# Mehrabian 7–38–55: preliminary source readiness - Object ID: `em:research-note:sha256:48131613b1b8e5c64e3be3076401d8b749833e06ac16ef089c432a1d6f833ab2` - Kind: `research-note` - Repository path: [`research/how-we-know/library-preliminary/mehrabian-7-38-55.md`](https://github.com/yoheinakajima/epistemedia/blob/f92846570180dfa4511263f8ba98ecd18f7772c9/research/how-we-know/library-preliminary/mehrabian-7-38-55.md) - Content digest: `04445f37619922b9770a6ad62eb674c0622b4a49d61f07fe74ee1e037151aa23` **Also filed under:** [Research Program](https://epistemedia.org/topics/research-program/) ## Source content # Mehrabian 7–38–55: preliminary source readiness Status: **HOLD — lead-level mapping only; not a dossier, verdict, admission, or publication recommendation** Task: EM-0027 Source-access window represented: 2026-08-23T02:18:40Z–02:44:51Z Accepted base for this synthesis: `e9ad62b18f21594258643694c55709f78b4f9a50` This packet records public, credential-free source-scout findings. It does not credit prior model-generated notes. Publicly readable scans are access leads, not proof of reuse rights. ## Provisional disposition **HOLD.** The two primary identities, the narrow experimental target, the paper-level location of the three-number proposal, the author's qualification, and a bounded propagation sample are traceable. Source closure is incomplete: publisher full text is closed, accessible scans are unlicensed, the `7%` origin is not reported in the two papers, and the relevant book editions and original materials are not closed. Minimum later gates are lawful stable access or an approved quote-minimal treatment for both version-of-record spans; independent verification of the integration paragraph; explicit representation of the missing `7%` derivation; 1971/1981 edition reconciliation; and review of every dependence edge. ## Circulated proposition versus tested propositions The circulated proposition is that “93% of communication is nonverbal,” commonly divided into 7% words, 38% tone, and 55% body language or facial expression. The 1967 experiments instead measured how undergraduates inferred positive, neutral, or negative attitude toward an addressee: 1. one-word semantic content crossed with vocal tone; and 2. the neutral word “maybe” in three tones, paired with head-only facial photographs. They did not measure all communication, natural conversation, factual comprehension, persuasion, memory, gestures, posture, dress, spatial behavior, or three simultaneously varied verbal/vocal/facial channels in a single experiment. ## Verified primary identities, access, and quotation boundary ### P1 — words and tone **Albert Mehrabian and Morton Wiener, “Decoding of inconsistent communications.”** - Journal identity: *Journal of Personality and Social Psychology* 6(1), 109–114 (May 1967). - DOI: [`10.1037/h0024532`](https://doi.org/10.1037/h0024532); PMID [`6032751`](https://pubmed.ncbi.nlm.nih.gov/6032751/). - Metadata: [Crossref](https://api.crossref.org/works/10.1037%2Fh0024532); [OpenAlex](https://api.openalex.org/works/https:%2F%2Fdoi.org%2F10.1037%2Fh0024532). - Access: the DOI redirected to an APA page that returned 403; PubMed exposed metadata but no abstract/full text. A transient [third-party scan](https://sergiosenna.com.br/wp-content/uploads/2026/01/mehrabian1967-decoding-of-inconsistent-communications.pdf) returned a six-page PDF, 563,059 bytes, SHA-256 `d2c6cac3f178ca32a85ec29a177c2c004890333221e708fb27d6379c9b8e4ee8`. - Edition: DOI, issue header, Crossref, PubMed, and scan identity agreed; no correction or version relation appeared in the captured Crossref record. - License/treatment: Crossref declared no license and OpenAlex classified the work closed. Do not archive or redistribute the scan; retain metadata and quote-minimal necessary spans. - Study boundary: 45 undergraduates prepared word materials; a separate 30 completed the main task. P1 reports categorical analyses and **does not report a `7%` verbal coefficient** or a three-channel regression. ### P2 — face and tone **Albert Mehrabian and Susan R. Ferris, “Inference of attitudes from nonverbal communication in two channels.”** - Journal identity: *Journal of Consulting Psychology* 31(3), 248–252 (June 1967). - DOI: [`10.1037/h0024648`](https://doi.org/10.1037/h0024648); PMID [`6046577`](https://pubmed.ncbi.nlm.nih.gov/6046577/). - Metadata: [Crossref](https://api.crossref.org/works/10.1037%2Fh0024648); [OpenAlex](https://api.openalex.org/works/https:%2F%2Fdoi.org%2F10.1037%2Fh0024648). - Access: the APA target returned 403 and PubMed listed only Ovid full text. The [university course scan](https://is.muni.cz/el/phil/jaro2018/PSA_034/um/Mehrabian.pdf?lang=en) uses a credential-free but session-dependent flow: the first anonymous request returned a 117-byte HTML/meta-refresh response and set an anonymous cookie; repeating the URL with that cookie returned a five-page PDF, 459,340 bytes, SHA-256 `f9d91dfd98725e902ef24a1b478c4bb8aac5be66c2be2b5486cc0c4253711f84`. Because the carrier depends on mutable anonymous session state, it receives no durable-access credit. - Edition: authoritative metadata and the issue header say *Journal of Consulting Psychology*, despite inconsistent secondary naming. - License/treatment: no open license; public course hosting does not establish reuse rights. Link and quote minimally; do not archive source bytes. - Study boundary: three groups totaling 62 female undergraduates; outcome was inferred like/dislike attitude. P2 reports `A_T = 1.50 A_F + 1.03 A_V` and says face received about 3/2 the vocal weight. Wider population applicability was proposed, not tested. ## Cross-study derivation audit P2's final paragraph combines its implications with P1 and **suggests** coefficients `.07`, `.38`, and `.55` for simultaneous verbal, vocal, and facial attitude communication. It calls the step an extrapolation to be tested. Neither experiment manipulated all three channels. - `1.50 / 1.03 = 1.4563`, close to `.55 / .38 = 1.4474`. - Reserving `.07` and allocating the remaining `.93` in P2's `1.50:1.03` ratio produces `.5514` and `.3786`. - That arithmetic is a reconstruction, not a derivation printed in either paper. - The origin of the initial `.07` is not disclosed in P1 or P2 and cannot be recovered from P1's reported ANOVA as a stated verbal weight. The formula's appearance in P2 is verified; the complete derivation is unresolved. ## Author qualification and edition trail 1. **Albert Mehrabian author page.** - URL: [kaaj.com/psych/smorder.html](https://kaaj.com/psych/smorder.html). - Access receipt: `200 text/html`, 3,499 bytes, SHA-256 `d2e9ce6de5935b1eefcf7ada00ed2f9da11332527660fd6b98be4c909e0a5a72`. - Boundary: describes “Total Liking” and limits the equations to inconsistent messages of feelings and attitudes, not other communication. - License/treatment: copyrighted HTML, no open license observed; link/paraphrase and quote minimally. The statement's original publication date remains unknown. 2. **Albert Mehrabian, *Silent Messages*.** - Author page identity: 1981 Wadsworth edition, pp. 75–80. - Inspected lead: [1971 Wadsworth scan](https://e-edu.nbu.bg/pluginfile.php/855150/mod_resource/content/1/Albert-Mehrabian%20-%20Silent%20Messages%201971%20-%20red.size.pdf), ISBN `0-534-00059-2`, 160 pages, 19,515,952 bytes, SHA-256 `3223d08c89749142d9774a49e6843ee4512ef9daa70928ed8667051ef8a123c1`. The previously recorded digest `94925f4d17b626bc029ce8ae8375c8468788a179cc879f336a930b7b1273ddeb` was not reproduced from this exact URL and receives no artifact-identity credit. - The 1971 edition states the approximate equation and its feeling-dimension boundary but does not supply the missing `.07` derivation. - License/treatment: all rights reserved; do not preserve or redistribute scan bytes. - Gap: 1971 and 1981 editions were not fully collated. ## Bounded propagation and correction leads These objects show circulation or correction; they are **not empirical support roots**. - [Highland Council, *Staff Connections*, issue 41 (December 2020)](https://www.highland.gov.uk/download/downloads/id/23149/staff_connections_dec_2020_issue_41.pdf): broadens the proposition to all communication/video policy. - [Hampshire Police and Crime Commissioner training deck (March 2022 URL)](https://www.hampshire-pcc.gov.uk/wp-content/uploads/2022/03/MASTER-010322.pdf): broadens it to 70–93% of all communication. - [University of Birmingham, “Audio visual for events explained”](https://conferences.bham.ac.uk/2020/01/20/audio-visual-for-events-explained/): broadens it to event communication/effectiveness. - [University of Hertfordshire, “The myths of body language”](https://www.herts.ac.uk/research/centres/research-blogs/psychology-and-sport-sciences/the-myths-of-body-language): institutional correction emphasizing the single-word feeling-judgment context. - [University of Liverpool Prosper, “Communication”](https://prosper.liverpool.ac.uk/postdoc-resources/explore/communication/): institutional correction describing the context-free 93% proposition as a myth. No open licenses were observed. The purposive circulation sample is `n=3` and cannot estimate prevalence. Correction pages are leads, not substitutes for the primary papers. ## Dependence, counterevidence, and gaps - P1 and P2 are distinct article identities within one linked research program: common author, University of California grant/pools, evaluative scale, artificial crossed-channel design, female-recorded stimuli, and shared use of “maybe.” - Participant and speaker overlap across papers is not reported. - P2 cites and integrates P1; the book and author page point back to both. - The three institutional recirculation objects repeat one lineage and add zero independent empirical roots for “93% of communication.” - No raw participant data, recordings, photographs, coding sheets, or analysis code were located. - Official publisher full text was inaccessible; third-party scans establish inspection, not reuse rights or durable availability. The independently rechecked P2 carrier was unstable. - No 1967 study tested three simultaneous channels. - P2 cites P1 as pp. 108–114, while authoritative identity is 109–114; preserve the likely citation discrepancy. - Seeded follow-up/replication literature was not evaluated; no field-level effect conclusion is authorized. ## Bounded corpus and cost - Nine content objects: two primary articles, one author page, one bounded book span, three propagation objects, and two correction objects. - Six metadata/access receipts: Crossref, PubMed, and OpenAlex for each primary article. - Estimated later review: 6–10 focused hours with lawful stable views, or 1–2 working days including edition collation and derivation review. - Stop before a general nonverbal-communication review; do not count propagation as evidence or reconstruct missing derivation as fact. The recommendation remains **HOLD** pending source/edition/derivation closure and exact-head independent review under a later task.