There have been recurring attempts to define Science Fiction, most of them satisfactory only to the author. Here is my take on this chestnut, reprinted from the May 1974 issue of APA:NESFA.  Although I have preserved the original text I have added footnotes for afterthoughts.
In a recent issue of APA-L (L-470 to be exact) I casually tossed off the following definition of science fiction: Science Fiction is that branch of of Fantasy in which the existence of science is used as a mythic rationale. Like most of my APA contributions it was written without any thought before hand.  (Indeed certain of my critics hold that thought and my APA contributions are usually unrelated. ) Having written this gem of illiterate prose I stopped to read what I had written (It might have been interesting, you know) and was surprised to read the makings of a good legitimate definition of SF. 
There are two major nits that I will pick. First of all "existence of science" is not quite the phrase I want and, secondly, the phrase "mythic rationale" is a neologism. Before Id emolish the definition irretrievably, however, let us look at what it is actually saying.
The definition advances three theses: that SF is a branch of fantasy, that all fantasy requires a "mythic rationale" , and that the appeal to the "existence of science" is the characteristic mythic rationale of science fiction. The first theseis, that SF is a branch of fanstasy, is meaningless unless we also have a definition of fantasy. The second thesis, however, implies an operational definition of fantasy.
The problem in defining fantasy is to distinguish it from fiction in general.  If fiction is a story that doesn't happen to be true  then what is the special character of fantasy? It is characteristic of general fiction that is tied pretty close to reality - allowing, of course, for the distorted perception of reality of the author. Fantasy, on the other hand, always contains an essential element of unreality - not of matter-of-factness, but of world view.  Dracula is not fantasy just because there are no vampires - it is fantasy because vampires cannot be; for there to be vampires in the classical sense the world must be essentially different in nature than we understand it to be.  This is, of course, a relativistic conception of fantasy, but I think that it is fair. To Homer the Odyssey and Iliad were simple narration, to us they are fantasy. 
This conception of fantasy does not require a "mythic rationale". Indeed it is possible to point to examples of stories which would call fantasy which do not. A good example is the fiction of Borges, which starts out apparently rooted in reality and then quickly wanders off in the most inexplicable directions.  Another kind of example is the what if essay or story. A good example is If Burgoyne Had Won At Saratoga.  Such a story accepts the standard world view - it makes use of no fantastic elements except the initial hypothesis - but the secondary universe created is so far from our reality that we cannot accept it as being within our reality. Such a tale must therefore be classified as fantasy. However it, too, does not need a "mythic rationale". 
None-the-less most fantasy does need such a rationale. If fantasy is to be explicable (and it need not be) then there must be a basis for understanding the element of essential unreality that it contains. In the "what if" story we get it for free. We start with a permissable variation in reality which is followed by action following the ordinary rules of fiction. However we extend the story to point that the world portrayed is no longer "our" world.  In a story about vampires we start out with an element that is outside our ordinary reality.  If we were not already familiar with vampires and with a conception of reality in which vampires made sense then the story would be inexplicable. Vampires and all of the rest of the paraphernalia of traditional fantasy are all familiar and explicable.
The reason that they are is that they are all creatures of a world view - the world of mythology and magic. It is not the particular mythology that is important - our imagination suffers no strain when hobbits and orcs are added to the list. Rather it is the entire world view of animism and magic that makes "sense" out of traditional fantasy.
What then of science fiction and "the existence of science as a mythic rationale"? The contention evidently is that science can be used ainstead of traditional magic and mythology as a basis for fantasy and that when this is done the result is science fiction. At first sight this seems to involve a paradox.  Science is, after all, part of our world view - part of our reality. But that science in science fiction is not used that way - the science in SF is not really science; it is pseudo-science.  Nor are the stories about science and scientists as they really are.
What is done is that science is used to justify the essential elements of unreality that is characteristic of fantasy. Naturally enough science cannot justify them. Instead the appeal is made to the existence of science as a rationale. (Now there is a cloudy sentence if I ever heard one.) In its simplist form this is done by a straight replacement of magic by science (Let me tell you a tale of a marvelous sorceror.... Now let me tell you a tale about a marvelous invention....) Much early science fiction had just this characteristic.
One of the common features of traditional fantasy is the creation of a place for the story to happen - a secondary universe , a place where the laws of magic are allowed to operate. This invention is all too necessary; it is hard to believe in trolls, wizards, an dragons in Orange County. In SF the future plays a similar role. Not only does the future have the great advantage of being not-here, but it is a place where the powers of science and technology will be greatly extended. The future is the Middle Earth of science fiction.
In essence I believe the definition is accurate and complete.  It explains why stories about science and scientists are not, per se, science fiction. It accounts for the relationship of fantasy and science fiction and provides a means for telling them apart - a test that seems rreasonably adequate. It explains why the future is so associated with science fiction but not insist that science fiction must be associated with the future. The only real complaint I have with it as a definition is that the wording is not sharp.
If all of this makes sense (which I believe it does) then the urge upon the part of some critics and authors to rename SF as speculative fiction is misguided. Speculative fiction is another branch of fantasy. It is, or can be, closely related to science fiction but it is a different breed of animal. That, however, is another story.... 
The above essay is overly long and is also incomplete. It does not come to terms with the neologism "mythic rationale". The idea, however, is simple. Prior to the enlightenment there was no real conception of the future as being different other than religious visions. With the industrial revolution came the idea of progress - the future would be like the present, only there would be more and it would be better. There was no concept of a fundamental transformation of life, other than the Marxist political transformation.
The twentieth century brought the notion of science and technology as a transforming principle, one that would induce radical changes in the ways that people lived and thought. This seems obvious now but it was not obvious in the mainstream culture 50 and 60 years ago. Science Fiction was a literary response to this vision of transformation.
Thus we have the vision of the future as essentially different from the here and now. It is this distancing from the mundane that allies SF and fantasy. For a coherent body of literature, however, there must be a common articulation of the basis for the distancing. This articulation is the rationale. In the case of fantasy it is magic and mythology. In the case of SF it is the vision of science based transformation.
In each case the literature is rich in myths. This is obvious in fantasy. It is less obvious in SF. However the "science" in SF is mythic; it is not and is not expected to be "real science". The mythic science in SF is explained by and justified by the appeal to the vision of transformation. Science and technology themselves are expected to change and be transformed.
 APA:NESFA is an APA (Amateur Press Association) magazine published
by the members of NESFA (New England Science Fiction Association).
 Things haven't changed much in twenty years.
 Critics haven't changed much either.
 Within this paper SF stands for Science Fiction. It is not short for "Speculative Fiction". It definitely does not stand for that ugly expression "Sci-fi".
 Myths and mythmaking were big when this was written. They're out of fashion now. However the old still works.
 One can take the view that "mainstream fiction" is a special category of fiction, Alexi Panshin, in SF in Dimension, distinguishes between mimetic fiction (an imitation of immediately observable life) and fiction in general.
 The notion of fiction as a story is problematic as is the notion of a true story. None-the-less the notions are serviceable even if the philosophers and literary theorists draw long faces and mumble and scratch themselves. One also has to distinguish between the artist who creates fiction that is not a story and the inept beginning writer who creates fiction that is not a story.
 Remembering always that fiction has to make sense; reality doesn't.
 Unreality here should be taken in the broadest sense as a displacement from the structured "here and now" waking world. The historic novel is not displaced from the structure of "here and now" because history is part of the structure. Reality is sometimes established by convention for the purposes of fiction. Consider, for example, the Western.
 This sentence is currently in the shop for repairs on its faulty transmission.
 This is mere twaddle. The strait jacket of mimetic fiction had not been invented when Homer wrote.
 The literati, aka the slime gods of academia, have discovered Borges, et al, and have invented the charming term "magic realism" to account for this form that does not fit within the established canon. Having long ago trashed fantasy as not being worthy of the mantle of James and Hardy they needs must smuggle it in via the back door.
 Nominally non-fiction, this work is set in a world in which Burgoyne did win at Saratoga. (It is my understanding that American History is no longer taught so the title and the concept lack force today.) Much of the work is concerned with the question of what history would have been if Burgoyne had lost - this should be our world but it isn't.
 In retrospect I think that there is a real problem with this division of fiction into mimetic fiction and fantasy.
 Most fiction is not set in "our" world; it is set in a genre world dictated by literary conventions.
 At least I hope it is.
 In modern genre fantasy traditional animism and magic have been replaced by one of two avatars, dungeons and dragons, and new age goo.
 Paradoxes are grist for the mill for SF. Go read The man who folded himself.
 Pseudo-science is a high standard which much SF falls short of.
 No argument here.
 Marvelous indeed. The names of the old pulp SF magazines tell the story - Astonishing Stories, Amazing Stories, Thrilling Wonder Stories, Startling Stories, Astounding Science Fiction. A major weak point of this essay is that it never comes to terms with "a sense of wonder". This sense of wonder, this touch of the marvelous, has been a part of story telling throughout the ages. It is only in modern times in realistic mimetic fiction that the sense of wonder has been expunged and erased.
 Tolkien has much to answer for.
 Not these days.
 Cute, Harter, cute.
 How about the tooth fairy and Santa Claus?
 "Not sharp" is a generous appraisal.
 The idea of combining science fiction and fantasy under the rubric of "speculative fiction" is, in my opinion, a serious misjudgement because it misses the essential common feature of both. Speculative will serve (albeit with some strain) for science fiction. It does not serve for fantasy. The common element is the displacement from the "here and now" and the subliminal emotion of awe.
This page was last updated September 30, 1997.