We admire Kline's attempt to illuminate the evolution of teaching via a taxonomy of different varieties, and by considering the adaptive pressures and costs that might lead to their emergence. At the same time, we doubt that Kline's theoretical distinctions are the best formulations.
Kline defines “stimulus enhancement” as occurring when “the teacher stimulates the pupil's interest in a stimulus or location” (target article, sect. 3.1, para. 4). In thereby characterising it as including cases in which a teacher intentionally draws attention to something, Kline departs from standard usage of this term (e.g., Whiten & Ham Reference Whiten and Ham1992) in comparative psychology, in which one agent's activities make salient to another some valuable information. Importantly, on this usage, enhancement can be provided even when an agent is oblivious to the presence of an onlooker – and so is cognitively undemanding. Since Kline includes as examples of stimulus enhancement cases of pointing that are typically thought to be cognitively difficult (Clark Reference Clark1996; Moore Reference Moore2013a; Tomasello Reference Tomasello2008), her taxonomy glosses over cognitive issues that have been considered foundational in the evolution of human cognition. While Kline motivates her functional approach by stating that behaviour (and not cognition) is the target of natural selection, a taxonomy that lumps together behaviours supported by different cognitive abilities and appearing in only distantly related clades is not intuitively a useful tool for understanding evolution. It may lead researchers both to over-estimate the relatedness of different behaviours on account of functional similarities, and to overlook the similarity of cognitively related behaviours performed with different functions.
It is also not clear to us that Kline's terminological distinctions are illuminating. For example, while she describes the flossing of teeth by long-tailed macaques (Masataka et al. Reference Masataka, Koda, Urasopon and Watanabe2009) as a form of stimulus enhancement, the same behaviour is also consistent with her criteria for direct active teaching – since it could well be characterised as a “non-verbal demonstration, punctuated with exaggerated movements, by an expert ... to a novice” (target article, sect. 3.1.5, para. 1). Indeed, we often engage in direct teaching by drawing others' attention to important features of objects – which suggests that Kline's categories are also not mutually exclusive. It is also hard to see why the cases of informative pointing that Kline counts as stimulus enhancement are not cases of active (albeit pre-verbal) teaching; and why the Warao father's adjustment of his son's wrist is a case of direct active teaching, and not evaluative feedback.
The confusions caused by these overlapping categories are unlikely to facilitate identification of cases of teaching in the animal kingdom. Moreover, they undermine our confidence that this new theoretical framework could be used to generate new scenarios for testing for the presence of teaching. Consequently, although Kline's categories are thought-provoking, it is not clear that they improve on the categories of social learning already described by others (e.g., Whiten & Ham Reference Whiten and Ham1992).
Despite this skepticism, we do not think that Kline has over-estimated cases of active teaching – at least among chimpanzees. Since chimpanzees are among our nearest living relatives, their teaching activities are of great interest for understanding the evolution of our own. Furthermore, we agree with Kline that intentional and “theory of mind” based teaching approaches sometimes overstate the social cognition that active teaching requires (Moore Reference Moore2013b), and so agree that “the constraints of cognition ... do not seem sufficient to explain why direct active teaching appears to be limited to humans” (target article, sect. 7, para. 1). But then why is more active teaching not found in chimpanzees?
It seems unlikely that researchers have simply been looking in the wrong place, because several groups (Dean et al. Reference Dean, Kendal, Schapiro, Thierry and Laland2012; Lonsdorf Reference Lonsdorf2006; Matsuzawa et al. Reference Matsuzawa, Biro, Humle, Inoue-Nakamura, Tonooka, Yamakoshi and Matsuzawa2001) have tried and failed to substantiate earlier reports (Boesch Reference Boesch1991). Kline's emphasis on adaptive value may hold out an answer here.
Boesch (Reference Boesch1991; Reference Boesch2012) has argued that chimpanzee mothers at Taï teach their children how to crack panda nuts. Because the Panda oleasa is particularly hard and difficult to crack, juvenile chimpanzees don't typically succeed until they are 8 years old. Because the chimpanzee interbirth interval is 5 years, Boesch argues that the demands of having two dependent offspring may push mothers to accelerate their offspring's learning. We find this explanation unlikely. Although the panda nut may be highly valued, it constitutes neither a large nor an ineliminable part of the Taï chimpanzee diet (Boesch & Boesch-Achermann [Reference Boesch and Boesch-Achermann2000, p. 210] themselves describe Panda nut consumption as “rare” and “irregular”). Therefore, there is likely to be little adaptive pressure for teaching this skill. Given the scant evidence of teaching in chimpanzees, and the failure of others to find further evidence supporting Boesch's reports, it seems advisable to doubt that it is really happening. Why would this be?
One answer favoured by Kline and others (e.g., Gergely & Csibra Reference Gergely and Csibra2005) is that behaviours that are both complex and difficult to learn through observation should lead to pressures for the emergence of teaching. Since naive captive individuals have already proven able to reinvent various wild “cultures” without social learning (Allritz et al. Reference Allritz, Tennie and Call2013; Huffman & Hirata Reference Huffman and Hirata2004; Menzel et al. Reference Menzel, Fowler, Tennie and Call2013), such opaque behaviours may not exist in chimpanzee culture. Therefore, non-teaching learning mechanisms may suffice for the propagation of contemporary chimpanzee technologies, including different forms of observational learning, individual learning, and inherited cognitive skills (Moore Reference Moore2013b; Tennie et al. Reference Tennie and Tomasello2009; Reference Tennie, Call and Tomasello2012). This may be true even for the most complex multi-tool sets (e.g., Boesch et al. Reference Boesch, Head and Robbins2009; Sanz & Morgan Reference Sanz and Morgan2007).
We suspect that chimpanzees have simply faced little adaptive pressure for tools and tool-sets more complex than those that they already possess. Since they were never forced to leave their ecological niches, simpler forms of learning and social learning always sufficed for them to acquire whatever tools, tool-sets and communicative devices they needed. This would explain the lack of pressure for active teaching, not to mention the comparative absence in chimpanzees of high-fidelity learning mechanisms such as imitation. Given her closing comments about the adaptive value of teaching, we think that Kline would agree with this conclusion – but it is not clear why we needed her theoretical framework to get there.
We admire Kline's attempt to illuminate the evolution of teaching via a taxonomy of different varieties, and by considering the adaptive pressures and costs that might lead to their emergence. At the same time, we doubt that Kline's theoretical distinctions are the best formulations.
Kline defines “stimulus enhancement” as occurring when “the teacher stimulates the pupil's interest in a stimulus or location” (target article, sect. 3.1, para. 4). In thereby characterising it as including cases in which a teacher intentionally draws attention to something, Kline departs from standard usage of this term (e.g., Whiten & Ham Reference Whiten and Ham1992) in comparative psychology, in which one agent's activities make salient to another some valuable information. Importantly, on this usage, enhancement can be provided even when an agent is oblivious to the presence of an onlooker – and so is cognitively undemanding. Since Kline includes as examples of stimulus enhancement cases of pointing that are typically thought to be cognitively difficult (Clark Reference Clark1996; Moore Reference Moore2013a; Tomasello Reference Tomasello2008), her taxonomy glosses over cognitive issues that have been considered foundational in the evolution of human cognition. While Kline motivates her functional approach by stating that behaviour (and not cognition) is the target of natural selection, a taxonomy that lumps together behaviours supported by different cognitive abilities and appearing in only distantly related clades is not intuitively a useful tool for understanding evolution. It may lead researchers both to over-estimate the relatedness of different behaviours on account of functional similarities, and to overlook the similarity of cognitively related behaviours performed with different functions.
It is also not clear to us that Kline's terminological distinctions are illuminating. For example, while she describes the flossing of teeth by long-tailed macaques (Masataka et al. Reference Masataka, Koda, Urasopon and Watanabe2009) as a form of stimulus enhancement, the same behaviour is also consistent with her criteria for direct active teaching – since it could well be characterised as a “non-verbal demonstration, punctuated with exaggerated movements, by an expert ... to a novice” (target article, sect. 3.1.5, para. 1). Indeed, we often engage in direct teaching by drawing others' attention to important features of objects – which suggests that Kline's categories are also not mutually exclusive. It is also hard to see why the cases of informative pointing that Kline counts as stimulus enhancement are not cases of active (albeit pre-verbal) teaching; and why the Warao father's adjustment of his son's wrist is a case of direct active teaching, and not evaluative feedback.
The confusions caused by these overlapping categories are unlikely to facilitate identification of cases of teaching in the animal kingdom. Moreover, they undermine our confidence that this new theoretical framework could be used to generate new scenarios for testing for the presence of teaching. Consequently, although Kline's categories are thought-provoking, it is not clear that they improve on the categories of social learning already described by others (e.g., Whiten & Ham Reference Whiten and Ham1992).
Despite this skepticism, we do not think that Kline has over-estimated cases of active teaching – at least among chimpanzees. Since chimpanzees are among our nearest living relatives, their teaching activities are of great interest for understanding the evolution of our own. Furthermore, we agree with Kline that intentional and “theory of mind” based teaching approaches sometimes overstate the social cognition that active teaching requires (Moore Reference Moore2013b), and so agree that “the constraints of cognition ... do not seem sufficient to explain why direct active teaching appears to be limited to humans” (target article, sect. 7, para. 1). But then why is more active teaching not found in chimpanzees?
It seems unlikely that researchers have simply been looking in the wrong place, because several groups (Dean et al. Reference Dean, Kendal, Schapiro, Thierry and Laland2012; Lonsdorf Reference Lonsdorf2006; Matsuzawa et al. Reference Matsuzawa, Biro, Humle, Inoue-Nakamura, Tonooka, Yamakoshi and Matsuzawa2001) have tried and failed to substantiate earlier reports (Boesch Reference Boesch1991). Kline's emphasis on adaptive value may hold out an answer here.
Boesch (Reference Boesch1991; Reference Boesch2012) has argued that chimpanzee mothers at Taï teach their children how to crack panda nuts. Because the Panda oleasa is particularly hard and difficult to crack, juvenile chimpanzees don't typically succeed until they are 8 years old. Because the chimpanzee interbirth interval is 5 years, Boesch argues that the demands of having two dependent offspring may push mothers to accelerate their offspring's learning. We find this explanation unlikely. Although the panda nut may be highly valued, it constitutes neither a large nor an ineliminable part of the Taï chimpanzee diet (Boesch & Boesch-Achermann [Reference Boesch and Boesch-Achermann2000, p. 210] themselves describe Panda nut consumption as “rare” and “irregular”). Therefore, there is likely to be little adaptive pressure for teaching this skill. Given the scant evidence of teaching in chimpanzees, and the failure of others to find further evidence supporting Boesch's reports, it seems advisable to doubt that it is really happening. Why would this be?
One answer favoured by Kline and others (e.g., Gergely & Csibra Reference Gergely and Csibra2005) is that behaviours that are both complex and difficult to learn through observation should lead to pressures for the emergence of teaching. Since naive captive individuals have already proven able to reinvent various wild “cultures” without social learning (Allritz et al. Reference Allritz, Tennie and Call2013; Huffman & Hirata Reference Huffman and Hirata2004; Menzel et al. Reference Menzel, Fowler, Tennie and Call2013), such opaque behaviours may not exist in chimpanzee culture. Therefore, non-teaching learning mechanisms may suffice for the propagation of contemporary chimpanzee technologies, including different forms of observational learning, individual learning, and inherited cognitive skills (Moore Reference Moore2013b; Tennie et al. Reference Tennie and Tomasello2009; Reference Tennie, Call and Tomasello2012). This may be true even for the most complex multi-tool sets (e.g., Boesch et al. Reference Boesch, Head and Robbins2009; Sanz & Morgan Reference Sanz and Morgan2007).
We suspect that chimpanzees have simply faced little adaptive pressure for tools and tool-sets more complex than those that they already possess. Since they were never forced to leave their ecological niches, simpler forms of learning and social learning always sufficed for them to acquire whatever tools, tool-sets and communicative devices they needed. This would explain the lack of pressure for active teaching, not to mention the comparative absence in chimpanzees of high-fidelity learning mechanisms such as imitation. Given her closing comments about the adaptive value of teaching, we think that Kline would agree with this conclusion – but it is not clear why we needed her theoretical framework to get there.