Monday, March 03, 2008

Extrinsic Identity

All this talk of thisness reminds me of a question Jack once asked:
Let's agree that there is the property of being such and such, where such and such is a particular object, and be agnostic for a moment about whether that property is a fundamental property or derivative somehow. At the same time, let's acknowledge the long-standing distinction between intrinsic properties and extrinsic properties, where whether an intrinsic property applies to some object depends only on how that object is, in some to-be-specified sense. Do you think that the property of being such and such is an intrinsic property, an extrinsic property, or sometimes one sometimes the other?

To which I responded:
I lean towards the extrinsic option (insofar as I have a grasp of the question at all). At least for 3-Dists, the best theory of identity is Nozick's "closest continuant" account, but which future object is the closest continuant of your present self is an extrinsic matter. (Fission cases illustrate this nicely. Either continuant alone would have been such-and-such, were it not for the other!) And counterpart theorists are going to say similar things about trans-world identity. I can't tell whether a counterfactual guy is Jack just by looking at him; I need to know whether there are any better candidates in the vicinity!

How might a world ontologist accommodate some kind of (derivative) 'thisness'? We've seen it must be extrinsic, because there is nothing in the duplicates Bob1 and Bob2 to distinguish them. But it's not entirely obvious how to fill out the details. My basic strategy is to model the identity facts on the semantic facts, as per the holistic account I offer here. Roughly: the distinct identities [thisnesses] of Bob1 and Bob2 consists in the extrinsic fact that their world contains two Bob-like duplicates. The perfect symmetry means that there are no grounds for assigning any particular distinct property to either one of them alone -- giving just one of them the peculiar property of being Bob1, say. That's no more possible than our name 'Bob1' referring to a determinate one of the two. Nonetheless, it may be determinate that 'Bob1' and 'Bob2' do not co-refer, and similarly that being Bob1 and being Bob2 are not co-exemplified, even if there's no further fact about which thisness is where. Does that seem objectionable? (It would be if these vague properties were metaphysically fundamental. But I think their 'derivative' status may afford more leeway...)


  1. Hey Richard, I've read through some of your earlier posts on this topic, and it seems you have a rather peculiar idea of what a "thisness" is supposed to be. On your construal, a thisness is a property which ensures that a thing will remain itself no matter what qualities you remove from it -- it's sort of a magical "personal identity" property for the thing it's a part of. To me, this sounds more like the usual definition of a substance (or perhaps an attribute, in Spinoza's terminology, or something like that), which I would definitely agree with you is a pretty sketchy notion. This is, however, not what Robert Adams, at least, has in mind in the article you mentioned in your previous post on this topic. To him thisnesses are merely supposed to be relations of "brute self-identity". That is to say, if you were to remove just a single quality of a thing (or change it in any other conceivable way -- e.g. its relation to other things or its place in the world or whatever), you would change its "thisness".

    Given this definition of thisness, I don't think Robert Adams' individuation of objects by their thisnesses is that different from your individuation by "world facts" about the objects. In fact, I think his individuation includes yours as a special case, because if there are facts about the world that individuate, say, Bob1 and Bob2, then these world facts would, of course, be part of what determined the "brute self-identity" of both Bob1 and Bob2, and thus would be part of what determined the thisnesses of each of them. On the other hand, Adams' individuation by thisnesses may be superior because of its generality. Indeed, who's to say that there _could_ be higher-order "facts about the world" that managed to differentiate Bob1 and Bob2? Hacking would, of course, say that such higher-order facts cannot possibly exist. (Although both of us disagree with Hacking.)

    The objection to the Adams' individuation might be, of course, that the definition of "thisness" becomes _so_ general that it becomes empty. But I, at least, don't think this is true. I think its generality is a plus. In my estimate, there _is_ such a thing as "brute self-identity" (even if I keep putting it in quotation marks..), and the nature of this relation of "brute self-identity" depends of the nature of the scenario (or "world", if you prefer) with which we are dealing.

  2. (By the way, my above reply was probably pretty misplaced; your post deals very explicitly with the issue I raise in the third paragraph -- that is, how the nature of "thisnesses" may depend on the world we are dealing with, with you claiming that it's always in terms of world facts.)

  3. Richard, I want to comment on two things, both concern your use of determinacy.

    First, on your view the property of Bob1 and the property of Bob2 are exemplified in the world, but it is indeterminate which object has the property of being Bob1. Is that right? I might have missed something. If yes, what is the source of the indeterminacy? Is the source semantic? Is it supposed to be metaphysically indeterminate as to which of two objects has the property of being Bob1 because it is semantically indeterminate which thing is called 'Bob1'? That would seem to me to reverse the order of explanation. It is strange to have the world be indeterminate because our language designed to describe the world is indeterminate. But maybe I have misunderstood you here.

    Second, is it a feature of your view that Bob1 and Bob2 exemplify all the same properties? I think it is. But then, I wonder how many things in the Bob1/Bob2 world instantiate the property of being Bob1. Here are the options: none, one or two-indeterminately. If none, then I suggest that Bob1 is not in the world. If one, then, Bob1 and Bob2 differ in their properties. This would be contrary to your world ontology I think, favoring constituent ontology. If two-indeterminately, then both Bob-like objects exemplify the property of being Bob1 indeterminately. In that case, I say that it is at best indeterminate whether Bob1 is in the world. At worst, it is determinate that Bob1 is not in the world. The thought behind the worst case scenario is that Bob1 couldn’t possibly be identical to a thing for which it is indeterminate whether it is Bob1.


  4. "Second, is it a feature of your view that Bob1 and Bob2 exemplify all the same properties? I think it is. But then, I wonder how many things in the Bob1/Bob2 world instantiate the property of being Bob1."

    Maybe I should let richard answer a question addressed to him, but I'd say that here you are already assuming "thisness" in the sense that I described above: that is, you are assuming that there _is_ such a thing as a "property of being Bob1". If there is such a thing, then obviously only one thing can exemplify this property: namely Bob1. This "property" is clearly equivalent to nothing less than Bob1's thisness (in the sense that I explained above).

    But why would you call this a "property" at all? It's not a property in the _qualitative_ sense in which properties are usually understood. That is, properties are usually understood to be things which _many_ objects can potentially exemplify, but "the property of being Bob1" (which is really no different from "the brute self-identity of Bob1") violates this criterion from the beginning.

  5. BV - I must disagree when you suggest that "if you were to remove just a single quality of a thing (or change it in any other conceivable way -- e.g. its relation to other things or its place in the world or whatever), you would change its 'thisness'."

    Adams opts for a moderate essentialism, according to which we still have many non-essential properties that could change without changing our identity or 'thisness'. (He mentions that Leibniz held the extreme view you mention; perhaps this is the source of your confusion. Note that Adams himself dismisses this "implausible thesis" [p.10].)

    [So, to clarify, a 'thisness' is defined by Adams on p.6. as being "the property of being identical with a certain individual".]

    Jack - I was thinking that the source of the indeterminacy is not itself semantic, but derives from a similar basis as the semantic indeterminacy does -- namely, the lack of any basis in the fundamental world facts for treating the two duplicates any differently.

    "Second, is it a feature of your view that Bob1 and Bob2 exemplify all the same properties?"

    No, I was thinking that exactly one of the duplicates exemplifies Bob1's thisness, and exactly one exemplifies Bob2's thisness, and no one object exemplifies both. This follows from the whole world description, on the same model as the semantic facts. (It's just that it's not determinate which is which.) You'll have to further explain why this might be thought to be in tension with my "world ontology"...

    [I should add that I'm sympathetic to BV's suggestion that thisnesses are not real "properties" in any deep sense, i.e. sparse properties. But I take it you're using the term in a looser sense, so I'm happy to try to accommodate that if possible.]

  6. A long reply, sorry.

    Richard, let me start with your second reply. On your view, if I understand it correctly, in the Bob1/Bob2 world there is exactly one thing that exemplifies the property of being Bob1. Is that right? If we were to list the properties of each object in the Bob1/Bob2 world, there would be exactly one object whose list included the Bob1 thisness?

    Now imagine a fallible deity, perhaps in a heaven above the Bob1/Bob2 world, pointing at one of the Bobs. The deity asserts: this is Bob1, while pointing at one of the Bobs. Your view is that the deity has either spoken falsely or has spoken truly—since the Bob either is the unique thing that exemplifies the Bob1 thisness or isn’t—so, it is not indeterminate whether the Bob that the deity is pointing to is Bob1? If there is a fact of the matter about which Bob the deity is pointing to, as I am supposing to be the case, then there is a fact of the matter about whether that Bob is Bob1. Is this right so far? If not, how could it be indeterminate of some determinate object whether it exemplifies a property that is determinately exemplified by exactly one object? (I can put the point in terms of a question: consider just one of the Bobs. Is it determinate or indeterminate whether this thing is Bob1? I thought you would say indeterminate—in which case one wonders whether Bob1 is in the world at all—but now I take you to be saying that it is determinate whether this Bob is Bob1.)

    What, then, are you saying in saying: “it’s just that it’s not determinate which is which”? It is too determinate, as shown by the deity example. For every object in the Bob1/Bob2 world, there is a fact of the matter, that is, it is determinate as to whether it exemplifies the property of being Bob1. What is true is that you cannot specify which Bob is Bob1 by qualitative means. If thisnesses are impossible to detect, then it might be that it is unknowable whether this Bob is Bob1. But there is a fact of the matter about which Bob is Bob1 because on your view exactly one thing exemplifies the thisness of Bob1. The one and only thing that exemplifies the Bob1 thisness is Bob1.

    Next, here is how saying that exactly one thing exemplifies the property of being Bob1 runs against your world ontology. Constituent ontology holds that if A and B have all the same constituents, then A and B are identical. In both the sparse and abundant senses, properties are constituents. It would be a problem for the constituent ontologist if intuitively Bob1 and Bob2 exemplified all the same properties and yet were intuitively distinct. But on your view, Bob1 and Bob2 differ in their properties. Why then the need for world ontology? Indeed, insofar as the Bob1/Bob2 case is a problematic case for the constituent ontologist, your view that Bob1 andBob2 differ in their properties is evidence for constituent ontology since it disarms a potentially problematic case.

    Finally, let me say something about your first reply. I don’t understand what the source of the indeterminacy is. If the indeterminacy doesn’t originate from the semantic consideration, then what? Is the world fundamentally indeterminate? Of course, the world is not qualitatively indeterminate. The world fully specifies the distribution of the qualitative properties. But does the world somehow fail to specify the distribution of the non-qualitative properties? It seems to me an unhappy consequence if the world is fundamentally and incurably indeterminate. Perhaps on your view the world isn’t fundamentally indeterminate, but then I am left wondering again about what the source of the indeterminacy is if it is neither the world nor the semantic considerations.

  7. There is no fundamental indeterminacy (as per the closing remarks of my post). Thisnesses aren't anything fundamentally real. But if we want to talk about them nonetheless, I'm offering a (somewhat convoluted) construction to allow us to do so. But the construction is not entirely determinate. We are adopting a rational procedure which does not yield a determinate true answer when given the fundamental facts and asked the question, "which duplicate shall we say has the property of being Bob1?"

    You ask: "how could it be indeterminate of some determinate object whether it exemplifies a property that is determinately exemplified by exactly one object?"

    In much the same way that it can be indeterminate of some determinate object whether it is denoted by a name that determinate denotes exactly one object. That is: it is determinate how many objects possess this feature (only one, certainly not two), but it is indeterminate which one object is so privileged.

    Compare the following distributions of Bob1's thisness between the two duplicates:
    (1) {here, -}
    (2) {-, here}
    (3) {-, -}
    (4) {here, here}

    The world facts allow us to rule out 3 and 4, and leave open 1 and 2 only. Hence (on my interpretation of indeterminacy), it is determinate that there is exactly one Bob1 thisness in the world, but indeterminate which of the two places it is to be attributed to.

    "in which case one wonders whether Bob1 is in the world at all"

    There is no content to such wondering. The fundamental world-description told you everything that's in the world. All else is wordplay.

  8. (Sorry for the extremely long post.)

    Richard, yes and no. I don't think there is a significant difference between moderate and extreme "essentialism", as you call it. It just depends on what you take the individuals/particulars to be (I will use the term "individuals" from now on). According to either doctrine, what we understand to be the relevant properties are all necessary constituents of any individual. It's just that most ordinary individuals are defined in such a way that a lot of properties are automatically excluded as irrelevant. Thus for instance, most individuals we talk about (i.e. people, cars, boats, stars, etc.) are, for instance, usually _defined_ to be time- and space-independent (at least the way people usually talk about them). This means that it's not considered a property of _me_ that I am sitting with a macbook in my lap right now (one can hear how artificial that statement sounds, as opposed to, say, "it's a property of me that I am conscious"). Of course, philosophers have made the distinction between accidental and necessary properties, but talk of the identity of ordinary objects, at least, will always be in terms of "essences".

    It has to be, if we are to be entitled to speak of an individual as having an identity at all. The alternative is to lapse into what you call a "substratum" theory, which (I think) is incoherent. Or else, to obtain a meaningless notion of identity (if two things that share no relevant properties, nor are thought to be made of the same "substratum", are nonetheless judged to be identical, what could possibly be the content of that judgment?)

    Anyway, I definitely should have chosen my words more carefully. The really significant difference lies been essentialism (of any degree) and substratum theory. Now, I (probably mistakenly, I now believe) thought that you took "thisness" to be a kind of substratum -- or better: a sort of qualitative _analog_ of a substratum. That is, you could remove every other property of an individual, and still be left with the same individual by virtue of its brute "thisness" property.

    Whether that is your understanding of thisness or not (and I see now it probably isn't), that is not what I take Adams' (or any reasonable person's) understanding of it to be. A thisness must necessarily attach to an individual that has properties. The properties are in part what define the individual after all -- but they are not enough to ensure particularity. All properties (properly construed) are general, but all individuals are particular. Thisnesses are thus what "particularize" the identity of individuals. They are, a priori, nothing but theoretical posits, but it's hard to see how they could not have metaphysical counterparts (otherwise, how do you secure particularity in nature itself -- he alternative is a view like Hacking's, where particularity is merely an artifact of description and theory).

    I have a lot more I want to say, but this message is already far too long.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)