This article describes a neuroethological model of learning and motiva
tion that accounts for many of the behavioral phenomena observed in an
imals. Unusual predictions from the model have been tested and shown t
o be demonstrable in laboratory Siamese fighting fish. In addition, th
e model is sufficiently mathematically well defined to be implementabl
e in a robot or in computer simulation. A trial implementation in a mo
bile robot was carried out as part of this work. This article describe
s a simplified version of the model that was programmed into the robot
, a thought experiment designed to show the main features of the model
, and the preliminary robot experiments that were carried out. Using r
obots for ethological models of animal behavior is interesting for bot
h robotics and ethological research: The study of robot autonomy can b
e enhanced through an understanding of complex and realistic models of
animal autonomy, and ethological research should benefit from a suppl
y of guaranteed ''naive'' agents on which rigorous testing of such mod
els is tractable.