I suppose there's sorites on the one hand, and the ship of Theseus on the other; you can ask if you still have a heap after taking away a grain, and if you still have the same heap. (People's intuition about the latter might very dramatically.) An external constraint -- this blob is the bronze, meaning all of it, that used to be the statue -- blocks the latter but not the former. If you've lost any, you have to say this is some of the bronze. — Srap Tasmaner

One could stash a heap of sand on the deck of the Ship of Theseus and that may make things more interesting ;-)

One issue with blobs, heaps and chunks is that they are modifiers that turn mass nouns into count nouns. A bag of flour (count noun) isn't quite the same as the flour (mass noun) that's in the bag. The amount of flour that's in the bag can be more or less, but the bag of flour is one. There is an issue, of course, if one buys a bag of flour: how much flour can be missing before it doesn't count as a full bag of flour anymore? And that may depend on what's written on the bag (10kg, say). This is where a specific issue of vagueness arises.

Back to the statue. If one chips away at the bronze statue of Hermes with a scissor, then there will come a point where what remains isn't a statue of Hermes anymore. Likewise if you would hammer it flat. There will come a point where it's not recognizably a statue of Hermes anymore. And the issue of vagueness also arises in this case. It may not be possible to say exactly after how many blows of the hammer the statue doesn't exist anymore, as opposed to its merely being a badly damaged statue (or its having turned into a statue of D.J.T., possibly.)

So, when a sortal concept such as statue provides criteria of persistence and individuation, those criteria can specify (in accordance with common understanding) how much stuff can be taken away, or replaced, or how much the form can be altered, etc., before the statue is deemed not to exist anymore. And issues of vagueness may arise in all cases. Sometimes, such issues are settled in court.

There's a sense of 'object' on which a heap of sand would not be considered an object, but you can perhaps invent a story where it would be natural to call it an object. — Fafner

I think vagueness issues are orthogonal to the question of the mass noun versus count noun distinction. Hence one might ask if a mountain that has two identifiable peaks really is just one mountain or two mountains. That is a separate question from the question whether mountains are material objects or rather are a sort of stuff. They clearly are the former as a matter of ordinary linguistic practice (since "mountain" is conventionally used as a count noun) but we could imagine a reformed language that would identify "mountain-stuff" as the presence of abnormal elevation along the surface a continuous terrain. We'd say: "there is much mountain-stuff over there..."

Likewise, I think heaps are discrete entities, but saying so just is a grammatical remark. It's not due to anything intrinsic to the structure of actual heaps (although regarding the vagueness issue about heaps, Timothy Williamson has strange ideas about them).

I'm thinking that for a given theory, some ways of refining or extending it will be natural and some will be ad hoc. So naturalness is also theory-relative. — Srap Tasmaner

I agree. Natural scientists aim at carving "nature" at its joints. But the natural joints just are the joints that show up in empirical inquiry when the inquirer privileges matter over form, and intrinsic properties over relational properties. Since forms and relations aren't any less real than material and intrinsic properties are, this all amounts to an arbitrary restriction of the definitions of "nature". Such a restriction of the concept of nature may reflect a "fundamentalist" tendency (as when one speaks of physics being a more fundamental science than biology is, say) or a reductionistic tendency.

Actually I think that would be a measure of how the theory refinement is done, not that it has been refined. — Srap Tasmaner

Agreed. One is more of an act of theoretical reason while the other is more of an act of practical reason. Either we are seeking to disclose affordances of the natural world, or we are seeking to produce new technical and/or social affordances (or we are engaging in some combinations of both theoretical and practical endeavors).

(On edit: I'm still talking about our general activity of defining and/or instituting and/or discovering sortal concepts.)

(and there's a further question about which set exactly defines this or that piece of matter: If I remover one molecule of copper from some piece of copper, would would it be no longer the same piece of copper?) — Fafner

It likely remains the same piece of copper but, if it didn't belong to you to start with, that might constitute theft!

I guess I'm just still unclear what we're supposed to have learned about "meaning." — Srap Tasmaner

I don't know. So far, I have mostly been concerned with clarifying possible misunderstandings. If there still is a mystery lurking it's quite possible that it is also a mystery for me.

So we're abolishing any distinction between natural and ad hoc sortals. — Srap Tasmaner

I don't think so since the substance-sortals that we use to individuate artifacts or other sorts of objects that we care about aren't merely ad hoc, on the one hand, and also "natural sortals" (e.g. concepts of natural kinds such as chemical elements of biological species) also respond to specific theoretical and technical interests that we have, on the other hand.

Yes, I think that's the sort of thing I'm saying. We could say that dummyness is theory-relative, and that Travis's argument is that there is no final theory possible, in which all dummyness has been eliminated. But there is the built-in corollary that we can always eliminate dummyness by a refinement of our theory, so I'm still unclear on what conclusion is to be drawn. — Srap Tasmaner

I think what you just said (including the corollary) is good enough a conclusion to be drawn!

Why isn't set membership enough? — Srap Tasmaner

If you are a mereological essentialist then set membership is sufficient (and furthermore necessary!) But even if you aren't a mereological essentialist -- as you probably shouldn't be on my view -- then, in some cases, set membership might the relevant identity criterion associated to a sortal concept. A married couple, say, doesn't remain the same married couple through replacement of one of the spouses!

Yet again it is a question of interest and purpose. Are we sure there will turn out to be sortals that are never dummy sortals? — Srap Tasmaner

A sortal that isn't a dummy is a sortal that is merely good enough for the job at hand. A similar question could be asked about meanings, generally. A explanation of the meaning of a world could be a dummy explanation, in a sense, if it would fails to disambiguate two different ways this word could be used to convey two separate meanings, in a specific context. If it is unambiguous in this context, then the definition is good enough. But even after disambiguation has occurred and two senses have been distinguished and explained, we can always imagine a new context -- Charles Travis excels at this -- where the word used in one of those previously explained senses is ambiguous again.

It's not perfectly clear that what you call here the "lump of bronze" that constitutes the statue is an object at all. It feels more like a mereological sum of bronze bits. Your question might still be ambiguous, but not between two further determinations of a generic sortal, but between the object and the stuff it's made of (which is not an object).

If I tell you to move the statue, I'll expect you to keep it in intact. If I tell to clear the snow from the front walk of our museum, it's okay for you to change the configurations of the bits of snow, let some of them melt, etc. Swapping object and stuff in those examples would have dramatic and peculiar consequences. — Srap Tasmaner

No contest there. In the case of artifacts, it's very much our own interests that determine what degree of change in material constitution can be tolerated before an object will count as being destroyed. Hence, our attitude towards historical artifacts, sacred relics, or functional artifacts may vary widely. Incidentally, that's precisely how Peter Simons most elegantly solves the paradox of the Ship of Theseus in Parts: A Study in Ontology, Clarendon Press, 2000.

According to Simons, ship is something of a dummy sortal since some people may be interested in Theseus's ship qua historical artifact, or buy it in order to make use of it as a fishing boat. Those two diverging interests would coordinate with two different substance-sortals. There is, in effect, two different sorts of ship at the same place and at the same time. The planks that are making it up at a time are something else entirely. One of those ships (the fishing boat) is resilient to the replacement of the planks (and may even mandate such replacements as part of necessary maintenance), while the historical artifact isn't.

What do you mean by 'dummy sortal'? — Fafner

Consider the case of material constitution. At a given time, a particular lump of bronze can materially constitute a particular statue of Hermes. Both the properties lump of bronze and statue are substance-sortal concepts because they have associated with them (among other things) principles of persistence and individuation. When one grasps what it is for an object to exemplify a given sortal concept, one thereby grasps what the conditions are under which such an object persists in time or is destroyed. (This may of course be in large part a matter of convention regarding objects of those types). Hence, if the statue is being hammered flat, it is destroyed but the lump of bronze that was constituting it persists under a different form. We can also imagine circumstances where it makes sense to say that a statue survives the exchange of its original material constituents (through successive acts of restoration, say).

Both the statue and the lump of bronze constituting it at a time are material objects. Material object, though, is a dummy sortal because there is a need for further specification in order that a principle of identity be applied to an object that falls under it. In fact, further specification of a dummy sortal is required in order that reference to a particular material object be secured at all, even in cases where the act of reference occurs by means of demonstration rather than definite description.

For instance, suppose I were to point to a statue of Hermes and ask you whether this "object" that I am pointing at might survive being hammered flat. You can't answer unless I would specify whether I am talking about the lump of bronze, or the statue, or something else maybe. If I insist that I am thinking merely of "this material object", whatever it is, then my question is meaningless (or, at any rate, confused).

So, my view (following Wiggins and Marcus) regarding event-types is similar. The general category event, just like the general category material object, is a dummy sortal since it isn't, in the general case, specific enough to determine conditions of persistence and individuation for events.

Wouldn't quotation marks signify an utterance? How could you know the sense without knowing the context of the utterance? — Mongrel

Well, of course, I assumed it was indeed implied that those words have reference only in the context of an utterance. Fafner was proposing an analogy with the case of the references of "the son of of Edward VIII" and "the father of Elizabeth II" that are the same in spite of their (Fregean) senses being different on some occasions of use. So, I made the point that, on my view, "Caesar's murder" and "Caesar's death" don't merely have different senses but also different references on account of murder and death being distinct event-types. This is a difference at the level of reference, and not at the level of sense, or so I would argue.

But I also made the concession, after more exchanges with Fafner and Srap Tasmaner, that there are pragmatic contexts where both words -- "murder" and "death" -- might be regarded as referring to the same event-type, for all practical purposes, and hence that one might refer to the same event as either "Caesar's murder" or "Caesar's death".

Around Ceasars murder... there were quotes. So it was kind of bizarre considering the title of the thread that a discussion ensued about the reference of the words. — Mongrel

Ah! Sorry. I thought you were referring to the original post in this thread. Fafner indeed used quotes around Ceasar's murder. I am unsure what it is that you find bizarre or that I may have missed regarding the significance of those quotation marks.

We utter sentences in order to express propositions. You have to look to context to know what proposition was expressed. You can't look at the words and know the reference, but just a possible reference. Did Frege disagree with this?

Frege wouldn't disagree. He made a big point that you can know the reference (Bedeutung) of a word only through grasping its sense (Sinn). Frege's explanation of sense (Sinn), though, is very much different from what has come to be known as linguistic meaning: something akin to a tacit rule, or definite description, that language users make use of to determine the reference. He also was reaching for an ideal language free of ambiguities and imprecision, but that's something else.

Exactly. Whereas the original question in this thread involved quotes. I think you have a tendency to ignore the significance of that. — Mongrel

I got involved in this thread when I reacted to what someone said, after the discussion was well on its way. But I just now looked back at the original post and I am unsure what you are driving at. What quotes?

I can't pass quoting another bit from Eric Marcus, Rational Causation. This is footnote 22 on p.195:

"Someone might object here that two people (e.g., J. J. Thompson(sic) and Anscombe), neither of whom could be said to grasp the ordinary concept of a murder better than the other, might disagree about the temporal boundaries of a murder and thus disagree about the principle of identity for murders. The concept of a murder, it will be said, is inherently murky. One might thus conclude that there is no determinate principle of identity for murder and perhaps for many other event-types as well. This might well be true; but if it is, it only strengthens the analogy with the principles of identity associated with different kinds of object. For one can analogously argue that an ordinary grasp of such concepts as ship, caterpillar, and person does not furnish one with definite answers to every question that might arise over the survival of a particular ship, caterpillar, or person. To argue that such murkiness undermines principles of identity for object-sortals and event-types across the board would, in effect, be to argue that there are no facts about the identity of objects and events. I take this to be an unacceptable result." --Eric Marcus (bold in the original)

Bloody murder! I had read that five years ago and quite forgotten that Marcus had chosen murder as an example. He used a bold type for murder to mark it as an event-type, a sortal concept.

"J. J. Thompson" is actually J. J. Thomson. Elizabeth Anscombe and him had had a disagreement over the interpretation of the doctrine of double effect.

I don't believe that conceptual inquiry is a way to 'disclose' the essential metaphysical nature of things (and therefore I also reject the idea of a synthetic apriori truth, at least on the traditional understanding of the term), and this is perhaps where our disagreement lies. — Fafner

I don't believe that either. You brought up the case of Kripke's a posteriori necessities regarding the essences of natural kinds, specifically. (I'd rather speak of essences since I view the claimed "identity" of water and H2O as a matter of necessary material constitution, and, following Wiggins, I am distinguishing identity and constitution). In response to that I insisted that what can be known only on the basis of experience need not be, for that reason alone, outside of the scope of inquiry about meaning, as I take Kant to have shown (as I once argued here and there).

I think that in some sense it is an arbitrary matter whether we say that two events are identical or not (like a death and a murder), and it is a confusion in my opinion to think that analyzing the concepts "death" and "murder" can tell you the 'real' answer from the perceptive of the events 'themselves' as it were (and please correct me if I'm wrong, but I read you as saying that there is an objective answer to this question, which is determined by the nature of the events in the world).

I agree that it is only in relation to a specific practical context, and our purposes in that context, that a death and a murder can be subsumed under the event-types (the equivalent of substance-sortals for events) that individuate them. My claim was that it isn't generally the case that they will turn out, under those pragmatic conditions, to identify the same event. And that's in part because 'event' is a dummy sortal. But I had made the concession, while responding to Srap Tasmaner's useful suggestion, that what is merely a dummy sortal, in one pragmatic context, might provide a fine-grained enough principle of individuation , in another context of inquiry (e.g. a murder investigation) to individuate the murder and the death in such a way that they are conceived to be identical events (under 'two different descriptions' we might say).

So, I am not seeking 'real' answers, but only objective truths. Objective truths can be truths about secondary qualities of objects, say, or socially instituted facts (such as the value of a currency) so objective and subjective aren't contraries.

I agree that concepts are world-involving as you said, but not by a way of reflecting the metaphysical essences of things. However, it is also not the case on my view (and I agree with you here) that "investigation into the objects, kinds and properties that our words purport to refer to is something that occurs outside of the bounds of the conceptual".

That's my main point.

There's a wonderful paper by Putnam called Rethinking Mathematical Necessity that explores this topic further. — Fafner

Thanks for this reference. It is Putnam's endorsement of Travis that motivated me to read Unshadowed Thought. Now, I think I may have developed some reservations regarding Travis's understanding of Frege, but that would be a topic for another time!

One problem with facts is how to tell when one fact is the same as another. There's no fat man in my doorway. There's no thin man in my doorway, either. Two facts or one or none? Perhaps there's just One Big Fact to which all true statements refer - which gets around the problem of fact-identity. — Cuthbert

According to the slingshot argument there indeed only is one single fact that all true sentences correspond to. The arguments, though, has also been viewed as a refutation of the correspondence theory of truth.

If I may just throw out a suggestion, in passing, regarding the requirement (or lack of a requirement) for there to be a convention in order that communication could be meaningful. I haven't managed to read all of the recent exchanges between StreetlightX and Srap Tasmaner. But it had seemed to to that StreetlightX was suggesting that words can be used in specific ways to achieve meaningful communication without there being an established convention for using them in that way; and Srap Tasmaner was disagreeing.

StreetlightX had also mentioned at some point, if I remember, that words are used meaningfully if the way they are used is, in a sense, conventionalizable. And that would be sufficient for effective communication to occur even if no convention already exists. That seems right to me.

But I was also reminded of Charles Travis's example of the plant with green leaves. Srap Tasmaner had brought it up to illustrate the occasion-sensitivity of meaning. (That was actually Fafner, sorry). In some contexts, if one were to request some green leaved interior plant and were offered a brown leaved interior plant that had had its leaves painted green, that would likely not satisfy her request. But if she rather had needed some decor prop for a photo-shoot, then it might have. The point of the request, which is assumed to be understood by the interlocutor, can determine what does or does not satisfy the predicate "...green" as meant by the requester. The question then is: must there be an established practice for using the predicate "...green" in just that sense for the question to be understood? And the answer would seem to be no. All that's required is that the interlocutor has a grasp of the point of the request, and this understanding may only requires something like agreement in form of life (including, possibly, a shared local culture).

Nevertheless, if there happens to be a general context where the request for a "green leaved plant" is meant to be understood to have the specific point mentioned above, then it might come to be expected that the predicate "green..." will be used in that way whenever such a general context reoccurs. This use will thus come to be conventionalized. And from that point on, the expression will be misused, and not merely misunderstood, when someone misinterprets it in that general context. What has happened, effectively, after the use has become conventionalized is that what was formerly a route from apprehension of the requester's intention to the apprehension of the (occasion sensitive) meaning of her utterance, now has become a route from the apprehension of the (conventional) meaning of her words to the apprehension of her communicative intention.

(Edited to add links to earlier comments)

I agree that at least on some cases we can know just on conceptual grounds that the same thing cannot satisfy two different description if it doesn't make sense to say that it does (e.g. to describe something both as an animal and as an inanimate object at the same time). But sometimes such identities can become aposteriori discoverable possibilities, as Kripke and Putnam have taught us about natural kinds. The interesting question here is what distinguishes the two cases and how can you know when you are confronted with the one or the other. And this brings me to another interesting thing that you said. — Fafner

I think this is just a misunderstanding, so let me try and clear that up separately.

On my view, which is indebted to Evans's and Wiggins's neo-Fregean re-appropriations of Putnam and Kripke on the semantics of natural kind terms (as well as the metaphysics of natural kinds) just because the manner in which our words reach out to their referents can't be determined through linguistic analysis alone (where such an analytical activity is construed to be achieved from the armchair only) it doesn't follow that investigation into the objects, kinds and properties that our words purport to refer to is something that occurs outside of the bounds of the conceptual. Empirical inquiry oftentimes is the proper way for us to clarify our concepts (or, our 'conceptions', as Wiggins would characterize the Fregean senses of natural kind terms) and to better anchor them into the essential natures we seek to disclose (when there are any). What Kripke would call a posteriori (metaphysically-)necessary, though, rather corresponds to what Strawson, Sellars or Kant would call a priori, or synthetic a priori, in the case of Kant; and to what Wittgenstein calls grammatical remarks. None of those inquiries are done from the armchair, but rather are reflective inquiries into our public language games. And those language games are world involving. It's only in this sense that they are 'a posteriori'.

(A little while ago, I had posted links to a couple older posts of mine, about Kant, Sellars and Haugeland, where I sought to explain the relavant sense of synthetic a priori. That water may turn out to be H2O, on that account, and therefore 'twin-water' not to be water, would be sythetic a priori even though it can only be known by means of experiments.)

Thinking of this as a "more complete description" led me almost immediately to the concern you noted, that I would need to posit "fully determinate facts" at the head of such sets, and that seems a bit dubious. — Srap Tasmaner

Before responding to this, and also to Fafner's most recent reply to me, I wanted to read again chapter 5 -- Events and States -- in Eric Marcus, Rational Causation, HUP 2012, a brilliant book that I had last read five years ago. It is especially relevant to our present discussion since it explains sortal concepts for substances, and event-type concepts for events, as both being in the business of providing identity principles for them. While Marcus was bringing up Wiggins's discussion of 'dummy sortals' (such as 'object' or 'thing'), I wrote down the following note this morning:

"Or 'mammal'?

Maybe we could say that 'mammal' is a determinable sortal. If does some of the job of individuation, but not all of it until it has been further determined into a species concept.

Maybe this idea provides the sought after stopping points for Srap Tasmaner's "cofinal tails". Determinable properties and dummy sortals don't determine such tails, but substance-sortals and event-types possibly do since they determine as fully as one might want *what* something is. Further, specifications (or further determinations of determinables) beyond such a natural stopping point only achieves the specification of merely accidental properties (including such things as the accidental microphysical realization or material constitution of events or substances.

Where occasion sensitivity might play a role, among other places, is where determination of some determinable has established membership in an equivalence class that is thereby fine-grained just enough to satisfy some practical purpose (including the founding of a meaning convention for effective communication, in some cases."

Here is the relevant paragraph from Rational Causation:

"I hold, then, that objects instantiate sortals and that to instantiate a sortal is at least in part for there to be a principle of identity that determines the conditions under which the object persists. Objects, however, do not instantiate principles of identity as such, but rather only insofar as
they are particular sorts of objects. Here is how Wiggins puts the idea: "If a is the same as b, then it must also hold that a is the same something as b,"7 where something is a quantifier ranging over determinate sortals. Or, if a is the same as b, there must be an answer to the 'same what?' question. To say that object is not a true sortal is to say that it is not a proper answer to the 'same what?' question. In saying that a is an object, we do not say what a is. Wiggins thus distinguishes 'dummy sortals,' such as object and thing from genuine sortals such as dog and table. Terms for dummy sortals share the grammar of terms for true sortals (e.g., they are modified by articles and quantifiers), but are not associated with a principle of identity." -- Eric Marcus, Rational Causation, p. 187 (bolds and italics in the original).

I'll finish re-reading the whole chapter before responding more fully to this and another reply of yours.

I'm not necessarily disagreeing, but I have two questions (which are related): a) are you claiming that one can know the Fregean reference solely by virtue of knowing the meaning of the relevant predicates? (which clearly you can't since you cannot know apriori whether "Caesar" and "the conqueror of Gaul" denote the same person) — Fafner

No, of course not. First I was talking about the references of the predicates -- "...died" and "...was murdered" -- and not the references of the singular terms. And our knowledge of the references of the terms (or predicates) can be empirical, or gained through testimony, or (on Gareth Evan's account -- see chapter 11 in The Varieties of Reference) derived from our being "consumers" in a socially instituted naming practice in which some other individuals -- the "producers" -- who are directly acquainted with the named individual (or with the designated property) are participating (or participated).

Secondly, my claim was that the two events are numerically distinct by dint of the predicates "...died" and "...was murdered" referring to different sorts of actions/processes regardless of anyone's knowledge of the references of those predicates. But, of course, our knowledge of them, when we know them, isn't generally derived a priori form linguistic meanings except in the special case of so called "nominal definitions". (Though, how much can be inferred from knowledge of linguistic meaning might vary depending how you relate the idea of linguistic meaning to the Fregean concepts of sense and reference, and what the scope of linguistic 'analysis' might be. If such analysis is allowed to cover the examination of public language games for instance...)

and b) Is "the conqueror of Gaul" a rigid designator on your account? Because if it is not (and it is plain that it isn't) then I think your criteria for the non-identity of 'x' and 'y' (in the quote) becomes vacuous. Because consider that it is a contingent fact that "Caesar" and "the conqueror of Gaul" denote the same person (and you can further substitute 'Caesar' with another description to eliminate all names); but this you can know only aposteriori, so it means that on your criteria 'x' and 'y' (if 'x' and 'y' are definite descriptions) denote the same entity if their terms happen to denote the same entity, and of course everyone will agree with that...

But all I was suggesting was that, just in case Caesar wasn't the conqueror of Gaul in the actual world, then it trivially follows that the event of Caesar's death is numerically distinct from the event of that other guy's death. And so it is, I am arguing, if the event- of process-forms that specify the two event types are different. (I draw the concept of an event- or process-form from Michael Thompson). The issue is ontological and not directly tied with issues of knowledge or reference. Let me add, though, that while thinking about this case, and about criteria of event individuation, I have gained a better appreciation of the ground one may have to claim identity between the historical events of Caesar's death and of Caesar's murder. I think the case being discussed, and the implicit surrounding narratives, can be further filled up in such a manner as to warrant either one of the two intuitions depending on the kind of 'sortal concept' (or rather, the kind of 'event- or process-form', for the category of events) that most perspicuously attaches to the events being talked about and thereby determines their criteria of persistence and individuation. But I'll say a bit more about that at a later time; it also connects with the issues of occasion-sensitivity.

I keep thinking, as I suggested in the other thread, that what we want here is sets of propositions ordered by entailment, but it looks like that would have to be relative to a set of assumptions or background knowledge or something. I want Pat's being 5 feet tall to buy you, as a single fact, everything it entails. A separate fact for everything Pat is taller or shorter than seems less than optimal. — Srap Tasmaner

I think this proposal raises a few problematic issues. First, through attempting to do away with the multiple realizability of coarse-grainedly specified facts, it tends to restrict our fundamental ontology to something akin to a 'fundamental' supervenience base like the set of microphysical facts, and that seems to me to be a profound error. Secondly, there is an issue regarding the nature of the entailment relations that are allowed for eliminating entailed 'facts' in favor of the 'real' or 'fully determinate' facts that entail them all. Is it only logical entailments from (ultimately) 'raw' empirical facts that are being allowed or are entailments that make use of premises expressing conceptual truths also allowed? The latter would seem to be the case if the fact that an apple is red is to be derived from the fact that it is crimson. One also needs the conceptual truth (if it is one) that all crimson things are red. (Or, at least, that crimson apples are red). Augmenting our special ontologies with a priori conceptual truths seems problematic to me since conceptual truths rather seem to belong to the transcendental background of those ontologies: the a priori conceptual truths in virtue of which empirical facts can be disclosed to us at all.

(In this post and the following one I had sketched a defence, following Haugeland, of a pragmatized Sellarsian/neo-Kantian conception of 'impure' synthetic a priori statements that are dependent on experience without them being objects of experience -- or without then 'arising from' experience, as Kant would say)

Consider the following case that has been discussed to illustrate 'contrastive' causal explanations: A pigeon has been trained to peck at red objects. A red apple that happens to be crimson is presented to the pigeon and the pigeon pecks at it. The fact that the apple is red, in conjonction with the fact that the pigeon has acquired a disposition to peck at red objects, causally explains why the pigeon pecked at the apple. The fact that the apple is crimson may also explain why the pecking occurred provided we are reminded that crimson is a shade of red. However, causal explanations of events seek not only to determine, on each occasion, why some specific effect followed some specific cause. It is no real explanation that 'explains' only one single occurrence. The real explanation of the fact that the apple was pecked at is that it was red, and this explanation unifies a whole range of similar phenomena (i.e. other instances of the same pigeon pecking at, or failing to pick at, various objects). Hence, the fact the the apple is crimson, unlike the fact that it is red, may fails to provide the real causal ground of its nomological relation with definite ranges of effects. (I may try to come up with a somewhat less contrived example. However, you can refer to the two posts liked above where I discussed an example from Newtonian mechanics and another one from the game of chess.)

There's also a very good paper by john McDowell "The Content of Perceptual Experience" (appears in "Mind, Value and Reality"), that argues for a very similar idea to Hornsby. — Fafner

Seconded. This is a paper that ought to figure in any anthology on the philosophy of mind.

Is that down to philosophical fashion? Or is there something we know that most of cognitive science doesn't? — Srap Tasmaner

Quite possibly. I think cognitive science, by an large, has not kept up with J. J. Gibson, among other anti anti-representationalist pioneers; although there had been a salutatory revival of a variety of embodied/embedded/situated/scaffolded paradigms recently (See Robert A. Wilson, Boundaries of the Mind: The Individual in the Fragile Sciences - Cognition, Cambridge University Press, 2004). Cognitive scientists still tend to be mired in psychologism and representationalism when they theorizes about mental abilities and mental states in a way that construe the mind and its operations as being sandwiched in between the "inputs" from the senses and the "outputs" of motor action. (Susan Hurley characterized this as "the classical sandwich model of the mind", which goes hand in hand with representationalism). This is not necessarily damaging when they seek to understand how specific cognitive abilities are being enabled by physiology, for instance, or what are favorable or unfavorable learning conditions for this or that skill being efficiently acquired, etc. But when they theorize about the very 'person-level' skills and phenomena (e.g. the ability to remember or what it is to believe this or that) that they seek to understand, then they are being biased in their theoretical undertakings by inchoate Cartesian-empiricist presuppositions. This is being argued forcefully in Max Bennett and Peter Hacker, The Philosophical Foundations of Neuroscience, Wiley-Blackwell, 2003.

Just to be clear, what anti-representationaists such as the authors mentioned above object to isn't the quite reasonable idea that brains and internal cognitive processes enable people to use language, represent the world to be thus and so, and act skilfully in the world. Neither are they denying that we are indeed representing to ourselves the world to be thus and so. Rather, they object to the construal of our abilities to represent the world (and act in it) as a matter of our (or our brains') constructing internal representations that have their contents independently of the form of our engagements in the social/material world (such as our playing language games, or our skilfully exploiting behavioral affordances).

Another paper that is quite relevant to anti-representationalism is Jennifer Hornsby's Personal and sub‐personal; A defence of Dennett's early distinction, Philosophical Exploratons, 3, 1, 2000

But what if somebody meant his murder by "his death?" Wouldn't the reference be the same in that case? — Mongrel

You mean the event she is thinking about and the reference of her utterance?

On edit: I think it's entirely possible that she means to be referring to Caesar's murder when she uses "his death" and be correctly understood thanks to the surrounding conversational context. For all that, someone else could refer to Caesar's death as such, and be referring to a distinct event. One might inquire, for instance: "I heard Caesar bled and suffered for many hours after he had been stabbed. Might not his dying have been abbreviated by his friends?" And this would surely not meant the same as: "Might not his being murdered have been abbreviated by his friends?" And the reason for this, I surmise, is because the murder and the death aren't identical events. (As well, we might say, of the murdering and the dying, which are the same two distinct events, correspondingly, being described from the progressive rather than the perfective grammatical aspects.)

Let me just make a mention of the elephant in the room. Caesar's murder was an action whereas his death wasn't. Intentional actions have a principle of unity quite unlike the principles of unity of natural events and processes. This has been stressed quite a bit in the philosophy of action since the publication of Anscombe's Intention. Here is another relevant paper that I haven't read yet but that is on my very short list: Wittgenstein on Actions, Reasons and Causes

You should change the descriptions to "the death of Caesar" and "the murder of Caesar", and then I think it will make more sense to think that they denote the same event (and you cannot really decide this just by analyzing the descriptions themselves, since it is after all possible for two different descriptions to denote the same event; e.g. "the death of Caesar" and "the death of the conqueror of Gaul"). — Fafner

That would be a valid objection if my criterion for saying that 'x' and 'y' refer to two distinct events is for them to have different Fregean senses. But that's not what my criterion of non-identity is. I'm rather saying that 'x' and 'y' refer to distinct events if (among other possible differentia of events) the predicates used to characterize the actions (or processes, etc.) that are predicated of some entities, in order to individuate those events, have different Fregean references. (Predicates, and not only singular terms, do have references as well as senses. They typically refer to properties, action forms, activities, etc.) In your example above, the same event is being referred to twice since the two singular terms "Caesar" and "the conqueror of Gaul" have the same reference. If they had had different references, then the events would have been numerically distinct, obviously. And so is it, on my view, if the references of the predicates has been different (as I'll argue some more below).

Since "murder" just means something like "violent death", then on your view it would follow that a person can die twice (if "murder" and "death" are two distinct things that happen to everyone who's murdered), which is be a pretty bizarre thing to say in my opinion.

I wouldn't say that Caesar died twice. I would say that he died because he was murdered. The 'violence' that is constitutive of the event's being a case of murder is the mens rea of the murderer(s). This mens rea isn't a constitutive part of Ceasar's dying. Hence, since the two events don't have the same constitutive parts, they are not the same. That is true (in this case, anyway) even when we restrict attention to what occurred in the actual world (and a fortiori if we consider the modal properties of those events).

I'm not claiming that dying and being murdered are always the same thing. I'm only claiming that in the particular case of Caesar the two descriptions happen to denote the same event (since they are non-rigid designators etc.). And there's nothing problematic in saying this. I'll try to illustrate this through your example. Crimson is a type of red, but it doesn't follow that a crimson apple has two distinct colors: crimson and red, but it has only one color that falls under two different descriptions (and this is consistent with the fact that being crimson and being red sometimes do refer to distinct colors).

I would not say that the apple has two different colors either, because counting as two colors a determinate quality and the determinable quality that it is a determination of would be misleading. That would be like saying that I have two pets: a cat and a mammal. But the fact that the apple is red isn't the same as the fact that the apple is crimson, is it? Likewise, (1) the fact that Pat is shorter than Chris isn't the same as (2) the fact that Pat is shorter than the Eiffel Tower. And neither of those facts are the same as (3) the fact that Pat is 5 feet tall. For all that, the properties ascribed to Pat in (1), (2) and (3) stand pairwise as determinable to determinate.

I don't agree, I think it is the same event under different description. And also I don't see the disanalogy between the two examples: why can't "...was murdered" and "...died" have the same reference just as "the son of..." and "the father of..."? — Fafner

Well, "the son of..." and "the father of..." will only have the same reference if you fill them up with singular terms in such a way as to turn them into complete definite descriptions.

As for "...was murdered" and "...died", there just is no way to fill those up and refer to the same event (or so would I argue). In order to achieve something similar to the previous case, you would rather need something like "the ... who was murdered at (some time and place)" and "the ... who died at (some time and place)". Then, yes, you could fill them up in such a way that they would refer to the same individual (under two different Fregean senses). But this individual would be a human being rather than a historical event.

The issue with Caesar's murder and Caesar's death is that they refer to two different things that happened to Caesar. For sure, the murder could not have occurred without the death also occurring, but, as I think you already noted, the converse isn't true. And this asymmetry isn't merely a matter of the modes of presentation (Fregean sense) of the events.

Another way to highlight the difference is to notice that "...was murdered" is a determination of the determinable "...died" rather in the same way in which the property "...is crimson" is a determination of "...is red". But it is clear that an apple's being red isn't the same thing as its being crimson under two different descriptions. Likewise for, say, Fido being a dog and Fido being a mammal. Being a dog is one determinate way for an animal to be a mammal and being murdered is one determinate way for someone to die.

Again, I think the root of the illusion of there being a common "thing" being referred to under different descriptions when two separate events occur at the same place and the same time is the fact that roughly the same individuals are involved in both cases and there is a tendency to identify what happens (the 'neutral event') with its 'raw' material supervenience base, as it were. But an event singles out not just the individuals involved while merely specifying some definite time interval when the action occurs. It ascribes to them some specific relations and/or action/process forms that those individuals are involved into, and leaves out others that are irrelevant to the constitution of the event. (Hence, say, an apple falling from a tree wasn't necessarily part of WWII even if it occurred right then and there).

On edit: I think the same error underlies Donald Davidson anomalous monism. He is right regarding mental events not being subsumable under universally quantified statements of laws under those descriptions (that is, qua mental events), but he is wrong about them being token identical with physical events that are so subsumable under those different descriptions. A physical event never is something mental that is being described differently. If something mental occurs, then when one proceeds to describe what is occurring in physical terms, one is thereby talking about something else. (Which is not to say that there aren't any relationships between mental events and some of their physical underpinnings).

Or my example of "Caesar's murder" etc. — Fafner

That doesn't seem to be quite the same since in the case of the earlier two examples (regarding the state of the weather, or George VI) the respective pairs of predicates, or singular terms, had the same Fregean reference (same Bedeutung) albeit different Fregean senses. But in the case of the events being referred to as "Caesar's murder" and as "Ceasar's dying" the difference in meaning runs deeper since the predicates "...was murdered" and "...died" have different Fregean references and not merely different Fregean senses. They are not two ways to single out the very same determination of the object (Caesar) but rather are ascribing different properties of him. On my view, there is no common 'event' that is being referred two under different descriptions. They are two numerically distinct events even though the very same individual is involved in both of them roughly at the same time and at the same place.

Characteristics of the apple; its sweetness and its redness separately make those corresponding statements true. It is that entity, that apple, which is both red and sweet. Just as with the two statements: that Caesar died and that he was murdered, it is the corresponding characteristics of that event that make them true. — John

Exactly. That's my main point. But then when you suggest that x and y are numerically the same 'event' you are conflating the physical movements where those events occurred (or some other such allegedly 'neutral' way to characterize what's going on), which are the particulars falling under the predicates "...exemplifying Caesar's being murdered", and "...exemplifying Ceasar's dying", on the one hand, and those particular events falling under the corresponding predicates, on the other hand. The former are particulars, which indeed are numerically identical, but the latter are structured propositions, which aren't.

OK, it seems now you are saying that it is a matter of interpretation as to whether he was murdered or justifiably assassinated, or something like that? If that's so, it's a different question, and could be gotten around simply by saying that he was killed. — John

Not at all. Part of the point is that events are particulars whereas being murdered (or being killed) are general concepts expressed by predicates. Claims and sentences have (minimally) subject/predicate form. What purports to "correspond" to claims such as to make them true must therefore have a similar structure.

Say, the claims that a particular apple is red, or that it is tasty, may both "correspond" to the same apple. This is just to say that it is the same apple that is being referred to in both cases, whereas the predicates (and the general properties ascribed) are different. But the apple itself doesn't make both claims true quite appart from its falling under the corresponding predicates. It's the same with events, since just like apples, those are particulars. "Caesar's dying" may seem to be referring to an event, and hence to a particular, but it's only really the individual Caesar (and also, possibly, a specific time, or the values of a time variable being quantified over) that are the particulars being referred to in the judgement expressed by "Caesar died". The "event" all by itself doesn't single out the general properties that it falls under such as to make claims about it true.

I can't see that, because assuming that Caesar was murdered then it is his being murdered that makes "Caesar was murdered" true, and that also makes Caesar's dying and Caesar's being murdered the very same event. — John

If x is the event of Caesar's dying and y is the event of Caesar's being murdered, then what makes x being a case of murder isn't the same thing as what makes x the same event as y.

What makes it the case that x is the same event as y, presumably, is that both occurrences are instantiated in the same region of space and time. (I am just trying to play along with the token-identity theory that seems to underlie your correspondence theory of truth). But what makes it the case that x is a case of murder is something else entirely. It depends on the significance of the concept of murder in a way its alleged identity with y doesn't.

They are the same event iff Caesar was murdered. — John

Precisely. In my post I only considered the actual world, not any alternative possibilities. And I assumed that Caesar was indeed murdered, and that x and y therefore denoted the same event/entity (on your own account!) It still does not seem to be the case that the actual event makes the claim that Caesar was murdered true. It is rather something specific about this event that makes the claim true: namely that the event happens to be falling under a specific concept expressed by the predicate ("...was murdered") in the claim.

If statement A and B do correspond to actuality, then they do, and if they do not, then they do not. Of course it is logically possible that they might not have both corresponded to the same event, but in that case actuality would have been different. Possibility has nothing to do with actual correspondence, though, as far as I can tell. — John

Maybe one way to explain what Fafner is driving at is this: If in the actual world A corresponds to x and B corresponds to y, where

A is the claim that Caesar died,
B is the claim that Caesar was murdered,
x is the event of Caesar's dying,
y is the event of Caesar's being murdered,

then, on the assumption that x and y denote the same numerically identical physical event, what make it the case that A is true is the very same thing that makes it the case that B is true, namely: x

However, it seems that there is something specific about x that makes it the case that B is true, namely that Caesar's actual death was a case of murder. But if it is only in virtue of x being a case of murder that x makes B true, then it would seem that what B "corresponds to" is something intensional about y, and not merely extensional; it is a concept under which the 'event' falls. And therefore it can't be y qua physical event that makes B true.

(On some accounts, it would rather be because the Fregean thought expressed by B is identical with the fact of Caesar's having been murdered that B is true. But then, this fact and the Fregean proposition expressed by B don't merely correspond to each other. They are identical.)

Pleasure, I suppose. — Terrapin Station

tl;dr

The object of agreement is different. For Wittgenstein as for Cavell, there is 'agreement in the form of life' at stake. It is not an agreement with respect to the conventional (by which I mean 'already-established') use(s) of language. That's the key difference. There are two analytic axes at work here: a language game and the form of life in which that language game operates. 'Agreement' operates at the latter level, as it were. — StreetlightX

The object of agreement seems to me to be the same sort of thing as the understanding Wittgenstein talked about when he said that there is a way to understand a rule that is not an interpretation. And it is regress blocking in the same manner.

This is a interesting thread and I wish I had the time to get involved in it. Recently, while browsing the items distributed by De Gruyter, I stumbled upon a book by Avner Baz -- When Words Are Called For: A Defense of Ordinary Language Philosophy, HUP 2012. I thought it might interest some of the participants in this thread. (Charles Travis is being discussed extensively). I've placed it near the top of my reading list. Here is the overview:

"A new form of philosophizing known as ordinary language philosophy took root in England after the Second World War, promising a fresh start and a way out of long-standing dead-end philosophical debates. Pioneered by Wittgenstein, Austin, and others, OLP is now widely rumored, within mainstream analytic philosophy, to have been seriously discredited, and consequently its perspective is ignored.

Avner Baz begs to differ. In When Words Are Called For, he shows how the prevailing arguments against OLP collapse under close scrutiny. All of them, he claims, presuppose one version or another of the very conception of word-meaning that OLP calls into question and takes to be responsible for many traditional philosophical difficulties. Worse, analytic philosophy itself has suffered as a result of its failure to take OLP’s perspective seriously. Baz blames a neglect of OLP’s insights for seemingly irresolvable disputes over the methodological relevance of “intuitions” in philosophy and for misunderstandings between contextualists and anti-contextualists (or “invariantists”) in epistemology. Baz goes on to explore the deep affinities between Kant’s work and OLP and suggests ways that OLP could be applied to other philosophically troublesome concepts.

When Words Are Called For defends OLP not as a doctrine but as a form of practice that might provide a viable alternative to work currently carried out within mainstream analytic philosophy. Accordingly, Baz does not merely argue for OLP but, all the more convincingly, practices it in this eye-opening book."

... "Avner Baz is Associate Professor of Philosophy at Tufts University. He has written about ethics, aesthetics, perception, judgment, and about the question of philosophical method in the works of Kant, Wittgenstein, Cavell, and John McDowell"

And then there is the "presidential" wedding.

Right. We know that. How about we analyze what it would amount to for it to make sense to you? Or is that a problem because you haven't read anything that you can regurgitate on that? You don't mean that you wouldn't say the same thing, right? — Terrapin Station

It really sounds to me like whenever the tea pot gets warm prime numbers above 17 suddenly get significantly heavier (except for 883, of course!). How about we analyze what it would amount to for this sentence to make sense to you? Of course, I'm not going ever to provide you with the slightest hint what it could possibly mean for me to say that. Why would I need to? It's all just a matter of arbitrary conceptual stipulation, after all. Or is that a problem because you are unaccountably prejudiced against tea drinkers?