About CA
Contact CA

Search Journal

Editorial Board

Permission to Reprint


Site Map


On Canon

  Craig Derksen & Darren Hudson Hick

Canon is a concept from aesthetics that has become a regular subject of commonplace discussions. The nature of canon, especially as it is used in these commonplace discussions, has not been subject to adequate philosophical scrutiny. We attempt to remedy that by placing canon in its historical and philosophical context, exploring and rejecting several common accounts, and presenting some basics of how canon works. We reject the accounts that place control with the author or the legal property holder, which appear to be the most commonly held accounts.

Key Words
canon; intellectual property; interpretation; truth in fiction


1. The truths of fiction

In his seminal 1978 paper, "Truth in Fiction," and also "Postscripts" in 1983, David Lewis sketches a theory about how we can make truth-functional sense of propositions about fictions.[1] Take the statement, "Sherlock Holmes has hands," for example.[2] Now, assuming that we take the statement to refer to the fictional character created by Sir Arthur Conan Doyle, we want to frame the domain of our proposition. For purposes of analysis, Lewis suggests that such a proposition should be regarded as being prefaced with the operator: "In such-and-such fiction…" or, in this case, "In the Sherlock Holmes stories…."[3] The fiction, Lewis suggests, is best understood not as a string of sentences but rather as "a story told by a storyteller on a particular occasion," in particular, a story told in a world where that story is known fact, that is, known to the storyteller, and not as a fiction.[4] Conveniently, in all but four of the Holmes stories written by Doyle, it is Holmes’s companion, Dr. Watson, who narrates the stories.

Lewis famously analyzes counterfactual truths by looking to possible worlds, worlds where things turned out differently than they did in ours. There are countless such worlds that contain an entity named Sherlock Holmes, and in many of these he is a detective. Now, in some of these, he is a man with two hands, and in others he is a sort of slug with no hands at all: the possibilities are endless. But, when we are talking about Sherlock Holmes, we aren’t normally talking about the slug-Holmeses or most of the other Holmeses; we’re talking about the worlds that align with what’s written in the stories we have read. That Holmes has more than one hand appears unproblematically confirmed by the text of "The Disappearance of Lady Frances Carfax" (1911) and The Valley of Fear (1915), both narrated by Watson. So, in all of the worlds that align with the story-as-told in Doyle’s Holmes stories, where these stories are told as facts known to the storyteller, it would be true that Holmes has hands.

Now, take the statement, "Sherlock Holmes is left-handed." What are we to make of this? Nowhere in any of the stories written by Sir Arthur Conan Doyle is it suggested that the detective is left-handed. Nor, however, is it said anywhere that he is right-handed or ambidextrous. That is, there are equally as many worlds where Doyle's stories (as written) are told as known fact where Holmes is is right-handed, left-handed, and ambidextrous. So, on Lewis’s theory, the matter is indeterminate: "What is true throughout [the worlds] is true in the stories; what is false throughout them is false in the stories; what is true at some and false at others is neither true nor false in the stories."[5]

So far as it goes, Lewis’s theory would seem to have a lot going for it, even if we don’t buy into the extremes of modal realism.[6] Towards the end of "Truth in Fiction," however, Lewis makes a curious claim:

I have spoken of Conan Doyle’s Holmes stories; but many other authors also have written Holmes stories. These would have little point without inter-fictional carry-over. Surely many things are true in these satellite stories not because of the explicit content of the satellite story itself, and not because they are part of the background, but rather because they carry over from Conan Doyle’s original Holmes stories. Similarly, if instead of asking what is true in the entire corpus of Conan Doyle’s Holmes stories we ask what is true in "The Hound of the Baskervilles," we will doubtless find many things that are true in that story only by virtue of carry-over from Conan Doyle’s other Holmes stories.[7]

Certainly this makes some sense. In A Study in Scarlet, the story of Holmes’s and Watson’s first adventure, it is established, by Watson’s estimation, at least, that Holmes is an accomplished violinist. So it should follow that in "The Stockbroker’s Clerk," which takes place some years later, Holmes is an accomplished violinist, though no mention is made of this. Of course, things can work in the opposite direction, too. In "The Greek Interpreter," we are introduced to Holmes’s elder brother Mycroft, whose deductive powers outstrip those of Sherlock. The events of this story take place several years after A Study in Scarlet. Though Mycroft goes unmentioned in the first Sherlock Holmes adventure, surely it is true that if Holmes has an older brother in "The Greek Interpreter," then he has the same older brother in the earlier Study in Scarlet.

Now, what are we to make of the story, "A Scandal of No Importance," in which Holmes, in the fall of 1897, reveals his romantic feelings for Watson? The story is written in six chapters by one "amalcolm1" and can be found on www.fanfiction.net, a repository of stories of familiar characters written by their fans. As a work of fan fiction, "A Scandal of No Importance" is relatively tame fare, but the story explains so much. In "The Blanched Soldier" (one of only two of Doyle’s stories that Holmes himself narrates) taking place in 1903, the detective states, "The good Watson had at that time deserted me for a wife, the only selfish action which I can recall in our association." Discounting "A Scandal of No Importance," the claim of desertion is odd. Watson has been married before several times, it seems, and has moved out before. What makes this action in 1903 so selfish? Well, Watson’s previous wife, Mary Morstan, died in 1894, and he remained single until 1903. So, Watson’s marriage in "The Blanched Soldier" would seem to be his first heterosexual romance since Sherlock declared his love for the doctor in 1897.

"But," you declare, "'A Scandal of No Importance' is fan fiction! It didn’t happen! At the very least, it’s not the same Sherlock Holmes as the one Doyle was writing about. It isn’t canon!" Exactly right.

2. Holmes and canon

Usually, when canon is discussed in the aesthetics literature, it is with reference to "the canon" or "the Western canon," the body of literature, music, and art generally considered the most influential or important to Western culture. Another historical use of canon in art and aesthetics refers to an evolving set of rules of beauty and art tracing back to Polyclitus’s system of proportions, spelled out in his fifth-century BCE treatise, titled The Canon, and exemplified in his sculpture of Achilles, also sometimes called The Canon. Neither of these uses of 'canon' is in the sense that interests us here, though both are related in meaning. Rather, the use of canon that concerns us refers to material accepted as making up part of a particular fiction, especially a story stretching over more than one work: roughly, the official, authorized, or accepted story, though this sketch of a definition will need refinement.[8]

It’s curious that Holmes has become the running case in philosophical discussions about fiction, as the first works of popular literature to garner serious discussions of canon, in the sense that interests us here, were the stories of Sherlock Holmes. Although Holmes has appeared in more movies than any other character, on stage, radio, and television, and has been co-opted across a variety of media by a host of other authors, the standardly accepted Sherlock Holmes canon consists of sixty adventures told in fifty-six short stories and four novels, all written by Doyle. Some have argued for additional works to be included in the canon, including two Holmes parodies written by Doyle and a selection of Holmes stories written by Doyle’s son and his biographer, but these are contentious.[9]

The Holmes canon has been a lively subject of serious literary debate for over a century, though it started with something of a joke. One of the earliest forays into scholarly consideration of Holmes was Ronald A. Knox’s "Studies in the Literature of Sherlock Holmes," first published in 1912 and often credited with birthing the field as a whole.[10] The now-legendary essay presents a tongue-in-cheek argument that the Holmes stories written after 1893’s "The Final Problem" should be considered apocryphal, as falling outside the canon.[11] In "The Final Problem," Doyle killed off Holmes, only to bring him back in 1901. The twentieth-century Holmes stories, Knox suggests, are not actually stories of Holmes but rather fabrications of the narrator’s (Watson’s) imagination:

The evidence against these stories may be divided into (a) those suggested by changes in the character and method of Holmes, (b) those resting on impossibilities in the narrative itself, (c) inconsistencies found by comparison with the previous narrative.[12]

In other words, Knox suggests that these stories should be excluded as apocryphal stories of Sherlock Holmes because they are inconsistent with the detective’s character, with other accepted stories, and within themselves.

Curiously, these are some of the central criteria upon which materials have been excluded from the biblical canon. Perhaps this should come as no real surprise: Knox himself was a Catholic priest. Theologian Kilian McDonnell notes that, for biblical canon, the question was not centrally a historical question of authorship but rather a theological one: "[T]he Apostolic Constitutions, a late fourth century canonico-liturgical compilation… warned that 'we must not rely on the attribution to the apostles, but attend to the character of the material and the correctness of the thought.'"[13] That a book was written by an apostle was less important than that it represented the person and divinity of Christ, that it was a coherent book, and that it meshed with already-accepted canonical materials.[14]

Thus the emphasis often placed on authorship makes the Holmes canon more the exception than the rule. One of the hottest ongoing debates in the domain of canon is what we might call the Star Wars universe. George Lucas had a hand in all six of the original movies, as screenwriter, director, or both, enough to likely qualify him as an author, perhaps the author, of these works.[15] But he had less to do with 2015’s Star Wars: The Force Awakens, 2016’s Rogue One: A Star Wars Story, and 2017’s Star Wars: The Last Jedi. Nevertheless, few would suggest that these later films are not canon. Nintendo’s Legend of Zelda game series presents a similar challenge: no single individual was instrumental in the creation of every game in the Zelda canon, although Takashi Tezuka comes closest. Authorship appears to offer neither necessary nor sufficient criteria for inclusion of a given work in a canon.

An ontology of characters might provide insight into an account of canon. If a work is canon, then it depicts the same character(s) as other works in the canon; if it falls outside the canon, it does not. In her analysis of fictional characters, Amie Thomasson suggests that character x (appearing in work W) and character y (appearing in a later work W') are the same fictional character only if the author of W' is well-acquainted with x, and intends to import x into W' as y.[16] This account is only intended to provide a necessary condition for x and y to be the same character and not a sufficient condition. In regard to sufficient conditions, Thomasson suggests that if the properties attributed to y are too different from those attributed to x in W, then x and y are not the same character.[17] Although Thomasson contends that she isn’t providing a sufficient condition, and the aberrant sort of case just discussed notwithstanding, she suggests that her criterion provides "a very good benchmark for whether or not we can reasonably claim that two characters are the same."[18] However, as things stand, your writing a Holmes story where the detective’s attributes are perfectly in line with those of Doyle’s detective will not make them the same character. You don’t have the requisite authority to make that happen.

3. What is canon?

Although authorship is not the sole determining factor of canon inclusion, there is nevertheless a matter of authority at play. And, as a matter of practice, such authority seems wrapped up with legal ownership. The Legend of Zelda series is legally owned by Nintendo. Legal ownership of the Star Wars franchise passed from George Lucas to Disney in 2012. Following acquisition, Disney has formally pronounced a list of canonical works and relegated the rest to Legends status. In 2011, Nintendo published The Legend of Zelda: Hyrule Historia, a sort of bible that laid out the official chronology of the Legend of Zelda canon, excluding a number of games. However, this official chronology came with a caveat:

As the stories and storytellers of Hyrule change, so, too, does its history. Hyrule’s history is a continuously woven tapestry of events. Changes that seem inconsequential, disregarded without even a shrug, could evolve at some point to hatch new legends and, perhaps, change this tapestry of history itself.[19]

So canon is open to change, seemingly at the whim of the legal authority. As a particularly good example of this rule, we might consider the Highlander franchise. The original movie, Highlander (1986), introduced audiences to the mysterious race of Immortals. The first sequel, Highlander II: The Quickening (1991), made drastic changes to the story set out in the original film, effectively rewriting key points in the story’s history, generally referred to as a "retcon" for "retroactive continuity." These changes were not well received. Critic Roger Ebert gave the film half a star, calling it “the most hilariously incomprehensible movie I’ve seen in many a long day.”[20] So the producers did the only reasonable thing they could. They premiered a TV show the next year, Highlander: The Series, and pretended the second movie had never happened, re-rewriting history. Then, a third movie was released, Highlander III: The Sorcerer (1994), that ignored both the TV series and the second movie. The fourth and fifth films then followed the TV series’ continuity, ignoring the second and third movies.[21] Similar retconning moves can be found in the original Halloween movie series, where the seventh and eighth films[22] ignore the events of the third through sixth movies in the series;[23] in the Rocky franchise, where later films effectively ignore the unpopular Rocky V;[24] and in the Exorcist film series, where the third movie ignores the second film, as do two later prequels which are incompatible with each other.[25]

All of these changes have been authorized by the legal authority. But the fickleness of the legal authority is not its only problem. A work being authorized by a legal authority does not cement its status as canon, either. The estate of Sir Arthur Conan Doyle has authorized any number of stories of young Sherlock Holmes, including the 1985 film, Young Sherlock Holmes; the 2011 graphic novel, Sherlock Holmes: Year One; and the Young Sherlock Holmes series of novels by Andrew Lane. Some of these directly contradict the content of Doyle’s stories, and all of them contradict one another. None, however, are widely accepted as canon despite being authorized by the legal authority.[26] Perhaps the most (in)famous example of an authorized work excluded from canon is The Star Wars Holiday Special (1978), released between the original Star Wars (1977) and The Empire Strikes Back (1980), which features Chewbacca’s extended family and Boba Fett. The show aired only once on television, in 1978, and has never been rebroadcast or officially released. Although Boba Fett would go on to appear in the film series, this iteration of Chewbacca’s family would not, and the Holiday Special would be wiped from official Star Wars continuity.[27]

When works in a series are particularly poorly received by fans, it is not unusual for legal authorities to overwrite them. Even Doyle did this when he brought Holmes back from the dead, bending to enormous pressure from fans.[28] At other times, a retcon may be motivated by other marketing reasons.[29] But, while fans, and other factors, can apply pressure or offer ideas, ultimately it seems only the legal authority has the power to decide what is and is not canon, sometimes by official pronouncement, as in the cases of Star Wars and The Legend of Zelda, or by releasing a new work that overwrites some past work, as with Highlander, Halloween, and Rocky.

So, with all this in mind, it might seem that it is the legal authorities who control canon and, so far as it goes, this is largely true. Today, for better or for worse, the authority over canon will largely rest with the person, persons, or corporate entity holding copyright over the works and characters in question. Copyright ownership consists in a certain sort of authority, centrally, the exclusive right to make or authorize copies of one’s protected works. By extension, copyright provides the owner with the exclusive right to make or authorize derivative works based on one’s protected work. It also provides protection for at least some fictional characters. Since the most obvious place where such extensions of copyright come into play is with regard to making or authorizing sequels, the connection between copyright and canon is fairly intuitive.

Certainly, copyright owners can and do use their authority to quash unauthorized sequels featuring their characters. In 2009, J.D. Salinger successfully sued the author of 60 Years Later: Coming Through the Rye, an unauthorized sequel to Catcher in the Rye, preventing the novel’s distribution in the United States.[30] J.K. Rowling and Warner Bros. have similarly worked in recent years to prevent publication or distribution of unauthorized sequels. Andrea Phillips notes: "At the end of the day, the core owner of the property in question is the one with the power to choose who profits and how much, or what becomes canon and what doesn’t."[31] Owning copyright in a work and its characters means getting to decide which stories are released, when protected stories are re-published in revised editions, and which of the many works featuring protected characters are authorized, and all of this is supported by an international legal system.[32]

Where an unauthorized work is released featuring characters protected by copyright or which would be treated as a derivative work in the law, copyright owners have the legal authority to quash that work. Where there is no copyright in a work or its characters, canon quickly becomes murky. One might not expect any certain answer to a question about the "Arthurian canon" or the "Robin Hood canon."[33] Although, thanks to the popularization of select works, certain versions of these characters, and also Santa Claus, Count Dracula, and others, may have become more cemented in the public imagination over time, the lack of any recognized authority makes canon a more difficult matter.

All this would seem to support the idea that, as one legal scholar puts it, "In the context of fan-based activities and their resultant works, the concept of canonicity draws a clear and nigh impenetrable barrier between the works of the original author and the fan works based upon it."[34] However, it now needs to be pointed out that the practice that establishes the copyright holder as the authority over canon is utterly contingent. None of the copyright-holder’s authority over canon necessarily follows from that copyright ownership. Copyright legally protects works and certain elements of those works, but there are no laws protecting canon. Nothing about the ownership of works or even fictional characters provides the copyright owner with the power of determining that his or her sequel, rather than some unauthorized work of fan fiction, is the real/true/better/right/canonical story of what happened to Harry Potter or Sherlock Holmes next. Those of us who concern ourselves with cases of canon do, as a matter of practice, typically treat the copyright owner as if he or she has this authority, but it is not a part of her copyright ownership, and not of any logical extension of that ownership.

Although the connection between copyright and canon is not altogether arbitrary, it is nevertheless altogether contingent. Why, as a matter of contingent fact, we do give such authority to the copyright holder is an anthropological question, and there is almost certainly a complex anthropological answer. Perhaps we recognize such authority because copyright owners hold an official position of authorizing sequels, and we simply conflate the two. Perhaps it is because we are lazy and would prefer to hand over the reins rather than be forced to make the decisions for ourselves. However, there seems no in-principle reason that fans might not rise up and wrest control over canon from the copyright holders, installing some new criterion for what will be canon, perhaps recognizing canon in the best stories, regardless of who has written them and whether they are authorized by the copyright holders.

In the case of legally unregulated characters, like Robin Hood or Santa Claus, we might imagine such criteria as the source of the stories, historical priority, or public consensus. Indeed, we can already see some cracks in the veneer. When George Lucas released the Special Edition of Star Wars in 1997, a great many fans were aghast at one change Lucas had made to the film: In the original 1977 release, Han Solo, questioned at gunpoint in the cantina bar by the bounty hunter Greedo, shoots Greedo dead before calmly strolling out of the bar; in the Special Edition, Greedo shoots first and misses, and Han fires the fatal shot a fraction of a second later. "Han shot first" became a rallying cry that hasn’t yet died down. In this case, fans are unwilling to accept Lucas’s attempt to change canon.

Holmes, too, is instructive. Although the detective’s copyright was held by the Doyle estate well into the twenty-first century, die-hard Holmes fans have generally rejected from the accepted canon anything but the stories written by Doyle himself.[35] Indeed, this blanket rejection may be the result of the Doyle estate’s indiscriminate licensing promiscuity, or it may be some reverence for the author himself. Regardless, the until-recent legal authority had by the author’s estate does not seem to have translated into recognized authority over the canon.[36]

Henry Jenkins notes that the modern fan fiction movement began in part to correct what fans saw as errors and oversights by the works’ creators, particularly female fans, and particularly fans of Star Trek. Indeed, Jenkins notes, "many fan writers characterize themselves as 'repairing the damage' caused by the program’s inconsistent and often demeaning treatment of its female characters."[37]

If control over canon were taken from the copyright holder, where could it go? This centrally depends on whether people want to maintain a single sense of canon, many senses, or no senses. The establishment of the biblical canon was intended to be driven by the authority of expertise, requiring careful and considered judgment by those most familiar with the materials, history, and repercussions of any determination and not authoritative fiat. It isn’t difficult to imagine a revised practice putting canon determinations in the hands of those with the greatest expertise. Alternatively, Elisabeth S. Aultman imagines a scenario where a group of participating fan-authors are collectively in charge of a canon, where that canon "would be determined by the number of up votes a given submission gets by members of the participating online community."[38] In Aultman’s imagined scenario, canon is determined by popular vote.[39] Again, it isn’t too difficult to imagine a revised practice distributing equal power and responsibility among those most invested in the outcome the determination. Of course, if fan fiction could impact canon, there would only be more reason for copyright holders to sue to forestall these derivative works, once again showing the interaction of the two. So perhaps this isn’t a power that fans want, after all. And, again, any new authority in canon determination could be just as arbitrary as our current arrangement.

4. How canon works

Usually, if we refer to the Holmes canon, we are picking out a set of works: the 56 short stories and four novels generally accepted as encompassing the official story of Sherlock Holmes. This appears similar, then, to talk of the Western canon. But, since we are centrally talking about an extended story, "canon" is also used as shorthand for the official chronology of events and the facts and rules derived, or, what happened in that story, when, why, and to whom. Typically, such facts are determined in-story but matters of canon may also be established by official pronouncement outside the narrative itself. Marvel Comics, for example, has established a number of canonical facts in its Official Handbook of the Marvel Universe, an encyclopedic series of periodicals with entries on each of the major characters, groups, and technologies appearing in Marvel Comics stories, where entries often outstretch details provided in-story.[40] In the same series, Marvel established the rule that for every four to five publication years in the real world, one year passes in-story in the Marvel Universe.[41] J.K. Rowling has similarly released a steady stream of facts about the Harry Potter universe outside of the novels themselves. In addition to publishing editions of textbooks discussed in-story, in 2012 Rowling launched the Pottermore website, unveiling secrets and histories for fans of the series. These revelations are generally, though not universally, accepted as canonical.[42]

As such, canon is tied up with philosophical issues of both truth-in-fiction and interpretation. While these two are often divided based on their subjects (concrete fictional facts versus more ephemeral claims about things like themes, meaning, and messages), they can also be divided based on approach. Truth-in-fiction usually takes a descriptive approach to the answers it offers by attempting to establish facts. Under current practice, interpretation often offers a hypothetical approach, proposing readings based on the content, form, and role of a story.

Although it is centrally a matter of truth-in-fiction, canon prescriptively establishes facts of a story or fictional world by determining what rules apply and what events should be included. Sometimes this will be a coarse-grained matter: Zelda: The Wand of Gamelon is not canonical, nor are the events that occur in it. And sometimes this will be more fine-grained: officially, the original Star Trek TV and film series takes place in the twenty-third century, though occasional statements in particular episodes would appear to place it in the twenty-second or twenty-eighth century; these statements are non-canonical, though the episodes in which they appear are otherwise canonical.

To say that a work, a fact, or a rule is canonical is normally to say that there was another viable option, given what has been descriptively stated of the story or fictional world. In other words, to say what is canonical is ordinarily to imply that an official or authoritative choice, a sanctioning, has been made either to include or to exclude some work, fact, or rule. As such, it is unusual for some straightforwardly descriptive fact of a story or fictional world to be referred to as canon or canonical.[43] It would be strange to ask, for instance, whether it is canonical that Hogwarts School of Witchcraft and Wizardry is in Britain, as this has never been a matter of debate. However, it would be perfectly reasonable to ask whether it is canonical that the Gotham City and Metropolis of DC Comics are connected by a bridge, or rather are hundreds of miles apart, a surprisingly flexible issue. Sometimes an apparent in-story discrepancy will prompt an authoritative decision about canon; sometimes such a decision will simply be prompted by creative or marketing decisions.

The authority that sanctions the facts of canon does not likewise authoritatively sanction interpretations. Disney’s prescriptive exclusion of the 1980s Droids and Ewoks cartoons from the Star Wars story is a matter of canon, but an official interpretive claim by Disney, say, that the Star Wars saga is centrally about the cyclical struggle of good and evil embodied in the story of the rise, fall, and redemption of Anakin Skywalker as a tragic hero, would not be canonical.[44] Although canonical interpretation is a widespread notion in biblical studies, it would seem to have little application in studies of literature. That being said, what qualifies as canon can have an enormous impact on interpretation. If amalcolm1’s "A Scandal of No Importance" is in the Holmes canon, then it would seem perfectly reasonable to interpret Sherlock’s feelings of betrayal towards Watson in "The Blanched Soldier" as those of a jilted lover and to place some importance on the fact that nowhere in the later story does the detective mention his love for the doctor. If the story falls outside the canon, however, this interpretation would seem altogether unfounded.

Perhaps it is not altogether surprising that canon has gone largely overlooked and underdiscussed in the aesthetics literature until quite recently.[45] There is, after all, no Dickens canon; there is no Tolstoy canon; there is no Shakespeare canon, at least, not of the sort we have been discussing. Rather, the term appears to be restricted in its use largely to works of contemporary popular fiction, and, at that, to a rather select body of such works. Although, for instance, Happy Days and All in the Family were immensely popular television shows, and although each raises many of the issues at the heart of canon, it would be at least a little odd to refer to the Happy Days canon or the All in the Family canon.[46] And although certain genres seem to predominate canon discussion, genre does not appear to offer any clear conditions, necessary or sufficient, in this regard. Indeed, there would seem to be no apparent rule for which works will garner considerations of canon except that talk of canon normally arises within a well-developed fan community, and often where there is a lively tradition of fan fiction.

Still, for such works where canon does come into play, it would seem to do a great deal of critical work. Where a fictional story stretches across individual works, be they novels, video games, episodes of a TV show, or some combination of these, what is true of that story or the world it describes may depend very much on canon. This will include the rules of that world, in turn determining any number of particular fictional truths. In telling us which works are stitched together, and which are excluded, canon adds a complicating factor for identity conditions for fictional characters, further grounding ontology in artistic practice. And although its work seems to logically precede that of interpretation, insofar as the first task of interpretation lies in identifying the thing being interpreted, canon can play a central role here, too. Moreover, insofar as canon can change, so too it seems can the domain of reasonable interpretations of works within that canon. In a very real sense, then, the malleability of canon can change what is true of a work long after that work is completed by its author. So, although the domain in which it operates is contingently small, canon is part of an artistic practice that widely impacts philosophical issues in fiction and must be accounted for. So, certainly, philosophers have some unique reasons to care about canon.

Canon matters to writers, filmmakers, and other creators because it is often important to those creators to control the nature and fates of the characters and worlds they create. Controlling canon is no small part of this. Canon matters to fans to the degree that they are invested in these characters and stories. And make no mistake: fans care very much about canon. It is in the interests of copyright owners, who may or may not be the creators, that fans care about canon and associate it with the copyright owner's authority. Doing so will keep them coming back for more official content and otherwise shunning what is not. Insofar as philosophers are interested in the relationship between artist and audience, and particularly in the power dynamics at play in this relationship, they have this reason to care about canon as well.[47]


Craig Derksen

Craig Derksen is Assistant Professor of Philosophy at California State University East Bay. He is very interested in the intersection of popular culture and epistemology.


Darren Hudson Hick

Darren Hudson Hick is the author of Introducing Aesthetics and the Philosophy of Art (2012, 2017) and Artistic License: The Philosophical Problems of Copyright and Appropriation (2017), and co-editor (with Reinold Schmücker) of The Aesthetics and Ethics of Copying (2016).

Published on March 27, 2018.


[1] David Lewis, "Truth in Fiction," American Philosophical Quarterly, 15 (1978), 37–46;  "Postscripts to 'Truth in Fiction'," in Philosophical Papers, Vol. I (Oxford: OUP, 1983), 286–280.

[2] Lewis’s central example is Sherlock Holmes.

[3] Lewis, "Truth in Fiction," 37-38.

[4] Ibid., 39-40.

[5] Ibid., 43. This has all been a rather quick sketch of Lewis’s analysis, which delves into much greater technical detail than we are outlining here. We have omitted, for instance, the parts of Lewis’s theory that allow him to declare it false that Holmes, for all his two-handedness, has a third eye in the middle of his forehead, which simply goes unmentioned in Doyle’s stories. Such detail is unnecessary for our purposes here, and for the analysis of such statements as "Sherlock Holmes has hands" and "Sherlock Holmes is left-handed," so we leave it aside.

[6] Although some accept Lewis’s theory as being on the right track (see, e.g. Gregory Currie, "What Is Fiction?," Journal of Aesthetics and Art Criticism, 43 (1985), 385–392; ref. p. 391), he is certainly not without his detractors (see, Peter Lamarque, Fictional Points of View (Ithaca: Cornell University Press, 1996), pp. 55–70). Nevertheless, we believe that the problem we present for Lewis’s theory raises similar problems for competing or refined versions of his approach.

[7] Lewis, "Truth in Fiction," 45.

[8] See, Kristina Busse and Karen Hellekson, "Introduction: Work in Progress," in Fan Fiction and Fan Communities in the Age of the Internet (Jefferson, NC: McFarland & Co., 2006), pp. 5–32; ref. pp. 9–10; Eckart Voigts-Virchow, "Pride and Promiscuity and Zombies," in Adaptation and Cultural Appropriation: Literature, Film, and the Arts, eds. P. Nicklas and O. Lindner (Berlin: de Gruyter, 2012), pp. 34–56; ref. p. 42.

[9] See, Peter Richard, "Completing the Canon," Sherlock Holmes Journal, 6 (1962), 10–14; Peter Richard, "Completing the Canon," Sherlock Holmes Journal, 6 (1963), 76–81; Peter Richard, "Completing the Canon—A Postscript," Sherlock Holmes Journal, 6 (1964), 114–115; Peter Haining, "Introduction," in Arthur Conan Doyle, The Final Adventures of Sherlock Holmes (Berkeley, CA: Apocryphile Press, 2001), pp. 11–15.

[10] Reprinted in R.A. Knox, "Studies in the Literature of Sherlock Holmes," Blackfriars, 1 (1920), 154–172.

[11] Knox never actually uses the word 'canon.'

[12] Knox, "Studies in the Literature of Sherlock Holmes," 157.

[13] Kilian McDonnell, "Canon and Koinonia/communion: The Formation of the Canon as an Ecclesiological Process," Gregorianum, 79 (1998), 29–54; ref. 33–34.

[14] See, Lee M. McDonald, "Identifying Scripture and Canon in the Early Church: The Criteria Question," in The Canon Debate, eds. Lee M. McDonald and James Sanders (Peabody MA: Hendrickson, 2002), pp. 416–439.

[15] See Darren Hudson Hick, "Authorship, Co-Authorship, and Multiple Authorship," Journal of Aesthetics and Art Criticism, 72 (2014), 147–156.

[16] Amie Thomasson, Fiction and Metaphysics (Cambridge: Cambridge University Press, 1999), p. 67.

[17] Ibid., pp. 68–69.

[18] Ibid., p. 67.

[19] Nintendo also states that the Link and Zelda identified in the chronology might not be the same Link and Zelda throughout but might be a series of descendants. Patrick Thorpe, The Legend of Zelda: Hyrule Historia, English Edition (Milwaukee, OR: Dark Horse Books, 2013), p. 68.

[20] Roger Ebert, "'Highlander 2' Plummets to the Depths of Absurdity," Chicago Sun-Times, (1 November 1991), p. 45.

[21] Highlander: Endgame, dir. Douglas Aarniokoski, (2000), and Highlander: The Source, dir. Brett Leonard, (2007).

[22] Halloween H20, dir. Steve Miner (1998), and Halloween: Resurrection, dir. Rick Rosenthal (2002).

[23] Halloween III: Season of the Witch (dir. Tommy Lee Wallace, 1982), which was never considered part of the canon; Halloween 4: The Return of Michael Myers (dir. Dwight H. Little, 1988); Halloween 5: The Revenge of Michael Myers (dir. Dominique Othenin-Girard, 1989); and Halloween: The Curse of Michael Myers (dir. Joe Chappelle, 1995). Only the original Halloween (dir. John Carpenter, 1978) and Halloween II, (dir. Rick Rosenthal, 1981) remain in-canon.

[24] Rocky Balboa (dir. Sylvester Stallone, 2006) ignores Rocky V (dir. John G. Avildsen, 1990) though it doesn’t flatly say anything to contradict it. Creed (dir. Ryan Coogler, 2015) explicitly refers to events from each of I-IV, and is consistent with Rocky Balboa, but there is no hint of Rocky V. Given the narrative parallels between Creed and Rocky V, that the latter fits into continuity but was simply not mentioned would be strange.

[25] The Exorcist III, dir. William Peter Blatty, 1990, and Exorcist II: The Heretic, dir. John Boorman, 1977; Exorcist: The Beginning, dir. Renny Harlin, (2004); and Dominion: Prequel to the Exorcist, dir. Paul Schrader, (2005).

[26] Contrary to recent claims by James O. Young, then, the character of Sherlock Holmes does not evolve with new imaginings of him, as in Guy Ritchie’s films starring Robert Downey, Jr., or Bill Condon’s Mr. Holmes (2015). These are different Holmeses. See James O. Young, "Appropriating Fictional Characters," in The Aesthetics and Ethics of Copying, eds. Darren Hudson Hick and Reinold Schmücker  (London: Bloomsbury, 2016), pp. 153–172.

[27] As would countless novels, comic books, video games, and other officially-licensed fictional works, collectively referred to as the "expanded universe" or "Star Wars Legends," bracketed off from the core canon by Disney in 2012. See http://www.starwars.com/news/the-legendary-star-wars-expanded-universe-turns-a-new-page, accessed February 19, 2018.

[28] See Darren Hudson Hick and Craig Derksen, "Righteous Art Anger," Journal of Aesthetics and Art Criticism, 70 (2012), 373–382.

[29] This has occurred several times in the superhero universe of DC Comics, a case that would require much more space to spell out than we have available here, but see Henry John Pratt, "Why Serials Are Killer," Journal of Aesthetics and Art Criticism, 71 (2013), 266–270; ref. 268

[30] Salinger v. Colting, 641 F. Supp. 2d 250 (S.D.N.Y. 2009).

[31] Quoted in Elisabeth S. Aultman, "Authorship Atomized: Modeling Ownership in Participatory Media Productions," Hastings Communications and Entertainment Law Journal, 36 (2014), 383–405; ref. 388

[32] Strictly speaking, there is no international copyright law. Each country has its own copyright law but these are harmonized according to international treaties to which nearly every nation is a member.

[33] Actually, there is ample extant academic discussion of the Arthurian canon, and more limited discussion of the Robin Hood canon, though in each case the term 'canon' is used with maximal inclusivity, being effectively synonymous with "stories about the Knights of the Round Table" or "stories about the Merry Men," respectively.

[34] Nathaniel T. Noda, "Perpetuating Cultures: What Fan-Based Activities Can Teach Use about Intangible Cultural Property," Creighton Law Review, 44 (2011), 429–461; ref. 453.

[35] It was only in 2014 that the U.S. Supreme Court declined to take up an appeal ruling that the character of Sherlock Holmes had outlived its copyright and fallen into the public domain, thus officially ending any legal control over the character.

[36] Roy T. Cook contends that "canonicity practices involve complex negotiations between producers and consumers." See "Canonicity and Normativity in Massive, Serialized, Collaborative Fiction," Journal of Aesthetics and Art Criticism, 71 (2013), 271–276; ref. 273). Cook points as evidence for this claim at Marvel Comics’ historical practice of awarding "No-Prizes" for readers who are able to offer explanations, however convoluted, for errors in published comics. However: (1) The No-Prize has largely drifted out of practice; (2) No-Prizes were never a matter of negotiation between producers and consumers, but were awarded at the publisher’s whim; and (3) Cook offers no evidence of No-Prize solutions being treated as canon.

[37] Henry Jenkins III, "Star Trek Rerun, Reread, Rewritten: Fan Writing as Textual Poaching," Critical Studies in Mass Communication, 5 (1988), 85–107; ref. 93.

[38] Elisabeth S. Aultman, "Authorship Atomized: Modeling Ownership in Participatory Media Productions," Hastings Communications and Entertainment Law Journal, 36 (2014), 383–405; ref. 388.

[39] Though, as it happens, Aultman’s hypothetical still conflates copyright and canon authority, rather than separating the two phenomena.

[40] That Wolverine’s claws were bionic implants was first established in Volume 1, Issue #15 of the Official Handbook of the Marvel Universe. This was later confirmed in-story, and still later retconned away.

[41] Jeff Christiansen, Official Handbook of the Marvel Universe A-Z, Vol. 2 (New York: Marvel, 2008), np. This rule has produced a complex litany of subordinate rules and rolling in-story retcons. There are actually several distinct chronologies in the Marvel Universe, with the central storyline, where the sliding timescale rule applies, taking place on what is called Earth-616. The Marvel cinematic universe takes place in a parallel world: Earth-199999.

[42] See, harrypotter.wikia.com/wiki/Canon, accessed February 19, 2018; Erin McCarthy, "15 Post-Harry Potter Revelations from Pottermore," Mental Floss (2015), http://mentalfloss.com/article/59817/15-post-harry-potter-revelations-pottermore, accessed February 19, 2018; Kat Miller, "'What Is Canon?' – Part 1: It’s All in J.K. Rowling’s Head," MuggleNet (2014), http://www.mugglenet.com/2014/08/what-is-canon-part-one-its-all-in-jk-rowlings-head/%20, accessed February 19, 2018; Keith Hawk, "'What Is Canon?' – Part 2: The Books or Not the Books – That Is the Question," MuggleNet (2014), http://www.mugglenet.com/2014/08/what-is-canon-part-2-the-books-or-not-the-books-that-is-the-question/, accessed February 19, 2018.

[43] Unusual, but certainly not unheard of. The terms 'canon' and 'canonical' can be used quite loosely in ordinary conversation.

[44] Disney has not made this claim. The thesis is, however, made by Amelia Hill in her article, "What Star Wars Is About: The Cyclical Struggle of Good and Evil," About.com, 4 January 2013, http://scifi.about.com/od/starwars/a/What-Star-Wars-Is-About-The-Cyclical-Struggle-Of-Good-And-Evil.htm, accessed February 19, 2018.

[45] A recent debate between Andrew McGonigal, Ross P. Cameron, and Ben Caplan, for instance, has focused on the question of whether fictional truths change as canon changes. Although our discussion here would certainly impact that debate, our focus has been on the foundational mechanics of canon itself. See Andrew McGonical, "Truth, Relativism, and Serial Fiction," British Journal of Aesthetics, 53 (2013), 165–179; Ross P. Cameron, "How to be a Nominalist and a Fictional Realist," in Art and Abstract Objects, ed. Christy Mag Uidhir (Oxford: OUP, 2013), pp. 179–196; and Ben Caplan, "Serial Fiction, Continued," British Journal of Aesthetics, 54 (2014), 65–76.

[46] Each of these shows takes place in a larger universe folding in a number of related works. Happy Days takes place in the same fictional world as Laverne and Shirley, Mork and Mindy, and a number of other sitcoms; similarly for All in the Family; and each raises continuity issues. For example, Chuck Cunningham, older brother to Richie and Joanie on Happy Days, was written out of continuity in later episodes, to the point where Mr. Cunningham refers to his "two kids."

[47] We are grateful to the anonymous reviewer for helpful comments.