Hooded crows are capable of
transitive inference formation Four crows were trained in multiple instrumental discrimination task: A+ B-, B+ C-, C+ D-, D+ E- (letters stand for stimuli, plus and minus indicate rewarded or non-rewarded stimulus). Stimuli were colored cards with a circle of the same colour drawn on reverse side. Circles' diameters were decreased in stimuli series. To preclude an influence of reinforcement history on test choice we calculated reward/penalty ratio for B and D after training. Then overcompensation phase was presented consisting of training pair with different presentations' frequency. After that reward/penalty ratio for D was 1.5 - 2.0 times greater than for B. If birds chose stimulus according reinforcement history in test they will prefer D; if birds used an inference-like process they will prefer B. In test crows strongly preferred stimulus B over D (83.1%). In second experiment four naive crows were trained with the same design but the circles' diameters were equal. The performance with BD in two crows was above chance (53.1%), and two other chose D with the greater reward/penalty ratio (80.0%). If crows have additional information about stimuli which presumably allow them to construct an ordered series they are able to solve transitive test using non-associative mechanisms. Keywords: crows, birds, transitive inference, visual discrimination |
|
|
|
|
|
|