Third-Manning the Representation Relation




OK

Originally uploaded by Pete Mandik.

I’ve been thinking about Chase Wrenn’s third-man argument against realization (here) and my own against non-reductive physicalism (here) and the less-than-fully-baked idea occurred to me to run a third-man against the so-called representation relation.

My recollection without looking anything up is that old-school third-man arguments go something like this:
Step 1. Start with some uncontested fact like that Chase is tenured.
Step 2: Conjoin it with a Dumb Principle like, nothing can be true of Chase (like being tenured) without there being some additional entity (like the property of being tenured) and some relation borne to that entity (like the instantiating-a-property relation).
Step 3. Generate an infinite regress by reusing the Dumb Principle to postulate, next, a third entity (like the property of bearing the partaking in the instantiating-a-property relation) and so on.
Step 4. Block the regress by rejecting the Dumb Principle

Adapting this to the present case, namely, to deride the representation relation, yields:
Step 1: Start with some uncontested fact like that Chase thinks that Buffy the Vampire Slayer is blonde.
Step 2: Conjoin it with a Dumb Principle like nothing can be thought about by Chase (like that Buffy the Vampire Slayer is blond) without there being some additional entity (like the non-actual possible state of affairs of Buffy the Vampire Slayer being blond) and some relation borne to that entity (like the representation relation).
Step 3. Generate an infinite regress by reusing the Dumb Principle to blah blah blah…
Step 4. Block the regress by rejecting the Dumb Principle

14 Responses to “Third-Manning the Representation Relation”

  1. R Brown says:

    Buffy is hot…

    *snaps out of it* uh, can you do this for the causal theory that requiers the thinker to be in a certain relation to actually existing things? I don’t hink you’ll get the regress then…

  2. Pete Mandik says:

    Richard,
    I thought that causal theories begin by presupposing that representation is a relation and then go on to posit the identification of the representation relation with certain causal relations. If this third-man argument works, it would block that initial supposition.

    Do you know of any arguments for a causal theory that don’t make such an initial supposition?

  3. Chase Wrenn says:

    Pete-

    I thought you’d been thinking along these lines before, and it was part of what made me think of arguing against the reality of realization in the way I did. So, I’m sympathetic.

    But I’m also worried about how the relevant Dumb Principle gets iterated in this case. I believe that Buffy is the Slayer. That state consists partly in my having a mental representation with the content that Buffy is the Slayer. For the representation to have the content is for it to bear a certain relation to the Outside World (which may or may not include intentionally inexistent Vampire Slayers).

    For the Dumb Principle to iterate, it has to be the case that I can’t bear that relation to the Outside World unless Outside World and I bear some further relation to some further thing, etc. ad infinitum. It isn’t obvious that just saying representation is a relation automatically commits us that far.

    How about this Wrennian argument against the representation relation? If there were a representation relation, it would make no causal difference to its relata. But a relation does not exist unless it makes such a difference. So, there is no representation relation.

  4. R Brown says:

    Hey Pete, I think you are right about that…it is certainly one of the ways I have been pressing to avoid the unicorn…but my point was more along the one that Chase made. I do not think that you will be able to iterate the dumb principle when what you are talking about is a causal reference relation.

    Chase, so do you think that being a brother is not a real relation? What about being to the left of?

  5. Pete Mandik says:

    Hi Chase,

    I don’t have anything particularly intelligent to say about criteria for existence, causal or otherwise. But I’ll say something anyway. I strikes me as coherent to doubt the existence of causation. It doesn’t, however, strike me as coherent to doubt whether anything at all exists. This assymetry makes me think that making causation a criterion on existence gets things exactly backwards. That said, I do feel the pull of “what difference would it make?” questions in contexts concerning ontic comittment, so I don’t mean to be totally dismissive of the causal criterion.

    Regarding the third-man argument against the representation relation, could you clarify something for me? When you make the following assertion, it’s not clear that you are embracing the relational account that I am targetting.

    You say:

    For the representation to have the content is for it to bear a certain relation to the Outside World (which may or may not include intentionally inexistent Vampire Slayers).

    There are lots of readings of that that wouldn’t be the sort of thing that I’m targetting. So, consider a social conceptual role semantics whereby what gives your thought the content it has is that you are the member of a languagage game playing community that gives you a jelly bean every time you utter “Buffy is blonde”. That would make having the thought that P very much involved with relations to stuff in the Outside World. But that wouldn’t be the sort of relational account that I am after.

    The relational account that I am after is one that reads off of ascriptions like “Chase thinks that P” that “thinks that” names a two place relation that relates something named by “Chase” to something named by “P”.

    So the clarification I’m requesting is whether you take it that the sort of relations to the Outside World you had in mind are the sort that I have in mind to target.

    At this point, it is perhaps worth pointing out that it doesn’t matter so much what the relation is or what the relata are. What matters is what prinicple motivates the relational view. So, by itself, even the view that in “Chase thinks that P”, “thinks that” names a two place relation that relates something named by “Chase” to something named by “P” is not something automatically doomed by the Third Man. What matters crucially is what the reason or principle is that motivates the relational analysis. The Third Man argument highlights the need for providing a principle that isn’t Dumb.

  6. Chase Wrenn says:

    Richard-

    Right. Brotherhood and being to the left are not part of the furniture of the universe. But fear not! It’s still true that there some people are to the left of their brothers, and ‘_ is brother of _’ and ‘_ is left of _’ are still polyadic predicates. The point is just that those predicates don’t pick out ENTITIES.

    To put it a bit more obscurely: When I say brotherhood isn’t a real relation, the emphasis should be on ‘real’, not on ‘relation’.

  7. Chase Wrenn says:

    Pete-

    Suppose for moment that having a content is expressing a proposition. If I think Slaying is cool, then I bear a relation to the proposition that Slaying is cool.

    (You know I don’t think that about contents and propositions, but some folks do, and I took you to be arguing against them.)

    The 3rd Man trouble only arises once that relation is reified and treated as if it were part of the furniture of the universe — something with a nature to be discovered and theorized about in some sort of empirico-conceptual way. Of course, I suppose a lot of people do that. Shame on them!

  8. Pete Mandik says:

    Chase,

    Right. I don’t disagree.

    But the positing of the second relata (that which is thought about) itself is in need of a justification that won’t lead to the dumb principle.

    Compare: In the standard third man case, the sorts of considerations that motivate reifying a property are the same justifications that motivate the property of having a property and thus, regress. Such motivations might include “well, a red ball and a red pen have to have something (read as some thing) in common”.

    Simlarly: “Well, if Chase thinks Buffy slays and Richard thinks Buffy slays, they have to have something (read as some thing) in common”.

    And off we go to infinity.

  9. R Brown says:

    Ahh, thanks Chase…so let me ask you, is the causal relation itself a real relation?

  10. Chase Wrenn says:

    RIchard-

    I’m inclined to deny that there is an entity corresponding to the causation predicate. As you might guess, motivating such a posit without inviting the Third Man to the party might be difficult.

    Some people think such a denial is inconsistent with embracing an Eleatic criterion of reality for properties and relations. But I don’t think it is. Probably that’s because I tend to think of properties and relations as having their causal powers built in, and so we don’t need causation AND properties for things to happen in the world. We just need the properties.

    [I am suppressing various more radical things I might say here. Maybe I'll post some of them on my blog soon.]

  11. R Brown says:

    Yeha, I guess I am one of those people, that was kinda themotivation for my question…

    At any rate, don’t you think that it is a bit odd that, on your view, causation is not a real relation, but ‘being 100 feet above the Earth’ is?

  12. Chase Wrenn says:

    I don’t think that’s odd at all, Richard.

    Remember, I’m not denying that ‘_ causes _’ is a dyadic predicate, and I’m not denying that it has satisfaction conditions. There are plenty of true causal claims, and plenty of false ones.

    What I’m denying is that there is a third entity, “the causal relation,” acting as the middle man between causes and their effects.

    (I should mention my ambivalence about the reality of ‘being 100 feet above the Earth’. It satisfies a necessary conditon for being a real relation, but I’m open to the possibility that there are others it doesn’t satisfy. So, I should say that causation isn’t a real relation, but ‘being 100 feet above the Earth’ might be.)

  13. R Brown says:

    Oh, I think I agree with that…I guess what gets me is that that means thatit is not a REAL relation, whereas it seems to me that that is just how relations are. Ducassee pointed out this fallacy long ago: don’t think of a relation as the same kind of thing as those which it holds between. But I don’t think that means they are any less real…Though I admit this is probably teminological.

  14. Chase Wrenn says:

    It probably is terminological, Richard.

    There are formal reasons we can allow that there are entities corresponding to all our predicates, even if the theory using those predicates is true. So, when I worry about “real relations,” I’m thinking of relational entities, designated by polyadic predicates, over and above the entities that satisfy the predicates.