What is happening here?

Well, thinking needs a subject (such as Alice). But Alice doesn't, in turn, have a subject - any chain of dependencies terminates with her. So in that sense, she is not dependent on anything further for her existence. — Andrew M

If I understand you correctly, we're talking about conceptual dependence here. That I can deal with. It even has a natural connection to Frege's saturated/unsaturated distinction: "___ is thinking it's going to rain" is unsaturated, incomplete, and therefore an abstraction, and therefore has only dependent existence.

I was seriously afraid that "independent existence" was going to lead to having to say what the ultimate constituents of the universe are! You have a comfort level with QM that I don't, so I thought that might not scare you as much as it does me; or, rather, it might be something I would rather not have to do just to talk about what ordinary sentences mean, but you might not mind!

Or, even if we're not doing that, we might take "dependent" in Alice's case to mean, you know, a planet with a breathable atmosphere and a food source, stars, the universe. Alice's existence isn't independent in any number of ways.

For Aristotle, ordinary objects (his primary substances) were the fundamental entities. — Andrew M

So here maybe we're talking about what is conceptually fundamental, and for what Sellars calls the "manifest image" (or for Strawson's "descriptive metaphysics") that is indeed going to be sensible objects and persons.

So the YouTube version: — Banno

Are we done now? Should I bother working on the rest of the article?

By the way -- I haven't intended to be advocating for Grice against Davidson, not exactly. He brings up Grice himself, situates how he intends to argue in relation to Grice, and suggests that the argument will along the way justify his approach. All of which is slightly odd since he already knows he's headed for a dead-end. Anyway, I've just been trying to work through that, not stand in for Grice.

I actually think Davidson's argument about what is and isn't needn't for first meaning is worth going over very carefully. I also think it's worth knowing if the main argument is sound, and if it is I am very curious how the formal semantics community has responded.

I was thinking of the passage that begins with Diogenes: he uses intention to pick out first meaning and then says we lose nothing by ignoring intention and going back to a radical interpretation model.

It's part of the argument that whatever Grice was on about, it might as well be extra-linguistic.

I haven't even gone through the main argument carefully, but don't you harbor any suspicion that he painted himself into the corner he'll end up in?

the relation between intention and meaning seems fraught. — Banno

Then what did you make of Davidson relying on it finally in his definition of "first meaning"?

Whose idea is this exactly? — JerseyFlight

I think we're talking about almost every cognitive scientist since Chomsky.

I don't have the text in front of me, but insofar as (1) talks about relationships, it's just going to be compositionality: we have a recursive engine for interpreting sentences based on how the semantic elements of the sentence are put together (syntax, logical constants). Sentences get their first meaning without reference to any other sentence.

So you're right that malapropism is built on a comparison between sentences, but the theory Davidson is imagining isn't. Unless I'm wrong.

Which might pretty much be that.

Should have said something about how Davidson treats intention in his definition of "first meaning", but we can come back to that.

I appreciate your comments but of course we are not discussing Plato, only the theory bearing his name that reifies meanings, ideas, numbers, propositions, concepts, what have you. Though I haven't read as much Plato as you, I too find his approach valuable and liberating; the theory bearing his name, less so.

The cessation of pain is, you're claiming, a particular experience different from not being in pain, like your Erfworld example. Another real life example is being in a room with a noisy old-fashioned window air conditioner that suddenly cuts off when you're in the middle of a sentence -- you find that you were speaking much more loudly than you realized in order to be heard over the noise.

I understand one part of it now: he brings up Grice only to dismiss him.

I dip into these matters only to distinguish them from the problem raised by malapropisms and the like. The problems touched on in the last two paragraphs all concern the ability to interpret words and constructions of the kind covered by our conditions (1)–(3); the questions have been what is required for such interpretation, and to what extent various competencies should be considered linguistic.

The first paragraph referred to is about ambiguity, and the order of the clauses in a conjunction. That includes this:

But part of the burden of this paper is that much that they ((i.e., competent interpreters)) can do ought not to count as part of their basic linguistic competence.

The second includes this:

Whether knowledge of these principles ought to be included in the description of linguistic competence may not have to be settled: on the one hand they are things a clever person could often figure out without previous training or exposure and they are things we could get along without.

We're a few pages in. Davidson has given the examples he finds challenging or puzzling, and explained how he's going to use the phrase "first meaning". Why does he bring up Grice? He has to. Grice's whole theory is based on distinguishing sentence meaning from speaker meaning, and Grice defends a view that we can still talk about sentence meaning as literal meaning. That is, in Grice's classic cases, we use words with their usual literal meaning to mean something different from what those words say, and we can be understood when we do this because there are rules that govern conversation.

Davidson is first of all making the point, made near the beginning of ever so many papers, that the case he wants to focus on is not covered by prior art. You might think that's something like, "This is worth talking about because, cool as Grice is, his theory doesn't cover my case." But what Davidson says in dismissing Grice is that the "abilities" (the word he'll use in the next paragraph) Grice describes are not specifically linguistic abilities, and anyway they're common sense, and anyway we don't "really" need them.

So what's going on here? Davidson is going to restrict the usage of "linguistic competence" to cover only the understanding of literal meaning, and not how we use language to communicate, if "communicate" is understood to mean letting others know what we mean (speaker's meaning) given a shared understanding of the literal meaning of our words. Of course he can choose to talk about whatever he likes, but to call this the only part of language use that is properly linguistic is tendentious and he knows it: he is dismissing all of pragmatics as having nothing essential to do with language. It's just common sense -- stuff a clever person could figure out -- and we "could get along without" it.

What does that last little comment mean? It means we -- i.e., Davidson -- can imagine a language that is fully disambiguated, does not rely on indexicals, has a prescribed sentence structure, has a stock of names large enough not to rely on any local speech community's usage, and is only ever used literally. (I probably left some things out, but you get the idea.) That is, while we don't speak an idealized Tarski-like language, we could, and the fact that we could means that whatever pragmatics has to say about language use is only about how we happen (strangely) to use the languages we happen (sadly) to have. In a perfect world, we wouldn't need it.

And there is an argument here about priority, which is why there's all the talk about first meaning and what comes first in order of interpretation. There are the usual two points here:

(i) If I only understand what you mean by figuring out that you did not mean what you said literally, I must have figured out the literal meaning of your words first -- temporal priority.
(ii) If what you mean is to be characterized as something different from the literal meaning of your words, then I rely on the literal meaning of your words to characterize what you said -- logical priority.

I don't see any reason to contest either of these points, but I will point out how un-Gricean it is to start here. People only blurt indicative sentences at each other because they intend to communicate. Grice doesn't have a principle of charity, which is an interpretive strategy, but a principle of cooperation, which binds speaker and audience in a shared enterprise.

This is not to say that there is something wrong with trying to understand the specific mechanism by which we communicate, but Davidson insists on describing how the machine works without acknowledging what the machine is for, and rules out discussion of what people use the machine for as irrelevant.

Malapropisms introduce expressions not covered by prior learning, or familiar expressions which cannot be interpreted by any of the abilities so far discussed. Malapropisms fall into a different category, one that may include such things as our ability to perceive a well-formed sentence when the actual utterance was incomplete or grammatically garbled, our ability to interpret words we have never heard before, to correct slips of the tongue, or to cope with new idiolects.

This is where the paper's argument properly begins, so this is where I will stop for now. On the one hand he's inclined to dismiss pragmatic considerations on principle, but he also thinks the analysis of malapropisms will justify this dismissal. What we may want to look out for is Davidson denying pragmatics the resources to explain malapropisms on the grounds that he has already ruled out pragmatics.

The argument seems at first blush to be that malapropisms cannot, by their very nature, be subsumed and accounted for by such conventions of language. Is that the whole of Davidson's argument, and is it cogent? — Banno

Hmmm. I had a quick read and a half, and will reread tomorrow.

We have options here.

(A) Is Davidson's argument valid? Is he right that his principles (1)-(3) cannot account for what I guess we'll call successful use of language?

(B) Are principles (1)-(3) actually descriptive of anyone's views? Davidson himself is certainly a candidate, but is anyone else? (Not attempting a survey, but we might want some of an answer for (C), coming up.)

(C) If his argument is valid, and we reject views well-described by his three principles, are there other approaches out there that handle malapropisms better?

(D) If we fail to find an existing theory that handles malapropisms happily, does that tell us anything general about the prospects for a grand theory of language, or only that an important dataset has been overlooked? (It's now been decades since ANDOE, so there is essentially zero chance that this dataset is still being overlooked in academic circles. I shudder to think how many papers this paper is directly responsible for.)

I will try to look carefully at the argument tomorrow. On the one hand, it wouldn't surprise me if Davidson's overall view of language use could be shown to be untenable, because it is a fanatically impoverished approach. On the other hand, Davidson has a way of constructing arguments that I can only describe as "untrustworthy", so I will tend to be a little skeptical of even Davidson's own views being refuted by Davidson.

When I first read this paper many years ago, I had not read Grice yet, so all of the nods to Grice went right by me. Now that I have read and thought about Grice a fair amount, the invocation of Grice here and there is just puzzling. I'll think about that too.

What is really shocking though is that he doesn't even mention David Lewis, who at the conclusion of Convention (published 1969) suggested that, while there are two general approaches to language, an idealized model like Tarski's or a more sociological approach like Wittgenstein or Grice , and he finds value in both, it is entirely possible that no one ever really speaks a language in the idealized model sense. (And then we get a little bonus mini-theory of what they might be doing.)

So this is a little bizarre right? Convention was not a nothing book. (And certainly by the mid-80s Lewis was well-known.) It carries a preface from Quine about how Lewis took up Quine's challenge to make some sense of the idea of convention and show its relevance to language -- he doesn't quite change his mind, as I recall, but he does admit that Lewis has made one hell of a case.

And then Davidson's argument turns overwhelmingly, he says, not on (1) and (2) -- closer to his own heart -- but on (3), and the very last line has a swipe at convention that could have come from a Quine paper some 30 or 40 years earlier!

So is the whole thing actually a veiled attack on Lewis? (I had a quick browse and for provocation there's at least a paper by Lewis from the mid-70s on "radical interpretation" and a quick scan of that suggests that Lewis was not exactly supportive of the program.)

That's a big pile of chitchat. I'll try to have something substantive to say tomorrow.

However, this is not the thread to hash it out on — JerseyFlight

No shit.

No one on this forum wants a lecture from you about how they should be spending their time instead; no one wants you to intrude in their thread to tell them you think it's pointless.

Please stop doing that.

Here's something I posted years ago:

There's a touching passage in Tarski's little Introduction to Logic that I'll quote in full here:

I shall be very happy if this book contributes to the wider diffusion of logical knowledge. The course of historical events has assembled in this country the most eminent representatives of contemporary logic, and has thus created here especially favorable conditions for the development of logical thought. These favorable conditions can, of course, be easily overbalanced by other and more powerful factors. It is obvious that the future of logic, as well as of all theoretical science, depends essentially upon normalizing the political and social relations of mankind, and thus upon a factor which is beyond the control of professional scholars. I have no illusions that the development of logical thought, in particular, will have a very essential effect upon the process of the normalization of human relationships; but I do believe that the wider diffusion of the knowledge of logic may contribute positively to the acceleration of this process. For, on the one hand, by making the meaning of concepts precise and uniform in its own field and by stressing the necessity of such a precision and uniformization in any other domain, logic leads to the possibility of better understanding between those who have the will to do so. And, on the other hand, by perfecting and sharpening the tools of thought, it makes men more critical--and thus makes less likely their being misled by all the pseudo-reasonings to which they are in various parts of the world incessantly exposed today.

That's Tarski writing from Harvard in 1940, having fled Poland before the German invasion.


Now will you please stop cluttering the forum with this drivel about what "true thinkers" should or shouldn't do. It is beyond tiresome. Do some philosophy or shut up.

I'll pitch in after I get a chance to re-read the paper, but it might be a few more days.

I think it's just a matter of grammar. I don't think there are metaphysical implications in that phrasing. — Andrew M

Then you are not my target audience!

As long as we take care not to reify such abstractions, is there really a problem here? — Andrew M

Right, this is the part of your position I've ignored: abstract entities have only dependent, not independent existence. By "reify" you mean precisely attributing independent existence to something that doesn't have it.

Here's one way we can talk. There is a general event type, someone thinking it's going to rain; there is a particular event type, Alice thinking it's going to rain; and then there are particular instances of that, Alice's thinking yesterday that it was going to rain. (Obviously lots of other ways to carve that up...)

That last is a particular, but in your terms it is not a concrete particular, not because of anything to do with types and instances but because every instance of Alice thinking it's going to rain is dependent for its existence on Alice existing, is inseparable from Alice. We separate what Alice is thinking from Alice only fictively, by means of abstraction.

And then two people thinking the same thing is still as simple as I want it to be, just a matter of using the same words to describe what you fictively detach as "what they're thinking." If you then generalize, you can talk about the idea "that it is going to rain" as what anyone you would describe as thinking it's going to rain is thinking.

One little question: in this analysis, every instance of Alice thinking it's going to rain is dependent for its existence on Alice existing, not on Alice independently existing, right? I'd love to stay away from saying what that's supposed to mean, but we don't seem to rely on it anyway. Do you agree?

So far we're juggling general vs. particular, abstract vs. concrete, and dependent vs. independent. There are obvious temptations to match them up (respectively) that I'm trying to be careful about.

If logic has an inherent contradiction, and math is based on logic, than math is self refuting — Gregory

The contradiction is in claiming that there can be any such barber. There cannot. Just because you can string words together grammatically doesn't mean you're describing something that can actually be.

That's why I like framing the problem as figuring out who would be in the set of all such barbers. You find that such a set is necessarily empty because the conditions for being a member are inconsistent. You might as well define a set of all the numbers equal to 4 and equal to 5. There's no contradiction in that; you've just defined a set that's necessarily empty.

Number 2 is wrong. The philosopher doesn't shave himself so he is in the set of those who shave all those who do not themselves. — Gregory

Let S be the set of all men who don't shave themselves.
P is a member of S.
To be a member of R, you have to shave all the members of S.
Since P is a member of S, to be a member of R he would have to shave himself.
But he doesn't.
Therefore P is not a member of R.

Suppose we have a town with three men: a barber (B), a philosopher (P) who doesn't shave himself, and a mathematician (M) who does.

Now define a set R as all and only men who shave all and only men who don't shave themselves.

1. M is never a member of R because he shaves a man who shaves himself.
2. P can't be a member either because he doesn't shave himself, so he'd have to shave himself to be a member, but he doesn't.
3. What about B? He would have to shave P and not M. No problem. If he shaves himself, he'd be out, like M, but if he doesn't, he'd be out like P. So B can't be a member no matter what he does.

So R = { }. No one shaves all and only men who do not shave themselves, therefore the barber does not shave all and only men who do not shave themselves. The three cases are exhaustive, in fact: no one can be a member of R whether they shave themselves or not. — Srap Tasmaner

I used to find the "third realm" argument (for instance, Frege's version) persuasive, but now I can't imagine why.

I've been neglecting the motivating role of quantifiers, so a quick note.

If Alice took her umbrella, we have to say Alice did something, right? The only alternative seems to be that she did nothing, but that's not the case, for she took her umbrella.

If I agree that, in taking her umbrella, Alice did something, while denying that there is something that Alice did, I am no more committed to saying that Alice did nothing than I am to saying there is nothing that Alice did.

By an abstraction, I simply mean something that does not exist independently of a concrete and particular context but can be considered independently of that context. Whether that be events, actions, thoughts, descriptions, or whatever. — Andrew M

I'm intending just the one sense. In considering something abstractly, we are being selective as you note, and there are different ways that might manifest, including at increasingly complex levels. But the key point is that it depends on something concrete. — Andrew M

So glad you've chimed in, @Andrew M!

Your approach (which you would say is broadly Aristotelian?) seems very sound: there is only one sense of "abstraction"; it is what we do when we consider a particular concrete context selectively.

I wonder, though, why is existence -- as in the first quote -- part of this story at all? If Alice is thinking it's going to rain, why even say that there is a thing, the thought that it is going to rain, that does exist, only it doesn't exist independently of Alice thinking it is going to rain, or of someone thinking it is going to rain?

I ask for two reasons:

(1) If I'm of a mind to deny that Alice thinking something entails there is something Alice is thinking , and you insist that we can consider what Alice is thinking independently of the concrete occasion of Alice thinking it, I do not need to deny this -- why would I? I only need to deny that us considering what Alice is thinking entails there being something we are considering.

(2) If the point is to emphasize our capacity to consider things selectively, and to describe this somewhat picturesquely as an ability conjure abstract entities for our consideration rather than being compelled always and only to consider the totality of the concrete situation, I will point out that we are already doing that all the time simply by using language in the first place.

Insofar as we want to ignore whatever else is going on with Alice except her thinking about the chance of rain and taking her umbrella, we say, "Alice is taking her umbrella because she thinks it's going to rain." "Considering selectively" is not a special thing we do sometimes with language; it's practically all we ever do.

But there does seem to be an exception to the idea that language is always selective: names of concrete particulars. When we refer to Alice, we mean everything about her, or at least intend not specifically to exclude anything about her.

The question then is whether, in creating "names" on-the-fly, we are referring to abstract particulars such as "Alice taking her umbrella" (an action or an event), and that question seems particularly acute when the name is anaphoric and thus somewhat open-ended: "what Alice said" or "what Alice did" or "what Alice was thinking".

And now we're sort of back where we started, but with a somewhat different focus.

When we see Alice grab her umbrella, we assume she thinks that it is going to rain. But that is an abstraction over her behavior, not some additional thing that has an independent existence. And we can consider that abstraction separately from its concrete context, i.e., as an abstract entity. — Andrew M

There are several points here that confuse me:

1. "Alice is grabbing her umbrella" is also an abstraction, right? We are leaving out whatever else is going on with Alice in describing her current behavior as "grabbing her umbrella".

2. "Alice thinks it's going to rain" is an abstraction in the same way (1) is -- we're not talking about whatever else may be going on in her mind -- but is it an abstraction in some other way? Is there another sense of abstraction in play here?

What I have in mind first is something like this: if we think of "grabbing an umbrella" as an action someone might perform, then we might take "thinking it's going to rain" not as another sort of action, to be correlated with "grabbing an umbrella", but as a disposition to perform the action "grabbing an umbrella". That seems like a different sort of abstraction -- one level up. (This disposition talk, as a way to deal with the mental, has a pretty low reputation these days, but I wanted to leave it as an option.)

If we take "someone grabbing an umbrella" as an event, I suppose we could think of "that same someone thinking it's going to rain" as another sort of event that could cause the first sort, but I'm inclined to think of it as a distal cause (if I'm using that term correctly!), something like the disposition above, a condition or situation that, if it obtains, makes it more likely that events of the "someone grabbing an umbrella" type occur.

3. Alice is a concrete entity and Alice's umbrella is a concrete entity; is "Alice grabbing her umbrella" an abstract entity? Is that what actions are? Or events? How do we capture the difference between "Alice grabbing her umbrella", a sort of abstract event that might occur, and "Alice is grabbing her umbrella" which, while an abstraction in the simple sense of (1) is pretty concrete -- it's a realization of "Alice grabbing her umbrella" after all.

4. Is there yet a third sense of "abstract" -- beyond actions/events, dispositions/conditions -- we apply to what someone is thinking, or what someone is doing, or what is happening? So we might have "that it's going to rain", "grabbing her umbrella", "Alice grabbing her umbrella" as the answers to what is she thinking? what is she doing? what is happening? Is this a different sense of "abstract" or are these again like simple (1)-type abstractions, leaving out everything but the content being thought, the type of the action being performed, the type of the event occurring.

Sorry to labor this so!

The tl;dr is that if what's being thought is an abstract entity, is that because there's something special about thinking? or because thinking is acting? or because thinking is an event occurring? Is it abstract because it's thinking, or because all our descriptions are abstractions?

They're something and not nothing, yes? And if not nothing, something, but not like bricks. In fact, not like any thing. They're something but not a thing. It seems like an absurdity. — tim wood

Perhaps you're trying to answer the wrong question.

As to what, exactly, let's give the science another fifty years or so. — tim wood

Which pair of questions do you think it more likely the science will find more tractable?

A1: What is a thought?
A2: What is the thing Alice is thinking when she is thinking about her grandmother's house?

B1: What is going on when we are thinking?
B2: What is going on when Alice is thinking about her grandmother's house?

I think idea are things, just exactly not like bricks, but rather as ideas. That is, they must be something. — tim wood

Why do you think that?



Do you think ideas are things because if you have an idea, there must be something that you have?

Are you trying to prove something about the world through grammar? — tim wood

That would be the Platonist you're thinking of, not me. I should have thought that was perfectly clear.

Because of the grammatical similarity between "Alice is kicking something" and "Alice is thinking something", there is a temptation to say that there is something Alice is thinking -- "there is" in a full-blooded sense: ideas are things like bricks, they exist independently of us, like bricks, all the rest.

Anaphoric constructions aside, we know what goes in place of "something" in "Alice is thinking something"; it's constructions like

"that the roof will never hold"
"of going to graduate school in the fall"
"about her grandmother's house".

Any of those look like things to you?

"Alice is thinking Descartes's cogito"? I don't think that's English.

For your other examples, yes! After I tacked on the what-clause thing, it occurred to me we can go directly for other anaphoric noun phrases:

"Alice is thinking the same thing as Steve."
"Alice is thinking a thought not unlike my own."

Can you think of an example that just sticks a regular noun phrase in there, and not some anaphoric construction like this?

Compare with:

If Alice is kicking something, must we conclude there is something that Alice is kicking?

Yes, for example, a ball. — Andrew M

Sure, but what goes in place of "something" in "Alice is kicking something"? It's a noun phrase of some kind:

(a) A proper name: "Alice is kicking Steve";
(b) An indefinite noun phrase: "Alice is kicking a ball";
(c) A definite noun phrase: "Alice is kicking the ball."

Can we do the same thing with "Alice is thinking something"? No, no, and no.

ADDED:
There are more options and one of the ones I left out has a fun sub-case:

(i) Alice is kicking what I left on the floor.
(ii) Alice is thinking what I am thinking.

Yes, counting requires that we have developed the idea of number, but that is not difficult. All we need is for our mate to be happier when we bring home 2-3 birds or melons instead of one. — Dfpolis

Read that as many times as you have to to see what's wrong here.


Thanks for the conversation and the reading recommendations. Cheers!

It is because sets are not things, but mental constructs (ways of grouping elements in our minds). Primitive shepherds counted sheep by tying knots in a string. ((...)) Why count sheep and not relations or sets? Because shepherds are not generally interested in possible relations between sheep, but in the number of sheep they have. — Dfpolis

The point of focusing on how we count was to point out how much of our conceptual apparatus (and not only that but other practices as well) must already be in place to do it.

So if a shepherd is interested in keeping track of how many sheep they have, that's not a property of the sheep (as "having white wool" is); that's the cardinality of the set of sheep they have. If they use knots on a string to keep track, that's Hume's principle, a one-to-one correspondence (a relation) between the set of sheep they have and the set of knots on the string. Thus shepherds must use sets and relations if they're interested in sheep.

And more than that, you have to know not to make a knot every time you see a sheep, even if it's one of yours, but only to make a knot, one knot, for each numerically different animal, either by keeping track or by artificial means like forcing them one-by-one through a gate.

There's no counting without both the mental constructs and the associated practices already in place. Does that matter to your position? What point were you making with the example of shepherds?

Yes; folk do lots of things that they probably shouldn't. — Banno

I'm sorry, are you saying there is something wrong with people giving vocal expression to their feelings, mood, and all the rest?

Wittgenstein’s rather unsympathetic response is exactly right, while entirely missing the point. — Banno

It's sociopathic.

(( I'm ignoring the rest of your post because I've actually read Wittgenstein. ))

Yeah. Many years ago I more or less gave up philosophy because of the rise of cognitive science -- I just didn't see anything much left for us to do. (I was young.)

My evolving view -- which one can see evolving in this very thread! -- actually reinforces my desire to mostly ignore brain science but keep doing philosophy, because brain science is not in my purview but there's still a ton to say about whatever System 1 dumps into our awareness and what we do with it.

And this still leaves two ways of reading Kant (should I ever get around to this), either as describing what System 1 throws our way or as the transcendental analysis of the conditions of possibility of experience -- in which case it can also be kind of a guide to what System 1 is constrained into coming up with.

Ha! I have a little tinnitus, which now that we're talking about it, I am experiencing again! (Tinnitus is famously mysterious this way.)

There's a point of Wittgenstein's that came up in the Moore's paradox thread that I found interesting: you can't take your own belief that p as evidence that p -- except when you can, but that requires moderately special circumstances. For instance, if you have, you know not how, picked the winner of every football game you've thought about for the last 11 years, you could take your, let's say, "intuition" that the Flamingos will defeat the Cormorants as evidence that they will. (Outlandish for clarity but ordinary might be informative.) But in normal circumstances, you don't, shouldn't, maybe can't treat your beliefs the way you treat things we do typically count as evidence. The way you can "read off" the state of the weather by stepping outside, you don't read off the state of the world from what you happen at the moment to think it is. Your beliefs are just your beliefs, and not evidentiary.

So I think this is the kind of thing LW is getting at, though it's perfectly obvious we can reflect on our inner state and say things like "I know what I'm looking at", "I know that I'm in pain", whatever, and in some conversational situations it's just the right thing to say. We say stuff. He just wants to block philosophers from thinking that the way we talk is a reliable guide to -- well, to any of the sorts of things philosophers like to talk about it. I'm putting that wrong, but I hate doing Wittgenstein exegesis.

Take counting and mathematical abstraction. After children count enough different kinds of things, they see that the relations of the numbers do not depend on what we count (abstraction) but on the act of counting alone. This is the basis for learning arithmetic. — Dfpolis

This is just the kind of thing I had in mind. Insofar as I have been pushing your position to look more like empiricism than you want, it's because I wonder how this works.

So if there are $n$ things on the table -- I can't even set up the question neutrally! -- we say there are $n$ things on the table, not $2^n$, because we don't count all the sets as things. This is the natural way to count, to us, but there is a view -- I'm going to call this Wittgenstein's view -- that all we can really say is "this is how we count" and there could be other ways. In particular, modern mathematics seems to require counting the sets as, well, "things", even while maintaining a distinction between element and subset. It might be tempting to say: by "how many things are on the table" we mean "how many elements does the set 'things on the table' have", but that's circular, and I don't see a simple way out.

Your adaequatio approach suggests that "how many things are on the table" means "how many of the sort of things we understand ourselves to be talking about are there on the table" and that feels right. But then we're back to this not being something a child could conceivably figure out through the exercise of natural reason but only in a context where the conceptual apparatus is already in place.

This is why your use of "awareness", glossed as "infallible knowledge by acquaintance", troubles me. I think some of this conceptual apparatus is just hardwired from birth and some of it we acquire not through the exercise of reason but through training by adults already conversant with a much more elaborate scheme than what they had at birth, and then some more we develop through the use of reason, and thus modify our own habits of conceptualization. However that side of the story works, a lot of the conceptualization of our experience is taking place below the level of awareness and thus without attention or reason. The way you distinguish knowledge from judgment suggests a similar hierarchy, it's just that "knowledge" in this case would map to psychological processes we are largely unaware of, and that sounds a little odd to my ear. Of course we have to call it something!

Yeah, I think there is overwhelming evidence that something like this must happen. I see what's on a chessboard dramatically differently from the way my kids do, and I know people who play a hell of a lot better than I do (or used to) see it differently still. System 2 as the minor leagues for System 1's habits is clear enough.

And I have a sense -- I think! -- of how kids can be trained to conceptualize as we do but it still feels like System 1's pump needs to be primed somehow, if not so much our individual System 1s, since we have adults to pour a cup of water in, then the "original", the core conceptual apparatus we've been passing down to our progeny for millennia -- and I don't see much alternative there besides "thank you, Darwin", which is not overly satisfying.

leaving what counts as a causal relationship vague intentionally — fdrake

A wise man ... (Yes, I always read your inline notes.)

You did a very nice job lining up simple material conditionals to necessity and sufficiency. You avoided counterfactuals and you avoided the word "because", both of which take some fancy footwork.

It often seems to me that when talking about proximate vs distal causes, people implicitly recognize that the last member of the set to show up has the role of completing the set -- like a "tipping point" state in an election -- so that it's not just necessary like all the others but, when it's the only one we're waiting on, it takes on the sufficiency of the whole set. That means what people mean by "cause" is often expressed by a conditional that is backwards from what "if ___, then ___" suggests. (And this tends to lead straight to counterfactuals.)

And this shows up in a phenomenon known as "perfecting the conditional". There's evidence (at least I think there is!) that English speakers tend to hear statements like "I'll give you $10 if you cut my grass" as "I'll give you $10 if and only if you cut my grass".

(( I was planning on doing this one, when I could get a moment, but I'm glad you got to it before me because you did it better than I would have! ))

Rather than reply point-by-point -- it may surprise you, but I'm not certain we're as far apart as it might appear, plus I'm not wedded to my position, plus I'm very interested in an Aristotelian take on all this, so -- rather than reply point-by-point, I wonder if I could get your view on what seem to be somewhat distinct questions:

1. How does a fluent adult speaker of a natural language perceive things as categorized in the terms of her native language, or in terms of the conceptual apparatus common perhaps not just to speakers of her native language but to many other people, or even all people?
2. How does a pre-linguistic child learn how to do this?
3. How did mankind begin doing this?

We're accustomed to say that "ontogeny recapitulates phylogeny" and thus in some ways expect the answers to (2) and (3) to be related, except we know there's an enormous difference, because children are trained to speak their native language by already competent adults. It is conceivable, therefore, that we could eventually come up with satisfying answers to (1) and (2) while (3) continues to elude us.

I'm not asking you to answer all three questions! But I am curious to know whether you think in terms at all like this, how you might see answers to these questions being related to each other, how you read Aristotle (and his interpreters -- thanks for the reading advice) as seeing the connections between these questions, or even whether the way we answer these questions has any bearing on this thread at all!

For repackaging I have in mind very familiar things we do in deciding how to carve up what we observe. If I show you a picture of a bunch of people and ask you what you see, there are a lot of right answers available! All of those answers are going to package the people in the picture (assuming that's even what you think about -- maybe you notice their clothes) in a certain way: a bunch of old folks, a bunch of white folks, a bunch of people. But suppose I then ask you how many women are in the picture? You might not have thought about this -- which may or may not mean you analysed the question and tried to guess what I was asking, depends on the circumstances. So now you need the "data" to be repackaged into [ the men in the picture ] and [ the women in the picture ]. Mostly you'll do that quite readily. But the categories used in the second version are still the province of System 1, your habits of identifying and classifying and conceptualizing, habits acquired largely through training to speak your native language; only if things get tricky will you settle down to analysing the data piece by piece, seeking justification for how you classify, etc.

I'm also thinking of slightly more specialized pursuits like painting. As I understand it, in introductory painting classes some considerable pains must be taken to get students to see the patches of color in their visual field: they tend to think apples are a uniform red all over because the object they know is kinda like that. With enough practice they can learn to overcome color constancy and actually see the effects of light and shadow and reflectivity that present the apple not just in shades of red but with blues and purples and whatnot. Again, even that repackaging is not the raw data, but matches up with the names printed on the tubes of paint!

Does that make sense?

It just says that they are not something we should talk about. — Banno

Except of course that we do. The question is what sort of hay philosophy can make of talk of our inner experiences.

So in §243 we find this:

But could we also imagine a language in which a person could write down or give vocal expression to his inner experiences -- his feelings, mood, and the rest -- for his private use? ---- Well, can't we do so in our ordinary language?

Does Wittgenstein turn out later to answer this "no" rather than the "yes" implied here? Been too long since I looked, but I can't think of a reason to say "no" myself. We have inner experiences and we talk about them in ordinary (i.e., public) language.

But then here is §246:

In what sense are my sensations private? --- Well, only I can know whether I am really in pain; another person can only surmise it. --- In one way this is false, and in another nonsense. If we are using the word “know” as it is normally used (and how else are we to use it?), then other people very often know if I’m in pain. --- Yes, but all the same, not with the certainty with which I know it myself! --- It can’t be said of me at all (except perhaps as a joke) that I know I’m in pain. What is it supposed to mean --- except perhaps that I am in pain?

Other people cannot be said to learn of my sensations only from my behaviour,--- for I cannot be said to learn of them. I have them.

The truth is: it makes sense to say about other people that they doubt whether I am in pain; but not to say it about myself.

I don't see any denial here that we have an inner experience of being in pain, or that this might be expressed by saying "I'm in pain". But he does want to deny that this is a cognitive experience, that that-I-am-in-pain is something I learn about myself, and something I could properly be said to know.

(I find the first part of the middle paragraph a bit of a head-scratcher: he is clearly not saying that others cannot be said to learn that I am in pain, know that I am in pain, doubt that I am in pain; they can, but I can't. I think you have to read the sentence backward: if we behave in a certain way to communicate to others what we know, and if me-being-in-pain is not something that I know, then my behavior cannot be how I communicate that I am in pain, and therefore you cannot learn that I know I'm in pain from my behavior. Still feels a little off.)

Okay, so for the topic of this thread: of course people have inner experiences we might properly describe as seeing the red triangular facing surface of an object; but it's not, or not normally, something they know, and certainly not something that by definition they infallibly know. The language of "sense impressions" or the like properly belongs to psychology, to a causal theory of our ability to use language, our acquisition of concepts, and so on. (Like the genetic learning algorithms you mentioned to me recently.) But there is another way to talk about what we know, and that's to do with reasons, justification, and the like; that's not a causal story.

The ambiguity of "because" and "why" make a mess of the distinction -- you could almost say foundationalism is exactly what you get if you systematically equivocate on the meanings of those words.

So meanings exist. — Olivier5

Here's a thumbnail sketch of an alternative approach you might find interesting. Sellars called it a kind of functionalism. The basic idea looks sound to me, but I haven't spent much time with yet. It does give a pretty natural account of translation.