraint I have in mind is this.There is a sense in which not even Cervantes canhave intended to compose Don Quixote. Andwhatever constrains Cervantes in this respectshould also constrain Menard, so long as what weimagine is that he intends to compose theQuixote. Now I do not see why we cannot imaginethat he intends something, analogous to whatCervantes intended, that results (howeverimprobably) in producing the Quixote. For I donot think we know enough about the world topredict the Quixote could just never appear inthis way; nor do I think we know what we wouldactually say about it if it did (despite someglib claims by certain philosophers in anenthusiasm of anti-Goodmanian fervor). And so Ithink if Menard intends to produce the Quixote,under this constraint on what that can mean, welland good.

In this regard, it might be helpful to thinkof composing as analogous to solving a problem inalgebra. One cannot set out to solve a particularalgebra problem and, simultaneously, intend thatthe expression of the solution just be ‘x–1’. Onecan, of course, intend to express the solution tothe problem as ‘x–1’. But then one will havealready worked out the solution and cannot besetting out to solve it. Composing a novel, poem,and so on, is like this: when setting out tocompose one does not know how the composition

172


will turn out. Borges is clearly in touch withthis idea, for he has Menard write the followingin a letter:

When I was ten or twelve years old, I read it, perhapsin its entirety. Later, I have reread closely certainchapters, those which I shall not attempt for the timebeing. I have also gone through the interludes, theplays, the Galatea, the exemplary novels, theundoubtedly laborious tribulations of Persiles andSegismunda and the Viaje del Parnaso [...]. My generalrecollection of the Quixote, simplified byforgetfulness and indifference, can well equal theimprecise and prior image of a book not yet written.(cf. Borges 1964: 41)

The algebra analogy can also serve to remind usthat, just as actually solving a problem inalgebra does not happen by accident but as aresult of the intention to solve a problem,composing the Quixote, for example, does nothappen by accident but as the result of theintentions of the author, however we end upcharacterizing them.

Borges suggests a second constraint on anyadequate formulation of authorial intention whenhe writes that Menard considered two ways hemight arrive at composing the Quixote; the first,dismissed by Menard as “too easy”, would be tobecome Miguel de Cervantes; the second would beto compose the Quixote out of his own (20thCentury) experience. Any way we have of coming to

173

James Hamilton

terms with either of these compositional routeswill have to deal with the issues:

A. what materials of writerly composition wereavailable to Menard or Cervantes;B. what aims could have been in view for Menard orCervantes;C. whether the materials deployed are appropriate tothe aims in whose service they were deployed.

This third issue initially may seem more a matterof critical judgment rather than any part of thecontent of literary ambition. So set it aside forthe moment.

The first two issues appear to be subjectsfor inquiry in literary history. Clearly Menardcould have Cervantes’ material available. In factBorges takes pains to stress what Menard did toacquire them. Menard might even be able to chooseto adopt Cervantes’ aims, but perhaps onlynostalgically. This is why, even if it isnonsensical on other grounds, Menard should thinkof the alternative of becoming Cervantes: were heto do so, his aims would just be those ofCervantes, and writing the Quixote withCervantes’ aims would be considerably lesschallenging.6 Alternatively, Menard might attempt

6 Not without challenges, however, as we shall see in thenext two sections. One way to put the matter here would bejust to say that it is unclear whether even Cervantes couldhave had the aim of composing Don Quixote, even though he

174


to put Cervantes’ materials to use to other, 20thCentury, ends. I suppose it must be somethinglike this that Menard chooses to do. But,characterized in this way, there is no particularphilosophical oddity in Menard’s project. It isquixotic, of course; and what makes the projectquixotic is that Menard chooses to use exactlythe same materials, in exactly the same ways, forthese other ends. I will soon show that there isa way to account for our sense Menard’s intentionis more than merely a little odd, that there isin fact something inconsistent or incoherent inthe offing, by appeal to the third issuedescribed above. For, I believe, we sometimeshave to attribute to authors, as part of theirnormal intentions, a specific intention withrespect to the appropriateness of materials toaims.

3.2. An abstract formulation of authorial intentionsWithin these constraints I now suggest thefollowing schematic of any authorial intention:

(a) I intend to assemble (literary) materials torealize some project;

did compose it and even though his composing it was noaccident.

175

James Hamilton

(b) I intend that the assembled materials will allow,induce or invite my audience, the readers, to havecertain reactions or reaction-types; and(c) I intend that the readers’ reactions come about asa result of the fact that the materials actually do fitmy aims and [some version of a judgment of fitnessclaim] and [possibly, some version of an adaptivityclaim].

Some general remarks are in order. Clearly thisformulation allows us to get at two of the themesI claim one has to contend with in order tounderstand what Menard is supposed to haveintended: (one) what materials of writerlycomposition were available to Menard orCervantes, and (two) what aims could have been inview for Menard or Cervantes.

I have introduced the third clause for tworeasons. First, I think an author would think shehad failed to realize her ambitions were she todiscover that readers had reacted exactly as sheintended but demonstrably had done so because ofother features of the composition than those shesaw as connected to those reactions, or that theyhad reacted as she intended but for reasons thathad nothing whatever to do with the compositionitself. Secondly, the third clause constitutes arelatively uncontroversial expression of thethird theme mentioned above as possibly necessaryfor understanding Menard’s and Cervantes’intentions: (three) whether the materials

176


deployed are appropriate to the aims in whoseservice they were deployed.

The second and third conjuncts of the thirdclause are meant to be placeholders forsubstantive beliefs that render robust even thisabstract formulation of authorial intentions.Without beliefs of the kind that would fill inthese gaps, we might well wonder if the thirdclause accurately captures all possible authorialambitions with respect to the products of theiractivities. Noting two things can bring out whatI have in mind.

First, in addition to intending that readers’reactions come about as a result of the fact thatthe materials actually do fit the author’s aims,an author might also intend that this come aboutas a result, further, that the readers haveunderstood – by exercising their own criticaljudgment – that the various conventions,grammatical devices and other writerly strategiesthat are the author’s ‘materials’ do fit theauthor’s aims. In contrast, another author maywish the readers’ reactions to come about as aresult of the fact her materials fit her aims,but not intend the readers be aware of how thathappens, that it not happen as a result of thereaders exercising critical judgment. Suchbeliefs as these are what I have called ‘judgmentof fitness claims’.

177

James Hamilton

Second, authors may hold beliefs, just asanyone else might, concerning the degree to whichadaptivity of literary materials to literary aimsis historically conditioned. Such beliefs may beweaker or stronger. A relatively weaker anduncontroversial belief of this kind is the viewthat artistic aims not possible in one timeperiod may become possible at a later time. Thiskind of belief is what fills in the space forwhat I have called ‘adaptivity claims’. Such abelief can become part of what gives fullexpression to the content of an author’sintentions in the sense that, without referenceto such a belief we may not be able tocharacterize adequately and fully the author’sintentions.

I have posed these last points as a pair ofoptions with respect to which an author mightadopt a point of view with regard to qualifyingthe intention that readers’ reactions come aboutas a result of the fact her materials do fit heraims. One might think we should settlephilosophically which ways of responding to eachof these options are correct. I do not think thisis so and, as I will now show, it is for thatreason I do not think we can determinephilosophically what makes Menard’s intentionsincoherent. What I do think philosophy can do forus here is make clearer what are the elements ofauthorial intention, what elements must be

178


present and, where options are available to anauthor, what those options consist in andseverally entail.

Suppose we adopt a fairly strong formulationof the judgment of fitness claim in order to givean account of literary ambitions. A strongformulation would hold that, to be described ashaving literary ambitions, an author must bethought of as intending to assemble materials insuch a way that allows, indeed invites, one’saudience not only to have certain reactions andreaction-types, as a result of the fact thematerials do fit the aims, but also to exerciseits own critical judgment about what materialsare appropriate to what aims.7 If we adopt thisstrong formulation, we may be led to say that onecannot reproduce – let alone set out to reproduce– an already existing work unless one has exactlythe same aims as the author of that existingwork. We certainly can be led to this conclusionif we also adopt a strong view about theadaptivity of materials and ends. If we – incontrast to the author – believe only one preciseset of means is adaptable to any precise set ofaims, then we are committed to the view that noauthor could attempt to invite her readership todo what she invites it to do by assembling for7 Here I am developing an idea about authorial intentionsregarding ‘fit’ (as between materials and aims) to be foundin Wollheim 1984.

179

James Hamilton

new aims the same materials already deployed inone work for another set of aims. She would haveto suppose them incapable of seeing that themeans are fit for earlier aims while supposingthem capable of seeing, and indeed inviting themto see, that the means she deploys are fit forher new aims.8 On these grounds there seems to bereal evidence of something inconsistent orincoherent in the Menard’s project (but thatinconsistency or incoherence has yet to bespelled out precisely).

To arrive at this point we have adopted astrong formulation of the judgment of fitnessclaim as part of the expression of normalauthorial intentions and adopted a strongposition on the degree to which adaptivity ofmaterials to aims is historically conditioned. Aweaker version of the judgment of fitness claimcould hold that the author intend only toexercise her own judgment concerning fitness ofmaterials to aims without expecting the same fromher readers. That is, this author could bedescribed as intending to assemble materials insuch a way that allows and invites her audienceonly to get the point (because the materials do

8 Indeed, this is Wollheim’s way of handling the matter. Onealso should note our reliance on the view that, to intend todo a one cannot know one cannot do a. This is likely to bethe least controversial part of Wollheim’s way of handlingthe matter.

180


fit the aims) without consciously deciding forthemselves that the materials fit the author’saims. This will still render Menard-like aims asquixotic if we still accept the strong adaptivityclaim. But if the author does not likewise adoptthe strong adaptivity claim,9 there will benothing incoherent about her intentions. And ifwe drop the strong adaptivity claim, not only areher intentions coherent, they also lose even theappearance of being quixotic. This suggests aline we can take in capturing the incoherency inMenard’s intentions.

3.3. Why (and how) Menard’s intentions areincoherentWe are able to demonstrate Menard’s intentionsare incoherent in part by attributing to him astrong version of the judgment of fitness claim(noted in italics). Menard, we may say, intends(in part) the following:

I intend that the readers’ reactions come about as aresult of the fact that the materials actually do fitmy aims AND that they have this reaction as a result ofexercising their own critical judgment that thematerials are appropriate to the aims.

9 The situation may be different if she merely has doubtsabout whether she can use the old materials for new aims.For a discussion of this issue, and related matters about‘side effects’ not directly addressed in the present essay,cf. Mele 1992.

181

James Hamilton

Then, to complete our account of what makes hisintentions incoherent, we attribute to him astrong version of the adaptivity claim. Menard,we may say, believes the following:

Only one precise set of means is adaptable to anyprecise set of aims.

Now, one might think we want another way toarrive at ‘Menardian incoherency,’ if I may callit that, that does not depend on his holding anyversion of the adaptivity claim. At least we maythink that it shouldn’t fall out that Menard’sintention is incoherent only in case theintending agent is committed to the strongadaptivity claim.

This is not because the strong version of theadaptivity claim is false. It is false. If thenotion of ‘literary means’ has scope to include,for example, the specific forms of directauthorial address used by both Henry Fielding andJorge Amado, whose uses of this same techniqueare to very different effects, then the strongversion of the adaptivity claim has to be false.10

10 Of course, as we have noted, there can be weaker versionsof the claim. It might be true that some fairly specificliterary techniques are adaptable to only a restricted rangeof effects. I do not dispute that. There is room for a lotof subtlety here.

182


The reason we may want Menard’s intention tocome out incoherent even if Menard believed noversion of the adaptivity claim, true or false,is that otherwise, Menard’s incoherence iscontingent on the merely accidental fact theintending agent happens to believe somethingabout the adaptivity of literary means toliterary (and other) ends. And this may seemextraneous. But, if this is what we want, Ibelieve we are mistaken. Look again at thepassage in which Menard writes about how heconceives of the project:

Once that image [of a book not yet written] (which noone can legitimately deny me) is postulated, it iscertain that my problem is a good bit more difficultthan Cervantes’ was. My obliging predecessor did notrefuse the collaboration of chance: he composed hisimmortal work somewhat à la diable, carried along bythe inertias of language and invention. I have taken onthe mysterious duty of reconstructing literally hisspontaneous work. (Borges 1964: 41)

Here Menard points out three “obstacles” to hisproject. The first is that he is allowed to makevariations “of a formal or psychological type”.The second is that he is “oblige[d]..to sacrificethese variations to the ‘original’ text and toreason out this annihilation in an irrefutablemanner”. To these “artificial hindrances”, Menard

183

James Hamilton

writes, there is added a final obstacle, “of acongenital kind”:

To compose the Quixote at the beginning of theseventeenth century was a reasonable undertaking [...]at the beginning of the twentieth, it is almostimpossible. (ibid.)

To be sure, this is not the strong adaptivityposition we have been working with to arrive atour account of Menardian incoherence; and so ouraccount must be adjusted. But it is clear fromthe passage that Menard holds some version of theadaptivity claim. I think we can capture theversion he holds as follows:

Only one precise set of means is reasonably adaptive toany precise set of aims.

It is because Menard believes this (also false)version of the adaptivity claim that he conceiveshis entire project with the intention toundertake, as though it were reasonable,something which also he believes in fact to beunreasonable. And this, at least, is a kind ofincoherence.

That he has this incoherent intention doesnot, however, make the appearance of a text,produced in this way, impossible. For, as I havenoted above, I do not think we know enough aboutthe world to predict Quixote, or two chapters and

184


part of a third of it at any rate, could justnever appear in this way or something like it.Nor do I know what we would/should actually sayabout it if it did.

So, I conclude, we can extract aninteresting account of authorial intentions fromBorges’ story and, by so doing, we can show whatis quixotic about Menard’s intentions. Becausethe incoherency of his intentions turns onempirical facts about what Menard believes, orcould believe, we cannot show his intentions mustbe incoherent, as a matter of logical orconceptual necessity. This fact is part of whatmakes visible the possibility that both Borges’implied account of authorial intention and hisdescription of one way in which authorialintentions can misfire – both prominent featuresof the story – are obscured for us by using thestory to motivate a debate in the ontology ofart.

4. An objection: the apparent need to resolve theontology debate

A plausible objection to the line I have beentaking is that my argument succeeds only becauseI have been overly vague about what counts as

185

James Hamilton

‘materials’ and ‘aims’.11 The objection mightproceed by suggesting that the best way to makethese notions precise is to state identitycriteria for materials and then ask whether aimssupervene on materials or materialsunderdetermine any aims realizable.

A favored way of talking about literary worksof art would then lead us to decide whether by‘materials’ we mean inscription- or sentence-types, sentence meaning, or utterance meanings.

If the materials Menard used are identifiedas inscription- or sentence-types, then Menardand Cervantes use the same materials and,clearly, these materials underdetermine aims.Sentence-types can bear multiple meanings and notevery utterance of a sentence-type realizes thesame communicative aim.

This suggests we include sentence meaning inthe identity conditions of materials. Butsentence meaning is determined by the semantics

11 Peter Lamarque suggested this line of objection to me incomments he made on an earlier version of this paper at theAmerican Society for Aesthetics, Pacific Division, in March1999. I am grateful to him for spelling this out so clearlyand allowing me to read his comments in detail. Readerscognizant of recent discussions of actual-author intentionsversus hypothetical-author intentions in the context ofsettling questions of interpretive scope will find the termsof this objection familiar. See, for example, Carroll 1992;Carroll 2000; Iseminger 1996; Levinson 1992; Stecker 1993;Stecker 1997; Trivedi 2001.

186


of a language so we can, once again, havedifferent realizable aims. Here the objector cansupply familiar cases involving indexicals toshow that sentence meaning underdetermines theaims speakers may have.

Suppose we identify materials with utterancemeaning. Now aims do look like they supervene onmaterials, for any change of aim would imply achange of materials.

None of these seem to be what Menard has inmind. Menard’s agonizing and endless drafts arenot consistent with identifying his materialswith inscription- or sentence-types. Menard’sexplicitly stated intention to compose theQuixote is not consistent with identifying hismaterials as sentence meanings. Nor finally, willit do to identify his materials with utterancemeanings: this latter would require Menard tobecome Cervantes, which, you remember, Menarddismisses as “too easy”! The objector can now addthat Borges’ narrator points out that the finalproduct of Menard’s efforts is not identical tothe original; it is, as he puts it, “infinitelyricher”, “more ambiguous”, it is in a differentstyle, it makes different claims about history,and so forth.

So, it may be said, there is a way to showMenard’s intention is incoherent of necessity.This would then justify us in looking at hisachievement, whatever it is he produces, free

187

James Hamilton

from any constraints having to do with hisintentions in producing it. His apparentachievement, as described by the narrator, is tohave produced something seemingly distinct fromthe Quixote, with different properties, includinga different meaning. The achievement can justabout be described as a new work built on thematerial of an earlier writer.

It is precisely here that we need to invoke adistinction between work and text. For example,one familiar story about this would insist that,if we take the identity conditions of thematerial to be text-identity, including not onlynotational but also semantic identity, then wecan describe Menard as reproducing the sametextual material as Cervantes. If we take work-identity to be distinct from text-identity and toincorporate features like utterance meaning andcircumstances of origin, then Menard can be givencredit for producing his own work different fromCervantes. There is nothing more perplexing aboutthis achievement than the fact that two speakerscan use the same sentence-types to make distinctstatements, with different aims and subject todifferent interpretations.

To sum up then, the first argument we havebeen tracking in stating the objection looks likethis:

(1) Menard either aims at producing

188


(A) the same sentence- or inscription-typesas Cervantes,

(B) the same sentence meanings as Cervantesor

(C) the same utterance meanings as Cervantes.(2) The narrator should be taken at his word withrespect to what Menard does and says about whathe is doing.(3) Menard agonizes over multiple drafts.(4) Premise (3) is inconsistent with alternative(A) in (1).(5) Menard insists he wants to “compose theQuixote”.(6) Premise (5) is inconsistent with alternative(B) in (1). (7) Menard dismisses the idea of becomingCervantes as “too easy” for his purposes.(8) Premise (7) is inconsistent with alternative(C) in (1).(9) Insofar as premise (1) exhausts thepossibilities with respect to Menard’sintentions, the narrator’s description ofMenard’s intentions is incoherent.

Having shown Menard’s stated aims are incoherent,we are free to consider independent argumentsconcerning whether his achievement is coherent.The argument rehearsed above as an example can bepresented thus:

189

James Hamilton

(10) Text-identity is determined by not onlynotational but also semantic identity. Work-identity incorporates features like utterancemeaning and circumstances of origin.(11) The narrator should be taken at his wordwith respect to what Menard produced.(12) So, we can gloss what happens in the storythis way: Menard reproduces the same textualmaterial as Cervantes but Menard produces his ownwork different from Cervantes’.(13) The achievement description in (2) iscoherent. Indeed it is the same sort ofexplanation one might give of some of theachievements of Shakespeare who famously used andreworked material of his predecessors while stillproducing original works.

And our objector’s conclusion is the conjunctionof (9) and (13): Menard’s stated aims areincoherent but the narrators’ description ofMenard’s achievement is coherent.

5. A response

In assessing these arguments the first thing wemight want to notice is that (9) and (13),conjoined with (2) and (11), imply that Borges’narrator is confused. This is surely possible; afictional narrator might be confused about many

190


things. What we must claim here is that Borges’narrator fails to notice, at least in the case ofMenard, that we cannot draw the same connectionsbetween intentions and attributions ofachievement we ordinarily do. Borges’ narratordoes think that Menard’s intended products arethe result of his intended activities.

There is a problem with this. If theinterpretive claim is that the narrator isconfused in this way, we want more than Borgespresents in the story by way of support for theclaim. There should be some independent textualevidence of the narrator’s disposition toconfusions of this kind. Borges, as I have justsaid does not provide the relevant evidence. Ofcourse, another possibility is that Borgeshimself is confused. Short of either of thesehypotheses, however, we might want to see if morecould be done by way of analysis of the story. Sofar, then, let us hold (2) and (11) fixed.

To clear away one unnecessary source ofcomplexity, let me note that I am not at allconcerned with the argument for (13) exceptinsofar as the entire argument is a symptom ofthings gone seriously awry. True enough, theobjector’s conclusion, as I have presented it,depends on the strength of the inference from(10) through (12) to (13). But (10) and (12) aresubject to considerable dispute amongphilosophers working on the ontology of works of

191

James Hamilton

art and Menard’s ‘achievement’ has been describedvariously as plagiarism (Tilghman 1982; Bailey1990), copying (Bailey 1990; Sparshot 1982: 179;Wreen 1990), and the coincidental writing of thesame work (Janaway 1992; Currie 1989: 120-124). Itake no position on this matter.

What I do want to take up is the fact that,as the objection is worked out, we are entitledto undertake an argument concerning whetherMenard has or has not produced a new work ofliterary art only because we have concluded thathis intention is incoherent and therefore we mayask about his achievement without reference tohis intentions or ambitions. But I have insistedthat, in any robust notion of authorialintention, we must understand the products ofthose intentions as having come about non-accidentally as results of activities undertakenunder the direction of the author’s intentions.If I am right we should be able to detectsomething fundamentally wrong with the argumentfrom (1) through (8) to (9).

And indeed we can. Here, again, is thatargument:

(1) Menard either aims at producing(A) the same sentence- or inscription-

types as Cervantes, (B) the same sentence meanings as

Cervantes or

192


(C) the same utterance meanings asCervantes.(2) The narrator should be taken at his word withrespect to what Menard does and says about whathe is doing.(3) Menard agonizes over multiple drafts.(4) Premise (3) is inconsistent with alternative(A) in (1).(5) Menard insists he wants to “compose theQuixote.”(6) Premise (5) is inconsistent with alternative(B) in (1). (7) Menard dismisses the idea of becomingCervantes as “too easy” for his purposes.(8) Premise (7) is inconsistent with alternative(C) in (1).(9) Insofar as premise (1) exhausts thepossibilities with respect to Menard’sintentions, the narrator’s description ofMenard’s intentions is incoherent.

Note first that premise (2) is what underwritesour confidence in premises (3), (5), and (7).These latter are what are appealed to indetermining that none of the ways of specifyingMenard’s materials listed in premise (1) can besustained. And that , in turn, is the substanceof the consequent of the conclusion, (9).

This argument is, however, only as strong asour confidence in the antecedent of the

193

James Hamilton

conditional that forms the conclusion. It mightbe more perspicuous to note that this antecedentfunctions as an assumed premise in the argument.But now, if we think it implausible that Borges’narrator (or Borges himself) is confused, weshould conclude that premise (1) does not exhaustthe possibilities with respect to Menard’smaterials.

This is a conclusion that I think we shouldalso arrive at quite naturally fromconsiderations of what authors themselves thinkof as their “materials.” The thought behind theobjection is right, that this term captures manydifferent kinds of things. The term is vague. Butthe correct way to make it more precise is tooffer detailed examples.

Borges had, I believe, already done this inhis story. It may be sufficient – both to showhow we should make the concept of writerly“materials” more precise and to demonstrate whatmust get obscured if we focus on ontologicalquestions about works of art – to list the kindsof things Menard took to be his (and Cervantes’materials). Remember what Menard says he did toacquire an “equal [to] the imprecise and priorimage of a book not yet written”:

…I have reread closely certain chapters, those which Ishall not attempt for the time being. I have also gonethrough the interludes, the plays, the Galatea, the

194


exemplary novels, the undoubtedly laborioustribulations of Persiles and Segismunda and the Viajedel Parnaso ...12

Among the materials Menard consults then, theother works of Cervantes, are where he will findthe multiplicity of sentence structures, thetechniques for achieving depth of character, thesimplifications of action, and the commitment todetailed descriptions (achievable in prose butnot in poetry – at which Cervantes was apparentlynot very good) that characterize the writing ofCervantes. Then too, there is a rhythm and soundin Cervantes’ prose that would be the appropriateobject of Menard’s attention. These elements, andthings like them, are what we need to think aboutwhen we consider what are an author’s“materials.”

But these elements, and things like them,disappear from view when we hold that thematerials an author deploys are inscription- orsentence-types, sentence meanings, or utterancemeanings. Under the influence of a conception ofauthorial intention that is friendly to theontology debate, a more robust and adequate

12 I might also note that, even here, Borges is having us on.La Galatea and the two surviving plays antedate the Quixote.But the Exemplary Novels and the Viaje del Parnaso both comesome ten or more years after Cervantes wrote Part I of DonQuixote. And the Travails of Persiles and Seigismunda waswritten a year after Part II of the Quixote.

195

James Hamilton

conception of authorial intentions – where thearticulation of those intentions makes appeal tosuch concepts as ‘materials’, ‘aims’, and ‘fit’ –becomes unavailable to us.

Bibliography

Agassi, Joseph. 1970. ‘Philosophy as Literature: The Case ofBorges’ in Mind 79: 287-293.

Bailey, George. 1990. ‘Pierre Menard’s Don Quixote’ inIyyun, the Jerusalem Philosophical Quarterly 39: 349.

Borges, Jorge L. 1964. ‘Pierre Menard, Author of theQuixote’ in Labyrinths: Selected Stories and Other Writings. NewYork: New Direction Publishings: 36-44.

Carroll, Noel. 1992. ‘Art, Intention, Conversation’ inIseminger, Gary (ed.) Intention and Interpretation.Philadelphia: Temple University Press: 97-131.

Carroll, Noel. 2000. ‘Interpretation and Intention: TheDebate between Hypothetical and Actual Intentionalism’in Metaphilosophy 31(1-2): 75-95.

Currie, Gregory. 1989. An Ontology of Art. New York: St. Martin’sPress.

Danto, Arthur C. 1973. ‘Artworks and Real Things’ in Theoria39(1): 1-17.

Danto, Arthur C. 1981. The Transfiguration of the Commonplace.Cambridge (Mass.): Harvard University Press.

Diamond, Cora. 1988. ‘Losing Your Concepts’ in Ethics 98(2):255-277.

Elgin, Catherine and Nelson Goodman. 1988. Reconceptions inPhilosophy and Other Arts and Sciences. London: Routledge.

Goodman, Nelson. 1968. Languages of Art. Minneapolis: Bobbs-Merrill Company.

196


Iseminger, Gary. 1996. ‘Actual Intentionalism vs.Hypothetical Intentionalism’ in Journal of Aesthetics and ArtCriticism 54(4): 319-326.

Janaway, Christopher. 1992. ‘Borges and Danto: a reply toMichael Wreen’ in British Journal of Aesthetics 32(1): 72-76.

Levinson, Jerrold. 1992. ‘Intention and Interpretation’ inIseminger, Gary (ed.) Intention and Interpretation.Philadelphia: Temple University Press: 221-256.

Lewis, David. 1978. ‘Truth in Fiction’ in American PhilosophicalQuarterly 15(1): 37-46.

Mele, Alfred. 1992. ‘Recent Work on Intentional Action’ inAmerican Philosophical Quarterly 29(3): 199-217.

Savile, Anthony. 1971. ‘Nelson Goodman’s Languages of Art: AStudy’ in British Journal of Aesthetics 11(1): 3-27.

Schier, Flint. 1986. Deeper into Pictures. Cambridge: CambridgeUniversity Press.

Sparshott, Francis. 1982. The Theory of the Arts. Princeton:Princeton University Press.

Stecker, Robert. 1993. ‘The Role of Intention and Conventionin Interpreting Artworks’ in Southern Journal of Philosophy31(4): 471-489.

Stecker, Robert. 1997. ‘Hypothetical Intentions and ImpliedAuthors’ in Stecker, Robert Artworks. Philadelphia:Pennsylvania State University Press: 186-212.

Tilghman, B.R. 1982. ‘Danto and the Ontology of Literature’in Journal of Aesthetics and Art Criticism 40(3): 293-299.

Trivedi, Saam. 2001. ‘An Epistemic Dilemma for ActualIntentionalism’ in The British Journal of Aesthetics 41(2): 192-206.

Wollheim, Richard. 1984. ‘On the question “Why painting isan art?”’ in Aesthetics: Proceedings of the 8th InternationalWittgenstein Symposium, Part I. Vienna: Hölder-Pichler-Tempsky: 101-106.

Wreen, Michael. 1990. ‘Once is Not Enough?’ in British Journal ofAesthetics 30(1): 149-158

197

James Hamilton

Zemach, Eddy. 1967. ‘A Stitch in Time’ in The Journal of ValueInquiry 1: 223-241.

198

Borges's Menard 2013 Text&Work

Documents