Florin Cojocariu

On the view I develop, grounding is a phenomenological achievement of organisms — structured, concern-laden encounter with the world, observable in basic minds that connect to their environment without representational content. Language coordinates with this connection; it does not create nor implement it. I develop these views within the Pattern Recognition Unity (PRU) framework, distinguishing experiential Pattern-Constellations from Linguistic Pattern-Constellations and exploring how they coordinate. Contra Mollo and Milliere’s (2025) view, I further argue that teleosemantics — a theory that earned its credibility on pre-verbal organisms selected against grounded functions like feeding and fleeing — loses its explanatory force when applied to systems selected against linguistic adequacy. The selection history of LLMs borrows its world-connection from human raters rather than constituting its own; its causal relation to worldly features is inherited ancestry, not the proximal coupling teleosemantics describes in the natural world. It follows that no evaluation conducted within the linguistic register can distinguish genuine grounding from sufficiently powerful linguistic mastery.

This paper develops a meta-language for the relationship between experience and language, motivated by a structural diagnosis: natural language — including the precise technical language of philosophy — lacks stable constraints to hold apart the experiential and linguistic poles of cognition without collapsing one into the other. Every attempt to specify what "grounding" or "reference" is produces another linguistic description, which is exactly the kind of thing that needs grounding. When we ask what connects the word "cat" to cats, every answer — causal chains, tracking relations, informational content — is itself a linguistic description of a connection living entirely \emph{inside} language and \emph{outside} the world, not the connection itself. The debate cycles because it uses the medium whose limitation it is about. The meta-language is built within the Pattern-Recognition Unity (PRU) framework, which rests on two theoretical claims with empirical support. First, Pattern-Constellations: organisms form unified attractor states that integrate sensory input and motor-affective response as a single learned configuration, not as separate processes wired together. Second, Pattern-Recognition Unity: every recognition event is already a reaction — pattern and recognition are inseparable aspects of one event, not a detection step followed by a response step. These claims ground the notation in a substantive account of pre-linguistic cognition rather than in stipulation. The notational devices — marking the experiential pole without redescribing it in linguistic terms, marking the linguistic pole as linguistic, representing the coordination between them as a structured relation between different kinds of thing — are not arbitrary conventions but expressions of the architecture the framework describes. This is representational hygiene built on cognitive theory, not formalism for its own sake, though the framework is designed to support further formal development.

This paper develops a meta-language for the relationship between experience and language, motivated by a structural diagnosis: natural language — including the precise technical language of philosophy — lacks stable constraints to hold apart the experiential and linguistic poles of cognition without collapsing one into the other. Every attempt to specify what "grounding" or "reference" is produces another linguistic description, which is exactly the kind of thing that needs grounding. When we ask what connects the word "cat" to cats, every answer — causal chains, tracking relations, informational content — is itself a linguistic description of a connection living entirely \emph{inside} language and \emph{outside} the world, not the connection itself. The debate cycles because it uses the medium whose limitation it is about. The meta-language is built within the Pattern-Recognition Unity (PRU) framework, which rests on two theoretical claims with empirical support. First, Pattern-Constellations: organisms form unified attractor states that integrate sensory input and motor-affective response as a single learned configuration, not as separate processes wired together. Second, Pattern-Recognition Unity: every recognition event is already a reaction — pattern and recognition are inseparable aspects of one event, not a detection step followed by a response step. These claims ground the notation in a substantive account of pre-linguistic cognition rather than in stipulation. The notational devices — marking the experiential pole without redescribing it in linguistic terms, marking the linguistic pole as linguistic, representing the coordination between them as a structured relation between different kinds of thing — are not arbitrary conventions but expressions of the architecture the framework describes. This is representational hygiene built on cognitive theory, not formalism for its own sake, though the framework is designed to support further formal development.

This paper uses Gertrude Stein's phrase "a rose is a rose is a rose" as a lens to examine how reference operates in language. Rather than treating the phrase as mere repetition, syntactic bracketing reveals two distinct recursive structures that highlight different modes of linguistic reference. The analysis introduces a formal distinction between words functioning as concepts (x^c) and as objects (x^o), proposing reference as a binary operation R(x^c, y^o) internal to language rather than a mapping between words and external things. Two bracketing patterns emerge: left-nested recursion that resolves toward stabilized reference, and right-nested recursion that defers closure indefinitely. These structures suggest that language mediates between poles of immediate presence and conceptual elaboration. The resulting model challenges classical reference theories—Frege's sense/reference distinction, Kripke's rigid designation, and Tarski's model-theoretic approach—by demonstrating how meaning can arise from syntactic operations alone. The approach aligns with structuralist insights while offering a precise formal mechanism: reference becomes a rule-governed transformation among signs rather than a bridge between language and world. This internalist view suggests that apparent tautologies like "I am I" or "this is this" may be structurally productive rather than semantically empty, opening new avenues for understanding self-referential language and circular definitions when external context is minimal.