Sunday, February 20, 2005

On "More on Verisimilitude"

A response to my post on verisimilitude has been put up at Studi Galileiani. In it, Paul Newall (aka Hugo Holbling) argues that his point wasn't that theory-ladenness is a problem for the realist as such, and that he wasn't really discussing realism; which is fair enough, I suppose - it wouldn't be the first time I've gotten caught up in my own musings in the middle of a discussion. It should be said incidentally, that I don't really take the question-begging charge seriously here; my argument was not primarily to establish historical realism but to argue that the considerations brought against it need not worry the historical realist (but that said, Newall has missed the point of my argument, which I discuss below). Some brief comments on other issues, to clarify my own points on realism, which I don't think Hugo has quite got right:

Hugo says repeatedly that his real point about theory-ladenness was that there can be no historical facts independent of theoretical constructs. I got that, however; my claim was that this has no bearing on historical realism. He says, for instance:

In the latter respect, I explained that traces are not "left" but made. (This excepts archaeological and geological evidence, which have problems of another - although similar - kind.) We may insist that whatever the distortion caused in recording them, the traces nevertheless contain an element of "what happened"; but the problem of theory-ladenness is that there is no factual portion that can be demarcated from the theoretical. The question-begging arises when we realise that the observational content of these traces has been assumed in order to reverse the direction of influence. That is, there is little reason (and no argument) to suppose that if an event occurs and leaves traces, none of which have "factual" content, a proposition or account coherent with the traces is necessarily coherent with the event - unless we have already assumed that the traces correspond to the event, which is what is at issue.


But my point was that there is nothing threatening to realism in the claim that traces are made; that there is no factual portion that can be demarcated from the theoretical is not a particularly troublesome issue, because (1) the historian doesn't have to assume crudely that traces correspond to the event, only that they are traces of the past; and (2) it follows as a matter of logical necessity that if an event occurs and leaves traces (I have no notion what is meant here by the 'factual content' of traces) that a proposition or account coherent with the traces is coherent with the event: namely, it is coherent with the event insofar as it leaves traces. No correspondence is needed.

Hugo goes on to say there are problems with the coherence account of truth, but as I noted in my post, one doesn't need a coherence account of truth, only a coherence account of approximate truth. That there are many accounts coherent with the traces is irrelevant; that was precisely my point. The historical realist is not committed to saying there are not, and no problem arises for historical realism if these many coherent accounts are all approximately true in virtue of their coherence with the traces of the past. (Again, as I kept noting in my post, this does presuppose that they are genuinely traces of the past; more on this below.) Hugo says that we cannot say that these are all the same account in different forms; my comments explicitly denied that they were. My point is that one does not need 'the same account' in order to be realist about the accounts one has; or, at least, that it has not been shown that one needs to do so.

More seriously, however, Hugo says:

(More importantly, perhaps, and to mention realism for once, if we hold that there exists a world independently of our thoughts about it then propositions or accounts concerning it take their truth value from correspondence with that world, regardless of how well they cohere or fail to cohere with other propositions, accounts or traces. Kirkham summarises this objection by stating that "on a realist ontology it is hard to deny that coherence is neither a necessary nor sufficient condition for truth.")


But Kirkham is simply wrong. There is nothing about a realist ontology, as such, that implies anything about whether coherence is a necessary or sufficient condition for truth; and there are very clearly realist ontologies, Whewell's, for instance, on which coherence and correspondence mutually imply each other: nothing we look at involves a separation of theory and fact (or Idea and Fact, as Whewell prefers to call them), and correspondence is just to consider those things qua Fact, while coherence is to consider them qua Idea. Since everything we actually could possibly call real is necessarily both (Whewell is very explicit about this), from the perspective of a Whewellian realist ontology the separation Hugo makes above is merely an attempt to smuggle the separation of fact and theory back in. This is (again, from such a realist point of view) a common problem with the discussion of theory-ladenness; it assumes that the realist is somehow really committed to positing things on the observational side of an observational/theoretical divide. But a realist like Whewell, who holds that anything we could possibly consider to be reality necessarily to be Conceptualized or Theorized Fact (roughly, on the Whewellian view, Concepts are specified Ideas; Theories are constructs of Concepts), with no real separation of Fact and Concept possible, simply does not do that. In other words, my point is that I don't think the line of thought that Newall put forward is taking the impossibility of a theory/fact separation seriously enough. But in any case, as I note above, I explicitly denied that I was putting forward a coherence account of truth.

As I've noted several times, my argument was conditional: given that there really are traces of the past, we are committed to historical realism about the past of which there are traces insofar as it leaves traces. While Newall denies that the only alternative to this sort of historical realism is historical skepticism, if my argument is right (and he has actually not dealt with it, since he hasn't yet taken into account its conditional character), it is so: either in conceptualizing things as traces of the past one conceptualizes them as being traces of the past, or one is committed to denying that there are any traces of the past. This is trivially true: either traces of the past are of the past, or there are no traces of the past. Now, if one conceptualizes something (a book, for instance) as a trace of the past, one is committed to the denial of skepticism about there being a past for the book; there is such a past, ex hypothesi. Only by denying that there are traces of the past, which implies skepticism about the past, can this be avoided. Or so it seems. The challenge, it would seem, to someone who rejected such realism, would be to create an account of history that is intelligible but did not think of history as presupposing traces of the past; and to someone who rejects the above dichotomy, to present an account of history that recognizes traces of the past but does not in so doing commit one to the basic historical realism noted above. The fact that the traces are made appears to be irrelevant, because it seems only to be relevant on a view in which this would be a problem for realism, namely, a view which holds that the realist is committed to an observational/theoretical divide and that something has to be entirely on the observational side to count as objective, real, or whatever. But as there is no reason to hold this, there is no reason for the realist to worry about whether traces of the past are made by the mind are not; she is not committed, qua realist, to saying that they are not. To say that historical statements are meaningless is, it seems to me, a form of skepticism about the past, since it necessarily denies that traces of the past are of the past - they can't be anything if historical statements are meaningless.

So I conclude that nothing Newall has put forward in his new post actually touches my argument: (1) I take the impossibility of the theory/fact separation seriously; (2) I did not put forward a coherence account of truth; (3) his argument on approximate truth makes the mistake of assuming that approximately true accounts have to be part of one account in order to be approximately true; (4) in short, (re both (2) and (3), Newall seems to me still to be making the mistake of not taking the approximateness itself seriously; (5) his question, "Where two other accounts are mutually exclusive, moreover, which of these possible approximately true versions are they to cohere partly with?" is thus not relevant to the question of whether something is actually approximately true, for much the same reason questions about progress in approximation aren't, namely, that it's a downstream question and thus not a problem for an account appealing to approximate truth; (6) therefore, my comment that nothing has been done to show that there is anything in any of this that properly speaking discredits historical realism still stands.

But, then, since I've been arguing that a form of historical realism is necessary for there to be history at all, and Newall says he isn't arguing about realism at all, it's not surprising that there's been some miscommunication. In any case, the primary purpose of my post was not to refute Newall but to rough out why I think the problems alleged against the historical realist are exaggerated; and it still does that quite well.