Discrimination learning and concept formation in the Arabian horse

Abstract

Discrimination learning and concept formation abilities were investigated in four mature Arabian horses. A series of two-choice discrimination problems were presented on stimulus panels that could open to allow access to food bowls. Selection of the correct stimulus resulted in food reinforcement, and an incorrect choice was not rewarded. The positions of the correct and incorrect stimuli were varied randomly during each test session, and the criterion for learning each problem was 85% correct for two consecutive sessions of 30 or 40 trials. Testing progressed through six discrimination problems. The first four were simple pattern discriminations, but the last two incorporated several different triangles as correct stimuli and thus involved the concept of triangularity. Two of the subjects successfully completed only simple pattern discriminations, one showed evidence of learning in the first concept problem, and one completed all six tests, including the two concept formation problems. The results demonstrate complex pattern discrimination ability in horses, and suggest that they may also have the ability to form and use concepts in problem solving.

The effect of stimulus height on visual discrimination in horses1

“…in the same location as the reward). In a more recent study using a visual discrimination task to assess intelligence and learning in horses the reward was presented in the same location as the stimulus, but both at nose height (Sappington and Goldman, 1994). The results of an early study into equine colour vision where stimuli were presented at ground level (Grzimek, 1952) differ from findings of subsequent studies involving higher level stimulus presentations (Pick et al, 1994;Macuda and Timney, 1999;Smith and Goldman, 1999).…”

The cognitive capabilities of farm animals: categorisation learning in dwarf goats (Capra hircus)

“…To avoid the need for tests in extinction, we analysed only Wrst-trial performance in the transfer problems to prove generalisation abilities because subjects may rapidly learn how to respond accurately to the novel symbols (Thompson and Oden 2000; Wright et al 1988). The last criterion in particular has not always been recognised in the previous studies on categorisation (Hanggi 1999;Range et al 2008;Sappington and Goldman 1994;Vonk and MacDonald 2002).…”

“…Categorisation in horses has been conWrmed by studies using two-dimensional (2D) geometric symbols as discriminative stimuli (Hanggi 1999;Sappington and Goldman 1994). Horses have also learned to categorise stimuli based on a relative size concept (Hanggi 2003).…”

The impact of visual perception on equine learning

“…Visual discrimination training has been used to assess learning in the horse with the stimuli generally being presented at a height of 1m or above (Sappington and Goldman, 1994;Flannery, 1997;Hanggi, 1999). The initial findings of Gardner (1937b), that stimuli placed at a low level resulted in fewer errors than those placed in a high position, appear to have been disregarded.…”

“…For example, the shapes used as discriminative stimuli in the study by Sappington and Goldman (1994), and those used in the relational discrimination tests by Flannery (1997), may have appeared less clear to the horse if presented on the ground. Further study is required in order to determine whether the positional advantage would be lost in the discrimination of more complex stimuli or whether the increased attention that appears to be paid to ground level stimuli would still facilitate learning.…”