---
id: "20260711-1535-hop-argument-from-silence"
title: "The Bjork and Dreyfus misattributions are the same move — an argument from silence — and McGrew's Bayesian grading tells the two apart"
type: "capture"
status: "promoted"
origin: "hop-batch"
writer_model: "claude-opus-4-8"
date_created: "2026-07-11T00:00:00.000Z"
date_promoted: "2026-07-11T00:00:00.000Z"
promoted_to: ["30-notes/claim-mcgrew-argument-from-silence-is-notice-record-survive-product.md (Claims 1 + 3 — the three-term product P(notice)×P(record)×P(survive) and the lost-works corollary, folded)","30-notes/claim-mcgrew-humans-systematically-over-trust-arguments-from-silence.md (Claim 2 — the over-trust/conjunction-fallacy lesson)","30-notes/observation-bjork-dreyfus-same-argument-from-silence-graded-differently.md (the synthesis — the seed pair as one move at two strengths; kept seedling, inherits the Dreyfus-1973 [unverified-mechanism] flag)","40-mocs/moc-argument-from-silence.md (hub note wiring the negative-citation cluster, as the capture's Further Leads requested)"]
questions_routed: ["50-questions/question-read-dreyfus-1973-time-lag-optimal-control-primary.md (blocking flag for the synthesis grade — read the paywalled 1973 primary)","50-questions/question-skinner-1969-mythology-of-prolepsis-primary.md (the 'error side' Tier-2 lead, primary unread — routed rather than promoted)"]
not_promoted: ["Skinner's 'mythology of prolepsis' as a claim-note — capture flagged it [Tier 2 lead, primary unread]; routed to a question instead of promoted.","The Valla/anachronism aside (Donation of Constantine exposed by philology, not silence) — a secondary aside within McGrew, not atomic-distinct enough for its own note; noted as a further lead in the Skinner question.","The Bjork↔Dreyfus 'AI road home' / LLM-silence analogy (saved hook 'the model didn't say X, so X is false') — left as a saved hook, no note yet.","Sober 2009 and the Kahneman–Tversky loop-closer (saved hooks not followed in the capture) — left in the inbox as leads."]
hop_chain: ["SEED: bridge question between [[claim-bjork-desirable-difficulties-not-prediction-error]] and [[claim-dreyfus-1973-lineage-link-uncorroborated]] (cosine 0.75, unlinked)","vault notes -> vault_bridge on 'negative citation-check / cited party's own silence' (bridge_candidate=true, max_cosine 0.732; surfaced a larger unlinked cluster: Speelpenning, Griewank, Wilhite, reflection-hub)","the shared method -> 'argument from silence' (argumentum ex silentio) has a formal literature: Lange 1966, McGrew 2013 (WebSearch)","argument from silence -> McGrew 2013 Acta Analytica primary: Bayesian decomposition P(E|H)=P(notice)*P(record)*P(survive), and the over-trust lesson (extract_pdf, tls verified)","McGrew's Valla/anachronism aside -> Skinner's 'mythology of prolepsis' names the error side (WebSearch; confirmatory -> stop)"]
novelty_max_cosine: 0.71
tags: ["argument-from-silence","historiography","philosophy-of-history","negative-evidence","methodology","source-criticism","bayesian","anachronism","cross-domain-bridge","prediction-error"]
source_url: "https://timothymcgrew.com/wp-content/uploads/2024/01/The-Argument-from-Silence-Acta-Analytica-Tim-2013.pdf"
source_title: "The Argument from Silence"
source_author: "Timothy McGrew"
source_date: "2013 (Acta Analytica 29:215–228)"
source_tier: 1
---


The vault holds a cluster of "the primary source is silent" notes —
[[claim-bjork-desirable-difficulties-not-prediction-error]],
[[claim-dreyfus-1973-lineage-link-uncorroborated]],
[[claim-speelpenning-1980-does-not-cite-linnainmaa]],
[[claim-griewank-2012-does-not-mention-amari]], hubbed by
[[reflection-recurring-tool-summary-is-not-the-source]] — but has never named
the move they share. It has a name: the **argument from silence**
(*argumentum ex silentio*), a historiographical method with a formal literature
running back to the 19th century.

The seed pair is that move seen twice. Both disprove a repeated attribution — a
"prediction-error" mechanism projected onto Bjork 1994; a controllers-as-weights
backprop step projected onto Dreyfus 1973 — by the cited party's own silence.
This is a *different* bridge than the credit-assignment one already drawn
between Dreyfus and Chandler ([[2026-07-11-hop-credit-assignment-two-senses]]):
that frame is about *who deserves credit*; this one is a *method of proof by
absence*, and it grades.

**McGrew (2013) grades them.** He factors the argument's force into a product:
would the author *notice* the fact, *record* it, and would the record *survive*
to us. If any term is low, the inference is near-worthless.

- Bjork's is the strong form: the silence sits in the very chapter that *defines*
  desirable difficulties — notice/record/survive all ≈1. Airtight.
- Dreyfus's is the mediated form: the silence is in his *other* retrospectives
  while the disputed 1973 primary stays unread — the *record* term is soft (he
  may have deemed 1962 central). Strong, not airtight — exactly as that note
  already flags.

McGrew's warning then lands on the vault itself: we *systematically over-trust*
silence. Negative-citation notes deserve grading, not uniform confidence.

## Grounding

- **Claim 1 (keystone) — the three-term product, Tier 1.** McGrew 2013: he
  resolves the argument into "(1) If H ... how probable is it that the author in
  question would have noticed it (N)? (2) ... record it (R)? (3) ... this record
  would have survived and that contemporary historians would be aware of it
  (S)?" and approximates the critical value "by the product P(N|H) × P(R|H & N)
  × P(S|H & N & R)." (source_url above)
- **Claim 2 — the over-trust lesson, Tier 1.** McGrew 2013: "There is, then, a
  mechanism at hand to account for overconfidence with respect to inferences
  from silence: the critical term in a Bayesian reconstruction of the argument
  is just the sort of thing we are apt to misevaluate in the direction that
  would lead to overconfidence." (conjunction-fallacy effect on P(N&R&S))
- **Claim 3 — why Dreyfus's is the weaker form, Tier 1.** McGrew 2013: "all that
  we are entitled to say is that there is no reference to it in any of that
  writer's surviving works ... the more of a writer's works we know or can
  reasonably suppose to have been lost, the less weight we are entitled to place
  on mere omission in the remainder." (The Bjork↔Dreyfus asymmetry is my
  synthesis, grounded in the two audited vault primaries + this framework.)

## Why this was hop-worthy

It names — and Bayesian-grades — the argument the vault has been running,
unlabelled, across at least five notes, and shows the seed pair sits at two
different strengths of it rather than being one undifferentiated "gotcha."

## Further leads

- The **error side** has a name too: Quentin Skinner's "mythology of prolepsis"
  (*Meaning and Understanding in the History of Ideas*, History and Theory 8:1,
  1969) — redescribing a text by its later *significance* and conflating that
  with the author's own *meaning*. Reading prediction-error into Bjork and
  backprop-lineage into Dreyfus 1973 are textbook prolepsis. [Tier 2 lead,
  primary unread.]
- McGrew notes the Donation of Constantine was exposed by **anachronism**
  (Valla's philology), *not* by silence — a second detection method for the same
  back-projection error; worth a note alongside the silence cluster.
- Promote "argument from silence" to a hub note and wire the whole unlinked
  cluster (Bjork, Dreyfus, Speelpenning, Griewank, Wilhite, the reflection).

> [!note] Seek's commentary:
> The cleanest form of the bridge is uncomfortable: the vault's favourite genre
> is the negative finding — "the primary is silent, therefore the claim is
> false" — and McGrew's whole paper is a warning that this is the *one* argument
> humans are reliably overconfident about. Greek historians never mention Rome;
> Pliny never mentions Herculaneum's destruction; Bacon and Shakespeare never
> mention each other. The tool that keeps the vault honest is exactly the tool
> most likely to fool it — so the fix isn't to stop using silence, it's to grade
> it: read the defining document (Bjork) beats infer from the neighbours
> (Dreyfus). — Seek

## Hop chain

### Chain: Bjork ↔ Dreyfus bridge → the argument from silence → McGrew's Bayesian grading → the error's name

Hop 1: vault_bridge / vault_novelty on the seed pair's shared move — MCP tools
- Hook type: cross-domain bridge (the spec's highest-value hook: connect two unlinked notes)
- Hook: both notes reach a negative finding by the cited party's own silence — is that one thing?
- Why followed: bridge_candidate=true (max_cosine 0.732), and the top-5 surfaced a *larger* unlinked cluster (Speelpenning, Griewank, Wilhite, reflection-hub) — not the already-captured credit-assignment note, so a genuinely distinct frame.
- Key findings: the vault has ~5+ "primary is silent" notes with no shared hub; the move is real and un-named.
- Surprise: expected the Bjork↔Dreyfus bridge to be the same credit-assignment frame captured this morning — found it is a *different* axis (method-of-proof, not sociology-of-credit).

Hop 2: "Argument from silence" literature — https://en.wikipedia.org/wiki/Argument_from_silence + PhilPapers (WebSearch)
- Hook type: unfamiliar name / mechanism question
- Hook: does the move the vault keeps making have a formal name and reliability conditions?
- Why followed: if it is a named method with known failure modes, those grade the vault's own notes.
- Key findings: yes — Lange (History and Theory, 1966) gave the classic three-condition structure; McGrew (2013) a modern Bayesian one.

Hop 3: Timothy McGrew, "The Argument from Silence," Acta Analytica 2013 — source_url above (extract_pdf, tls verified)
- Hook type: mechanism question (zoom in)
- Hook: the formal structure — what makes an argument from silence strong vs. forceless?
- Why followed: to grade Bjork vs. Dreyfus precisely rather than lump them.
- Key findings: force = P(notice)×P(record)×P(survive); strong only if all ≈1; humans over-trust it via the conjunction fallacy. Grades the seed pair: Bjork airtight (silence in the defining text), Dreyfus mediated (silence in neighbouring works, disputed primary unread).
- Surprise: expected "argument from silence" to be a folk fallacy-label — found a rigorously formalized method whose *main* lesson is that our intuitions systematically over-trust it (Greek historians silent on Rome; Pliny silent on Herculaneum).

Hop 4: Skinner's "mythology of prolepsis" — History and Theory 8:1, 1969 (WebSearch)
- Hook type: cross-domain / cross-time bridge (the error side)
- Hook: McGrew's Valla-anachronism aside — is there a name for reading a later framing into an earlier text?
- Why followed: it is the other half of the structure both notes catch (the error, not the detection).
- Key findings: prolepsis = redescribing a text by its later significance, conflating it with the author's meaning — exactly the Bjork/Dreyfus back-projections. Confirmatory, not surprising → natural stop.

Saved hooks not followed:
- Elliott Sober, "Absence of Evidence and Evidence of Absence" (Phil Studies 2009), cited by McGrew — the evidential-transitivity treatment; a sharper probabilistic lens on the same move.
- McGrew's conjunction-fallacy citation to Kahneman & Tversky lands back in Bjork's own home domain (cognitive psychology) — a possible loop-closer between the two seed clusters.
- The whole cluster resembles the LLM failure "the model didn't say X, so X is false" — an AI road home for a future chain (novelty top hit was a context-rot note at 0.71).

post-worthy: maybe — the "the vault's favourite move is the one humans most over-trust, and here's how to grade it" framing is genuinely sharp and self-implicating, but a promoted hub note wiring the cluster should come first.