Transitive inference (TI) effects have been demonstrated in several no
nhuman species using a nonverbal version of the task in which A is bet
ter than B is represented in a simple simultaneous discrimination, A+B
-. Following five-term training (e.g., A+B-, B+C-, C+D-, D+E-), the ch
oice of B over D on test trials is taken as evidence of TI. Recently,
differential value transfer from the S+ to the S- in a simultaneous di
scrimination has been proposed as a noncognitive basis for these effec
ts. Two experiments were conducted to control for differential value t
ransfer. The results suggest that differential value transfer is not t
he only basis for nonhuman TI performance. An alternative account base
d on spatial mapping is discussed.