Teaching is fundamentally cooperative: teachers, be they ants or psychology professors, invest time and energy in helping others to learn. The proximate mechanisms by which they do so may be many and varied, but ultimately all are likely to result in fitness benefits for both pupils and teachers (Thornton & Raihani Reference Thornton and Raihani2008). Here we argue that the debates that have stifled our understanding of the evolution and diversity of teaching mirror those in the broader literature on cooperation and stem largely from confusion between proximate and ultimate explanations.
Historically, most approaches to the study of teaching have conflated proximate mechanisms and ultimate function by specifying a priori psychological mechanisms without which behaviour cannot be classified as teaching, even if it demonstrably helps others to learn. Thus, what Kline terms “culture-based definitions” stipulate a need for formal, Western classroom techniques, mentalistic definitions insist on the use of theory of mind, and the natural pedagogy approach (which Kline places within the functionalist school) specifies that teaching must involve ostensive cueing and metacognition. By imposing restrictive mechanistic pre-requisites, these approaches automatically exclude the possibility of teaching not only among most animals but also among many in human groups including non-Western societies or people with socio-cognitive impairments such as autism. Consequently, they are of limited value for understanding how teaching evolves and is implemented across human and nonhuman societies.
The comparative functionalist framework (Caro & Hauser Reference Caro and Hauser1992; Hoppitt et al. Reference Hoppitt, Brown, Kendal, Rendell, Thornton, Webster and Laland2008; Thornton & Raihani Reference Thornton and Raihani2008), which Kline builds on in her article, provides a solution to this problem by following in Tinbergen's footsteps to consider adaptive value and proximate causation as separate but intertwined questions (see Tinbergen Reference Tinbergen1963). Here, teaching is treated as a functional category of behaviour that serves to promote learning in others and can be underpinned by a variety of different mechanisms (Thornton & Raihani Reference Thornton and Raihani2008). Comparative research has not only demonstrated that taxonomically diverse animals, including invertebrates, birds, and mammals, perform behaviour that functions to help others learn but has also revealed the psychological mechanisms by which this is achieved. Thus, we now know that diverse mechanisms, including reflexive responses to observable cues of pupils' competence (as in meerkats; Thornton & McAuliffe Reference Thornton and McAuliffe2006) and active Pavlovian conditioning (as in pied babblers; Raihani & Ridley Reference Raihani and Ridley2008), may underpin teaching in different species. As Kline highlights, a similar functionalist framework to study human teaching is now essential to uncover the variety of psychological mechanisms employed across cultures and to understand how these evolved.
Like the teaching literature, the study of cooperation has suffered from confusion over proximate and ultimate levels of explanation. For instance, some researchers have treated seemingly cooperative behaviour in nonhuman animals as diagnostic of psychological mechanisms known to be prevalent in humans. In one highly publicised example, laboratory rats learned to open a door to release a trapped conspecific and were duly credited by the authors with empathetic concern (Ben-Ami Bartal et al. Reference Ben-Ami Bartal, Decety and Mason2011). Although the trapped animal clearly benefitted from being liberated, the study provided little direct evidence that its rescuers were motivated by psychological representations of the unfortunate captive's distressed emotional state (Vasconcelos et al. Reference Vasconcelos, Hollis, Nowbahari and Kacelnik2012). Indeed, the assumption that rescue behaviour necessarily involves empathy is akin to assuming that teaching behaviour requires theory of mind. Proximate mechanisms cannot be assumed but rather must be identified through experimentation. For example, a recent follow-up to the rat empathy study found that rescue behaviour could be explained by a motivation for social contact on behalf of the rescuers (Silberberg et al. Reference Silberberg, Allouch, Sandfort, Kearns, Karpel and Slotnick2014). In work on teaching, playback experiments show that adult meerkats deliver age-appropriate hunting lessons to their pups not by reasoning about pups' knowledge states, but by responding reflexively to age-related changes in their begging calls (Thornton & McAuliffe Reference Thornton and McAuliffe2006). Similar experimental scrutiny of human psychological mechanisms is critical if we are to understand not only the role of seemingly computationally complex processes such as theory of mind, but also lower-level cognitive responses and hence the minimal cognitive constraints on the evolution of human cooperation and teaching (Thornton & McAuliffe Reference Thornton and McAuliffe2012).
Just as we cannot infer psychological mechanisms from the existence of seemingly helpful behaviour, we must also be cautious in extrapolating adaptive functions from proximate mechanisms. This issue is particularly prevalent in literature on strong reciprocity, which is often presented as both a proximate and an ultimate explanation for human cooperation (see Scott-Phillips et al. Reference Scott-Phillips, Dickins and West2011). Strong reciprocity is defined as a psychological predisposition to reward cooperators and punish cheats (Fehr & Fischbacher Reference Fehr and Fischbacher2003). People often exhibit such tendencies in anonymous, one-shot laboratory games in which there is no obvious scope for direct reciprocity or reputation-based benefits. Consequently, some theorists argue that this psychological altruism implies evolutionary altruism – that is, that actors derive no fitness benefits from their helpful actions – and that therefore these predispositions must evolve as a result of group-level rather than individual-level benefits (e.g., Bowles & Gintis Reference Bowles and Gintis2004). However, like mentalistic views of teaching, this argument conflates psychological and evolutionary goals. The comparative functionalist framework illustrates the need to analyse these goals separately: ants, meerkats, and babblers may not be driven by the psychological goal of helping others to learn, but their behaviour is nevertheless favoured by selection because it achieves this goal and consequently benefits both teachers and pupils (Thornton & Raihani Reference Thornton and Raihani2008). Conversely, humans' psychologically altruistic preferences may be selected because they ultimately benefit the actor, for instance through increasing perceived attractiveness or prestige (Hardy & Van Vugt Reference Hardy and Van Vugt2006; Sylwester & Roberts Reference Sylwester and Roberts2013; Van Vugt & Iredale Reference Van Vugt and Iredale2012). Thus, just as neural and hormonal mechanisms mediating parents' love for their offspring are ultimately self-serving, psychological altruism may also yield self-serving benefits, even if we are not consciously motivated by them.
Kline neatly illustrates the need for functionalist frameworks that consider the adaptive function and proximate mechanisms of teaching as distinct but complementary questions. If other researchers follow her lead, we will be well on our way to a truly integrative understanding of the proximate and ultimate drivers of teaching and cooperation across species and cultures.
Teaching is fundamentally cooperative: teachers, be they ants or psychology professors, invest time and energy in helping others to learn. The proximate mechanisms by which they do so may be many and varied, but ultimately all are likely to result in fitness benefits for both pupils and teachers (Thornton & Raihani Reference Thornton and Raihani2008). Here we argue that the debates that have stifled our understanding of the evolution and diversity of teaching mirror those in the broader literature on cooperation and stem largely from confusion between proximate and ultimate explanations.
Historically, most approaches to the study of teaching have conflated proximate mechanisms and ultimate function by specifying a priori psychological mechanisms without which behaviour cannot be classified as teaching, even if it demonstrably helps others to learn. Thus, what Kline terms “culture-based definitions” stipulate a need for formal, Western classroom techniques, mentalistic definitions insist on the use of theory of mind, and the natural pedagogy approach (which Kline places within the functionalist school) specifies that teaching must involve ostensive cueing and metacognition. By imposing restrictive mechanistic pre-requisites, these approaches automatically exclude the possibility of teaching not only among most animals but also among many in human groups including non-Western societies or people with socio-cognitive impairments such as autism. Consequently, they are of limited value for understanding how teaching evolves and is implemented across human and nonhuman societies.
The comparative functionalist framework (Caro & Hauser Reference Caro and Hauser1992; Hoppitt et al. Reference Hoppitt, Brown, Kendal, Rendell, Thornton, Webster and Laland2008; Thornton & Raihani Reference Thornton and Raihani2008), which Kline builds on in her article, provides a solution to this problem by following in Tinbergen's footsteps to consider adaptive value and proximate causation as separate but intertwined questions (see Tinbergen Reference Tinbergen1963). Here, teaching is treated as a functional category of behaviour that serves to promote learning in others and can be underpinned by a variety of different mechanisms (Thornton & Raihani Reference Thornton and Raihani2008). Comparative research has not only demonstrated that taxonomically diverse animals, including invertebrates, birds, and mammals, perform behaviour that functions to help others learn but has also revealed the psychological mechanisms by which this is achieved. Thus, we now know that diverse mechanisms, including reflexive responses to observable cues of pupils' competence (as in meerkats; Thornton & McAuliffe Reference Thornton and McAuliffe2006) and active Pavlovian conditioning (as in pied babblers; Raihani & Ridley Reference Raihani and Ridley2008), may underpin teaching in different species. As Kline highlights, a similar functionalist framework to study human teaching is now essential to uncover the variety of psychological mechanisms employed across cultures and to understand how these evolved.
Like the teaching literature, the study of cooperation has suffered from confusion over proximate and ultimate levels of explanation. For instance, some researchers have treated seemingly cooperative behaviour in nonhuman animals as diagnostic of psychological mechanisms known to be prevalent in humans. In one highly publicised example, laboratory rats learned to open a door to release a trapped conspecific and were duly credited by the authors with empathetic concern (Ben-Ami Bartal et al. Reference Ben-Ami Bartal, Decety and Mason2011). Although the trapped animal clearly benefitted from being liberated, the study provided little direct evidence that its rescuers were motivated by psychological representations of the unfortunate captive's distressed emotional state (Vasconcelos et al. Reference Vasconcelos, Hollis, Nowbahari and Kacelnik2012). Indeed, the assumption that rescue behaviour necessarily involves empathy is akin to assuming that teaching behaviour requires theory of mind. Proximate mechanisms cannot be assumed but rather must be identified through experimentation. For example, a recent follow-up to the rat empathy study found that rescue behaviour could be explained by a motivation for social contact on behalf of the rescuers (Silberberg et al. Reference Silberberg, Allouch, Sandfort, Kearns, Karpel and Slotnick2014). In work on teaching, playback experiments show that adult meerkats deliver age-appropriate hunting lessons to their pups not by reasoning about pups' knowledge states, but by responding reflexively to age-related changes in their begging calls (Thornton & McAuliffe Reference Thornton and McAuliffe2006). Similar experimental scrutiny of human psychological mechanisms is critical if we are to understand not only the role of seemingly computationally complex processes such as theory of mind, but also lower-level cognitive responses and hence the minimal cognitive constraints on the evolution of human cooperation and teaching (Thornton & McAuliffe Reference Thornton and McAuliffe2012).
Just as we cannot infer psychological mechanisms from the existence of seemingly helpful behaviour, we must also be cautious in extrapolating adaptive functions from proximate mechanisms. This issue is particularly prevalent in literature on strong reciprocity, which is often presented as both a proximate and an ultimate explanation for human cooperation (see Scott-Phillips et al. Reference Scott-Phillips, Dickins and West2011). Strong reciprocity is defined as a psychological predisposition to reward cooperators and punish cheats (Fehr & Fischbacher Reference Fehr and Fischbacher2003). People often exhibit such tendencies in anonymous, one-shot laboratory games in which there is no obvious scope for direct reciprocity or reputation-based benefits. Consequently, some theorists argue that this psychological altruism implies evolutionary altruism – that is, that actors derive no fitness benefits from their helpful actions – and that therefore these predispositions must evolve as a result of group-level rather than individual-level benefits (e.g., Bowles & Gintis Reference Bowles and Gintis2004). However, like mentalistic views of teaching, this argument conflates psychological and evolutionary goals. The comparative functionalist framework illustrates the need to analyse these goals separately: ants, meerkats, and babblers may not be driven by the psychological goal of helping others to learn, but their behaviour is nevertheless favoured by selection because it achieves this goal and consequently benefits both teachers and pupils (Thornton & Raihani Reference Thornton and Raihani2008). Conversely, humans' psychologically altruistic preferences may be selected because they ultimately benefit the actor, for instance through increasing perceived attractiveness or prestige (Hardy & Van Vugt Reference Hardy and Van Vugt2006; Sylwester & Roberts Reference Sylwester and Roberts2013; Van Vugt & Iredale Reference Van Vugt and Iredale2012). Thus, just as neural and hormonal mechanisms mediating parents' love for their offspring are ultimately self-serving, psychological altruism may also yield self-serving benefits, even if we are not consciously motivated by them.
Kline neatly illustrates the need for functionalist frameworks that consider the adaptive function and proximate mechanisms of teaching as distinct but complementary questions. If other researchers follow her lead, we will be well on our way to a truly integrative understanding of the proximate and ultimate drivers of teaching and cooperation across species and cultures.
ACKNOWLEDGEMENTS
Alex Thornton was supported by a BBSRC David Phillips Fellowship (BB/H021817/1); Nichola Raihani was supported by a Royal Society University Research Fellowship.