@Leontiskos

Again, Soames explains Frege this way:

"In general, when we want to refer to the thought expressed by a particular sentence, we use an expression such as "that S" or "the proposition that S". The use of the expression indicates that something is being said about a thought (proposition)." ibid pg. 22

- Early Frege did not distinguish sense and reference and may have spoken of propositions simply as thoughts, but this is not true for later Frege, which is what my IEP quote reflects. But none of this really helps you with your vague term "proposition," for "thought" is equally vague (as the quote from the OP shows, where Frege literally says that propositions need not be thought).

You seem bound determined to be vague and ambiguous in your claims. I'll leave you to it.

I'll leave you to it. — Leontiskos

Uh, ok. Thanks.

@Pierre-Normand
Hi! I was hoping to get some clarification from a professional. Did Frege think propositions were thoughts? Abstract objects, but basically thoughts?

Hi! I was hoping to get some clarification from a professional. Did Frege think propositions were thoughts? Abstract objects, but basically thoughts? — frank

Now you are getting into metaphysics, I was told that this shan't be done for this discussion.

Now you are getting into metaphysics, I was told that this shan't be done for this discussion. — schopenhauer1

I don't think abstract object implies any ontological commitment. It just means it's not a physical thing, but it's not a mental object like if you're picturing a flower in your mind. An abstract object is something I could be wrong about, for example if I say that 2 squared is 5.

I don't think abstract object implies any ontological commitment. It just means it's not a physical thing, but it's not a mental object like if you're picturing a flower in your mind. An abstract object is something I could be wrong about, for example if I say that 2 squared is 5. — frank

Why would that not be ontological? You are literally discussing someone's understanding of what an object is. Pondering the nature of objects/entities seems to fit that pretty well.

Have you read the OP?

Frege says, “A proposition may be thought, and again it may be true; let us never confuse the two things.” (Foundations of Arithmetic)
— J

Here is IEP:

For this and other reasons, Frege concluded that the reference of an entire proposition is its truth-value, either the True or the False. The sense of a complete proposition is what it is we understand when we understand a proposition, which Frege calls “a thought” (Gedanke). Just as the sense of a name of an object determines how that object is presented, the sense of a proposition determines a method of determination for a truth-value.
— Frege | IEP — Leontiskos

I don't think those two quotes speak against the claim that Frege is equating propositions with thoughts in the specific sense in which he understands the latter. The quote in Foundations of Arithmetic seems meant to distinguish the thinking of the thought from its being true. This is quite obvious if we consider the thought (or proposition) that 1 = 2.

The IEP quotation seems to use "proposition" rather in the way one would use "sentence", or well formed formula. So, the distinction that is implicit here parallels the distinction between a singular sense (i.e. the sense of a singular term) and a nominal expression in a sentence. If we think of the proposition as what is being expressed by the sentence — i.e. what it is that people grasp when they understand the sentence, what its truth conditions are — then, for Frege, it's a thought: the sense (Sinn) of the sentence.

Thank you!

Why would that not be ontological? You are literally discussing someone's understanding of what an object is. — schopenhauer1

I get that. It's just that when I've tried to pin down the ontological significance of "abstract object", the answer seems to be that this question doesn't need to be answered. We're more in the realm of just identifying what we can't do without.

Hi! I was hoping to get some clarification from a professional. Did Frege think propositions were thoughts? Abstract objects, but basically thoughts? — frank

I began writing my earlier response before I read your request to me. Let me just specify that I hardly am a professional. I did some graduate studies in analytic philosophy but I don't have a single official publication to my credit. Furthermore, a few posters in this thread would seem to have a better grasp on Kimhi and on Frege than I do, and that includes @Leontiskos even though we may currently have a disagreement.

:up:

I don't think those two quotes speak against the claim that Frege is equating propositions with thoughts in the specific sense in which he understands the latter. — Pierre-Normand

If "thought" is understood in a specialized sense then, sure, if you like. Again, my point is that these invisibly specialized senses of "proposition" and "thought" are not getting us anywhere.

Frank is just trying to uphold his idea that in "real life" we don't talk about unasserted declarative sentences, without actually going through the work of defending it.

I get that. It's just that when I've tried to pin down the ontological significance of "abstract object", the answer seems to be that this question doesn't need to be answered. We're more in the realm of just identifying what we can't do without. — frank

Well that's what has become the discussion oddly. We must talk about how the logic is presented but then nothing that would ground the logic itself (thought, psychology, epistemology, metaphysics, science)? Odd. This has been a theme it seems of a lot of language philosophy. For example the Tractatus asserts atomic facts, states of affairs, and objects, but grounds them in nothing. So the whole theory of correspondence of propositions to states of affairs in the world is taken on simply taking the original assertions at face value. There are no proofs for it. There is no evidence for it.

If, for example, propositional thinking is said to be part of the human animal, that needs to be explained. I was under the impression that philosophy involves a certain amount of rigorous explanatory power, not simply various assertions.

Frank is just trying to uphold his idea that in "real life" we don't talk about unasserted declarative sentences, without actually going through the work of defending it. — Leontiskos

I talked about that a lot. I don't think you saw it.

If "thought" is understood in a specialized sense then, sure. Again, my point is that these invisibly specialized senses of "proposition" and "thought" are not getting us anywhere. — Leontiskos

I would have thought that those specialized senses of "thought" and "proposition" struck at the core of Frege's thinking about logic in general, and his anti-psychologism in particular. Frege famously conceived logic as "the laws of thought". And he understood those laws in a normative (i.e. what it is that people should think and infer) rather than an empirical way (i.e. what it is that people are naturally inclined to think and infer). We might say that Frege thereby secures the essential connection between logic as a formal study of the syntactic rules that govern the manipulation of meaningless symbols in a formal language and the rational principles that govern the activity of whoever grasps the meanings of those symbols.

- :up:

---

You appear to be agreeing that we can't have unasserted propositions in real life, even if the assertion is only hypothetical or potential. — frank

The answer here is, 'No, that's not what Russell was saying at all, whether we take your "proposition" in Russell's sense or Frege's sense.'

This thread has taken plenty of care in avoiding ambiguous language like "proposition," and this is because the language of the thread is in no way controlled by Frege's usage. It is controlled by Kimhi's usage, the publicly available sources, and common usage.

If, for example, propositional thinking is said to be part of the human animal, that needs to be explained. — schopenhauer1

I think there are two approaches to the OP, and maybe in philosophy in general:

1. You can start with ontological biases and let those views form the limits of investigation.
2. You can start by trying to map out where you are, and let the borders remain foggy.

Both approaches have virtues. What's at stake with the concept of propositions is the communing part of communication. It's all about agreement. If we ever really do arrive at agreement, then we would say we're thinking the same thoughts. That's what a proposition is supposed to be: that thing we can agree or disagree on.

If you start with a strongly materialistic bias, you're likely to lean toward behaviorism, which says that we never really agree on anything. All agreement is really just a certain kind of behavior. If we follow this approach, we're never even going to arrive at the OP, because it's about thinking.

We might say that Frege thereby secures the essential connection between logic as a formal study of the syntactic rules that govern the manipulation of meaningless symbols in a formal language and the rational principles that govern the activity of whoever grasps the meanings of those symbols. — Pierre-Normand

Where, historically, would you say that essential connection gets dissolved or weakened?

I don't think those two quotes speak against the claim that Frege is equating propositions with thoughts in the specific sense in which he understands the latter. The quote in Foundations of Arithmetic seems meant to distinguish the thinking of the thought from its being true. — Pierre-Normand

Well, I think this avoids the force of my point a bit. Frege said, "A proposition may be thought..." He did not say, "A thought may be thought..." There are apparently good reasons why he didn't use the latter formulation, and they point a difference between a proposition and a thought. Specialized senses can "solve" this, I suppose, at least up to a point. (Nor am I convinced that IEP is using 'proposition' in a non-Fregian way, but I don't want to get bogged down in this.)

Well, I think this avoids the force of my point a bit. Frege said, "A proposition may be thought..." He did not say, "A thought may be thought..." There are apparently good reasons why he didn't use the latter formulation, and they point a difference between a proposition and a thought. Specialized senses can "solve" this, I suppose, at least up to a point. (Nor am I convinced that IEP is using 'proposition' in a non-Fregian way, but I don't want to get bogged down in this.) — Leontiskos

In the original German text, Frege wrote: "Man nehme nicht die Beschreibung, wie eine Vorstellung entsteht, für eine Definition und nicht die Angabe der seelischen und leiblichen Bedingungen dafür, dass uns ein Satz zum Bewusstsein kommt, für einen Beweis und verwechsele das Gedachtwerden eines Satzes nicht mit seiner Wahrheit !"

In Austin's translation, "may be thought" corresponds to "das Gedachtwerden", which literally means "the being thought" or "the being conceived." Frege is referring to the act of a proposition or thought being grasped or entertained by someone. "Satz" can be either used to refer to the syntactical representation of a thought (or of a proposition) or to its content (what is being thought, or what proposition is entertained).

If we attend to this distinction between the two senses of "Satz" Frege elsewhere insists on, and translate one as "declarative sentence" and the other one as "the thought", then, on the correct reading of the German text, I surmise, a literal translation might indeed be "A thought may be thought, and again it may be true; let us never confuse the two things." and not "A declarative sentence may be thought...". The latter, if it made sense at all, would need to be read as the elliptical expression of "(The content of thought expressed by) a declarative sentence may be thought," maybe.

In other words, Frege's text is making a distinction between the Gedanke (thought or proposition) being entertained and its being true, rather than focusing on the sentence (Satz) itself as the object of thought. (And then, the thought being judged to be true, or its being asserted, is yet another thing, closer to the main focus of this thread, of course.)

This intended reading, I think, preserves the philosophical distinction Frege is drawing in this passage between the mental act of thinking (grasping the thought) and the truth of the thought itself. Translating "Satz" in this context as "declarative sentence" would blur that important distinction, since Frege is interested in the thought content (Gedanke) rather than the linguistic expression of it (Satz). And, unfortunately, the English word "proposition" carries the same ambiguity that the the German word "Satz" does.

On edit: Regarding your earlier question, one clear target Frege had in mind was the psychologism he had ascribed to Husserl, which threatened to eliminate the normative character of logic and another possible target might be formalism in the philosophy of mathematics that turns logical norms into arbitrary syntactic conventions.

On edit again: "The locus classicus of game formalism is not a defence of the position by a convinced advocate but an attempted demolition job by a great philosopher, Gottlob Frege, on the work of real mathematicians, including H. E. Heine and Johannes Thomae, (Frege (1903) Grundgesetze Der Arithmetik, Volume II)." This is the first sentence in the introduction of the SEP article on formalism in the philosophy of mathematics.

Would you say an AI can assert a proposition? Or are we just treating it as if it can?

Coincidentally there is an article in a recent New Scientist lamenting the use of the word "force" in physics . "...language affects our wordless imaginations so deeply that it can’t fail to influence the way we think."

The notion of force being used in this thread is nothing to do with changing the motion of objects. I suspect this is misleading us.

Illocutionary "force" is not something that inheres in sentences, but what people do with sentences. Denoting is not something that inheres in names, but what people do with names. This performative aspect of language is hidden by talk of sentences having force.

(I) probably never should have wandered into speculations about different kinds or modes of “force,” even though Kimhi himself often seems to do that. — J

Well, it seems that when the notion of "force" is clarified, it doesn't do what Kimhi wants. He's reliant on ambiguity. But further, he seems not to consider the developments of logic and metalogic since Frege - and they are profound.

"Does a strong formalism such as Frege's invalidate whatever can be said or thought about p in ordinary language?" — J

There's an old adage concerning someone complaining about geneticists studying fruit fly when elephants were so much more important. But you can get more fruit fly into a laboratory, and they breed faster, so it makes sense to work with them. What is learned can then be appleid to elephants, if needed. Fruit fly simplify the process so that certain aspects can be examined in detail. Formal logic does much the same thing, but with language. It models certain aspects in a way that can be examined productively. If Kimhi wants to understand elephants, he could do worse than to start with fruit fly.

It is also a mistake to think that what Frege has to say is anything like the present state of play of formal logic. His discussion of truth is a case in point, his argument having been superseded by several generations of subsequent work.

And the discussion in this thread of modus ponens was just plain muddled.

It's hard to be sure, but there seems to be a profound misapprehension concerning what logic is, underpinning the Kimhi's work and much of the writing on this thread. I'm not inspired to go down that path.

Would you say an AI can assert a proposition? Or are we just treating it as if it can? — frank

While this issue isn't entirely off topic, and hence would be worth discussing, it also is liable to send us off into tangential directions regarding the specific nature of current LLM-based AI agents, the issue of their autonomy (or lack thereof), their lack of embodiment (and the issue of the empirical grounding of their symbols), how they are being individuated (as machines, algorithms or virtual agents in specific conversation instances), to what extend their assertions have deontic statuses (in the form of commitments and entitlements to claims), and so on. Hence, that might be a topic for another thread even though links could be drawn to the topic being discussed here. (I've already explored some of those issues, and connected them to some ideas of Robert Brandom and Sebastian Rödl, in my conversations with GPT-4 and with Claude, and reported on them in my two AI threads).

it also is liable to send us off into tangential directions regarding the specific nature of current LLM-based AI agents, the issue of their autonomy (or lack thereof), their lack of embodiment (and the issue of the empirical grounding of their symbols), how they are being individuated (as machines, algorithms or virtual agents in specific conversation instances), to what extend their assertions have deontic statuses (in the form of commitments and entitlements to claims), and so on. — Pierre-Normand

I'm not seeing that as tangential. I think it would highlight, especially for the naive, like me, the difference between a formal language where apparently there are propositions that aren't thought of as asserted in any way, and ordinary language, where the listener always thinks of what's being asserted in the light of who asserts it, or in what setting it's asserted. But I'll go with your judgment.

And since @frank brought up the topic of non-rational-animal issuers of "assertions", Kimhi writes on page 137 of Thinking and Being:

"But now one can ask: in virtue of what is the forceless combination Pa associated with the truth-making relation that a falls under the extension of P, and thus with the claim Pa, rather than with the truth-making relation that a does not fall under P (or falls under the extension of ~P), and thus with the opposite claim ~Pa? This question cannot be answered, since Pa does not display an assertion, and therefore there is nothing that associates it with the positive rather than the negative judgment. The association of the proposition Pa, conceived as forceless, with one of these conditions as its truth-making relation, understood in terms of the object the True rather than the other, smuggles assertoric force back into the predicative composition Pa—force which Frege denies is there."

I need to give much more thought to this before I can comment intelligently but, for now, I just want to point out that the great gangsta philosopher Ali G already had raised this issue decades ago in a conversation with then INS Commissioner James Ziglar.

"Ain't the problem though that 99% of dogs can't speak English so how does they let you know who is carrying a bomb ..."

"How does you know they ain't pointing to say this one definitely ain't got a bomb in it."

I'm not seeing that as tangential. I think it would highlight, especially for the naive, like me, the difference between a formal language where apparently there are propositions that aren't thought of as asserted in any way, and ordinary language, where the listener always thinks of what's being asserted in the light of who asserts it, or in what setting it's asserted. — frank

Maybe I misunderstand you but I don't see the distinction between a thought being merely entertained (or grasped) and it being asserted to line up with the distinction between (merely) formal and natural languages. Furthermore, I view conversational AI agents as users (to the little extent that they are) of natural language even though the underlying algorithm processes "symbols" blindly and manipulates them without regard to their semantic content. Ascriptions of semantic content to the responses of LLM-based AI agents, and interpreting them as having assertoric force, is a matter of taking a high-level intentional stance on their linguistic behavior with a view of rationalizing it.

Maybe I misunderstand you but I don't see the distinction between a thought being merely entertained (or grasped) and it being asserted to line up with the distinction between (merely) formal and natural languages. — Pierre-Normand

I think that when you entertain a thought, you're imagining it as an assertion, even if there was no such event. For instance, if you contemplate "the cat is on the mat", in terms of thought, all you have is a sentence that could mean all sorts of things. It's not truth-apt. To the extent that meaning is truth conditions, it's meaningless. Am I wrong?

Ascriptions of semantic content to the responses of LLM-based AI agents, and interpreting them as having assertoric force, is a matter of taking a high-level intentional stance on their linguistic behavior with a view of rationalizing it. — Pierre-Normand

And I think that's what we do all the time when we communicate. We rationalize. I don't think you can really see any meaning in the output of an AI unless you take it as having assertoric force. It's a reflexive part of communication.

I think that when you entertain a thought, you're imagining it as an assertion, even if there was no such event. For instance, if you contemplate "the cat is on the mat", in terms of thought, all you have is a sentence that could mean all sorts of things. It's not truth-apt. To the extent that meaning is truth conditions, it's meaningless. Am I wrong? — frank

No, I agree with you. I also tend to view the pragmatic role of assertion to be paradigmatic and central in understanding what propositions (or thoughts) are. But I haven't yet engaged closely enough with Kimhi's text (and with most contributions in this thread) to properly assess his challenge to Frege.

And I think that's what we do all the time when we communicate. We rationalize. I don't think you can really see any meaning in the output of an AI unless you take it as having assertoric force. It's a reflexive part of communication.

I agree. It relates to Davidson's constitutive ideal of rationality (which grounds language interpretation) and it also relates closely to Frege's context principle. If thoughts were merely entertained but never asserted (or denied) then they would lack content since there would be no rational ground for interpreting them one way or another.

Cool, thanks!

First of all, thanks for your helpful response.

It's hard to be sure, but there seems to be a profound misapprehension concerning what logic is, underpinning the Kimhi's work and much of the writing on this thread. I'm not inspired to go down that path. — Banno

The common misunderstanding of logic is why I’m asking you to stay involved in this thread. This website is a type of public school for enthusiastic amateurs who need teachers. High aptitude for formal logic is rare, and therefore you can do much good by sharing some fruits of your gift.

Oftentimes it appears as if the nose-bleed section isn’t listening to the dugout, but they are.