A foundational gap in Wilfrid Sellars’ account

Analysis of ch. 10 of Chrucky’s dissertation by chatGPT:

This passage critically engages with a foundational tension in Wilfrid Sellars’ theory of perception and conceptual learning — namely, how it’s possible to acquire the concept of a physical object if such a concept is needed in order to perceive one in the first place. The evaluation exposes a gap between:

Empirical experience, especially as Sellars models it using theoretical entities like sensa, and
The acquisition of conceptual and linguistic capacities, particularly the language Sellars calls “Behaviorese.”

Let’s break down and evaluate the core insights and implications:

1. Sellars’ Presupposition of Concept Possession

The critique opens by noting that Sellars assumes the Ryleans already possess a language (“Behaviorese”), and hence the concepts required to talk about behavior and physical objects. But this move presupposes the very thing needing explanation: how these concepts could be acquired in the first place. This raises the fundamental philosophical problem:

How is it possible to acquire the concept of a physical object if that concept is necessary to perceive one?

This is a version of the Myth of the Given problem (which Sellars famously criticizes): it points out that perception itself seems to be conceptually structured, but then questions how concepts get started.

2. The Role of ‘Sensa’ and Natural Evolution

Sellars does try to explain the emergence of language/concepts via natural and social evolution, and he posits that pattern-governed behavior leads to concept acquisition. However, the critique asks:

What are these “perceptible elements”?
Are they raw sense data (sensa)?
Is the learner passively conditioned or actively constructing?

This raises serious issues about agency, innateness, and the nature of perceptual input. Sellars introduces sensa as non-conceptual elements which underlie perception, but the critique rightly notes that:

Sensa are not objects of sense cognition, but theoretical postulates.
There’s no clear story of how these become organized into the conceptual framework of physical objects without already having some conceptual apparatus in place.

3. The Gap Between Process Ontology and Conceptual Perception

The paper then highlights a deeper ontological problem: Sellars’ Scientific Image is built on a Heracleitean process ontology — a world of processes, not stable objects. So how does a perceiver carve out physical objects from this flux?

The introduction of Representational Systems (RS) is a useful move by the author: to bridge the gap, there must be some pre-linguistic representational ability (rs-concepts). Animals, for instance, can seemingly rs-see physical objects. That suggests:

The human concept of a physical object supervenes on a more primitive representational structure.

But this implies a hierarchy of concept formation:

rs-concepts: primitive, possibly non-linguistic representations
ur-concepts: pre-linguistic but structured by logical constructions
linguistic concepts: full-blown concepts used in language like Behaviorese

Thus, Sellars seems forced to postulate innate representational mechanisms that exist prior to language — a move that leans toward rationalist or nativist theories of mind (e.g., Chomsky, Fodor) and seems to contradict his empiricist behaviorist tendencies.

4. Evaluation of the Critique

This is a strong and sophisticated critique of Sellars on several fronts:

It presses the regress problem: how do we explain the first concept acquisition?
It identifies the hidden reliance on innateness in Sellars’ account (ur-concepts or rs-concepts).
It demands a more complete epistemology of perception — one that explains not just linguistic usage, but how experience gets structured in a way that allows for linguistic conceptualization at all.

The author’s identification of the need for rs-concepts as a precursor to linguistic ones is particularly insightful and anticipates current debates in cognitive science and developmental psychology (e.g., core cognition, object permanence in infants, animal cognition).

Final Thought

If Sellars is committed to rejecting the Myth of the Given, but also needs non-conceptual sensa and pre-linguistic conceptual structure to explain perception and language acquisition, then he is either:

In tension with his own commitments, or
Obliged to flesh out a much more robust account of innate representational structures and how they become linguistic.

This passage rightly identifies a foundational gap in Sellars’ account — and in doing so, raises important questions for anyone working at the interface of epistemology, philosophy of language, and cognitive science.

Would you like to explore how this critique compares with more contemporary thinkers on concept acquisition (e.g., Brandom, McDowell, or even enactivist theorists)?

yes

ChatGPT said:

Great — let’s dive into how contemporary thinkers have tried to respond to or extend Sellars’ project, particularly regarding concept acquisition and the nature of perception. We’ll focus on:

Robert Brandom: inferentialism and the social character of concepts
John McDowell: conceptual content of experience and the “space of reasons”
Enactivist theories: cognition as embodied interaction with the world

1. Robert Brandom: Inferentialism and the Social Game of Giving and Asking for Reasons

Brandom builds directly on Sellars and defends a view known as inferentialism:

To have a concept is to be able to use it in inferentially articulated practices (e.g., making claims, giving reasons, drawing consequences).
Language is a normative social practice. It’s not just about labeling experiences but playing a game where claims can be challenged, justified, or withdrawn.

How Brandom addresses the gap:

Brandom would agree that there’s no pre-conceptual “given” that can serve as a foundation.
Concepts arise within socially governed inferential roles — essentially, learning a language is learning a role in a norm-governed game.
But: this doesn’t explain how you get into the game in the first place. Brandom tends to presuppose the existence of participants already capable of minimal discursive participation — so the critique you raised about Sellars still bites here.

Tension:

Brandom’s account seems to require already-in-place discursive capabilities — it’s not clear how the leap from animal rs-concepts to human conceptual normativity happens.

2. John McDowell: Mind and World – Experience is Already Conceptual

McDowell takes a different tack: he embraces the idea that perception is already conceptually structured. In Mind and World, he argues:

There’s no non-conceptual input (“Given”) that justifies knowledge.
Instead, perception is already infused with conceptual capacities — when you see a red apple, you see it as a red apple because you already have those concepts.

How McDowell addresses the gap:

He accepts that animals may have perceptual sensitivities, but denies that our kind of perceptual experience (experience with justificatory power) is possible without conceptual capacities.
McDowell suggests a “minimal empiricism”: experience is where the world exerts rational constraint on our thinking, but only because our capacities have been shaped by education into the space of reasons.

Tension:

Critics argue McDowell doesn’t sufficiently explain how conceptual capacities develop in the first place — he idealizes the result (a rational agent) but doesn’t give a genetic account of how one gets there from infancy or animal experience.

3. Enactivist and Embodied Theories: Concept Acquisition Through Sensorimotor Engagement

Thinkers like Francisco Varela, Evan Thompson, and Alva Noë represent a break from the Sellarsian model:

Perception is not about representing a world of objects, but about skilled interaction with a dynamic environment.
Concepts are not imposed on perception, but emerge from embodied activity in the world.

How enactivists address the gap:

They argue the split between sensa and concepts is artificial. Organisms are always enacting a meaningful world through their actions.
Infants and animals learn through embodied engagement — not by passive conditioning, but through dynamic coupling with the environment.
This model suggests rs-concepts aren’t internal “pictures” or constructions but structured habits of interaction.

Implications:

Enactivism avoids the problem of how to bootstrap concepts from raw data because it rejects the idea of raw data altogether.
It offers a possible answer to your critique: the capacity to see physical objects arises from action-oriented processes, which are developmentally elaborated into conceptual practices in humans.

Comparative Summary:

Thinker	Role of Concepts	How Are Concepts Acquired?	Relation to Sellars’ Gap
Sellars	Conceptual perception; no “given”	Pattern recognition + evolution, but under-specified	Gap exists: unclear how sensa → concepts
Brandom	Concepts = inferential roles	Entering social practices	Gap remains: presupposes discursive agency
McDowell	Experience is conceptual	Education into rational practices	Gap shifted: not explained but normalized
Enactivists	Concepts as embodied interaction	Emergent from active sensorimotor life	Gap dissolved: no need for internal construction from ‘sensa’

Would you like to develop this into a deeper comparative paper? Or zoom in on one of these views (like McDowell’s conceptualism vs. enactivism)?

Reason

A foundational gap in Wilfrid Sellars’ account —