Natural category discrimination in chimpanzees (Pan troglodytes) at three levels of abstraction

Abstract

Two adult chimpanzees were presented with a series of natural category discrimination tasks on a touch screen computer, in which the discriminations varied in degree of abstraction. At the concrete level, discriminations could be made on the basis of single perceptual features, but at the more abstract level, categories were more inclusive, containing exemplars with variant perceptual features. For instance, at the most abstract level, the chimpanzees were required to select images of animals rather than nonanimals, and exemplars within both categories were perceptually diverse. One chimpanzee showed positive transfer at each level of abstraction but required more sessions to reach criterion as the discriminations became more abstract. The other chimpanzee failed to demonstrate consistent significant acquisition of a concept. The results indicate that unlike other apes and black bears, tested previously, chimpanzees found the most abstract discriminations the most difficult to acquire. Analyses of the features of pictures that yielded high or low accuracy revealed no significant differences on several key features, suggesting that the presence of facial features, eyes, or specific coloration did not control responding. In addition, the chimpanzees performed more accurately with photos judged as less typical exemplars of the category by human raters. However, responses to pictures of particular species suggest that chimpanzees may rely on perceptual similarity to familiar exemplars when acquiring experimenter-defined natural categories.Keywords Chimpanzees . Concept formation . Levels of abstraction . Natural categoriesThe formation of natural categories is critical to survival for species faced with challenges such as distinguishing food items from nonpalatable objects and prey from predators. Given the fundamental importance of natural concepts in determining the appropriate response to novel objects, it is expected that most animal species would form natural categories, despite their notoriously fuzzy boundaries (Malt, 1993;McCloskey & Glucksberg, 1978). Natural categories may be inferred from perceptual and functional attributes that are necessary, sufficient, and perhaps prototypical for a member of that category. A deep literature grounded in cognitive, comparative, and developmental psychology has established that categories vary to the extent that they are tied to singular perceptual features or abstracted from patterns of features, functions, and attributes (Fize, Cauchoix, & Fabre-Thorpe, 2011;Gelman & Medin, 1993;Herrnstein, 1990;Mervis & Rosch, 1981;Roberts & Mazmanian, 1988; Rosch, Mervis, Gray, Johnson, & Boyes-Braem, 1976;Vonk & MacDonald, 2002, 2004Vonk & Povinelli, 2012; Zentall, Wasserman, Lazareva, Thompson, & Ratterman, 2008). To the extent that categories vary in inclusiveness, they will vary in the degree of abstraction that is required to identify category membership. For instance, a concrete-level category may include members of only a particular species, whose physical features...

Matching based on biological categories in Orangutans (Pongo abelii) and a Gorilla (Gorilla gorilla gorilla)

“…This is consistent with the prediction made for humans (Keil, 1988; Rosch et al, 1976), and possibly with findings from chimpanzees (Tanaka, 2001, although see Vonk et al, 2013), but not from squirrel monkeys, pigeons (Roberts & Mazmanian, 1988) or gorillas (Vonk & MacDonald, 2002). The gorilla subject in the present experiments seemed to learn the concrete level discrimination more rapidly despite the fact that she was simultaneously learning the DMTS task for the first time.…”

“…At the most abstract level, such as ‘animal’, superordinate category exemplars share even fewer perceptual features within a category but also share even fewer features with members of other abstract categories. The category ‘mammal’ may be considered superordinate to the basic level ‘dog’, but, in keeping with previously published studies (Roberts & Mazmanian, 1988; Vonk & MacDonald, 2002; Vonk & MacDonald, 2004; Vonk, Jett & Mosteller, 2012; Vonk et al, 2013), ‘mammals’ will be considered intermediate with ‘animal’ being considered the most abstract level category.…”

“…The terms concrete and subordinate, intermediate and basic, and abstract and superordinate have been used interchangeably. The former terms will be used exclusively throughout the remainder of this manuscript in keeping with previous research (Vonk & MacDonald, 2002; Vonk & MacDonald, 2004; Vonk, Jett & Mosteller, 2012; Vonk et al, 2013). …”

“…Thus both species of Great Ape demonstrated the ability to learn concepts at each level of abstraction, but the orangutans appeared to learn the basic or intermediate level discrimination the most readily while the gorilla learned it with the most difficulty. More recent follow-ups have shown chimpanzees to have the most difficulty with the most abstract discriminations (Vonk et al, 2013) in that they required more sessions to reach criterion as discriminations became more abstract, and were less likely to show significant transfer. In contrast, black bears showed significant transfer at all levels of concept discrimination, even when trained on the most abstract problems first (Vonk, Jett & Mosteller, 2012).…”

“Bear-ly” learning: Limits of abstraction in black bear cognition

“…In addition, the bears showed some degree of transfer to the control images depicting large and small foods (Vonk & Johnson-Ulrich, 2014). Thus, we have demonstrated rapid learning equivalent to that of great apes also tested in similar tasks (e.g., quantity discrimination, ambiguous cue paradigm, McGuire et al, 2017; natural category discrimination, Vonk, Jett, Mosteller, & Galvan, 2013;Vonk & Johnson-Ulrich, 2014;Vonk & MacDonald, 2002;2004), but we have not shown that bears can learn truly abstract constructs that are not tied to observable perceptual features, such as demonstrating relational learning or causal reasoning. Here, we present data from attempts to train bears to perform a matching-to-sample (MTS) task, and a serial list memory task.…”