Saturday, December 12, 2009

Beings of reason

Extend English by adding a new kind term "ersoh", and allowing quantification over ersohs. Now the following stipulations generate a semantics for ersohs as well as for a relation "horsesucceeds":

  1. Every ersoh horsesucceeds a unique horse.
  2. Every horse that dies at a time after which there are at least three minutes (i.e., at a time such that time does not come to an end in less than three minutes after the horse's death[note 1]) is horsesucceeded by a unique ersoh.
  3. If e horsesucceeds h, then e exists forever (if time has no end) or until the end of time (if time has an end) from the moment three minutes after the death of h.
  4. An ersoh e1 in w1 is identical with an ersoh e2 in w2 if and only if the horse that e1 horsesucceeds in w1 is identical with the horse that e2 horsesucceeds in w2.
  5. If P is a physical-property predicate (e.g., "has mass of 2kg") that it makes sense to apply to a physical object at particular time, then e satisfies P at t if and only if (a) e exists at t and the horse h that e horsesucceeds satisfies P at t*, where t*=f(t,td,tb) where td is h's time of death, tb is the beginning of h's existence, and f is some specified one-to-one function[note 2] that maps times from td+3min to times between tb and td.
  6. The only basic predicates that an ersoh satisfies are is an ersoh, horsesucceeds h and P where P is handled as in (5). All other basic predicates are unsatisfied by every ersoh. More complex predications to ersohs are to be reduced to these in some Tarskian way.

I may have omitted something needed for a complete semantics of ersoh-talk. If so, add it, keeping to the spirit of the proposal. The point is that we can easily generate a complete semantics for a language that supposes ersohs.

Now, surely, there are no ersohs in a strict sense of "are". But we can meaningfully talk about them in a way that reduces to talk about horses.

Here, then, is one of my major philosophical intuitions: Any entities talk about which reduces to talk about other entities have the same ontological status as ersohs. They don't really exist. Hence, if reductionism about persons is true, then in a strict sense of "are", there are no persons.

But what is this "strict sense"? Well, the above gives an ostensive definition of it: the basic entities are, while ersohs aren't.

[Minor typos fixed.]


David Manley said...

This is a nice test-case for thinking about quantifier variance. My intuition in this case is that stipulations don't always succeed. If I were a semantic god, deciding what semantic values should be assigned to terms given the pattern of use that subvenes, I'd say 'esroh' fails to refer and all attempts to use it crash semantically.

The reason is that, to make the stipulation succeed, the meanings of the predicates involved 'weighs n kg', and the meaning of the quantifier itself, would have to shift. This would require interpretive gymnastics all over other areas of the language. It's not formally impossible, but will yield counterintuitive results; but the relevant intuitions that would be violated are among the facts on which meaning supervenes. Perhaps there is a possible language in which the meaning of the quantifier (or some quantifier-like term, defined by its inferential role) means what the English quantifier would have to mean if this stipulation were to work. BUT-- such a language is not otherwise like English. The most charitable semantics for English plus these stipulations makes them fail.

I hope that makes sense!

Alexander R Pruss said...

We do get some predicate shifting in other cases that people tend to think are OK. For instance, it's plausible that electric charge applies in the first instance only to particles (and it seems they have it essentially). A larger material object, then, is negatively charged iff the sum of the charges of the parts is negative. So we have the primitive property of charge that particles have, and the non-primitive property of total-charge that non-particles have that is defined (stipulated?) in terms of the primitive property. Granted, in this case it may seem not so bad as for esrohs. For particles also have total-charge just as the big objects do.

Of course this might just be a good argument against composition.

Alexander R Pruss said...

One could also make there be no predicate shifting in my case by saying that no non-existence-type positive primitive predicate applies to an ersoh other than "is ersoh" (thus, exists-at-t applies, but that's all). However, then one might want to identify ersohs with states of affairs (the state of affairs of a horse having died at least three minutes ago). But one could modify the identity conditions so that wouldn't be plausible.

David Manley said...

Right-- and noting that esrohs might just turn out to be states of affairs makes me suspect that states of affairs are just beings of reason!