Untaggable papers — a study

Papers whose stored full text cannot be faceted. A first-class, queryable category (untaggable tag), not a silent skip.

7 untaggable
1,022 corpus total
0.7% untaggable share
0 failed on clean text
Key finding. Every untaggable paper here failed at text extraction (garbled source layer), not at faceting. On clean text the method faceted 1015/1015. So this 0.7% measures digitisation/sourcing quality, not tagging-method coverage. Characteristic signal so far: untaggable papers skew older (median year 1994 vs 2017 for tagged) and paywalled (14% personal-copy vs 4% among tagged) — i.e. this correlates with age & access, not reliability. The reliability signal lives elsewhere: in the facet profile of papers that tag successfully (bold claim + no stated limitations + untested/failed replication) — a separate scan.

By reason

Extraction failuren
interleaved sources (two papers mixed)5
interleaved columns (two columns read as one)1
corrupt text layer (jammed/dropped characters)1

The papers

iddisciplineyearlicensereasondetail
breitmeyer2005-primingneuroscience2005personal-copycorrupt-layercorrupt-text-layer: text extracted as fragmented figure/table remnants (run-1 garbled paper); no clean prose. Not enriched.
nash1950-equilibriumeconomics1950open-accessinterleaved-columnsTHREE PAPERS IN ONE FILE and the middle one is line-interleaved. The staged text is a PNAS page range: the tail of an Eilenberg/MacLane homo
grover1996-searchquantum-computing1996arxiv-openinterleaved-sources
janisse1970-attitudinal-effectssocial-psychology1970reference-onlyinterleaved-sources
Wilson1994-reactivation-of-hippocampal-ensemble-memorieneuroscience1994not-redistributableinterleaved-sources
akp-nar2025-critical-appraisal-of-evidence-inmedical-education2025interleaved-sources
rosenbaum1982-a-priming-method-for-investigatingcognitive-psychology1982interleaved-sources

© 2026 Shir Sivroni
Meta-Tagging Project · untaggable study · regenerate: python TOOLS/make_untaggable_report.py
Built from corpus 1022 rec . 86,783 anchors . 2026-09-01 15:40:57 . bb6e90.