We start with raw interaction. — frank

Well that's a question, right?

Your OP was based in part on the idea that this is something we do, that we're wired to do, to structure our experience as conversation. That leaves room for the idea that we more or less project this structure onto the rest of the world (or our experience of it). Projecting structure in this way is a mainstream issue for epistemology, most especially from Kant forward. And in this particular version, there are people we can point to who do project voices onto the rest of the world: we call them schizophrenics. That leaves a choice for how you proceed: do you propose a view that encompasses "hearing voices" as a variation of some kind on what we all do?

I want to focus on negotiation — frank

So start there. One way of thinking about a person's interaction with the rest of the world, is to note that the rest of world resists and surprises. If you falsely believe that boulder will never roll down the hill onto your house, when it does your belief becomes untenable.

So to keep with your approach, you tell the world what you're thinking and it might agree, disagree, or say nothing. Or something else. It has some input, and you revise your belief to take that input into account.

David Berkowitz believed his neighbor's dog told him to kill. Something has gone wrong in the negotiation here.

I think you'll also want to be able to distinguish between your dog telling you she wants to go out and your dog telling you to kill the president.

Could you say more about this? — frank

For an easy example, languages actually, literally evolve. No metaphor.

Do you have a Grice reading recommendation? — frank

Studies in the Way of Words.

Thinking of them in terms of equivalence classes adds an extra layer of complexity for little gain — fdrake

Maybe so, but it's a structure I'm fond of, and it gets close to my intuition of classifying as taking a shortcut. You see that this rock is similar to that one in ways that matter to you, so maybe what worked for moving that one will work for moving this one. (But maybe not for lots of reasons.)

It's a learned behavior. When kids are just starting to read (already knowing their letters and how to talk) they treat each word as a unique challenge to be sounded out or guessed at, even if they've already read that word several times, even on the same page. Learning to see words as tokens of types is learning how to take a shortcut -- say what you said last time and see if dad corrects you or makes you try again. (The move from attention to habit.)

Eventually we learn, one can hope, how to be careful with how we classify, with the cognitive shortcuts we take.

I'm on board. I think this could be a really fruitful idea to explore. Couple points:

1. Sometimes you can find that a metaphor or an analogy works because the things compared share an underlying mechanism that is literally, not figuratively, the same. This is always worth investigating.

2. From the other direction, there is Grice's theory that grounds "non-natural meaning", the way the English word "cloud" means [[cloud]], in "natural meaning", the way dark clouds mean it's going to rain. Most folks shy away from connecting those directly as he does.

3. Some of us do speak to things regularly. (When I find something where it shouldn't be, there's a fair chance I'll speak to it --"You're not supposed to be there," "What are you doing here?" -- that sort of thing.) Why do we do that?

4. There are lots of idioms that recognize what you're talking about: listen to what your body is telling you, numbers don't lie (i.e., they tell the truth), etc.

If there is a type-token distinction that is parsed in some way and not another, it can only be with an eye to doing something with it; one fixes distinctions in place so as to be able to make intelligible moves in discourse. — StreetlightX

I keep circling around what I think of as a thoroughly naturalist and nominalist approach something like this: the difference between, say, a particular triangle and an "abstract" triangle is not that the latter is a different sort of object at all. The "abstract" triangle is still a particular, the one you imagine, the one printed in the book or drawn on the blackboard. The difference is in how you handle it. If you ignore none of its particularity, that might be taking it, say, as a work of art. But if you ignore many of its particular features -- its particular materiality, the thickness of its lines, etc. -- then you can treat it as an abstract triangle.

@Nagase can answer this view though quite readily by pointing out that I am now relying on types or classes of actions to explain (away) types or classes of objects. Can I then try to explain these away following the same procedure? It looks like any attempt to avoid classes and types altogether is doomed to fail, even if we can avoid treating them as objects per se.

And there's an analogy here, perhaps more than an analogy, to the problem of talking about concepts. Frege himself makes the point several times that when you talk about a concept, you're treating it as an object, so you're never talking about the concept as concept. That you cannot do; you can only show how it works. (I'm also convinced that no Fregean should think propositions are objects.)

Where I want to end up is with an explanation of how the ideal actually comes to have a role in our lives. Grice speculates that maybe we never quite mean anything, in the strictest sense, but we approach the ideal of meaning something and deem that success. Lewis also, in Convention, reaches the surprising conclusion that maybe no one ever does speak a language in the Fregean sense -- again, we only approach this as an ideal. We huddle together in vaguely defined equivalence classes speaking languages that are near kin to each other. (And then there's Davidson: "We must give up the idea of a clearly defined shared structure which language-users acquire and then apply to cases. And we should try again to say how convention in any important sense is involved in language; or, as I think, we should give up the attempt to illuminate how we communicate by appeal to conventions." My bias is showing but how do you get away with that last clause in a paper that doesn't so much as mention David Lewis?)

That appeal to equivalence classes again looks like it demands something we've just said we can't have, real ideals, real types to ground the equivalence, and that backing off to our practices instead is no help. My suspicion is that it does help because the project of communal living gives you a choice: provisionally deem someone to be speaking a language you can understand or give up.

https://en.wikipedia.org/wiki/Law_of_noncontradiction

So if we say speakers of a shared language with a shared neurology — apokrisis

Should have addressed that. The full version is: "If you don't know that's red, either you don't speak English or there's something wrong with you."

And the actual story is that both levels of semiosis are constraints on our habits of interpretance. — apokrisis

That's the lines I was thinking along, but it's hard to say "When you talk about concepts you're also talking about how we use words" (details to be filled in) without it coming off as "When you talk about concepts you're only talking about how we use words." (And to top it off: when you talk about things, you're only talking about "our" concepts of things. Yuck.)

This tangent is strangely on-topic.

Not if you just assert that the apple is red to speakers with a shared neurology. No need to take things to the Whorfian extreme on colour perception. — apokrisis

I just can't get on board with the Whorf-Sapir thing. I know there's still controversy, but I like to think of it as refuted for color perception.

I think LW had a bit about "Because I speak English" being a perfectly good answer to "How do you know that's red?" A little like Austin answering "How do you know that tree is real?" with "Well, it's not fake."

In one sense this is just annoying, a sort of pretended obtuseness. So my reasonable sentence above ("This is red to speakers of English") turns out to be an explanation of how the word "red" is used, not an explanation of how perception works.

But I have some residual affection for this move, and I think it might be because it is sort of anti-Whorfian. There's a presumption that whoever you are and wherever you're from, you can't really be struggling with the concept of [red], so you must need help with the word "red". (There's something else here but I can't quite put my finger on it. Will mull ...)

When we say 'the apple is red' we don't mean 'the apple is red to me.' If Tom, god bless him, were to see the apple, he'd see red as well. But we also don't mean by 'the apple is red' 'the apple is red to tom and me'. — csalisbury

Hmmm. You seem to be working your way up to "The apple is red to speakers of English," which is not only not an unreasonable thing to say, but just the sort of thing people say when teaching a language.

The point is not that the type itself explains the relationship between its tokens. Rather, it is that in order to explain the relationship between the tokens, we will generally have recourse to some type, though not necessarily the type of which they are tokens. — Nagase

Can you take another run at this? This says that to explain the relationship between tokens we will generally have recourse to something that does not itself explain the relationship. The only sense I can make of that is that objects don't talk, people do.

But if you've ever seen an afterimage, the perceptual evidence seems pretty stacked up in favour of a relational logical form. — jkg20

I don't know... Even if you think of the paradigm of perception as being a relation between a perceiver and an object, something like P(S,0), afterimages aren't really like this, are they? It's not that the perceiver is pointing her perceiving device at a different object; it's that the device itself has a temporary modification. To get someone to see an afterimage, you even tell them to look at a blank wall or something. It's just there's an overlay now. So the perceiving relation might take an optional parameter, something like P(S, O, [X]), where X is some object you have an afterimage of. You could think of the X not as an object you're seeing at all, but as an indicator that you're seeing the wall in a special, X-modified way. Not much different from looking through a filter of some kind.

It could be that the initial impulse for the adverbial account is to capture the difference between changes to what you're looking at and changes to the equipment you look with, changes in what you see, on the one hand, versus changes in how you see, on the other. I don't know how well it works, or whether it can reasonably be extended to a general account, but a distinction like that seems reasonable and motivated to me.

Again, note the if: IF the phenomenal experience is ineffable, then it is irrelevant to the discussion. — Banno

Can I believe that the content of the experience, its qualia as the kids say, is ineffable, but the fact of whether the experience took place at all is effable, and relevant?

For instance, suppose I believe that what a piece of music sounds like cannot be put into words, not in English anyway. Whatever. I could still ask you to put on my headphones and listen to something and say if you liked it. I expect you to have an experience I don't expect you to be able to put into words, but you and I will have no trouble at all determining whether you had the experience. Is that inconsistent?

Who — Corvus

Hume's Dialogues Concerning Natural Religion is key text for a lot of modern chitchat on the topic.

Just trying to get a handle on this irrelevance.

If I invite you to look at something, I'm inviting you to have a particular phenomenal experience, aren't I? That phenomenal experience, according to you, is irrelevant to something, so I'm trying to figure out what difference it could make for you to accept or refuse my invitation.

If I complain that you didn't even look, does it make sense for you to explain that your phenomenal state is irrelevant?

Suppose you and I are driving down the road, you glumly staring out the passenger window. I, looking out the windshield since I'm driving, see a hot air balloon, and I say something like, "Wow, look at that!" Without turning, you say, "Yeah, that's awesome." (Maybe you're a sulky teenager in this scenario.)

Your phenomenal state is irrelevant, so it doesn't matter that you responded without turning to look at what I was remarking upon, right?

Hanover seems to think I deny internal experiences. I don't. I'm saying that if they can not be shared, spoken about, then they are irrelevant. — Banno

Irrelevant to what, is the question.

The upshot of the above was supposed to be this: whatever types are for, it can't be for deciding whether two objects count as the same sort of thing.

That just leaves the examples that involve talking about things in some general way. Some of the standard examples can be readily dealt with as a kind of shorthand (where we're in essence just talking about extensions), and eliminated, at least for countable domains. That just leaves the mathematical examples that even Quine, great foe of second-order logic, thought required sets as first-class objects.

Math is always the odd case.

Whatever all the objects on the table might be, we're rarely tempted to count as objects, just like those, the various sets of objects we could conjure. (So that instead of three objects there would be eight?) Whatever sets are, they're different.

What is it that makes it the case that this particular inscription is a token of, say, rouge? — Nagase

But what role does the type play in determining whether two given inscriptions are (intended to be) tokens of the same type? We can imagine an effective procedure for comparing two inscriptions directly and determining whether (following some community standard, ignoring differences of typeface, for instance) they're intended to be the same.

Type plays no role in the comparison. How could it? If it were necessary instead to compare each inscription to an abstract type, rather than comparing them directly to each other, then we would seem to need some meta-type to enable comparing the given token to a type. We'll never get there.

What is the "crack whore conundrum"? — Bitter Crank

A manuscript Ludlum's publisher rejected?

A DST functions like the expression 'the lion' in the sentence 'the lion is dignified': the singular term 'the lion' refers distributively to particular lions existing in space and time: hence, a distributive singular term. — StreetlightX

That looks a whole lot like what others would call the extension of the predicate "... is a lion."

I wonder if there's a place here for the medieval distinction Lewis revives in Convention:

If I expect every driver to keep right, in sensu composito, then I have one expectation with general content. I expect that every driver will keep right. It does not follow that if Jones is a driver, I expect that he will keep right, for I might not realize that he is a driver. Indeed, I might even realize that Jones is a driver and still not expect that he will keep right, for I might fail to draw the proper conclusion from my general expectation. If, on the other hand, I expect every driver to keep right, in sensu diviso, that I have many expectation, each with nongeneral content. I expect of Jones, a driver, that he will keep right. Of Morgan, too. And so on, for all the drivers there are. I need not know that Jones, Morgan, and the rest are all the drivers there are; I might falsely believe there are other drivers who do not keep right. Or I might altogether lack the general concept of a driver. Generality in sensu composito and generality in sensu diviso are compatible and often coexist; but it is possible to have either one without the other.

(Lewis will argue that the expectations that underwrite convention are primarily general in sensu diviso.)

Oh it's not technical at all. He's a far more accessible writer than Sellars, but then almost everyone is.

Wow, that sounds fascinating — frank

Sarcasm?

Do you just type really fast, or are you in fact an AI?

But what is pressure? Is there such a thing? — frank

Pressure is a nice example because statistical mechanics was invented for just this purpose, and would be one of the prime examples of explaining how an aggregate entity -- some given volume of a gas -- behaves based on understanding how its constituents behave, the many individual gas molecules. Sociologists and economists dream of being able to pull off something like that! (Full disclosure: just started reading Schelling's Micromotives and Macrobehavior.)

We're headed rapidly toward stuff I don't understand. If you ask next what kinetic energy is, and whether there really is such a thing, I'll be shrugging.

My inner nominalist keeps saying the problem is there's no such thing!

There are a lot of issues that bear a resemblance to this one: the debate in the social sciences over methodological individualism leaps to mind. The problem may not be so much with talking "as if" social groups have a kind of agency you know they really don't -- maybe that's just a harmless shorthand -- but with missing the mechanisms by which the actions of individuals add up to the "actions" of a group.

The suspicion, in a general way, would be that explanations that lean on abstractions in the wrong way are not explanations at all. What exactly happens when you turn up your recessor coils? They're more intense. Okay. More intense how? What does that actually mean? The answer better not be, now they have more intensity.

Notice that it won't do to just say that there are no types, just inscriptions that we are regularly disposed to react in a certain behavioristically specified way, since there must be a regularity for us to respond to. That is, there must be something that accounts for the similarity of the various X inscriptions, and this can't be just our behavior, as we are dispositionally inclined to regularly react to Xs because they are similar, not the other way around. — Nagase

But language is a much messier affair than this. In a language such as English, there is a considerable range of sounds that count as a given phoneme. Not just anything, but also not all that sharply circumscribed because we change what will count based on context. There are allophones allowable when singing that would seem strange in everyday conversation. Toddlers utter sentences in which the prosody is right and just a couple of the phonemes are close to standard, and that counts. You use different allophones when whispering or screaming, and so on.

Paul Grice tells a story about an Oxford college that hired a new don they were very excited to get. The only trouble was he had a dog, and dogs were forbidden. So they held a meeting and "deemed" his dog a cat. Grice then wryly comments (this is all apropos his theory of meaning) that he suspects we do a lot more deeming than we realize.

That's one side: the requirement for regularity is not particularly strict, and is responsive to the project of communal living. On the other side: would natural regularity account for our behavior? Whence our disposition to respond similarly? There is still a logical leap in counting numerically distinct objects howsoever similar as "the same", as members of a class, tokens of a type, or exemplars of a property. We must still deem them the same.

I think it comes down to the nature of abstraction, and to what we ignore. A bee is attracted to a certain sort of flower, not to some one particular flower. Nature operates with types. But what does that mean? The bee must still sup from a particular flower rather than a type of flower. The particular flower has to have certain properties and the rest -- its exact height, position, blah blah blah -- are ignored. The required properties in turn must fall in some range, rather than having one specific value.

Sorry to ramble, but I find types thoroughly confounding.

What matters, in other words, is the 'matter-of-factual' (graphic, inscriptual, physical) relations between X, Y and Q, which show but do not say, the relation between X and Y. Which is another way of saying that Q does not denote and does not signify. It is a physical pattern which we respond to in rule governed ways. — StreetlightX

I'm almost certain Quine mentions Church when he introduces "virtual classes" (I'd forgotten that's what he called them). What caught Quine's attention about Church's lambdas is that they're anonymous: if you write "(λx . x is red)(that apple)" you're doing something a little different from writing "red(that apple)" -- as Frege might have -- because in the lambda version there's no function called "red" at all. You just concatenate the indicated strings in the indicated way. (You can give functions names, as Lisp does, for convenience, but it's in essence just a kind of shorthand.) Pretty clear why this would be appealing to a nominalist: no temptation to hypostatize when you don't have a name to work with.

Yeah, commitment and entitlement make a nice pair of terms, because there are natural points of contact with your speech community there: what you've committed to and what you're entitled to are clearly not entirely up to you but negotiated, so to speak, and over longer and shorter terms, with other members of your speech community (which I've come to think of also as your epistemic community). All of that fits nicely too with the game theory analysis we get from Lewis of reciprocal expectations, which is where the "contract" comes in. Good stuff.

I'm not sure there's any harm in adding belief talk here. But if the idea is that explicit appeal to induction will explain everything, that might not work out.

Say your cat has formed an inductive generalization that A is always followed by B, and given A concludes that B must be coming. That seems to explain the expectation of B.

But what explains your cat making the inference at all? The cat's belief that A is always followed by B, together with the appearance of A -- still gives you nothing. Either the cat makes the inference for a reason -- hard to see that working out -- or the cat is caused to make the inference. No explanation in the offing for that.

So if bare expectation is considered unsatisfactory, shifting to belief talk instead leaves bare unexplained inference. Maybe that's progress, I don't know.

which I think should be called 'expecting' — Janus

It's a nice word. You can do a whole lot with just "expectation" and "preference" and translating everything else into those two.

I don't think that it poses any problem for my stance. — Sapientia

Wasn't meant to. Just chiming in. Endless one-on-one around here can get tiresome.

Let A = some state of affairs (whatever that turns out to mean)

Philosophers would usually call "that <A>" a fact, by which it is meant that A obtains. This is a little unfortunate, because it makes the grammar of "<S> believes that <A>" puzzling.

If someone S has a belief regarding A, it could be that A does or does not obtain, that A is possible or not, likely or not, etc.

It seems most natural to call the object of S's belief this obtaining or not obtaining of A (or its possibility, likelihood, etc.), rather than A itself, but a little ambiguity here might be harmless.

As I suggested a long time ago, it might be fruitful to treat A as something like the indirect object of S's belief: S believes of A that it obtains (is possible, is likely, etc). You can do this with the elements of A as well, something like, "S believes of the cat that it is on the mat."

(Two reasons to lean this way: we can separate the referential and predicative functions of the proposition; it's one simple, not overly unnatural step toward the lambda calculus -- (λx . x is on the mat)(the cat) -- and we like lambdas.)

Gandalf isn't named "Power Wizard" for instance. — Bitter Crank

Pointless necro here:

Actually he more or less is. Look up where Tolkien got the name "Gandalf".

Just chiming in to say I think "commitment" is the magic word here. This is exactly the word I was about to reach for over in the "Belief" thread to explain my sense of beliefs as something like rules or norms you follow in thinking and acting.

And I think of commitment as placing your bet, or running your experiment. There's a strong current of pragmatism running beneath all this that I find increasingly appealing.

Would you say it's unreasonable to say that John actually was an atheist due to the lack of behavioral consequence? — frank

Oh I see no reason not to include what he thinks. There may be no outward behavior, if that's what you mean, but I wouldn't demand that. You can keep a belief secret.

Could you tell me why you're pointing this out? — frank

Sorry-- you suggested an analogy between believing and slapping. I think there is only a surface, grammatical similarity there.

But aren't those more the consequences of belief? — frank

I don't have a full-blown theory to offer, but I think this is the right stuff to look at.

When you walk down the sidewalk in a big city, you're behaving as if the buildings you walk by won't fall on you, as if the cars you walk by won't explode, and so on. We could you say you behave as if you hold such beliefs. Do you? If asked, you might assent. Would we say such beliefs cause you to walk down the sidewalk, or even that they are reasons for doing so? Doesn't sound quite right. And yet attributing such beliefs to you makes sense. And if you did not hold beliefs such as these, would you behave the way you do?

Given your belief that DJT is a good man, you might also vote for him, campaign for him, give him money, etc.

Btw, grammatical form isn't necessarily logical form. Classic example:

• The king's carriage was pulled by black horses.

vs.

• The king's carriage was pulled by four horses.

"Black" is a type of horse, in some sense, but "four" isn't.