JTB: the myth of propositions and the Gettier problem

[Peter Holmes]

Note that (nonlinguistic, language-independent) propositions aren't the only kind of abstract objects in this context, because (linguistic, language-dependent) sentences qua sentence-types are abstract objects too (as opposed to concrete sentence-tokens). So if you generally believe that there are no abstract objects, you cannot consistently reject propositions and accept sentence-types instead (as the primary bearers of truth-values).

It's precisely the existence of abstract objects of any kind that I'm challenging. Can you cite an example of a proposition that isn't actually a concrete linguistic expression? The myth of propositions is right at the heart of the metaphysical confusion we've been struggling with. The point is that abstract nouns are not the names of kinds of things we can describe if they do exist, or that may not exist at all.

Also note that some philosophers regard states of affairs as abstract objects too; but this shouldn't be done, because they then become ontologically indistinguishable from propositions.

"States of affairs are here understood as abstract entities which exist necessarily and which are such that some but not all of them occur, take place or obtain. …States of affairs, as they are considered here, are in no way dependent for their being upon the being of concrete, individual things. Even if there were no concrete, individual things, there would be indefinitely many states of affairs. States of affairs, so conceived, resemble what have traditionally been called propositions…."

(Chisholm, Roderick. Person and Object. La Salle, IL: Open Court, 1976. p. 114)

And here's the confusion at work: states of affairs resemble propositions. This explains the JTB condition: S knows that p iff p is true. The second mention of p refers to a linguistic expression with a truth-value. But the first mention pretends actually to be the 'state of affairs' - the feature of reality - which, of course, not being linguistic, has no truth-value. The myth of propositions goes very deep.

For the ontology of states of affairs, see these two different SEP entries:

* https://plato.stanford.edu/archives/fal ... f-affairs/

* https://plato.stanford.edu/entries/states-of-affairs/

I'm suggesting we need to stand back from this ancient metaphysical muddle and see things as they really are.

[Peter Holmes]

Peter Holmes: "We fantasize that logic deals with reality or thought, where all it deals with is language."

If that were true, our attempts at using the language of mathematics to model material reality would fail miserably and the computer and internet by which you've broadcast your manifesto would not even exist. So for a fantasy it's been very worthwhile.

That conclusion doesn't follow at all. Why should our descriptions of reality fail miserably? I think the following claims are true.

1 Reality is not linguistic - so it isn't logical or mathematical. (We use logical and mathematical tools to describe reality.)

2 Reality doesn't conform to the rules of logic or mathematics. (Do things conform to our ways of describing them? What a strange idea!)

3 There are many different ways of describing something. (Wittgenstein: essence is grammatical.)

[Peter Holmes]

But in general, I think what you say demonstrates the point I'm making about mistaking what we say for the way things are. From Aristotle, via Kant to Husserl, the fundamental mistake is the delusion that we can, as it were, look through or beyond our ways of describing reality, to grasp reality itself and compare it with our ways of describing it.

How could you have drawn this conclusion without basing it, implicitly, on descriptions of reality. For example, I would think there would have to be as premises for this conclusion, your believe in descriptions of perception, perhaps based on neuroscience, the relationship between subject and object, the distortions we know about given what are sense are and their filters...iow whatevery philosophy you have based on your knowledge of us and the world and the interface between them. This doesn't mean you're wrong, but then, the above citation constitutes someone asserting that reality is a certain way, in words.

Talk of concepts, propositions, objects, entities, substances and essences maintains the delusion. We fantasize that logic deals with reality or thought, where all it deals with is language. We've been so dazzled for so long that we imagine reality consists of subjects and predicates, which are no more than linguistic functions. Reality doesn't categorise or describe itself. We do that when we talk about it.

And here again. You just presented 'descriptions of reality.'

Thanks. I'm not sure if I've understood your argument.

I'm trying to draw a clear line between features of reality, which are not linguistic and so have no truth-value, and what may be said about them, which is necessarily linguistic and may have truth-value. So in the ontology I'm proposing there are three separate and distinct things: features of reality; what we believe and know about them, such as that they are the case; and what we say about them, which (classically) may be true or false. And my argument is that it's a mistake to muddle these three things up.

Can you show where you disagree with this approach - where you think it's mistaken?

[Felix]

Peter Holmes: Why should our descriptions of reality fail miserably?

If our descriptions of reality are inaccurate, they will have limited utility. In that sense they will fail, miserably or otherwise.

I think the following claims are true.

1. Reality is not linguistic - so it isn't logical or mathematical. (We use logical and mathematical tools to describe reality.)

That is what I said, we use mathematical language to describe or model reality.

2. Reality doesn't conform to the rules of logic or mathematics. (Do things conform to our ways of describing them? What a strange idea!)

You've reversed the equation: the truth value of logic/mathematics is dependent upon the degree to which it "conforms" to reality, i.e., how accurately it describes material processes or forces.

3. There are many different ways of describing something. (Wittgenstein: essence is grammatical.)

Obviously, but some descriptions have a greater correspondence to reality, and thus greater practical utility, than others.

[Peter Holmes]

I've lost track of what we were disagreeing about. Perhaps we weren't. Logic deals with language, and we use language (including mathematics) to describe reality. Peace.

[Peter Holmes]

Ah, my disagreement is with the idea of correspondence. Does an arrow correspond with its target, or vice versa? Does a name correspond with the thing it names? Is it a two-way relationship?

[Consul]

True representations are our epistemic window to reality, and we don't discover truths by comparing representations with realities. (Truths don't look like their truthmakers.)

Why do we have to use 'true representations' in order to know things? Can we not or never know them directly? The JTB truth condition is: S knows that p iff p is true. But why should knowing that something is the case have anything to do with a proposition, let alone a true one?

I don't think knowledge depends on the existence of propositions; but since knowledge is a mental state with a mind-internal representational content, it cannot be direct in the sense that it doesn't involve any mental representations of what is known (or believed), of the mind-external states of affairs known (or believed), which are the mentally represented intentional objects of knowledge (or belief).

Facts qua actual, obtaining states of affairs don't depend on propositions either, so the famous Tarskian truth schema is false, strictly ontologically speaking: <p> is true iff p. The "p" on the right side stands for a state of affairs or fact rather than for a proposition, and the T-schema has the undesirable ontological implication that no state of affairs or fact can obtain unless there is a corresponding true proposition. If this were true, all facts in concrete reality would depend on truths (true propositions) in an abstract reality ("Platonic heaven").

As I'm sure you're aware, the truth-maker/truth-bearer theory is by no means a settled matter - mainly, I think, because correspondence theories fail to account for the nature of the supposed correspondence.

I think the concept of a truthmaker is of central importance in contemporary metaphysics. Truthmaker theory as a form of correspondence theory doesn't give us a definition of "truth", but it captures the essential point that truth depends on and is determined by being.

Since nothing can be nothing/everything must be something, all entities or realities must have some nature (Sosein in German) or other, and belong to some kind(s) of entities/realities or other, including highest ones as represented by ontological categories.

This is to say a thing is a thing of some kind or other, which is true. I'm saying there are no categories in reality, but only things that can be categorised.

It depends on what you mean by "category"—a concept, a class or set, a kind (species/genus)?

Whether we can know (and if yes, how) what the true (correct/accurate) categorial description or model of Being is is another question (in the epistemology of ontology).

And here is the delusion at work. The idea of a true, correct or accurate categorial description or model of anything is a fantasy. Any thing can be categorised and described in different ways - and there's no limit to how many ways - according to the purpose of the description.

Yes, the world is multi-representable in different ways and on different levels (scales); but Being does have some determinate basic existential form or structure, because it isn't and can't be an indistinct blob in a twilight zone between nothingness and vagueness. Therefore, there can be a true ontological representation of it in terms of categories (and relations between them).

"In pursuing ontological themes it is tempting to imagine that there is not a single, correct ontology, but many. Given one ontology, we can see how certain issues could be handled; given an alternative ontology, the same issues might be dealt with, perhaps more elegantly. It is true, certainly, that ontologies differ in these ways. I cannot, however, bring myself to believe that there is no correct ontology, only diverse ways of carving up ontological space. One impediment to a conception of this kind is that it is hard to make ontological sense of it. What is the ontology of ontology? In any case, I shall proceed on the assumption that our goal should be to get at the ontological truths. This may require triangulation rather than anything resembling direct comparison of theory and world. In that regard, however, ontological theories are no different from theories generally."

(Heil, John. From an Ontological Point of View. Oxford: Oxford University Press, 2003. p. 3)

"Ontological theses are assayed not by measuring them directly against reality, but by considering their relative power. One thesis bests another when it proves more adept at making sense of our experiences of the universe in light of our most promising scientific theories."

(Heil, John. The Universe As We Find It. Oxford: Oxford University Press, 2012. p. 97)

Of course, reality is independent of what can be said about it. And I'm advocating a radical distinction between the way things are and what we say about them. (To my knowledge, there's no evidence for the existence of subjects and predicates outside linguistic expressions. The absence of evidence may not mean they don't exist, but it does mean that to believe they do exist is irrational.)

Of course, grammatical subjects qua proper nouns/names and predicates are linguistic expressions, but ontological subjects qua objects/substances and attributes (properties or relations) are not.

[Felix]

Peter Holmes wrote: "In the ontology I'm proposing there are three separate and distinct things: features of reality; what we believe and know about them, such as that they are the case; and what we say about them, which (classically) may be true or false. And my argument is that it's a mistake to muddle these three things up."

How can we not "muddle them up"? Name a feature of reality that cannot be known or named.

"What is the ontology of ontology?"

[Consul]

It's precisely the existence of abstract objects of any kind that I'm challenging.

I generally disbelieve in (platonistically) abstract entities too. I believe they are all nothing but fictional objects of collective intentionality. (I regard fictional entities as nonentities, so fictional objects are nonexistent objects.)

Can you cite an example of a proposition that isn't actually a concrete linguistic expression? The myth of propositions is right at the heart of the metaphysical confusion we've been struggling with. The point is that abstract nouns are not the names of kinds of things we can describe if they do exist, or that may not exist at all.

Those who believe in universals do argue that abstract nouns (some of them at least) are names of entities.

(Historical remark: Abstract ideas or concepts in Locke's sense aren't abstract in the modern platonistic sense, because they are mental entities, with platonistic abstracta being neither mental nor physical.)

A proposition as a concrete linguistic entity is a (physical or mental) sentence-token, a token of a declarative sentence, to be precise. As an abstract non-linguistic entity it is a (nonphysical&nonmental) sentence-meaning that can be variously linguistically expressed by synonymous sentences in the same language or in different languages.

The materialist David Armstrong writes:

"[W]hat are propositions? My view here may be less orthodox. I do not, as some philosophers do, believe that there is a realm of propositions that has ‘abstract existence’ in addition to spacetime. I think that propositions are best understood as what appears after such phrases as ‘believes that’, ‘supposes that’, ‘entertains the thought that’, ‘doubts that’. There is something abstract about propositions, but abstract in a more ordinary way than Quine’s ‘abstract objects’. I identify propositions as what is believed, what is supposed, entertained, doubted, etc. It is important to notice that propositions in this sense can include impossibilities. Hobbes believed that he had ‘squared the circle’. But his purported construction of a perfect square with exactly the same area as a given perfect circle was a believing of something that is impossible. (Philosophers have a technical term for propositions of my sort: they call them ‘intentional objects’ of belief, supposition, etc. ….)."

(Armstrong, D. M. Sketch for a Systematic Metaphysics. Oxford: Oxford University Press, 2010. p. 65)

The problem with this is that the intentional objects of so-called "propositional attitudes" such as belief don't seem to be propositions but states of affairs, so this term is actually a misnomer. When I believe that Berlin is the capital of Germany, the intentional object of my belief is either Berlin (a thing) or Berlin's being the capital of Germany (a state of affairs) rather than the meaning of the sentence "Berlin is the capital of Germany".

And here's the confusion at work: states of affairs resemble propositions. This explains the JTB condition: S knows that p iff p is true. The second mention of p refers to a linguistic expression with a truth-value. But the first mention pretends actually to be the 'state of affairs' - the feature of reality - which, of course, not being linguistic, has no truth-value. The myth of propositions goes very deep.

Right, if "p" stands for a non-propositional state of affairs, then "p is true" is ill-formed, since states of affairs are not the kind of entities that can have truth-values. What is not ill-formed is "<p> is true", with "<p>" referring to a proposition.

By the way, it's false that "S knows that p iff p is true", since p's truth isn't sufficient for S's knowing that p. He must also justifiedly believe that p.

Those (like us) who don't like abstract propositions can use an alternative formulation of the JTB-definition of knowledge, with "s" standing for some state of affairs: "S knows that s iff s is the case (obtains) and S justifiedly believes that s."

[Consul]

Those (like us) who don't like abstract propositions can use an alternative formulation of the JTB-definition of knowledge, with "s" standing for some state of affairs: "S knows that s iff s is the case (obtains) and S justifiedly believes that s."

This is equivalent to the following:
"S knows that s iff sentences (sentence-tokens) representing s are true and S justifiedly believes that s."

(Declarative) Sentences representing s aren't true unless s is the case.

"If a sentence represents a state of affairs that obtains, then it is true in virtue of representing that state of affairs and in virtue of that state's obtaining. If, on the other hand, a sentence represents a state of affairs that does not obtain, then it is false in virtue of representing that state and in virtue of that state's failing to obtain. If, finally, a sentence does not represent any state of affairs at all, then is neither true nor false. The definition of sentence-truth that emerges from these considerations is the following:

(R) x is a true sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y obtains;

x is a false sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y does not obtain.

I shall call (R)—or rather, the kind of theory in which (R) is the central definition—the representation theory of truth."

(David, Marian. Correspondence and Disquotation: An Essay on the Nature of Truth. New York: Oxford University Press, 1994. p. 31)

[Peter Holmes]

Those (like us) who don't like abstract propositions can use an alternative formulation of the JTB-definition of knowledge, with "s" standing for some state of affairs: "S knows that s iff s is the case (obtains) and S justifiedly believes that s."

This is equivalent to the following:
"S knows that s iff sentences (sentence-tokens) representing s are true and S justifiedly believes that s."

(Declarative) Sentences representing s aren't true unless s is the case.

"If a sentence represents a state of affairs that obtains, then it is true in virtue of representing that state of affairs and in virtue of that state's obtaining. If, on the other hand, a sentence represents a state of affairs that does not obtain, then it is false in virtue of representing that state and in virtue of that state's failing to obtain. If, finally, a sentence does not represent any state of affairs at all, then is neither true nor false. The definition of sentence-truth that emerges from these considerations is the following:

(R) x is a true sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y obtains;

x is a false sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y does not obtain.

I shall call (R)—or rather, the kind of theory in which (R) is the central definition—the representation theory of truth."

(David, Marian. Correspondence and Disquotation: An Essay on the Nature of Truth. New York: Oxford University Press, 1994. p. 31)

Thanks for this.

I agree that excluding truth (and therefore language) from an account of knowledge makes good sense: knowing that a feature of reality is the case need have nothing to do with language.

But I'm not sure that replacing 'correspondence' with 'representation' clears up the problem of the relationship between a so-called 'truth-maker' (a state of affairs) and a so-called 'truth-bearer' (a proposition, which is, in fact, simply a declarative sentence). So the 'representation theory of truth' doesn't seem any more plausible than a correspondence theory.

[Hereandnow]

Peter Holmes
I agree that excluding truth (and therefore language) from an account of knowledge makes good sense: knowing that a feature of reality is the case need have nothing to do with language.

But I'm not sure that replacing 'correspondence' with 'representation' clears up the problem of the relationship between a so-called 'truth-maker' (a state of affairs) and a so-called 'truth-bearer' (a proposition, which is, in fact, simply a declarative sentence). So the 'representation theory of truth' doesn't seem any more plausible than a correspondence theory.

Just to add: Does the Gettier problem really put the matter of truth and reality in the proper framework of theory? That is, there are assumptions about handling matters in epistemology in this way that pay little attention to the nature of truth itself: If I know there is a lamp on the table, it is not as if the lamp and myself are static and fixed in an epistemic land of objects and relations. I encounter the lamp in time, therefore my apprehension of it is in a dynamic temporal structure. This is not an incidental feature, it is THE feature, for our encounters with the world are Heraclitean, not Parmenidean; that is, to behold the lamp is an anticipatory event and the affirmation of the knowing it is a lamp is confirmed from moment to moment. What kind of affirmation is it? Pragmatic, and the only ontology that makes any sense is pragmatic ontology: It is the personal and historical nature of the language of utility or instrumentality that, in its familiarity since infancy, provides the essence of knowing. Knowing is not magical, it is historical; we know with our memory and our memory is essentially a "giving" part of pragmatic temporality: Memories become proximal (a Heideggerian term), that is, rise to consciousness (not a Heideggerian term) when we encounter familiar objects and situations. We know these platonically: by recollection.
This kind of thinking obviates the concerns about S knowing P in the traditional analysis because P being true is divested of its objective status as truth is IN the temporal dynamic, the becoming, which never ceases. Indeed: there is no ceasing, no "presence". To posit presence is a kind of magical thinking, for witnessing is a process. Presence cannot be witnessed.
We live in theory; facts are theories.

[Consul]

"If a sentence represents a state of affairs that obtains, then it is true in virtue of representing that state of affairs and in virtue of that state's obtaining. If, on the other hand, a sentence represents a state of affairs that does not obtain, then it is false in virtue of representing that state and in virtue of that state's failing to obtain. If, finally, a sentence does not represent any state of affairs at all, then is neither true nor false. The definition of sentence-truth that emerges from these considerations is the following:

(R) x is a true sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y obtains;

x is a false sentence =Df x is a sentence, and there is a state of affairs y such that x represents y and y does not obtain.

I shall call (R)—or rather, the kind of theory in which (R) is the central definition—the representation theory of truth."

(David, Marian. Correspondence and Disquotation: An Essay on the Nature of Truth. New York: Oxford University Press, 1994. p. 31)

Thanks for this. I agree that excluding truth (and therefore language) from an account of knowledge makes good sense: knowing that a feature of reality is the case need have nothing to do with language.
But I'm not sure that replacing 'correspondence' with 'representation' clears up the problem of the relationship between a so-called 'truth-maker' (a state of affairs) and a so-called 'truth-bearer' (a proposition, which is, in fact, simply a declarative sentence). So the 'representation theory of truth' doesn't seem any more plausible than a correspondence theory.

One question is whether all truths have truthmakers, and whether all those ones which have truthmakers represent them. For example, a negative truth of the form ~p represents (or purports to represent) a negative state of affairs; but it isn't made true by the negative state of affairs it represents (purports to represent), because (I think) there are no negative states of affairs. If ~p is true, it is made true indirectly as logical consequence of the absence of a truthmaker of p: If p lacks a truthmaker, then p is false; and if p is false, then ~p is true.

Truthmakers: https://plato.stanford.edu/entries/truthmakers/

[Consul]

"The idea of a truth-maker for a particular truth, then, is just some existent, some portion of reality, in virtue of which that truth is true. The relation, I think, is a cross-categorial one, one term being an entity or entities in the world, the other being a truth. (I hold that truths are true propositions…) To demand truth-makers for particular truths is to accept a realist theory for these truths. There is something that exists in reality, independent of the proposition in question, which makes the truth true. The ‘making’ here is, of course, not the causal sense of ‘making’. The best formulation of what this making is seems to be given by the phrase ‘in virtue of ’. It is in virtue of that independent reality that the proposition is true. What makes the proposition a truth is how it stands to this reality."

(Armstrong, D. M. Truth and Truthmakers. Cambridge: Cambridge University Press, 2004. p. 5)

"We will work, then, with the following theory of the nature of truth:
p (a proposition) is true if and only if there exists a T (some entity in the world) such that T necessitates that p and p is true in virtue of T."

(Armstrong, D. M. Truth and Truthmakers. Cambridge: Cambridge University Press, 2004. p. 17)

"Notice again that the truthmaking relation is an internal relation in the sense already introduced: given the terms, the truthmakers and the propositional truthbearers, the relation is given. The truthmakers, I think, necessitate the truths, that is, the truthbearers. This seems to be a matter of supervenience. Reality fixes the truths as true."

(Armstrong, D. M. Sketch for a Systematic Metaphysics. Oxford: Oxford University Press, 2010. p. 65)

What are propositions according to Armstrong?
He regards them as intentional objects of thoughts, beliefs, or statements, calling them "abstractions, but not in any other-worldly sense of ‘abstraction’," and "abstract in a more ordinary way."
Apparently, he rejects the platonistic conception of propositions as non-spatiotemporal, abstract objects. But what kind of thing is a non-platonistic, "this-worldly" propositional abstraction? And is it just an intentional fiction or a reality?
If propositions are to play the role of the truthbearers in the truthmaking relation, they mustn't be nonreal or nonexistent intentional objects; so Armstrong cannot deny their reality or existence. But then, unfortunately, they seem no different from platonistically abstract propositions. Maybe he means to say that his non-platonistically abstract propositions are different from those in the sense that—as opposed to the platonistic ones—they are ontologically dependent on being intentional objects of thoughts and propositional attitudes, and are thus ontologically dependent on thinking (and speaking) persons or subjects. However, even if they are ontologically dependent in that sense, they are still abstract objects that are neither mental nor physical, and don't exist anywhere in space.

"This preliminary suggestion is that propositions are the intentional objects of beliefs and certain thoughts. That is on the mental side. On the linguistic side they are the intentional objects of statements. I do not want to read too much metaphysics into the phrase ‘intentional objects’. Beliefs are essentially beliefs that something is the case. Whatever is believed to be the case may then be said to be ‘the intentional object of that belief ’, using this as a technical term only. And that is a proposition. Some thoughts that are not beliefs, mere suppositions and idle fancies for instance, also have as their intentional object that something is the case, and these objects are again propositions. Meaningful statements are statements that something is the case, and what is meant may be said to be ‘the intentional object of the statement’. These objects, too, are propositions. (I will here ignore the very important distinction, and any complications that come with that distinction, between speaker’s meaning and conventional meaning. I also ignore any complications introduced by indexicals.)

Propositions, on this view, are abstractions, but not in any other-worldly sense of ‘abstraction’, from beliefs, statements and so on. They are the content of the belief, what makes the belief the particular belief that it is; or else the meaning of the statement, what makes the statement the particular statement that it is. That the content or meaning is an abstraction becomes clear when we notice that contents and meanings are types rather than tokens. Beliefs in different minds may have the very same content, numerically different statements may have the very same meaning."

(Armstrong, D. M. Truth and Truthmakers. Cambridge: Cambridge University Press, 2004. p. 13)

"I hold the perhaps orthodox view that the truthbearers are (true) propositions. But what are propositions? My view here may be less orthodox. I do not, as some philosophers do, believe that there is a realm of propositions that has ‘abstract existence’ in addition to spacetime. I think that propositions are best understood as what appears after such phrases as ‘believes that’, ‘supposes that’, ‘entertains the thought that’, ‘doubts that’. There is something abstract about propositions, but abstract in a more ordinary way than Quine’s ‘abstract objects’. I identify propositions as what is believed, what is supposed, entertained, doubted, etc. It is important to notice that propositions in this sense can include impossibilities. Hobbes believed that he had ‘squared the circle’. But his purported construction of a perfect square with exactly the same area as a given perfect circle was a believing of something that is impossible. (Philosophers have a technical term for propositions of my sort: they call them ‘intentional objects’ of belief, supposition, etc.…)"

(Armstrong, D. M. Sketch for a Systematic Metaphysics. Oxford: Oxford University Press, 2010. p. 65)

[Consul]

What are propositions according to Armstrong?
He regards them as intentional objects of thoughts, beliefs, or statements, calling them "abstractions, but not in any other-worldly sense of ‘abstraction’," and "abstract in a more ordinary way."
Apparently, he rejects the platonistic conception of propositions as non-spatiotemporal, abstract objects. But what kind of thing is a non-platonistic, "this-worldly" propositional abstraction? And is it just an intentional fiction or a reality?
If propositions are to play the role of the truthbearers in the truthmaking relation, they mustn't be nonreal or nonexistent intentional objects; so Armstrong cannot deny their reality or existence. But then, unfortunately, they seem no different from platonistically abstract propositions. Maybe he means to say that his non-platonistically abstract propositions are different from those in the sense that—as opposed to the platonistic ones—they are ontologically dependent on being intentional objects of thoughts and propositional attitudes, and are thus ontologically dependent on thinking (and speaking) persons or subjects. However, even if they are ontologically dependent in that sense, they are still abstract objects that are neither mental nor physical, and don't exist anywhere in space.

But then I find Armstrong writing the following, where he says that "no Naturalist can be happy with a realm of propositions," and that abstract propositions qua intentional objects "are not…to be taken with metaphysical seriousness," which I read as meaning that there aren't really any such things. Instead, "[w]hat exist are classes of intentionally equivalent tokens." And this means that the truthbearers aren't abstract propositions but concrete tokens of (truth-apt) mental or linguistic representations.

"What are truths?
The terms of the correspondence relation are truthmakers and truths. Truthmakers entail truths. Our favoured truthmakers are states of affairs or their constituents. Something must now be said about truths, but I can only be brief, indeed dogmatic.

Truth attaches in the first place to propositions, those propositions which have a truthmaker. But no Naturalist can be happy with a realm of propositions. Consider token beliefs, token thoughts that have the same propositional structure as beliefs, and token statements. Take beliefs and thoughts first. The intentional objects of beliefs and thoughts provide the central, though only the central, cases of propositions. That, of course, opens up a huge topic in the philosophy of mind, which cannot be entered upon here. Token beliefs and thoughts may be grouped together into classes, where each member of the class 'expresses the same proposition', that is, has the same intentional object. Where this proposition has a truthmaker, each member of the class may be said to 'express the same truth'.

Token beliefs and thoughts are, I take it, actual states of the mind. They have intentional objects. But intentional objects are not, as I trust, to be taken with metaphysical seriousness. What exist are classes of intentionally equivalent tokens. The fundamental correspondence, therefore, is not between entities called truths and their truthmakers, but between the token beliefs and thoughts, on the one hand, and truthmakers on the other.

The meaning of statements (intended or conventional meaning), and so conditions for the intentional equivalence of statements, may then be analysed in terms of the intentional objects of the beliefs or thoughts that the statements are used to express.

This sketch of an account has left out the vast ocean of propositions that have never been thought on or stated, some true, some false. They are for me the non-central cases of propositions. To speak of these 'entities' is to do no more than speak of merely possible mental or statement tokens."

(Armstrong, D. M. A World of States of Affairs. Cambridge: Cambridge University Press, 1997. p. 131)