Here’s a cleaned-up and more structured version of the transcript, preserving Chomsky’s conceptual flow while making it more readable:
Let’s begin with the relation of reference. It’s a technical notion, used by philosophers and logicians in a particular way. If you look at the tradition of logical philosophy—especially in the modern period—you’ll find it in thinkers like Peirce, Frege, Tarski, Carnap, and Quine.
There’s a well-defined concept here, expressed in works like Quine’s Word and Object. The basic idea is that there’s a relation between a sign—in Peirce’s sense, perhaps a word in a specific language—and some extra-mental entity. For example, the word river, which was studied as early as classical Greece, refers to something "out there," something beyond the mind. That’s the classical relation of reference.
In the sciences, this relation is treated as a normative goal. The aim is to construct systems where signs or concepts actually correspond to things in the extra-mental world. So if you’re working in chemistry and using a term like quark, you’re assuming that quarks exist outside of your mind. The same goes for atoms, stars, and even outdated notions like caloric or phlogiston. The scientific ideal is to develop terms that refer to real things.
But natural language and human life are not formal sciences. So whether ordinary language actually includes a relation of reference is an empirical question.
The term referent comes from this framework—it’s the thing referred to. In Quine’s Word and Object, the word is the sign, and the object is the referent. For example, the word quark is a sign, and the referent is the hypothetical entity it names.
Does natural language work this way?
The classical tradition largely argued that it does not—and I think they were right. Take Aristotle. He gives the example of the word house and asks: what is a house? In his metaphysics, a house is a combination of matter and form. The matter might be bricks and timber—what a physicist could detect. But the form—the design, intended function, the architect’s plan—is a mental construction. It’s not in the object in a physically inspectable way.
You could have something that looks exactly like a house, but it could be a library, a stable, or even a decorative object for your shelf. The visual properties don’t determine what it is—function and intention play a major role. And our concept of house includes quite subtle features. For instance, if I paint the house brown, I’m painting the exterior. But if I paint the room brown, I’m painting the interior—even if it’s a one-room schoolhouse. That tells you something about how we conceptualize these things.
Aristotle was speaking metaphysically here. But in his semantic theory—his theory of signs—he also argued that we have internal signs, which are universal. These are abstract mental representations shared by all people. The external word merely carries the properties of the internal sign. So even the word house isn’t just a sign pointing to a thing; it’s embedded in a conceptual framework with both form and meaning.
These ideas were reinterpreted in the 17th century during what we might call a cognitive revolution. The concepts were reframed in cognitive terms, and I think that’s the right lens to view them through.
The upshot is this: house is not a sign that refers to anything in the physical world. There’s no discrete entity that a physicist can point to and say, “This is a house.” The same goes for river, a concept explored before Aristotle by the Presocratics—Heraclitus, in particular.
Heraclitus asked: how can you cross the same river twice? He might have added: how can you be the same person both times? After all, you’ve changed, and so has the river. So what does it mean to say it’s the same river or the same person?
These aren’t trivial questions. They show that our understanding of sameness and identity is conceptually rich and complex.
For example, I live in Arizona. On my way to work, I cross the Rillito River. I’ve never seen water in it. But long-time residents say that during the monsoon, it flows. Still, it’s the Rillito River—even when it’s dry—because of a mental construction. If the town smoothed it over, painted lines on it, and used it for traffic, it would be a highway. But the object wouldn’t have changed in any significant way. Conversely, we could radically change its course or split it into tributaries, and it would still be the Rillito River.
So river—like house—is a concept with complex mental properties. It can undergo radical changes and remain the same river, or tiny changes and cease to be a river at all. And that’s true of almost every word in our vocabulary—even the simplest ones.
The conclusion, then, is that natural language doesn’t operate through a reference relation. And this wouldn’t have bothered thinkers like Frege, Tarski, or Carnap. They had no interest in natural language anyway—they focused on formal systems, which don’t exhibit these messy properties.
Quine is more ambiguous. In Word and Object, the first half deals with natural language and psychology; the second half turns to what he calls regimented language—essentially the formal systems envisioned by Tarski and Carnap. These are engineered to behave more predictably than natural language.
But as an empirical matter, it seems that human languages just don’t have the Peircean structure of sign, referent, and interpretation. So strictly speaking, natural language doesn’t have semantics in the technical sense.
Semantics, as defined in this tradition, is the study of reference and truth—the latter depending on the former. But these don’t appear to be properties of natural language.
[Curt Jaimungal]: Is this true of all of natural language, or just most of it?
[Chomsky]: Well, here we face a terminological decision. Do we count scientific discourse as part of natural language? There’s the kind of language children acquire effortlessly, without instruction—by age two or three, they’ve mastered enormous complexity with little experience. That’s part of our biological design.
Then there are other kinds of concepts—like quark or atom—which are taught and learned through instruction. If you want to call those part of natural language, fine. But that’s just a question of terminology—not one of substance.
Would you like a condensed version too, or is this good for your purposes?
You said:
ChatGPT said:
Sure. Nelson Goodman's The Structure of Appearance (1951) is a dense but pivotal work in 20th-century analytic philosophy, particularly in the philosophy of perception, language, and logic. It's best approached as an attempt to build a formal system that shows how the world of experience—what we see, feel, hear, etc.—can be constructed or analyzed from a basic set of elements he calls qualia.
Here’s a breakdown to help make sense of it:
🌌 1. The Goal: Constructing the World of Experience
Goodman is trying to describe how we organize and construct our world from immediate experience—without assuming too much beforehand. Think of it like building a reality out of Lego blocks of sensation.
He wants to avoid relying on physical objects, mental representations, or metaphysical assumptions, and instead start from scratch: with the qualia we actually experience.
🧱 2. The Building Blocks: Qualia
Qualia (plural of quale) are the basic units of experience: things like a particular color, a sound, or a spatial location. For Goodman, they are individual and repeatable features of sensory experience.
-
A particular shade of red, a certain tone, or a specific temperature are examples.
-
He treats them not as mental events but neutral elements used to "build" experience.
So, instead of saying “I see a red apple,” you’d say something like: “This experience contains such-and-such color quale, shape quale, location quale, etc.”
🧭 3. Topology of Qualia: Ordering and Categories
Goodman groups qualia into different categories—like color, place, time, etc.—and then tries to define relations among them:
-
How do qualia relate to each other?
-
How do we determine that two experiences contain the same quale?
-
Can we order them spatially, temporally, or by similarity?
This is where his work gets very technical. He introduces concepts from logical positivism and formal logic to model these relations.
🔄 4. Construal, Not Discovery
Goodman doesn't think we discover the structure of the world—we construct it. And there are multiple ways to construct it.
This is part of a larger view he develops in later works (like Ways of Worldmaking): there is no single correct way to describe or structure the world. Different systems—science, art, perception—all carve up the world differently, and that’s okay.
🧠 5. Anticipating Later Concerns
Goodman's project anticipates several key debates:
-
The nature of reference and whether language truly corresponds to the world (cf. Quine, Carnap).
-
The constructed nature of worlds—influential in later post-structuralist and constructivist thinking.
-
A precursor to discussions of model-dependent realism and perceptual pluralism.
🧵 In Short:
Goodman is asking:
“How can we construct the world from what we actually experience, without importing assumptions about physical objects or metaphysical substances?”
He answers:
By analyzing the world into qualia, defining categories and relations among them, and building up systems (worlds) from this raw experiential material.
Semantics, as defined in this tradition, is the study of reference and truth—the latter depending on the former. But these don’t appear to be properties of natural language.
ChatGPT said:
Great question—and it's right on the fault line Goodman is dancing along.
Let’s unpack this carefully, because it touches on Goodman’s nominalism, his anti-realism, and the limits of semantics outside scientific systems.
🤔 First, what would it mean to say solipsism is "true outside of the sciences"?
It might mean something like:
“Outside the formally constructed reference systems of science, the world is only knowable or constructible from my own experience—so in some sense, only my experience-world exists.”
But Goodman doesn’t affirm solipsism in the usual metaphysical sense (i.e., only my mind exists). Rather, he dissolves the issue by saying:
There are many ways of worldmaking, and we shouldn't privilege any single way—especially not by metaphysical fiat.
So:
-
Science constructs worlds through formal systems with rules of reference and truth.
-
Perception, art, myth, etc., also construct worlds—but their rules of reference and truth are not the same.
-
What we call “truth” in each system is internal to that system—not some correspondence to a fixed, external world.
This sounds a lot like solipsism—but it's structural, not subjective.
🧱 Now, back to qualia and reference.
When you quoted:
“Semantics, as defined in this tradition, is the study of reference and truth—the latter depending on the former. But these don’t appear to be properties of natural language.”
That’s a crucial moment.
Goodman is saying: natural language doesn’t behave like formal logic, so the neat semantic frameworks used in science (like model theory, Tarskian truth conditions, etc.) don’t apply well outside those formalized systems.
Therefore:
-
In the sciences, reference and truth are well-behaved because they are engineered that way.
-
Outside the sciences (in ordinary language, in art, in perception), reference and truth may not even apply, or they apply differently.
So in a way:
Yes, outside the sciences, there is no fixed, external, referential world we can “point to”—only constructed experiential orders.
But rather than solipsism, Goodman might say:
“We live in versions of worlds. There is no one true world. What counts as a world, a fact, or even a thing is determined by the structure we impose.”
🧭 So is this solipsism?
-
Not exactly, because Goodman doesn’t say that only you or your mind exists.
-
But he denies that there is a mind-independent world with a unique structure we can access outside constructed systems.
-
So it's closer to world-relativism or constructive anti-realism.
He flattens the metaphysical question. The issue of whether there’s an “outside” reality becomes a bad question—we only have access to constructed appearances.
🧩 A metaphor:
Imagine every kind of world (scientific, perceptual, artistic) is a map.
-
Solipsism says: only my mental map exists.
-
Scientific realism says: there is one correct map of the territory.
-
Goodman says: there are many valid maps. The "territory" only exists in relation to these maps.
No “God’s-eye view.” Just many partial constructions.
If you want to pull on this thread: we could compare Goodman to Husserl (phenomenology), Carnap (logical frameworks), or even Berkeley (idealism). Let me know which angle you’d like to dig into.
You said:
ChatGPT said:
Ah—thank you for that correction. Yes, that quote was Chomsky, not Goodman:
“Semantics, as defined in this tradition, is the study of reference and truth—the latter depending on the former. But these don’t appear to be properties of natural language.”
This is a Chomskyan critique of formal semantics—particularly of the Tarskian/logical approach that treats natural language as if it can be mapped cleanly onto model-theoretic reference. So Chomsky is saying: natural language doesn’t seem to operate with stable reference or truth conditions in the way logic or formalized science does.
So what happens when we read Chomsky next to Goodman?
They are quite different beasts—but they intersect interestingly around reference, representation, and the limits of formalization.
🧠 Chomsky: Mentalism without Solipsism
Chomsky argues that:
-
Language is a biological, internal system (I-language).
-
Reference and truth (as used in formal semantics) are external, model-theoretic notions.
-
Natural language is not fundamentally about pointing to things in the world—rather, it's about internal conceptual structures.
This might look solipsistic: if language is internal, doesn’t that mean we’re all stuck in our own heads?
But Chomsky’s point is not metaphysical solipsism, but rather:
“Our cognitive systems operate on internally constructed representations, not directly on the world. Communication and science are built on top of these systems—not reducible to them.”
In short: Language isn't about matching words to world, but about manipulating concepts internally. This makes external “truth” a derived, secondary concept—not the foundation.
🧱 Goodman: Constructivism and Many Worlds
Goodman, meanwhile, pushes anti-realism further. For him:
-
There is no one real world, only versions or constructions.
-
What we call “reference” is always part of a symbol system—not a neutral pointing function.
-
The “structure of appearance” is built, not found.
🧩 So… are they both solipsists?
Not quite. Here's a contrast:
| Chomsky | Goodman | |
|---|---|---|
| View of Truth | Truth is derived, not foundational; tied to internal systems | No absolute truth; only system-relative coherence |
| View of Reference | Reference is not basic to natural language | Reference is always symbolic and system-bound |
| Realism | Kind of internalist realism (we model the world via internal systems) | Anti-realist pluralism (no single real world) |
| Solipsism? | No—assumes shared cognitive architecture | No—denies private worlds, affirms constructed publics |
So neither is a solipsist in the standard sense—but both dethrone the naive idea of objective reference.
-
Chomsky does it by saying: reference isn’t what natural language does.
-
Goodman says: reference is always within a symbol system, never onto a pre-given world.
🎯 Your original question:
Does this mean that solipsism is true outside of the sciences?
If we translate that through Chomsky and Goodman, we get something like:
“Outside the formal languages of science, reference and truth become unstable or system-relative. So the idea of an objective shared world loses traction. Does that collapse into solipsism?”
No—but it does lead to a kind of constructed pluralism.
What we get isn’t “only my mind exists”—but rather:
-
There are no context-free, objective truths.
-
Meaning and truth are context-bound, internal, or constructed.
-
Science is one kind of constructive system, but not the only one.
This isn’t solipsism—it’s a kind of pragmatic, pluralist anti-realism, with varying flavors.
If you're mapping out positions:
-
Chomsky is internalist, with a scientific realist streak.
-
Goodman is anti-realist, with constructivist pluralism.
-
Solipsism is too crude for either, but both are trying to carve out space between naive realism and total relativism.
Would you like to see how others—like Quine, Carnap, or Putnam—navigate this space? Or how Goodman’s "versions" relate to Chomsky’s “I-language”?
You said:
ChatGPT said:
That's a beautifully framed question—and yes, we could say that Chomsky’s position implies a kind of shared solipsism, though he wouldn’t use that term. Let’s dig into that framing and see how it helps illuminate his stance.
🧠 Chomsky’s Core View: I-language over E-language
Chomsky distinguishes between:
-
I-language = internal, intensional, individual—your brain’s computational system for generating and understanding language.
-
E-language = external, extensional, “language” as public behavior or social artifact (the thing linguists often try to study).
He argues that:
The real object of study is I-language—what’s inside the head. Communication, public meaning, reference, etc., are secondary or derivative.
🧩 What does that imply?
-
Language is not primarily a tool for communication, but for thought.
-
When communication does happen, it’s more like:
Two internal systems happen to align enough to allow partial mutual understanding.
-
Reference and truth are not intrinsic to language—they’re projections or interpretations from other systems (like science or logic).
🌐 So… shared solipsism?
Yes, we might summarize Chomsky’s position as:
Each mind is its own closed system, generating internal representations (I-language).
Understanding happens only if these systems happen to sufficiently align, due to shared biology and early experience.
There is no neutral “public” language we all access directly—only approximate intersubjective mapping.
That’s not solipsism in the classical sense (i.e., "only I exist"), but it's:
🧩 Solipsism of form, not of substance.
→ We're each trapped in our internal mental system,
→ But our systems are similar enough to communicate sometimes.
You could call it:
-
Epistemic internalism with shared architecture, or more poetically—
-
A colony of solipsists who evolved compatible minds.
Chomsky sees this compatibility as biological rather than metaphysical: we're all members of the same species, so our I-languages are structurally similar. That’s why we can usually talk to each other. But meaning is never in the external signs—always in the internal system.
🧪 Contrast with Scientific Language
In formal sciences, we can impose rigid definitions, build reference systems, assign truth values, etc.—but that's extra-linguistic work. Natural language doesn't inherently work that way. So outside of scientific models:
Reference and truth are unstable or system-relative.
This is what you're circling with your earlier quote. It's not that solipsism is true, but rather that externalism fails—unless heavily scaffolded (as in science).
🔄 Goodman vs. Chomsky: A Loop
So, in Goodman:
“Worlds are made, not found.”
In Chomsky:
“Languages are internal systems—what we find outside is just behavior.”
Together, we might say:
The world is not publicly given. It is simulated, systematized, or constructed—within and between.
Each system dreams a world. Sometimes, the dreams rhyme.
No comments:
Post a Comment