Parfit’s “Argument from Below” vs. Johnston’s “Argument from Above”

First, a sort of apology: this post is a bit lengthy, and it may be a bit more “metaphysicsy” than the usual on PEA Soup, but it’s on a topic that has an important bearing on the ongoing debates over the role metaphysics, and in particular personal identity theory, has on our practical concerns, in particular our future-directed self-concern.  It has to do with an exchange between Derek Parfit and Mark Johnston, stemming from Johnston’s argument (in, among other places, “Human Concerns Without Superlative Selves,” from the Dancy collection Reading Parfit) that personal identity can still have non-derivative importance, even on Parfit’s reductionist view, according to which the facts about personal identity just consist in more particular facts about brains, bodies, and mental/physical continuity, and even where those more particular facts themselves don’t have non-derivative importance.  In “The Unimportance of Personal Identity,” Parfit replies to Johnston’s objection.  Here I simply want to try and track the dialectic between Johnston and Parfit, and I’ll explain why I think that Parfit’s response doesn’t adequately meet Johnston’s concern.  (Because I want to keep the post as short as possible, though, what I say here will be quite compressed and will assume knowledge of Parfitian reductionism, so readers not familiar with the basics of the debate may not want to read any further.)

Johnston first gets Parfit to admit that his reductionism is actually compatible with a further fact view, one according to which the facts of personal identity might still consist in facts that are distinct from facts about mental and physical continuity.  Since Parfit’s reductionism can’t be an analytic reductionism – it can’t be that talk of persons has the same truth value as talk of mental and physical continuity – it must be an ontological reductionism.  And of the various types of ontological reductionism – identifying, eliminative, and constitutive – Parfit’s is constitutive: so persons are wholly constituted by brains, bodies, and mental/physical events, although persons are still distinct from their brains, bodies, etc., so facts about persons may involve further facts.  On Johnston’s view, for instance, persons are essentially human beings, wholly constituted by brains and bodies, but such that the facts about human beings are further facts from facts about brains and bodies.  In this way, then, persons are like the statues in the famous statue/lump case.  Parfit thus adjusts his statement of reductionism slightly, not so that it involves no further facts simpliciter, but so that it involves no independent or separately obtaining facts, where the fact of personal identity “is not a further difference in what happens” (“The Unimportance of Personal Identity,” in Henry Harris, ed., Identity (Oxford: OUP, 2001), p. 19).

OK, so on to the main exchange.  In discussing the original teletransportation case, Parfit claims we shouldn’t care if the same brain is preserved between me and the person who walks out of the transporter on Mars, just as long as Relation R (psychological connectedness and/or continuity) obtains between us.  Parfit suggests that we can call the survivor me if we like, but he actually thinks the identity of the survivor is indeterminate.  Johnston disagrees, but suggests that, in trying to figure out what matters with respect to these cases, we need to consider both possibilities (indeterminacy and determinacy).  If the identity of the survivor is in fact indeterminate, Johnston argues that we should merely extend the scope of our ordinary self-concern to include our R-related successors, and not, as Parfit has argued, switch over our ordinary self-concern to track Relation R, even in the non-exotic cases.  But what if the identity of the survivor of teletransportation is determinate, i.e., he is not me?

Here Johnston gives his Argument from Above: if fact (a) consists entirely in facts (b) and (c), then even if (b) and (c) don’t matter, (a) might, in which case (b) and (c) will have only derivative importance (insofar as they constitute (a)).  We can contrast this with Parfit’s Argument from Below (which he reaffirms in “The Unimportance of Identity”): if fact (a) consists entirely in facts (b) and (c), it can only be (b) and (c) which matter.  Thus, if one has no non-derivative concern for (b) and (c), one should not have a non-derivative concern for (a) either.  Now the specific facts in the teletransportation case are as follows: if (a) the survivor of teletransportation isn’t me, that fact just consists in the facts that (b) he’s not physically continuous with me, and (c) the R that obtains won’t have its normal cause.  For Parfit, because we just don’t have a non-derivative concern for either (b) or (c) (and Johnston can agree that we don’t), then the fact that the survivor won’t be me shouldn’t matter as well.  Johnston, however, maintains that even if the lack of physical continuity and the normal cause of R don’t matter non-derivatively, that doesn’t yet mean that the fact the survivor won’t be me doesn’t matter non-derivatively.

Now this might seem just to be a standoff, but here Johnston gives two reasons why Parfit’s position is something of a disaster.  First, it yields the reductio of nihilism: given that all the facts of the world consist, ultimately, in facts about microphysical particles, and because we don’t have non-derivative concerns for those particles, then all values based on them shouldn’t matter either, which is absurd.  Second, it commits the fallacy of composition: just because the parts of which some fact consists have no non-derivative value, that doesn’t mean the thing they consist in has no such value.

Parfit’s reply is two-fold: he gives an example of the disagreement between Johnston and himself and assumes his take on the example is more plausible than Johnston’s, and then he tries to respond directly to the nihilism reductio.  I’ll here focus solely on the first response.  He gives the example of something’s being “alive.”  If “alive” is understood in a reductionist way, then the fact I’m still alive, despite having undergone irreversible brain damage, say, just consists in the fact that my heart and other organs are still functioning.  If we decide these lower level facts don’t matter, then the fact I’m still alive shouldn’t matter either.  On Johnston’s view, even though it may not be non-derivatively good that our hearts beat, it is good to be alive, and so we may be rational in hoping that our hearts continue to beat, given that their beating is what constitutes our being alive.

Parfit simply believes his argument is more plausible here, and his crucial point is this: if fact X consists entirely in more particular facts, then (1) X is not an independent or separately obtaining fact, and (2) X is also, in relation to those other facts, merely a conceptual fact.  It’s not a fact about our concepts, of course; rather, it’s a fact about the application of our concepts to some feature of the world.  The challenge, then, is this: why should a merely conceptual fact matter, when the lower level facts don’t?

Take, then, the beating heart case.  Suppose we agree that the fact your heart is still beating doesn’t matter, but we also hold that it does matter that you’re still alive. If we do that (and we agree that the fact of your being alive just consists in the fact that your heart is beating), then “we’re treating language as more important than reality” (p. 32), and that can’t be plausible.

Nevertheless, it’s not at all clear to me that (2) follows, that, if X consists entirely in more particular facts, X is, therefore, just a conceptual fact.  Consider the statue/lump case, for example.  The statue is wholly constituted by the lump of clay, so facts about the statue consist entirely in facts about the lump.  Now surely the lump doesn’t matter non-derivatively.  Nevertheless, (a) facts about the statue aren’t just conceptual facts, it seems, and (b) it’s certainly rational for me to care non-derivatively about the statue.

So it seems that not all wholly constituted facts are conceptual facts, and this seems to be the real source of the disagreement between Johnston and Parfit: are the facts about identity more like the facts about “alive” or more like the facts about statues?  If there are different kinds of wholly constituted facts (some conceptual, some not), then Parfit has to explain why “identity” is merely conceptual.  And if all wholly constituted facts are conceptual, we need an argument to that effect.  At this point, then, it looks as if Johnston has the better of the exchange.  Well, he does, that is, unless someone can show me that (a) wholly constituted facts are merely conceptual, and (b) there’s some argument to that end that I’ve missed.  So please enlighten me.

37 Replies to “Parfit’s “Argument from Below” vs. Johnston’s “Argument from Above”

  1. Ok – this is going to be unfortunately a vague attempt to try to explain what might be going on at least in my mind. I’m not altogether sure how useful ‘the fact’ talk here is or how much weight it can carry.
    Anyway, one way of looking at the issue is to start, instead of the facts on different ontological levels, from the logical criteria for applying the terms, that is, the norms of their correct use. From this perspective it is easier to see how some terms are more basic than others. The criteria for applying psychological and physical terms, on which ascribing relations like continuity and R are based, seem to be something quite direct. We can point to the parts of the body at different times and observe or hear accounts of the mental states, and track whether they go on being the same.
    In this respect, the term of identity seems different and more complex, like a cluster term. There doesn’t seem to be similar directly observable criteria for saying when one person is the same one at a different time like there are for her features. We then need to consider to what degree each of the more basic criteria apply, and whether given the sum of that we are warranted to make the claim of person sameness.
    I always thought that something like this must be behind Parfit’s view. The norms of use of ‘sameness’ that exist conventionally in our linguistic community then only guide us in the normal circumstances – to the kind of instances and aggregates of the basic continuities for which we need to use the notion in ordinary life. In the kind of cases we meet in philosophical literature, these norms just do not have definative answers on the basis of how the more basic terms apply as a whole. This seems like an instance of general conceptual vagueness.
    The criteria for applying psychological and physical continuity do serve us still well in the hard cases. So, the only thing for us is to say that it is indeterminate, by the norms of use, whether the terms apply here and maybe go on to propose a sharpening of the terms.
    So, for me, this looks like a good way to understand the case. If we are deflationists about truth and fact talk, then I guess this would make the higher-order facts conceptual. At least the truth of the identity claims depends on the conceptual criteria given by the applibicality of the lower-level terms. I’m not sure what *mere* there is to this. At least it avoids adding many metaphysically robust facts in the world. What about the concerns? I don’t see why we couldn’t have concerns that track the applibicality of the terms on both levels, that is, why we couldn’t care about both identity *and* the psychological and physical continuity. (Maybe some care about only the other, and some about neither.) If the story above is correct, then this would just mean that our continuity concerns have definative targets in the hard cases whereas the identity concerns are left ‘blind’. In the easy cases, where the logical criteria in terms of bundles of the continuity terms for sameness are either clearly satisfied or unsatisfied, then the concerns would both have a target.
    I hope this makes at least some sense.

  2. David,
    This is a really interesting post. Maybe you can help me with what looks like an inconsistency. I probably missed something important.
    You ascribe to Parfit a form of ontological reductionism that is non-eliminitivist. You say this,
    “Parfit’s is constitutive: so persons are wholly constituted by brains, bodies, and mental/physical events, although persons are still distinct from their brains, bodies, etc., so facts about persons may involve further facts.”
    But then we find (one of) Parfit’s complaint(s) to Johnston described in this way,
    “if fact X consists entirely in more particular facts, then (1) X is not an independent or separately obtaining fact, and (2) X is also, in relation to those other facts, merely a conceptual fact.”
    So here’s my question. If the constitutive facts about persons do not preclude persons from being distinct (ontologically, I take it, and not simply as a matter of language use) from what constitutes them, then why do constitutive facts about X (in this case, about being alive) preclude the fact that I am alive from being ontologically distinct (rather than merely conceptually distinct) from what constitutes that fact? What did I miss here?

  3. Jussi: I’m not sure I follow you entirely, but perhaps you can answer me this: on your view, is there any difference between the “alive” case and the “statue” case? That is, is your analysis stricly applicable just to the identity case (after all, you’ve specified everything in terms of psychological/physical continuity), or is it more generally applicable to other so-called constitutive relations?
    Mike (and if any other Mikes get involved, I’m switchin’ to last names): if I understand you correctly, I think you’ve put your finger on precisely the problem (and so you’ve put my point in a much more succinct fashion). It is as if Parfit is moving back and forth between two forms of reductionism, constitutive and eliminative (although it’s not clear to me that eliminative reductionism would be such that facts about the supervenient properties would be “merely conceptual” facts).
    Another way to put my final question in the post, then, would be this: can ontological distinctness, where that further fact isn’t separately obtaining, still involve something more than merely conceptual facts (facts about the application of our concepts)? (Surely the answer is yes, in which case Parfit’s reply fails.)

  4. David, that looks to me like a pretty serious objection to Parfit, if he is equivocating on forms of reduction. I’ve got to go back to the Johnston paper in the Reading Parfit collection.

  5. David,
    sorry about being unclear. That’s what happens when I do metaphysics and philosophy of language. It was meant to be a general account. Surely we have a lot of concepts where the application of the given term depends on to what degree a group of other concepts apply – say, well-being for instance. The statue case is different because we are talking about one particular case – the statue. If we talk which lumps are in general statues, then we might need to go through various terms and check to which lumps they apply. But, it makes no sense to ask of many lumps which ones are certain one statue. Maybe it actually does in modal cases. We can imagine changes in the lump that constitutes the statue and ask in which conditions that lump would still be the same statue. Then we do need to use various criteria in the thought-experiments.
    Sorry, there has to be something wrong with the question starting ‘ontological distinctness, where the further fact isn’t separately obtaining’. The former seems to imply that there is a further separately obtaining fact. If x is ontologically distinct from y, then some further fact than y’s existence must obtain, no?
    Also, I’m starting to get very uneasy about the conceptual fact business. Bachelorhood is nothing above being a man and unmarried (well, it is a bit put let it pass). Thus, there even is a reductive analysis of the term, so that bachelors don’t add anything to our ontology. But, still, me being a bachelor is not a conceptual fact about the application of terms. It’s a fact about me.

  6. Jussi’s last comment seems to the point here. Just because the limits of some concept C make it the case that x, y and z constitute one way of being C that doesn’t seem enough to argue that the importance of being C is reducible to the importance of x, y and z considered apart from their connection to C. It may be that Parfit’s thought standing behind the conceptual fact move is that if a fact is conceptual, then it is stipulated or arbitrary. (I’d have to carefully read the Parfit to see if this suggestion is a correct statement of the motivation.)
    You might think that this doesn’t follow from a fact’s being conceptual. For example if C is a natural kind, there may be facts about the most natural way to divide C from not-C which make it the case that something which is x, y and z counts as C. If so, counting something with x, y and z as C will not be arbitrary. And the conceptual nature of such facts would not show they did not matter.

  7. I don’t see the worry with conceptual facts. Certainly, to take a cliched example, we have the conceptual distinction between the morning star and the evening star. And presumably the following propositions differ only conceptually,
    1. The morning star is pleasing to observe.
    2. The evening star is pleasing to observe.
    What’s more, as far as I can see, it can make perfect sense to affirm (1) and deny (2). So the conceptual difference can yield a perfectly intelligible difference in value–in this case aesthetic value.

  8. Mike: As I suggested in my previous response, I’m not yet sure if the charge of equivocation is correct. If Parfit is indeed appealing to another (non-constitutive) form of reductionism here (e.g., identifying or eliminative), it would have to be one in which all the facts about persons are merely conceptual facts, and I’m not sure either of the alternatives is necessarily committed to that view.
    Jussi: You wrote:
    “Sorry, there has to be something wrong with the question starting ‘ontological distinctness, where the further fact isn’t separately obtaining’. The former seems to imply that there is a further separately obtaining fact. If x is ontologically distinct from y, then some further fact than y’s existence must obtain, no?”
    “Separately obtaining,” for Parfit, refers to facts about persons qua separately existing entities, or facts about persons describing something additional that happens *in the world*, beyond just facts about brains, bodies, and interrelated mental and/or physical events. If Johnston’s right, then, there can be further facts about persons — which will be facts about human beings — beyond those referring to brains and bodies, but that (a)aren’t facts about “separately existing entities,” yet (b) are also not mere conceptual facts. If Parfit’s right, you can have further facts that aren’t about separately existing entities, but then those further facts are merely conceptual, or you can have further facts that aren’t merely conceptual, but then those further facts are about separately existing entities; you can’t, though, get both (a) and (b).
    Another example might help here. Parfit talks about viewing a grouping of trees on a hill, and then finding out that, because there’s a grouping of trees on the hill, there’s a “copse” on that hill. He’s now learned something new, but what he’s learned isn’t about the world (about “reality”); it’s instead about our concepts: he’s learned that the concept “copse” applies to that grouping of trees. But of course copses are wholly constituted by individual trees. So we are constitutive reductionists about copses, allowing that there can be further facts about copses but where these further facts are merely conceptual (they aren’t about separately existing “copse” entities).
    My question, though, is this: why think that all further facts on constitutive reductionism are merely conceptual? Johnston’s example of human beings is enlightening: as a human being, I may be wholly constituted by my brain and body. Nevertheless, there are some facts about me that are further facts beyond facts about my brain and body. For one thing, if my body were destroyed, I could still survive if my brain were transplanted to another body (so the facts about my survival don’t consist entirely in facts about my current body, or facts about my current brain & body combination). And facts about me aren’t simply facts about my brain. For example, I don’t weigh 6 pounds. But these facts about me don’t seem merely conceptual: they aren’t about the application of our concepts to reality; instead, they’re directly about reality. When you find out how much I weigh, for instance, you’re not just finding out how to apply the concept “person.”
    I hope that this helps a bit. And about the bachelor comment: we’re *analytic* reductionists about bachelorhood, so the facts about bachelors have the same truth value as facts about unmarried men (suitably construed). But that can’t be the type of reductionism Parfit has in mind (as Johnston points out).

  9. David,
    here’s an intuition test case which I think does a pretty good job in deciding between Johnston and Parfit here. It comes from Chalmers and Jackson. Say that all you knew were the complete description of the universe on the basic particle level and you had all the possible semantic knowledge there is to be had (i.e., you knew to which kind of particle aggregates molecule terms apply, to which kind of molecule aggregates different substance terms apply, to which kind of substance aggregates different artifact terms applied, and so on). Now, if the Parfitian was right given all this knowledge you could infer whether a at t1 and b at t2 were an identical person (Parfit’s own view would be that in most cases you would even though in some hard cases the semantic knowledge would leave the issue undeterminate). If Johnston was right, then you would need some further sui generis knowledge of constitution relations that are not included in the two sorts of knowledge we have. My intuition here is with the Parfitians. For one I don’t really see a plausible epistemology for a direct knowledge about constitution relations – semantic knowledge is much easier to come by.
    I do still feel that there is something deeply wrong with your construal of conceptual facts. If some question is a matter of whether a certain term applies this cannot be sufficient for making the correct answer a merely conceptual matter. For one, this would mean that all truths are conceptual truths. And second, surely the terms must already come with truth-conditions and the world has to co-operate on the question whether those conditions are satisfied. Even in the copse case – the trees could have been otherwise but now they happen to be contingently in some way – this is what we learn from the world and it is not mere conceptual knowledge.

  10. Jussi: Oddly enough, I don’t think Parfit holds the sweeping view (anymore) that you suggest he would hold. Let me quote him at length (from “The Unimportance of Identity,” p. 33):
    “It is not true… that, if we knew how the particles moved in some person’s body, and understood our concepts, we would thereby know, or be able to work out, all of the relevant facts about this person. To understand the world around us, we need more than physics and a knowledge of our own language.”
    He’s simply maintaining that there are some cases where, “relative to the facts at some lower level, the higher-level fact is . . . merely conceptual.” These merely conceptual facts can’t be morally or rationally (non-derivatively) important, then, if the lower level facts aren’t (non-derivatively) important.
    So the upshot is that he’s now accepting, along with Johnston, that there are some further facts *about persons*, but he’s denying that these further facts are anything more than merely conceptual, and so denying that they are rationally or morally important.
    Second, this ain’t *my* construal of conceptual facts! It’s Parfit’s. But let me see if I can explain it further by responding to your questions. First, why think that this construal would render all truths conceptual truths? Suppose it’s true that “there are 11 trees within that 2500 square foot patch of ground.” Surely that’s not (or not merely) a conceptual truth. But now consider the statement “the previously referred-to 11 trees together constitute a copse.” If Parfit’s right, then the truth of the second sentence, yes, is a merely conceptual truth, insofar as what makes it true is *solely* whether or not our concept “copse” correctly applies. The truth of the former claim depends crucially on whether or not there *are* 11 trees grouped together in the referred to patch (as well as, perhaps, whether or not various concepts are properly applied), i.e., it depends on what actually happens in the world.
    Second, then, even if these concepts come with truth-conditions attached, what the world “does” is irrelevant to whether or not those conditions apply; the only thing that matters is certain conditions about our language.
    Think of it in terms of the types of information produced by certain facts. Some facts give you information about the world (and perhaps about our concepts), whereas some facts give you information solely about our language. Facts about the trees themselves, their proximity to one another, and so forth give you information about the world. The fact that some grouping of trees is a copse, however, gives you information *only* about our language. It’s thus a merely conceptual fact.

  11. “Facts about the trees themselves, their proximity to one another, and so forth give you information about the world. The fact that some grouping of trees is a copse, however, gives you information *only* about our language. It’s thus a merely conceptual fact.”
    David,
    It looks like what you want to say is that whether or not something is a conceptual fact is relative to some received ontology. Isn’t that what you have in mind? Imagine an ontology that countenanced things like copses but not trees. Relative to such an ontology “facts about of trees” might be considered purely conceptual. We might say, relative to that ontology, there really aren’t any trees. The concept of a tree is a linguistic invention, an abstraction from copses. What are really there are copses.
    You seem to be making this point in saying, on the contrary, that the trees have some ontological status (they’re in the accepted ontology) and the copse-talk is conceptual.

  12. That Parfit quote is interesting, but I’m not sure that its purpose is to show that identity would be over and above the physical. I suspect that the point is pretty much like Chalmers – given the zombie possibility the physical and semantic knowledge leaves out the facts about conscious states. And, they seem to be important in assessing psychological continuity. I guess a better way of putting the question is if you knew all physical facts in the history, all facts about concsious states and all the semantic knowledge would you be able to deduce how different persons endured in that history of the world or would you need some further sui generis constitution or identity knowledge. If you don’t think that the physical, semantic, phenomenological knowledge is enough for identity knowledge, then you need an account of what the further knowledge is and how we can be in touch with it.
    And, sorry, the more you explain Parfit’s view of conceptual facts the more incredible it sounds.
    1. I can’t see why the argument for that when you face a copse you come to know a conceptual fact – learn only about application of terms – cannot be then be run to trees. When you face a trunk and some leafs, you learn that the word ‘tree’ applies to them, and what you thus learned is about concepts not trees. And, we can run this for trunks and leafs. We can even run it for protons – we learn to call the maker of the vapor trail a ‘proton’, and thus we have learned only about the application of words. But, now we’ve lost all the empirical knowledge. That has to be a reductio for Parfit’s view of conceptual facts.
    2. “Second, then, even if these concepts come with truth-conditions attached, what the world “does” is irrelevant to whether or not those conditions apply; the only thing that matters is certain conditions about our language.” This does not make much sense to me. Here’s a problem. If how the world is does not make contribution to the truth-conditions of ‘there’s a copse’ being satisfied, then it would be true whatever way the world was. Thus, it would be necessarily true. But surely that’s a contingent claim that either is true or false depending on how the world is. The conditions about our language would remain the same in the case that there were not trees but only painted cardboard boxes in front of me. So, the claim would be true in either case with real or fake trees – I would only learn the same fact about application of our terms. Surely, no-one believes this. I want to be so charitable to Parfit that not even he can hold this view.
    3. I think we are also speaking past each other about facts – has happened to me here earlier. I can’t make sense of some facts being about the reality, about whatever that is and some facts being about language. Facts are what the world is, and true sentences and propositions describe those facts. Some of the facts in the world are just how we use the language and some just how the trees are in the world.

  13. Mike: You’re way of putting it (“that whether or not something is a conceptual fact is relative to some received ontology”) sounds right to me, for the reasons you’ve given. Jussi, does that help?

  14. David,
    unfortunately not really. I just cannot square contingently true claims like ‘that’s a copse’ with the idea that these are conceptual facts, after all in most cases we know the application criteria for our terms a priori. But thanks for all the patience with me and good attempts to help. I think I need to go to the original sources and see what they really say about conceptual facts. Maybe there is a paper to be written here that they are both conceptually confused about conceptual facts : )
    I’ve started to think that a good way of catching the Johnston view would be some sort of anti-realist minimalism a la Crispin Wright. On this view you could say that facts about identity are not reducible to mere physical, psychological and semantic facts. Rather, there are epistemic norms of warranted assertibility for identity claims – for when it is ok to say that ‘these and these features constitute the enduring person’ or ‘he’s the same guy we met at last year’s APA’. But, then with truth minimalism, there is no more to truth or facthood of such claims than that the identity discourse internal epistemic norms are satisfied. That is we have warranting evidence to make the claims on basis of physical and psychological continuity. The view would still be metaphysically lightweight. We get further identity and person facts but not FACTS or distinct existences.

  15. Jussi,
    What exactly is the source of confusion here? Maybe it is the word ‘conceptual fact’; maybe that’s an unhelpful way to label what is being described. But the idea is relatively simple. All that is being assumed–as far as I can see–is an ontological eliminativist position. There are any number of them. Take Merrick’s eliminativist position. On Merrick’s view the ‘folk ontology’ of chairs, tables, statues, stars, rocks, etc, is radically mistaken. There *literally* are no such things. But obviously, it does not follow that we cannot use the language of folk ontology–we could hardly get by without it. All we need to recognize is that when we think we are talking about and seeing macrophysical objects like chairs, we really aren’t: in fact we are seeing (I mean, under perfectly normal conditions, no hallucinations, perceptual distortions, etc.) “things arranged chairwise”.
    Whatever one thinks of that ontology, it makes perfect sense to say that facts about chairs are (perhaps for lack of a better term) conceptual facts. There literally are no chairs. Those facts–facts we all assent to such as “chairs are for sitting in” and “some chairs are comfortable”–are not about anything that literally exists. This follows directly from eliminativism about folk ontology.
    This is the sort of view, I think, that David has been describing. I don’t see any source of incoherence in it.

  16. Sorry I haven’t been more helpful, Jussi. I’ve been wondering as well if, as you suggested previously, we aren’t just talking past one another in our “fact” talk.
    At any rate, I don’t think Johnston would have a huge problem with your most recent suggestion, although he’s not interested in keeping things “metaphysically lightweight” as much as he is in discounting the importance of the metaphysics in justifying our patterns of self-concern. Indeed, his main point in that article is to defend the view that self-concern — being specially concerned for the future person that is *myself* (as opposed to my R-continuer) — is reasonable independently of the metaphysics of identity, in part insofar as it stands in a coherence relation with other of our concerns and commitments.
    Mark (sorry to have missed your earlier comment!): I doubt Parfit thinks it’s merely a matter of stipulation or aribtrariness that renders conceptual fact C unimportant. It’s just that if x, y, and z wholly constitute C, and x, y, and z are facts about the world, whereas C is just a fact about the application of some concept to x, y, and z, then (according to Parfit), if x, y, and z don’t matter, neither could C. If the facts about the world don’t matter, and C is just a fact about our way of *talking* about the world, and not a further difference in what happens in the world, then C can’t rationally or morally matter either.

  17. David,
    The last sentence of your reaction to my earlier comment seems to me to capture the idea I was suggesting might stand behind Parfit’s division (emphasis on ‘might’ since I haven’t gone to look at the relevant Parfit):
    If the facts about the world don’t matter, and C is just a fact about our way of *talking* about the world, and not a further difference in what happens in the world, then C can’t rationally or morally matter either.
    I guess I was thinking that the idea lurking in the background is precisely that some facts are underwritten by the world whereas others are just underwritten by ways of talking. My thought was that facts of the latter sort were supposed to be thought of as less important because how we talk is arbitrary; we could have chosen to talk some other way and if we had we would have made no mistake. But in that case (the train of thought continues) such people would be making no mistake to ignore the “conceptual facts” in making moral judgements.
    I don’t endorse this way of thinking (the way I’m suggesting might be what Parfit is thinking) largely because I don’t endorse this way of thinking about facts as based on ways of talking as opposed to how the world is. For instance, sometimes the ways we talk are as they are because there is a natural division of the constituent level facts which explains why we count some of them as constituting the higher level fact and don’t count others. (I actually think this is one upshot of some of Quine’s more famous arguments.) In other words, if we talk the way we do because there are such natural divisions, it is not all that obvious that these same divisions might matter morally.
    On Johnston and minimalism, I don’t think that the sort of minimalism that Johnston would favor would wind up being anti-realist.

  18. Mark,
    I don’t have any problems with ontological reductivisms as such, and maybe not even conceptual facts as you characterised them. Well, at least, I accept the negative claim you make that claims on a topic of ontological reductivism are not about anything that literally exists. Usually though ontological reductivism leads to either error theory or semantic, analytic reductivism. Either our claims attempt to be something in the world and it doesn’t deliver, or they are about, contrary to the appearances, about the lower-level things. But, now a third alternative is given for our ontological reductivist – these claims are not about what they appear to be about but rather about the application of the terms. It is this last positive claim I cannot make sense of. My claim ‘that’s a chair’ would not be about the ‘folk chair’ which does not exist, not about the particles arranged chairwise, but instead about in which circumstances I’m allowed to use the word ‘chair’. And that seems to get things completely wrong. But, now I’m table pounding…
    Mark’s second last paragraph pretty much sums up my second big worry, so a big ‘thank you’!

  19. Sorry, should have added a third alternative for the ontological reductivist – fictionalism, which seems to come closest to Mark’s description of the chair situation. We can continue to entertain chair talk and chair fact talk even though there are no chairs. When I say ‘chair’ I talk of something like an unicorn but it’s ok cause it’s useful. Even now this fictional fact wouldn’t be about application of terms so I’m still not seeing the sense in which it is conceptual. Nor am I seeing the way in which this is reductionism. Nothing is reduced to something else. Chairs are just eliminated from our ontology. Come to think of it error theory does the same.

  20. Mike: Well, neither Johnston’s nor Parfit’s (recent) view is ontological eliminativism (although Parfit’s earlier view in Reasons & Persons probably was). They’re both now advancing ontological constitutive reductionism (whew!), according to which there are such things as persons, and there are facts about persons that are “further” facts than those just about brains, bodies, and mental/physical continuity, but nevertheless persons are wholly constituted by brains and bodies. This is akin to statues being wholly constituted by their lumps of clay, but nevertheless there are facts about the statue (when it came into existence, say) that are further facts than facts about the lump.
    Jussi, I assume you were talking to *Mike*, not Mark, in your more substantive responses?

  21. Jussi,
    I’m assuming ‘Mark’ refers to Mike. Or else I have been reductively eliminated. I think you’re right that what might be called ‘folk fictionalism’ goes nicely with reductive eliminativism. Perhaps you don’t like the use of ‘reductive’ in contexts of eliminativism, but it is standard usage; no need to court confusion with yet more terms.
    On a different point, I’m not sure what the following means or why it is necessary to convey the worries that Parfit has,
    “But, now a third alternative is given for our ontological reductivist – these claims are not about what they appear to be about but rather about the application of the terms.”
    The claims need not be *about* the application of terms at all. I don’t see why Parfit’s point could not be made without claiming that. Rather the claims *involve* a semantic decision to apply certain linguisitic items to certain macrophysical “objects” even after we learn that such objects don’t exist. But this part of the discussion seems to quibble. To get at Parfit’s point about value, reductive eliminativism gives a good motivation.

  22. David, thanks. The constitutive claim does make Parfit’s point more difficult to establish. Further, it makes claims about the reduced, but not eliminated object (statue) no more conceptual it seems than claims about its constitution (the clay). Absent elimination, there really are two objects, not one, of course. So once I was found and now I am lost!

  23. Mike, (hope I got this right this time, sorry everyone!).
    thanks. Maybe I was reading too much to this passage from David:
    ‘Some facts give you information about the world (and perhaps about our concepts), whereas some facts give you information solely about our language. Facts about the trees themselves, their proximity to one another, and so forth give you information about the world. The fact that some grouping of trees is a copse, however, gives you information *only* about our language. It’s thus a merely conceptual fact.’
    But, you are right Parfit need not make claims like this. But there again is a wrinkle. Say we thought this as you suggest:
    ‘Rather the claims *involve* a semantic decision to apply certain linguisitic items to certain macrophysical “objects” even after we learn that such objects don’t exist.’
    First, I have my Quinean worries about the coherence of thinking that it’s ok to say that I saw a fine copse yesterday while denying that copses exist. I would be quantifying there over copses whilst denying that there are such things. But, let that pass.
    The bigger worry is that I’m not sure we can under this understanding of conceptual facts construct the difference between Parfit and Johnston on teh original issue. Think of the constitution relation itself. I take it that we do not want to take these relations to be ontologically robust (for one it is a relation between robust things and things that do not constitute further independent facts). Then, I’m not sure it would make sense to talk about constitution relations being constituted out of something (about what?) and also this would threaten regress anyway. So, what we end up needing to say that it is a semantic decision to to apply the linguistic item ‘constitution relation’ when we say that ‘such and such things constitute the enduring personal identity’. But, if that’s true, then the fact which was supposed to be a constitutive has now collapsed into a conceptual one – the Parfit position. And, now we have to Parfitian views a simple one and a complicated one but both agree that what identity comes down to is conceptual facts.

  24. Jussi/David,
    I think (as David mentioned) that we cannot quite attribute to Parfit the eliminativist position. But–now having read the Johnston paper–it seems to be something close to this.
    The constitutive relation between persons and brains/bodies is not one on which (1) persons are eliminatively reduced to brains/bodies NOR is it (2) one on which there is non-identity with “superlative further facts”. It is rather (3) one on which there is non-identity with “ordinary further facts”.
    But what is the difference between (2) and (3)? It is true under both (2) and (3) that you can survive even if all or most of your body/brain do not. Under (2) there is something more to us than our bodies/brains (e.g., there is a soul or mental substance or Cartesian ego or something). This (say) Cartesian ego is the superlative further fact. That’s how we survive, the mental substance continues on or something. Under (3) there is never anything more to us than our bodies and brains. We survive, on this account, if a physically continuous (though not necessarily identical) descendent survives.
    The person on this third account is an additional thing but not a superlative additional thing. Just as in the case of the statue and the lump of clay: again the statue is not a superlative additional thing. And this is what is a little unclear to me. This non-superlative additional thing is supposed to be of a different ontological category (i.e., both the person and the statue are of this different ontological category) from that in which it consists. So there are two things, but not two things of the same ontological category, and one of those is consists in the other.

  25. Mike,
    thanks for the clarfications. At some point these debates always remind me of the medieval logicians’ task of making sense of three being one…
    I wonder if one way of making sense of the option (3) would be to start from the idea that what is essential for enduring personhood is psychological continuity – having the same mental states. This doesn’t yet commit us to any sort superlative further facts – you don’t need cartesian minds for mental states. Yet they are on physical level multibly realisable. Maybe the thought is that my brains could bit by bit be replaced by silicon based stuff, that could be planted to a new body, and yet still the thing is me due to the continuity of the mental. In this way the endurance conditions of me as a person and me as a human animal come apart (and there are modal differences too). In that sense we are distinct but there are no two things in any more robust sense.

  26. Maybe this will help if filling out option 3:
    The argument for non-identity with respect to statues and clay is fairly straightforward. They have different modal properties so they must be distinct by Leibniz’s law. (The clay would survive smashing, the statue would not.)
    Similarly, modal differences between persons and their constituents will also generate an argument for non-identity. With persons and their brains and bodies, I’m not any longer sure what the relevant modal differences are according to Johnston, but the stuff at the end of Jussi’s post above are of the right sort to do the work.

  27. Mark’s right about the modal differences. With persons, those differences are that I would survive without my body (by having my brain transferred into another body, say), so I’m not identical to my body (nor am I identical to the combination of my brain and body), but I would weigh more than my brain (and so am not identical to my brain). (This is the way Johnston runs the differences.)

  28. And so (I should finish) facts about my survival (qua human being/person) are further facts than facts about just the survival of my body, or about the survival of my brain and body, or about the survival of just my brain, even though I (qua human being/person) am wholly constituted by my brain and body. “Persons” and “brains/bodies” are from two different ontological categories.

  29. Mark,
    This argument is similar to the one Gibbard ran against Kripke on contingent identity claims. You write,
    “The argument for non-identity with respect to statues and clay is fairly straightforward. They have different modal properties so they must be distinct by Leibniz’s law”.
    Suppose someone claims that the clay is in fact identical to the statue. Kripke’s claim (we know) is that if the terms flanking the identity sign in identity claims are rigid designators, then the identity claim is necessarily true (though perhaps not a priori knowable). So if the clay and the statue are (rigidly) named Ted and Alice, then, since Ted = Alice (and, indeed, necessarily so) it looks like LL does not cover modal properties.
    Why wouldn’t LL cover modal properties?
    Take Bernard Ortcutt and the man in the brown hat. They also have different modal properties. The man in the brown hat has the property of being believed to be a spy and the dispositional property of evoking suspicion in Jones and others. B. Ortcutt does not have the modal property of being believed to be a spy or the dispositional property of evoking suspicion. But certainly Ortcutt = the man in the brown hat. The modal properties are different, but that does not entail that the identity fails.

  30. Mike,
    the last case seems different because the modal properties are different propositional attitudes we have towards the same being under different descriptions. It’s well known that for those the Leibniz’s law doesn’t hold – a case of the Cartesian Masked Man Fallacy. De re necessities and possibilities are usually thought to be different in this respect.
    In the statue/lump case I wonder why we would, if we had the two object view, be pushed to say that Ted=Alice. I mean we could name the lump before its making to a statue Tim and the statue once its formed Alice. Now, Tim and Alice can rigidly designate distinct things as reflected by their different modal properties. Tim existed before Alice and Alice’s destruction does not imply Tim’s destruction. Maybe you need some sort of four-dimensionalism for this but seems like a live option.

  31. My central point is that modal properties (narrow them to de re modal properties, if you like) don’t seem to be covered by LL. Gibbard proposed–contra Kripke–that Lumpl and Goliath (his names for the clay and the statue that now coincide spatiotemporally) are identical *but might not have been*.
    Certainly a desperate Kripkean might insist that these are not identical on the basis of differing dispositional or counterfactual properties. But the more reasonable position, I think, is to deny that those modal properties are covered by LL and allow for Gibbard’s sort of contingent identity.

  32. I think we can give a straight counterexample to Parfit.
    Suppose Theseus is going on a long trip and he asks you to take care of his ship while he’s gone. He’ll give you a thousand minas if the ship is still floating when he gets back. He’ll count it as the same ship if it’s continuous with the present ship; you can replace any part but not substitute a whole ship at once.
    Well, the ship is pretty rickety and you have to replace a lot of parts over the next few years, a little at a time of course. But at each replacement, it seems to me that
    (a) the ship consists in (is constituted by) nothing but its timbers and their relations;
    (b) you do not care about the continuation of any given timber, or any given relation between timbers;
    (c) you do care, and have reason to care, about the continuation of the ship.
    It is also true that you have reason to care about a disjunction of timber-continuations, since you can’t substitute them all at once; but that doesn’t help Parfit, since the ship is not constituted by any of the disjunctions you have reason to care about.
    Isn’t that a case where Parfit’s view fails, and Johnston’s view succeeds?

  33. Heath,
    I’m not sure. Parfit might say that Theseus has no grounds for caring whether it is *the* ship he sees when he returns. We are not even sure of the criteria here. If you do enough of the changes at a time you might think that it’s a new ship with parts from the whole and in many cases the answers here may be vague and indeterminate. What Theseus should be worried about is that there is some ship that resembles the one he left you with. What you might have to worry about is some ‘floatability-making’ and ‘cargo-carrying’ features that are physical. But sometimes you can retain these while loosing the same-ship-making features. That might be a bit of a stretch but at least it goes somewhere towards an answer. What would be the point of caring about sameness and whether some such criteria are satisfied?

  34. Mike,
    (Apparently some people use Leibniz’s law for the identity of indiscernibles, so just to be clear I should say I am using it as a name for the indiscernibility of identicals.)
    I’m actually sort of a stark raving fanatic about the indiscernibility of identicals insofar as I think that <1>properly understood it applies even to intentional properties. The general idea is that you can think of the same thing in two ways and not know that it is the same thing. Thus, you might truly say of Orcutt that he has the property of being suspected by you to be a spy when thought of as the man in the brown hat. And the man in the Brown hat has that property too. You can then explain why most people count “I think that Orcutt is a spy,” as false either because it suggests as implicature that when you think of him as Orcutt you think of him as a spy. Or you could build that into the semantics of “I think that Orcutt is a spy.” But if that is your preferred semantics for attitude attributions, the property attributed when you substitute ‘the man in the brown hat’ for ‘Orcutt’ will be a different property. So I don’t think that even intentional properties require us to reject the indiscernibility of identicals. Obviously this is controversial and there are long debates in the literature about it. But my saying this is not original with me or especially idiosyncratic.
    In any case, not every modal property is an intentional property. And the ones in question with the statue and the clay were not intentional. So we would need a further argument that all modal properties are not covered by LL. (And I think that limiting the case to de re necessity actually makes the case for LL stronger.) Again, the issue is controversial, and you rightly note that the Gibbard contingent identity paper argues that one could coherently think other modal properties too were exempt from Leibniz’s Law/the indiscernibility of identicals. But it isn’t Johnston’s view (he has a paper on this in Phil Studies in the early 1990s). And I don’t think you have to be especially desperate to deny contingent identity. For purposes of the dialectic here, it looks like Parfit has agreed with Johnston’s position on the issue.

  35. Mark, you write,
    “The general idea is that you can think of the same thing in two ways and not know that it is the same thing. Thus, you might truly say of Orcutt that he has the property of being suspected by you to be a spy when thought of as the man in the brown hat. And the man in the Brown hat has that property too.”
    But this seems to invite some well-known Quinean problems with opacity. I might know that the man in the brown hat murdered Sue and not know (or have any clue or any evidence) that I am the man in the brown hat. Certainly I could fail to know that. But denying that I know this can’t be explained away by appeal to implicature. At most what could be said is that, given some reasonable epistemic closure principle and given the knowledge of the identity claim, I could come to know that it was I who killed Sue.
    I agree that Parfit does not appeal to contingent identity in his discussion. But appealing to further facts in a different ontological category seems to me even less plausible than contingent identity. I’m not so certain what the ontological categories are, are you? I’m not sure what the term means other than “it is a different kind of thing altogether” and that’s unhelpful. The only relevant examples that come to mind are Ryle’s alleged instances. So I’m not quite sure what Parfit is referring to there.

  36. Mike,
    I guess I think that there are many well-developed treatments of opacity post-Quine that handle the problems relatively well and which don’t force us to give up LL, but there are plenty of people better placed to argue for that than I. One thing to note though, is that not every approach involves using implicature to do the needed work, though I suppose the ones I tend to favor do use implicature.
    On the stuff about categories, I think you don’t need to have a story about categories to use the test which invokes LL, and in fact you can use the test to figure out when things are of different categories if you have some need for categories. Just ask, would A survive such and such? Would B survive such and such. If you get different answers A and B are not identical. That will be true even if A and B are both in the same category. If they are in the same place at the same time, you might think that the fact that they are not the same thing is best explained by their being in different categories of thing. But I suppose you could deny that if you did not like categories and still not give up LL.

Comments are closed.