DEFENDING ‘THE ARTIST’S THEORY’: WOLLHEIM’S LOST IDEA REGAINED?

The paper considers an argument of Richard Wollheim’s, originally presented in a 1976 symposium with Goodman and Wiggins, which disappeared when the symposium contribution was ‘reprinted’ in the supplementary essays to the expanded edition of Art and Its Objects (Wollheim, 1980). It lays out the argument’s original context, locating its objectives by means of a comparison with Goodman’s autographic/allographic distinction, with its attendant discussion of the ‘history of production’, and presents Wollheim’s defence of ‘the artist’s theory’. This defence coheres in interesting ways with Wollheim’s aesthetics emphasis on the importance of the artist’s intention (suitably understood) as part of a specification of what the work itself is. This conception reinforces the importance Wollheim grants both to the fulfilled intentions of the artist and to a suitably positioned, suitably informed, and suitably sensitive spectator. Both should be modelled as operating under the aegis of the artist’s theory, a notion this ‘missing’ argument serves to emphasize.

however, to find that a section had not made the transition into Version B, especially since that section contained an interesting passage of argument.
Both Version A and Version B appeal to our intuitions concerning the nature of authorship in highlighting what Wollheim urges is an implausibility in Goodman's position. But, in Version A, Wollheim also offers a brief argument to sustain or bolster these intuitions, although it is directed at Goodman's picture of the identity-conditions for artworks in what Goodman calls 'allographic arts' (see below) and it is only fully intelligible against that background.
Version B presents simplified forms of some arguments from Wollheim's original, Version A, paper. The argument that is missing in Version B is, nevertheless, considered important, if not -for certain readers -pivotal in Version A: Some may find the preceding argument convincing. But some who do not, may yet find the principle of the argument convincing. It is to those who find the principle of the argument convincing […] that the rest of this paper is directed. 8 And 'the rest of the paper' is more than half.
One question here, then, might be why Wollheim did not include this argument, when re-drafting this material into two essays. 9 The context of the Bristol symposium was a relevant factor here just because it more or less required Wollheim to discuss Goodman's work. 10 Further, the decision to include Wiggins (and Wiggins's pet topic, identity) in the symposium more or less set the remainder of the agenda -that Wollheim would discuss some of Goodman's identityrelated claims.
Still, any resolution here, explaining 'why' Wollheim behaved as he did, in the absence of evidence, must be little more than idle speculation. Instead, I shall urge that the missing section contains a fundamental insight of Wollheim's, one which, had it been more widespread, might have clarified a particular, puzzling example, and really before it took hold of the collective imagination of philosophical aesthetics in the Anglo-American analytic tradition.
Moreover, considering its place, or lack of it, points to further insights from Wollheim.  Wollheim,'Criteria' ,37. 19 Perhaps a more revealing question would enquire why he included it the first time through, and would lay the blame on Goodman's presence at the Bristol Conference -for the topic might seem both more pertinent and more interesting to Goodman than to the rest of us. 10 And to go beyond what he had said in reviewing Languages of Art initially: the review is reprinted in Richard Wollheim, On Art and the Mind (Harmondsworth: Allen Lane, 1973), 261-89.

II. THE GOODMAN-WOLLHEIM DEBATE
To locate the lost argument in its context, and hence to understand it more fully, we should return to the dispute between Wollheim and Goodman. At the centre of this debate is Goodman's categorization of some arts as autographic and others as allographic, and the identity-related implications of this categorization. The idea was that, for works in autographic arts (like painting and cut-sculpture), identity-questions were in effect centrally questions about numerical identity for material objects; and were answered in terms of what Goodman called the 'history of production' 11 -such that to identify a work as 'such-and-such a Michelangelo'is to draw a direct connection to 'the artist's hand'.
A painter, to produce a work, must end up with a painting (must 'finish the painting' , 12 Goodman says), but for music, 'the composer's work is done when he has written the score'. 13 And writing the score is not actually instantiating the work, for what one ends up with is not yet the artwork itself, but something closer to a recipe for the work. 14 One must still play the music in order to confront the artwork, yet that work is nevertheless made in making the score. So, for works in allographic arts (like novels, poems, dances, and musical works), identity-questions did not turn on the 'history of production' , since (as the music case shows) we can agree that there is such-and-such a piece of music (the score has been written) even though no one has heard it -it has not yet been played. Thus identifying the artwork in the allographic arts (like music) does not require that one be confronting the artwork at that time. Hence there need be no direct connection to 'the artist's hand'. So perhaps there is no need to refer to history of production to identify the work itself -the artist need have no direct connection to that performance or that inscription. In fact, for Goodman, the point is definitional: in order that an art form be allographic, '[w]hat is necessary is that identification of the […] instance of a work be independent of the history of production'. 15 Such works could be understood as a series of words, or movements, or sounds, in a particular order. The work is identified in terms of (say) the words used: so one needs 'exact correspondence as sequences of letters, spaces and punctuation marks'.  Goodman, Languages, 122;Nelson Goodman, Problems and Projects (Indianapolis: Hackett, 1972), 101. Goodman has reprinted sections of his works: thus Problems and Projects contains relevant sections of Languages of Art; and Nelson Goodman, 'Comments on Wollheim's paper' , Ratio (1978), 49-51, reprinted in idem, Of Mind andOther Matters (Cambridge, Mass.: Harvard University Press, 1984 But now reproducing, say, those words seems simply to repeat the poem or novel; in the language that Wollheim had deployed, it produces another token of the same type. 16 There need be no direct connection to the poem's author, but each is (equally) the artwork in question. Hence, were this correct, works of literature would be impossible to forge, since '[a] forgery of a work of art is an object falsely purporting to have the history of production requisite for the (or an) original of the work'. 17 And, allographic works have no such connection.
Further, identification of works of literature would not depend, in this way, on the 'history of production' or on 'the artist's hand'. 18 Goodman had used the connection to the possibility of forgery to introduce these coinings ('autographic' , 'allographic'). For, the argument runs, if I were to produce something indistinguishable from an artwork in an autographic art (such as painting), I shall have forged that painting; or, at best, copied it. In either case, my 'work' is not that artwork. But if the object I produce is indistinguishable from an artwork in an allographic art such as poetry or music, I have simply made another instance of that artwork, for example, another inscription of that poem, or novel, or piece of music. Hence allographic art is, for Goodman, unforgeable.
In both versions (but especially Version A), Wollheim argues that Goodman is wrong on the place of the 'history of production' for work-identity. 19 The artist's hand has a clear role in obviously autographic arts, such as painting. 20 Thus Wollheim remarks that '[i]t is common ground to Goodman and myself that, when a work of art is an individual, identity depends on history of production' , 21 '[t]he history of production of Titian's Bacchus and Ariadne is relevant to its identity'. 22 Yet these points, connecting responsibility to identity, apply as clearly to, say, literary works as to paintings. So Wollheim insists on the parallel conclusion for the relation of 'the history of its production to the identity of Earlier (Art and Its Objects, 4-5, § 5), Wollheim had drawn a distinction between two kinds of artworks: some were particular objects (like paintings), some were multiples (like novels, poems, and musical works). And he characterized the multiple kinds by means of a type/token contrast explicitly drawn from Peirce. See ibid., 74-84, § 35-37: this is Wollheim's initial discussion of 'criteria of identity and individuation appropriate to, say, a piece of music or a novel' (ibid., 74).
Hamlet: if Shakespeare did indeed write Hamlet in 1600, then necessarily Hamlet was written by Shakespeare and in 1600'. 23 In this way, identifying the object as (say) Donne's 'Valediction: Forbidding Mourning' is recognizing that it is a poem, in seventeenth-century English, and in the 'metaphysical' genre. These features supply some categorial constraints for its appreciation. But to recognize that, say, it is not ironic may lead the reader back to Donne's authorship. Certainly this is a feature typically invoked in identifying a poem -hence in distinguishing this poem from that one. And, as Goodman grants, 'determination of authorship  Goodman,'Comments' ,50,Of Mind,140. Of course, for Goodman, that is irrelevant for our thinking about the other kind of artworks. 25 Goodman, 'Comments' , 50, Of Mind, 140. 26 Goodman, Problems, 141. This concession is not in the 'Comments' paper. But what is Goodman's considered response? Where is it? In Problems, 141, Goodman claims to be working on it: 'I am currently (1983) exploring some of these matters for a paper on interpretation and identity'. If the outcome is Nelson Goodman and Catherine Z. Elgin, Reconceptions in Philosophy (London: Routledge, 1988), it simply amounts to denying the case. one from Menard, commenting on the different import of the two passages. He continues: Equally vivid is the contrast in styles. The archaic style of Menard -in the last analysis a foreigner -suffers from a certain affectation. Not so that of his precursor, who handles easily the Spanish of his time. 29 But what has changed here? Not the words, of course. Rather, the meaning of the Menard differs from that of the Cervantes: we can accurately say things of the Menard which would not be true of the Cervantes -different reasons for our judgements become open to us.
This case raises exactly the problem that Goodman here presents as of dubious relevance. For, as Borges describes the case, the work of literature -that is, a work we expect to be in an allographic art -would not be identifiable solely on the basis of 'those words in that order'. Each is in the same position in that respect. On the contrary, seeing this as a work by Menard involves the 'history of production' , and therefore so does recognizing it as one by Cervantes.
Goodman's own reaction here is revealing -he simply denies that the case as Borges presents it makes sense: To deny that I read Don Quixote if my copy, although correctly spelled in all details, happens to be accidentally produced by a mad printer in 1500, or by a mad computer in 1976, seems to me utterly untenable. 30 But this passage assumes the answer to the most important question -namely, that what he held was indeed (a copy of ) Don Quixote. If we were sure that this text's origin was a mad printer or a mad computer, we would be happy -I imagine -to deny that this 'text' was of a different work, because we would doubt (or deny) that it was a work at all. My computer has just thrown out a text orthographically indistinguishable from Cervantes's masterpiece: once I know for sure that this is just an accidental result of some glitch, the resultant object is -in a clear sense -naturally occurring (it is simply the result of the working-out of causal forces, devoid of intention). I am certainly surprised that my computer's glitch has resulted in a text orthographically indistinguishable from the Cervantes.
But has my computer produced another copy of Cervantes's masterpiece (as Goodman seems to assume)? Well, the cracks in my wall could not produce a particular sonnet, no matter how much the cracks looked like that poem. For the poem is a meaning-bearing object; and that is not within the scope of the cracks in the wall, which are the mere working out of causal forces (in the case Estetika: The Central European Journal of Aesthetics, XLVII/III, 3-26 under consideration). Only the size of the coincidence would restrain me from saying the same thing about the computer-driven Cervantes. But that is philosophically irrelevant. So one could well deny that what the computer produced is a copy of Cervantes, a token of that type. And the same thing could be said if the 'text' were instead produced by the madness of (say) my mad-printer ancestor in 1500: he did not intend to make a novel; and certainly not a novel in sixteenth-century Spanish (which he did not speak). For, as I imagine this, it too was merely the working out of causal forces (devoid of intention), although now forces in my ancestor's brain: it is also a (kind of ) naturally-occurring event -at least, no one is responsible for it. So these cases could be 'recognized heads of exception' to the ascription of any meaning here.
Yet blanket rejection of this sort seems precisely Goodman's position here.
Indeed, Goodman and Elgin contend 'that the two supposed works are actually one.
[…] What Menard wrote is simply another inscription of the text'. 31 But how can they argue for this straight-out rejection? If their position is (a) that the allographic/autographic contrast is fundamental -which Goodman had not previously asserted -and (b) that it is explained through a differential relation to history of production, it follows for Goodman that the 'works' of Menard and of Cervantes, being 'works' in an allographic art form, must in fact be different transcriptions of the same work. But that simply reiterates the original distinction in the face of the (apparent) counter-case: these seem like different artworks, and for roughly the same reasons as those in Danto's gallery (quoted earlier) -namely, that we can (truly) offer different artistic judgements of each.
Further, as Wollheim points out for this case, such objects would seem to lack a history of production, in the relevant sense, 32 since (for Goodman) they cannot appeal to that history to resolve either interpretative-or identity-questions, in exactly the places we would require them (in practice).

III. WOLLHEIM'S FIRST RESPONSE
Yet Wollheim does not simply reject this Goodmanian line of argument, but instead, mobilizes our intuitions. Moreover, he explicitly chooses his own example, rather than Borges' , 'so as to attract greater plausibility to' 33 the supposition (as he puts it in Version A). So Wollheim posits a reworking of the idea behind this story by Borges's 34 in a short lyric poem written by an Elizabethan poet, and an orthographically indistinguishable one composed by a Georgian poet. 35 We grant, of course, that the second poet operates as poets always do; in particular, that he operates 'in complete ignorance of his predecessor' 36 -or at least in relative ignorance of that predecessor, so that we are content to call him an author in this respect. So there is no question that the second merely transcribes the work of the other. Then Borges's preferred solution -two works with different properties, despite being orthographically indistinguishable -seems most appealing: each wrote his poem by his methods.
But even those who maintain that there is only one poem in this case still need to pay attention to history of production even to conclude 'that the Tudor poet wrote it […] and the Georgian poet merely wrote it out'. 37 For that reference is required to determine the priority. If that line too were rejected, it is unclear what realistic options would remain open to us. Clearly, someone is the author of this poem. So: we could stick at the fact that two poets wrote down the same lines in the same spelling and simply refuse to countenance the further question, Whose poem is it? 38 Yet that is not really an option. 39 If we cannot ask whose poem it is, neither writer can be responsible for it. And that seems absurd. As Wollheim recognizes,40 were such a position correct, 'poems would not have a history of production: they would be more found than made'. 41 (But this conclusion might attract someone who thought poems timeless abstract objects: see § 5.) Here, then, both Versions A and B deploy a powerful example -concerning these 'confusable counterpart' 42 poems of the Tudor and Georgian poets -in which the 'history of production' seems important to our recognizing which poem we are confronting; and hence how it should be confronted.

IV. WOLLHEIM'S 'LOST' ARGUMENT
But Wollheim wishes also to explain the force of our intuitions here. Moreover, he explicitly aims 'to make this point in the context of what is for Goodman an allographic art' 43 -since, after all, they broadly agree for autographic arts (at least, for those autographic arts in which the artwork is a particular object, such Ibid. 39 Wollheim says this position could only be adopted 'implausibly' (Ibid.,141). 40 Ibid., 170. 41 Ibid., 171. 42 Danto, Transfiguration, 1. 43 as paintings and cut-sculptures). Thus -in the passage whose disappearance I have been plotting -Wollheim turns to the novelist, highlighting two strands to such an artist's thinking: 'the thought that what he made is his and the thought that what he made is a novel'. 44 For an artist in a relevant art, such as a novelist makes it, and is therefore responsible for it, in two senses of the word: first, he is responsible because he brought it into being -it is his creation; and, second, he is responsible because any praise or blame attaching to it is his. But what has he made? Here we see the power of Wollheim's two-part formulation. For (a) 'what he has made is his' , and also (b) 'what he has made is a novel': that is what he is responsible for. Yet what exactly is that? For a novel in this sense cannot just be a collection of words, since its author could not with justice claim responsibility for the words or the sentences. As Wollheim notes, nothing less than the whole is credible as his, as nothing less could be assigned to him: 'evidently no novelist thinks that a word is his, that a phrase is his, that a paragraph is his '. 45 At least, no sensible novelist should think these things. 'All these, he must surely recognise, belong to the language.' 46 Hence his claim to responsibility must be for the novel as a whole.
In sum, Wollheim is arguing here that (i) since the novelist, as author, is responsible for his novel, and (ii) since his responsibility can only be for the novel (and not for any smaller unit, such as a paragraph), then the novel cannot be identified with anything smaller -in particular, not just with words (as Goodman hoped). In this sense, as Wollheim puts it, the concept of the novel cannot be a 'macro-grammatical' concept. 47 But that is precisely to deny what Goodman had urged. This, then, is the 'lost' argument, presented semi-formally. As Wollheim correctly describes it, its conclusion involves disidentifying the particular novel that he has written down from any other particular novel that has a different history of production, even if by some grotesque chance the two exhibited sameness of spelling. 48 For that would not be his novel. So, identifying his work simply as a collection of words misses the sense of it as his novel.  51 At the least, this reinforces our locating the work in its precise place in the narrative of art history in this art form (here, the novel). For it shows the connection of the novelist's specific work to the larger narrative of the history of the novel. Then how novelist X regarded his novel is partly a reflection of the state of the novel at the time of its composition, partly a reflection of the author's view of the state of the novel at that time, and partly a reflection of his conception of the place of this novel beside other novels. One upshot of the 'lost' argument is to stress that artistshere typified by our novelist -bring to their creative activity a concern with the state of the art form and a view (perhaps implicitly) about it.
Further, if the artist's theory is 'to control the artist's output [in this way]' , 52 such that it is revealing (in a later example) 'that Verdi composed his operas under the concept "opera"' , 53 and if that theory is also to 'allow us to explain or (better) understand that output, then it must surely specify the criteria of identity for a work of art within that art in which the artist works'. 54 For that is saying what is involved in creating an opera, rather than some other kind of artwork; and also that in this 'criterial specification' (as Wollheim called it) 'reference is made to the history of production'. 55 And, of course, once this is granted, the outcome of the 'Tudor versus Georgian poem' case as two different poems (see above) -which otherwise depended just on appealing to our intuitions -follows automatically.  53 Wollheim, Art and Its Objects, 171. 54 Wollheim, 'Criteria' , 36-7. 55 Ibid., 37. 56 Wollheim, Art and Its Objects,171. For that downplays the pervasiveness of the artist's theory: rather, Wollheim insists: 'The truth is that Verdi composed his operas under the concept "opera"'. 57 And, in saying this, he is stressing the importance of the concept opera: that 'the concept was regulative for what he wrote'. 58 And this regulative role, always there, might be missed.
In stressing this point explicitly in Version B, was Wollheim conceding that it was missing -or had been missed by readers -in Version A? Clearly, the point is implicit in much that he does offer in Version A: but it is not explicit either in the thought that 'the artist's theory' should 'control the artist's output' 59 when applied from the perspective of the artist or 'allow us to explain or (better) understand that output' 60 when applied from the audience's perspective. And, in Version A, these comments immediately precede the 'lost' argument, where Wollheim stresses that, for its author, his novel 'can be understood and appreciated only in the light of his having made it' 61 that is, in terms of his authorship. And only something like the regulative conception can bear that weight. For only then does his conception of the work as a novel inflect directly how it is to be understood. But that point is not very explicit. And that may be a weakness in this formulation of the 'lost' argument.
The power of these considerations (if they are to be granted) must be stressed.
If they are correct, then for all artworks (and especially for the identification of such works), the 'history of production' is relevant. And hence it is relevant even for those multiple works that Wollheim treats using the type/token contrast. Now Wollheim's introduction of the type/token contrast is fairly cursory. But, no matter how we work out that type/token contrast, such that the tokens instantiate artworks, these considerations should guarantee that there must be a 'history of production' for the types. Hence it cannot be that types are not made.

VI. CONNECTION TO WOLLHEIM'S VIEW OF THE MIND
Wollheim did not follow through in his discussion of poems and, especially, novels from the 'lost' argument. But we can readily imagine that some of the points he makes in other cases (and especially in respect of pictures) would apply there mutatis mutandis. Of course, his primary concern in both Version A and Version 'eternal' -hence these artworks (musical works) are at best discovered, not made. Such an emphasis is incompatible with a commitment to the relevance of the 'history of production': if the 'history of production' is relevant to identity here, identity cannot be understood in some timeless or eternal way. At the very least, it would be absurd to regard such a position as the default -see Dodd, Works of Music, 8: 'the account is prima facie correct and must be accepted as long as it is not defeated'; see also the perceptive review by Ben Caplan, British Journal of Aesthetics 47 (2007): 445-6. 64 Wollheim, 'Criteria' , 48. 65 Ibid. For if we take any particular painted surface, it is always possible that suitably sensitive, although insufficiently informed, spectators will be able to see different things in it. 68 But such disagreements must be resolvable (at least in principle) if the ascription of meaning to the painting is to make sense: that is, as a bulwark against subjectivism: at this stage […] the intention of the artist has a role to play. For, of those things which the spectator can see in the painting, it is correct for him to see only […] those things which the artist intended. 69 And, after all, the artist is (sometimes) rightly seen as a spectator of his own work. 70 So, as Wollheim summarizes a key thesis from Painting as an Art: 'the meaning of a painting combines what a sensitive spectator can see in it and what the artist intended it to convey.' 71 And, of course, the same could be said mutatis mutandis for the meaning-bearing aspect of other art forms. Moreover, both elementsartist's intention and suitable spectator -involve, in different ways, 'the artist's theory'. 72 Concerning intention, it is explicit, in that a full characterization of the artist's intention will make reference to 'the artist's theory' in locating his work (as he sees it) in the narrative of art history (as he sees it); for the spectator, it involves being suitably informed, for example, about the narrative of art history. Yet Wollheim has an additional thought here, that these must be actual intentions of the artist. 73 So, for Wollheim, the creative process (surely an aspect of artistic intention in this context) is 'not stopping short of, but terminating on,  Richard Wollheim, Formalism and Its Kinds (Barcelona: Fundació Antoni Tàpies, 1995), 34. 69 Ibid., 34. See also ibid., 35: 'This point about correct and incorrect perception of a painting and the role of the artist's intention in deciding between them is most readily, most effectively, made in the case of figurative painting. But it applies no less cogently to abstract painting.' 70 See Wollheim, Painting as an Art, 43: 'An artist must fill the role of agent […] but he must also fill the role of spectator' -more bluntly, 'The artist is essentially a spectator of his work' (ibid., 39). 71 Wollheim, Formalism, 36. He is 'recapitulating, but in a highly compressed fashion, and therefore without benefit of argument, what I have set out at greater length, and, I hope, more persuasively, in Painting as an Art '; ibid., 34. 72 Thus, Painting as an Art, 36: 'if we are to understand when and why painting is an art, we must consider it from the perspective of the artist.' This means, in context, from the perspective of the artist's theory: for that connects the artist's view of his work with his view of the art form into which it is a contribution (and his conception both of the historical narrative and the current state of that art form). 73 Various writers, most notably Levinson ( 'Wollheim on Pictorial Representation' , in idem, Contemplating Art, Oxford: Clarendon Press, 2001, 250), have identified Wollheim as an actual intentionalist (in the current terminology). But the evidence is scanty: many of the seemingly telling expressions -'fulfilled intention' (Formalism,36) For it has always been a constraint on Wollheim's aesthetics -and a sound one -that, as far as possible, the discussion of intention inside philosophical aesthetics should parallel that outside it.  Wollheim,Art and Its Objects,185. 75 See Stanley Cavell, Must We Mean What We Say? (New York: Scribners, 1969), 181: 'it no more counts towards the success or failure of a work of art that the artist intended something other than is there, than it counts, when the referee is counting over a boxer, that the boxer had intended to duck.' 76 Thus, in Painting as an Art, 19, Wollheim draws 'a distinction within intentions between those that are realised or fulfilled in the work and those which, although they contribute to the making of the work, are not realised in it'. That is, between actualized and failed intentions: the first are primarily embodied in the artworks. So one's intentions will usually be addressed explicitly only for the second case. 77 Ibid., 18. 78 Ibid., 9. 79 Ibid.,37. 80 Note that this is (presented as) a 'starting point'.
the starting point for debate should resemble that deployed in the philosophy of science, when one judges accounts of 'the scientific method' partly by how well they describe the practice of acknowledged scientists such as Newton, Darwin, or Einstein. Just so, the default position for philosophical aesthetics must be the 'canon' of art from the art world. For this is the base from which we approach what is or is not art. For instance, a work which the art world finds (or seems to find) relatively uncontentious -such as Tracy Emin's Unmade Bed (1999) -should (mostly) appear in our deliberations as an example of art; our starting point should not be that the general public finds it contentious. Yet beginning from the 'canon' of the art world, and its placing of certain works, will inevitably import some narrative of art history relative to these works -their precursors, their genre, and such like.
Again, we may need to review such claimed examples: that might revise our view of them, or of the narrative of art history thereby offered. Or again, we might motivate a discussion here by considering how to develop in argument a case for the art-status of such a work. In any case, much of the discussion would be comparative: by noting how our discussions of this work mirror our discussions of other, uncontentious artworks. But these considerations also apply (if implicitly) to the artist's perspective on the work. Further, it is usually more revealing to look at real justifications offered for works the general public can find problematic -for instance, Carl Andre's Equivalent VIII (1966), the firebricks in Tate Modern (London). Andre suggests that he does for texture (as embodied, in this case, in the firebricks) what Turner did for colour. 81 Making out this case, the comparative aspect is crucial, as is the connection to the past of the art form. For our treatment of them explains our treatment of the new case -if only as rebellion against the past! Thus, as Andre suggests, perhaps 'the Venus de Milo would just be a stone woman if nobody knew about sculpture'. 82 Yet, in each case, the argument deploys features from the narrative of the relevant parts of art history; and does so to show the conceptual connection of this work (however unusual) to other artworks. And recognizing 'the artist's theory' is just granting that this connection is mobilized by the artist in determining what to do to create his work and when it is finished.

VII. 'THE ARTIST'S THEORY' RECONSIDERED
Can we offer a more plausible picture of 'the artist's theory' than that presented previously? What needs modification? There seem to be two large questions: the first derives from the intentionalism which seems both suggested by talk of an 'artist's theory' and endorsed by Wollheim. The second is more specific.

Goodman offers standard objections: he is (rightly) 'as suspicious of talk about an artist's theory as […] [he is] of talk about a speaker's grammatical theory' 83
or, he might have added, the speaker's 'theory of meaning' for his language, in the sense from Dummett and Davidson. 84 He continues: Some artists and some speakers of a language do have such overt theories; and in a few cases an artist's theory may have a recognisable and noninhibiting effect on what he produces. 85 But, for Goodman few artists and few speakers have such theories explicitly, and what it could mean to say they have the theories implicitly puzzles me as much as what it would mean to say that the planets have the laws of motion implicitly. 86 Goodman also finds problematic the idea that theory plays in the nonverbal arts any such role as Wollheim maintains; for a theory, as I understand it, consists of statements of words. 87 Goodman's first worry, while legitimate, arises from his placing too much weight on the term 'theory' here; and perhaps that term was not well chosen. Certainly a useable conception of 'the artist's theory' must make it accessible to spectators Defending 'the Artist's Theory': Wollheim's Lost Idea Regained? 90 Richard Wollheim, 'Philosophy and the Arts' , in Modern British Philosophy, ed. Bryan Magee (London: Secker & Warburg, 1971), 186. See also Wollheim,Painting as an Art,18: 'the word ['intention'] is used by philosophers of art either far too narrowly or far more broadly than seems reasonable elsewhere.' 91 As Wollheim says, applied to painting: 'No total preconception of the picture that is independent of all engagement with the medium is a serious possibility' (Wollheim,Painting as an Art,19). 92 Wollheim, 'Philosophy and the Arts' , 186. 'Both these two doctrines [intentionalism and anti-intentionalism] share a common assumption: that there exist certain inner states of a certain kind -states which occur frequently in the process of makingand which can be understood independently of the product in which they issue.' Wollheim clearly rejects this assumption. 93 Wollheim, Art and Its Objects, 171. 94 Wollheim, 'On Pictorial Representation ' , 13. 95 Here, of course, I mean that we cannot guarantee that this 'prompting' will not be necessary. 96 And, as Wollheim notes: 'The spectator's experience must concur with the artist's intention, but it does not have to do so through knowledge of it'; Painting as an Art, 96. 97 Wollheim,'On Pictorial Representation' ,27. between intending to do such-and-such and (succeeding in) doing it and by its connection to being responsible for what one does, since these are features shared with the art case. But how should it be made out? To generate a less 'realistic' intentionalism, I suggest a so-called hypothetical intentionalist treatment, taken defeasibly. This most concessive version achieves all we need: on it (for a painting), understanding that work, or grasping the meaning embodied in it, is conceived in terms of what would be most justifiably ascribed to the artist: on the basis of the perceptible features of the painting, a complete grasp of its context of production, and a full knowledge of the artist's intentions as to how the work was to be taken, approached, or viewed. 98 And we can imagine parallel accounts drawn up for other art forms mutatis mutandis. My version will be robust in recognizing from the beginning that there are no finite totalities of such properties for the author. Hence talk of 'complete' or 'full' here seems misplaced: we can only make sense of such-and-such as incomplete when we also say what is not included. But that need not worry us in practice. For our project is a matter of picking up the best 'reading' we can.
Of course, one of the 'recognized heads of exception' will reflect that the work did not realize the author's plan: here, we might modify any ascription of the author's (explicit) plan as to how the work should be both recognized and appreciated -the work is not as the author planned it. And our concern is with the work: to be 'a novel-reader, not a mind-reader'. 99 That is to say, our focus is on what the artwork embodies. But that work is still intentionally understood -the author's intelligence is still behind it, even if we modify the precise place in the narrative to which he ascribed it. Moreover, this provides us with the intellectual resources to deal with, say, irony or allusion: in such cases, no direct reference to the author's (actual) life may be needed -although that move is not disqualified.
Our position here is profitably contrasted with a moderate actual intentionalist such as Noël Carroll, 100 who will agree with us that intention-ascription does not require planning 'in the head' of the artist. But such a moderate actual intentionalist insists, with some justice, that one must not put words into the artist's mouth, and hence must not ascribe to him or her thoughts not his or hers. Yet, since we are not discussing what such-and-such an artist did think or consider, our account is not necessarily refuted by evidence that the person did not think about such-and-such for himself or herself. There are two key cases here: Here, Johnson is offering a stronger 'reading' of the passage. And what is resolved is the meaning of that passage. Since this clearly might occur and, in at least some cases, Johnson's position will be the right one, I do not see how an actual intentionalism based only on what the artist would say (if asked) is defensible.
This case illustrates that, rather than being about (for example) what the makers or the audience thought, reference to 'the artist's intention' here, as is typical, amounts to commentary on the artworks themselves, hence, is true or false of them. This accords with a general commitment to artistic properties as properties of the artworks.
The other case asks whether, faced with (say) diary evidence of what the artist claims to have intended in the artwork, a hypothetical intentionalist might still ascribe conflicting intentions to the work, and with justice. Here, our moderate actual intentionalist will typically insist on following that diary evidence. In this case, I recognize that the diary evidence imagined here is relevant; but surely it is not determinative. For that diary evidence might resemble Goldsmith's first thought in the case above. So one still considers that (putative) evidence in the light of other evidence. Doing so might, in a particular case, lead one to set the diary evidence aside. (It will be important to resolve such examples case by case, rather than looking for a once-and-for-all resolution, to cover all cases.) Our conclusion, then, is towards a more flexible and evidence-responsive intentionalism than some, which recognizes much of that evidence as public.
This should offer a defensible intentionalism.
Let us now explicitly confront Goodman's objections. Then, if we are hypothetical intentionalists of the kind I have been sketching -which, for me, is just intentionalists, since the same pattern of 'filling-in' takes place in respect of other intention-ascription -could we not take a similar view of the artist's theory?
Could we not look to the view of that art form ascribable to our artist on the basis of our best reconstruction, taking into account what he did, as well as the concepts from his period, genre, and so on? And without assuming that the same strategy will be best in all cases? This seems a plausible strategy.
Doing this might (for instance) involve the following changes, all of which would produce a picture of 'the artist's theory' more plausible than previously, Defending 'the Artist's Theory': Wollheim's Lost Idea Regained? and hence bolster support for the principle behind Wollheim's argument -thus responding to Goodman's first worry. So perhaps our abstract model should (a) shed the term 'theory' here, since it seems (at least now) to import more baggage than that of which it rids us; (b) use instead a term like 'conception' (or to look for another), since a key element is how, say, Verdi views (or conceptualizes) the art form of opera -for this should feed into what he then does in trying to write opera. And it does seem that such a conception might both be extracted (for our purposes) from what Verdi did and hence be taken as operative in his thinking; (c) treat the whole in terms of insights from Wittgenstein, 102 insights which Wollheim was usually careful to accommodate and which stress the public character of what is fundamental in philosophy, rejecting the idea of a deep logical 'skeleton' that grounds our understanding and that philosophical analysis might uncover. Instead: -'nothing is hidden' 103 (also 'what is hidden, for example, is of no interest to us'): 104 the kinds of 'conception' identified above would be available to us as spectators, at least defeasibly; -'Look and see' 105 -which is the right advice only if what is needed is before our eyes (and certainly not requiring analysis to be uncovered to us).
Wollheim might well agree with this, given his tendencies to be appreciative of, and careful not to assert anything contrary to, positions in Wittgenstein's later philosophy as he understood them; and especially if we grant that what is generated from 'the hypotheses of psychoanalysis' 106 counts for these purposes as amongst what is in the public domain. That will be one way to ensure that, as Wollheim put it, our account leaves the meaning of the painting where we want it to be: neither in the head of the artist nor in the head of the spectator, but firmly on the surface of the painting. 107 And, mutatis mutandis, for the other arts.

IX. CONCLUSION
By re-invigorating Wollheim's 'lost' argument, this paper has stressed the importance of authorship for our understanding of artworks as meaning-bearing objects, in particular, for our understanding of the works as reflecting human intelligence -as intentional in this sense. While Wollheim certainly offers similar sets of considerations in other works (some of which were explicitly drawn on here), to my mind he nowhere presents any underlying matters as clearly or succinctly as in the 'lost' argument, which has given us a focus here. I have drawn on Wollheim's conception of the artist's theory in this sense, which provides the context in which the authorship should be understood. That, in turn, has exploited Wollheim's conception of the psychological and of agency. Hence -if the argument for the importance of 'history of production' considerations really does go through for all art forms (at least defeasibly) -the prospect of articulating 'the artist's theory' will be informative across the board: in all cases, making sense of an artwork as an artwork 108 will grant the connection to the artist's conception both of that art form and of his or her work's place in it.
Understood in line with an extension of (hypothetical) intentionalism, such an engagement with 'the artist's theory' replicates in a more forceful fashion Wollheim's concerns both with the artist's intentions properly understood (since the artist's theory simply reflects his or her aim, understood in the context of a narrative of the relevant art form) and with the demands of the suitably informed and suitably sensitive spectator, who may eventually be suitably prompted. For (especially) that spectator must take a view -in line with 'the artist's theory'of the place of this work in the narrative of that art form in the appreciation of the relevant artwork. And even when there is no pre-existing art form, an artist's theory for a proto-art (say, what would become opera, in the case of early works of Monteverdi, perhaps) can be drawn up, perhaps with hindsight, as a full-blown artist's theory. That recognizes, of course, that we can make sense of ascription of the artist's theory even in contexts where, if asked, the artist would not mention such a theory; or would deny holding one. That is, in the context of our hypothetical intentionalism, suitably generalized. For this conception exploits precisely the importance of authorship (and its attendant emphasis on responsibility), in the context of adherence to the artist's theory, as had been stressed in Wollheim's 'lost' argument.