Fish had said, while discussing an attempt by Richard Dawkins to characterize the distinction between scientific trust and acceptance of religious pronouncement:
It was at this point that Dawkins said something amazing, although neither he nor anyone else picked up on it. He said: in the arena of science you can invoke Professor So-and-So’s study published in 2008, "you can actually cite chapter and verse."
With this proverbial phrase, Dawkins unwittingly (I assume) attached himself to the centuries-old practice of citing biblical verses in support of a position on any number of matters, including, but not limited to, diet, animal husbandry, agricultural policy, family governance, political governance, commercial activities and the conduct of war. Intellectual responsibility for such matters has passed in the modern era from the Bible to academic departments bearing the names of my enumerated topics. We still cite chapter and verse — we still operate on trust — but the scripture has changed (at least in this country) and is now identified with the most up-to-date research conducted by credentialed and secular investigators.
To this Casey replies:
Really slowly: the list of items Fish mentions here (in bold) are prescriptions based on divine commands. The chapter and verse Dawkins refers to are prescriptions based on arguments. They're just reported second hand.
Those things are hugely different.
And this is all intended to show why Casey is holding up Fish's argument as an example of a straw man.
If you actually read Fish's claim in context, it doesn't take much to find out the problems with Casey's argument. They are three.
(1) Fish notes that Dawkins and Pinker would make just this response (or something very similar to it), in the very next paragraph. Casey's response can only be a reason for taking Fish to be a straw man if Fish isn't taking into account the point that the response raises. But Fish does. One gets the feeling that Casey simply stopped reading right here, because Fish's raising of precisely this point in response is quite clear and obvious and, as I said, is the entire subject of the very next paragraph.
(2) Casey clearly has missed the point of Fish's argument, or he wouldn't have responded in this way. That "those things are hugely different" merely begs the question against Fish, because Fish's argument is simply that there is an important way in which they are not. That is, Casey has responded to Fish's argument by treating Fish's conclusion as if it were an assumption. Both of these problems, (1) and (2), tie in with the third point.
(3) Casey seems to have missed the actual structure of Fish's argument. Fish's points here are obviously (I) that it is ironic that Dawkins would use the language of religious faith to characterize the superiority of scientific rationality over religious faith; and (II) that the use of such language shows that there is a common ground on which the two sides can be compared, namely, the question of rational trust. This is all simply a set up, establishing background (context, terms, reason for approaching the argument this way), for Fish's actual argument, which is that the kind of response Casey makes here is circular. The next paragraph gives the 'hugely different' response to the way Fish has described the issue, the paragraph after that gives Fish's reason for rejecting this response, which then leads to the main point of the post.
So Casey's argument that Fish is strawmanning is a strawmanning argument: his characterization of the argument is so off that his knocking it down actually doesn't affect Fish's argument at all, and the mischaracterization is egregious, since it requires claiming that Fish doesn't take into account a point that Fish immediately goes on to take into account. Fish's argument is not hugely well-developed (although it doesn't put itself forward as rigorous and thorough), but it certainly doesn't have the problem Casey suggests.
I actually think a lot of accusations of straw man fallacy run the danger of being examples of the fallacy themselves. It is interesting, though. Straw man fallacy is a very slippery category of fallacy. Many informal fallacies actually admit of a fairly straightforward characterization in terms of the ends of the argument, but straw man is more elusive. You aren't committing a straw man by simplifying your opponent's argument; nor are you even committing it if you oversimplify it. If I just oversimplify an argument, I am not thereby committing the straw man fallacy -- for instance, people often oversimplify their own arguments, and it is odd to think of anyone strawmanning themselves (straw man is a fallacious attempt at refutation), and I might oversimplify my opponent's argument but not in any way that's relevant to the point at hand. The basic idea of the straw man is easy to grasp -- one reason people appeal to it so often is that it has a very vivid image attached. Instead of knocking down the real man, I construct a straw man that looks roughly like him and knock that down as if it were the real man; but, of course, it isn't the real man, and in knocking down the straw man I haven't touched the real man. If it were that obvious when talking about arguments, there wouldn't be a problem, but in the case of arguments we are actually faced with continual questions about how much oversimplification is required to make an objection simply irrelevant to the actual argument. Lots of objections that are not obviously relevant to an argument may nonetheless be relevant; and a toy-model version of an argument may make it easier to see what is wrong with the original argument. (I occasionally have complaints against analytic philosophers and their use of toy versions of the arguments they are examining, but even I would never say that analytic philosophy consists of almost nothing but the straw man fallacy. Sometimes you do learn useful things from the really simplistic toy version.) There are still obvious cases -- this one for instance, which criticizes an argument for not considering a point it explicitly and obviously considers -- so the straw man fallacy does seem to latch on to something. But characterizing it is certainly tricky.
UPDATE: John Casey has clarifies in the comments that he didn't intend the post to sound like he was claiming that Fish was strawmanning, but was trying to make a different point; he's done some revising to avoid the misleading impression. That's good, of course -- although in a weird way I'm disappointed to lose an example of straw man straw man -- and thanks to John for clarifying.