The third possible relationship is a zero or non-significant
correlation between performances on the two tasks, which
would imply that the processes behind them are independent.
These are the three hypotheses we contrast in an experiment
on captive pigeons. To make sure that our results are not
artefacts of captive conditions, we also test flocks of urban
pigeons on the innovation task, social learning being impossible
to distinguish from individual learning in the uncontrolled
setting of the field (Lefebvre 1986). Cognitive tasks
that force a food-deprived animal housed alone in a cage to
interact with an unfamiliar apparatus may have questionable
validity. If the task can be solved by animals in the field,
under normal conditions of hunger, social interactions and
feeding options other than the one offered by the apparatus,
we can be more confident that it assesses abilities animals
use in their usual environment.