5 ideas
Primatologist Thelma Rowell studied sheep and fed her flock with one more bowl than there were animals. The spare bowl meant no sheep was forced to fight for food, so any competition or cooperation she saw was a choice, not something the setup produced. Her aim was to give sheep the same chance to be socially complex that primates had been given, rather than assuming dullness and designing studies that could only confirm it.
Robert Rosenthal told student experimenters that their lab rats had been bred as either maze-bright or maze-dull, though the rats were assigned at random. The rats labeled bright learned the maze faster, because students who believed in them handled them more gently and attentively. Despret reads this as more than contamination: the rats responded to being trusted, so the experimenter's expectation became part of what the animal could actually do.
Rigid dominance hierarchies in primates were mostly recorded under captivity, provisioned feeding, or crowding, where scarce resources force repeated contests. Rowell found little stable hierarchy in forest baboons foraging freely in Uganda. Despret argues that such hierarchy tells us about the stress of the research conditions and the observers' focus on conflict, not a fixed trait of the species.
When an animal's result turns out to depend on the experimenter, as when Clever Hans read his questioners' tiny body cues instead of doing arithmetic, science usually throws the result out as an artifact. Despret treats it instead as a finding about what the animal can do: Hans showed an extraordinary sensitivity to human bodies. Viewed this way, a contaminated result points to a capacity that the original question was not designed to ask about.
An experiment can only get rich answers if its question matters to the animal from within its own world of relevance, its Umwelt. A good question also leaves the animal room to refuse, redirect, or redefine the task. When an animal fails a test, the first thing to ask is whether the test was meaningful to the animal, before concluding that the animal lacks the ability.