My last post on whether or not Maedhros threatening to kill Elrond and Elros was canonical has generated a lot of wonderfully thought-provoking comments. Not surprisingly, many of these have been about canon: what it is, how it is defined, and at what point to we pass from “canon” to “AU.” This is a matter to which I have given a lot of time and attention over my years in Tolkien fandom, so in wake of the discussion on Take Pity upon Him, I thought I’d put some of my more recent ideas down as I continue moving toward that (perhaps unattainable) goal of defining “canon.”
When I first started studying Tolkien’s works and writing stories based on them, I had this idea that, as I studied more, I’d move closer to being able to define canon definitively; that is, to produce a final and unequivocal judgment on how things really went down. Instead, I’ve found that the opposite has happened. Pandemonium remarked the same in the comments on Take Pity upon Him: “As I’ve examined JRRT’s work, canon becomes more and more nebulous to me.” Yet, at the same time, I feel better equipped now than I did four years ago to analyze what JRRT wrote in terms of “canon,” even if–at the end of my study–I don’t end up with any answers at all.
Sometimes, I think that we discuss canon and how to define it without really differentiating the ways that authors use JRRT’s writings to form judgments on their relative truth. This leads to arguments where something that is clear fact to one author, debatable to another, and false to a third, and all three fans are trying to prove each other wrong without considering whether they might all be right. I don’t think that canon can be so neatly summed up as “is” or “is not”; it occurs on a continuum, and different people will draw the line between “canon” and “not canon” in different places–even multiple places–along that continuum. I am going to attempt to summarize some common ways–complete with made-up terms!–that I think authors put together information from the texts to develop their definitions of canon.
Defining Canon
Canon. “Canon is synonymous with fact. It is not arguable. An author who violates this canon unintentionally has made a mistake; an author who violates this canon intentionally has written an AU.
Precious little from JRRT’s texts are canon by this definition. That which qualifies tends to be basic facts that would either be difficult/impossible to distort or lie about (such as the date of a major battle in which a literate culture particpated) or which are so frivolous that no one would logically have motivation to lie about them. Hair color, if definitively stated, is one such detail, perhaps ironically since this fandom is prone to fights over characters’ hair colors. But if Fëanor’s hair color is stated definitively to be black (The Silmarillion, “Of Fëanor,” §6), why would a loremaster or historian have reason to lie about this? Geographical details, dates, and physical descriptions all tend to fall into this category … in other words, mostly boring stuff.
Personal canon. Personal canon is an individual author’s appraisal of what parts of the texts are “fact” and employs any or all of the analyses discussed below and then some. For example, some authors have determined from reading the HoMe that JRRT’s final word on Gil-galad’s parentage puts Orodreth as his father. For these authors, Orodreth as Gil-galad’s father is personal canon; it is not an indisputable fact and so not simply canon, but it is a detail in the texts that these authors have analyzed and found to be true. For other authors, their personal canon is that Fingon was Gil-galad’s father.
It is important to point out that personal canon must come from the text (i.e., is not paracanonical or extracanonical) and is different from personal verse.
Personal verse. Personal verse is the sum total of all that an author believes to be true about the world in which she or he writes. It involves facts from the text (personal canon), as well as facts that the author develops based on and independent of the texts (paracanon, extracanon et al; see below).
For example, a personal verse might use the personal canon that Orodreth was Gil-galad’s father. It might also operate on the idea that Maedhros and Fingon were lovers and that confusion about Gil-galad’s paternity arose because historians associated with the House of Fingolfin were encouraged to conceal Fingon’s homosexuality and so distorted facts when they thought they could get away with it so that it appeared that Fingon had a wife and fathered a child.
Paracanon. In my original comment to Rhapsody about this, I called this extra-canonical. I’ve reassessed this term and think that paracanonical describes better what I mean, but to keep things as confusing as possible, I am using extracanonical elsewhere for something different. Paracanon is arrived at by putting together facts from one’s personal canon and drawing conclusions based on those facts. Likewise, paracanon cannot conflict with other facts in one’s personal canon.
The important aspect of paracanon is that the conclusions are fact-based. They are not merely whims or inventions. The author has analyzed a body of facts from the text and has, from this analysis, developed a personal canon. In putting those facts together, certain conclusions can sensibly be drawn. This is paracanon.
Naturally, for every dozen authors, you will end up with a dozen paracanons!

The Maedhros/Fingon pairing is a paracanon. Any fan of this pairing can tick off a dozen personal canon facts that makes this pairing, for them, a logical interpretation based on these facts. The pairing comes from putting those facts together and deciding that romantic involvement between the characters is the preferred conclusion to draw from those facts.
At the same time, other authors will put together personal canon facts to develop the paracanon that Maedhros and Fingon remained close (platonic) friends throughout their entire lives. Both paracanons are justified with JRRT’s texts and don’t involve any invention on the author’s part.
Some authors will necessarily “stretch” further than other authors in developing paracanons. However, the mechanism is the same.
Extracanon. Extracanon develops ideas outside of but in accordance with the texts. In other words, extracanonical facts in an author’s personal verse do not have any strong basis in the texts. Neither do they directly contradict the texts that an author uses in his or her personal canon.
Original characters are perhaps the best and most common example of extracanon. Their presence does not contradict the texts in most cases, but the texts don’t give us any information about them either.
Pandemonium’s The Apprentice is a good example of extracanon being used in this fashion. Sámaril is not a canon character. But neither does his existence as an apprentice in the Gwaith-i-Mirdain defy canon in any way.
Other extracanons place canon characters in settings other than what JRRT described. Erestor gets a lot of extracanonical treatment. He frequently ends up in Gondolin; in my By the Light of Roses, he ends up in Formenos. There is no canon support for either of these ideas. Neither does canon dispute them, however.
Making Fëanor a chronic nail-biter or Túrin’s favorite color black or Amarië the daughter of an important Vanyarin scribe are all extracanonical.
As with paracanon, different authors will have different comfort levels when it comes to how far they’re willing to go in inventing extracanonical details.
Pericanon. Pericanon analyzes and interprets the texts using concepts from psychology, mythology, sociology, science, and other “real world” disciplines. Because JRRT intended his stories to serve as a history or mythology for our world, and Arda corresponds with our solar system, then much of what we understand about our world can also be applied to Arda and, thus, becomes a sort of canon.
Authors using pericanon might use it to choose one text over another for their personal canons (such as using the more scientifically accurate ideas from Myths Transformed in describing how Arda operated outside a mythological framework) or add extracanonical details (such choosing to have Maedhros threaten to kill Elrond and Elros based on his psychological state at the time).
My assertion that homosexuality is canon is based on pericanon: If Elves and Men are human (in JRRT’s own words [Letter 153]), and homosexuality is normal behavior among humans, then lacking anything in the texts that makes an exception for Elves and Men, homosexuality would have occurred in their populations as well.
Pretty much everything Pandemonium writes uses pericanon to develop and explain not only the science of Arda but its cultures. My Another Man’s Cage uses pericanon in that I was often informed by psychology in how I developed the characters extracanonically.
I should note that pericanon uses our understanding of our world to enhance existing information from the texts, not to challenge or contradict them. To challenge the texts requires …
Historiocanon. Historiocanon is the process by which some authors challenge the texts and develop interpretations that do not take the texts at face value. Historiocanon justifies deviating from the texts where historiographical analysis causes concern about authorial bias or inaccuracy.
Pericanon can influence historiocanon when our understanding of how the world works calls us to question the accuracy of the texts. JRRT acknowledges this himself in Myths Transformed (HoMe XII) when he expresses doubt that readers would believe that scientifically sophisticated cultures (like the Eldar) would believe primitive and implausible cosmogonical myths.
Historiocanon is based on an understanding of Arda as our own solar system and, also, the JRRT’s texts as an ancient history/mythology of our own world and so subject to historical analysis. Historiocanon can hinge on the following (please note that this is an incomplete list):
- the narrator possesses bias (such as Pengolodh’s vilification of the Fëanorians in light of his service to Turgon, who was opposed to them)
- the narrator is relying on hearsay or could not possess accurate knowledge about the subject (such as Pengolodh writing about Fëanor’s death, which occurred before he was born, or about Lúthien’s plea to Mandos, during which none from Middle-earth were present)
- knowledge of how the world works makes the event as reported impossible (such as Maedhros hanging on Thangorodrim for fifty years)
Pandemonium’s Risk Assessment uses historiocanon to offer alternate explanations about lembas. My An Ordinary Woman uses historiocanon to argue that Lúthien’s exceptionality in, well, everything was more a case of hero worship and wishful thinking by her people than truth.
Pericanon and historiocanon are both, of course, personal canons as well: They require accepting Arda as our own solar system and a world subject to many of the same natural laws. Historiocanon also requires accepting as personal canon that the texts are historical or mythological accounts and can be analyzed using historiography. I think it makes sense, then, that these forms of canon will be the most controversial in terms of concept alone (not individual use) and won’t be used by everyone. However, they are valid ways to develop interpretations of canon.
Alternate universe. By definition, alternate universe (AU) requires the deliberate changing of a canon detail to affect the outcome of a story. Juno Magic’s Lothíriel is an AU because it adds a tenth walker to the Fellowship. My For What I Wait is AU because it is based on the premise that Fëanor outlived all of his children.
Both of these stories change canon facts. There were nine members of the Fellowship; it is hard to argue–though perhaps not impossible–that a tenth would have been completely overlooked by the many people who observed or were involved with the Fellowship. That Fëanor died shortly after the Battle-under-stars is another fact that would be extremely difficult to argue against. The AU aspects of both stories are not justifiable using any of the above-discussed canons. They are simply changes to the canon that the reader will have to accept and that are essential to the story.
It is important to note that AU cannot be justified by canon. Positing that Lúthien was less than perfect, as I do in “An Ordinary Woman,” is not AU because it makes sense from a historiocanonical perspective, which can be defended using Tolkien’s texts. Deciding that Erestor grew up in Gondolin is not AU because it does not counter a canon fact; it is extracanonical. Writing Maedhros and Fingon as lovers is not AU because it can be defended using evidence from the texts. However, I think that the term and label “AU” is misapplied as often as it is used correctly.
When Does “Canon” Become “AU”?
I am hardly the first to tackle this topic. Earlier this year, we had a discussion on the SWG Yahoo! group about how to define AU. This prompted a series of posts and discussions elsewhere (many of which I didn’t even know about until researching this post). I will link these discussions throughout my post, but it seems that they come to some of the same conclusions.
First of all, that the “AU” label is misused in the Tolkien fandom. I’ll discuss this further in a moment.
Second of all, that there is a strong desire, in discussions of canon, to move beyond the “is canon”/”is not canon” dichotomy and to recognize at least a third way to classify ideas used in fan fiction. Marta called this “extra-canonical” in her post On Canon and Fanfic, and this term (and the concept it defines) was echoed throughout the discussions following her post. So my own idea of a continuum between “canon” and “AU” is hardly original to me.
So why so many differentiations when Marta made good use of the single term “extra-canonical”? Mostly as a demonstration of how many different methods fans use to arrive at the extra-canonical (by Marta’s definition of the term) details that they use in their stories. I don’t expect the terms I’m using here to make it into popular usage. They’re awkward and hard to distinguish between for anyone who doesn’t make a regular habit (as I do) of thinking and writing about these things. In other words, for most people, they’re useless.
However, I think there is an important point to be made with them. As I defined each term, I often qualified that different authors would have different comfort levels with how far (or in what direction) they wanted to take various interpretations. Perhaps the most salient example is that of the paracanon about Maedhros and Fingon. Proposing that the texts suggest close friendship requires less stretching than suggesting that the characters were lovers, even though both interpretations utilize similar analyses. Yet I know that readers and authors will consider some details “canon” and others “AU,” even when the same methods were used to construct them. People are fond of lamenting that AU is hard to define. I don’t think that it is, if we recognize that accepting all of the above as legitimate analyses of Tolkien’s texts and understand that our willingness to accept (or not) an interpretation derived from them reflects more about how we see canon than the actual canonicity of the interpretation.
I also wonder if people’s comfort differs between the different ways of interpreting the texts that I’ve mentioned here. For example, maybe I’m not willing to stretch far in terms of paracanon. Maybe I like my interpretations of the texts to as innocent and obvious as possible. But maybe I’m willing to accept more in terms of extracanon: If you want to add all sorts of original characters and off-the-wall facts about the canon characters, then this doesn’t bother me. So Maedhros/Fingon feels wrong to me, but I’m okay with Fingon having, once upon a time, studied herb lore, lived with the Fëanorians in Tirion, and been engaged three times to three different women before the Darkening. Looking at the different ways that we shape our personal canon from the texts will, hopefully, aid me in approaching these questions in the future.
Which brings me to the other point about the misuse of the term “AU.” There is the popular complaint that some authors use the “AU” label to deflect any criticism about the wanton flouting of canon in their stories. Several people made this point in the posts I’ve linked here; Roh Wyn goes as far in Can(n)on Fodder to differentiate between canon deviations: non-canonical, where “some important detail has been altered, and this alteration affects all the downstream activities events or characters so that the entire story is different from canon”; and un-canonical, “stories that essentially break canon. … [T]hey don’t merely change a few canonical details. These fics change the basic premises of canon, so that the ultimate story bears little relation to the original.” My understanding of Roh Wyn’s uncanonical is that these are those stories that change details from the text because the author doesn’t know better (or doesn’t want to do the research to find out) or because the author simply likes the changed version better than the textual version but doesn’t want to think about how to make the work within the general canon framework Tolkien has established; for example (to borrow Roh Wyn’s example) because s/he wants Aragorn and Boromir to be twins but doesn’t want to have to do the work to make that plausible. So it just is–much in the way that Legolas has been married off to many teenaged unicorn-riding princesses–and the reader is expected to accept it without explanation or question.
Others bring up how “AU” is used as a defense against the so-called “canon police” or “canatics,” who are depicted as fans who hunt through stories looking for any detail that does not jive with their particular interpretation of the texts. In a comment on her rantastic Is AU a negative label?, Juno writes,
In my rant I didn’t discuss the validity of labels such as “canon” or “AU” as such. They definitely can have their uses. But they also pose problems. There are no fixed, exact rules about what is and what is not “AU” or “canon”. Actually, there IS no one canon, really; canon is not determined by physical laws or divine laws, canon is always the result of the interpretation of an individual and thus … fluid. Therefore, labels can be misleading. Especially in LOTR fandom, especially about new authors I’ve noticed the tendency to label what I would call “canon stories” as AU, simply because some kind self-appointed canon-police scared them and made them feel insecure about their stories.
In my original post on the SWG that started this whole discussion, I admit to doing just that. I am not alone in this either. But I’m also willing to admit that this comfortable deflection of attacks from canatics does a disservice to actual AU stories and the very valid approaches to the texts that I and others take in developing personal verses that give thoughtful treatment to Tolkien’s writings.
All of the terms I discuss above are valid ways of approaching and interpretting Tolkien’s texts, and none of them are AU. Yet I’m sure that many of us can think of examples of each where the author or her/his critics would suggest such a label: the Maedhros/Fingon pairing (or Celegorm/Aredhel, for that matter), a story told by an original character or heavily featuring original characters, a story that challenges the truth behind Laws and Customs among the Eldar. I think the temptation–when encountering a story that uses an interpretation unfavorable to us as readers–is to discount that story as “uncanonical” or to suggest that the author needs to label it as “AU” rather than giving thoughtful consideration to the means by which authors use facts from the texts to arrive at different interpretations or conclusions.
And this brings me full-circle back to Pandemonium’s comment about how the study of Tolkien’s texts makes recognizing a definitive “canon” more and more difficult. Personally, in all but a few instances, I’m ready to be done with the term “canon” for good. It’s misleading. It doesn’t exist in the form that we think it does, though it’s a nice idea–that with enough study and effort, we can devise a compendium of facts about Tolkien’s world that allow stories to be graded in terms of canonicity–like many of the fancies to which humankind has been prone over the millennia.
I doubt that one humble heretic like me will ever have such influence, though. In the meantime, though, if I can encourage even a few people to resist the temptation to jab pointy fingers and shriek, “AU!!” and, instead, stop and think and question how the author arrived at a particular conclusion, then I will consider my overwrought analysis a success.
Filed under: Fandom and Online Life, Tolkien by Dawn
19 Comments »