The comparative psychology of uncertainty monitoring and metacognition

Abstract

Researchers have begun to explore animals' capacities for uncertainty monitoring and metacognition. This exploration could extend the study of animal self-awareness and establish the relationship of self-awareness to other-awareness. It could sharpen descriptions of metacognition in the human literature and suggest the earliest roots of metacognition in human development. We summarize research on uncertainty monitoring by humans, monkeys, and a dolphin within perceptual and metamemory tasks. We extend phylogenetically the search for metacognitive capacities by considering studies that have tested less cognitively sophisticated species. By using the same uncertainty-monitoring paradigms across species, it should be possible to map the phylogenetic distribution of metacognition and illuminate the emergence of mind. We provide a unifying formal description of animals' performances and examine the optimality of their decisional strategies. Finally, we interpret animals' and humans' nearly identical performances psychologically. Low-level, stimulus-based accounts cannot explain the phenomena. The results suggest granting animals a higher-level decision-making process that involves criterion setting using controlled cognitive processes. This conclusion raises the difficult question of animal consciousness. The results show that animals have functional features of or parallels to human conscious cognition. Remaining questions are whether animals also have the phenomenal features that are the feeling/knowing states of human conscious cognition, and whether the present paradigms can be extended to demonstrate that they do. Thus, the comparative study of metacognition potentially grounds the systematic study of animal consciousness.

Getting It Together: Psychological Unity and Deflationary Accounts of Animal Metacognition

“…Figure 2 shows the results of the dense-sparse test (Smith, Shields, and Washburn 2003) given to a rhesus monkey.…”

Information Seeking in Animals: Metacognition?

“…Experiments reported during the past 15 years have led to the suggestion that metacognition may be found in nonhuman primates (Hampton, 2001;Hampton, Zivin, & Murray, 2004;Kornell et al, 2007;Smith, Shields, Allendoerfer, & Washburn, 1998;Smith, Shields, Schull, & Washburn, 1997;Smith, Shields, & Washburn, 2003) and in a bottlenosed dolphin (Smith, Schull, Strote, McGee, Egnor, & Erb, 1995). These experiments generally require the subject either to make a psychophysical response or to take a memory test.…”

“…Smith et al (2008) argued that associative accounts of uncertainty responding fail to explain why monkeys transferred the uncertainty response accurately to novel stimuli and tasks (Couchman et al, 2010;Kornell et al, 2007;Washburn, Smith, & Shields, 2006) and why monkeys continued to use the uncertainty response when its use was not reinforced until after the end of a block of trials (Couchman et al, 2010;Smith, Beran, Redford, & Washburn, 2006). Further, monkeys do not always make the uncertainty response under stimulus conditions that the associative model suggests they should ) and do make the uncertainty response to difficult discriminations that involve only abstract relational comparisons (Smith et al, 2003).…”

Why the question of animal consciousness might not matter very much

“…Consider some of the evidence of uncertainty monitoring in other species of animal, recently discussed at length by Smith, Shields and Washburn (2003). The authors show how some animals in a state of uncertainty in a forced-choice situation (the pattern is dense/the pattern isn't dense; the tone is a high one/the tone isn't high) will choose adaptively in favor of some third alternative when given the chance (e.g.…”

“…humans will couch their self-explanations in meta-cognitive terms). Since uncertainty behaviors in humans display both higher-order thought and phenomenal consciousness, Smith et al (2003) argue, it is likely that similar states are present in monkeys and dolphins in such circumstances. Uncertainty monitoring can be explained without resort to attributions of metacognitive processes, however.…”