Manolo Martínez (2025), Philosophy and the Mind Sciences, vol. 6 — DOI-10.33735/phimisci.2025.11321
Key takeaways
- A state counts as a representation when it shows adaptations for transmitting information for its own sake, i.e. for efficiently trading off rate, distortion and coder complexity — and that trade-off is what grounds the properties cognitive scientists attribute to representations.
Seeded from the abstract and a skim, before the reading in NOTE-165. What follows is what the work says about itself.
Martínez proposes a framework for theorizing about representation in cognitive science built on two theses. The Transmission thesis says representations are primarily signals whose function is transmitting information for its own sake, not as a by-product of some other role. The RDC Trade-Off thesis says those signals aim at an efficient balance among three budgets: rate (transmission and storage cost), distortion (fidelity), and the computational complexity of the coders. Evidence that something is a representation is therefore evidence of adaptations for efficient transmission, and the familiar marks of paradigmatic representations are argued to be products of such rate–distortion–complexity adaptations.
Standing in the record
Filed from the 2026-09-26 sweep of the papers-feed tracker for philosophy: 20 seconds of active reading over 1 session. It was filed Deferred, unread. NOTE-165 is the close reading of 2026-09-26, and it placed the work: Active — worth reading as the clearest short statement of a rate–distortion–complexity (teleo)semantics of representation, provided it is read as the position paper it says it is (§1: “comparatively light” on objections); the defence is in Martínez 2019b/2024 and in work the paper says is to come.
When seeded, before the reading in NOTE-165, the entry gave a priority for a deeper reading: medium — short dwell (20 s), but it is a clean, formal account of representation that bridges information theory and philosophy of mind; the skim captures the two theses and the surface idea, while the technical §§3–5 remain unread.
What the seeding asked a deeper reading to check (NOTE-165 is that reading):
- Offers an information-theoretic criterion for representation-hood that transfers naturally to learned representations in ML (bottlenecks, rate–distortion framings of representation learning); check whether Martínez discusses learned networks directly (Bengio et al. 2013 is cited).
- The “representational surface” is a testable framing: check §4–5 for what empirical signatures would count as adaptations for rate or complexity efficiency.
- A deeper reading should check how the Transmission thesis avoids counting every efficiently-coded signal (e.g. in engineered systems) as a representation, and how it handles misrepresentation.
Access when seeded: Journal landing page (abstract, keywords, published 2025-06-18) and the open-access HTML galley (full text, CC BY 4.0) fetched with curl; read §1, §6, §7 and the section heads. Crossref confirms date and volume.