Notes·citation verification·29 July 2026

Indexed Is Not Reachable

A citation-verification note.

The task

Verify a citation before submitting a manuscript. The target:

Schredl, M. (2007). Personality correlates of flying dreams. Imagination, Cognition and Personality, 27(2), 129–137. doi:10.2190/IC.27.2.d

The author has no institutional subscription. This is the ordinary condition for independent researchers, and it is the condition under which most citation errors survive: verification is expensive, so people cite from secondary descriptions and hope.

The paper was needed for one clause in an introduction. Not the full text, not the statistics. The abstract would have been enough.

Six lookups, six failures

Unpaywall aggregates institutional repositories, preprint servers and author self-archives worldwide. Zero locations is not “we did not find one.” It is closer to “there is not one.”

The Semantic Scholar response carried an explicit note: the abstract field had been elided at the publisher's request. So the abstract existed in the corpus and was withheld at serve time. This is the part worth noticing. The paper was fully indexed everywhere and reachable nowhere.

The remaining option was to email the author and wait.

Cito

get_paper(doi="10.2190/IC.27.2.d")

Returned the abstract in full, plus structured metadata: year: 2007, corpus_id: 146630268, cites: 26, fwci: 0.2868, institutions: Central Institute of Mental Health, topic classification, citations by year.

Same corpus identifier as the Semantic Scholar record that returned null. Cito reads abstracts from a local snapshot rather than requesting the field from the public API at query time, so serve-time field elision does not apply to it.

Elapsed: under a second, against an unknown number of days waiting on an email that might not arrive.

What the abstract actually said

This is where the note stops being about retrieval.

The manuscript already contained a sentence built on this citation: flying-dream report associated with low neuroticism, falling-dream report with high neuroticism. That pairing had been taken from a later paper by the same author describing his own earlier work.

The abstract does not mention falling dreams. Not once.

“Their occurrence in a student sample (N = 444) was related to low neuroticism, openness to experience, boundary thinness, dream recall frequency and playing an instrument.”

“The traits associated with flying dreams positively support the continuity hypothesis at the level of emotions.”

The falling half was either in the full text and absent from the abstract, or the author had compressed two findings into one when citing himself thirteen years later. Either way it could not be asserted on the evidence in hand, and it was removed from the manuscript along with a dated note recording why.

A second lookup settled a smaller question. The author's own later reference list dates the paper 2008. Crossref's publisher-deposited metadata gives published-print: 2007-10; the Crossref created timestamp is 2008-03-24, which is the likely origin of the error. The manuscript now says 2007.

The point

Two errors were about to reach a journal submission. Both came from trusting a secondary description of a primary source, which is what researchers do when the primary source is three clicks and a paywall away. Neither error was careless. One of them originated with the author of the cited paper.

The distance between “this paper is indexed” and “I can read enough of it to cite it responsibly” is where citation errors live. Closing that distance is not a convenience feature. It is the difference between a verified reference list and a plausible one.

Cito closed it in one call.

Retrieval logs and the resulting manuscript diffs are available on request.·All notes·Docs·Search