Blog powered by Typepad

« Post-docs at ANU | Main | New SEP entries »

November 21, 2005


Justin Fisher

I don't think I understand Dave's intuitive characterizations of revelation and humility in terms of "our knowing what a property is". Here are three ways of (mis)understanding this, which surely aren't the ones that he intends:

(1) "Sure, I know what the property of being an electron is, it's the property of being an electron. What else did you think it might be?"

(2) "Sure, I know what the property of being Newman is; it's the property of being the guy I dubbed Newman a moment ago."

(3) "Sure, I know what the property of being an electron is, it's the property of playing thus and such a role..."

I'm sure that the defenders of the revelation/humility distinction will think that none of these do well enough to demand that we "know the essence of" the relevant properties, or somesuch, but I don't think I have a good sense of what this somesuch would mean.


Well, (2) and (3) are arguably not very good answers, since they're arguably false (since in each case, something could arguably have had the latter property mentioned without having the former). In any case, certanly there is no claim that this definition reflects the ordinary sense of "knowing what". As Lewis says in the key passage quoted in section 7 of Daniel's handout, the sense in question is an "uncommonly demanding" sense. There's not much point debating ordinary language here, but I think there's a demanding but not wholly unnatural way of understanding the expression according to which someone who has the concept of heat without knowing that it is molecular motion doesn't know what property heat is (it's a notion of that sort that is made more precise by the 2D definition). Of course all this is consistent with the claim that such a person may know what property heat is in a less demanding sense. In any case, the more precisely defined notion is the one that matters, and nothing here rests on the use of the phrase "knowing what".

Justin Fisher

I agree that the stuff that Dave does later rests upon his 2D-analysis of revelatoriness rather than the intuitive characterization in terms of "knowing what a given property is". Still, I was hoping I could get an intuitive grasp on what this restrictive sense of "knowing what" was supposed to be, as this seems to be the standard way of formulating what is at issue in these debates. I'm afraid I still don't have a clear sense of it (and I suspect that what this reflects is a problem in these debates and not a problem in me).

I'm afraid that Dave's 2D-analysis didn't really help me get much more clear on this. This analysis says that "a revelatory concept is one such that it picks out the same property in all [epistemically possible] worlds considered as actual". I'm puzzled by trying to judge sameness of properties across various epistemically possible worlds, each considered as actual.

Consider two epistemically possible worlds: Earth and Putnam's Twin Earth (upon which it turns out that the local lakes, rivers, etc are actually filled with XYZ). What's to stop someone from individuating properties as follows? In Earth, considered as actual, the property of being water is (or at least is coextensive with) the property of being H20, while in Twin Earth, considered as actual, the property of being water is (or is coextensive with) the property of being XYZ; in both of these epistemically possible worlds, our water-concept picks out the same property, namely the property of being water (or maybe Dave would prefer to put it: 'the property of being the (actual) local watery stuff'). If someone says this, then she would conclude that our water-concept is revelatory, but I take it that Dave doesn't want to accept this conclusion.

I imagine that this violates some tacit presumption that Dave has about how properties are to be individuated across epistemically possible worlds, but it would be nice to have this presumption stated clearly, and to have it justified.

Relatedly, I'm worried that Dave might be committed to saying that the concept 'actually being a philosopher' satisfies his first formulation of the 2D-analysis of revalatoriness stated above (picking out the same property in every epistemically possible world taken as actual) but not his second formulation (having its extension across metaphysically possible worlds be independent of which world is taken to be actual). The intuitive reason for this is that the first formulation ignores the subjunctive modal profile of a concept (e.g., whether the concept designates rigidly or descriptively), while the second formulation relies upon it.

I guess Dave will probably deny that this case actually satisfies the first formulation (probably because of his "modal analysis of properties"), but I think it will be useful to see this denial spelled out.


Here I'm more or less following Lewis's 2D account of "knowing what a property is". The worlds considered as actual are regular (centered) metaphysically possible worlds. The properties picked out by a concept in a world considered as actual are just regular properties, and the criteria for transworld identity of properties are just the regular criteria. Of course those criteria depend on one's account of properties. But we might as well work with the modal conception of properties as something like functions from worlds to classes of individuals in those worlds.

For the concept being water, if Earth is considered as actual, this picks out the property of being H2O, while if Twin Earth is considered as actual, it picks out the property of being XYZ. Of course these are distinct properties. As for "actually being a philosopher", if the idea is that this expression picks out the class of possible individuals who are philosophers in the actual world, then it will satisfy neither the first nor the second criteria for revelatoriness. It picks out different properties in different worlds considered as actual (if A is a philosopher in W but not in W', then if W is considered as actual, the expression picks out a property that A has in both worlds, while if W' is considered as actual, the expression picks out a property that A lacks in both worlds). Correspondingly its extension across metaphysically possible worlds will vary depending on which world is taken to be actual. If we individuate properties by their extension across metaphysically possible worlds, then these two criteria won't be able to come apart.

Jason Zarri

As for the property "actually being a philosopher", I think there is another way of taking it. While might take "the actual world" in a descriptive sense to mean "whichever world happens to be (or is considered to be) actual" we could also take it as rigidly designatig *this* very possible world, irrespective of actuality. This is a point Plantinga discusses in "Essays on the Metaphysics of Modality". So if we follow Plantinga and call our home world Kronos, the property "actually being a philosopher" comes out as the 'world-indexed' property "being a philosopher in Kronos" on this reading. This will then take everyone who is actually a philosopher and pick them out in any world in where they exist, irrespective of whether or not they are a philosopher in that world. For no matter how things turn out, it is true that they are a philosopher _if_ Kronos is actual. In this way the term "being a philosopher in Kronos" rigidly designates an entire class of individuals.


Sure, given that W1 is the actual world, then "actually being a philosopher" designates the property of being a philosopher in W1, and designates that property in all worlds (considered as counterfactual). But the point of bringing in the 2D apparatus here is that one can also consider hypotheses according to which worlds other than W1 are actual (even if it's necessary that W1 is actual, it's not a priori that W1 is actual). Relative to W2 considered as actual, "actually being a philosopher" will designate the property of being a philosopher in W2, and will designate that property in all worlds (considered as counterfactual). This is compatible with the Plantinga point, as evaluation in worlds considered as actual is a different sort of evaluation from ordinary possible-worlds evaluation. For formal definition of this sort of evaluation in terms of the logic of the "actually" operator, see Davies and Humberstone's "Two Notions of Necessity".

Jason Zarri

Thanks for your response. I have to admit I'm an undergraduate, and I'm not sure if I could follow a full formal definition. What I was trying to get at (even though I probably didn't express it as clearly as I could have)is that even though any possible world could have been actual (which should be vacuously true on any reading)Kronos is "the lucky world", the one that really is (Just as we might say that even though anyone could have invented the lightbulb, Thomas Edison is the lucky person who really did). So unless we accept a sort of indexicalism about 'actual', I feel that Kronos should have a special staus when we use terms like 'actual world'-- of course we can _consider_ any world as actual, but only Kronos is, and not just "in a world", but absolutely.

Timothy J Scriven

Thanks, I found this very useful. It's given me food for thought in relation to my argument for error theory, good can't be adequately distinguished from bad due to certain absences of relational properties, therefore good and bad are meaningless. As someone who's only just now doing an introductory course to philosophy I find it almost awe inspiring how seemingly massively distant branches of philosophy are relevant to each other.

The comments to this entry are closed.