As best as I can gather, Joshua's definition of science fiction is this, "Science fiction depends on scientific principles, or extrapolation of scientific principles. Aliens? Scientifically they are plausible, so they can exist." Now, I'm sure that he would argue that that isn't a formal definition, and I agree, so let me start off by saying I'm not going to nit pick it to death. I'll accept that implicitly he means things like, "plausible [at the time they are written]", or "plausible [broad speculation, as opposed to the narrow speculation of ordinary fiction which eshews the fantastic]", and he just left out a bunch of details covering special cases with the understanding that an intelligent and fair minded-reader would understand what he meant. Even so, even with various similar additions, the definition will not stand as an absolute definition of science fiction, and in fact I hope to show that ultimately it has no real meaning.
Likewise, as best as I can gather, Joshua's definition of fantasy is, "Fantasy, on the other hand, is defined by including elements that are flat-out impossible to explain. It's not about imagery, it's about including stuff that cannot be."
Now, first, it's worth noting that there is something very important we both argree on. The genera isn't defined by the imagery. You can make a pretty decent argument that it is (at least better than Joshua's argument IMO), but since neither of us are arguing for that, I don't have to refute it at this time. All I have to refute is the notion that the above definition can usefully distinguish science fiction from fantasy, or for that matter science fiction from anything.
The problem with the definition is that it depends on the reader's understanding of what is plausible. Joshua wrongly assumes that everyone (or even anyone) will agree on what is reasonable speculation based on the laws of science, and wrongly assumes that everyone (or even anyone) will agree on what is flat out impossible to explain. Thus, Joshua's definition depends on the interpretation, knowledge, biases, and beliefs of the particular reader.
Put simply, Joshua expects that every reader has the exact same degree of suspension of disbelief as he does, and so thinks that its possible to come to any reasonable agreement on what is plausible or not. The fact of the matter is that what is actually classified or not classified as plausible depends not on what is actually scientifically plausible, but on the general "common sense" acceptance of the conventions of the genera. But 'common sense' understanding is by no means universially shared, and in fact the more scientific knowledge you actually have the more this common sense notion of what is scientifically plausible breaks down. In fact, the science fiction authors themselves are far more aware than the general public that they are writing stories that don't hinge on scientific plausibility at all, and have at all times thrown out scientific plausibility whenever it suited the needs of the story. So in actuality, whether or not something is considered science fiction has almost nothing to do with science and never has.
Not that I'm saying that many SF authors aren't deeply interested in science and its implications, or that many authors (the so called 'hard sci-fi' writers) don't try to be rigorous about the science, but for the most part 'hard sci-fi' is terribly unpopular even among fans of science fiction. Gregory Benford (to name one of the most prominent) is a name which doesn't necessarily ring a bell even among fans of science fiction.
To show you what I mean, take the example of 'Aliens'. Now, I am by profession a programmer that writes bio-informatics software in a genetics laboratory. I know a little bit about the genetic code, and its is to say the least enormously complicated. To me, the odds that life just spontaneously generated itself althroughout the galaxy is about as likely as monkey typing out the code for MS Windows, or a bowl of petunias and a whale suddenly popping into being in orbit around the planet. It's just so bloody unlikely that if it were to be the case, it would be a miracle - even in a universe as vast and complicated as our own. And my objections don't have to end there. The question arises that if our galaxy does have several hundred or thousand galactic civilizations in it, why haven't we heard from them? Why wasn't there radio traffic filling the heavens the first time we aimed an instrument at the sky? Why aren't they here already? Why weren't they colonizing Earth millions of years ago? You could of course answer something like, "Well, you see they communicate using this form of energy that we haven't discovered yet.", but to me whether true or not that is about as scientific of an answer as, "God did it." If you are going to start suggesting that something is possible because it depends on something that we haven't discovered yet, then anything and everything becomes possible. How does magic work? "Well it employs forces that we haven't discovered yet." It's a stock answer that answers everything, and at the same time answers nothing.
FTL travel? "Well it employs forces that we haven't discovered yet."
Wouldn't the ammount of energy required to move a ship between two star systems perclude trading in commodities? "Well, you see they have this means of generating nearly unlimited energy in a way that we just haven't discovered yet."
Wouldn't a power source that produced that much radiant energy melt the ship/lightsaber/blaster instantly if it leaked only .0001% of its energy as heat? How do they defy the 2nd law of thermodynamics? "Well, you see they have discovered this means of directing energy so efficiently that no waste heat is produced"
And so on and so forth. To which I respond...
Casting a fireball with a wave of the hand? "Well you see it employs forces that we haven't discovered yet."
And there is another problem. What someone considers real is pretty subjective to. Joshua seems to believe that everyone agrees on what is magic and what is real, when in fact nothing could be further from the case. If someone writes a story that features UFO abductions, I might consider it fantasy. Someone else might consider it science fiction. A third person might believe its a true to life drama in the vein of a Tom Clancy geo-political thriller. On the other hand, if someone writes a story that features miraculous heaing, someone might consider it a true life account because he has a theological explanation for the event, someone else might believe its a science fiction account because he has a paranormal explanation for the event, and someone else might consider it a fantasy because he has a magical explanation for the event and doesn't believe in magic.
And BTW, there are science fiction stories about people being awakened from cryogenic slumbers by far future faith healers who discovered psychic powers, but need the person to program a computer for them because its become a lost art so its not like I'm talking about mere theoretical classification needs. These are real issues. By Joshua's definition, there is absolutely no way to classfy that except for yourself, and what good is a definition that only works for yourself? Well, possibly a definition that works only for yourself is good for something, but even if it does a definition that works only for yourself doesn't match with the statement, "I'm advocating genre definitions that are based entirely on the text, and not on the reader's interpretation of it."
It gets even worse. Because whether or not a reader is willing to suspend disbelief on a particular issue and say, "Yeah, I'm willing to accept that as plausible." is an emotional responce. At what point does bending the rules of science suddenly get to you and you say, "You know, I was willing to accept using nukes to deflect an asteroid the size of Texas even though I know that really that's improbable and a couple 300kt nukes probably wouldn't dent a rock that size, but once I realized that they intended to do it from near earth orbit rather an a couple of AU's out where the angle of deflection just might be large enough, that was the straw that broke the camel's back for me." What was the difference between the two law breaking plot elements? Only that eventually the odds got too low to keep believing in it. One librarian comes along and shelves it in science fiction. The next comes along and shelves it in fantasy.
Joshua's inclination at this point in the past has been to shrug and say, "Well, the work must be 'space opera' then." First, that's a misuse of the term, but I'm not going to quibble. Secondly, the fact of the matter is that if we were to classify every science fiction book by whether it rigorously followed the science, we'd end up with nothing left in science fiction. We'd have to call Iain M. Banks, 'Look to Windward', a 'space opera', and David Brin's "Uplift Wars' a 'space opera', but if we call a works that seriousness a 'space opera' (especially works by people with as impressive of scientific backgrounds as Brin) then we probably are going to have to call everything in the genera a space opera, or at the least someone will be able to call everything a 'space opera' because there will be someone who can't suspend enough disbelief to believe that the science is in fact plausible.
So much for a definition that depended only on the text and not the reader.