Nonhuman Minds: Animal, Artificial or Other Minds
Cognitio 2011
Young researchers conference in cognitive science
Montréal, July 3rd, 4th and 5th 2011
Theory of Mind, Chimpanzees, and the Lafayette-Leipzig Dialectic
Kyle Ferguson
Abstract:
For over three decades, comparative and developmental psychologists have investigated an ability called Theory of Mind (‘ToM’, hereafter). ToM is a social-cognitive phenomenon in which a subject attributes mental states to another and appeals to those attributions when predicting or explaining behavior. ToM researchers don’t debate whether their subjects are cognitive creatures. Instead, they debate over the particular cognitive abilities their subjects possess.
ToM research generally takes place in two arenas. First, researchers in the developmental arena investigate the kinds of mental states that young children know about and aim to discover the developmental sequence of ToM understanding. For years, researchers largely agreed that children do not exhibit ToM understanding until around the fourth year of life, at which time children succeed on verbal false-belief tasks. More recent studies, however, purportedly show that 12-month-old infants exhibit ToM understanding. Second, in the comparative arena, chimpanzees have attracted the most scholarly attention. The phrase ‘theory of mind’ originates in the Premack and Woodruff’s (1978) landmark study on chimpanzees’ understanding of others’ intentions or goals. Researchers in either arena face significant obstacles, especially when dealing with nonverbal subjects such as human infants or chimpanzees. How, after all, could researchers actually detect ToM understanding in nonverbal subjects?
One of the most important debates in ToM research is between two of the world’s leading comparative cognition laboratories - one lab in Lafayette, Louisiana, led by Daniel Povinelli, the other in Leipzig, Germany, led by Michael Tomasello – and it rests on a methodological worry of the kind mentioned above. The Leipzig group claims to have shown in various studies that chimpanzees possess parts of ToM: chimpanzees appear to attribute perceptions, intentions and knowledge to others, but not beliefs. The Lafayette group has responded that the Leipzig studies cannot rule out rival hypotheses, according to which chimpanzees merely reason about others’ observable behavior of others but not about their underlying mental states.
As Povinelli sees it, there are two possible cognitive systems: call them ‘Sb’ and ‘Sb+ms’; the former system lacks psychological concepts and is restricted to reasoning about others’ behavior, while the latter additionally reasons about others’ mental states. Povinell’s methodological challenge, then, is to design an experiment that is sensitive to the difference between these two systems. The challenge, to be sure, rests on an intuitive distinction between reasoning about mental states and reasoning about behavior, and it gains more apparent strength when we realize that predictions of another’s behavior using Sb would rarely, if ever, diverge from those generated by Sb+ms. Until such an experiment, Lafayette reasons, Leipzig’s claims to positive evidence are unjustified.
In this paper, my main objectives are (i) to critically examine this methodological challenge and its role in the Lafayette-Leipzig dialectic, and (ii) to raise some doubts about the validity of the challenge by undermining the distinction upon which it rests. First, I argue that without specifying how the posited difference between Sb and Sb+ms is expressed in observable behavior, Povinelli’s challenge is empty and the debate at hand is empirically irresolvable. The Lafayette group takes itself to have operationalized the difference in at least one study, but I argue that this design reveals a problematic assumption about the very nature of psychological concepts; viz. the assumption that one first learns about the mind from one’s own case not from observing others. Finally, I argue, appealing to an insight from David Lewis (1972), which provides an independent reason to think that the difference between Sb and Sb+ms is merely notational. If this is right, a fundamental assumption in ToM research should be reevaluated.
ToM research generally takes place in two arenas. First, researchers in the developmental arena investigate the kinds of mental states that young children know about and aim to discover the developmental sequence of ToM understanding. For years, researchers largely agreed that children do not exhibit ToM understanding until around the fourth year of life, at which time children succeed on verbal false-belief tasks. More recent studies, however, purportedly show that 12-month-old infants exhibit ToM understanding. Second, in the comparative arena, chimpanzees have attracted the most scholarly attention. The phrase ‘theory of mind’ originates in the Premack and Woodruff’s (1978) landmark study on chimpanzees’ understanding of others’ intentions or goals. Researchers in either arena face significant obstacles, especially when dealing with nonverbal subjects such as human infants or chimpanzees. How, after all, could researchers actually detect ToM understanding in nonverbal subjects?
One of the most important debates in ToM research is between two of the world’s leading comparative cognition laboratories - one lab in Lafayette, Louisiana, led by Daniel Povinelli, the other in Leipzig, Germany, led by Michael Tomasello – and it rests on a methodological worry of the kind mentioned above. The Leipzig group claims to have shown in various studies that chimpanzees possess parts of ToM: chimpanzees appear to attribute perceptions, intentions and knowledge to others, but not beliefs. The Lafayette group has responded that the Leipzig studies cannot rule out rival hypotheses, according to which chimpanzees merely reason about others’ observable behavior of others but not about their underlying mental states.
As Povinelli sees it, there are two possible cognitive systems: call them ‘Sb’ and ‘Sb+ms’; the former system lacks psychological concepts and is restricted to reasoning about others’ behavior, while the latter additionally reasons about others’ mental states. Povinell’s methodological challenge, then, is to design an experiment that is sensitive to the difference between these two systems. The challenge, to be sure, rests on an intuitive distinction between reasoning about mental states and reasoning about behavior, and it gains more apparent strength when we realize that predictions of another’s behavior using Sb would rarely, if ever, diverge from those generated by Sb+ms. Until such an experiment, Lafayette reasons, Leipzig’s claims to positive evidence are unjustified.
In this paper, my main objectives are (i) to critically examine this methodological challenge and its role in the Lafayette-Leipzig dialectic, and (ii) to raise some doubts about the validity of the challenge by undermining the distinction upon which it rests. First, I argue that without specifying how the posited difference between Sb and Sb+ms is expressed in observable behavior, Povinelli’s challenge is empty and the debate at hand is empirically irresolvable. The Lafayette group takes itself to have operationalized the difference in at least one study, but I argue that this design reveals a problematic assumption about the very nature of psychological concepts; viz. the assumption that one first learns about the mind from one’s own case not from observing others. Finally, I argue, appealing to an insight from David Lewis (1972), which provides an independent reason to think that the difference between Sb and Sb+ms is merely notational. If this is right, a fundamental assumption in ToM research should be reevaluated.