What’s in a Meow? A Study on Human Classification and Interpretation of Domestic Cat Vocalizations

Why it matters and research question

Can humans identify whether a meow was recorded while a cat waited for food, was isolated, or was brushed? Does cat experience relate to performance? Distinguishing perceived meaning from verified context helps prevent overconfident interpretations of feline needs. No earlier reading required.

Method

225 adults completed an online questionnaire and heard six single meows: two from each context, selected using acoustic features to represent a typical example (called a “medioid”) and an unusual example. The recording pool came from ten Maine Coon cats living with one owner. Participants chose among the three contexts, rated perceived emotion and reported experience and empathy. Later analyses focused on the typical examples.

Findings — evidence

• For the three typical examples, correct classification was 40.44% for waiting for food, 26.67% for isolation and 32.89% for brushing (Table 1). Typical examples were identified more accurately than unusual examples.
• Cat owners outperformed nonowners for isolation and brushing, but not significantly for food waiting (Table 2). These are associations, not evidence that ownership causes better decoding.
• Emotion-descriptor analysis using all responses showed no clear separation. Among correctly classified responses, isolation sounded more negative and brushing more positive (§4/Figure 3). This selection matters: it does not demonstrate reliable emotion recognition across all listeners.

Limitations and interpretation — Lumi

Only six recordings were tested, with one typical exemplar per context. Many human listeners do not compensate for such sparse sound sampling or the single-breed, single-household recording pool. Human judgments do not establish what the cat actually felt. The online volunteer sample and self-reported experience/empathy also limit causal conclusions.

The authors report that typical-example classification was not significantly above chance, but the text labels chance “0.33%”; three equally likely choices imply one-third (33.33%), not 0.33%. This is a source-level numerical inconsistency, not a formal correction found today. Keep the raw percentages and attributed author claim distinct; no independent reanalysis was performed. Do not infer a broad inability to understand cats from this narrow task.

Reading route

Simple Summary → §2.2 Cat Vocalizations → Table 1 → §4 paragraph before Figure 3 → Discussion. Notice the restricted sound set and the shift to correctly classified responses.

Guided reading — Lumi

Explanatory synthesis of the existing note, not new research. The archived check covered the full article, but not an independent audit of supplements/audio; no new source audit or reanalysis is claimed.

1. Separate recognizing a situation from translating a sound

• What to look for: Listeners assigned isolated meows to food waiting, isolation or brushing, then rated perceived emotion.
• What it means: The known answer was the recording context, not a verified sentence or the cat’s internal emotional state. A plausible emotional description is a different outcome from correct classification.
• Why it matters: This tests human interpretation under restricted listening conditions, not a universal cat dictionary.

2. Count the sounds, not just the listeners

• What to look for: The questionnaire involved 225 adults but only six meows, drawn from ten Maine Coons in one household. Each context had a typical and an unusual example.
• What it means: The typical example, called a “medioid,” represents a central acoustic example rather than every sound in that context. Later analyses focused on these typical sounds.
• Why it matters: More listeners improve our picture of responses to those recordings; they do not supply missing diversity of cats, households or calls.

3. Follow which responses enter each analysis

• What to look for: Emotion descriptions did not clearly separate contexts across all responses. Differences appeared when considering correctly classified responses.
• What it means: Selecting successful answers changes the question: how successful listeners described sounds is not how reliably everyone recognized emotion.
• Why it matters: Keep that restricted result separate from claims that humans generally decode feline feelings.

4. Keep experience and uncertainty in view

• What to look for: Ownership was associated with better classification in some contexts, not all; the note also flags inconsistent chance notation in the paper.
• What it means: Ownership was not experimentally assigned, and the numerical inconsistency has not been resolved by reanalysis here.
• Why it matters: Neither ownership nor confident listening guarantees accuracy. Adding familiar cats or visual context is a reader’s proposed follow-up, not a procedure tested in this task.

Two reflection questions

1. Would recognizing your own cat’s meows be a different skill from classifying an unfamiliar cat’s calls?
2. How would adding many cats and many recordings per context change what we can generalize?

Connections

• Investigates: Human interpretation of meows — accuracy and perceived emotion are separate outcomes.
• Investigates: Vocal communication — human reception of cat calls.
• Compares with: 2009 McComb - Cry embedded within the purr — context classification of meows versus perceived urgency and sound editing of purrs; neither is a universal translation test.

Source access and verification scope

Verified 2026-09-27 Asia/Seoul. Publisher checked for exact title, all nine authors, date, DOI, journal and article number. PMC substantive methods, tables/results, discussion and license inspected, including the omitted middle of the initially truncated extraction. Supplements/audio were not independently audited. Publisher spells author Giorgio Presti; a search aggregator rendered Giorgio Lo Presti, which was not adopted.

Publisher version notes (https://www.mdpi.com/2076-2615/10/12/2390/notes) list the December 14, 2020 version-of-record PDF and later HTML updates, most recently September 1, 2025, without explanatory correction content. Publisher/PMC inspection and DOI-specific correction search revealed no formal correction notice; this is a bounded check, not a guarantee that none exists. Do not mistake HTML update dates for a new study or publication year.