There's a minor tradition within the science fiction community of using mimetic and mimesis to mean the opposite of the fantastic. The oldest such uses that I could find (with a quick and profoundly less than exhaustive search) date to the early 1970s, and the casual employment of the term in those contexts makes me suspect that it has a longer history within the SF world as a way to point toward what more generally gets called within that community "mainstream" or (less frequently) "mundane" fiction.
All of these terms are problematic for various reasons, but what struck me this time about the words mimetic and mimesis was how their meaning in this context relates to and in some ways contradicts a few others I can think of -- the classical idea of mimesis as "imitation", particularly "imitation of nature"; Erich Auerbach's influential mid-twentieth century study Mimesis: The Representation of Reality in Western Literature; and, particularly, Ronald Sukenick's use of "mimetic fiction" in various essays collected in one of the many books I currently have out from the library, In Form: Digressions on the Act of Fiction.
Despite its history within the SF world, I'm not convinced that "mimetic fiction" is the best term for that thing for which there is, admittedly, no perfect term. Using mimesis to describe this thing, though, seems even less perfect to me than most of the other terms, because I'm more persuaded by Sukenick's use of it to mean fiction that tries to hide its illusions. In an essay on Gerald Graff, Sukenick makes this wonderfully efficient statement: "Mimetic fiction depends on the suspension of disbelief; nonmimetic fiction does not."
Almost all genre fiction, of course, depends on the suspension of disbelief, so if we accept Sukenick's definition, then the vast majority of SF is, in fact, mimetic.
This use of the term makes sense to me because it does two things. First, it does not deviate significantly from how the term has traditionally been used -- mimetic fiction in this sense seeks to give the reader a feeling, at least while reading the text, that there is a fundamental reality to the world conjured by the words. Sukenick writes:
The key idea is verisimilitude: one can make an image of the real thing which, though not real, is such a persuasive likeness that it can represent our control over reality. This is the voodoo at the heart of mimetic theory that helps account for its tenacity. Though such schizoid illusions are fostered by concepts of imitation, one cannot have control "over" that of which one is a part, or even formulate it completely -- one can only participate more deeply in it.It doesn't seem to me that we have to accept Sukenick's preference for fiction that shuns verisimilitude in order to see that his distinction between mimetic and non-mimetic fiction is a useful one, which is the second reason I find it persuasive -- it describes in a more coherent and less problematic way than other terms I can think of what feels to me like a fundamental difference between types of writing. It illuminates some of what is different between, for instance, Tender Buttons and Dubliners -- but also some of what is similar between Dubliners and Tarzan of the Apes.
There is a good argument, too, that using mimesis as a way to distinguish SF from non-SF hides (or at least obscures) some of what SF seeks to do -- we're back at the suspension of disbelief. The creation of believable worlds from at least ostensibly unbelievable material. Such fiction relies upon verisimilitude, which is a point made by writers and critics for decades: that SF is realism on steroids. It seeks to create in the reader's mind what John Gardner called a "vivid, continuous dream". It may not be the representation of what we believe to be "reality", but it is an attempt to represent an imagined reality in a way that the reader does not reject as preposterous. This is worldbuilding. Philip K. Dick once offered advice on "How to Build a Universe That Doesn't Fall Apart Two Days Later" -- a nonmimetic writer might offer exactly the opposite advice.
Even the classical notion of mimesis as "imitation of nature" can apply to SF -- particularly to core science fiction, which deliberately tries to extrapolate from known science, and is often based on ideas of nature as we currently understand it. But even in something like Judith Berman's "Rather Cranky Post on Verisimilitude in Fantasy", we're still talking about the imitation of nature.
There is a need for a value-neutral term for non-SF. I'm just not convinced that "mimetic fiction" is what it should be.
Update (7/7/09): Hal Duncan has posted a thoughtful extension of some of these ideas in connection with some other discussions that have been going on around the intertubes.