Psychological Effects of Cultural Upheavals from Millions of Song Lyrics Over 100 Years
cs.CL
Quick verdict
Claims century of lyrics show war, 9/11, COVID shifted artists' language; credible within-artist design, but effects are tiny and secular trends confound.
Summary
The paper analyzes roughly 3 million song lyrics (1903–2026) from LRCLIB, matched to MusicBrainz metadata, to test whether three US cultural upheavals — the 1965 Vietnam escalation, 9/11, and the onset of COVID-19 — coincided with shifts in artists' language. Using LIWC-22, the author scores self-references, collective-references, analytic thinking, and cognitive processing within symmetric five-year pre/post windows, separating within-artist change from between-artist compositional change via group-mean centering in linear mixed models with random intercepts for artist, genre, and language. Self-references fell within-artist after Vietnam (t=−3.35) and 9/11 (t=−7.38; marginally after COVID-19), while cognitive processing rose after all three events; collective-references rose after 9/11 (t=13.66). Lagged 12-month-bin models suggest reduced self-focus persisted roughly a decade after Vietnam and 9/11. The Meaning Extraction Method yields seven themes, with 'life and freedom' and 'moral conflict' rising and 'courtship and nightlife' falling post-upheaval. Main limitations: minuscule effect sizes (Cohen's f² mostly ≤0.0005), no correction across roughly 24 tests, no placebo windows or pre-trend checks to rule out secular lyric trends, noise from machine-translated lyrics and chorus repetition, and no stated code or data release.
Review complete
Assessment at a glance
4
risk flags identified
● high 1
● medium 2
● low 1
5 strengths · 5 weaknesses · 5 author questions
AI text signal = Binoculars zero-shot detector score over 20 sampled passages (Fast-DetectGPT co-reported) — absence does not prove human authorship, and a flag does not imply misconduct.
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Within-artist self-references — a marker of psychological distancing — decreased significantly in the five years after the Vietnam escalation (t=−3.35) and 9/11 (t=−7.38), and marginally after COVID-19 (p=.053).
- Cognitive processing terms rose post-upheaval for all three events, which the author reads as artists' increased meaning-making.
- Within-artist reductions in self-focus persisted roughly a decade after Vietnam and 9/11 (Figure 2 lagged analysis).
- Content moved post-upheaval toward 'life and freedom' and 'moral conflict' and away from 'courtship and nightlife', corroborating the style-based findings.
- Within-artist and between-artist effects frequently diverge or reverse in sign, so designs that conflate them misestimate upheaval effects.
Strengths
Unmatched scale with conservative filtering
32,254,478 LRCLIB lyrics reduced to 3,064,753 analyzable songs (1903–2026) via tiered MusicBrainz matching (46.9% retained; duration-mismatched pairs excluded) — the pipeline is documented step by step, from fastText language ID to nllb-200 translation to one-row-per-recording deduplication.
Within/between decomposition is the real contribution
Group-mean centering (Enders & Tofighi, 2007; Mundlak, 1978) reveals opposite-sign effects — 9/11 self-references at −0.14 within-artist but +0.18 between-artist (Table 2) — which a conflated pre/post design would average into a single meaningless estimate.
Transparent effect-size and sample reporting
Layered robustness checks
Style and content converge
Weaknesses
No counterfactual for secular trends
Vanishingly small, uncorrected effects
LIWC validity on sung, translated text
Selection and coverage bias
Covers as independent observations
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
Report preview
Scope & Claims
The paper asks whether three dated US upheavals — the Vietnam escalation of March 8, 1965, the September 11 attacks, and the onset of COVID-19 (January 10, 2020) — left measurable traces in the language of popular music. From 32,254,478 LRCLIB lyrics, 3,064,753 songs spanning 1903–2026 survive preprocessing and enter LIWC-22 scoring of four dimensions: self-references, collective-references, analytic thinking, and cognitive processing. The headline claims are that within-artist self-references fell after Vietnam (t=−3.35, p=.001) and 9/11 (t=−7.38, p<.001; marginally after COVID-19, p=.053); cognitive processing rose after all three events (t=3.07, 2.07, 7.92); reduced self-focus persisted for roughly a decade (Figure 2); and content moved toward 'life and freedom' and 'moral conflict' while 'courtship and nightlife' receded. Claims are framed as correlational, which is appropriate.
The remaining sections cover Methodology Assessment, Statistical Evidence, Results & Reproducibility, Limitations & Risks, and Verdict & Recommendations.
- All 5 strengths & 5 weaknesses, analyzed in detail
- 4 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
What did the study find about song lyrics after 9/11 and other major events?
Analyzing about 3 million song lyrics (1903–2026), the author reports that singers reduced first-person-singular words (‘I’, ‘me’) in the five years after the Vietnam escalation and 9/11 — a marker of psychological distancing — and increased cognitive processing words (‘because’, ‘should’) after all three events studied (Vietnam, 9/11, COVID-19). The shifts were statistically significant but extremely small in magnitude.
Where did the lyrics data come from and how many songs were used?
Lyrics came from LRCLIB, a public lyrics database, matched to MusicBrainz metadata for release dates, genre, and language. Of 32.25 million collected lyrics, about 3.06 million songs released between 1903 and 2026 entered the final analyses after filtering for usable matches, release dates, and scorable text; non-English lyrics were machine-translated into English first.
Does this study prove that wars or pandemics change how musicians write?
No. The design is correlational: it compares the same artists' lyrics before versus after each event, but it has no counterfactual period, and song lyrics have drifted steadily over the past century for many reasons. The author explicitly labels the findings non-causal, and reviewers note that secular time trends could partly explain the patterns.
Is the dataset or analysis code publicly available?
The paper text does not state a public code or data release. The two underlying sources (LRCLIB lyrics and MusicBrainz metadata) are publicly downloadable, but LIWC-22, the text-analysis tool used, is commercial software.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Figures 1–2 are present only as captions; year-level trends and CI patterns could not be inspected from the extracted text.
- Supplementary Tables S1–S6 and the online supporting information are not included; English-only and thematic-control results are taken from in-text summaries.
- Table 2 extraction is dense with occasional ambiguous column alignment (e.g., the Vietnam 'Informative' count reads 23,779); exact digits were used as extracted.
- No data/code availability statement appears in the extracted text; one may exist elsewhere in the arXiv version.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.