Tuesday, April 19, 2011

Two problems of temporary intrinsics

I've been thinking about the problem of temporary intrinsics and don't see much of a problem.  There are two kinds of ways of formulating the problem, and I think they are basically different problems, and neither is particularly compelling.


Formulation 1:
  1. Socrates at t0 is bent.
  2. Socrates at t1 is straight.
  3. Socrates at t0 = Socrates at t1
  4. So, Socrates at t0 is bent and is straight.  (Which is absurd.)
I think this is a linguistic paradox rather than a metaphysical problem, and hence deserving of being linguistically defused.  "Socrates at t0 is bent" is awkward English.  The normal word order is "Socrates is bent at t0."  But if we rephrase (1) this way (and (2) analogously), the argument becomes invalid.  Nothing untoward follows from Socrates being bent at t0 and Socrates being straight at t1.  All that follows is that he is bent at t0 and straight at t1.

For the argument to be valid, we need to parse (1) as: "Socrates-at-t0 is bent", and (2)-(4) analogously.  But what is this Socrates-at-t1?  Suppose we say that it's just Socrates under another name.  Then we should deny (1), since Socrates isn't bent--he's dead (unless by "is bent" we mean "is bent at some time or other", in which case (4) tells us that "Socrates-at-t0 is bent at some time or other and is straight at some time or other", which is unproblematic).  Of course, if "Socrates-at-t0" is just another name for Socrates, we can say "Socrates-at-t0 is bent at t0".  But no untoward consequences follow from the claims that Socrates-at-t0 is bent at t0 and that Socrates-at-t1 is straight at t1.  We end up saying that Socrates-at-t0 is straight at t1, which sounds weird, but that weirdness only comes from this weird "Socrates-at-t0" name we've used.  It's like the weirdness of saying: "Ivan the Terrible was actually a pretty nice kid" (which for all I know is true).

I think what is going on here is this.  We sometimes speak in the historical present with a contextually implicit time.  We say things like: "September 1, 1939.  Germany invades Poland.  The Polish defenses crumble."  The two sentences following the contextual introduction of September 1, 1939 are to be understood as saying that Germany invades Poland and the Polish defenses crumble on that date.  We do the same thing spatially.  For instance, we can be describing the course of the (imaginary) Borogove River which comes from Oklahoma to Texas.  We've just described it in Oklahoma.  We now say: "Texas.  The Borogove is very silty."  We mean that it is silty in Texas.  In the case of "Socrates-at-t0", the "-at-t0" determines the context of evaluation for the historical present "is bent."  So all we are saying is that Socrates-at-t0 is bent at t0.  And no paradox ensues.

Now, there is another reading.  We sometimes adopt a metaphor of a individual being split into multiple individuals, either by means of time or role.  Thus we say things like:
  1. Late Plato disagrees with Middle Plato on whether all the serious problems of philosophy are solved by positing the Forms.
  2. Smith the Rhetorician loves this argument, but Smith the Philosopher hates it.
When we adopt this fiction, we do not allow intersubstitution--that would be inappropriate mixing of metaphor with reality, like when someone says that the lights came on for her after she read so-and-so's paper and we ask if they were incandescent or fluorescent.  In other words, on this metaphorical reading of (1) and (2), we will reject (3).

Granted, the perdurantist can take "Socrates-at-t0" and "Socrates-at-t1" to literally refer to two entities, and then reject (3).  But that kind of metaphysics is not at all required by the argument.

So, in the first formulation, the argument can be defused purely on linguistic grounds.  This point applies also to my favorite formulation of the problem:
  1. The young Socrates is ignorant.
  2. The old Socrates is wise.
  3. The young Socrates is the old Socrates.
  4. So the young Socrates is wise and the old Socrates is ignorant.
Formulation 2:
  1. Presentism is true or the application of a temporarily applicable predicate to x is never correctly explained in terms of x's instantiation of a non-relational monadic property whose choice is dependent only the predicate (and not on the time of application).
  2. The predication of shape (say) predicates is correctly explained in terms of the object's instantiation of corresponding shape properties.
Notice that while the first formulation could grip a non-philosopher, (11) is simply a constraint on philosophical theories of predicate application.  There seems to be very little cost in denying (12) and its parallels, since (12) and its parallels simply do not state any sort of ordinary intuition--they are a substantive claim about how to explain predication.

10 comments:

Leo said...

Thanks for the post.

If you're interested, David Oderberg offers a somewhat similar response to the problem of temporary intrinsics here.

Alexander R Pruss said...

Thanks for the reference!

Sarraclab said...

Hmm, would there be something wrong with taking Leibniz's law as only applying to synchronic identity? Wouldn't that be a quick fix?

Leo said...
This comment has been removed by the author.
Leo said...

David: Why would we need to have recourse to such a move when we have other resolutions available that leave Leibniz' Law fully intact?

Sarraclab said...

Leo: Hm, I'm not sure the solutions leave Leibniz's Law intact. It seems to follow from this solution that two things can be diachronically identical without sharing all their properties, which entails that Leibniz's law doesn't apply to diachronic identity, no? So I'm saying it might be a good enough solution to just state that upfront.

Leo said...

David: I don't see how that follows. Suppose that we interpret times along the lines of modal operators, so that "Fido is barking at t" becomes "at t(Fido is barking)." Just as it does not violate Leibniz' Law to say that both "possibly(Fido is barking)" and "actually(Fido is not barking)" are true, so it does not violate LL to affirm both "at t1(Fido is barking)" and "at t2(Fido is not barking)." Or am I missing something?

Leo said...

To put it another way, if Fido-at-t1 and Fido-at-t2 are identical, and Fido-at-t1 barks at t1, then Fido-at-t2 barks at t2. No challenge to Leibniz' Law here.

Leo said...

Whoops! That should read "... then Fido-at-t2 barks at t1." My bad ;-)

Sarraclab said...

Yes, that's a sort of relationist approach, and I tend to agree with Lewis about its problems, about it making intrinsics relational, which seems funny. But yes I agree that modifying Leibniz's law probably isn't an easier or less problematic solution.