More than a Fluke: Lessons Learned from a Failure to Replicate the False Belief Task in Dolphins

Abstract

Critical to advanced social intelligence is the ability to take into consideration the thoughts and feelings of others, a skill referred to as Theory of Mind (ToM) or mindreading. In this article, we present a critical review of the comparative methodology and utility of the nonverbal FBT along with a description of an attempted FBT replication conducted with a bottlenose dolphin prior to the implementation of the more successful approaches used currently. Attempting to replicate Tschudin’s (2001, 2006) methodology with dolphins highlighted several flaws that may explain the failures of socially complex mammals to display competency: (1) reliance on a containment invisible displacement procedure that is difficult for non-human animals and especially dolphins to follow, (2) a complex procedure which demands extensive training time, (3) a long trial duration with several moving parts which taxes the animal’s memory and attention, and (4) a restricted number of two-choice FBT test trials, which limits statistical power given the small pool of trained animals. Although recent research paradigms for primates have corrected for some of these flaws, it is critical that comparative psychologists address these limitations for other species or taxa to be tested validly. Future research in ToM understanding through a false belief approach should move toward more ecologically valid designs and appropriate implicit measures that facilitate comparative approaches that can be replicated.

Fish out of water: Insights from a case study of a highly social animal that failed the mirror self-recognition test

“…Furthermore, because an important step in the MSR test is cessation of prior social response to the animal's own reflection, it follows then that it is not a barrier that the reflection lacks some properties of a conspecific (e.g., scent in the case of wolves) and perhaps even helps the animal to advance more quickly through that stage in MSR precisely because some conspecific cues are lacking. Thus, we do not concur that a vision-based self-cognizance test is preemptively inappropriate for canids, but we conclude rather that a test that relies on responding to a visually detected, ecologically irrelevant mark is likely inappropriate (Hill et al, 2018;Vonk, 2019) for this and similar species.…”

“…Given the paucity of information on wolves and MSR, we aimed to rectify the absence in the scientific record of specific details regarding gray wolf performance at various stages of the MSR test. In revisiting the wolf MSR test, we also anticipated identifying features of the test that may need to be modified to better assess this in wolves and other similar, highly socially complex predators (Hill et al, 2018;Vonk, 2019). Understanding the specific stage(s) of the MSR test (as delineated by Plotnik et al, 2006) that wolves failed (if, indeed, they failed, because differences in intraspecific variation, rearing conditions, etc., in our study relative to Bekoff's study [Bekoff 2014] could contribute to them passing) is important for focusing future research.…”

“…A suite of ecologically relevant self-cognizance tests that consider life history and sensory characteristics and that can also be scaled along a potential self-recognition spectrum are needed (Hill et al, 2018). Tests regarding olfactory and auditory self-recognition have been proposed as potential MSR analogs (Bekoff, 2001(Bekoff, , 2014Cazolla Gatti, 2016;De Veer & Van Den Bos, 1999;Horowitz, 2017).…”