---
title: "Steven Weinberg argues Whig history is legitimate specifically for science, because scientific progress lets historians say who was right"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://www.nybooks.com/articles/2015/12/17/eye-present-whig-history-science/"
source_title: "Eye on the Present—The Whig History of Science"
source_author: "Steven Weinberg (Nobel laureate physicist)"
source_date: "2015-12-17"
source_quote: "in scientific history we really can say who was right"
source_tier: 2
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-whig-history-of-ai.md, 2026-07-12 (headless)"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-11-hop-whig-history-of-ai.md"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["whig-history","historiography-of-ai","presentism","philosophy-of-science","weinberg"]
---


In "Eye on the Present — The Whig History of Science" (*New York Review of
Books*, 17 December 2015), the Nobel physicist Steven Weinberg takes the
minority position against the historians' consensus that Whig history
([[claim-butterfield-1931-coined-whig-history-as-present-referenced-past]]) is
a fallacy to be avoided. His argument: physics, chemistry, and biology are
different from politics or art because "in scientific history we really can
say who was right." The story of a scientific episode "makes no sense unless
we recognize that some were wrong and some right… only from the perspective of
our present knowledge." He calls the pure-historicist alternative — describing
past science strictly on its own terms, without judging it against what came
later — "nonsense."

This does not license unlimited presentism. Weinberg's exemption is narrow: it
covers judging *whether a past claim turned out to be correct*, not importing
a *later concept or intent* the original actors never had. That distinction
does the real work when applied to the vault's own AI-history corrections
([[claim-werbos-1974-no-credit-assignment-language]],
[[claim-complex-systems-issue-one-paired-nettalk-with-cellular-automaton]]):
judging that Werbos's equations amount to a gradient computation is
present-knowledge appraisal of the Weinberg-licensed kind, while attaching the
retrospective label "spatial [[entity-credit-assignment|credit assignment]]" — a phrase and concept his
1974 thesis never used — is the anachronism Butterfield warned against, not
the kind Weinberg defends. See [[moc-backpropagation-origins]].

> [!note] Seek's commentary:
> The interesting tension isn't "Weinberg vs. Butterfield," it's that Weinberg
> gives a principled boundary for when retrospective judgment is fair — and
> the vault's myth-ledger, read carefully, has mostly been staying inside
> that boundary already (correcting concept-importing, not correctness
>-judging) without ever naming the boundary explicitly. Naming it is the
> value of this note. — Seek