Complete Artworks without Authors

Abstract

Genetically complete yet authorless artworks seem possible, yet it is hard to understand how they might really be possible. A natural way to try to resolve this puzzle is by constructing an account of artwork completion on the model of accounts of artwork meaning that are compatible with meaningful yet authorless artworks. However, I argue that such an account of artwork completion is implausible. Therefore, I leave the puzzle unresolved.

1. Introduction

Let us focus on an unlikely pair of examples: “Golden Age” American comics like “The Ski-Trail Murder” from Boy Comics #40 (1948), and the psychedelic James Bond spoof Casino Royale (1967), which is loosely based on Ian Fleming’s 1953 novel with the same name. While not all comics and films are artworks (e.g., instruction manuals and videos), I assume that Golden Age comics of this kind and Casino Royale are artworks.

I begin by setting out a plausible template for accounts of the authorship of artworks (what I will henceforth just call authorship). Then, I consider three accounts based on recent discussions of authorship that conform to the template. I argue that, relative to these accounts, it is unclear that the artworks mentioned above have authors, and it is possible for artworks produced in a manner similar to how they were produced to lack authors. And, as a more general matter, it strikes me that any account of authorship that both conforms to the template and is otherwise reasonable will have similar implications. I take this all as (defeasible) evidence that it is possible for artworks to lack authors.

If authorless artworks really are possible, then it should be possible for them to be genetically complete, or so I will argue. ¹ But, at the same time, I will show that, relative to a plausible approach to artwork completion, it is impossible for authorless artworks to be genetically complete. A natural way to try to resolve this puzzle is by constructing an account of artwork completion on the model of accounts of artwork meaning that are compatible with meaningful yet authorless artworks. I argue, however, that such an account of artwork completion is implausible. Therefore, I leave the puzzle unresolved. ²

2. Two Accounts of Authorship

Let us call properties like being a comic and being a film that correspond to particular artforms artform properties. ³ Where F is an artform property, if an individual or group of two or more individuals is an author of an F, this is so because that individual or group is appropriately related to an intention with the right content that plays the right causal/explanatory role. So, we have three issues to consider:

• • Content: what is the content of author-making intentions?

• • Role: what is their characteristic causal/explanatory role?

• • Relation: what is it to be appropriately related to an author-making intention?

We can view accounts of authorship as consisting of content, role, and relation conditions, where these conditions are answers to the questions above. The first account I will consider draws on Livingston’s (2005, chap. 3) discussion of authorship:

• • Content: an author-making intention is an intention to make an F with certain artistically relevant properties.

• • Role: such an intention contributes to the production of an F in a way that renders any individual appropriately related to it ultimately responsible for the F, at least in part.

• • Relation: an individual is appropriately related to an author-making intention when the individual possesses it.

While Livingston does not formulate things in just this way, I will refer to the above account as Livingston’s account for the sake of convenience.

Some points of clarification are in order. The relevant intentions are intentions to make Fs in contrast to unspecified intentions to “make some art” (2005, p. 43). While Livingston does not provide an account of artistically relevant properties, he points to the following as paradigmatic examples: having aesthetic value, exhibiting skill in the manipulation of expressive media, and expressing attitudes (2005, p. 90).

It may be that the role condition as presented above is underspecified (e.g., it does not speak directly to how intentions with the right content are formed). But it is the connection to responsibility that will be of particular interest in what follows. I take it that part of what it is to be ultimately responsible for a particular F is to be ultimately responsible for the fact that the particular is an F. ⁴ Consider a case of a collaboratively produced print with a single author and a single mere contributor (a print technician). In this case, the mere contributor has various intentions that contribute to the production of the print. But the thought is that these intentions do not do so in a way that renders that individual ultimately responsible, even in part, for it. So, the mere contributor fails to satisfy the role condition as specified in Livingston’s account. Ultimate responsibility instead traces back to the author alone. As Mag Uidhir (2013, p. 62) puts the point, in this case, the mere contributor’s intentions figure in the production of the artwork only as “proxy” for the author’s intentions.

Implicit in the above account is the idea that individuals rather than groups are candidate authors of artworks. Still, an artwork can have multiple authors. Livingston suggests that two or more individuals are collaborating authors of an F when they collectively intend to make an F with certain artistically relevant properties, and this intention contributes to the production of the F in a way that renders each of them, at least in part, ultimately responsible for the F. (The second part of this condition could alternatively be formulated in terms of collective responsibility.) Livingston appeals to Bratman’s (2014, chap. 3) account of collective intention in this context. Given Bratman’s approach, if two or more individuals collectively intend to φ, then each individually intends to do so. Hence, if two or more individuals collectively intend to make an F with certain artistically relevant properties, each individual has an intention with the same content. And, if their collective intention contributes to the production of the F in a way that renders each of them, at least in part, ultimately responsible for the F, then their corresponding individual intentions do as well.

There are, however, alternative approaches to collective intention. Consider Gilbert’s (2014: chap. 2) account. According to Gilbert, if individuals A and B collectively intend to φ, this is so because they have a joint commitment to φ “as a body,” where this means that they are jointly committed to emulating, by virtue of their actions, a single φ-er. Pace Bratman, Gilbert claims that if two or more individuals collectively intend to φ, it need not be the case that any of them individually intends to φ. What is required (among other things) is that each is individually committed to φ-ing with the other.

Bacharach & Tollefsen (2010, 2011) propose an alternative to Livingston’s take on collaborative authorship that appeals to Gilbert’s account of collective intention. And they focus on collective intentions to make Fs by way of certain processes or procedures, rather than collective intentions to make Fs with certain artistically relevant properties. Building on their discussion, consider the following account of artwork authorship:

• • Content: an author-making intention is the intention to make an F by way of certain substantive processes or procedures.

• • Role: such an intention contributes to the production of an F in a way that renders any individual appropriately related to it ultimately responsible for the F, at least in part.

• • Relation: an individual is appropriately related to such an intention when the individual either possesses the intention or is among a group that possesses it.

While B&T do not offer an account of authorship but instead only a necessary condition for collaborative authorship, I will refer to the account above as B&T’s account. Implicit in this account is the idea that individuals rather than groups are candidate authors of artworks. Suppose that a group has a collective intention with the right content that plays the right role relative to this account. In this case, each of the individuals in the group rather than the group itself is an author, even if none of them has an individual intention with the right content.

Here is a point of clarification before we continue. There are medium-specific processes and procedures such that an individual cannot rationally intend to make an F unless that individual intends to make an F by way of those processes or procedures. For example, it seems that an individual rationally intends to make a painting only if that individual intends to make something partly on the basis of the manipulation of paint. This is why the content condition, as specified above, appeals to substantive processes or procedures. A process or procedure is substantive in the relevant sense only if an individual can rationally intend to make an F without intending to make an F via that specific process or procedure. Returning to painting, presumably an example would be the intention to apply paint to the canvas not with brushes but squeezed directly from the tube. The rough idea is that an individual intends to adopt a substantive process or procedure in making an F only if their plan for making an F goes beyond doing whatever is minimally required to make an F. In the absence of this qualification, the process/procedure element in the content condition as specified above would not do any work.

What should we make of the idea of authorless artworks given these accounts of authorship? I begin with Livingston’s account and focus on Golden Age comics, which serve as a useful test case for accounts of authorship given their standardized but distinctive production process. Reflection on these comics suggests that authorless artworks are possible, if not actual, given Livingston’s account.

In the economic heyday of comic production, various comic studios in New York City adopted an approach to comic production that Will Eisner describes as “an Egyptian galley going down the Nile” (Hajdu, 2008, p. 27). With the advent of the comic book in the mid-1930s, subsequent comics production attempted to expedite the creative process by adopting an assembly line model popularized by, among others, the Iger and Eisner comics studios. While variations were numerous and commonplace, typically, a writer would devise a basic narrative conceit; a penciler would generate a drawn series of pages that conform roughly to the writer’s prompt; an inker would subsequently embellish the pencils to ensure their visibility during the photographic component of the production process; a letterer would insert speech balloons, narrative captions, and other text; and a colorist would then indicate the color design of the pages to be implemented by those working at the printing press, where color separations were done by hand. At any juncture in this process, revisions might be undertaken and steps might be skipped or repeated. The abrupt removal of pages, panels, or plots to ensure conformity with publishing aims was commonplace. Similarly, the use of assistants and apprentices at each stage was a regular practice, with established pencillers typically focusing on figure drawing and action scenes, while studio assistants might be tasked with penciling or inking backgrounds and perhaps lettering or coloring work. Certain artistic successes were made possible through virtuosity and coordination in writing, penciling, and other elements of comic craft.

Relative to Livingston’s account, it seems likely that some comics produced in this manner had no contributors who intended that the final product have particular artistically relevant properties. If so, then there are comics that fail to satisfy the content condition as specified by this account of authorship. Lev Gleason Productions, which adopted the assembly line studio model popularized by Eisner, Iger, and others, provides a useful example in the form of “The Ski-Trail Murder,” published in Boy Comics #40 (1948). ⁵ The story, credited to Virginia Hubbell, bizarrely compresses a romance, a western, and an FBI drug bust on the ski slopes into three and half pages, suggesting enormous editorial intervention and a model of production focused solely on the superficial adequacy of pages for inclusion in a comics magazine. Art credits are given to Mike Roy and John Belfi, with the latter doing “finishes” over Roy’s pencils. No colorist or letterer is credited.

It is, of course, possible that this and other comics produced by Lev Gleason Productions did have contributors with intentions with the right content relative to Livingston’s account. But clearly the studio could have produced comics in the absence of such intentions. The moral is that authorless artworks are possible, if not actual, given Livingston’s account of authorship. ⁶

Let us turn to B&T’s account of authorship, which does not require that an individual is an author of a comic only if that individual intends to make a comic with certain artistically relevant properties. On this account, a comic is authorless if none of its contributors intended to make a comic by certain substantive processes or procedures, and none of them were jointly committed to doing so.

Beginning with the issue of joint commitment, perhaps studio head Lev Gleason routinely expressed readiness to the others to take on a joint commitment to make comics as a body. Perhaps Gleason was individually committed to promoting the object of this commitment to the best of his ability in coordination with them. But it would not be surprising to learn that none of the others had individual commitments to either make a comic in coordination with the others or grant Gleason the authority to make them party to joint commitments to do certain things as a body. For instance, the extent to which Hubbell is in collaboration with Roy and Belfi in the production of “The Ski-Trail Murder” seems limited.

Did any of the contributors to the comics produced by the studio have individual intentions with the right content relative to B&T’s account? For each of these comics, perhaps Gleason intended to make that comic. Is it the case, however, that, for each comic produced by the studio, Gleason intended to make a comic via certain substantive processes or procedures? (Recall that an individual intends to adopt a substantive process or procedure in making an F only if their plan for making an F goes beyond doing whatever is minimally required to make an F.) Consider the assembly line procedure by which the comics were produced. Perhaps in each case Gleason intended to make a comic using this procedure, as he set up the studio’s assembly line in the first place. But does the assembly line method count as a substantive procedure in the relevant sense? It is hard to say in the absence of a substantive characterization of the substantiveness of comic-making processes and procedures. Until this issue is resolved, we are not going to be in a good position to establish whether “The Ski-Trail Murder” lacks authors given B&T’s account, or even whether it is possible for similarly produced comics to lack authors given this account. ⁷

At this point, I think it will be helpful to shift from Golden Age comics to Casino Royale. Happily, we know more about just how Casino Royale was made compared to “The Ski-Trail Murder” and the like. This film may very well lack authors relative to B&T’s account, or so I will argue. At the very least, it is possible for films produced in a similar manner to lack authors given this account. Importantly, my case for these claims does not hinge on the issue of what counts as a substantive process or procedure for making a film—we can put this issue to the side. Whenever I claim below that contributors lack (individual or collective) intentions to make a film by way of certain substantive processes or procedures (and thus fail to satisfy the content condition), this is because they lack the intention to make a film, period. So, it is not the case that they fail to satisfy the content condition because, while they intend to make a film, they do not have particular substantive processes or procedures concerning how to do so in mind.

Casino Royale is an example of what Livingston calls a “traffic jam” film—roughly, a film that is the unintended result of disparate intentional and unintentional activities (1997, p. 138; 2005, p. 80; 2011, p. 224). ⁸ Casino Royale had seven directors and twelve screenwriters (credited and uncredited), and it was filmed at three different British studios. Directors were responsible for different stand-alone vignettes and instructed not to communicate with each other. Almost no one involved in the film’s production, save for producer Charles Feldman, ever saw a complete screenplay, which was constantly under revision in any case. The matter of how the vignettes might fit together was considered only after principal filming was well underway, where one of the directors, Val Guest, was eventually assigned the role of “Additional Sequences,” that of creating transitional sequences between the vignettes in an effort to bring some coherence to the film. The final product is an episodic film with jarring lapses in continuity that resulted from the lack of coordination between the many contributors. ⁹

Feldman stated, “I think that film drove me crazy. I didn’t know what had been shot and what hadn’t been shot… I lost control” (Richardson 2015, p. 178). Joe McGrath, a director who was fired before his section was complete, stated: “There’s no control. Nobody has an overall feeling for the firm and what is happening” (Richardson 2015, p. 75). While the film was in production, the actress Ursula Andress reported, “I’m in a daze. I do not know what I’m supposed to say. I do not know which script, which director, which producer, or which scene. It is confusing (Richardson 2015, p. 81).

Let us return to B&T’s account. In this case, Casino Royale lacks authors, provided that none of its contributors intended to make a film using certain substantive processes or procedures, and none of them were jointly committed to doing so. Beginning with the issue of joint commitment, consider, for example, Feldman and the various screenwriters and directors. Perhaps Feldman expressed readiness to the others to take on a joint commitment to make a film as a body. Perhaps he was individually committed to promoting the object of this commitment to the best of his ability in coordination with them. But, given the disjointed way in which Casino Royale was made, it would not be surprising to learn that none of the other members of this group had individual commitments to either make a film in coordination with the others or grant Feldman the authority to make them party to joint commitments to do certain things as a body. And, if it turns out that the contributors did have the relevant individual commitments, clearly, a film could be produced by a similarly disjointed method in the absence of such commitments.

Provided that Casino Royale does not have collaborating authors relative to B&T’s account, might the film instead have a single author in this case? It is true that Feldman was largely responsible for the division of labor among the contributors, including the odd setup with the many directors. But, while he might have had an intention with the right content relative to B&T’s account, it just seems wrong to say that Feldman, on his own, was ultimately responsible for the film. ¹⁰ If Feldman were, then producers, as a matter of course, would count as being at least in part ultimately responsible for the films they produce. But they aren’t. For example, in an auteur film (where the auteur is not, in addition to their other roles, a producer), the producer is not even in part ultimately responsible for the film—ultimate responsibility instead resides wholly with the auteur.

What about the idea that Feldman is among a group of non-collaborating authors, say Feldman plus the various directors? The idea is that, while they did not have a collective intention with the right content that played the right role, each member of the group had an individual intention with these features. This proposal works only if each member of the group individually intended to make a film. But we can imagine that, besides Feldman, none of them did. Perhaps they were instead just concerned with fulfilling the terms of their contracts (i.e., filming various scenes with certain actors following a partial script) and escaping the project with their reputations intact. ¹¹

Putting all of this together, we now have a recipe for authorless artworks relative to B&T’s account with the following ingredients: no collaborating authors of the F, as no group of two or more contributors to the F has a collective intention to make an F; no single author of the F, as no single contributor to the F is ultimately responsible for it; and no group of two or more non-collaborating authors of the F, as any candidate group includes individuals who do not intend to make an F. It is also possible for comics and other artworks that aren’t films to satisfy these conditions, but I will not pursue this matter further here.

3. Another Account of Artwork Authorship

Above, I argued that authorless artworks are possible, if not actual, relative to two accounts of authorship based on recent discussions of authorship that conform to our template. In this section, I argue that the same is true of another such account, one that appeals to the grounding profile of art form properties.

For any art form property F, there are properties that Fs have fully in virtue of which they are Fs. For example, for anything that is a comic, there are properties the thing has fully in virtue of which it is a comic. Plausibly, the aim of an account of comics is to specify what these properties are, and the same goes for films, paintings, and so on. The account of authorship I have in mind draws on Mag Uidhir’s (2011, 2012, 2013, chap. 2) discussion of authorship, where G is a property that some Fs have at least partly in virtue of which they are Fs:

• • Content: an author-making intention is the intention to make a G.

• • Role: such an intention contributes to the production of a G in a way that renders any individual or group of two or more individuals appropriately related to it ultimately responsible for the G, at least in part.

• • Relation: an individual or a group of two or more individuals is appropriately related to such an intention when the individual or group possesses it.

While Mag Uidhir does not put things in just this way, I will refer to the above account as Mag Uidhir’s account.

Two points of clarification. First, note that on the previous accounts of authorship only individuals are candidate authors. But, on Mag Uidhir’s account both individuals and groups are candidates. However, I will continue to speak as if only individuals are candidate authors to simplify our discussion. Second, if an individual is ultimately responsible for something being a G, then, if that thing is an F partly in virtue of being a G, then the individual is at least in part ultimately responsible for the thing being an F. Responsibility for Fs flows from responsibility for those things having the features that make them Fs. ¹²

Suppose that for art form property F, there is some collection of properties such that, for anything that is an F, it is an F fully in virtue of having these properties. While not an explicit consequence of the account as stated above, I take it that in this case a given F has an author only if each of these properties is such that some contributor (or group of contributors) to the F intends to make something with that property. Otherwise, it would be possible for the F to have an author, yet no individual (or group of individuals) is ultimately responsible for it. In this case, among Fs with authors, there could be, as it were, merely partially authored Fs rather than fully authored ones. But, to the extent that this distinction even makes sense, it seems necessary that any authored artwork be fully authored.

It seems that comics produced by outfits like Lev Gleason Productions do not satisfy the condition just described. Plausibly, there is a collection of properties such that, if something is a comic, then it is a comic fully in virtue of having these properties. But it is not the case that, for each of these properties, comics produced by the studio had contributors who intended to make things with that property. Explaining why will require some stage setting.

Following Cowling and Cray (2022, chap. 2), the property of being a comic is maximal in the sense that something is a comic only if it is not part of a slightly larger comic. This is not something special about comics—as Sider (2001) notes, many ordinary sortal properties are maximal, such as being a house. Compare my house with the mereological difference between it and one of its windows, house-minus. It seems clear that house-minus and similar entities aren’t houses—alas, I only have a single house to call home!

Whether something is a house depends in part on what’s going on outside the borders of that thing, and the same goes for all maximal properties. So, maximal properties are extrinsic. And, for any maximal property P, there is some intrinsic property P* such that if something has P, then it has P partly in virtue of having P*, and partly in virtue of having the extrinsic property of not being a part of a slightly larger P*. Depending on what property P is, it may be that if something has P, then it has P fully in virtue of having these properties alone (e.g., where P is the property of being a rock). Or it may be that if something has P, then it has P fully in virtue of having these properties only together with some further extrinsic property (e.g., where P is the property of being a house).

Suppose that when something has property I_H it has what it takes, intrinsically, to be a house, which presumably includes having a structure of a certain type. As I indicated above, there is more to what it takes, extrinsically, to be a house than not being a part of a slightly larger I_H. Suppose that, when something also has property E_H, then it has the necessary extrinsic profile, which plausibly includes something like being regarded as a house. In this case, if something is a house, this is so fully in virtue of the fact that it is an I_H, it is not part of a slightly larger I_H, and it is an E_H.

In the case of being a comic, I take it that its associated intrinsic property is being a sequence of juxtaposed images. ¹³ And comics are like houses in that there is more to what it takes, extrinsically, to be a comic than not being a part of a slightly longer sequence of juxtaposed images. Suppose that, when something also has property E_C, then it has the necessary extrinsic profile, which plausibly includes certain representational and historical/institutional properties. And let I_C be the property of being a sequence of juxtaposed images. In this case, for anything that is a comic, that thing is a comic fully in virtue of the fact that it is an I _C, it is not part of a slightly longer I_C, and it is an E_C.

For any comic produced by Lev Gleason Productions, plausibly one or more contributors to the comic intended that it be an I_C. But it seems unlikely that any contributor intended that any one of them not be a part of a slightly longer I_C. And it seems unlikely that any contributor intended that any one of them be an E_C, whatever this property comes to exactly. This is not to say that there are not conditions under which comic contributors have such intentions. A contributor to a comic might intend that it not be a part of a slightly longer I_C in response to pressure from an editor to add another panel. Or an individual might decide to make a comic rather than a sonnet partly on the basis of having formed the intention to work in a tradition that traces back to eighteenth-century Britain rather than thirteenth-century Italy. The point is just that these aren’t intentions that comic contributors have, as a matter of course, and we have no reason to think any of the contributors in Lev Gleason Productions, in particular, had such intentions. And, if each comic produced by the studio (miraculously) had contributors with intentions with the right content relative to Mag Uidhir’s account, clearly, the studio could have produced comics in the absence of such intentions. So, authorless comics are possible, if not actual, given Mag Uidhir’s account. Similar considerations apply to Casino Royale.

Given Mag Uidhir’s account, it seems that the default status for comics and films in general is that they are authorless, as contributors normally just aren’t going to have all the relevant intentions. So, there is nothing really special about Golden Age comics or Casino Royale in this context—we could have illustrated the same points with many other examples of comics or films. And there’s more—plausibly, all art form properties are maximal, and, if something has an art form property, it has it partly in virtue of the fact that it has certain representational and historical/institutional properties. So, given Mag Uidhir’s account, it seems that the default status for artworks in general is that they are authorless!

4. A Puzzle

It strikes me that any account of authorship that both conforms to our template and is otherwise reasonable will be compatible with authorless artworks. ¹⁴ So, let us suppose that such artworks are really possible. As I noted at the outset, in this case, we should expect it to be possible for such artworks to be genetically complete. The rationale here is straightforward. Consider a possible authorless artwork and an authored counterpart. Both were produced by creative processes constituted in part by their contributors. It is possible for the process at issue with each artwork to reach an appropriate endpoint. If the creative process by which an authored artwork was produced reaches such an endpoint, that artwork is genetically complete. The same should, therefore, be true of authorless artworks—if the creative process by which an authorless artwork was produced reaches an appropriate endpoint, that artwork also is genetically complete. ¹⁵

At this point, however, we run into a problem. Given the plausible psychological approach to artwork completion, it is impossible for authorless artworks to be genetically complete, or so I will argue. So, we face the following puzzle: it seems that genetic completion depends on authorship in that it is necessary that an artwork is genetically complete only if it has an author; yet it also seems that genetic completion is independent from authorship in that it is possible for there to be genetically complete yet authorless artworks.

According to the psychological approach to artwork completion, whether the creative process by which an artwork was produced has reached an appropriate endpoint is ultimately a psychological matter. Different psychological accounts specify different psychological states as being the ones relevant to artwork completion. Trogdon and Livingston (2014), for example, propose an account in terms of certain cognitive dispositions and the mental states in virtue of which artists have these dispositions.

On the psychological approach, it is the psychological states of the artists of an artwork that matter to whether it is genetically complete, rather than the psychological states of others closely related to the artwork (e.g., critics, audience members, or technicians who contribute to its production). But why should this be the case? The answer is that an artist of an artwork (in the sense of ‘artist’ relevant here) is an author of it. When an individual is an artist of an artwork, their psychological states in particular are relevant to the artwork’s completion status because that individual, as an author of the artwork, is such that their psychological states ultimately direct the creative process by which the artwork is produced. It is their psychological states that both initiate this process and determine (at least in part) whether it reaches an appropriate endpoint. Putting all of this together, the psychological approach in effect says that an artwork is genetically complete only if its authors have certain features. When viewed in this way, it is obvious that this approach rules out authorless artworks.

Now, perhaps you are not sympathetic to the psychological approach. Rohrbaugh (2017), for example, argues that the psychological approach makes implausible predictions about the first-person perspective of artists. While I will not go into the details here, I think this objection fails, as it apparently hinges on treating certain opaque contexts (e.g., contexts in which an artist questions whether their artwork is complete) as being transparent. Still, it is worth noting that there are no complete yet authorless artworks relative to certain non-psychological accounts as well, including an account similar to Rohrbaugh’s own non-psychological account.

Let an artistic plan be a (perhaps schematic) plan for making something with certain artistically relevant properties. Here’s a proposal similar to one Rohrbaugh (2017) proposes: if an artwork is complete, this is so because (i) the artwork has an artist or artists; (ii) either the single artist individually or the multiple artists collectively have an intention whose content includes an artistic plan for the artwork; and (iii) the artwork satisfies that plan. The overall thought is that for the creative process by which an artwork was produced to reach an appropriate endpoint is for the artwork to satisfy the artist’s plan for it. Whether an artwork satisfies an artistic plan is not itself a psychological matter, and this is the sense in which the account of artwork completion is a non-psychological account.

Note that, in addition to artistic plans, there are curatorial plans. Suppose that a curator’s plan for an artwork includes displaying it in an exhibition next to another artist’s artwork to highlight similarities between them. Given the above account of artwork completion, why is the artist’s plan for the artwork (whatever it is exactly) directly relevant to its completion status rather than the curator’s plan? The answer is that the artist rather than the curator is an author of the artwork—the artist’s plan rather than the curator’s ultimately directed the process by which the artwork was produced. Or consider a case in which an artist’s plan for their print is thus-and-such, and they have commissioned the help of a print technician in implementing their plan. Imagine that, after conversations with the artist, the technician comes to embrace the same plan for the artwork. But it is the artist having this plan rather than the technician having it that is directly relevant to its completion status, and the fact that the former rather than the latter is an author of the artwork explains this asymmetry.

What should we ultimately say about this puzzle? I confess that I’m not sure. I will close by sketching one way you might try to resolve it. The idea I have in mind is this: there is a prima facie plausible account of artwork meaning that is compatible with meaningful yet authorless artworks, and we can construct a corresponding schematic account of artwork completion. This account departs from extant accounts of artwork completion in interesting ways and is compatible with complete yet authorless artworks. ¹⁶

What an artist intends to do with their artwork, in contrast to, say, what a curator intends to do with it, is potentially relevant to its meaning. This contrast is grounded in the fact that the artist, rather than the curator, is the author of the artwork. When an author intends to say or convey thus-and-so with their artwork, call this a meaning intention. Consider a meaningful poem. Suppose that it has an author with certain meaning intentions, and these intentions are successful. Importantly, success in the relevant sense is not to be understood in terms of whether the intended meaning of the poem (i.e., what the poet tries to say or convey with it) is included in its actual meaning (i.e., what the poem actually says). On one approach, a meaning intention is successful when there is the right kind of uptake of the intended meaning by the appropriate audience (Stecker 1997, chap. 9).

In this case, the poem has the meaning it has partly in virtue of the fact that the author’s meaning intentions are successful. But there are various contextual factors and conventions that are relevant as well. For example, if the poem contains the word “sleep,” then, not only is what the word means in ordinary discourse relevant, but so too is the fact that in poetry, it is associated with death in particular. If there are, say, multiple unintended anagrams in the poem, this too would be relevant to its meaning, as there is an artistic convention according to which anagrams are common and artistically significant in poems. And there are authorial intentions in addition to meaning intentions that are potentially relevant as well, such as the intention to make a poem.

Let us say that an artwork satisfies the meaning intention condition just in case it has an author with a successful meaning intention with respect to it. Following Stecker (1997, chap. 9, 2003, chap. 2), if an artwork has a meaning, then, if it satisfies the meaning intention condition, it has the meaning it has fully in virtue of this fact together with certain conventional and contextual factors. And, if it has a meaning but does not satisfy the condition, then the artwork has the meaning it has fully in virtue of certain conventional and contextual factors alone.

Suppose that a meaningful artwork does not satisfy the meaning intention condition. Why might it not satisfy the condition? One possibility is that, while there is some author of the artwork with a meaning intention, no author has a successful meaning intention with respect to it. Another possibility is that, while there is an author of the artwork, no author has a meaning intention with respect to it, period. But there is a third possibility as well, one that Stecker as far as I can tell does not consider: the artwork just does not have an author. So, I take it that the above account is compatible with meaningful yet authorless artworks.

We can use this account as a model for a schematic account of artwork completion. When an author intends not to work on their artwork further, let us call this a completion intention. And let us say that an artwork satisfies the completion intention condition just in case it has an author with a successful completion intention with respect to it. Importantly, success in the relevant sense is not to be understood in terms of whether the artwork is in fact complete or not. On one approach, a completion intention is successful when it is backed by the right kind of reason. Clearly, a reason to abandon the artwork would not count as the right kind of reason.

If an artwork is complete, then, if it satisfies the completion intention condition, it is complete fully in virtue of this fact, together with certain conventional or contextual factors. What sorts of conventional or contextual factors might be relevant? Plausibly, one relevant contextual factor is that no contributor to the artwork continues to work on it. Perhaps another is the truth of the following counterfactual: were the possibility of working on the artwork further to become salient to contributors and they were capable of working on it further, they would not do so. If a contributor to an artwork goes on to make minor changes to it (e.g., adding a missing signature or correcting typos), this does not count as continuing to work on it in the relevant sense.

Suppose that a complete artwork does not satisfy the completion intention condition. In this case, it is complete fully in virtue of certain conventional or contextual factors on their own. And why might it not satisfy this condition? It is not because, while there is some author of the artwork with a completion intention, no author of it has a successful one. And it is not because no author of the artwork has a completion intention with respect to it, period. In either case, the artwork would be incomplete. Instead, the artwork does not satisfy the completion intention condition because it is authorless. So, our schematic account of completion is compatible with complete yet authorless artworks.

There is a problem, however, with this approach to artwork completion. Consider the conventional and contextual factors, whatever they are exactly, fully in virtue of which a complete artwork is complete when it lacks authors. These factors are enough on their own to make the artwork complete. It seems that the same factors are operative in cases of complete artworks that have authors. The upshot is that the proposal is self-undermining, as apparently, the completion intention condition does not play an essential role in artwork completion. In this case, it seems that conventional or contextual factors on their own fully explain why complete artworks are complete, even when the condition is satisfied. Facts to the effect that an author intends to stop working are explanatorily otiose.

Note that the account of artwork meaning under discussion does not face a corresponding problem. It is true that on this account there are cases in which conventional and contextual factors fully explain on their own why an artwork has a meaning. But, when the meaning intention condition is satisfied, such facts do not fully explain on their own why the artwork has the meaning it has. The crucial difference here is that, while there is a distinction between an artwork having some meaning or other (i.e., being meaningful) and having a particular meaning, there is not a corresponding distinction with respect to artwork completion. ¹⁷