Thursday, August 20, 2009


Imagine a language like English but that has no disjunction—it still, however, has conjunction and negation. There is a sense in which speakers of that language are not missing out on any aspect of reality. It may or may not be true that the English "p or q" expresses the same proposition as "it's not both the case that not p and that not q", but there seems to be an intuitive sense in which whatever reality is pointed out by "p or q", it is also pointed out by that disjunction-free circumlocution. Hold on to that sense of "not missing out".

Now, take a language like English but devoid of indicative and subjunctive conditionals. However, the language has disjunction, negation and conditional probabilities ("it is (very, somewhat, not at all, etc.) likely that ... given ..."). Are the speakers of that language missing out on any aspect of reality? Is there any irreducibly conditional aspect of reality?

Let me grant for the sake of argument that the English indicative "q if p" is not logically equivalent to "q or not p". But are there aspects of reality that can only be captured by indicative conditionals, with disjunction and negation not being adequate to the task? I do not know of any. That is a weak argument for their non-existence, I know. However, I tend to be fairly self-conscious about my use of conditionals, and I talk about a large variety of subjects, so there is some evidence here. Certainly, it does not appear that science requires an indicative conditional. "Either the reaction occurred or the chemicals were not mixed" seems to do just as well as "The reaction occurred if the chemicals were mixed."

There may be conditional obligations. But a conditional obligation either uses a sui generis conditional that is neither the indicative nor the subjunctive, or else can be expressed disjunctively: "Either I should do what I promised or I cannot reasonably do what I promised."

If Molinism is true, the subjunctive conditional is irreducible, and expresses features of reality that cannot be otherwise expressed. But I think Molinism is false.

It is tempting to think dispositions cannot be expressed in any way other than by subjunctive conditionals. But it's really hard to express dispositions even by means of subjunctive conditionals, given finkish problems.


Brandon said...

Perhaps it depends on what conditional locutions we are considering. "If p, q" is very generic: it doesn't establish any particular sort of relation between p and q. But there are more domain-specific conditional locutions, e.g., when considering explanation

q because p

"The reaction occurred if the chemicals were mixed" is consistent with the relation being symmetrical -- perhaps the two are merely correlated effects of something else. But "The reaction occurred because the chemicals were mixed" establishes an asymmetry between the two propositions. "Either the reaction occurred or the chemicals were not mixed" won't preserve the asymmetry, and English doesn't seem to have a disjunctive locution that would.

Of course, that's not strong enough to allow us to conclude that there is any "irreducibly conditional aspect of reality"; we'd need to know not merely that English in fact does not have domain-specific disjunctions to correspond to domain-specific conditionals but that it couldn't in fact do so because they can't be made. And that would be a hard thing to argue either way, I'd imagine.

Alexander R Pruss said...

I've never thought of "because" as a conditional. Do you want to expand?

Brandon said...

Hmmm; I've never thought of q because p as a candidate for not being conditional. I take q because p to be merely a specification of q if p, where the underlying relation is causal (or, more broadly, causally explanatory). (Likewise, q given p is merely a specification of q if p where the underlying relation is logical, e.g., {insert conclusion here} given {insert conjunction of premises here}.) Otherwise how would we explain the fact that q because p is often represented as q if p? It seems the most natural way to take this is to say that q if p is the genus, so to speak, and q because p is a species in which q and p have a particular kind of relation to each other.