The Social Epistemology of Anthropology: Insights from Judgment Aggregation

Anthropological writings have, at times, been vague in the approach used to gather evidence of cultural and social beliefs of the peoples studied, and the method of representing the data to the reading public. This paper employs the theory of judgment aggregation in critiquing anthropological theory and practice. It will be structured in three parts: first, I will present the theory of judgment aggregation as constructed by Christian List and Philip Pettit; second, I will sketch some epistemological methods used by anthropologists, and assess their attitude toward the notions of judgment aggregation and group agency; and finally, I will apply List and Pettit’s arguments about effective group organization to anthropological practice of representing its studied peoples’ beliefs and judgments by proposing three possible changes in method that will allow for more accurate and faithful interpretations and descriptions. Dagan Douglas Reed College douglasd@reed.edu https://doi.org/10.7710/2155-4838.1177 Volume 9, Issue 1 Res Cogitans 2 | eP1177 Res Cogitans Judgment Aggregation An emerging subfield of epistemology, social epistemology has sought to uncover the social element inherent in certain forms of knowledge.1 It expands from epistemology’s exclusive focus on individuals into “investigating the epistemic effects of social interactions and social systems.” The field of judgment aggregation2 investigates “the epistemic quality of group doxastic attitudes (whatever their provenance may be),” and “the epistemic consequences of adopting certain institutional arrangements or systemic relations as opposed to alternatives” as well.3 List called this the “radical form” of social epistemology, wherein “certain multi-member groups themselves are taken to be epistemic agents capable of acquiring beliefs and knowledge.”4 Judgment aggregation, where judgments are binary expressions of attitudes,5 is concerned with the establishment of group doxastic attitudes (judgments and beliefs) out of individual doxastic attitudes.6 Juries determining defendants’ guilt or innocence, expert panels of scientists recommending policy, and bank committees forecasting future opportunities are just the sort of groups List and Pettit want to explore.7 Much of their work on judgment aggregation since their famous 2002 article “Aggregating sets of judgments: An impossibility result” has been clarification of the aggregation procedure, defense of the existence of group agents, and suggestion for certain configurations of groups using certain aggregation procedures to produce certain results. I will now briefly sketch their views on each of these first two matters, 1 Alvin I. Goldman and Thomas Blanchard, “Social Epistemology,” in Stanford Encyclopedia of Philosophy (Winter 2016 Edition), Edward N. Zalta (ed.), section 1, https://plato.stanford.edu/ archives/win2016/entries/epistemology-social. The authors place its formal beginnings in the 1970s and 1980s. 2 One of the many types of social epistemology to have developed since its inception. Alvin I. Goldman, “A Guide to Social Epistemology,” in Social Epistemology: Essential Readings, ed. by Alvin I. Goldman and Dennis Whitcomb (Oxford: Oxford University Press, 2011), 11. 3 Standford Encyclopedia of Philosophy, “Social Epistemology,” section 2 and 4.1-4.2. 4 Christian List, “Group Knowledge and Group Rationality: A Judgment Aggregation Perspective,” in Social Epistemology: Essential Readings, ed. by Alvin I. Goldman and Dennis Whitcomb (Oxford: Oxford University Press, 2011), 223. 5 Christian List and Philip Pettit, Group agency: the possibility, design, and status of corporate agents (Oxford: Oxford University Press, 2011), 37-38. 6 “A Guide to Social Epistemology,” 17. 7 “Group Knowledge,” 223. Douglas | The Social Epistemology of Anthropology commons.pacificu.edu/rescogitans eP1177 | 3 and return to the third in my conclusion. Aggregation procedures “are mechanisms a multi-member group can use to combine (‘aggregate’) the individual beliefs or judgments held by the group members into collective beliefs or judgments endorsed by the group as a whole.”8 Groups like juries and federal agencies seek to make decisive judgments on important propositions whilst adhering to the dual constraints of responsiveness to the judgments of their composing individual members, and collective rationality across their judgments and over time.9 One example of many different possible procedures that such groups may implement is systematic majority voting on each proposition in question: the proposition is considered by itself by each member of the group, who give a binary affirmation or denial of its veracity; then, the ‘majority’ judgment on that proposition becomes the group’s judgment on it. In addition to responsiveness and rationality, the procedure, or “aggregation function” (“a mapping that assigns to each profile of individual attitudes towards the propositions on the agenda the collective attitudes towards these propositions”10), must also meet three more minimal conditions that List and Pettit claim we would expect “of plausible group attitude formation”: universal domain, anonymity, and systematicity.11 Universal domain is when “any possible profile of individual attitudes towards the propositions on the agenda” may be input to the function. Anonymity is when “all individuals’ attitudes are given equal weight in determining the group attitudes,” as in a secret ballot. And systematicity is when “the group attitude on each proposition depends only on the individuals’ attitudes towards it, not on their attitudes towards other propositions, and the pattern of dependence between individual and collective attitudes is the same for all propositions.” List and Pettit show that any organization attempting to meet all of these standards at once will fail to do so, an issue they dubbed the discursive paradox.12 To avoid it, they suggest ignoring at least one of these conditions, so that groups can still function while not being perfect. I return to this idea in the conclusion. 8 “Group Knowledge and Group Rationality: A Judgment Aggregation Perspective,” 222. 9 Christian List and Philip Pettit, “Aggregating sets of judgments: An impossibility result,” Economics and Philosophy 18 (1) (2002): 91, 96. Rationality itself is bound by three of its own constraints, which are completeness, consistency, and deductive closure. See pages 97-98 for definition. 10 Group agency, 48. 11 Group agency, 49. 12 “Aggregating sets of judgments,” 95-96, 100. Volume 9, Issue 1 Res Cogitans 4 | eP1177 Res Cogitans Judgment aggregation depends heavily on the notion of group agency, which might not be readily accepted as existing. List and Pettit claim that in everyday speech, we regularly ascribe to groups’ and organizations’ motivations and actions, such as the FBI investigating such and such suspects, or Goodwill wanting to hire this or that many employees. Not only do we find it acceptable to speak in this manner despite knowing that these groups exist only as collections of their members under certain organizing principles, but also, List and Pettit claim, we could not make sense of these groups as merely collections of individuals. They argue that these groups have agency apart from though arising from the agency of the individual members. List states that thinkers [M]ay be prepared to treat certain groups as agents, provided some stringent conditions are met.... In particular, to be an agent, a group must exhibit patterns of behaviour vis-a-vis the outside world that robustly satisfy certain rationality conditions... In short, a necessary condition for epistemic agency in a group is an institutional structure (formal or informal) that allows the group to endorse certain beliefs or judgments as collective ones; and the group’s performance as an epistemic agent [i.e. how they perform at acquiring beliefs or knowledge] depends on the details of that institutional structure.13 This definition allows groups, from corporations to tribes, to be considered group agents.

commons.pacificu.edu/rescogitans eP1177 | 3 and return to the third in my conclusion.
Aggregation procedures "are mechanisms a multi-member group can use to combine ('aggregate') the individual beliefs or judgments held by the group members into collective beliefs or judgments endorsed by the group as a whole." 8 Groups like juries and federal agencies seek to make decisive judgments on important propositions whilst adhering to the dual constraints of responsiveness to the judgments of their composing individual members, and collective rationality across their judgments and over time. 9 One example of many different possible procedures that such groups may implement is systematic majority voting on each proposition in question: the proposition is considered by itself by each member of the group, who give a binary affirmation or denial of its veracity; then, the 'majority' judgment on that proposition becomes the group's judgment on it.
In addition to responsiveness and rationality, the procedure, or "aggregation function" ("a mapping that assigns to each profile of individual attitudes towards the propositions on the agenda the collective attitudes towards these propositions" 10 ), must also meet three more minimal conditions that List and Pettit claim we would expect "of plausible group attitude formation": universal domain, anonymity, and systematicity. 11 Universal domain is when "any possible profile of individual attitudes towards the propositions on the agenda" may be input to the function. Anonymity is when "all individuals' attitudes are given equal weight in determining the group attitudes," as in a secret ballot. And systematicity is when "the group attitude on each proposition depends only on the individuals' attitudes towards it, not on their attitudes towards other propositions, and the pattern of dependence between individual and collective attitudes is the same for all propositions." List and Pettit show that any organization attempting to meet all of these standards at once will fail to do so, an issue they dubbed the discursive paradox. 12 To avoid it, they suggest ignoring at least one of these conditions, so that groups can still function while not being perfect. I return to this idea in the conclusion. Judgment aggregation depends heavily on the notion of group agency, which might not be readily accepted as existing. List and Pettit claim that in everyday speech, we regularly ascribe to groups' and organizations' motivations and actions, such as the FBI investigating such and such suspects, or Goodwill wanting to hire this or that many employees. Not only do we find it acceptable to speak in this manner despite knowing that these groups exist only as collections of their members under certain organizing principles, but also, List and Pettit claim, we could not make sense of these groups as merely collections of individuals. They argue that these groups have agency apart from though arising from the agency of the individual members. List states that thinkers [M]ay be prepared to treat certain groups as agents, provided some stringent conditions are met…. In particular, to be an agent, a group must exhibit patterns of behaviour vis-a-vis the outside world that robustly satisfy certain rationality conditions… In short, a necessary condition for epistemic agency in a group is an institutional structure (formal or informal) that allows the group to endorse certain beliefs or judgments as collective ones; and the group's performance as an epistemic agent [i.e. how they perform at acquiring beliefs or knowledge] depends on the details of that institutional structure. 13 This definition allows groups, from corporations to tribes, to be considered group agents.

Anthropology
Anthropology is a discipline founded on the desire to understand what it means to be human. 14 It seeks to uncover not merely the factual reality of human existence, but lived human experience as well. But in uncovering, it must also reveal, which necessitates an understanding on the part of anthropologists on just how to represent their subjects of study, which are often people themselves with their own beliefs and values.
In this section, I will present anthropologists' conceptions of the discipline's aims and methods first, and then contrast their attitudes on group agency and belief 13 "Group Knowledge and Group Rationality," 223.
commons.pacificu.edu/rescogitans eP1177 | 5 aggregation to that held by List and Pettit. Following this, I will conclude the paper by critiquing the anthropological conception using the insights gained through judgment aggregation, and propose some applications of List and Pettit's theories of effective group organization to reforming anthropological practice.
One anthropologist, John Hugh Marshall Beattie, wrote in 1959, "Social anthropologists, then, study both what people do and what they think about what they do." 15 He typifies these thoughts about behavior as being either about "what they actually do," or about "why they ought to do." Arguing that the anthropologist cannot from mere observation accurately establish the significance of social practices, but must somehow get inside the heads of the individuals whose culture these practices belong to, Beatie asserts that "social relationships cannot be intelligibly conceived or described apart from the expectations, intentions, and ideas which they express or imply…. Behaviours can have no social significance apart from what it means to somebody, and unless such 'meanings' are taken into account nothing remotely resembling sociological understanding is possible." 16 The kind of understanding the anthropologist gathers from purely interpretative work, while not no understanding whatsoever, is guesswork, which may be more often wrong than right.
At his article's conclusion, Beattie posed a crucial question: "Is any understanding of particular societies or cultures possible…?" to which he insisted, "It is to be achieved, if it is achieved at all, by coming to understand the dominant beliefs and values of the people being studied… for only then can the investigator represent to himself -and perhaps to others -what it would be like to be a member of that society." 17 The act of representation is critical to anthropological practice, for it is through their textual records that the anthropologist imparts to their readers, the rest of the world, their research done in what is often a very little known setting. Because of the ever-shifting nature of human circumstances, if the anthropologist fails at first to get things right, there may not be a future opportunity for setting the record straightwhat has been committed to memory will be there for posterity. It is imperative, then, that anthropologists clearly set out methods for obtaining the correct beliefs their subjects hold.
Res Cogitans Yet one not insignificant strain of anthropological theory heavily emphasizes the interpretive approach, best encompassed by the work of the famous Clifford Geertz. Geertz (1973) advocated an ethnographical practice of so-called "thick description," a revelation of the meanings and intentions imbued in something like the winking of an eye, distinguishing a mocking wink of parody from a signaling wink of treachery, among many other possible facets. 18 Indeed, the concern with meaning agrees with the notion of anthropology that Beattie had put forward, but where Geertz departs from this is in shifting the focus from observation to interpretation in analysis.
Implicit in the observational approach to understanding culture was a reliance on the subject's interpretations of events, to which the anthropologist could not gain access but through interview. Bronislaw Malinowski noted this in his description of the ethnographer's task (1939): "Empirically speaking the field worker has to collect texts, statements, and opinions, side by side with the observation of behavior and the study of material culture." 19 In leaving this paradigm, Geertz admitted, "In short, anthropological writings are themselves interpretations, and second-and third-order ones to boot. (By definition, only a 'native' makes first order ones: it's his culture.) They are, thus, fictions; fictions, in the sense that they are 'something made,' 'something fashioned,' -the original meaning of fictiō -not that they are false, unfactual, or merely 'as if' thought experiments." 20 There will indeed be an element of craft in the forging of an ethnographical account, but Geertz's approach seems prone to a loss of integrity of the event an accuracy of the account: substituting the "source" with "scholarly artifice." 21 Whereas the observational program at least follows a hierarchy of information from source (i.e. the events as they happened), to first order to second order (i.e. the anthropologist's) interpretation, the interpretive program cuts out almost wholly the first order.
If the importance of belief aggregation were contested in anthropology, it would seem that skepticism regarding group minds has long prevailed. as not requiring the concept: "At the present time, we must follow the opinion of the majority of psychologists in dismissing the theory of the group mind as unnecessary, and therefore regard all the thinking and feeling which occurs in a culture as done by individuals." 22 Marc Augé (1995) stated, "The social begins with the individual; and the individual is the object of ethnological scrutiny." 23 Malinowski called the positing of a "group mind" a "fallacy," claiming, "The group, after all, is but the assemblage of individuals." 24 Yet in parts of their work, the social epistemological notion of group agency would seem to benefit their work if implemented.
For example, Malinowski illustrates the pressure on bachelors to conform to the "influence of tradition" in the proper acquisition of marriage and family. 25 The bachelor takes on for himself the values of his culture or society, and these go on to determine a great deal of how he conducts his life, such as what traits he demands in a spouse, how he attends to his household, and what kind of sexual activity he engages in. Interestingly, Malinowski conceives of this social force not amorphously: "Yet here we once more we do not deal with the group and the individual, but we would have to consider a whole set of human agglomerations: the group of the two principal actors (i.e., marriage), the prospective family, the already developed families of each mate, the local community, and the tribe as the bearer of law, tradition, and their enforcement." 26 It echoes loudly List's condition of group agency, than "is an institutional structure (formal or informal) that allows the group to endorse certain beliefs or judgments as collective ones." So while initially protesting against the necessity of the concept, it would seem that implicitly, anthropology has trouble fending it off in theory and practice.

Critique
Having provided some criticism of anthropology's pure approach, I will now provide three proposals for a refinement of anthropological method based on these 24 "The Group and the Individual," 90. 25 Ibid., 93-94. 26 Ibid., 94.

| eP1177
Res Cogitans principles: two based on solutions to the discursive paradox, and a third based on a formal distinction of attitudes of List's devising.
List and Pettit proposed as a solution to the discursive paradox a relaxing of one or more of the minimal conditions group agents should abide by. For the practice of anthropology, I see two of their strategies being potentially effective means of clarifying the judgments of the studied group agents: relaxing anonymity, which they dubbed "the authority strategy," and relaxing systematicity, or "the priority strategy." 27 Under the authority strategy, informants are explicitly distinguished from one another by their reliability as sources. Certain informants are acknowledged as experts concerning some propositions, whose opinion on these matters is supported by some relevant factor such as experience, access to an exclusive domain of life, formal position, and so on, that the other informants lack. The result would be that on some propositions, not every informant would have an equal say on the truth of the matter; some votes would count for more or less than other votes, based on their capacity to know the truth of the matter.
This solution is intuitively palatable in social research. If, for example, one wanted to know about some long-standing cultural tradition of a tribe, it is far more likely that tribal elders would know of these traditions and understand their meaning and purpose to a greater extent than children in the tribe. So, it would be more advantageous for the anthropologist to consult elders on traditions than to consult children.
Anthropologist of Western Apache, Keith Basso, discovered a linguistic phenomenon of 'wise words' in use among certain tribe members, but known to most of the non-wise words using tribespeople. Western Apache wise words are the usage of metaphors that compare undesirable behavior in some people to the perceived habits of some natural thing, such as animals like the coyote or a force like lightning. 28 Basso found 27 List and Pettit regarded this relaxing as the most promising for effective group organization. Group agency, 55. that the speakers of wise words, "who form collectively a kind of intellectual elite, are typically well along in years and because of their advanced age are not expected to participate in the full round of daily activity that occupy most younger members of Western Apache society." 29 While all Western Apache can understand or be made to understand these metaphors, typically, only elders are capable of creating new wise words that catch on with the rest of the tribe. One informant told Basso at the end of his stay among the tribe, It's too bad that you didn't try to learn about wise words before. When I was young, old people around here used to make them up all the time. Only a few of them did it and they were the best talkers of all. Some people would try but they couldn't do it so they stopped trying.... Only the good talkers can make them up like that. They are the ones who really speak Apache.... I don't know how they do it. It's something special that they know. 30 In learning about wise words, Basso would have gotten the most complete understanding of the practice from the elders capable of formulating them, and so could have safely prioritized testimony from those people and treated other Apache testimony as lesser in importance, though not necessarily altogether irrelevant. Where informants are clearly in a privileged position of knowledge, anthropologists could capture the cultural or social data accurately by relying on these expert accounts.
The second alternative anthropologists could pursue is by implementing the priority strategy. In coming to understand a complex cultural or social phenomenon, anthropologists could divide their inquiry into atomic propositions that they then assign each a level of importance. 31 This allows the anthropologist to make sense of disparate interpretations by targeting the inquiries that would likely achieve the most consensuses among the informants, and using those truths to derive the truths of the propositions of secondary importance. Continuing with the example of wise words, Basso could have devised a scheme of two premises and one concluding proposition that would gain him insight into the practice. In understanding how they work, he might ask the informants if "Wise words compare people to natural things," "Wise words compare unfavorably," and "If wise words compare people to natural things, then wise words compare unfavorably." Basso could have called the first 29 Ibid., 100. 30 Ibid., 118. 31 Group agency, 56.

| eP1177
Res Cogitans two propositions premises, and the last proposition the conclusion. He might have decided that the premises were the more significant questions or the most likely to provide unambiguous answers, and so used the responses to the premises to deduce the truth values of the conclusion, regardless of what responses he got to it. As we know that the two premises are true, and unambiguously so, there is little chance that informants would throw Basso into doubt by giving unintentionally false answers, whereas the conclusion prioritized might have left open such misinterpretation.
The third and final strategy I propose anthropologists have recourse allows them to cleave to the interpretive method, but while making use of a formal distinction between different collective attitudes. List typified three separate collections of judgments, which are often conflated and cause confusion "in the social sciences" 32 These attitudes have different properties that relate to the degree of unity in judgment between the members and so can be used as a shorthand for the potential accuracy of the aggregate judgments. The types are "aggregate," "common," and "corporate" attitudes. He explains that aggregate attitudes "need not be more than constructs made by an observer," and that ascriptions can be made "to collectives independently of their agential status." 33 Aggregate attitudes are like summaries of collective attitudes that are legitimized through the use of an aggregation rule. 34 An ascription could be like, "The (x) tribe believes in the (y) deity," ascertainable by observing the majority of the tribe engaged in a worshiping of the deity. List distinguishes two types of aggregation; 35 the relevant one here is the second, behavioural aggregation, which is when "the aggregate attitudes are determined, not directly as a function of the corresponding individual attitudes, but indirectly, as an 'emergent' property of the individuals' patterns of behaviour, which, in turn, may reveal their attitudes." Thus witnessing tribespeople bowing before an idol could be taken as evidence for a behavioural aggregation judgment of the tribe's belief in said deity. For the sake of comparison, I will mention one more collective attitude. In contrast to aggregate attitudes, corporate attitudes are not easily reducible to the beliefs of the members and cannot be ascribed so easily, if at all, on the basis of observation alone. 36 Ascriptions require a shared consciousness about the collective on the part commons.pacificu.edu/rescogitans eP1177 | 11 of the members that is identical with group agency. The members usually explicitly decide attitudes. The high level of organization of such groups makes judgments concerning them much stronger, but the kinds of groups that corporate attitudes to be assigned to are restricted.

Conclusion
In this paper, I have explained the theory of judgment aggregation and explored the ways that anthropology has hitherto dealt with its concepts. After critiquing the practice of anthropology using List and Pettit's ideas, I proposed some methods that anthropologists could implement to more accurately represent the views of their studied subjects.