One italicized word
Here's Frege:
That's from "On Sense and Reference," and the emphasis on "the" is Frege's. Such passages are everywhere; here's one from "Thought":
Again, all emphasis Frege's.
I find this to be pretty convincing stuff. Whatever else you want to say about it, this is what logic seems to require, much like the distinction between object and concept.
But now here's Grice from "Meaning Revisited", emphasis his:
Look at that: "a" not "the".
I have loads to say about that one suggestive little change to Frege's account, but I'm curious to see what other people think first.
The same sense is not always connected, even in the same man, with the same idea. The idea is subjective; one man's idea is not that of another. [...] This constitutes an essential distinction between the idea and the sign's sense, which may be the common property of many people, and so is not a part or a mode of an individual mind. For one can hardly deny that mankind has a common store of thoughts which is transmitted from one generation to another.
In light of this, one need have no scruples in speaking simply of the sense, whereas in the case of an idea one must, strictly speaking add whom it belongs to and at what time.
That's from "On Sense and Reference," and the emphasis on "the" is Frege's. Such passages are everywhere; here's one from "Thought":
If other people can assent to the thought I express in the Pythagorean theorem just as I do, then it does not belong to the content of my consciousness, I am not its owner; yet I can, nevertheless, acknowledge it as true. However if what is taken to be the content of the Pythagorean theorem by me and by somebody else is not the same thought at all, we should not really say 'the Pythagorean theorem', but 'my Pythagorean theorem', 'his Pythagorean theorem', and these would be different, for the sense must go with the sentence.
Again, all emphasis Frege's.
I find this to be pretty convincing stuff. Whatever else you want to say about it, this is what logic seems to require, much like the distinction between object and concept.
But now here's Grice from "Meaning Revisited", emphasis his:
First, the operation of such creatures as I have been talking about is at least in certain circumstances going to be helped and furthered if there is what one might think of as shared experience. In particular, if psychological states which initially attach to one creature can be transmitted or transferred or reproduced in another creature (a process which might be called ?-transmission), that would be advantageous. Obviously, the production of communication devices is a resource which will help to effect such transfers.
If one accepts this idea, then one could simply accept that for the process to be intelligible, understandable, there will have to be correspondences between particular communication devices on the one hand, and psychological states on the other. [...] Whether direct or indirect, the correspondences would be between utterances or utterance-types on the one hand, and types of psychological states on the other, where these would include, for example, the belief-types to which the beliefs of particular people belong: not Jones's belief that such-and-such, but a belief that such-and-such.
Look at that: "a" not "the".
I have loads to say about that one suggestive little change to Frege's account, but I'm curious to see what other people think first.
Comments (53)
If it is the same for anyone who can understand it then it is common to all who think. What I think is interesting, is that it is both true, and also ideal, as are a great many mathematical and geometric ideas. That is, they're common to all who can grasp them, but they can only be grasped by a rational mind. So in that sense their reality is intelligible rather than corporeal - which is rather similar to the outlook of objective idealism. (See for a discussion of 'intelligible objects' this passage on Augustine on Intelligible Objects.)
Compare this passage from Ed Feser:
Some Brief Arguments for Dualism
As for Grice - isn't his argument a form of psychologism? He is attempting to provide an account which rests on so called 'psychological states', whereby such states then underwrite or form the basis of ideas, as such. Which, I suspect, is ultimately in the service of some form on naturalism, although it's hard to say on the basis of the above.
Maybe? But the main problem with psychologism is just that Frege has such a strong argument against it.
Compare the example of phonemes (or cheremes) which are explicitly defined as equivalence classes of sounds (or gestures). You could think of propositions, for instance, as equivalence classes of utterances, thus utterance-types, and thoughts as equivalence classes of psychological states, e.g. belief-types.
One question is, how do you get these equivalence classes rolling? What is required to be able to take something as a member of a class, or as a token of a type?
But where we want to start is the observation that, in uttering an allophone (within my speech community) for /a/, you are taken to have uttered the phoneme /a/. It might be helpful to put off the question of universals for a little while and look at how that transaction works.
I haven't quite grasped the distinction yet. Grice speaks of 'the belief-types' and 'the beliefs of particular peoples'. Isn't that use of 'the' the equivalent in his nomenclature to Frege's 'the sense'?
Maybe? But why go through the type business at all? Why not just say, as Frege does, that we each have the belief that such-and-such?
Oh, it's not that difficult to deny that.
I'll bite, against my better judgement...
Whether it's difficult remains to be seen. It's clear enough what Frege gains by not denying it; what do you gain by denying it?
What you gain is that you say something that's true rather than something one would simply like to be true. ;-)
"What you gain is that you say something that's true to you rather than something one would simply like to be true to one."
Fixed that for you.
That goes without saying. Of course, the reason it's true is that in my judgment, the proposition matches facts.
Then why ask me in the first place?
Just stick with inquiries directed at folks whose judgment you're interested in.
You misunderstand. I was not insulting or dismissing you.
On your own view, you cannot tell me the truth, only your truth. No matter what I claim, my truth cannot contradict your truth. (Something happens in my brain; something happens in your brain. Period.) It is in that sense that your truth is and can be of no consequence for me. Even if you wanted to provide an argument for why I should take your truth into account, how would you proceed? We are incapable of both assenting to the same premise.
All that anyone can tell you is propositions that match facts in their judgment. And that's what I did. We can and certainly do have different judgments, and we can't somehow get beyond the fact that we're making judgments about how propositions link up with facts. Propositions can't somehow match up with facts or not independent of us. Meaning is something that we do as individuals. Objectively, the sentences that we count as propositions are just text marks or sounds.
You could only care about your own judgments about the relationship of propositions and facts, but that's probably not the case since you're participating on a message board like this.
My mind was blown the first time I came across a person who even considered the notion that an idea might be somehow owned by the stuff in an individual's skull. And Frege was part of the discussion... it was about abstract objects like numbers.
How would Frege's view fit with a platonic account?
Leaving aside whatever there's room to debate, Frege's a pretty thorough platonist.
The owning ideas thing he explains with a telescope: there's the actual object out there it's pointed at, say the Moon -- this will be the reference -- then there's the image on the mirror, which is not the object but is still objective -- that's the sense or the thought -- and then there's the retinal images of whatever individuals look through the telescope, which are subjective and unshareable, and that's what Frege calls ideas.
I was thinking about meaning today. I have a problem with the concept of a sign. It's supposed to be a signifier/signified combo. I don't think an isolated sign has any meaning, though. I think it has to appear in a complete thought (a complete sentence?) in order to be meaningful. Could be off topic?
Scenario (1): A and B have a box of propositions, and they each have opinions about which ones are true. They take turns sorting them into boxes marked "true," "false," and "not sure." Maybe neither of them have some special status that allows them to know what is true, but they can at least see the different ways they sort the propositions. A might be surprised to see B put something in the "true" box that he wouldn't, but B might convince A that he should, because of some others he put in.
Scenario (2): A has his own box of propositions and B has his own box. They might as well sort at the same time, without paying any attention to each other, and they can even share the boxes they sort into. Doesn't matter. What difference could it make to A what B does with his box of propositions?
I took you, perhaps mistakenly, as going for scenario 2, rather than scenario 1.
A third scenario you may find more congenial is suggested to me by Grice's talk of types.
Scenario (3): It's more like kids each sorting their own collections of baseball cards. They can each have a copy of the same card-- not numerically the same, but same player, year, series-- and they can have cards that they count as the same in different ways. "Do you have a Clayton Kershaw?" can be answered "yes" whichever one of the various Clayton Kershaws that have been issued you have.
Scenario 3 is more appealing than 1 in some obvious ways, so long as we can make the type stuff work.
That's the idea. It's not the object referred to but still objective.
That would be a version of Frege's context principle. It can get a little weird.
Oh yeah--am I a platonist? Not by temperament. But I find it hard to talk about language, logic, and mathematics without drifting toward a Fregean sort of platonism. I'm not quite convinced that means you have to be what's usually called a platonist, but there's something there that has to be taken seriously.
Objective? Third-person data as opposed to first-person data?
Quoting Srap Tasmaner Weird how?
Quoting Srap Tasmaner
I don't know if you've really taken it seriously unless you've pondered how it fits into the bigger picture.
If by "third-person" you mean public, then I think yes.
There are various ways of formulating contextualism and some of them conflict with compositionality. I can't imagine giving up compositionality. I don't even know what the alternative would be.
I'm not sure what this means.
Or maybe you meant mind-independent. The reflection of the moon in the mirror is there whether anybody's looking or not.
Quoting Srap Tasmaner
I think I was arguing against compositionality. The parts, to the extent that they have meaning, gain that meaning from their place in the whole.
Quoting Srap Tasmaner
I meant how some form of platonism fits with your overall ontology. Dualist? Monist? Both?
I'm not a Gricean, but I@m trying to follow the logic of Grice's thought. He goes through the type business because he's trying to find good generalisations. Partly he's trying to solve one problem of Fregean 'sense', which is that it hovers in no-man's-land, an inbetweenie:
So I think this shows Frege not merely saying that each of us does something unique, but rather saying there's something intermediate: that 'sense' is something some but not all of us will share. Grice's version of that is 'implicature'. Both of them, 'sense' and 'implicature', seem to me interesting but dodgy concepts: they are the work of analytical people trying to pin down something slippery, contextual and often feeling-based or feeling-related. What do you think is the right way to look at it?
First, it seems like you're still thinking about "true" (and "false") as something other than a judgment we make, as individuals, about propositions and their relations. That's because you're saying things like "some special status that allows them to know what is true." If you were using "true" as a synonym for "(a particular sort of) judgment," and then made a substitution, you'd see that you're saying "some special status that allows them to know what judgment they're making."
Aside from that, yes, I'm saying something much more like (1), although of course it depends on the person whether they care what other people are doing or not. But (1) is clearly what we do most of the time when we're interacting with other people philosophically, for example. We wonder why someone is saying that something is true or false when we clearly reach a different conclusion. We wonder if they're not using words in some obviously different way, etc.
And re types, yes, I'm saying something more akin to your Grice example.
In a way, "no man's land" is exactly the right phrase, because nothing here is the sole and unshareable property of any man. I can understand why people think stuff out here "dodgy," but just look at what's here: meaning, information, patterns, mathematical objects, transitions, tendencies, dispositions, institutions, -- I could go on and on and on. We may nurse a view that we are particulars and all we ever really, in whatever sense you think you can make that work, talk about are other particulars, but I think every time you open your mouth you make use of stuff in no man's land. I think it's rather the point of language.
"Mind-independent" is not a phrase I have any use for, I think.
Re: compositionality, I don't see how you recursively generate expressions without it.
Re: my ontology, I don't have one.
OK. "Mental object" is what math people call the "idea" Frege speaks of. That's as opposed to "abstract object" which I suppose is his "sense."
I don't know if you know the paper where David Chalmers argues for a contemporary Fregeanism where 'sense' pretty much becomes 'intension'. Here it is.
But the question is, what are A and B making judgments about?
Frege has a clear answer to that: the proposition, the thought, which is objective. I'll grant that this is basically a posit, but like any posit it serves a purpose. If A and B disagree about whether a proposition is true, they have to assign different truth-values to one and the same thing. That thing cannot be any particular inscription of the proposition, but the proposition itself.
The thought expressed by a sentence is also what Frege says you get when you understand the sentence, and you get it without remainder. It is what is communicated, what is transferred from A to B, what A and B can have different opinions about. This is the idea behind scenario (1) in which there is a single, shared, publicly available box of propositions for A and B. It's what you said it would be easy to deny. (Starting to feel a little icky about talking about propositions as if they're objects.)
So the question is whether scenario (3) can be made to work.
As is, it's just an intuition pump, right? I mean, baseball cards are manufactured; they are by design identical. The analogy is going to fail almost immediately. The questions that replace the built-in identity are a little problematic: what would make two utterances instances of the same utterance-type, two beliefs instances of that same belief-type? What's a type? It feels like you need something from scenario (1) (or nearby) to get this going.
Here's what I'm tempted to do: agree with Grice that this is what happens-- to talk about the tree, we need each to have a belief that the object we're looking at is a tree, not the belief. Don't posit, not yet anyway. (The idea is to avoid using Frege's machinery at all.) Accept that what we have here is all we need to talk about the tree. Then look for an explanation for how two numerically distinct beliefs can count as beliefs of the same type right here, in the transaction between two members of a linguistic community. We honestly don't need them to be instances of the same type, not for this part, although it's pretty obvious why that would be helpful. Right now all we need is for A and B to agree to treat their numerically distinct beliefs as instances of a belief-type.
Grice is almost certainly going to get here with a (probably infinite) chain of intentions, so that can get a little weird.
I'd like to come at it sideways, by the comparison with phonemes. How does someone "decide" that the allophone you actually utter will count as a /d/? This is already a little wrong, because the range of allophones is itself already determined by the speech community. It still looks like we're trying to figure out how conventions work.
One shot at this might be this: when you utter a sound, I have to take it as an allophone of some phoneme we use in our speech community or not. If possible, I'll take it as one of ours, because (a) intentions, and (b) why not? You can provisionally, experimentally take the sound as a phoneme. Which one? Again, you have to decide whether that phoneme with the others around it make a morpheme, and again if possible you will, because (a) intentions, and (b) why not? You do that provisionally and experimentally, all the way up to the complete utterance, and see if it seems to work. I'd say there's a tiny bit of evidence we do this in the way we read over typos, mentally substituting the right letter because we're pushing toward taking the utterance as valid. You could think of this as the principle of charity, but you might also wonder what choice we have but to proceed this way.
Does this actually work? Has any of Frege's machinery been smuggled in here anywhere?
I think that's right, bearing in mind that he's going to take mathematical objects as, well, objects, just like physical objects. Sometimes he describes the sense of a complete (i.e., referring) expression as the way the object referred to is presented. Example: "2 + ..." is an incomplete expression, a function. Put an object in the blank, and you get a complete expression like "2 + 3". "2 + 3" refers to 5, but not the same way that "5" refers to 5, or "7 - 2" refers to 5. This is supposed to explain why equations can be informative. "2 + 3 = 5" tells you that the references of the two expressions are the same, but it remains that "2 + 3 = 5" expresses a different thought from "5 = 5" or "7 - 2 = 5". The thought expressed is the sense.
I do not, and thanks for the tip!
Also Soames presents an awesome argument for why we can't dispense with propositions (a sort of netherworld object) without denying that there is such a thing as agreement. It's in Understanding Truth.
I think I've been pushing a Frege-inspired version of this in chatting here with Terrapin. Certainly positing propositions do give you a way to agree and disagree, etc. So they're certainly sufficient, but I want to see more clearly whether they're necessary, which is what I'm working at above. I do wonder whether starting from the conditions of communication could eventually get you a version of Frege's machinery.
Insofar as truth goes, they're making judgments about the relation of a proposition to something else.
I state that as "something else" because not everyone uses the same "something else." That depends on the truth theory that someone subscribes to in the sense of correspondence versus coherence versus consensus etc.
Propositions, as the meanings of the sorts of sentences that can be true or false are not objective on my view, because meaning isn't objective. Of course Frege posited that they were objective, because Frege was anti-psychologism . . . which in my opinion was one of the dumbest moves that philosophy ever made. Not that that was only Frege's fault. I just mean the move away from psychologism in general.
If we're talking about the correspondence approach, the judgments are about the relation of a proposition to facts. Most facts on my view are objective.
So how do people compare judgments? I judge P[sub]me[/sub] true, you judge P[sub]you[/sub] true. We're not even talking about the same proposition. (In fact Frege argues that would actually be me judging P[sub]me[/sub] true[sub]me[/sub] and you judging P[sub]you[/sub] true[sub]you[/sub].)
As I said, if you can establish that P[sub]me[/sub] and P[sub]you[/sub], if not instances of P simpliciter, are members of some equivalence class (which we could then define to be P if we wanted), then you would have a meaningful way of comparing my judgment of P[sub]me[/sub] and your judgment of P[sub]you[/sub].
Until you do that, it's just me saying my apple's red and you saying your banana's yellow.
Are you asking me to literally report what we do? What we do should be obvious if you spend time talking to other people.
It's not like you saying your apple is red and me saying that my banana's yellow. It's like you saying your apple is red and me saying, no, you're apple is purple, where for all we can tell at least initially, we're both using the sound "apple" to "point at" the same objective thing, we're both using "red" and "purple" to "point at" the same objective things, etc.
Do you say "No, your apple is purple" because for all you can tell at least initially, when I said "My apple is red," I meant what you would have meant if you had said, "Your apple is red"?
Could I make a suggestion here, which is that the term 'objective' is somewhat misleading in this context. This is because mathematical proofs are not actually 'objective' in the sense of being 'inherent in an object or situation'. But they are not subjective, either. I think the problem here is that our modern use of 'objective' entails a certain class of truths, which is subtly different to a priori, rational or logical truths. I don't have a suggested alternative to 'objective' but I am just pointing out that I think it's a poor descriptor for the kinds of truth that Frege is wanting to elucidate.
This had not occurred to me. I think overwhelmingly I use objective/subjective to mean something like public/private, just because of the contexts in which I'm making the distinction. For instance, here the idea is that when you understand a sentence you have grasped something that anyone can, thus something public, as opposed to whatever images and so forth the sentence might call to your mind and your mind alone, which would be private.
'Common to all who think', would be a way of putting it. The same as saying, '7' is the same for anyone who can count (and meaningless to anyone who can't) That's why I drew attention to the example of 'the triangle' - that is a concept, which must have the same meaning for any observer. But what I'm wanting to say, is that those kinds of facts are of a different order to objective judgements. We will, for example, call on them when wishing to make an objective judgement - 'she said there were only 5 left, but I counted them, and there were definitely 7'. We rely on numerical judgements all the time to assess what is objectively the case, but numerical reasoning in some sense precedes objective judgement; objective judgement relies on our ability to count or to make rational judgements. That's the sense in which I'm saying that numerical judgements aren't 'objective'; it's not as if they're 'subjective', either. I think they're something more like 'transcendental', in the Kantian sense.
That is why we have an overwhelming tendency to defer to science when it comes to adjudicating what is or is not a matter of fact. Scientific judgement at the end of the day deals with matters that can be made subject to quantitative analysis. Qualitative questions are of a different order; how do we measure them? The rules of maths, and the rules of logic, can be brought to bear on almost any subject. This is why scientific judgement can be said to be 'public', in the sense that it comprises findings which produce measurable data that others can observe (notwithstanding the so-called 'reproducibility crisis'.) And that underlying assumption is buried very deep in our naturalistic culture.
Sorry if I'm rambling. I'll leave it at that for now.
This is absolutely true of course, and you may have to narrow the context all the way down to the occasion of utterance, and even then you may have to appeal to the intention of the speaker to disambiguate an expression.
There is something mechanical about this process though, which may be why it's of slightly more interest to linguists than philosophers. (Perhaps wrongly.)
1. Does communication presuppose complete disambiguation?
2. The real trouble seems to come once disambiguation is done, assuming it can be: when I understand something you say, have I acquired the content of your utterance, as a sort of payload?
I really cannot imagine doing philosophy by deciding ahead of time what I'll quantify over, if it comes to that. I have the same physicalist or naturalist prejudice most philosophers do, but that only decides all questions if you also believe in a form of reductionism that looks pretty suspect. Sometimes you're stuck with your theoretical entities.
Other commitments are certainly a factor in choosing one theory over another, but I'd say the main thing is always explanatory power: does the theory make sense of our collective intuitions? does it clarify murky cases? does it include what it should and exclude what it should? The opposites of those (and whatever else goes in there) are bad.
Should also have said something here about category mistakes.
That's cool. Could you give an example of the type of thing you want to explain?
I'm not usually too impressed by explanatory power. Explanations come and go. Each has some power, I suppose. I'm more drawn to a geometric approach. Pretend I'm an eliminative materialist. What would I have to conclude about the presuppositions of communication? What would I say about content?
What's the opposite of being eliminative? What would I see if I stood in that position? To make a square out of it, what view partakes of both of the above opposites (there should be two of them to make a square)?
Yes to both questions. Again, isn't obvious that that's how people operate in these situations?
There's the sentence I actually utter: "My apple is red."
There's the sentence you imagine uttering: "Your apple is red."
Do they have the same meaning? Express the same proposition? Are they equivalent in some other way?
I'm just still trying to figure out what all this means. Maybe if you clarified what you mean by "objective" and "subjective" -- I may have guessed wrong -- that might help.
I believe I understand how Frege's view works; I don't understand how your view works. If you want to just explain it, that would be fine.
We have to clarify what we're asking re whether the have the same meaning. Obviously, as a nominalist, I don't believe that they're literally the same. We're not talking about a numerical identity. (And I wish we didn't have to clarify this, but some people are confused into thinking that they must be numerically identical.)
But we assume that they're similar enough that they might as well be the same until there's a good reason to believe otherwise.
Quoting Srap Tasmaner
I've done that many times, but once again: subjective = mental phenomena, objective = the complement --things aside from mental phenomena.
That's not far from my earlier suggestion for how we can Grice's types rolling, but I still think that this similarity needs grounding, and we probably want a little more than an assumption to do it. And we need to explain how similarity does us any good.
Here's how I see the dilemma.
Option 1 (Frege's): propositions, meaning and truth are not subjective mental states or events.
Pros: meaning and truth are public and shareable; communication works as advertised -- understanding is grasping the same meaning as the utterer; logic works as advertised -- if A asserts P and B asserts ¬P, they're talking about the same thing.
Cons: entails a third (platonic) realm of entities (?) that are neither physical objects or subjective mental states or events.
Option 2 (psychologism): meaning, truth, etc. are subjective mental states or events.
Pros: does not entail the third realm.
Cons: communication and logic do not "literally" work as they do in Option 1: A and B cannot be in the same subjective mental state, thus A and B cannot "literally" understand each other's utterances, cannot both assert or deny the same proposition, etc.
You can of course just plump for option 1 or 2 and accept the consequences: accepting option 1 entails accepting a third realm many find implausible; option 2 leaves you hanging out with the freshmen asking, "How do I know your blue is the same as my blue?"
If that's not enough, there is further motivation for crafting a third option: there is a sense in which Option 1 explains nothing, but simply redescribes what we want to explain, with the needed theoretical entities (meanings, propositions, etc.) and a framework showing how they are related; Option 2 goes wrong not by relying on mental states and events, but by not engaging the theoretical framework of Option 1 at all.
We could modify, or clarify, Option 1 somewhat: it's platonic entities people are hesitant about, and it's not really clear what Option 1 is committed to in the way of entities. (Elsewhere, Frege is committed to numbers as objects, etc.) Propositions and concepts are not treated by Frege entities at all. But is the sense of a proposition? We talk about it as if it were, but perhaps there is a way of refining our presentation of Option 1 so that the population of theoretical entities is smaller and more acceptable.
We could modify Option 3 along the lines contemplated earlier in this thread, by gathering utterances and mental states into types or equivalence classes, with the intention of plugging this into the framework of Option 1 in place of the theoretical entities there. We somehow already do something like this with phonemes (or cheremes), for instance.
Re option 2, I don't see your cons as cons. We simply have to have theories of communication, understanding, etc. that reflect what's really going on given that 2 is the case.
Re how it's known that the blues are the "same," it's not something that can be known, but more importantly, it doesn't matter that we can't know this. There is nothing of practical importance that hinges on knowing this.
Re your third option, yeah, we can use the standard way of talking about this stuff, contra psychologism, as a useful fiction. There's nothing wrong with that as such, as long as we acknowledge that it's just a fiction that we're engaging in to make it easier to talk about the topics at hand. Hence why I'd normally talk about the same meaning, say, without explaining nominalist issues, etc.
I think psychologism is prima facie implausible as an account of how we talk about mathematics, for one thing. Now the psychologismist, if they weren't just going to deny this -- I expect you will -- could respond that Frege's machinery was developed especially to formalize mathematics, and so there's no surprise that it works there, but also no reason to think it works at all anywhere else. But then the question is, what's different about mathematics? If the response is that mathematics is just convention, that it's all true by definition, something like that, that leaves unexplained how such conventions could possibly arise, conventions for which Frege's account does actually work. And if you could have such conventions as the basis for mathematics, why not for other things, why not for natural language?
Quoting Terrapin Station
Talking as if something were something else is very close to something counting as something else, and I still want to know how that works. As I've said, I think there's a kind of start in the way phonemes work -- there's a whole range of sounds that will count as the phoneme. That involves selecting certain features and ignoring others. Pitch, for instance, is irrelevant in English.
That "selecting certain features" part makes it sound like we're headed right back toward the Fregean machinery. But maybe not, or not only that. At the very least, we're talking about counting numerically distinct objects or events as instances of the same thing, and in a sense it doesn't matter how "objectively" similar or different they are -- counting two things apparently identical in every way, that as far as we can tell are copies of each other, as instances of some thing-type is still a leap. And it's that leap that is the basis for whatever else we do.
So I'm still stuck at the move from utterance to utterance-type, belief to belief-type, thing to thing-type.