This arises from a discussion over on liviapenn's journal and continued on alixtii's. In the interests of full disclosure, I shall reveal that it is based on a theory from cognitive psychology called Prototype Theory, which argues that most categories are non-essentialist (ie. there is not a list of necessary and sufficient characteristics that will separate all and only members of the category from everything else in the world). Categories have fuzzy boundaries (ie. it’s not clear where some things stop being members of the category and start becoming something else) and are organised around “best examples” (known as prototypes, hence the name of the theory). The basic idea is thus that central members of a category exhibit a set of characteristics that everyone agrees are typical for a member of that category (so birds, for instance, have feathers, fly, sing, lay eggs, and are small enough to fit in your hand), whereas less central category members lack some of these characteristics (penguins and emus don’t fly, emus and turkeys don’t fit in the palm of your hand, ducks don’t sing – but they’re all still birds). In fact, less central members may have no characteristics in common with each other at all, but will still share some characteristics with prototypical members. Prototype theory arises from Wittgenstein’s theory of categorisation as a matter of “family resemblances” as alluded to by alixtii, in which members of a family as a whole have physical features in common, but individual members of the family may not. So Uncle Fred has blue eyes, blond hair and pigeon toes; I have pigeon toes and red hair and green eyes, and my sister has red hair, green eyes and perfect feet. She shares no single identifying characteristic with Uncle Fred, but both of them share characteristics with me, and within the family as a whole both red and blond hair, green and blue eyes, and pigeon toes are common, but the distribution within individuals differs). One of the linguistic tests for whether something is a central category member or more of a marginal member is how far it is substitutable for the superordinate term. Thus in the sentence “Twenty or so birds twitter on the telegraph poles outside my window every morning” the words “robins, thrushes, sparrows” can be substitued for “birds”, but “duck, turkey, eagle, penguin” can’t (or at least not without causing surprise). And so it is with fanfic. I would argue that there is a list of prototypical characteristics that central members of the category “fanfic” display, such that everyone would recognise them as fanfic. And I suspect the “conceptual analysis” discussion is more about people proposing central characteristics than saying this particular characteristic alone is sufficient and necessary to define fanfic. So what might these central characteristics be? I propose the following list, which is emphatically not given in order of significance ( and always bearing in mind that less central examples of “fanfic” may exhibit no more than one of these characteristics):


1. It is not written for publication
2. It is derivative
3. It is about a media product
4. It is a written text
5. It explores emotional/sexual relationships in greater depth than in the original source
6. It hits a kink
7. It utilises certain tropes (hurt/comfort, cavefic, aliens made us do it, soulmates)
8. It utilises certain stylistic features (this one is bound to be controversial, but it’s not difficult to come up with a list of classic fanfic stylistic features, such as the use of epithets, which explains why it’s possible to say that The Da Vinci Code “reads like fanfic”)
9. It is written by fans for other fans (if it is drawerfic, the imaginary readership is still fans of the source and not general readers)
10. It is character-centric



If we look at the outliers, we see that R&G [ETA: Rosencrantz and Guildenstern Are Dead]is a derivative written text about a media product while anthropomorfics are not written for publication and utilise certain recognisably fanifccish tropes (actually, it would be possible to argue that anthropomorfics are pastiches of fanfic in general, or that their source text is fanfic as a whole, but I digress). Note that there is an overlap with things like fanvids, which brush the very edges of the category (they’re not written, but they have a number of other characteristics in common with fanfic), which leads us into the theoretical waters of semantics. The salience of the central characteristics is affected by what other categories you are contrasting things with – “written” becomes most salient when contrasting fanvids with fanfic, “not for publication” when considering media tie-ins, “by fans for fans” when considering R&G. In fact, as sallymn pointed out, any film or TV (or radio) adaptation is derivative and therefore doesn’t contrast with fanfic on this dimension; and the same is true of a remake of a film or a TV series. If “derivative” is taken as the sole defining criterion of fanfic, then all these things have to be included; only by considering other typical features of fanfic can we consider the ways in which such things are not like fanfic.