Ernst Fehr on Human Altruism. An Interdisciplinary Debate
2005 (27) Issue 1
In the foregoing decade, two related developments in the behavioural sciences have drawn the attention of social scientists, particularly economists. The first is the use of laboratory experiments in the investigation of human behaviour. Although the use of such experiments has a longer history, only in the last decade has ’experimental economics’ become a sub-discipline of economics with which economists of just about all colours are familiar; indeed, experimental results regularly feed into, and influence, other areas of economics previously untouched thereby. The second development is inseparably related to the first, for it stems from results of experimental research, particularly that conducted by Ernst Fehr and a host of collaborators (to mention a few: Urs Fischbacher, Simon G _achter, Herbert Gintis, Joseph Henrich). These results and the theoretical developments they support are significant because they are at odds with the expectations of those led by some versions of orthodox economic theory. It is this development in the human sciences, and Fehr’s work above all, to which the current issue is dedicated.
Towards the end of 2004, the editors of this issue invited a number of internationally renowned scholars to comment on Fehr’s research into human altruism. We are grateful that many of those asked took up the offer and we present their commentaries in the second part of this volume. In light of the panoptic nature of Fehr’s work, we were intent on soliciting contributions from a wide variety of perspectives. Fehr’s research has implications for such an array of academic disciplines (including economics, anthropology, biology and neurological science) that a volume which seeks to do justice to its sub ject matter is compelled to draw on disparate sources. We are grateful, too, to Ernst Fehr for the encouragement and advice he gave us in executing this pro ject.
The first section of this issue contains two papers which introduce the theme. In the first contribution, Human Altruism â€“ Proximate Patterns and Evolutionary Origins, Ernst Fehr and Urs Fischbacher offer a detailed overview of the research programme to the development of which they have been instrumental. They present a plethora of results â€“ particularly from ultimatum games and public goods experiments â€“ and introduce the novel conceptual apparatus they use to depict the behaviour of their laboratory sub jects. Of particular note are the concepts: strong reciprocity, altruistic punishment and altruistic rewarding which many of the ensuing commentaries discuss. When it comes to explaining the behaviour, Fehr and Fischbacher introduce the motives of inequity aversion and reciprocal fairness, and they discuss neurophysiological origins of the behaviour observed in the laboratory. Finally, they enquire into the evolutionary origins of altruism. They ask whether one can identify a mechanism through which the benefits of altruistic behaviour are conferred on those whose behaviour is altruistic rather than on others, for unless this be the case, the evolution of altruistic behaviour cannot be explained.
In a second introductory article, Herbert Gintis explores the relationship between developments in behavioural game theory and conventional economic theory. Developments in the former reach back into the 1950s and early 1960s (with Maurice Allais’, Daniel Ellsberg’s and Vernon Smith’s research) and continue in their most recent form in the work of Fehr and his colleagues. Contrary to the claims of many, the results of behavioural game theory do not, Gintis argues, undermine conventional economic theory. Amongst other things, Gintis observes that the rationality postulate of economic theory makes no claims about the content of people’s preferences; rather rationality requires that preferences be consistent but not that individuals be self-interested; consequently, the other-regarding preferences of behavioural game theory do not contradict economic theory. Gintis also looks at behaviour in experiments in which contractual promises are neither perfectly nor costlessly enforced. In such situations in particular, strongly reciprocal behaviour is manifested and leads to efficiency gains over behaviour which is purely self-regarding. Gintis’ piece is an important contribution and will enrich ongoing debates between ’theorists’ and ’experimenters’ about the relationship between economic theory and experimental results.
The first two contributions offer both the specialist and those less initiated in the behavioural sciences an excellent basis for following the commentaries on Fehr’s work which occupy much of this issue. The commentaries can be divided into (i) those which focus on the evolutionary origins of altruism and (ii) those of a conceptual nature which examine the conceptual apparatus developed by Fehr (in particular the concept strong reciprocity ).
In Altruists with Green Beards, Fehr and Fischbacher examine Robert Frank’s ’green beard’ explanation of the evolution of cooperative behaviour. Frank’s hypothesis assumes that cooperators have characteristics (’green beards’) which allow others to distinguish them from non-cooperators. Green beards will cooperate with one another and reap the gains of mutual cooperation whereas they will not cooperate with the beardless (or with those who bear beards not green in colour). Fehr and Fischbacher reject Frank’s hypothesis. First, it does not allow for impostors (’mutants’) who, although in possession of a green beard, do not have a behavioural disposition to cooperate. This raises the issue whether impostors can be detected as impostors, something upon which Fehr and Fischbacher cast doubt. This being the case, impostors would have an easy time infiltrating a population of genuine green beards and taking advantage of their cooperative nature. Fehr and Fischbacher conclude their critique with experimental results which show that defectors, not cooperators, benefit most in prisoners’ dilemmas with green beards.
In the paper which follows, Robert Frank responds to Fehr and Fischbacher and contests their interpretation of the green beard hypothesis: whereas Fehr and Fischbacher consider brief interactions between strangers, Frank’s hypothesis involves people with much mutual familiarity, e.g., close friends; it is their green bearded nature, not that of strangers, which can be identified and form the basis of mutual cooperation. Although it might not be immediately clear from the interchange, the differences between Frank, and Fehr and Fischbacher concern the unit of evolutionary selection: Frank focuses on individual selection whilst Fehr and Fischbacher find group selection to be a plausible alternative.
Christopher Stephens pursues the theme of group selection in his Strong Reciprocity and the Comparative Method. Strong reciprocity (henceforth SR), he holds, is a behavioural trait which cannot easily be explained with the concepts of evolutionary biology. Yet he is sceptical about testing for the adaptivity of SR and hence about finding an explanation for its existence. Stephens’ point is methodological: SR is distinctly human and so one cannot draw comparisons between human and non-human species (be the latter extinct or living) to learn more about SR; and in lieu of comparative material, it is difficult to test the explanatory power of various adaptive hypotheses.
Jason McKenzie Alexander likewise examines the evolutionary foundations of SR. He takes issue with Fehr’s hypothesis that SR be attributable to ’specialised cognitive machinery’ which enables humans to distinguish those with whom we interact infrequently from others with whom we interact frequently and over a long time. Alexander argues that no such cognitive apparatus is necessary; that which we have for attributing beliefs and desires to others suffices. For instance, if A is a strong reciprocator and rewards B for cooperating, A must, amongst other things, believe that B has actually bestowed benefits on others. This and others examples suggest that no further apparatus need have evolved to explain SR. Alexander highlights the importance for altruism researchers to pay attention to individuals’ beliefs (be they true or false).
Terence Burnham and Dominic Johnson aver that if SR arose by natural selection, then it would be maladaptive in any conceivable laboratory experiment which aims at modelling interaction between anonymous strangers. If SR had evolved through group selection it would be maladaptive in its laboratory manifestations since the participants in the experiments do not constitute a group in the relevant sense; the evolution of SR and its manifestation in the laboratory are not analogous. Burnham and Johnson suggest that, in the environment of modern societies, SR is a dysfunctional trait, comparable to the human craving for sweets; we rely upon it ’at our peril’. Moreover, they argue that SR is most parsimoniously explained by individual selection, while Fehr views group selection theory as a plausible alternative to theories of individual selection.
Alex Rosenberg and Stefan Linquist hold it to be unlikely that a complex type of behaviour like human cooperation be a genetically encoded trait. They reject the view, common in evolutionary biology, that it suffices to present a ’how possible’ explanation if one wants to understand a ubiquitous phenomenon like human cooperation. Such explanations posit reconstructions which show how a given phenomenon could potentially have arisen without identifying the casual mechanisms which actually gave rise to the phenomenon. Their article challenges the orthodox narrative about the crucial role of hunting for the evolution of co-operation. In contrast to their closest relatives in the animal kingdom, humans are a spectacularly successful species and their success lies in their having found solutions to the problem of attaining cooperative equilibria in social interactions. Other, closely related species like chimps, gorillas and Homo erectus did not stumble on such solutions. Looking to the future, Rosenberg and Linquist argue that gene-sequence data will help to reconstruct the early stages of the cultural transmission of cooperation and they expect computational genomics to be soon in a position to give macromolecular explanations for those nucleotide sequences which enable humans to develop the capacity to co-operate.
With the four contributions which follow, we move from evolutionary pastures to conceptual issues in Fehr’s work. Each piece examines the term “strong reciprocity“ and enquires whether the behaviour which Fehr characterises as strongly reciprocal is rightly denominated ’altruistic’. Anton Leist thinks not and tries to inject conceptual clarity into discussions of “altruism“ by contrasting it with “self-interest“, “selfishness“ and “fairness“. For Leist, attention needs to be paid to the motives of individuals if one is to characterise their behaviour as altruistic. Fehr’s work, argues Leist, is too agnostic regarding the motives of sub jects. Consequently, Leist interprets the results of ultimatum, public goods and trust experiments in light of the possible motives involved. One motivating force which Leist suggests is a desire for ’social recognition’. He argues for an approach to human behaviour which takes more account of the social relations in which people stand and which condition their behaviour in contrast to an experimental approach which, in Leist’s view, abstracts from such relations.
Hans Bernhard Schmid argues that Fehr’s use of the term “altruism“ (and cognate terms) presupposes one, but only one possible, interpretation of the results of his third-party punishment experiments. Schmid urges that we look not to “altruism“ as an interpretational category in understanding sub jects’ behaviour but rather to “nostrism“, a term he takes from Ortega y Gasset. Nostrism reflects the way an individual’s identity is constituted by her group-belongingness. Schmid looks at the shared identities of sub jects in third party punishment experiments. He notices that although experimenters try to keep instructions to sub jects abstract and value-free, those in the third party punishment experiment contain suggestive formulations: the two sub jects who can transfer money to one another are, for instance, described as members of a group or a team; the third party’s instructions likewise present the other two players thus. The instructions, argues Schmid, frame the experiment in a way which constitutes the identities of sub jects and which establishes certain norms between them; if these norms are transgressed, the third party enforces them through punitive measures. Schmid’s thoughts raise the question whether it would make a difference if, instead of being described as a team, sub jects were termed ’competitors’ or ’adversaries’. Fehr and Fischbacher address issues concerning group membership and identity in their first contribution to this issue. Such issues raise questions about applying the results of experiments, for to predict the behaviour of people ’outside the laboratory’, one has to enquire into their own sense of identity and group allegiance.
Mark Peacock, Michael Schefczyk and Peter Schaber also take a look at Fehr’s conceptual repertoire. They take issue with Fehr’s definition of “altruism“ because, by giving a ’behavioural’ definition of altruism, Fehr disregards motives. Motives, however, are essential in ascriptions of altruism in the eyes of the authors: if an individual’s action benefits others and is costly to the individual herself (as the behavioural definition holds), it is nevertheless only altruistic if she intended to benefit those others. Without this additional condition, altruistic acts can be distinguished neither from those with a deceptive appearance of altruism, nor from those through which an individual benefits others unintentionally. Peacock, Schefczyk and Schaber question the term “altruistic punishment“ accordingly and suggest that such punishment may be described as retaliatory. The authors also examine the causes which Fehr describes behind the behaviour of sub jects, one such cause being ’social preferences’. The social preferences of individual X take the welfare of others, Y 1 . . . Y n , into account and hence a decrease in their welfare decreases that of X. But if social preferences cause the behaviour which Fehr calls ’altruistic’, one can argue that individuals with social preferences are no different to those typically portrayed in microeconomic models, for both maximise their own utility even if the arguments of their respective utility functions differ. Only if one examines the motives of individuals can one distinguish the two.
Jon Elster draws out threads from Fehr’s work which are suggested by, but not developed in, the latter. In Fehr on Altruism, Emotion and Norms, Elster suggests refinements to the concepts trust and norms. He also addresses the theme of punishment and notes that Fehr’s sub jects harbour irrational beliefs when in the role of a third-party punisher in prisoners’ dilemma experiments: punishers punish defectors differentially according to the behaviour of other players; if both players defect in a two-player game, the third party is more lenient on both than she is on a player who unilaterally defects. Is it rational to vary the severity of punishment according to matters outside the control of the person to be punished? This seems to transgress our intuitions regarding moral responsibility. Elster also asks what motivates punishers. He suggests that people do not punish out of fear of being punished for not punishing defectors; instead, people punish spontaneously and not from an external incentive. Elster adverts to a peculiarity of Fehr’s work, namely that ’biological altruists’, as Fehr calls them, seem to punish in order to make themselves feel good (because the punishment activates reward-related structures in the brain). Punishment is costly and benefits others because it acts as a disciplining device which reduces the incidence of defection in the population. It therefore fulfils the criteria for biological altruism. Nevertheless, Elster notes, punishment is more like revenge than altruism when considered according to the motives of punishers.
In one of the next issues of Analyse & Kritik Ernst Fehr will reply to the points raised by the critics.
Table of Contents
Title: Human Altruism - Proximate Patterns and Evolutionary Origins
Author: Ernst Fehr / Urs Fischbacher
Abstract: Are people selfish or altruistic? Throughout history this question has been answered on the basis of much introspection and little evidence. It has been at the heart of many controversial debates in politics, science, and philosophy. Some of the most fundamental questions concerning our evolutionary origins, our social relations, and the organization of society are centered around issues of altruism and selfishness. Experimental evidence indicates that human altruism is a powerful force and unique in the animal world. However, there is much individual heterogeneity and the interaction between altruists and selfish individuals is key for understanding the evolutionary dynamics as well as the proximate patterns of human cooperation. Depending on the environment, a minority of altruists can force a majority of selfish individuals to cooperate or, conversely, a few egoists can induce a large number of altruists to defect. Current gene-based evolutionary theories cannot explain important patterns of human altruism pointing towards the need for theories of cultural evolution and gene-culture coevolution.
Title: Behavioral Game Theory and Contemporary
Author: Herbert Gintis
Abstract: It is widely believed that experimental results of behavioral game theory undermine standard economic and game theory. This paper suggests that experimental results present serious theoretical modeling challenges, but do not undermine two pillars of contemporary economic theory: the rational actor model, which holds that individual choice can be modeled as maximization of an objective function subject to informational and material constraints, and the incentive compatibility requirement, which holds that macroeconomic quantities must be derived from the interaction and aggregation of individual choices. However, we must abandon the notion that rationality implies self-regarding behavior and the assumption that contracts are costlessly enforced by third parties.
Title: Altruists with Green Beards
Author: Ernst Fehr / Urs Fischbacher
Abstract: If cooperative dispositions are associated with unique phenotypic features ('green beards'), cooperative individuals can be identi ed. Therefore, cooperative individuals can avoid exploitation by defectors by cooperating exclusively with other cooperative individuals; consequently, cooperators ourish and defectors die out. Experimental evidence suggests that subjects, who are given the opportunity to make promises in face-to-face interactions, are indeed able to predict the partner's behavior better than chance in a subsequent Prisoners' Dilemma. This evidence has been interpreted as evidence in favor of green beard approaches to the evolution of human cooperation. Here we argue, however, that the evidence does not support this interpretation. We show, in particular, that the existence of conditional cooperation renders subjects' choices in the Prisoners' Dilemma predictable. However, although subjects predict behavior better than chance, sel sh individuals earn higher incomes than conditional cooperators. Thus, although subjects may predict other players' choices better than chance evolution favors the sel sh subjects, i.e., the experimental evidence does not support the green beard approach towards the evolution of cooperation.
Title: Altruists with Green Beards: Still Kicking?
Author: Robert H. Frank
Abstract: In earlier work, I proposed the 'adaptive standard of rationality', according to which narrow self-interest models can be broadened by positing additional tastes, but only upon a plausible showing that those tastes do not hamper resource acquisition in competitive environments. This proposal is related to the green beard hypothesis from biology, according to which altruism might be adaptive if its presence could be reliably signaled by some observable feature, such as a green beard. In their contri- bution to this issue Ernst Fehr and Urs Fischbacher o er theoretical arguments and describe laboratory experiments whose results they interpret as refuting my version of the green beard hypothesis. In this response, I argue that their theoretical arguments and experimental evidence pose no threat to the green beard hypothesis.
Title: Strong Reciprocity and the Comparative Method
Author: Christopher Stephens
Abstract: Ernst Fehr and his collaborators have argued that traditional explanations of human cooperation cannot account for strong reciprocity. They provide substantial empirical evidence that strong reciprocity is an important phenomenon that cannot be explained by the traditional models of kin selection or reciprocal altruism. In this note, however, I argue that it will be di cult to test speci c adaptive explanations of strong reciprocity because it is apparently unique to humans. Consequently, it is di cult to employ the comparative method, which is one of biology's best tools for testing adaptationist claims.
Title: The Evolutionary Foundations of Strong Reciprocity
Author: Jason McKenzie Alexander
Abstract: Strong reciprocators possess two behavioural dispositions: they are willing to bestow bene ts on those who have bestowed bene ts, and they are willing to punish those who fail to bestow bene ts according to some social norm. There is no doubt that peoples' behaviour, in many cases, agrees with what we would expect if people are strong reciprocators, and Fehr and Henrich argue that many people are, in fact, strong reciprocators. They also suggest that strongly reciprocal behaviour may be brought about by specialised cognitive architecture produced by evolution. I argue that specialised cognitive architecture can play a role in the production of strongly reciprocal behaviour only in a very attenuated sense, and that the evolutionary foundations of strong reciprocity are more likely cultural than biological.
Title: The Biological and Evolutionary Logic of Human Cooperation
Author: Terence C. Burnham / Dominic D. P. Johnson
Abstract: Human cooperation is held to be an evolutionary puzzle because people voluntarily engage in costly cooperation, and costly punishment of non-cooperators, even among anonymous strangers they will never meet again. The costs of such cooperation cannot be recovered through kin-selection, reciprocal altruism, indirect reciprocity, or costly signaling. A number of recent authors label this behavior "strong reciprocity", and argue that it is: (a) a newly documented aspect of human nature, (b) adaptive, and (c) evolved by group selection. We argue exactly the opposite; that the phenomenon is: (a) not new, (b) maladaptive, and (c) evolved by individual selection. In our perspective, the apparent puzzle disappears to reveal a biological and evolutionary logic to human cooperation. Group selection may play a role in theory, but it is neither necessary nor sufficient to explain human cooperation. Our alternative solution is simpler, makes fewer assumptions, and is more parsimonious with the empirical data.
Title: On the Original Contract: Evolutionary Game Theory and Human Evolution
Author: Alex Rosenberg / Stefan Linquist
Abstract: This paper considers whether the available evidence from archeology, biological anthropology, primatology, and comparative gene-sequencing, can test evolutionary game theory models of cooperation as historical hypotheses about the actual course of human prehistory. The examination proceeds on the assumption that cooperation is the product of cultural selection and is not a genetically encoded trait. Nevertheless, we conclude that gene sequence data may yet shed signi cant light on the evolution of cooperation.
Title: Social Relations Instead of Altruistic Punishment
Author: Anton Leist
Abstract: Ernst Fehr's experimental research on altruistic behaviour aims at superseding the classical homo oeconomicus in micro-economic behaviour theory. This essay discusses Fehr's results from two points of view: rst, in regard to the understanding of social action associated with the term "altruism"; second, in regard to the 'anthropological' strategy of research that is based on the laboratory method. Against the emphasis on altruism it will be argued that it misleads into providing a distorted description of social acting, and that, due to insu cient clarity about motives for acting, Fehr's empirical results give evidence not of altruism but rather of phenomena of social recognition. The objection against the anthropological strategy will be that it makes visible only local phenomena within prevailing social conditions and that it thus assumes more than it explains.
Title: 'Nostrism': Social Identities in Experimental Games
Author: Hans Bernhard Schmid
Abstract: In this paper it is argued that a) altruism is an inadequate label for human cooperative behavior, and b) an adequate account of cooperation has to depart from the standard economic model of human behavior by taking note of the agents' capacity to see themselves and act as team-members. Contrary to what Fehr et al. seem to think, the main problem of the conceptual limitations of the standard model is not so much the assumption of sel shness but rather the atomistic conception of the individual. A much-neglected question of the theory of cooperation is how the agent's social identity is determined, i.e. how individuals come to think of themselves and act as members of a group. Considering as an example one of Fehr et al.'s third party punishment experiments, I shall argue that the agents' identities (and thus the result of the experiment) are strongly in uenced by the way the experiment is presented to the participants, especially by the collectivity-related vocabulary used in the instructions.
Title: Altruism and the Indispensability of Motives
Author: Mark S. Peacock / Michael Schefczyk / Peter Schaber
Abstract: In this paper we examine Fehr's notions of "altruism", "strong reciprocity" and "altruistic punishment" and query his ascription of altruism. We suggest that, pace Fehr, altruism cannot be de ned behaviourally because the de nition of altruism must refer to the motives of actors. We also advert to certain inconsistencies in Fehr's usage of his terms and we question his explanation of altruism in terms of 'social preferences'.
Title: Fehr on Altruism, Emotion, and Norms
Author: Jon Elster
Abstract: I discuss recent work by Ernst Fehr and his collaborators on cooperation and reciprocity. (i) Their work demonstrates conclusively the reality and importance of non-self-interested motivations. (ii) It allows for a useful distinction between trust and blind trust. (iii) It points to a category of quasi-moral norms, distinct both from social norms and moral norms. (iv) It demonstrates how social interactions can generate irrational belief formation. (v) It shows the potential of punishment for sustaining social norms and for overcoming the second-order free rider problem as well as obstacles to group selection. (vi) It o ers a provocative experimental basis for the `warm-glow' explanation of altruistic behavior. I conclude by suggesting some experiments that might allow for further developments of the theory.