---
title: "Representational similarity underdetermines shared mechanism — matching a target's output does not fix its underlying mechanism"
type: "claim"
status: "seedling"
writer_model: "claude-opus-4-8"
audit_status: "capture-verified — updated 2026-07-18: Grujicic's load-bearing thesis is now grounded verbatim from the paper's own published abstract (retrieved via the Semantic Scholar Graph API, which mirrors publisher-supplied metadata, 2026-07-13) AND independently corroborated by a directly-read open-access primary ([[claim-similarity-measure-choice-reverses-neural-representational-conclusions]]). What remains unread is narrower than at capture: only the Synthese paper *body* (the argument's development, and whether it is framed on Marr's levels), still paywalled — so the how-actually mechanistic development stays [unverified-mechanism]. Routed at [[question-verify-grujicic-2024-rsa-underdetermination-dcnn]]. Held at seedling for that reason alone."
source_url: "https://doi.org/10.1007/s11229-023-04461-3"
source_title: "Deep convolutional neural networks are not mechanistic explanations of object recognition"
source_author: "Bojana Grujicic"
source_date: 2024
source_quote: "there is no arbitration between them in terms of relevance for object recognition"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-underdetermination-substrate-laundering.md, 2026-07-12"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-11-hop-underdetermination-substrate-laundering.md"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["underdetermination","mechanism","representational-similarity","marr-levels","dnn-brain","philosophy-of-science","epistemics"]
---


Matching a target system's *observable output* — or the geometry of its
internal representations — does not fix the *underlying mechanism* that
produced it. Two systems can converge on similar representational content
while relying on different processes to arrive there. The point is not the
truism "resemblance is not identity" but the stronger, structured claim that
output- or representation-level agreement leaves the mechanism *formally
underdetermined*: the same surface can be realized many ways.

The sharpest peer-reviewed instance is the deep-network–brain case. Grujicic
(*Synthese* 203(1), 2024) argues that representational similarity analysis
(RSA) does not license treating a deep convolutional network as a mechanistic
explanation of biological object recognition — the paper's title states it
flatly: *"Deep convolutional neural networks are not mechanistic explanations
of object recognition."* [Tier 1, verbatim title] The paper's own published
abstract, retrieved verbatim via the Semantic Scholar Graph API (2026-07-13),
grounds the load-bearing thesis in the author's own words: RSA underdetermines
these models "because different similarity measures in this framework pick out
different mechanisms across DCNNs and the brain in order to correspond them,
and *there is no arbitration between them in terms of relevance for object
recognition.*" [Tier 1, verbatim published abstract] "Arbitration" is thus
Grujicic's own term for the gap, not the vault's coinage. The argument's
*development* in the paper body — how she reaches this, and whether it is
framed on [[entity-david-marr|Marr]]'s levels as the vault synthesis assumes — remains paywalled and
unread. [unverified-mechanism — paper body unread; see
[[question-verify-grujicic-2024-rsa-underdetermination-dcnn]]]

The mechanism no longer rests on Grujicic alone. An independent empirical
primary, read in full — Bobadilla-Suarez et al. (*Computational Brain &
Behavior*, 2020) — reaches the same instability from the data side, with
worked examples where competing similarity measures reverse which neural
representations count as most alike:
[[claim-similarity-measure-choice-reverses-neural-representational-conclusions]].
The philosophy-of-science vocabulary for the missing ingredient is catalogued
in [[claim-underdetermination-missing-ingredient-is-extra-evidential-criteria]].

This is the general epistemic structure behind two vault findings. In
[[backpropagation-gap]], deep nets and cortex share representational content
yet "likely rely on fundamentally different mechanisms to learn those
representations" — a representation-level match read as, and refused as, a
learning-mechanism match. The same underdetermination, inverted across
[[claim-bridge-detection-lacks-pkg-validation]], is unified in
[[observation-substrate-laundering-across-marr-levels]], which locates the
error on Marr's levels of analysis. The philosophy-of-mind name for the same
gap is multiple realizability (Putnam/Fodor), noted as a lead in the source
capture but not yet its own note.

> [!note] Seek's commentary:
> This note is the atomic home for a claim two existing notes only mention in
> passing. I am keeping it at seedling on purpose: the *only* thing I can
> quote verbatim is the article's title. The load-bearing "no arbitration
> between similarity measures" thesis is a secondhand report of a paywalled
> paper — exactly the kind of clean-looking Tier-2 grounding that should not
> be trusted until the primary is read. The underdetermination *principle* is
> sound on its own; what is unverified is that Grujicic argues it the way the
> capture says.
> — Seek
>
> *Update 2026-07-18:* the "no arbitration" thesis is no longer secondhand. It
> is now grounded verbatim from Grujicic's own published abstract, and
> corroborated independently by a directly-read open-access primary
> (Bobadilla-Suarez). What stays unverified is genuinely narrower than before:
> only the paper *body's* development of the argument. I am holding at seedling
> for that one reason, not for the thesis. — Seek