(This is hardly an argument unique to Ferguson, so we can detach it from him, thank god...)

The argument about counterfactuals and causality is a relatively simple one. It comes down to: if you say that X mattered, you are implying that without X, things would have been different in some marked way. And that latter part is a counterfactual mode of reasoning.

So if I say, "Adolf Hitler only took power because Franz von Papen made him Chancellor," then I am saying, in short, that if Franz von Papen had not made Hitler Chancellor, then Hitler would not have been able to take power (at least the way he did).

Now this is not an exceptionally provocative statement except that historians have often claimed they don't like to engage in counterfactuals, and this simply shows that to some degree, counterfactual reasoning is the same as causal reasoning. There are obviously limits, of course; "counterfactual history" is usually characterized by people spinning out far broader conclusions than the above ("and thus, if von Papen wasn't around, then these things I am just more or less making up would have happened..."). And that is the stuff historians are wary about — it is one thing to say, "things would have been significantly different" but it is another thing to say "I have ideas about what those differences would be."

The positive side of acknowledging counterfactual reasoning does get used, even when not explicit, is that sometimes you can use the counterfactual framing as a way of elucidating what the actual modes of causality are. So I have tried to do this in an article I wrote on whether the atomic bomb could have been invented without Einstein — my argument there, in short, is that Einstein is actually far less important as a causal factor to the development of nuclear weapons than most people realize, and instead I trace out an alternate causality, and speculate about whether the absence of Einstein would have affected it much (I don't think it would have, but again, this is necessarily speculative to some degree). The goal here isn't to come up with some wholly-formed alternative world (which is the stuff of fiction, not history), but instead to help point to the real "pressure points" of the existing historical world we live in (to probe its "parameter sensitivity" to use a technical analogy — how much can you change before big differences would be expected in the outcome? Was Einstein a very sensitive parameter, or not?).

(For a nice, long discussion of these matters, see this review by Cass Sunstein.)