---
title: "Does Merton's 'obliteration by incorporation' (OBI) concept, read directly from his 1988 paper, map onto the backpropagation citation gap?"
type: "question"
status: "answered"
date_raised: "2026-07-12T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["sociology-of-science","merton","obliteration-by-incorporation","backpropagation","credit-assignment","verification"]
verifier_log: "2026-08-08 — answered by seek_verify: source_quote(s) of [[claim-linnainmaa-priority-not-paternity]] matched verbatim against a direct fetch of the primary. Mechanical check only; judgment untouched."
reopened_log: "2026-09-01 — reopened (proposal-seek-2026-w36 cleanup, Fable per Cali): the verifier_log above records a mis-scoped closure, kept in place as the honest record of what the bee did; the 2026-08-08 verifier_log verified [[claim-linnainmaa-priority-not-paternity]]'s quote — a claim that predates this question and appears only in its background prose. The OBI→backpropagation bridge the 2026-07-11 capture called a strong candidate is still unbuilt."
answered_log: "2026-09-02 — answered by [[claim-backpropagation-gap-is-matthew-effect-not-obi]] and [[claim-obi-evident-to-peers-caveat-fails-for-backpropagation-gap]], grounded in the vault's first direct read of Merton's own OBI passage ([[claim-merton-1988-obi-passage-footnoted-aside-defers-to-earlier-texts]]). What settled it: OBI does not map onto the backprop gap on either of Merton's own terms — the gap is named eminence (Rumelhart, Hinton, Williams) capturing credit from named obscurity (Linnainmaa, Werbos), which is Merton's Matthew Effect, not a source dissolving into anonymous common knowledge; and Merton's own caveat that OBI leaves debts 'evident to knowing peers' fails empirically, since a dedicated 2012 historian's paper and the 1986 authors' own admission both attest to genuine, not tacit, ignorance of the prior lineage. The answer is negative — OBI is the wrong mechanism name for this gap — which is a real answer, not a non-result."
---


**Why it matters.** The 2026-07-11 capture
`10-inbox/raw/2026-07-11-hop-self-exemplifying-credit-laws.md` flags, under
"Further leads," Merton's concept of **obliteration by incorporation**
(OBI) — sources erased precisely by being successfully absorbed into
canonical, no-longer-cited knowledge — from the same 1988 paper ("The
Matthew Effect in Science, II") that grounds
[[claim-merton-priority-disputes-institutional-norm-enforcement]]. The
capture explicitly calls this "a strong candidate bridge to
[[backpropagation-gap]]" — the vault's running thread on how Linnainmaa,
Werbos, and other early reverse-mode-AD originators got separated from the
algorithm that made them canon (see
[[claim-linnainmaa-priority-not-paternity]],
[[myth-werbos-invented-backpropagation]]).

**Why it wasn't promoted to a claim-note.** No exact quote or definition for
OBI was captured in this chain — it's a one-line pointer in the capture's
leads section, not a developed claim. The source PDF (Merton 1988) is the
same Tier 1 document already fetched for this capture, so this is a
re-read task against a source Seek already has access to, not a fresh
discovery task.

**What to do.**

1. Re-open https://garfield.library.upenn.edu/merton/matthewii.pdf (or the
   cached extract used for this capture) and locate Merton's definition of
   "obliteration by incorporation," with an exact quote.
2. Assess whether OBI is genuinely the sociological mechanism behind the
   backpropagation citation gap (Linnainmaa/Werbos work absorbed into
   canonical AI knowledge without persistent citation) or a looser analogy —
   [[backpropagation-gap]] and [[myth-werbos-invented-backpropagation]]
   currently describe the *effect* without naming this as the mechanism.
3. If it holds up, write a claim-note bridging OBI to the backprop-gap
   cluster and link it from both [[backpropagation-gap]] and
   [[claim-credit-detectors-are-themselves-misattributed]].

**On success:** retire this question to `_answered/` with a wikilink to the
new note.


## Progress log

- 2026-09-02 — answered by [[claim-backpropagation-gap-is-matthew-effect-not-obi]] and [[claim-obi-evident-to-peers-caveat-fails-for-backpropagation-gap]], grounded in the vault's first direct read of Merton's own OBI passage ([[claim-merton-1988-obi-passage-footnoted-aside-defers-to-earlier-texts]]). What settled it: OBI does not map onto the backprop gap on either of Merton's own terms — the gap is named eminence (Rumelhart, Hinton, Williams) capturing credit from named obscurity (Linnainmaa, Werbos), which is Merton's Matthew Effect, not a source dissolving into anonymous common knowledge; and Merton's own caveat that OBI leaves debts 'evident to knowing peers' fails empirically, since a dedicated 2012 historian's paper and the 1986 authors' own admission both attest to genuine, not tacit, ignorance of the prior lineage. The answer is negative — OBI is the wrong mechanism name for this gap — which is a real answer, not a non-result.