29 July 2008

In Which a Problem remains for McDowell, but I satisfy Myself that McDowell's older Position can Avoid the Difficulty

Looking at my blog now, there is way too much text. My last two McDowell posts are pretty much just two dense walls of obscure prose. To try to improve matters, I'm sticking this post behind an animated gif of Yuyuko eating rice. It's a follow-up post for my follow-up post on my earlier post on "Avoiding The Myth of the Given". Yuyuko is a video game character, and the rice is from Hidamari Sketch. Hitting the "Escape" key will stop gifs from looping in Firefox. (It's actually the same as hitting "Stop"; it's just that Firefox defaults to graying out the stop icon on pages that are fully loaded.)

Oh, and incidentally, "The Dark Knight" is still fantastic on a second viewing, and on a third. It's also worth seeing in IMAX if you can manage it; the city-shots are gorgeous.

In my last McDowell post, I encountered a dilemma: Even if one does not follow McDowell in limiting the conceptual capacities which can be involved in the content of an intuition to "common and proper sensibles", but allows that any conceptual capacities a thinker has might be involved in the content of an intuition given to that thinker, there seemed to be a problem about noninferential judgements based on recollected intuitional content. "I might see a pink ice cube on Tuesday, learn to recognize what ice cubes look like on Wednesday, and then, on Thursday, noninferentially gain knowledge that the thing I saw on Tuesday was made of ice (because I remember what it looked like, and I now have the ability to recognize cubes made of ice as being made of ice when I see them)."

I think that my problem was that I was working with a false picture of how "intuitional content" can be "recollected". Memory is not like a strip of film; remembering how something looked is not a matter of re-inspecting an image which was gained when the object was initially seen. Memory has more of a constructive element to it than that sort of picture allows. When I recollect something, I do not simply "re-present" how things were given to me; I re-construct them. Remembering how something looked involves the imagination. (This is why "false memories" can be such a problem -- the imagination cannot simply be excluded from the faculty of memory, so having accurate memories is a matter of having a disciplined imagination, and not simply "not imagining things".)

(It is worth noting that the sort of memory I am concerned with here is only one way in which we use memory-talk. If I remember my PIN number, this is just to say that I know what my PIN number is. I do not have to imagine the numbers to remember my PIN number; here, "remembering" is just knowing at various moments in a stretch of time. And the same holds for determinate (propositional) bits of memory, such as "I remember that the Washington Monument is white." This analysis won't do as a way of describing what it is to remember how something looked, in the sense that concerns me; that sort of memory is not just knowledge mentioned idiomatically. For if I remember how the cube looked, then this memory can provide a noninferential grounding for my judgement that the cube was made of ice. But if remembering how the cube looked was just a matter of knowing various things about the cube, then the grounding it provided would have to be an inferential one. But the knowledge that the cube was made of ice is not gained through inference: I simply am able to recognize ice cubes as being made of ice when I have them in clear view. It is not that I am able to gain knowledge of a thing's being made of ice when I have certain other bits of knowledge about it, such as might be made explicit as premises in an argument whose conclusion is that the thing is made of ice. If pressed for why I think the cube was made of ice, I might only be able to say that it looked like it was made of ice. But looks-talk can't provide inferential grounds for is-talk, as Sellars showed. This is all by way of preliminary throat-clearing.)

So, on the picture I was puzzled by, the dilemma was that the content of an intuition given to me can't involve the activity of conceptual capacities which I don't possess, and yet it seemed that this same intuition's content had to involve concepts which I gained later. For I might be unable to recognize things as made of ice when I have an ice cube in view, yet my having had a clear view of the ice cube might allow me to know that it was made of ice when I "look back on it" later, once I have gained the capacity to recognize things as made of ice. I think my problem was that I thought that memory must involve the re-presentation of the same intuitions which were my original having-in-view of the ice cube, such that limits on my conceptual capacities at the time of the original viewing would limit what concepts could be in play in the content of the recollection of the cube. But if the image which I call to mind when I remember what the cube looked like is a product of my imagination, then whatever conceptual capacities I have presently might be drawn upon in the construction of the image presented to me in recollection. (I'm worried about tripping up on my language here, since recalling how something looked has to be something I do, and not merely something that happens to me (as is the case with vision), yet I still want to draw on McDowell's language of conceptual capacities being passively drawn on in the content of what is "seen" (here, metaphorically, in a "mental image", rather than an actual one). My conceptual capacities are passively drawn on in my active attempt to recall how something looked, as opposed to being actively drawn on in an active attempt to produce an image (which is what is generally called "imagining something").)

Now, I can easily see a worry here: If the first time a conceptual capacity is drawn on is in imagining how things are (as an element of attempting to remember how things are), then why should the result be capable of providing a justification for a judgement as to how things are? For the product of the imagination might seem a mere subjective fancy. I think that there is a valid worry in this: trying to gain noninferential knowledge from how I recollect things to have been is not as reliable as trying to gain noninferential knowledge from how things appear to me in experience. In memory, it is more common to dream things up that never happened than it is in waking life.

But, this sort of being mislead by a faulty memory is intelligibly erring. For the problem is that what I took to be part of the content of what I had seen was not actually part of the way things were. That which was supposed to justify my noninferential judgement turned out to not be adequate for the task demanded of it. But this shows that adding the imagination to the picture is sufficient to dissolve the dilemma I faced earlier: If I can be misled by the contributions of my imagination, then I can also be lead to licit conclusions by them, as with my ice cube example. They are the sort of thing that is a candidate for being a reason for judging that things are thus-and-such. Having bad reasons for a belief is not simply having exculpatory excuses for having the belief; it is a case of reasoning badly rather than appealing to something outside of the space of reasons.

And so I am able to again hold the opinion that if one does not limit the conceptual capacities which can be involved in experience, as McDowell didn't used to, then one can avoid a problem which McDowell's most recent position has to face.


Pierre-Normand said...

Here is a suggestion as to how to understand McDowell’s new position according to which not every concept figuring in non-inferential perceptual judgment need also figure in the content of an experience grounding it. I first introduce a necessary and sufficient condition for such a concept not to figure in the content of experience and then I suggest some illustrative examples to motivate this formula.

When an x is perceptually recognized as an y, its being an y is not part of the content of the experience if and only if this further recognition contributes nothing to the way this y is being perceptually individuated and tracked over time and no new (forms of) perceptible properties can be ascribed to it just on account of its being an y.

For instance, when I learn to recognize some birds to be cardinals, there is no specific property that I can see cardinals to have that I cannot easily conceive some other (common) kind of bird to also manifest.

By contrast, when I see an animal to be a bird, then this might be on account of its manifesting (or its appearing suitably constituted so as to be able to manifest) some form of behavior that not only makes it manifest to me as a bird (as being red would contribute to make it manifests as a cardinal), but such a predicative forms is partly constitutive of the way birds are manifested to me.

Consider chess figurines and the functional roles they are called to play in the game of chess. As I attend to a game being disputed by two players, my recognizing some piece as a bishop can be a case where _bishop_ (the functional/individuating concept) figures in my experience, over and above the more generic concept of a chess piece because I can see and expect it to behave in certain ways (some of which bishops might share with other kinds of pieces, albeit in different combinations). If I were to see what I took to be a pawn being moved in the same way, then I would no longer be able to see it as a pawn, either because I acknowledge to having misidentified its role in the game, or because a player made a mistake, or maybe the players no longer abides by the rules that are constitutive of the game.

This last example also suggests to me that McDowell’s revised account is still hospitable to the possibility that one’s experiences must be enriched as a result of refinements in one’s perceptual conceptual skills. As someone at first learns the rules of chess and must apply the rules clumsily in a reflexive and self-conscious manner, she might only be able to individuate the pieces in her experience as wooden figurines with distinguishing names and shapes. At that stage, recognizing a piece as a bishop adds nothing (or very little) to her visual experience of the chess figurine. It only enables her to known (possibly infer) that some specific rules apply to *it*. It is only as she become proficient at the game that the concrete material tokens withdraw in her experience behind more abstract (but still manifest) functional ‘roles’ and significant chess positions. When she recognizes a figurine as a bishop she can then see that it threatens the opponent queen and this is a genuine enrichment of her experience. (“Threatens” is now a new form of relational predicate that she masters and can apply non-inferentially in experience specifically to chess pieces.)

A similar story might be told, mutatis mutandis, regarding the observation of positrons in cloud chambers.

This account might also illuminate further how reliance on the Myth might be avoided. When one learns to recognize birds that are cardinals as cardinals, one might in the beginning only be able to arrive at inferential judgments. “This is a red bird with a dark face, so, this must be a cardinal.” (If I can be allowed this enthymeme.) At a later stage one recognizes cardinals non-inferentially. There is no improvement possible of this conceptual skill as it bears on one’s grasp of the natural kind _cardinal_, beyond the way cardinals just *look* (without reliance on the scalpel and test tube, that is) One can even forget *that* cardinals are red and have dark faces and not be any worst at identifying them on sight. Think of learning to recognize people. One can actually get worse at knowing explicitly features while one gains familiarity with a human face.

The case of concepts that are genuinely part of intuitional content is different. Beyond mere unarticulated recognition of the kind, one gains (possibly tacit) mastery of interrelated skills with new forms of predication, identity conditions, etc. This complex and rationally articulated set of conceptual skill then becomes interwoven into one’s ability to recognize, individuate and track objects.

So, we have two cases of recognition: unarticulated (or minimally articulated, through just being able to name the recognized kind) and articulated. Only the account for the former kind threatens reliance on the Myth. But this is because we haven’t contemplated the possibility that in the case of that sort of recognition, one might indeed only be entitled to exculpation for the deliverance of ones intuitions. This is not to deny that the non-inferential judgments they ground can be genuine cases of knowledge. It is rather that experience alone is insufficient at providing justification. Because the conceptual skill involved in this sort of recognition is so thin, one must in addition rely on the knowledge that one is brutely reliable. (“How do you know this is a cardinal?”, “Because it looks like one to me.” This also is an enthymeme. The suppressed premise is that one can rely on the thin recognitional skill one has.) The case of the “articulated” recognitional skill is different. Here, one likewise gains knowledge non-inferentially, but one can unpack ones judgment and explain why an X looks like one and why that makes it an X rather than a Y.

Enough for now… I mostly meant to postpone commenting further on McDowell’s retraction of his former view regarding the propositional form of intuitional content. I am still struggling with this. I’ll just retract straightaway my earlier suggestion that Evans’s “fundamental Idea” of an object might figure in the content of an experience prior to its propositional articulation. It now seems to me that it might be precisely the other way around. One cannot even have a discriminating knowledge of an object until ‘after’ (conceptually, not temporally) the ‘I think’ of transcendental apperception has been involved in the articulation of an intuitional content.

Daniel Lindquist said...

"When an x is perceptually recognized as an y, its being an y is not part of the content of the experience if and only if this further recognition contributes nothing to the way this y is being perceptually individuated and tracked over time and no new (forms of) perceptible properties can be ascribed to it just on account of its being an y."

I don't think this works. Take Conway's "Game of Life": recognizing a group of pixels as a "glider" entails recognizing a new kind of object as present in the grid, with its own conditions for individuation and tracking its movement over time. But by your condition, this would mean that "glider" would have to figure in the content of an experience of seeing a cellular automaton playing out the "Game of Life". I think this points to a more general problem: gaining novel recognitive capacities can entail recognizing novels sorts of objects, and not merely recognizing old objects in a new way.

If I learn to recognize birds with long necks and white feathers as swans, the way that I individuate swans is just the way I previously individuated (those token) birds. But suppose I learn more about swans, such as that they are socially monogamous: they form distinct male-female pairings for purposes of protection and egg-rearing (though not so much for mating, it turns out -- Google taught me something about swans tonight).

If I learn this sort of thing about swans, then I might become able to recognize pairings of swans. Tracking and individuating these pairings presumes my being able to track and individuate individual swans, but is neither reducible to it nor sufficient for it. The conditions for being a pairing of swans are not identical to the conditions for being a swan, or even for being a pair of swans simpliciter, but are something more complicated (they have to be two swans that raise baby swans together, protect the female's eggs from predators, etc.). Recognizing two swans (x) as a pairing of swans (y) changes the way I individuate and track that pairing of swans (that y) over time. I will continue to track the individual swans in the same way as I previously had, but the new phenomenon, the pairing of swans, is tracked in a new way -- it comes into being after the two swans separately had existed for some time, and can cease to exist despite both swans continuing to exist (suppose we transplant one to another location and they find a new mate). So, concepts like "pairing of swans" (or whatever the proper term is for a unit of socially monogamous waterfowl is) will have to feature in the content of my experiences of pairings of swans, and not just be involved in the recognitive capacities which I bring to bear on what is already "visually present to me". This seems counter to McDowell's intent -- for if "pairing of swans" is part of the content of an experience, then it seems that "swan" will have to be, too. (Since if I can experience a grouping of birds as a pairing of swans, then I must be able to identify each bird individually as a swan; if the former does not require any recognitive capacities involving concepts not present in the content of the experience, then the latter can't, either. But then species of birds would be among the concepts which can figure in the content of an experience, which McDowell explicitly wants to deny.)

"For instance, when I learn to recognize some birds to be cardinals, there is no specific property that I can see cardinals to have that I cannot easily conceive some other (common) kind of bird to also manifest."

What about the property of being a member of the kind "cardinal"? If I can imagine other birds which have this property but are not cardinals, then this just shows that my property "being of the kind 'cardinal'" does not pick out all and only those birds which are members of the kind "cardinal". But this seems to be a perfectly good property, and I don't see why it can't count as a "perceptible property" for those who are sufficiently skilled in birdwatching.

It occurs to me that the distinction between "perceptible properties" and other properties, without relativizing the distinction to a particular subject, threatens to fall into the Myth of the Given. There are no properties which, simply as such, are perceptible properties; one has to become proficient in the use of the relevant concepts to gain any particular capacity to noninferentially gain knowledge of any particular sort. Thus Sellars's stress on the fact that one has to learn to be able to account for lighting etc. to be able to see what color an object is. And McDowell at least used to deny that there was a set "limit" on what properties could become perceptible properties, as his example of electrons in a cloud chamber shows.

I suspect that McDowell's new limitation on the concepts that can figure in the content of an experience may presume an argument for there being certain concepts which one has to have simply to be able to perceive at all; if McDowell thinks this is so, then that might explain his distinction, as being that between properties which one necessarily can recognize and properties which one can't necessarily recognize, simply by virtue of being a sentient language-user.

If there's any convincing argument of this sort to be found, then McDowell at least has not given it to us; he's become increasingly confident in his appeals to Kant's B-Deduction (such as in the Haddock & MacPherson volume on Disjunctivism, where his only hedge is "or perhaps Strawson's reconstruction of it"), so I suspect he might hold something like this view, and simply hasn't published on it explicitly.

This would tie in to his comments in section 9 of the first "Afterword" to "Mind and World" about it being possible to rehabilitate "interesting analytic truths", and to pick out necessary features of any possible worldview. I simply don't think that the sort of talk which McDowell suggests can be rehabilitated can be rehabilitated: any belief can be revised, and this scuttles the attempt (in its strong form -- arguments like Kant's B-Deduction can still show us how some of our concepts hang together in a way that makes it no longer mysterious that we hold some claims true (such as "every event has a cause"), but this requires the presumption that we hold certain other claims true). "Kant's Theory of Science" is the best work I've read on this topic; despite the name, it's mostly about what to do with Kant after Quine and Davidson (with some very nice constructive results). It's a pity it appears to be out of print; my local library has a copy. Incidentally, Davidson mentions it at the beginning of "Laws and Causes", which I find utterly incomprehensible as a response to McDowell-type worries about the "fourth dogma of empiricism". I read the book hoping it would make Davidson's objection clearer, but there is no argument whatsoever in that book for the necessarily nomic character of causal relations; indeed, he discusses Kant's failure in the "Metaphysical Foundations of Natural Science" to provide a transcendental grounding for the particular laws he wants to enshrine; that every event must have a cause doesn't lead to the conclusions about physical laws that Kant wanted it to.

Looking at section 9 of the first Afterword, it's worse than I remembered. I should probably post on it. McDowell seems to ignore the possibility of revising what we take experience to show us, and also presumes that we can say in advance what might show up for us in experience (in a strong enough sense that analytic truths can be those whose truth can't be called into question by any possible experience). Both of those strike me as bad moves (as does endorsing something like "On Certainty"'s "hinge propositions"); I want to stick to the good Sellarsian line that any judgement can be revised (though not all at once).

This comment is threatening to become really huge; I've not yet addressed most of your remarks! (Don't worry, I read the whole thing before I started writing this comment.) I think I will put a break here and come back to it later. (I don't think your "enthymatic" treatment of exercises of recognitive capacities will work; if I know I have the capacity to recognize cardinals, this can only be because I know a cardinal when I see one. So the former cannot justify the latter: if what is in question is whether or not the thing I claim is a cardinal is a cardinal, I can't appeal to the fact that I can recognize cardinals to justify the claim, since whether this particular claim is an exercising of that capacity is just what is in question. I note that your distinction does hold for third-person views: I can infer from Bob says he sees a cardinal over there and Bob is able to recognize a cardinal when he sees one to There's a cardinal over there without any problems, and the material inference from Bob says he sees a cardinal over there to There's a cardinal over there can indeed be treated as an enthymeme. This is because what I take Bob to be claiming is independent from what I think about Bob's capacities for making reliable claims of the relevant sort, whereas what I take my own claims to be is directly related to what I take myself to be capable of veridically judging. This is because I regard my own claims as true, but I don't have to think this about Bob's claims. Notice that only another person can raise a doubt about whether what I take myself to see is in fact how things are; in my own case, I do not doubt what I take myself to see. If I doubted, then I would not take myself to see it, but merely to seem to see it (or to possibly having seen it or merely appeared to see it). When I ask, in a Cartesian-skeptical vein, whether things "really are" as I view them to be, I take a third-person view of my own beliefs. Which is just the problem with that sort of skeptical question.)

That last parenthetical was supposed to be a short addition to the comment, and, well, look at it. Yeah, best break this off here and come back to it.

Pierre-Normand said...

There are very many good points and good objections in your last post (my enthymematic treatment is indeed flawed). Let me just address now a misunderstanding.

Some of the proposals in my last comment were much influenced by Wiggins’s theses about the individuation of continuants in _Sameness and Substance Renewed_ (in particular the thesis of sortal dependence of identity judgments) and Haugeland’s (Sellars inspired) thesis regarding the ground for the objective significance of empirical concepts advanced in his _Truth and Rule Following_ which appeared as the last chapter of his _Having Thought_. My use of the game of chess drew inspiration from that work. One of Haugeland's key points in this paper (opposing Dennett’s reliance of the idea of patterns disclosed by the intentional stance) is that, precisely as you say, bringing into view new kinds of objects through committing ourselves to standards that are constitutive of them (the rules of chess, say) need not consist simply into our seeing *old* objects as new ones.

I might thus have been careless with my use of the phrase “seeing as” which indeed invites a reading that implies numerical identity. But I meant this use of the phrase to also cover cases of material constitution, when, for instance, one would say that a piece of clay is recognized as a statue although both objects are distinct and might later proceed to have independent careers. They just happen to coincide spatially at a time. Your examples of the Conway game of Life and the Swan pairs are very perspicuous, but they also illustrate my main point, it seems to me: I indeed meant to suggest that, when my criterion (“if and only if...") is satisfied, then there *is* numerical identity between x and y (since the individuation conditions are then the same). I have suggested that, in that case, for an x to be a y is not part of my experience of that x (on my gloss of McDowell’s position). It might for instance be the case that the concepts _y_ is a temporal restriction of the sortal concept _x_ (adult vs. human being) or that _y_ and _x_ have a species-to-genus relationship.

If, on the other hand, my criterion is not satisfied on account of _y_ being a sortal concept with different individuation conditions than _x_ then, of course, the x and the y that can be seen are not numerically identical. But that was part of my point. It would be on account of this *manifest* non-identity that _y_ must then figure in my experience if I am to perceive that object at all. My experience of the y cannot be shared by one who merely can see x. The latter experiencer just lacks the appropriate receptive structure.

The second part of my criterion ("and no new form...") was also meant to leave it open that _y_ might be a species of _x_, and in that case the individuation conditions of the object are identical, but my grasp of the intuitional form of an _y_ (the way I perceptually am receptive to y’s) enables new forms of predication. It would then be on account of the latter that this is a case where _y_ must figures in my experience.

So far, I just meant to clear up a misunderstanding. However, since when I made my proposal, I’ve thought of a new and possibly better way to ground the distinction (and modify its formulation, accordingly) between the part of the content of the non-inferential judgment that figures in experience and the part that doesn’t. It did not occur to me very clearly, until now, that the way McDowell conceives of mere recognitional abilities that are brought to bear onto what experience independently bring into view makes axercises of the latter quite similar to identity judgments. This now seems to me to be quite significant; since I am tempted to follow Evans in thinking that perceptual judgments (singular demonstrative thoughts) are identification free in the same sense that genuine first personal thoughts are. Self-ascriptions of beliefs or actions are example of the latter. They leave no room for inquiring “I believe that p, but is it really I who believes that p?” I will not here go further into Evans’s account (or Sebastian Roedl’s slightly different account) as to why the two utterances of “I” cannot be thought by the speaker to refer to two different persons. Evans similarly proposes that a demonstrative thought must be temporaly dynamic if it is to be thought at all. This is, as I have indicated in an earlier post, required by the generality constraint. We can discuss this more, but if you are prepared to grant me that much, then it follows that the same demonstrative singular thought can be expressed more than one time as one visually tracks an object and predicate something of it. It’s being the same thought means that it is identification free, for, if the thought is true and remains so while one tracks the object, then there is nor room for the question “This glass cube is pink, but is it this glass cube that is pink?” In normal cases, is the same singular sense that is being expressed twice in both utterances of “this glass cube”.

Similarly, when I recognize someone as N.N. and see that she is F, then my judgment N.N. is F is a demonstrative judgment of the form {This person N.N. is F}. (“is” is here the copula, and “is F” a predicate) The concept of a person need not be explicitly articulated in the judgment, but it is a consequence of Wiggins’s sortal dependence of identity judgments that such a concept be part of it. The judgment is identity dependent for there is room to ask: “Is this person really N.N.” I thus know that N.N. is F if the person that I demonstratively identify is N.N. *and* that person is also F.

This suggest to me that the ground McDowell has for denying that _cardinal_ is part of the content of my experience of a bird might be that he conceives experience to consist into demonstrative singular contents while acts of predication that are (in some sense to be explained later) “identification” dependent, although they can rest on the demonstrative part of someone’s perceptual judgment, also depend on such “identification” judgments the satisfaction of which do not depend on the demonstrative component of the thought. Hence, when I see that a cardinal is perched, my thought might have the form: {this animal, a cardinal, is perched}. _Animal_ is normally unarticulated but it povides the identity criteria necessary for the singular content of the thought (_this (something with such and such identity criteria) is being perched_) to be contentful -- to be, that is, of a definite substance with a definite history. The thought also has the "identification" component _this is a cardinal_. I must further spell out in what sense I take this form to be assimilated to a identity judgment when the predicate (here a natural kind term) is itself a sortal. This has to do, I believe, with the way natural kind terms are distinguished from the general class of predicates that stand for properties of substances. But this has been long enough... (I will also need to comment further on how, in Haugeland, predicative form and identity criteria, and the skills required to apply both, hang together as two sides of the same coin in the disclosure of an empirical domain of objects.

Pierre-Normand said...

In French, ice is called "glace". So I expressed my thought: _Ce cube de glace_ with "This glass cube"!

It's also happened at least twice to me to type "de" in an English text while I meant "of".

Many other typos and improprieties in the message above despite my having proofread it three times before posting it...

Pierre-Normand said...

This is an imprtant correction however:

"The judgment is identity dependent for there is room to ask: “Is this person really N.N.” I thus know that N.N. is F if [I know that] the person that I demonstratively identify is N.N. *and* [I know that] that person is also F."

Daniel Lindquist said...

I have that Wiggins paper, but I haven't read it; I mean to at some point, since I enjoyed what little of Wiggins I've read. I should probably try to get through some more Hagueland before the fall, come to think of it....

I think I have a clearer grasp on what your biconditional was supposed to be saying now. The way cardinals are individuated is the same way that birds and animals generally are individuated, and so recognizing an animal as a cardinal doesn't change how I individuate it. Which is not the case with recognizing two swans as a pairing of swans. (To use Brandom's terminology, tracking individual swans is P-P necessary for tracking pairings of swans, but not P-P sufficient -- one has to be able to do more than just identify particular swans to track pairings of swans). And you're proposing that only a concept relevant in the individuation of an object is a concept which functions into the content of an experience of that object. This does at least seem to give a measure of sense to some of McDowell's examples, and how the distinction between "intuitional form" and "intuitional content" is supposed to be related to his new views.

I'm not sure I like the metaphysical weight all this seems to bring in its tow, but perhaps I'm just being skittish; even if it is "metaphysical" in a bad sense, it does sound Aristotelian enough to be something McDowell might think, and Wiggins and Hagueland are certainly both thinkers McDowell is friendly to. (McDowell's "quietism" would then seem to be threatened, but this would not be new -- it's not at all clear how some of McDowell's more Kantian pronouncements square with "quietism".)

As a first pass at poking at the metaphysics-in-the-bad-sense I see threatening to pop up:
"_Animal_ is normally unarticulated but it povides the identity criteria necessary for the singular content of the thought (_this (something with such and such identity criteria) is being perched_) to be contentful -- to be, that is, of a definite substance with a definite history. The thought also has the "identification" component _this is a cardinal_."

It's not clear to me why "cardinal" should only be a quote-identification-unquote component, rather than an identification component. If the way this cardinal is individuated & tracked is the same as how this animal is individuated & tracked, then why should one description have a privileged place? "Cardinal" would provide the same identity criteria, in this case, as "animal" or "bird" would. (If this bird, this animal, is a cardinal, then anything that is a different cardinal is also a different animal and a different bird, and vice-versa all around. So I don't see why one particular concept should have a privileged role in giving individuation conditions for the object.)

I also suspect that someone might be able to identify cardinals without having our concept "animal" -- perhaps he regards birds, mammals, reptiles, fish, fungi, and trees as all falling under the same genus, without making a distinction between animals and non-animal lifeforms. So our concept "animal" wouldn't be available to him to provide individuation conditions for the various living beings he is able to track, though he might individuate birds and cardinals just as we do. Generally, I suspect that attempts to pick out particular concepts as the concepts which provide individuation conditions for an object are going to be vulnerable to holism worries.

(In such a case, I should think it would be easy to teach him to use "animal" in the way we use it (roughly, animals are the lifeforms that are not rooted and can move themselves, which sort of aspects of lifeforms he should already be able to track). To use Brandom's terms again, his current conceptual capacities might be P-V sufficient to master the concept "animal", though he doesn't presently have such a concept. I think this sort of fact can explain why certain concepts seem like concepts that "have" to be possessed by a speaker who can wield certain other concepts -- possession of the one sort of concept demonstrates the practical possession of all the pragmatic capacities necessary to wield the further concept, provided one makes a distinction which already has the capacity to make; the feeling that certain concepts are "a priori" is then the slip from this sort of pragmatic conditional to the idea that because one could wield a particular concept, one does wield it, inexplicitly, and what one "learns" when one is taught to use the concept is merely to bring the concept to consciousness.)

"Evans similarly proposes that a demonstrative thought must be temporaly dynamic if it is to be thought at all. This is, as I have indicated in an earlier post, required by the generality constraint. We can discuss this more, but if you are prepared to grant me that much, then it follows that the same demonstrative singular thought can be expressed more than one time as one visually tracks an object and predicate something of it. It’s being the same thought means that it is identification free, for, if the thought is true and remains so while one tracks the object, then there is nor room for the question “This glass cube is pink, but is it this glass cube that is pink?” In normal cases, is the same singular sense that is being expressed twice in both utterances of “this glass cube”."

Assuming I understand you (I only know Evans secondhand), I agree with this. But I'm not sure how it's supposed to make sense of McDowell: it seems to me that "This ice cube is pink, but is it this ice cube that is pink?" is not a question that makes sense to ask, but neither is "This ice cube is pink, but is this ice cube pink?" The distinction between the predicative and the substantive terms in the judgement doesn't seem to make a difference, here. (They both look to me like variants on Moore's Paradox -- they have the form "P, but is it the case that P?".)

A rough attempt at a summary of where I think the conversation stands at the moment: Your proffered views are starting to seem plausible as being the sort of thing McDowell has in mind with his new distinction, but I can't see that they are satisfactory on their own account -- I still don't see how McDowell can avoid my worries about falling back into the Myth of the Given. Which is weird, because that's the sort of thing he's normally sensitive to, and his old views seem to me to be more satisfactory than his current ones!

Enjoying the discussion; lots to think about in all this.

J said...

Kant Koachin'!

Alas, there are no necessary arguments for the synthetic a priori itself (as with most Kantian conceptualism). It can be posited, suggested as an alternative to empiricism of various sorts, or given a cognitive spin, yet not proven.