1    

Using evolutionary thinking to cut across disciplines: the example of the argumentative theory of reasoning

Hugo Mercier Philosophy, Politics and Economics Program University of Pennsylvania 313 Cohen Hall 249 South 36th Street Philadelphia, PA 19104 [email protected] http://sites.google.com/site/hugomercier/

To be published in: Zentall, T. & Crowley, P. (Eds.) Comparative Decision Making. Oxford University Press.

2     Psychology often has a strange way to divide labor between its sub-fields. The very same psychological mechanism can be studied by different academic groups that rely on different assumptions and reach opposite results. Moreover, different sub-fields are also apt to carve up the mind in different ways, which sometimes seem based more on introspection than on principled theory (Cosmides & Tooby, 1994; Tooby & Cosmides, 1992). As a result communication between these sub-fields—to compare results for instance—is bound to be fraught with difficulties. Reasoning is a good case in point. If reasoning is defined loosely as an effortful, conscious mechanism of reflection that allows the use of formal rules, then one soon realizes that it has been studied by the psychology of reasoning, but also under the umbrella of judgment and decision making, social psychology, moral psychology, crosscultural psychology, developmental psychology, and doubtlessly several other disciplines. Yet there has been little cross-fertilization: the conclusions painstakingly reached in one field often fail to reach other disciplines. Psychologists of reasoning have uncovered severe failings: when confronted with logical tasks, people try to reason but they often stumble and fail to solve even trivial problems.1 Yet reasoning can also allow people to solve the very same problems—if people are smart enough, if they have enough time (Evans, 2006) and, more importantly, if they have the right knowledge (Stanovich & West, 1999). Similar conclusions have been reached in the study of decision making in which more than 30 years of ‘heuristics and biases’ have unearthed a wealth of decision making mistakes (Gilovich, Griffin, &                                                                                                                 1  It is not very sensible to make blanket statements about the performance of a given cognitive mechanism. The point here is that reasoning performs poorly when it comes to the tasks it is classically supposed to achieve, such as solving simple deductive problems, something it does reliably poorly (Evans, 2002)

3     Kahneman, 2002). Yet here as well, reasoning is hailed as the solution to get people out of these cognitive traps (Kahneman, 2003). The study of argumentation—as we will see, a crucial aspect of reasoning—seems to have reached even more conflicting conclusions: “Two facts about argumentation seem beyond dispute: (1) young children are good at it (Mercier, 2011b); and (2) adolescents and adults are bad at it (Kuhn, 2005, 2009).” (Moshman, 2011, p.192) Although I will later dispute point (2) (pace Moshman), the field is clearly ambivalent regarding reasoning’s performance. Equally conflicting are the conclusions of social psychologists. While some join the psychologists of reasoning and decision making in seeing reasoning as a way to correct biases (Gilbert, Pelham, & Krull, 1988), others emphasize the detrimental effect it can have on our decisions (Dijksterhuis, 2004; Kunda, 1990; T. D. Wilson et al., 1993). Over the past years, moral psychology has gone from seeing reasoning in an overwhelmingly positive light (Kohlberg, 1987; Piaget, 1997) to a serious skepticism, stressing that reasoning is often used merely to provide post-hoc justifications for intuitive judgments (Haidt, 2001). Similarly, cross-cultural psychology went from casting reasoning as a nearly unequivocal good (Cole, 1971; Luria, 1934) to a more neutral position that sees different ways of reasoning as being equally successful, but at different tasks (Buchtel & Norenzayan, 2009). Before trying to account for what seems like a bewildering pattern of performance, we must first make sure that different disciplines are referring to the same cognitive mechanism. Happily, many domains of psychology seem to be converging on dual process theories of the mind (for review, see Evans, 2008). Dual process models

4     divide cognitive mechanisms in two categories. The bulk of these mechanisms belong to ‘system 1,’ which is characterized as being automatic and relying on associations and heuristics. System 1 processes—or intuitions—tend to work quickly and to require little effort, as they do not usually rely on working memory (Evans, 2003). By contrast, system 2 mechanisms are said to be controlled and based on rules, but also to be slow and effortful, taxing working memory. They are usually thought to be conscious. These definitions are rather vague—a more specific alternative is elaborated in section 2—yet they are sufficient to show that when the different disciplines mentioned above talk about reasoning, they refer to some system 2 mechanism. So the disagreements over reasoning’s performance are unlikely to stem from a severe misunderstanding about what reasoning is. An alternative explanation is that people disagree about the function of reasoning. If one group of researchers thought that the function of feet was to walk, while another defended the hypothesis that their function was to manipulate objects, they would unsurprisingly reach opposite conclusions about the performance of feet. Yet this isn’t likely to be the case with reasoning: there seems to be a broad agreement that reasoning serves to better individual cognition. By correcting mistaken intuitions, reasoning is supposed to allow people to reach better beliefs and to make better decisions (Evans & Over, 1996; Kahneman, 2003; Stanovich, 2004). This can be called the classical view of reasoning. Clearly, reasoning can do these things: people sometimes reason their way to better beliefs and better decisions. Yet the empirical literature demonstrates that reasoning does not always lead to such positive outcomes—far from it. Reasoning often fails to correct misguided intuitions (Denes-Raj & Epstein, 1994). Sometimes it even

5     makes things worse, for instance by making people overconfident in their mistaken intuitions (Koriat, Lichtenstein, & Fischhoff, 1980). There is a major mismatch between what is commonly thought to be the function of reasoning and the performance of reasoning: reasoning doesn’t do well what it is supposed to do. One possibility would be that reasoning is simply a poorly functioning mechanism. This hypothesis however faces two problems. The first is that psychologists studying other cognitive mechanisms keep coming up with demonstrations of their quasioptimal behavior (e.g. Balci, Freestone, & Gallistel, 2009; Spellman, 1993; Trommershauser, Maloney, & Landy, 2008). Why would reasoning be such a flagrant exception? And how could such a flawed mechanism correct seemingly much more efficient intuitions? The second objection is that reasoning does not simply introduce random error, as would be expected of a shabby mechanism. Reasoning misleads people in predictable ways, mostly by strengthening—instead of correcting—misguided intuitions. The regularities in reasoning’s mistakes are a good sign that the mismatch between function and performance does not come simply from poor general functioning. Instead, we should reconsider the function of reasoning. Psychologists often rely on their intuitions or on naïve theories to determine what is the function of psychological mechanisms (Cosmides & Tooby, 1994; Tooby & Cosmides, 1992). In some cases their intuitions are likely to not be too far from the correct answer. For instance, when Marr postulates that the main function of the visual system is to create a representation of the outside world, this is likely to be a reasonably good approximation of the actual function of the visual system (Marr, 1982). However, evolutionary psychologists have argued that it is preferable to rely on evolutionary theory

6     to generate hypotheses about the function of cognitive devices (e.g. Barkow, Cosmides, & Tooby, 1992). Following the heed of evolutionary psychologists, Sperber relied on evolutionary theory to attribute to reasoning another function. As outlined in section 1, he used the framework of the evolution of communication to suggest that the function of reasoning is argumentative: producing arguments to convince other people and checking other people’s arguments (Sperber, 2000, 2001). On the basis of this hypothesis it has been possible to account for the seemingly bewildering pattern of performance mentioned above. The goal of the present chapter is to briefly present this evidence and show how recruiting evolutionary thinking can help discern broad trends emerging from different disciplines. I will start by summarizing the evolutionary rationale for the argumentative theory of reasoning. Section 2 specifies how reasoning is defined in this theory. All the remaining sections—3 to 13—review the predictions of the argumentative theory regarding the performance of reasoning, drawing from many areas of psychology.

1/ Argumentation and the evolution of reasoning

Within the Primate order, humans rely on communication to an unprecedented extent. They derive enormous benefits from communication: information about food, about dangers, about other people, about techniques, etc. But communicating also entails risks: people can be lied to, manipulated, misled. Yet, overall, communication has to be beneficial both to senders and receivers. The logic behind this conclusion is very simple: if senders do not benefit from communication, they stop sending (i.e. they evolve to stop

7     sending). Likewise, if receivers do not benefit from communication, they stop receiving (Dawkins & Krebs, 1978; Krebs & Dawkins, 1984). Given the dangers faced by receivers, there must exist some mechanism to ensure that communication is mostly honest—since dishonest communication tends not to be beneficial. Different solutions can be found throughout the animal kingdom (Maynard Smith & Harper, 2003), but humans mostly rely on the filtering of communicated information by mechanisms of epistemic vigilance (Sperber et al., 2010). Two of the major mechanisms of epistemic vigilance are trust calibration and coherence checking. People do not trust others uniformly: they grant more weight to information coming from people they deem to be competent and benevolent. People also test the coherence of what they are told against their previous beliefs, and tend to reject incoherent messages. Yet both mechanisms reject too much information: when someone you don’t trust enough tells you something that clashes with your beliefs, you are likely to reject the message, thereby missing out on some potentially valuable information. Sperber (2000, 2001) suggested that argumentation is a solution to this trust bottleneck. Senders can provide reasons to accept the messages they want to transmit. Receivers can then evaluate these reasons to determine whether they should accept the conclusion or not. As a result, more information passes between sender and receiver, making them both better off. Reasoning is a mechanism that chiefly evolved to allow people to find and evaluate arguments in dialogical contexts (Mercier & Sperber, 2009). The hypothesis that reasoning has an argumentative function—and therefore a social function—fits very well with current thought about the crucial role of the social environment for human evolution. (R. W. Byrne & Whiten, 1988; R. I. M. Dunbar, 1996;

8     Hrdy, 2009; Humphrey, 1976; Tomasello, Carpenter, Call, Behne, & Moll, 2005; Whiten & R. W. Byrne, 1997). In particular, cooperation may have been the main driver behind human evolution (Dubreuil, 2010; Sterelny, In press), and high levels of cooperation require extensive communication, which argumentation can facilitate.2 Others have suggested social functions for reasoning. Some of these suggestions are close to the proposal defended here (Billig, 1996; Gibbard, 1990; Haidt, 2001). Another hypothesis has been that argumentation evolved to show off one’s reasoning skills (Dessalles, 2007; Frankish, 2011). Undoubtedly, argumentation is sometimes put to such use, but the same thing is true of just about any skill, from language in general to athletic performance. However, such displays are relevant because the skills displayed have other uses, and so we are back to the question of what is the original function of reasoning (Mercier & Sperber, 2011a). A different suggestion is that reasoning may have a dialogic structure, but that its function is still to better individual cognition (GodfreySmith & Yegnashankaran, 2011). However, the pattern of performance mentioned above is also incompatible with this perspective, since it should predict that reasoning works for the betterment of individual cognition. Finally, it has also been suggested that reasoning for individual ends is merely an exaptation of reasoning for social ends (Evans, 2011; Frankish, 2011). This later hypothesis is thus very similar to the classical view of reasoning according to which the main function of reasoning is to help the lone reasoner reach better beliefs and better decisions. In several of the following sections, I will argue                                                                                                                 2  A correlate to this argument is that argumentation would not be useful if people had the same beliefs or the same interests. Small groups of our ancestors may have had a lot of shared baggage, both in cultural and genetic terms. However, they must still have had beliefs at odds with each other fairly often, due to differences of interests, access to information, expertise, etc. And even close relatives such as siblings or parents – children do not have perfectly aligned incentives.

9     that some of the features of reasoning are not compatible with the classical view: reasoning’s features serve an argumentative function well, but an individualistic function poorly.

2. Intuitive and reflective inferences

In line with the evolutionary theory suggested above, it is possible to delineate a view of reasoning that would be more principled than other dual process models, while maintaining the insight that reasoning constitutes a partially separable mechanism (for a detailed exposition, see Mercier & Sperber, 2009, 2011b). The basic process in this perspective is the intuitive inference. Very generally, a “process of inference is a process, the representational output of which necessarily or probabilistically follows from its representational input” (Mercier & Sperber, 2011b, p.58). In intuitive inferences, people are not aware of the reasons for drawing the inference. Nearly all of our inferences are intuitive and they can bear on many topic, from naïve physics to mate selection. Some inferences—that are particularly crucial for humans—bear on representations. For instance, mentalizing allows people to make inferences about other people’s representation: what they believe, desire, etc. Among the inferences bearing on representations, one subset bears on arguments: these inferences evaluate whether a given representation is a good reason to accept another representation. For instance, when confronted with “I think therefore I am,” most people intuitively infer that the premise (“I think”) supports the conclusion (“I am”): they have an intuition that this is a good argument. These intuitive inferences about arguments are the heart of reasoning. Yet

10     reasoning also requires a different type of inference, which can be dubbed a reflective inference. When an intuitive inference delivers a positive verdict regarding an argument, it is possible to accept the argument’s conclusion because of the argument. This mental act of accepting a belief because we have found a good enough reason to support it is what is commonly referred to as reasoning. Because reasoning—both the intuitive inferences bearing on arguments and the reflective inferences that may follow—is a metarepresentational mechanism, it gives reasoning an appearance of generality. To draw an analogy, using mentalizing one can attribute just about any thought one can entertain to anyone. Similarly, reasoning can process arguments about Camembert and quarks, sometimes in one stride (Fodor, 1983; Sperber, 1994). But reasoning is still a specialized, modular mechanism. It has a specific function: finding and evaluating reasons. It is about a specific object: representations (as opposed to, say, perceptual information). And it performs a specific operation: gauging the degree of support of one representation for another. Given that the current proposal is inspired by an evolutionary argument, it is good that it should fit with the evolutionarily inspired notion that the mind is massively modular (Sperber, 1994). While holding on to the insight that reasoning is a distinct psychological ability, the argumentative theory of reasoning differs from other dual process theories in several ways. The traits that are usually posited for each system in typical dual process theories are only unreliable indicators of whether reasoning is used. On the one hand, intuitive inferences can be slow, effortful and conscious, as when we’re looking for someone in a crowd using a typical intuitive mechanism: face recognition. On the other hand, reasoning can be fast and effortless, as when we’re involved in a freewheeling argument.

11     As mentioned above, reasoning is also in large part unconscious: people are generally not aware of why they think a given argument is strong or poor. Finally, the argumentative theory also characterizes reasoning in a more specific fashion than most dual process models. It is probable that what is usually referred to as system 2 comprises in fact several mechanisms: reasoning but also parts of planning, imagination, consequential thinking and possibly others (Mercier & Sperber, 2011a). At the time being, an algorithmic description of reasoning following from the argumentative theory hasn’t been fully specified (for a first stab, see Mercier, in press), but there is no a priori reason it can not be compatible with any of the existing theories of reasoning (Johnson-Laird, 2006; Oaksford & Chater, 2001; Rips, 1994). The remaining of the chapter reviews the predictions made by the argumentative theory about reasoning’s performance and some of its most striking features.

3/ Argumentation skills

The most straightforward prediction of the argumentative theory is that reasoning should be good at doing what it evolved to do, namely finding and evaluating arguments in dialogical contexts. This prediction does not allow distinguishing between the argumentative theory and the classical view of reasoning, since they could make the same prediction. However, it is necessary to test it since many scholars have questioned the value of argumentative skills and, if their doubts were truly founded, the current theory would be falsified.

12     Argument production is often thought to suffer from two main flaws: superficiality and confirmation bias. People would tend to find arguments that only “make superficial sense” (Perkins, 1985, p.568) and that only support their point of view (Kuhn, 1991). The question of bias is addressed in the next section. As for the apparent superficiality of arguments, it is in fact not incompatible with the predictions of the current theory, for two reasons. The first is that reasoning should not be expected to invest a lot of time and energy in finding foolproof arguments. In a discussion, one can try several times before convincing one’s interlocutor: there is no penalty for not starting out with the best argument. On the contrary, the interlocutor can make the work of reasoning easier by explaining the source of her disagreement, thereby allowing the speaker to suggest more appropriate arguments. As a result, we should expect an improvement in argument quality as the discussion progresses; which leads us to the second reason why people’s arguments are often deemed by psychologists to be superficial. In most experimental settings, there is no interaction, participants are not given the opportunity to progressively refine their arguments. Psychologists thus study almost exclusively what is likely to be the less refined products of reasoning. When the interaction is left to run its course, “participants . . . appear to build complex arguments and attack structure. People appear to be capable of recognizing these structures and of effectively attacking their individual components as well as the argument as a whole” (Resnick, Salmon, Zeitz, Wathen, & Holowchak, 1993, pp. 362– 63). Kuhn and her colleagues have observed similar improvements in reasoning as the result of sustained debate (Kuhn & Crowell, 2011; Kuhn, Shaw, & Felton, 1997).

13     Turning to argument evaluation, many people claim that it is also affected by the confirmation bias: people would spontaneously discount arguments whose conclusion clashes with their previous views (e.g. Klaczynski, 2000). However, the explanation of the confirmation bias to be offered in the next section does not apply to argument evaluation, which should be substantially more objective. The problem is that it is very hard, in a typical experimental setting, to disentangle argument evaluation and argument production. Except in formal domains, arguments are barely ever conclusive: conviction is more often the outcome of a sustained debate rather than of a single argument. When a participant is confronted with an argument and asked to evaluate it, she may evaluate it relatively objectively but, since she will not have been entirely convinced, she will then engage in a search for counterarguments. The search for counterarguments will display a confirmation bias that is bound to taint the evaluation of the argument. The best way to study argument evaluation is in the context of a dialogue: if argument evaluation were as biased as some suggest, people would resist almost any attempt at changing their minds. Yet as will be shown in section 7, people often change their mind when they argue with others. Finally, it should be noted that despite these methodological shortcomings, it is possible to show that when participants are motivated they grant more weight to strong arguments than to weak ones—although they may still have an overall bias towards their own view (Petty & Wegener, 1998).3

4/ The confirmation bias                                                                                                                 3  I can only scratch the surface of the issue here, and the interested reader is referred to the debate between Mercier and Sperber (2011a, 2011b) and Harrell (2011), Kuhn (2011) and Wolfe (2011). See also, on the role of learning for argumentation skills, Mercier (2011b).  

14     When people try to convince someone else, they are mostly interested in arguments supporting their position and going against that of the interlocutor. If one of the functions of reasoning is to produce arguments to convince others, we should expect reasoning to display a confirmation bias, which consists in the “seeking or interpreting of evidence in ways that are partial to existing beliefs, expectations, or a hypothesis in hand” (Nickerson, 1998, p. 175). According to the argumentative theory of reasoning, the confirmation bias is a feature of reasoning, not a flaw. Alternative explanations for the confirmation bias are difficult to sustain. The confirmation bias is not only observed in emotionally charged situation, but also in abstract reasoning tasks (e.g. Evans, 1996). It does not result from a lack of effort: asking people to be more objective (Lord, Lepper, & Preston, 1984), or paying them to reach the correct answer (Johnson-Laird & R. M. J. Byrne, 2002) has little effect. More importantly, the confirmation bias does not reflect a lack of ability, or the intrinsic difficulty of falsification. When people are confronted with statements they disagree with, they very easily find ways to falsify them—thereby confirming their own initial hunches (Cowley & R. M. J. Byrne, 2005; Dawson, Gilovich, & Regan, 2002; Sacco & Bucciarelli, 2008). The argumentative theory also makes two interesting predictions about the confirmation bias. First, it should mostly be observed in the production of arguments and not in their evaluation. The presence of the confirmation bias in argument production is established beyond reasonable doubt. Its absence or, at any rate, strong attenuation in argument evaluation is much harder to demonstrate (Mercier & Sperber, 2011b). It is however strongly suggested by the good results of group reasoning reviewed in section 7.

15     The second prediction is that the confirmation bias should only affect reasoning, and not other cognitive processes. For instance, a predator detection mechanism that would have a systematic tendency to confirm early judgments—even judgments that there are no predators around—would clearly be detrimental to fitness. In line with this idea, several tasks that used to be explained in terms of confirmation bias are now being described as resulting from sound intuitive heuristics. The 2, 4, 6 task—a task designed to evaluate hypothesis testing skills—was thought to demonstrate a confirmation bias (Wason, 1960), whereas in fact it reflects the process of a sound positive testing heuristic (Klayman & Ha, 1987). Likewise, the failure of most participants to solve the Wason selection task—a task designed to test people’s ability to falsify a statement—was pinned on problems grasping falsification (Wason, 1966), whereas in fact it merely reflects the operation of intuitive pragmatic mechanisms (Sperber, Cara, & Girotto, 1995). In both cases, it is reasoning that displays a confirmation bias: whatever hunch the intuitions provide reasoning fails to check, instead finding arguments in its support (Poletiek, 1996; Roberts & Newton, 2001). The prevalence and robustness of the confirmation bias are thorns in the side of the classical view of reasoning: why would reasoning be endowed with a feature that systematically leads to epistemic distortions? The problem is compounded by the extent of the damage that can be wreaked by the confirmation bias (see next section). It should be stressed, however, that the confirmation bias does not have to have negative epistemic consequences. In the proper group setting, the bias can become a form of division of cognitive labor, with each participant exploring the pros of her ideas and the cons of the other, rather than each having to exhaustively research the pros and cons of every

16     possibility. This explains in part the good performance of reasoning in group reviewed in section 7.

5/ Motivated reasoning

People often anticipate potential disagreements by preemptively finding arguments defending their decisions or beliefs. When they do so, the confirmation bias is given free reins, as people are unlikely to critically evaluate their own arguments. As a result, people end up accumulating arguments supporting their original intuition. The outcomes of this process have been documented in many experiments. A first consequence is belief polarization. When participants are left to reason about an attitude object, their attitudes become stronger in the direction of their initial hunch (e.g. Tesser, 1978). A second consequence is overconfidence. When people reason about their answers to general knowledge tests, they are apt to mostly find arguments supporting their initial intuition, making them unduly confident (Koriat et al., 1980). Another consequence of reasoning alone is belief perseverance. When people start reasoning about a belief they have formed, they create a scaffold of arguments around it. If the initial motivation for the formation of the belief is shown to be erroneous, the scaffold allows people to hang on to discredited beliefs (e.g. Guenther & Alicke, 2008; Ross, Lepper, & Hubbard, 1975). All of these effects can be put under the general umbrella of motivated reasoning (see, for review Kunda, 1990; Mercier & Sperber, 2011b). Some scholars have argued that motivated reasoning is but a special type of reasoning, which could be opposed to a

17     more objective type of reasoning (Kruglanski & Freund, 1983; Kunda, 1990). It is true that certain factors can restrain the power of the confirmation bias for lone reasoners. However, the most efficient way to attenuate motivated reasoning seems to be accountability (see for instance the studies listed in support of the existence of objective reasoning in Kunda, 1990). In specific conditions, having to justify one’s actions to an audience can make people anticipate potential counterarguments and set higher criteria for the arguments they generate. As a result, they may find themselves unable to satisfactorily defend their initial intuition and thus change their mind—often for the best, but not always (Lerner & Tetlock, 1999). The effects of accountability are compatible with the present view. The argumentative theory does not suggest that the lone reasoner engages in wishful thinking. On the contrary, the goal of internal argumentation is to make sure that some minimally decent arguments are available to defend our beliefs or actions. When participants find themselves unable to produce such arguments, they do change their mind. Many experiments have demonstrated that participants can be made more likely to engage in questionable behavior when excuses are provided. For instance, they cheat more on a test when the existence of free will is questioned: if there is no free-will, it’s not really their fault and they have a handy excuse to cheat (Vohs & Schooler, 2008). An increase in questionable behavior can only be obtained if reasoning, in the control condition, is unable to find an excuse. If reasoning could always find excuses, introducing a new excuse in the experimental condition should not make any difference. Accountability does not radically change the operation of reasoning, it simply raises the internal criteria used to decide what is a good argument.

18    

6/ Reason-based choice

Because of its confirmation bias, reasoning can lead to poor epistemic and practical consequences. But maybe reasoning would be able to drive people towards better decisions when they have weak intuitions? Unfortunately, that does not seem to be the case. When participants are faced with choices for which they lack strong intuitions, more reasoning often leads to worse decisions. The choices made after reasoning can be objectively worse (Dijksterhuis, 2004), less in line with expert judgment (T. D. Wilson & Schooler, 1991), or they can lead to less satisfaction for the participant herself (T. D. Wilson et al., 1993). The argumentative theory is in a good position to explain this poor performance of reasoning. When intuitions are weak, reasoning cannot simply find arguments supporting a preexisting hunch. Instead, reasoning looks for arguments that could support different intuitions. As a result, the intuition that is the easiest to justify ends up being chosen. Being easy to justify, however, is not the same thing as being the best option. Within the field of decision making, an important strand of research has examined reason-based choice “the idea that individual choice behavior under preference uncertainty [i.e. when intuitions are weak] can be better understood when seen as based on the available reasons or justifications for and against each alternative” (Simonson, 1989, p.158). Reason-based choice fits with the prediction of the argumentative theory. To provide an example, the introduction of a strictly dominated alternative in a choice set can change people’s choices (the attraction effect, see Huber, Payne, & Puto, 1982). If

19     participants are ambivalent between a cheap mediocre beer (A) and an expensive good one (B), the introduction of a beer that is not better than B but that is more expensive (C) will make people pick B because it becomes the easiest choice to justify (Simonson, 1989). Many poor decisions can be explained—in whole or in part—as reason-based choices (for a short review, see Mercier & Sperber, 2011b, sections 5.2 and 5.3). It should be stressed, however, that the research on reason-based choice has focused on its negative consequences but that it can also lead to positive outcomes. First, there is often an overlap between being easy to justify and being a good decision. For instance, when a student tries to solve a mathematics or a physics problem, the answer that is easiest to justify is likely to be the best—at least if the student has been taught the correct principles of mathematics and physics. This proviso is crucial: the efficiency of reasoning in cases of weak intuitions does not depend so much on the ability of the reasoner as on the cultural knowledge she has accumulated. The second positive consequence of reason-based choice is social. By picking the option that is most easily justified, decision makers may be better judged by others. For instance, reasoning leads people to choose unsatisfying electronic gadgets laden with useless features (Thompson, Hamilton, & Rust, 2005). Yet when someone chooses the more complex gadgets she is likely to be perceived—ironically—as more technology savvy (Thompson & Norton, 2008).

7/ Good performance in group

20     When people reason on their own, reasoning often leads them astray, either through the confirmation bias or through reason-based choice. But the argumentative theory also predicts that reasoning should lead to good outcomes when it is used in its normal conditions. The normal conditions for reasoning—the conditions for which it evolved (Millikan, 1987)—are those of deliberation (Mercier & Landemore, in press). Deliberation can be understood here as an exchange of arguments between at least two people who disagree about something but also seek a satisfying solution: figuring out who is right. In these conditions reasoning does work well. A good example is that of the Wason selection task, which generally produces approximately 10% of correct answers despite being logically trivial. When participants have to solve the task in group, their performance improves dramatically, reaching for instance 80% in one study (Moshman & Geil, 1998). For logical or mathematical tasks, it is observed that ‘truth wins’: as long as one participant has understood the problem, she is nearly always able to convince everybody that her answer is correct (Laughlin & A. L. Ellis, 1986). As a result, groups vastly outperform individuals. According to the argumentative theory, groups are able to reach better outcomes than individuals because in deliberative settings arguments can be thoroughly evaluated. The confirmation bias of the different participants is held in check by other participants who do not share their point of view. Poor arguments are rejected and good ones carry the day, leading the group to the best answer. The increase in performance in group settings could have other explanations, however. Groups could generally motivate people to spend more effort in the task. This is not the case: for most tasks, group performance is below expectations, in large part

21     because group members put in less effort than when they perform as individuals (e.g. Hill, 1982). Another explanation could be that group members simply pick out the smartest (Oaksford, Chater, & Grainger, 1999) or the most confident (Opfer & Sloutsky, 2011) individual and follow her lead. However, transcripts demonstrate that a substantial amount of discussion is necessary to reach an agreed upon solution: the individual with the correct answer has to convince everyone, simply stating the answer has little effect (e.g. Trognon, 1993). Moreover, groups can converge on a good answer that no member possessed prior to the discussion, giving rise to the assembly bonus effect, when groups outperform their best members (e.g. Laughlin, Bonner, & Miner, 2002; Michaelsen, Watson, & Black, 1989; Moshman & Geil, 1998, Sniezek & Henry, 1989). The very good results obtained for logical and mathematical tasks can also be obtained—if maybe in an attenuated form—in other types of tasks, such as inductive problems (e.g. on letters-to-numbers problems, Laughlin, VanderStoep, & Hollingshead, 1991). More importantly, they are also observed outside of the laboratory, in politics, law, science, business (see section 13) and schools (see section 11).

8/ Poor performance in group

The argumentative theory does not predict that reasoning always leads to felicitous outcomes in groups. Groups can violate the normal conditions for the use of reasoning just as much as individual reasoners. If everybody agrees to start with—or if dissenting voices are not heard—arguments will not be critically examined by the group. Instead, a scaled up version of individual polarization is likely to take place. Group members are

22     apt to generate different arguments to support the consensual view, arguments that will only strengthen every group member’s convictions. Thus the argumentative theory can explain group polarization—although other psychological mechanisms, such as conformity, are also likely to play a role (Isenberg, 1986). It should be stressed that group polarization is not a ‘law’—contra Sunstein (2002). It mostly occurs in very specific circumstances: when a group argues over a topic the members agree about—when there is disagreement, depolarization is more often observed (e.g., Vinokur & Burnstein, 1978). Reasoning is typically triggered by a disagreement, real or anticipated: people who agree with each other should not start arguing. Debate between like-minded people is thus somewhat artificial. For instance, when a jury agrees that a defendant is guilty, but must debate and justify the amount of the fine, reasons why the defendant should be fined pile up and deliberation leads to polarization with increased fines (Schkade, Sunstein, & Kahneman, 2000). In this case, deliberation is imposed on the jury by institutional constraints, and voting may have been more efficient.

9/ Moral psychology

The strongest echoes of the argumentative theory are found in the domain of moral psychology. As many other branches of psychology, the study of moral judgments and decisions is now dominated by dual process models. These models concur in positing that intuitions and emotions are the main drivers of moral judgments, but they differ in the exact weight to grant reasoning: relatively small in Haidt’s view (Haidt, 2001), more

23     important for Greene and his collaborators (e.g. Paxton & Greene, 2010). In line with Haidt’s view, the argumentative theory claims that most of individual reasoning consists in the post-hoc rationalization of preexisting intuitions. But it also suggests, as does Greene (e.g. Paxton, Ungar, & Greene, in press), that moral reasoning can change even deep-seated moral intuitions. Finally, it specifies that reasoning is more likely to influence moral intuitions in the context of a discussion than in private reasoning. That reasoning is often used to rationalize and justify pre-existing intuitions has been demonstrated in many experiments (e.g. Bandura, Barbaranelli, Caprara, & Pastorelli, 1996; Chance & Norton, 2008; Uhlmann, Pizarro, Tannenbaum, & Ditto, 2009). In a direct demonstration of the effects of reasoning, participants who could not reason—because of an increased cognitive load—deliverer fairer judgments, as they could not find justifications for unfairly favoring their own position (Valdesolo & DeSteno, 2008). The results of group reasoning are harder to interpret. Importantly, the prediction of the argumentative theory is not that reasoning should consistently lead groups towards more moral decisions. Instead, reasoning should, in the right conditions, lead groups to better decisions, whether these are more moral or not. Here “better” has the cynical sense of fitness enhancing: an immoral decision that may allow individuals to increase their fitness is a good one in that sense. For instance, when groups are confronted with economic games, they behave more in line with the predictions of game theory, but they are not more altruistic (e.g. Bornstein & Yaniv, 1998). Still, it can be argued that group reasoning often leads to more accurate moral judgments. What do I mean by more accurate moral judgments? A judgment that does a better job at tracking future behavior.

24     Sarah thinks that John is not a very good person, and so she refrains from interacting with him. If John was likely to try to deceive Sarah, it seems that the moral judgment has served its function—by contrast with thinking well of a conman for instance (see Mercier, 2011c). On the wider scale of societal changes, several authors have argued that narratives play a more important role than arguments in bringing about moral change (Bloom, 2010; Haidt & F. Bjorklund, 2007). The argumentative theory predicts that reasoning is at its best in interactive dialogues, not in unidirectional public speeches, and so it may not be surprising that narratives and appeals to emotions are more frequent in the later. However, it can also be argued that ‘everyday talk’, in which argumentation can function effectively, is a crucial driver of large scale moral change (e.g. Mansbridge, 1999).

10/ Cross cultural psychology

Psychologists ought to be wary when they claim to be studying universal cognitive mechanisms. Because the majority of psychological experiments are conducted within WEIRD (Western Educated Industrialized Rich Democratic) countries—more specifically on American undergraduates—it may not be warranted to extend their conclusions to other cultural groups (Henrich, Heine, & Norenzayan, 2010). As most evolutionary theories, the argumentative theory makes predictions that should apply to nearly all (non-pathological) humans. In the case of argumentation, some may suspect that its practice has been developed, or at least that it is mainly relied on in modern, Western populations. Maybe other cultural groups rely more on solitary reasoning and

25     are better at it? The two main groups that people have suspected of being endowed with reasoning abilities differing from those of Westerners are illiterate populations and Eastern cultures, which I examine in turn (the more extensive argument is in Mercier, 2011a). Sadly, there is a rich history in early anthropology and cultural psychology of denying the members of illiterate populations the ability to reason in the sense discussed here. For instance, Luria and his colleagues carried out experiments trying to demonstrate that illiterate Russian peasants could not solve even trivial syllogisms such as “In the Far North, where there is snow, all bears are white. Novaya Zemla is in the Far North. What colors are bears there?” (Luria, 1976, p.107). If the performance was indeed abysmal, the diagnostic was mistaken as in different circumstances similar populations can solve these problems. In particular, explicitly setting problem in a hypothetical world substantially improves performance (Dias, Roazzi, & Harris, 2005). Anecdotally, Cole and his colleagues (Cole, Gay, Glick, & Sharp, 1971) observed in the African group they were studying that “when engaged in group discussion, there was no difficulty in responding to such oral syllogisms” (p.186). As far as we can ascertain, illiterate populations exhibit patterns of reasoning similar to those observed among WEIRD people, including the boost offered by reasoning in group, even on abstract tasks. Other historians and anthropologists have claimed that there was a “lack of argumentation and debate in the Far-East” (Becker, 1986, see also Morrison, 1972; Nakamura, 1964). If, by lack of taste or ability for argumentation, such a large group as the East-Asian people failed to regularly engage in argumentation, this would surely be a deadly blow to the theory. This is not the case. Contrary to what has been suggested,

26     East-Asian languages are perfectly able to express logical relationships (Harbsmeier, 1998). Many strands of East-Asian culture disparage argumentation, yet this is not true of all of them—indeed, there is a rich tradition of Chinese rhetoric (e.g. Lloyd, 2007). And even those who pleaded against argumentation did it with forceful arguments (e.g. Hansen, 1992). Many results in cross-cultural psychology show that Easterners and Westerners can sometimes exhibit different behaviors when faced with some psychological tests (Nisbett, Peng, Choi, & Norenzayan, 2001). Importantly, these results do not demonstrate differences in competence, only in proclivity to rely on different cognitive mechanisms (e.g. Norenzayan, Smith, Kim, & Nisbett, 2002). Moreover, the majority of these experiments are conducted in individual contexts, and the argumentative theory does not predict that reasoning has to be activated in such contexts. It is therefore not surprising to observe cross-cultural (or inter-individual) variations in the tendency to rely on reasoning in individual contexts. To the best of my knowledge, the critical experiments involving group reasoning have not been replicated in non-Western settings. Some pilot data, however, suggests that Japanese students are just as able as Westerners to reap the benefits of argumentation: their performance on the Wason selection task increased dramatically when they solved it in small groups (Mercier, 2011e).

11/ Developmental data

Properties of reasoning relevant for the present purposes—those that are predicted by the argumentative theory—seem to be observable across all cultures. However, it could still

27     be that people in the entire world learn to reason in such a way, whether it is through explicit teaching, through observations of the interactions around them or through simple trial and error. Reasoning could begin as a mostly individual process, only to be co-opted during development to serve more argumentative purposes. Indeed, several authors have criticized the argumentative theory for its lack of attention to development (Kuhn, 2011; Moshman, 2011; Narvaez, 2011). It is true that an integrated theory that would be both evolutionary and developmental still remains to be spelled out. Still, it is possible to search in the patterns of reasoning in children for similarities with what can be observed in adults. First, it should be stressed that children display argumentative skills from very early on. As soon as toddlers start to form sentences, around 24 months of age, they produce justifications and arguments (e.g. Dunn & Munn, 1987; Kuczynski, Kochanska, Radke-Yarrow, & Girnius-Brown, 1987). Children also display a confirmation bias when they form arguments (e.g. Stein & Albro, 2001). More importantly children are also able to reap the benefits of reasoning in groups. Indeed, some of the strongest evidence demonstrating the efficiency of reasoning in appropriate group settings comes from developmental and educational research. The effects of group reasoning for children have mostly been studied within two traditions. The first is a neo-Piagetian research program that emphasizes the value of socio-cognitive conflict. This research has demonstrated that when children have to discuss a cognitive task with a peer, they generally outperform children solving the same task individually (e.g. Doise & Mugny, 1984; Perret-Clermont, 1980). The second tradition is that of collaborative learning, which has focused on gathering data from long-term projects in

28     school settings. In the experimental condition, students solve a variety of problems in groups rather than individually. Reviewing the relevant literature, Slavin remarked that “research on cooperative learning is one of the greatest success stories in the history of educational research” (Slavin, 1996, p. 43). Interestingly, the argumentative theory also explains why children can outperform adults in some decision making tasks. As noted earlier, reasoning sometimes leads to worse decisions—mostly because of motivated reasoning or reason-based choice. To the extent that children reason less in situations in which reasoning is not especially warranted, it is only to be expected that they should avoid some mistakes. Thus, children are less likely to commit the sunk cost fallacy (Morsanyi & Handley, 2008), they discount irrelevant information more easily (Klaczynski, submitted) and they are less sensitive to some framing effects (Reyna & S. C. Ellis, 1994). The overall similarity in the patterns of reasoning exhibited by children and adults is quite striking, and reinforces the claims of the argumentative theory (see, for an extensive review, Mercier, 2011b).

12/ Expertise

A strong argument against the classical view of reasoning stems from the poor performance of reasoning in individual reasoning tasks. It could be argued, however, that this poor performance merely reflects a lack of expertise. Maybe some people— experts—are better able to make use of reasoning and counteract its flaws and biases? By and large, that does not seem to be the case. Reasoning has the same traits in experts and laypeople (for review, see Mercier, 2011d). The main difference may be in the quantity

29     of arguments people can muster. The problem, however, is that if experts are as biased as laypeople, this trove of arguments is liable to only strengthen the effects of the confirmation bias. Thus people who are more knowledgeable about a topic tend to polarize more when they are left to reason on their own (Tesser & Leone, 1977). When people who are more knowledgeable are asked to list thoughts on a political issue, not only do they list more arguments supporting their side, they also list less arguments going in the other direction: a crowding out of argument that amplifies the confirmation bias (Taber & Lodge, 2006). Tetlock (2005) has also observed that the extra arguments experts are able to muster can make them more overconfident. Happily, experts also benefit from reasoning in group. Beyond experimental results (see for instance Lombardelli, Proudman, & Talbot, 2005), group discussions are central in business, science, law and politics, as explained in the next section. However, the same conditions apply to groups of experts and to groups of laypeople: debating with like-minded people can be dangerous. Indeed, the more knowledgeable the group members, the stronger is the polarization in a group of like-minded peers (Vinokur & Burnstein, 1974).

13/ Outside the laboratory

There is a large literature in experimental psychology—reviewed in sections 3 to 12 above—supporting the claims of the argumentative theory. However, some may question the ecological validity of these findings: maybe things work differently outside the laboratory. In particular, some may question the good performance of reasoning in group:

30     maybe external factors stop groups from performing well outside the laboratory? For instance people often have a representation of science being driven by lone geniuses rather than committees (Shapin, 1991). Yet ethnographic studies have recognized the lab meeting (the actual place where scientists discuss their ideas, not the lab of the psychology experiments) as the most crucial environment for fashioning theories and experiments (K. Dunbar, 1995). Groups are used to reach felicitous decisions in many domains outside of science. Corporations rely on teams at all levels, “from the shop floor to the executive suite” (Bainbridge, 2002; Cohen & Bailey, 1997). Indeed it could be argued that in these more naturalistic settings, groups are even apt to perform better than in the laboratory, as people get to know each other and to acknowledge each other’s strengths and weaknesses (Michaelsen et al., 1989). Group reasoning is also central in the judicial process. The adversarial system can be seen as a form of group reasoning that tries to make the best of the confirmation bias (see Van Koppen & Penrod, 2003). More straightforwardly, juries are asked to deliberate, not to vote, and deliberation can play an important role, allowing a verdict initially defended by a minority of jurors to be the final verdict of the jury (Kalven & Zeisel, 1966; Sandys & Dillehay, 1995). For lack of a benchmark, it is difficult to gauge the efficiency of jury reasoning, but some studies have shown that the verdict can track relevant properties of the case being judged (Sloan & Hsieh, 1990). In politics, reasoning in group is often seen through the prism of presidential or congressional debates. There are reasons why such debates may not lead to optimal outcomes: the participants are not as much interacting with one another as they are addressing a wider audience, television viewers for instance (Mercier, 2011d). By

31     contrast, deliberative democracy stresses the participation in debates of normal citizen. This movement, which has become a major force in political science, was originally a theoretical exercise regarding the best way to form opinions (e.g. Elster, 1998; Habermas, 1987). But hundreds of field experiments have now been conducted demonstrating the potential of deliberation among citizens. When citizens are brought together to deliberate they often end up with more informed beliefs, more convincing conclusions and, where relevant, more compelling policy proposals (e.g. Barabas, 2000; Fishkin & Luskin, 2005; Gastil & Dillard, 1999). The groups also often homogenize, bringing both sides of the political spectrum closer together (Luskin, Fishkin, & Jowell, 2002). I have stressed here the positive effects of reasoning in group outside the laboratory, because they may be more surprising than the pitfalls of reasoning. However, these are also well attested, from egregious motivated reasoning in the judicial system (Braman, 2009) to group polarization and groupthink among like-minded politicians (Janis, 1982).

14/ Conclusion

The present chapter begins by noticing how different sub-fields of psychology can reach results that are at odds with each other. Reasoning is a case in point: across (and sometimes even within) disciplines, opposite conclusion regarding the efficiency of reasoning can be found. Some hail it as a way to correct mistaken intuitions while others stress its weaknesses compared to intuitive mechanisms. I suggested that the adoption of an evolutionary perspective can help solve these dilemmas by bringing two crucial

32     clarifications. The first is a more principled way to carve up the mind, to isolate a specific cognitive mechanism (cf. section 2). The second is a more principled way to ascribe a function to this cognitive mechanism. In the case in hand, Sperber (2000, 2001) has suggested that reasoning should be thought of as a specific metarepresentational mechanism. He also suggested that its function is argumentative: to find and evaluate reasons in dialogic situations. Based on this hypothesis, we examined findings from many branches of psychology to see if they could be better accounted for from this perspective. This chapter has reviewed the evidence accumulated so far, showing how the argumentative theory can explain a wealth of findings in reasoning, decision making, social psychology and other areas of psychology. Evolutionary hypotheses can naturally lead to reviewing evidence that spans outside any given sub-field of psychology. Because evolutionary psychologists usually claim that the traits under study should be universal, they often pay attention to crosscultural variation (or lack thereof) (e.g. Buss, 1989; Sugiyama, Tooby, & Cosmides, 2002). Because evolutionary psychologists must claim that some relevant traits do not emerge purely from learning, they often pay attention to the amount of learning required to acquire a trait: if the trait is purely learnt (reading, chess), there is no need for an evolutionary explanation (e.g. D. F. Bjorklund & Pellegrini, 2002). Finally, because evolutionary psychologists make predictions about fitness enhancing traits, they are also inclined to look beyond the laboratory, to ‘real life’ behavior (e.g. Daly & M. Wilson, 1988; Thiessen, Young, & Burroughs, 1993). The central hypothesis of the argumentative theory of reasoning has also been checked against evidence from these domains. Broad trends across disciplines that had previously gone mostly unnoticed can now be brought

33     to the fore. It is only to be hoped that the theory may facilitate further cross-disciplinary dialogue.

34    

Bainbridge, S. M. (2002). Why a Board? Group Decisionmaking in Corporate Governance. Vanderbilt Law Review, 55, 1–55. Balci, F., Freestone, D., & Gallistel, C. R. (2009). Risk assessment in man and mouse. PNAS, 106(7), 2459–2463. Bandura, A., Barbaranelli, C., Caprara, G. V., & Pastorelli, C. (1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology, 71, 364-374. Barabas, J. (2000). Uncertainty and ambivalence in deliberative opinion models: Citizens in the Americans Discuss Social Security Forum. Annual Meeting of the Midwest Political Science Association. Barkow, J. H., Cosmides, L., & Tooby, J. (1992). The Adapted Mind. Oxford: Oxford University Press. Becker, C. B. (1986). Reasons for the lack of argumentation and debate in the Far East. International Journal of Intercultural Relations, 10(1), 75-92. Billig, M. (1996). Arguing and Thinking: A Rhetorical Approach to Social Psychology. Cambridge: Cambridge University Press. Bjorklund, D. F., & Pellegrini, A. D. (2002). The origins of human nature: Evolutionary developmental psychology. Washington, DC: American Psychological Association. Bloom, P. (2010). How do morals change? Nature, 464(7288), 490. Bornstein, G., & Yaniv, I. (1998). Individual and group behavior in the ultimatum game: Are groups more “rational” players? Experimental Economics, 1(1), 101–108.

35     Braman, E. (2009). Law, Politics, and Perception: How Policy Preferences Influence Legal Reasoning. Charlotte: University of Virginia Press. Buchtel, E. E., & Norenzayan, A. (2009). Thinking across cultures: Implications for dual processes. In J. S. B. T. Evans & K. Frankish (Eds.), In Two Minds. New York: Oxford University Press. Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioural and Brain Sciences, 12, 1-49. Byrne, R. W., & Whiten, A. (1988). Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes, and Humans. New York: Oxford University Press. Chance, Z., & Norton, M. I. (2008). I Read Playboy for the Articles. In M. S. McGlone & M. L. Knapp (Eds.), The Interplay of Truth and Deception: New Agendas in Theory and Research. Routledge. Cohen, S. G., & Bailey, D. E. (1997). What makes teams work: Group effectiveness research from the shop floor to the executive suite. Journal of management, 23(3), 239–290. Cole, M. (1971). The Cultural Context of Learning and Thinking: An Exploration in Experimental Anthropology. Methuen. Cole, M., Gay, J., Glick, J. A., & Sharp, D. W. (1971). The cultural context of learning and thinking. New York: Basic Books. Cosmides, L., & Tooby, J. (1994). Origins of domain specificity: The evolution of functional organization. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 85–116).

36     Cowley, M., & Byrne, R. M. J. (2005). When falsification is the only path to truth. In B. G. Bara, L. Barsalou, & M. Bucciarelli (Eds.), (pp. 250-255). Stresa, Italy: Mahwah, NJ: Erlbaum. Daly, M., & Wilson, M. (1988). Homicide. New York: Aldine de Gruyter. Dawkins, R., & Krebs, J. R. (1978). Animal signals: Information or manipulation? In J. R. Krebs & N. B. Davies (Eds.), Behavioural Ecology: An Evolutionary Approach (pp. 282-309). Oxford: Basil Blackwell Scientific Publications. Dawson, E., Gilovich, T., & Regan, D. T. (2002). Motivated reasoning and performance on the Wason selection task. Personality and Social Psychology Bulletin, 28(10), 1379. Denes-Raj, V., & Epstein, S. (1994). Conflict between intuitive and rational processing: when people behave against their better judgment. Journal of Personality and Social Psychology, 66(5), 819-829. Dessalles, J.-L. (2007). Why We Talk: The Evolutionary Origins of Language. Cambridge: Oxford University Press. Dias, M., Roazzi, A., & Harris, P. L. (2005). Reasoning from unfamiliar premises: A study with unschooled adults. Psychological Science, 16(7), 550-554. Dijksterhuis, A. (2004). Think different: the merits of unconscious thought in preference development and decision making. Journal of Personality and Social Psychology, 87(5), 586-98. Doise, W., & Mugny, G. (1984). The Social Development of the Intellect. Oxford: Pergamon Press.

37     Dubreuil, B. (2010). Paleolithic public goods games: why human culture and cooperation did not evolve in one step. Biology and Philosophy, 25(1), 53–73. Dunbar, K. (1995). How scientists really reason: Scientific reasoning in real-world laboratories. The nature of insight, 396. Dunbar, R. I. M. (1996). The social brain hypothesis. Evolutionary Anthropology, 6, 178– 190. Dunn, J., & Munn, P. (1987). Development of justification in disputes with mother and sibling. Developmental Psychology, 23, 791-798. Elster, J. (1998). Deliberative Democracy. Cambridge: Cambridge University Press. Evans, J. S. B. T. (1996). Deciding before you think: Relevance and reasoning in the selection task. British Journal of Psychology, 87, 223-240.     Evans, J. S. B. T. (2002) Logic and human reasoning: An assessment of the deduction paradigm. Psychological Bulletin, 128(6), 978-96. Evans, J. S. B. T. (2003). In two minds: dual-process accounts of reasoning. Trends in Cognitive Sciences, 7(10), 454-459. Evans, J. S. B. T. (2006). The heuristic-analytic theory of reasoning: Extension and evaluation. Psychonomic Bulletin and Review, 13(3), 378-395. Evans, J. S. B. T. (2008). Dual-processing accounts of reasoning, judgment and social cognition. Annual Review of Psychology, 59, 255-278. Evans, J. S. B. T. (2011). Reasoning is for thinking, not just for arguing. Behavioral and Brain Sciences, 34(02), 77–78. Evans, J. S. B. T., & Over, D. E. (1996). Rationality and Reasoning. Hove: Psychology Press.

38     Fishkin, J. S., & Luskin, R. C. (2005). Experimenting with a democratic ideal: Deliberative polling and public opinion. Acta Politica, 40(3), 284–298. Fodor, J. (1983). The Modularity of Mind. Cambridge, Massachusetts: MIT Press. Frankish, K. (2011). Reasoning, argumentation, and cognition. Behavioral and Brain Sciences, 34(02), 79–80. Gastil, J., & Dillard, J. P. (1999). Increasing political sophistication through public deliberation. Political Communication, 16(1), 3–23. Gibbard, A. (1990). Wise Choices, Apt Feelings. Cambridge: Cambridge University Press. Gilbert, D. T., Pelham, B. W., & Krull, D. S. (1988). On cognitive busyness: When person perceivers meet persons perceived. Journal of Personality and Social Psychology, 54(5), 733-740. Gilovich, T., Griffin, D., & Kahneman, D. (2002). Heuristics and Biases: The Psychology of Intuitive Judgment. Cambridge, UK: Cambridge University Press. Godfrey-Smith, P., & Yegnashankaran, K. (2011). Reasoning as deliberative in function but dialogic in structure and origin. Behavioral and Brain Sciences, 34(02), 80– 80. Guenther, C. L., & Alicke, M. D. (2008). Self-enhancement and belief perseverance. Journal of Experimental Social Psychology, 44(3), 706-712. Habermas, J. (1987). A Theory of Communicative Action. Boston: Beacon Press. Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108(4), 814-834.

39     Haidt, J., & Bjorklund, F. (2007). Social intuitionists reason, in conversation. In W. Sinnott-Armstrong (Ed.), Moral Psychology (pp. 241-254). Cambridge, MA: MIT Press. Hansen, C. (1992). A Daoist Theory of Chinese Thought. New York: Oxford University Press. Harbsmeier, C. (1998). Language and logic. Science and Civilisation in China. Cambridge, MA: Cambridge University Press. Harrell, M. (2011). Understanding, evaluating, and producing arguments: Training is necessary for reasoning skills. Behavioral and Brain Sciences, 34(02), 80–81. Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world. Behavioral and Brain Sciences, 33(2-3), 61-83. Hill, G. W. (1982). Group versus individual performance: Are N + 1 heads better than one? Psychological Bulletin, 91, 517-539. Hrdy, S. B. (2009). Mothers and Others. Cambridge, MA: Belknap Press. Huber, J., Payne, J. W., & Puto, C. (1982). Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis. The Journal of Consumer Research, 9(1), 90-98. Humphrey, N. K. (1976). The social function of Intellect. In P. P. G. Bateson & R. A. Hinde (Eds.), Growing Points in Ethology (pp. 303-317). Cambrige, Massachusetts: Cambridge University Press. Isenberg, D. J. (1986). Group polarization: A critical review and meta-analysis. Journal of Personality and Social Psychology, 50(6), 1141-1151. Janis, I. L. (1982). Groupthink (Vol. 2). Boston: Houghton Mifflin.

40     Johnson-Laird, P. N. (2006). How We Reason. Oxford: Oxford University Press. Johnson-Laird, P. N., & Byrne, R. M. J. (2002). Conditionals: A theory of meaning, pragmatics, and inference. Psychological Review, 109, 646-678. Kahneman, D. (2003). A perspective on judgment and choice: Mapping bounded rationality. American Psychologist, 58(9), 697-720. Kalven, H., & Zeisel, H. (1966). The American Jury. Chicago: University of Chicago Press. Klaczynski, P. A. (2000). Motivated scientific reasoning biases, epistemological beliefs, and theory polarization: A two-process approach to adolescent cognition. Child Development, 1347–1366. Klaczynski, P. A. (submitted). When (and when not) to make exceptions: Links among age, precedent setting decisions, conditional inferences, and argument evaluation. Klayman, J., & Ha, Y. (1987). Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review, 94, 211-228. Kohlberg, L. (1987). The psychology of moral development. San Francisco: Harper & Row. Van Koppen, P. J., & Penrod, S. (2003). Adversarial versus inquisitorial justice: Psychological perspectives on criminal justice systems. New York: Springer Us. Koriat, A., Lichtenstein, S., & Fischhoff, B. (1980). Reasons for confidence. Journal of Experimental Psychology: Human Learning and Memory and Cognition, 6, 107118.

41     Krebs, J. R., & Dawkins, R. (1984). Animal signals: Mind-reading and manipulation? In J. R. Krebs & N. B. Davies (Eds.), Behavioural Ecology: An Evolutionary Approach (Vol. 2, pp. 390-402). Oxford: Basil Blackwell Scientific Publications. Kruglanski, A. W., & Freund, T. (1983). The freezing and unfreezing of lay-inferences: Effects on impressional primacy, ethnic stereotyping, and numerical anchoring. Journal of Experimental Social Psychology, 19(5), 448-468. Kuczynski, L., Kochanska, G., Radke-Yarrow, M., & Girnius-Brown, O. (1987). A developmental interpretation of young children’s noncompliance. Developmental Psychology, 23(6), 799–806. Kuhn, D. (1991). The Skills of Arguments. Cambridge: Cambridge University Press. Kuhn, D. (2005). Education for Thinking. Harvard, MA: Harvard University Press. Kuhn, D. (2009). Adolescent thinking. In R. M. Lerner & L. Steinberg (Eds.), Handbook of adolescent psychology (3rd ed., Vol. 1, pp. 152–186). Hoboken, NJ: Wiley. Kuhn, D. (2011). What people may do versus can do. Behavioral and Brain Sciences, 34(02), 83–83. Kuhn, D., & Crowell, A. (2011). Dialogic Argumentation as a Vehicle for Developing Young Adolescents’ Thinking. Psychological Science, 22(4), 545. Kuhn, D., Shaw, V. F., & Felton, M. (1997). Effects of dyadic interaction on argumentative reasoning. Cognition and Instruction, 15, 287-315. Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480498.

42     Laughlin, P. R., & Ellis, A. L. (1986). Demonstrability and social combination processes on mathematical intellective tasks. Journal of Experimental Social Psychology, 22, 177–189. Laughlin, P. R., Bonner, B. L., & Miner, A. G. (2002). Groups perform better than the best individuals on letters-to-numbers problems. Organizational Behavior and Human Decision Processes, 88, 605-620. Laughlin, P. R., VanderStoep, S. W., & Hollingshead, A. B. (1991). Collective versus individual induction: Recognition of truth, rejection of error, and collective information processing. Journal of Personality and Social Psychology, 61, 50-67. Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of accountability. Psychological Bulletin, 125, 255-275. Lloyd, G. E. R. (2007). Towards a taxonomy of controversies and controversiality: Ancient Greece and China. In M. Dascal & H. Chang (Eds.), Traditions of Controversy (pp. 3-16). Amsterdam: John Benjamins Publishing Company. Lombardelli, C., Proudman, J., & Talbot, J. (2005). Committees versus individuals: An experimental analysis of monetary policy decision-making. International Journal of Central Banking, May, 181-205. Lord, C. G., Lepper, M. R., & Preston, E. (1984). Considering the opposite: A corrective strategy for social judgment. Journal of Personality and Social Psychology, 47, 1231-1243. Luria, A. R. (1934). The second psychological expedition to central Asia. Journal of Genetic Psychology, 41, 255–259.

43     Luria, A. R. (1976). Cognitive Development its Cultural and Social Foundations. Cambridge, MA: Harvard University Press. Luskin, R. C., Fishkin, J. S., & Jowell, R. (2002). Considered opinions: Deliberative polling in Britain. British Journal of Political Science, 32(03), 455–487. Mansbridge, J. (1999). Everyday talk in the deliberative system. In S. Macedo (Ed.), Deliberative politics: Essays on democracy and disagreement (pp. 211–42). New York: Oxford University Press. Marr, D. (1982). Vision: A Computational Investigation into the Human Representation and Processing of Visual Information. San Francisco: Freeman. Maynard Smith, J., & Harper, D. (2003). Animal Signals. Oxford: Oxford University Press. Mercier, H. (in press). Looking for arguments. Argumentation. Mercier, H. (2011a). On the universality of argumentative reasoning. Journal of Cognition and Culture, 11, 85-113. Mercier, H. (2011b). Reasoning serves argumentation in children. Cognitive Development. (26)3, 177-191. Mercier, H. (2011c). What good is moral reasoning? Mind & Society.  10(2), 131-148. Mercier, H. (2011d). When experts argue: explaining the best and the worst of reasoning. Argumentation. 25(3), 313-327. Mercier, H. (2011e). [Replication of Moshman and Geil (1998) with Japanese students]. Unpublished raw data. Mercier, H., & Landemore, H. (in press). Reasoning is for arguing: Understanding the successes and failures of deliberation. Political Psychology.

44     Mercier, H., & Sperber, D. (2009). Intuitive and reflective inferences. In J. S. B. T. Evans & K. Frankish (Eds.), In Two Minds. New York: Oxford University Press. Mercier, H., & Sperber, D. (2011a). Argumentation: its adaptiveness and efficacy. Behavioral and Brain Sciences, 34(2), 94-111. Mercier, H., & Sperber, D. (2011b). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(2), 57-74. Michaelsen, L. K., Watson, W. E., & Black, R. H. (1989). A realistic test of individual versus group consensus decision making. Journal of Applied Psychology, 74(5), 834-839. Millikan, R. G. (1987). Language, Though and Other Categories. Cambridge: MIT press. Morrison, J. (1972). The absence of a rhetorical tradition in Japanese culture. Western Speech, 36, 89-102. Morsanyi, K., & Handley, S. J. (2008). How smart do you need to be to get it wrong? The role of cognitive capacity in the development of heuristic-based judgment. Journal of Experimental Child Psychology, 99(1), 18-36. Moshman, D. (2011). Evolution and development of reasoning and argumentation: Comment on Mercier (2011). Cognitive Development. (26)3, 192-195. Moshman, D., & Geil, M. (1998). Collaborative reasoning: Evidence for collective rationality. Thinking and Reasoning, 4(3), 231-248. Nakamura, H. (1964). Ways of Thinking of Eastern Peoples: India, China, Tibet, Japan. Hawaii: University of Hawaii Press.

45     Narvaez, D. (2011). The world looks small when you only look through a telescope: The need for a broad and developmental study of reasoning. Behavioral and Brain Sciences, 34(02), 83–84. Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomena in many guises. Review of General Psychology, 2, 175-220. Nisbett, R. E., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of thought: Holistic versus analytic cognition. Psychological Review, 108(2), 291310. Norenzayan, A., & Heine, S. J. (2005). Psychological universals: What are they and how can we know. Psychological Bulletin, 131(5), 763-784. Norenzayan, A., Smith, E. E., Kim, B. J., & Nisbett, R. E. (2002). Cultural preferences for formal versus intuitive reasoning. Cognitive Science, 26(5), 653–684. Oaksford, M., & Chater, N. (2001). The probabilistic approach to human reasoning. Trends in Cognitive Sciences, 5(8), 349-357. Oaksford, M., Chater, N., & Grainger, R. (1999). Probabilistic effects in data selection. Thinking and Reasoning, 5, 193-243. Opfer, J. E., & Sloutsky, V. (2011). On the design and function of rational arguments. Behavioral and Brain Sciences, 34(02), 85–86. Paxton, J. M., & Greene, J. D. (2010). Moral Reasoning: Hints and Allegations. Topics in Cognitive Science, 2(3), 511–527. Paxton, J. M., Ungar, L., & Greene, J. D. (in press). Reflection and reasoning in moral judgment. Cognitive Science.

46     Perkins, D. N. (1985). Postprimary education has little impact on informal reasoning. Journal of Educational Psychology, 77, 562-571. Perret-Clermont, A.-N. (1980). Social Interaction and Cognitive Development in Children. London: Academic Press. Petty, R. E., & Wegener, D. T. (1998). Attitude change: Multiple roles for persuasion variables. In D. Gilbert, S. Fiske, & G. Lindzey (Eds.), The Handbook of Social Psychology (pp. 323–390). Boston: McGraw-Hill. Piaget, J. (1997). The moral judgment of the child. New York: Free Press. Poletiek, F. H. (1996). Paradoxes of falsification. Quarterly Journal of Experimental Psychology, 49A, 447-462. Resnick, L. B., Salmon, M., Zeitz, C. M., Wathen, S. H., & Holowchak, M. (1993). Reasoning in conversation. Cognition and Instruction, 11(3/4), 347-364. Reyna, V. F., & Ellis, S. C. (1994). Fuzzy-trace theory and framing effects in children’s risky decision making. Psychological Science, 5(5), 275–279. Rips, L. J. (1994). The Psychology of Proof: Deductive Reasoning in Human Thinking. Cambridge, MA: MIT Press. Roberts, M. J., & Newton, E. J. (2001). Inspection times, the change task, and the rapid response selection task. Quarterly Journal of Experimental Psychology, 54, 10311048. Ross, L., Lepper, M. R., & Hubbard, M. (1975). Perseverance in Self-Perception and Social Perception: Biased Attributional Processes in the Debriefing Paradigm. Journal of Personality and Social Psychology, 32(5), 880-802.

47     Sacco, K., & Bucciarelli, M. (2008). The role of cognitive and socio-cognitive conflict in learning to reason. Mind & Society, 7(1), 1-19. Sandys, M., & Dillehay, C. (1995). First-ballot votes, predeliberation dispositions, and final verdicts in jury trials. Law and Human Behavior, 19(2), 175–195. Schkade, D., Sunstein, C. R., & Kahneman, D. (2000). Deliberating about dollars: The severity shift. Columbia Law Review, 100, 1139–1176. Shapin, S. (1991). “The mind is its own place”: science and solitude in seventeenthcentury England. Science in Context, 4(01), 191–218. Simonson, I. (1989). Choice based on reasons: The case of attraction and compromise effects. The Journal of Consumer Research, 16(2), 158-174. Slavin, R. E. (1996). Research on cooperative learning and achievement: What we know, what we need to know. Contemporary educational psychology, 21(1), 43–69. Sloan, F. A., & Hsieh, C. R. (1990). Variability in Medical Malpractice Payments: Is the Compensation Fair? Law and Society Review, 24(4), 997–1039. Sniezek, J. A., & Henry, R. A. (1989). Accuracy and confidence in group judgment. Organizational behavior and human decision processes, 43(1), 1-28. Spellman, B. A. (1993). Implicit learning of base rates. Psycoloquy, 4, 61. Sperber, D. (1994). The modularity of thought and the epidemiology of representations. In Lawrence A Hirschfeld & Susan A Gelman (Eds.), Mapping the Mind: Domain Specificity in Cognition and Culture (pp. 39-67). Cambridge: Cambridge University Press.

48     Sperber, D. (2000). Metarepresentations in an evolutionary perspective. In D. Sperber (Ed.), Metarepresentations: A Multidisciplinary Perspective (pp. 117-137). Oxford: Oxford University Press. Sperber, D. (2001). An evolutionary perspective on testimony and argumentation. Philosophical Topics, 29, 401-413. Sperber, D., Cara, F., & Girotto, V. (1995). Relevance theory explains the selection task. Cognition, 57, 31-95. Sperber, D., Clément, F., Heintz, C., Mascaro, O., Mercier, H., Origgi, G., & Wilson, D. (2010). Epistemic vigilance, 25(4), 359-393. Stanovich, K. E. (2004). The Robot’s Rebellion. Chicago: Chicago University Press. Stanovich, K. E., & West, R. F. (1999). Discrepancies between normative and descriptive models of decision making and the understanding/acceptance principle. Cognitive Psychology, 38(3), 349-385. Stein, N. L., & Albro, E. R. (2001). The origins and nature of arguments: Studies in conflict understanding, emotion, and negotiation. Discourse Processes, 32(2&3), 113-33. Sterelny, K. (In press). The Evolved Apprentice. Cambridge, MA: MIT Press. Sugiyama, L. S., Tooby, J., & Cosmides, L. (2002). Cross-cultural evidence of cognitive adaptations for social exchange among the Shiwiar of Ecuadorian Amazonia. Proc Natl Acad Sci U S A, 99(17), 11537-42. Sunstein, C. R. (2002). The law of group polarization. Journal of Political Philosophy, 10(2), 175-195.

49     Taber, C. S., & Lodge, M. (2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50(3), 755-769. Tesser, A. (1978). Self-generated attitude change. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (pp. 289-338). New York: Academic Press. Tesser, A., & Leone, C. (1977). Cognitive schemas and thought as determinants of attitude change. Journal of Experimental Social Psychology, 13(4), 340-56. Tetlock, P. E. (2005). Expert political judgment: How good is it? How can we know? Princeton: Princeton University Press. Thiessen, D., Young, R. K., & Burroughs, R. (1993). Lonely hearts advertisements reflect sexually dimorphic mating strategies. Ethology and Sociobiology, 14(3), 209–229. Thompson, D. V., & Norton, M. I. (2008). The social utility of feature creep. In A. Lee & D. Soman (Eds.), Advances in Consumer Research (pp. 181-184). Duluth, MN: Association for Consumer Research. Thompson, D. V., Hamilton, R. W., & Rust, R. T. (2005). Feature fatigue: When product capabilities become too much of a good thing. Journal of Marketing Research, 42(4), 431-442. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences, 28(5), 675-691. Tooby, J., & Cosmides, L. (1992). The Psychological Foundations of Culture. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.), The Adapted mind: evolutionary psychology and the generation of culture (p. 19).

50     Trognon, A. (1993). How does the process of interaction work when two interlocutors try to resolve a logical problem? Cognition and Instruction, 11(3&4), 325-345. Trommershauser, J., Maloney, L. T., & Landy, M. S. (2008). Decision making, movement planning and statistical decision theory. Trends in Cognitive Sciences, 12(8), 291-297. Uhlmann, E. L., Pizarro, D. A., Tannenbaum, D., & Ditto, P. H. (2009). The motivated use of moral principles. Judgment and Decision Making, 4(6), 476–491. Valdesolo, P., & DeSteno, D. (2008). The duality of virtue: Deconstructing the moral hypocrite. Journal of Experimental Social Psychology. Vinokur, A., & Burnstein, E. (1974). Effects of partially shared persuasive arguments on group-induced shifts: A group-problem-solving approach. Journal of Personality and Social Psychology, 29(3), 305–315. Vinokur, A., & Burnstein, E. (1978). Depolarization of attitudes in groups. Journal of Personality and Social Psychology, 36(8), 872-885. Vohs, K. D., & Schooler, J. W. (2008). The value of believing in free will. Psychological Science, 19(1), 49. Wason, P. C. (1960). On the failure to eliminate hypotheses in a conceptual task. Quarterly Journal of Experimental Psychology, 12, 129-137. Wason, P. C. (1966). Reasoning. In B. M. Foss (Ed.), New Horizons in Psychology: I (pp. 106–137). Harmandsworth, England: Penguin. Whiten, A., & Byrne, R. W. (1997). Machiavellian Intelligence II: Extensions and Evaluations. Cambridge: Cambridge University Press.

51     Wilson, T. D., & Schooler, J. W. (1991). Thinking too much: Introspection can reduce the quality of preferences and decisions. Thinking, 60(2), 181-192. Wilson, T. D., Lisle, D. J., Schooler, J. W., Hodges, S. D., Klaaren, K. J., & LaFleur, S. J. (1993). Introspecting about reasons can reduce post-choice satisfaction. Personality and Social Psychology Bulletin, 19(3), 331. Wolfe, C. R. (2011). Some empirical qualifications to the arguments for an argumentative theory. Behavioral and Brain Sciences, 34(02), 92–93.