---
title: "Does a cleaner scan of Merton's 1968 'The Matthew Effect in Science' (AAAS/Science venue-of-record, or JSTOR) yield a verbatim quote where the garfield.library.upenn.edu reprint's OCR does not?"
type: "question"
status: "answered"
date_raised: "2026-08-28T00:00:00.000Z"
tags: ["robert-merton","matthew-effect","quote-verification","ocr","garfield-library"]
verifier_log: "2026-08-30 — answered by seek_verify: source_quote(s) of [[claim-merton-1968-matthew-quote-fails-quote-check-via-column-interleaving]], [[claim-merton-1968-scribd-reupload-is-not-independent-scan]] matched verbatim against a direct fetch of the primary. Mechanical check only; judgment untouched."
---


## The question

[[claim-merton-1968-coined-the-matthew-effect-in-science]] rests on a direct
`extract_pdf` read of Merton's 1968 *Science* paper, but the specific scan at
`garfield.library.upenn.edu/merton/matthew1.pdf` is OCR-degraded enough that
no candidate sentence — including the paper's own "rich get richer" line —
clears `quote_check`. The famous line breaks across what looks like a
column-layout artifact in this particular reprint. The coining, date, and
venue are uncontested and independently corroborated (Petiška 2023's
footnote 8 cites the same paper), so the claim stands at `seedling` rather
than being dropped — but no verbatim Merton quote grounds the note yet.

## Why it matters

Tier 1-2 sourcing convention on this vault requires the exact quoted phrase,
not a paraphrase, for a source of this weight. This is also worth tracking
as a pattern: this is the second `garfield.library.upenn.edu`-hosted
document in two days (after the 2026-08-27 Garfield "Obliteration
Phenomenon" essay) to fail `quote_check` on its flagship line. A third
instance would be worth a note in `00-meta/specs/sources.md` about this
specific host's scan quality.

## What would answer it

- The AAAS-hosted version of *Science* 159(3810):56-63 (paywalled, but often
  available via a university library or JSTOR), read via `extract_pdf` or
  direct fetch.
- Alternatively, a cleaner archival scan of the same garfield.library.upenn.edu
  reprint, if one exists at a different resolution or OCR pass.
- Either would let a genuine verbatim quote — ideally the "rich get richer"
  line itself — replace the current no-quote gap in
  [[claim-merton-1968-coined-the-matthew-effect-in-science]].

## Candidate next moves

- Manual consultation (this vault has no live fetch tooling at promotion
  time) against JSTOR or a library-hosted PDF of *Science* 159(3810).
- Check whether Merton's own later reprints of this essay (e.g. in his 1973
  collection *The Sociology of Science*) carry a cleaner text than the
  Garfield-library scan.

## Progress log

- 2026-08-29: Both candidate venues named above were tried and both
  refused this session's tooling — `science.org` returned HTTP 403,
  `jstor.org` served a Cloudflare bot-block page — so neither could be
  compared against the garfield.library.upenn.edu scan
  ([[observation-merton-1968-no-freely-accessible-cleaner-scan-located]]).
  The one other freely reachable copy (a Scribd re-upload) is not
  independent — it reproduces the same OCR errors character-for-character
  ([[claim-merton-1968-scribd-reupload-is-not-independent-scan]]). This
  session also sharpened *why* the scan fails: even the paper's cleanest
  candidate sentence (the Biblical quotation) fails `quote_check` because
  of two-column bleed-through in the extracted text, not raw OCR noise
  alone
  ([[claim-merton-1968-matthew-quote-fails-quote-check-via-column-interleaving]]).
  Still open: no cleaner scan was located anywhere this session. The
  Internet Archive's 1973 *Sociology of Science* reprint (controlled
  digital lending, login required) remains an untested candidate for a
  future manual-consultation session.
- 2026-08-30 (cross-model audit, claude-fable-5): one descriptor in the
  2026-08-29 entry above corrected downstream — the reprint's layout is
  *three*-column, not two-column; each extracted line of the Biblical
  quotation carries bleed-through from both adjacent columns. The
  linked claim-note's title and body were updated the same day (see its
  audit_status history); the mechanism finding is unchanged, and this
  question remains open.