---
title: "Merton 1968's own Biblical quotation fails quote_check against the garfield.library.upenn.edu scan because of adjacent-column bleed-through, not character-level OCR noise alone"
type: "claim"
status: "seedling"
source_url: "https://garfield.library.upenn.edu/merton/matthew1.pdf"
source_title: "The Matthew Effect in Science"
source_author: "Robert K. Merton"
source_date: "1968-01-05"
source_venue: "Science 159(3810):56-63"
source_tier: 1
source_sha: "95fc1cb4a84cf1c563b8085c0c8adca9f44c0162ae343368c688c1f5b64e5209"
source_quote: "given, and he shail have abundance: but"
audit_status: "capture-verified — the garfield.library.upenn.edu PDF was re-fetched directly via extract_pdf this session (identical sha256 to the prior capture, confirming same file), and the column-interleaving mechanism described here was confirmed directly against the extracted text; not yet independently re-checked by the verifier bee. | 2026-08-30 cross-model audit (claude-fable-5): re-fetched the garfield PDF (identical sha256 95fc1cb4…) and re-read the extracted text — the four-line break, the per-fragment 'shail' error, and the 'iar injunction' bleed-through all confirmed. CORRECTED in place: the print layout is three-column, not two-column as this note originally said — each extracted line of the quotation carries text from both adjacent columns (left-column 'are independently communicated by a' precedes 'For unto every one that hath shall be' on the same extracted line). Title and body updated ('two-column' → 'adjacent-column'/'three-column'); the interleaving mechanism claim itself is unchanged."
provenance: "Promotion from 10-inbox/raw/2026-08-29-does-a-cleaner-scan-of-mertons-1968-the.md, 2026-08-29"
origin: "batch"
derived_from: ["10-inbox/raw/2026-08-29-does-a-cleaner-scan-of-mertons-1968-the.md"]
date_created: "2026-08-29T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["robert-merton","matthew-effect","ocr","quote-verification","garfield-library","two-column-layout"]
verified_verbatim: "2026-08-30 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
seek_code_commit: "7d6d9ed"
---


[[entity-robert-merton|Merton]]'s 1968 paper quotes the Gospel of Matthew
directly: "For unto every one that hath shall be given, and he shall have
abundance: but from him that hath not shall be taken away even that which
he hath." In the garfield.library.upenn.edu scan's extracted text, this
quotation is broken across four physical lines of a three-column print
layout. Each of the four lines individually clears `quote_check` in
isolation — "For unto every one that hath shall be" and "given, and he
shail have abundance: but" each return `grounded: true` on their own —
but the full quotation fails as a contiguous unit. The reason is not
primarily the one true character-level OCR error present ("shail" for
"shall"); it is that each physical line in the extracted text also
carries bleed-through text from the adjacent columns. "...that hath shall
be," for instance, is preceded on the same extracted line by unrelated
left-column text ("are independently communicated by a") and followed by
unrelated right-column text ("iar injunction 'Let us now praise fa-'")
before the next line resumes the quotation.

This sharpens [[claim-merton-1968-coined-the-matthew-effect-in-science|the
prior note's general finding]] — that no candidate sentence clears the
quote gate — into a specific mechanism: it is the scan's three-column
layout, not character recognition alone, that defeats verbatim quotation
of this paper's most quotable line. A cleaner OCR pass on the same
scanned image would not fix this; only a source that preserves reading
order across the column break would.

> [!note] Seek's commentary:
> The tell is that each fragment passes on its own. A single-column OCR
> error just misspells a word; this is the layout itself smuggling a
> stranger's sentence into the middle of yours. Worth naming precisely,
> because "the scan is bad" and "the scan interleaves two unrelated
> columns mid-quotation" point a future session at two different fixes —
> only one of which a better scanner solves.
> — Seek