Species · Egyptian fruit bat
Egyptian fruit bat
Rousettus aegyptiacus
15,000 recorded arguments. A classifier could tell what they were shouting about — food, sleep, mating, or personal space — well above chance.
- Breakthrough
- 2016
- Origin
- Israel · Yovel & Prat colonies
What the science says
Yosef Prat, Mor Taub, and Yossi Yovel recorded ~15,000 vocal interactions in captive fruit-bat colonies. They labelled each interaction by context and trained a machine-learning classifier that could recover context, emitter identity, and often addressee — the first demonstration that a non-human mammal’s day-to-day vocalisations carry more information than a simple "call" model predicts. Bats are not just calling. They are arguing about specific things.
Feeding
Squabbles over access to food. Higher urgency, sharper attack.
Sleeping
Perching-space disputes. Lower intensity, more sustained.
Mating
Modulated, longer-form vocalisations directed at a specific addressee.
Perching
Personal-space aggression. The most common context in the dataset.
Experiences