Richerson et al. argue that widespread cooperation between human non-relatives and Darwin's observation that individuals “sacrifice themselves for the common good” (Darwin Reference Darwin1874, pp. 178–79; see epigram at start of target article) are best explained by a mechanism of cultural group selection (CGS). In the following, I elaborate on how ontogenetic mechanisms may also have contributed to our organization in groups and how both cooperation and the sub-phenomenon of “self-sacrifice” may be understood as parts of temporally extended behavioral patterns exhibited either by a group or by individuals.
Behavior of groups, as well as of individuals, takes time. Since time is limited, possible activities compete. Natural selection chooses those activities that correlate either with Phylogenetically Important Events (PIE) that ultimately enhance fitness or ontogenetic proxies of these PIEs (Baum Reference Baum2012). Temporally extended activities have parts that consist of smaller-scale behavioral patterns. Thus, natural selection acts on behavioral patterns of varying complexity. Those patterns might be extended in time, as well as in social space; that is, they can be exhibited by an individual or by a group. More local behavioral patterns have other consequences than more extended patterns embracing those local patterns. To estimate an activity's effect on an individual's or a group's fitness, we need to weigh the long-term effects of the sum of local acts against the long-term effect of the behavioral pattern as a whole. The key to understanding cooperation or self-sacrifice lies in the realization that the effect of the pattern as a whole does not equal the added effects of its parts, let alone does a particular part affect fitness in the same direction as the whole pattern. Focusing on the consequences of parts (i.e., single local acts) can lead us up the garden path when explaining cooperation.
The following two examples illustrate how the effect of an individual's local act can differ from the effect of an individual's or a group's extended behavioral pattern: An individual's local act, such as an alcoholic's acceptance of a drink at a party, is promptly followed by relaxed and friendly social interactions, that is, proxies of fitness-enhancing PIEs. We call the acceptance of the drink “impulsive.” The alcoholic's refusal of the drink is promptly followed by proxies for fitness-decreasing PIEs such as social disapproval or simply the absence of relaxed social interactions. If an alcoholic refuses a drink, we say that he or she is showing “self-control.” Despite the short-term consequences, an extended pattern of self-controlled behavior leads to enhanced fitness. Even if every single refusal of a drink is punished by proxies for fitness-decreasing PIEs, the extended pattern of abstinence from alcohol will correlate with being healthier, having more well-functioning relationships, and a better economy – self-controlled behavior (as an extended pattern) enhances fitness, assessed in the long run (Baum Reference Baum2015; Rachlin Reference Rachlin2004).
A parallel mechanism selects the behavior of a group. Each particular engagement in a cooperative activity, such as recycling, might be locally costly for the individual. If, however, enough group members cooperate, those activities lead to fitness-enhancing PIEs or their proxies. The effect of the more extended behavioral pattern is in the interest of the individual as well as of the group. Of course, those individuals who in fact recycled are not necessarily those whose fitness actually benefits. In most instances, however, the whole group benefits, which means that those who contributed at their own immediate cost often do indeed benefit in the long run. An individual's acts that lead – if they are not punished – to PIEs advantageous for his or her own fitness but disadvantageous for the fitness of other group members, are labeled “selfish,” whereas acts that are in the short run costly for the individual but beneficial for the group – including the individual – are called “cooperative.” How many instances of “self-sacrifice” are advantageous for the individual's fitness can be understood when analyzing them as parts of rates of cooperative behavior, which correlate with rates of consequences, such as reciprocity.
Even if altruistic behavior is not central to their argument, Richerson et al. “are inclined to think that evolution has built an element of […] true altruism into our social psychology” (sect. 6.2, para. 2). If “true altruism” is defined by the absence of fitness-enhancing consequences on the actor's behavior, here, Richerson et al.'s account is not reconcilable with the view of altruistic behavior as a temporally extended behavioral pattern that may be learned over an individual's lifetime (Rachlin Reference Rachlin2002). An individual act of cooperation, as well as of altruism, each of which might even be followed by fitness-decreasing PIEs in the short run, can nevertheless be fitness-beneficial when repeated over time. Just as biological selection has acted on groups of organs constituting organisms, extended behavioral patterns (i.e., groups of acts) can be selected together.
How do the abstract patterns of behavior resist being broken up by selfish behavior?
Richerson et al. posit that group-variable culturally transmitted social norms with group-level functionality are evidence of CGS. Teaching and following norms and rules are practices that function to overcome selfish and impulsive behavior. We have rules such as “brush your teeth twice a day” to ensure that our offspring act in a self-controlled manner and to maximize their fitness in the long run, and we have rules such as “help others” to ensure the coherence of the group.
In different environmental situations, natural selection occurring within groups and between groups may have different relative strengths. Even if cooperation, defined as “working together toward a shared aim” (Wehmeier and Hornby Reference Wehmeier and Hornby2000), was selected because it commonly enhanced the individual's long-term fitness, it neither follows that cooperation enhances all cooperators' fitness, nor that all acts that we classify as parts of cooperative behavior do so. A truly self-sacrificing act can come about because it is generally advantageous to cooperate, and making decisions on a case-by-case basis is costly. Just as a single drink in a pattern of abstinence does not significantly change the correlation between that pattern and PIEs (e.g., good health), a single fitness-reducing act of an individual will not significantly change the correlation of cooperative behavior and its fitness-enhancing effect resulting from the group's collaboration.
Richerson et al. offer a coherent framework which is supported, complemented, and potentially slightly modified by the proposed considerations of behavioral patterns and their consequences in different time frames.
Richerson et al. argue that widespread cooperation between human non-relatives and Darwin's observation that individuals “sacrifice themselves for the common good” (Darwin Reference Darwin1874, pp. 178–79; see epigram at start of target article) are best explained by a mechanism of cultural group selection (CGS). In the following, I elaborate on how ontogenetic mechanisms may also have contributed to our organization in groups and how both cooperation and the sub-phenomenon of “self-sacrifice” may be understood as parts of temporally extended behavioral patterns exhibited either by a group or by individuals.
Behavior of groups, as well as of individuals, takes time. Since time is limited, possible activities compete. Natural selection chooses those activities that correlate either with Phylogenetically Important Events (PIE) that ultimately enhance fitness or ontogenetic proxies of these PIEs (Baum Reference Baum2012). Temporally extended activities have parts that consist of smaller-scale behavioral patterns. Thus, natural selection acts on behavioral patterns of varying complexity. Those patterns might be extended in time, as well as in social space; that is, they can be exhibited by an individual or by a group. More local behavioral patterns have other consequences than more extended patterns embracing those local patterns. To estimate an activity's effect on an individual's or a group's fitness, we need to weigh the long-term effects of the sum of local acts against the long-term effect of the behavioral pattern as a whole. The key to understanding cooperation or self-sacrifice lies in the realization that the effect of the pattern as a whole does not equal the added effects of its parts, let alone does a particular part affect fitness in the same direction as the whole pattern. Focusing on the consequences of parts (i.e., single local acts) can lead us up the garden path when explaining cooperation.
The following two examples illustrate how the effect of an individual's local act can differ from the effect of an individual's or a group's extended behavioral pattern: An individual's local act, such as an alcoholic's acceptance of a drink at a party, is promptly followed by relaxed and friendly social interactions, that is, proxies of fitness-enhancing PIEs. We call the acceptance of the drink “impulsive.” The alcoholic's refusal of the drink is promptly followed by proxies for fitness-decreasing PIEs such as social disapproval or simply the absence of relaxed social interactions. If an alcoholic refuses a drink, we say that he or she is showing “self-control.” Despite the short-term consequences, an extended pattern of self-controlled behavior leads to enhanced fitness. Even if every single refusal of a drink is punished by proxies for fitness-decreasing PIEs, the extended pattern of abstinence from alcohol will correlate with being healthier, having more well-functioning relationships, and a better economy – self-controlled behavior (as an extended pattern) enhances fitness, assessed in the long run (Baum Reference Baum2015; Rachlin Reference Rachlin2004).
A parallel mechanism selects the behavior of a group. Each particular engagement in a cooperative activity, such as recycling, might be locally costly for the individual. If, however, enough group members cooperate, those activities lead to fitness-enhancing PIEs or their proxies. The effect of the more extended behavioral pattern is in the interest of the individual as well as of the group. Of course, those individuals who in fact recycled are not necessarily those whose fitness actually benefits. In most instances, however, the whole group benefits, which means that those who contributed at their own immediate cost often do indeed benefit in the long run. An individual's acts that lead – if they are not punished – to PIEs advantageous for his or her own fitness but disadvantageous for the fitness of other group members, are labeled “selfish,” whereas acts that are in the short run costly for the individual but beneficial for the group – including the individual – are called “cooperative.” How many instances of “self-sacrifice” are advantageous for the individual's fitness can be understood when analyzing them as parts of rates of cooperative behavior, which correlate with rates of consequences, such as reciprocity.
Even if altruistic behavior is not central to their argument, Richerson et al. “are inclined to think that evolution has built an element of […] true altruism into our social psychology” (sect. 6.2, para. 2). If “true altruism” is defined by the absence of fitness-enhancing consequences on the actor's behavior, here, Richerson et al.'s account is not reconcilable with the view of altruistic behavior as a temporally extended behavioral pattern that may be learned over an individual's lifetime (Rachlin Reference Rachlin2002). An individual act of cooperation, as well as of altruism, each of which might even be followed by fitness-decreasing PIEs in the short run, can nevertheless be fitness-beneficial when repeated over time. Just as biological selection has acted on groups of organs constituting organisms, extended behavioral patterns (i.e., groups of acts) can be selected together.
How do the abstract patterns of behavior resist being broken up by selfish behavior?
Richerson et al. posit that group-variable culturally transmitted social norms with group-level functionality are evidence of CGS. Teaching and following norms and rules are practices that function to overcome selfish and impulsive behavior. We have rules such as “brush your teeth twice a day” to ensure that our offspring act in a self-controlled manner and to maximize their fitness in the long run, and we have rules such as “help others” to ensure the coherence of the group.
In different environmental situations, natural selection occurring within groups and between groups may have different relative strengths. Even if cooperation, defined as “working together toward a shared aim” (Wehmeier and Hornby Reference Wehmeier and Hornby2000), was selected because it commonly enhanced the individual's long-term fitness, it neither follows that cooperation enhances all cooperators' fitness, nor that all acts that we classify as parts of cooperative behavior do so. A truly self-sacrificing act can come about because it is generally advantageous to cooperate, and making decisions on a case-by-case basis is costly. Just as a single drink in a pattern of abstinence does not significantly change the correlation between that pattern and PIEs (e.g., good health), a single fitness-reducing act of an individual will not significantly change the correlation of cooperative behavior and its fitness-enhancing effect resulting from the group's collaboration.
Richerson et al. offer a coherent framework which is supported, complemented, and potentially slightly modified by the proposed considerations of behavioral patterns and their consequences in different time frames.
ACKNOWLEDGMENT
I wish to thank William M. Baum, Per Holth, Kalliu Couto, Florian Lange, and Johan Viklund for helpful feedback on earlier drafts of this commentary.