For more than 40 years, numerous studies have shown that people seem to value a good they possess more than an equivalent good they do not possess. This endowment effect (Thaler, 1980) is observed in particular in the economic and psychological literature through the Exchange Paradigm (from now on designated as EP). The EP consists of two stages: The experimenter first checks whether a group of participants are indifferent to possessing an object A or to possessing an object B. Then, the remaining participants in the study are randomly endowed with one of the two equivalent goods A and B and after a time (in which the participant does some work or exercise) are then given the opportunity to exchange the endowed good for the other good that they did not receive. The endowment effect is reflected in a refusal to exchange the received good.

This refusal to exchange is observed in young children (Harbaugh et al., 2001; Horowitz and McConnell, 2002; Lucas et al., 2008; Da Silva et al., 2014) and also in certain non-human primates (Lakshminaryanan et al., 2008; Kanngiesser et al., 2011; Brosnan et al., 2012; Flemming et al., 2012, to cite a few). It is considered a rationality bias with respect to the prescription of standard economic theory which states that individuals' preferences over goods are independent of whether or not they posses it. In EP, the endowment of one of the two goods should not change the initial lack of preference of individuals between the two goods. Thus, if an individual has no preference between good A and good B and they own one of them, they should have no preference between keeping it or trading it. In EP, we should find a similar proportion of participants agreeing or refusing to exchange the received good for the other when they initially have no preference between the two objects. The endowment effect reflects this “exchange asymmetry” (Plott and Zeiler, 2007; Marzilli Ericson and Fuster, 2014) in favor of the given good.

The traditional explanation of this behavior is given by prospect theory and in particular by the notion of loss aversion (Kahneman et al., 1990; Tversky and Kahneman, 1991). Psychologically, a gain triggers a feeling of satisfaction, a loss triggers a feeling of dissatisfaction. Comparing the satisfaction level for a gain of a value X to the level of dissatisfaction triggered by the loss of a value of −X we observe that the level of dissatisfaction is higher than that of the satisfaction (Thaler, 1980; Knetsch, 1989; Kahneman et al., 1990). According to some authors, loss aversion would be the result of an evolutionary process based on the importance of overvaluing these goods when bargaining to acquire more resources. This would allow the maintenance of more offspring than people who correctly value (or undervalue) what they possess (Heifetz and Segev, 2004; Huck et al., 2005). This would be caused by an ancestral impulse to “defend one's own territory.” Authors suggest there might be a genetic predisposition to acquire and to preserve goods (food, shelters, tools and territory) which allows the individual and collective survival (Stake, 2004; Gintis, 2007)¹. An alternative explanation of a psychological nature is given with the singular status of the feeling of ownership of an object. The endowment effect would be caused by a simple possession effect which would favor a strong attachment and a privileged relationship with the object (see for a review Morewedge and Giblin, 2015).

To our knowledge, in the literature, few studies have emphasized that EP involves a particular interaction between two protagonists. In a study on children 4 years old, Lucas et al. (2008) found a link between the endowment effect and success in the theory of mind (TOM) task. It is thus possible that there is a link between the endowment effect and the participant's interpretation of the experimenter's intentions and beliefs. In order to understand participants' interpretations, we must first understand what are these interactions. In EP, an experimenter hands out an object and a recipient receives it. The experimenter is very often a teacher and the recipients are students. The relationship governing the interaction between them is hierarchical, it is unbalanced. The social status of the two people is not the same. The first phase of interaction is the experimenter's endowment to the participants. This interaction certainly has a strong contextual effect. It is rare for a professor to give something to their students in a university course and this certainly surprises the participants. Moreover, we can observe that all participants accept this gift. The second phase of interaction is the request for an exchange. This request is also unusual, at least in Western customs. Without making a pragmatic analysis of the task, Plott and Zeiler (2007); Klass and Zeiler (2013) hypothesize that the participants consider the object given to them “is a gift from the experimenter, even though the experimenter might simply intend to convey that subjects now own [the object]” (Plott and Zeiler, 2007, p. 1454). This interpretation of the endowed object as a present given by the experimenter would lead them to prefer this object to another and to feel obliged to refuse the exchange. They observe an absence of endowment effect when the endowed object is not physically given (see also Knetsch and Wong, 2009) or is given randomly in concordance with the other object. In turn, they observe a strong endowment effect when the experimenter explicitly gives the object as a gift. Many of the results in the literature can be explained using this interpretation.

In EP, the endowment effect also seems to disappear when the goods given and exchanged are goods of exchange, such as money (Svirsky, 2014). Moreover, individuals experienced in markets seem less subject to the endowment effect (List, 2003; Engelmann and Hollard, 2010). However, exchange goods such as money are not usually used as gifts, and market professionals who are used to receiving goods and reselling them are certainly less inclined to consider the endowment as a present. Finally, in the only study we know of using EP with a non-Western population, Apicella et al. (2014) show that members of hunter-gatherer tribes (Hadza Bushmen of Northern Tanzania) who are highly exposed to modern society and markets are reluctant to trade in EP. Hadza with little exposure to modern society and markets do not show the same reluctance to trade. It is possible to analyze these observed differences as a different social response to receiving a gift from the Western experimenter². The Hadza who have experience of Western social norms would behave like Westerners when they receive a gift the Hadza who are not familiar with the customs of the Western world would behave according to their own social norms by exchanging and sharing it with the rest of the tribe (see for example Marlowe, 2004). Thus, we support the hypothesis of Plott and Zeiler (2007) that EP would encourage participants to represent the question of exchange in terms of whether or not to exchange a gift received.

As noted by Horak (2018), the vast majority of cross-cultural economic experiments provide evidence that culture can moderate economic behavior, but few studies explain the reasons for this behavioral difference across societies in the field of decision-making. We argue that culture, through the social norms of individuals, shapes representations and implicatures which differ on the expectations of the experimenter, but that the underlying cognitive constructs are the same for all individuals. More generally, EP illustrates singular situations of interaction between the experimenter and the participant which are giving and exchanging. Such a paradigm is not only interesting for psychologists and economists but also represents an emblematic paradigm for anthropologists. Indeed EP illustrates experimentally the anthropological paradigm of the “gift” with its triptic “giving,” “receiving,” and “giving back” (discussed since Mauss, 1924; Malinowski, 2018) which can be easily studied in different cultures. Indeed, Mauss (1924) in his study of Maori gift giving (the “spirit of the gift”) suggests that, in many societies, the objects given are ultimately “inalienable,” that is, they cannot be entirely detached from the giver, but carry within them something of the personality of that person. They carry significant implications for the future actions of the receiver and for the relationship between them. This originality of Melanesian societies was observed by Henrich et al. (2005) in a study of the ultimatum game³ among the Au and Gnau of Papua New Guinea in which the majority of participants reject the seemingly generous offers of the proposer when it represents more than 50%. The explanation provided by Henrich et al. (2005) is as follows:

This specificity of gift and return-gift (reciprocity norm), as a key factor in social interactions, present in all Melanesian societies (Godin, 2015; Tcherkézoff, 2016), certainly produces a different interpretation of the experimenter's gift than that of the Western participants gift. In an EP effected in a Melanesian context, a Melanesian participant who receives an object from a Melanesian experimenter will also tend to perceive this endowment as a gift. However, this gift is not perceived as “a present” (as it is for Westerners) but as the source of the exchange that allows the interaction between the two individuals (here the experimenter and the participant) to come alive. It will be the same later in the question of exchange with an equivalent object, the participant can either accept it or refuse the exchange. We assume that there is no endowment effect in EP if the interaction involves two Melanesian individuals.

The Kanak of New Caledonia in the South Pacific is a native Melanesiane population located for the most part in the Northern Province and in the province of the Loyalty Islands. Kanak society has several levels of customary authority, from the 4,000 or 5,000 family clans to the eight customary areas that make up the territory. At the head of clans are clan chiefs and clans constitute 341 tribes, with a chief at the head of each one. The tribes are then grouped into 57 customary chiefdoms, with chiefs at their head, and forming the administrative subdivisions of the customary areas. In Kanak society, giving and exchanging play a primordial role in private interactions but also in the structure of society, notably with the ceremony of the custom. The Kanak population is all the more interesting to study because it is partially bicultural. Thus, the Kanak are also perfectly familiar with all the social codes of French Western society (notably through the obligation for all French children from the age of 3 to attend the school of the French Republic). The cross-cultural studies on Kanak populations are rare, however we can note Jamet et al. (2014).

After a pragmatic analysis of the possible implicatures of Western participants confronted with EP, we will explain the particularity of Kanak society and why Kanak participants should not be subjected to the endowment effect in EP. In our experiment we will distinguish the French Western context and the Kanak context in order to look at the adaptive capacity of the Kanak between the two cultures.

The literature on reasoning and decision making offers numerous examples in which behaviors or responses given by participants, initially judged to be erroneous, reveal a coherence with respect to the inferred representation of the participants to the requested task. These representations can be explained by the different pragmatic implicatures coming from the violations of the conversational maxims of cooperation of Grice (1975) (see Dulany and Hilton, 1991; Schwarz et al., 1991; Sperber et al., 1995; Baratgin and Noveck, 2000; Macchi, 2000; Politzer and Macchi, 2000; Baratgin, 2002, 2009; Bagassi and Macchi, 2006; Baratgin and Politzer, 2006, 2007, 2010; Macchi and Bagassi, 2012; Politzer, 2016; Macchi et al., 2019, 2020; Bagassi et al., 2020; Baratgin et al., 2020, for examples). The experimental paradigms are constructed through speech acts and the gestures of the experimenter and are, as in any communication fact, pragmatic in nature (Sperber and Wilson, 1986, 2002). The participants make assumptions about the experimenter's expectations and use the simplest interpretation procedure which consists in inferring from the communicative stimulus the intention most relevant to their own point of view Sperber (1994). However, what is relevant to the participant may be different from what the experimenter actually intends to communicate. Questions are relevant when they lead the person to whom they are directed to respond in a relevant way (i.e., questions that require the least cognitive cost for the most contextual effect). The exchange question in EP, can be interpreted differently depending on the social norms of the participants. In particular, we will be interested in the different interpretations of the question of exchange in a Western population (French population of Metropolitan France) and in a Kanak population in the South Pacific. The issue of social norms in the act of giving is important to understand. Before looking specifically at the issues of gift and exchange in French and Kanak society, we will briefly review the various theories of social norms of politeness.

In social interactions, norms of politeness are crucial. They help reinforce collaboration through rules that are known to all members of society. These rules lead to implicit expectations in actions and words from people interacting together (Geraci, 2020; Geraci and Franchin, 2021; Geraci et al., 2021). Since Leech (1983), it is admitted that communicative exchanges are subjected to the fundamental postulate “to be polite” in our interactions. He proposes a series of maxims and sub-maxims based on the “cost” and the “benefit” in relation to the interaction between self and other. Thus, if the benefit is higher to the receiver than the cost, then it seems to be more polite. On the contrary, if the cost is higher than the benefit to the recipient, then it seems to be less polite.

The model of politeness theory described by Brown and Levinson (1987) is the reference for most studies on politeness (for a critique see Fraser, 1990; Watts, 2003). The model is also intended to be “universal.”⁴ The authors use the notions of face from (Goffman, 1967, 1971) to define their pragmatic theory of politeness. The face is: “the positive social value a person effectively claims for himself by the line others assume he has taken during a particular contact” (Goffman, 1967, p. 267). Brown and Levinson (1987) maintain that each individual has two types of face; a negative face and a positive face. The positive face is the individual's desire to be appreciated and approved in their social interaction (valuing the image of the individual) and the negative face is their desire for freedom of action and to impose their ideas and desires (protection or defense of “[their] territory”). Brown and Levinson (1978, 1987), also use the hypothesis of Goffman (1967) that, for an individual, any interaction with others is a potential source of conflict. The rules of politeness correspond to ritual constraints, codes and strategies which aim at preserving the face (positive or negative for Brown and Levinson, 1987) of the interlocutors by avoiding conflicts. These norms thus assume the essential functions of facilitation and regulation of social interactions and are, following a learning phase at the youngest age, strongly internalized (Talwar et al., 2007). Throughout the course of the interaction, the interlocutors perform a number of verbal or non-verbal acts that potentially constitute threats to the positive or negative face of both interlocutors. The objective for the participants is to reduce as much as possible these Face Threatening Acts (FTA).

The aim of this study was to propose a new explanation for the endowment effect observed in EP. The endowment effect would be due to the respect of the social norms in force in the individuals' society. Our results seem to be coherent with this hypothesis. Kanak participants are subject to the endowment effect only when the context of the experiment involves interaction with a French experimenter and in a communication made in French (WCK). The endowment effect found is comparable to the one obtained under the same conditions with French participants from metropolitan France (WCF). On the other hand, when the experiment involves an interaction between Kanak (experimenter and participants), in a communication expressed in the vernacular language of the participant, the endowment effect disappears (KCF). Acceptances and refusals of the exchange are balanced in the same proportions as the participants' preferences of the two objects. It is important to note that this change in behavior is observed while EP is performed with Western objects, in a Western location and without any other explicit information. The lack of an endowment effect found is comparable to the situation in which the ceremonial context of Kanak exchange is accentuated by the experimenter's words and the performance is conducted in the tribe with prototypical customary objects (KCK). These results have several important implications, not only for understanding the endowment effect found in EP by Knetsch (1989) but also on other aspects discussed below.

These results corroborate the criticisms of Plott and Zeiler (2007) on the traditional explanation of loss aversion (Thaler, 1980; Knetsch, 1989; Kahneman et al., 1990). But they are also in dissonance with alternative explanations, whether it be the evolutionary one of a defense of the territory (Heifetz and Segev, 2004; Huck et al., 2005) or those stemming from a particular attachment to the object (see for a review Morewedge and Giblin, 2015). Indeed, the Kanak participant does not seem to have any aversion to exchanging the object in the KCF or KCK condition and does not show any indication of a particular desire to own it. In the Western context, explanations in terms of loss aversion and territorial defense seemed relevant. In fact, our explanation was not visible because it blended with these ones, since the model of Brown and Levinson (1978, 1987) explains the refusal of the exchange with both loss aversion or defense of territory. It is the use of a different social context that distinguishes all these explanations. Our study is a new example of the importance, for cognitive psychology, to take into consideration the points of view coming from other disciplines like anthropology (Sperber and Hirschfeld, 1999). They illustrate, indeed, the ideas of Mauss (1924); Malinowski (2018) on the exchange of gifts in many traditional societies in the World, especially in Oceania.

With this experiment, we provide new experimental arguments in favor of a pragmatic explanation of the endowment effect observed in EP. Culture, and in particular the social norms of individuals, shapes the pragmatic interpretation of the experimenter's offer as a gift (as Plott and Zeiler, 2007, were the first to make the hypothesis). In a French context, the least onerous interpretation of this unusual action is to consider the object as a present. In the Kanak context, the interpretation of the gesture of giving is different, it is an introductory exchange in order to build or consolidate a social bond; it is a usual (even anodyne) gesture in the Kanak world. The proposal of exchange in the French context is confusing. It illustrates a violation of the social rules expected after the offering of a present. It causes an updating of the implicatures of the participant regarding their expectations of the experimenter which results in a refusal of the exchange. In the Kanak context, the experimenter's exchange proposal does not cause an update of the Kanak participant's implicatures. The latter continues to infer that the experimenter wishes to further strengthen the social bond. There is no obligation to refuse this proposal, but one can also decide that the exchange will take place later with another object (of higher value).

According to this explanation, the refusal to exchange in the French context should not be interpreted as a lack of rationality on the part of the participant. Similarly, the Kanak participant in the Kanak context, who does not produce the endowment effect, should not be interpreted as behaving more rationally. In both situations the participant's decision is consistent with their implicatures and representations of the experimenter's expectations. For participants the value of the object is not the same because the object does not have the same meaning and value. When they are asked for their preference between two equivalent materiel objects in P_Group, the objects are simply physical objects (objet_physical). Then, depending on the French or Kanak context, they have either the status of a present (objet_present), or the status of an exchangeable good serving as a social link (objet_social). They symbolically represent the gift that must continually nourish and recreate the social bond, precisely because what circulates (when and if it circulates) is the result, not the cause, of the social bond itself (Godbout, 2009). In all these situations, we are exactly as in the Bayesian situation of a different subjective probability judgment of the same event according to the context (Baratgin, 2002, 2009; Baratgin and Politzer, 2006, 2007, 2010).

This observation of the variability of the status of the object by the participant according to the cultural and social context is to our knowledge a new experimental result. This result illustrates Searle and Willis (1995) definition of “social objects”: Social objects are created by the fact that we consider or count a physical object as something that goes beyond the physical structure of that object, thus conferring on it a social status in a certain context - for example, by virtue of collective recognition, a piece of paper counts as a fifty-euro bill in the context of the economy. In the Western French context of EP, the object is considered by the Kanak participant as object_present which is difficult for them to exchange, whereas it becomes object_social, in the Kanak context [this absence of an endowment effect when the object is perceived as an exchange good is, moreover, in line with the observations of Svirsky (2014) of an absence of endowment effect for money]. The gift of the object in Kanak society (as analyzed by Mauss, 1924) can be interpreted as an institution (in the sense of what makes the cohesion of society as defined by Searle and Willis, 1995) resulting from the self-transcendence of the social relations that the gifts themselves are expressly designed to create and according to which individuals orient their behavior (Cedrini et al., 2020). For a renewed reading of Mauss's work on exchanges (Tcherkézoff, 2016).

Our results also offer a new example of the flexibility of bicultural individuals, observed in other contexts (Gardner, 1985; LaFromboise et al., 1993). In particular, it corroborates the results found by Chuah et al. (2007, 2009) in the ultimatum game. The study compared the decision-making behavior of participants from Malaysia and Great Britain, taking into account the location of decision-making. It was observed that the amount offered was generally higher in the Malaysian treatment group (Malaysian offerers and responders) than in the British treatment group (British offerers and responders). However, when the groups were crossed, the Malaysian proposers generally offered lower amounts to the British but not to the other Malays. The British, however, did not change their behavior.

Finally, our results argue that, contrary to the assumption of economic theory that rational agents are self-interested, individuals' decision making is strongly influenced by social interactions such as social concerns for justice, fairness, and reciprocity (Gouldner, 1960; Henrich et al., 2005; Fiddick et al., 2013; Geraci, 2020; Culpeper and Tantucci, 2021; Geraci and Franchin, 2021; Geraci et al., 2021). In particular, our work is further evidence of the need to broaden the range of regions for cross-cultural investigation for cognitive psychology and experimental economics (Henrich et al., 2010; Masuda et al., 2020). This opening should also be done for the study of populations from holistic societies, the great majority of which are from Asian countries (see for example Masuda and Nisbett, 2001; Nisbett et al., 2001; Nisbett and Masuda, 2003; Nisbett and Miyamoto, 2005; Choi et al., 2007; Yama et al., 2007; Nakamura et al., 2018).

The essential proposition that has been developed and tested in our study is that the answer to the offer of exchange in EP crucially depends on the social norms at play in the contextualized interactions between experimenter and participant. In EP, two key elements of social interaction shared by all human societies are brought into play: gift and exchange. This paradigm, which seems disconcertingly simple, is much more than an experimental paradigm that allows us to evaluate the endowment effect. It is a paradigm offering the possibility of understanding the core of social interactions in all human societies. In this paper, EP makes it possible to account experimentally, for the first time in psychology, for the particularity of exchange-based interactions between Kanak. It may also allow for the study of more detailed predictions made by anthropologists on interactions linked to filiation (Godin, 2015, 2018).

The Kanak, who are partially bicultural, show a flexibility, depending on the context, to give a response either in accordance with French social norms of politeness or in accordance with Kanak social norms of exchange (gift and return-gift). This result, however, may seem to contradict certain results in the literature. First of all, the behaviors, similar to the endowment effect, observed in certain primates (Lakshminaryanan et al., 2008; Kanngiesser et al., 2011; Brosnan et al., 2012; Flemming et al., 2012) cannot be explained in terms of the social norms of politeness (of a Western society). It is the same to explain the appetence of very young children (2 years old) to keep an object that they have just received (Gelman et al., 2012; Hood et al., 2016). Indeed the concept of property and especially that of transfer of property, are completely acquired only from 4 to 5 years old (Blake and Harris, 2009, 2011; Nancekivell et al., 2013; Davoodi et al., 2020). This is the age when children, unlike apes, respect property as a cooperative arrangement, in which they inhibit their tendency to take the property of others on condition that others do the same (Kanngiesser et al., 2020). The acquisition of social norms starts from the age of 3 years old (Schmidt et al., 2016), but the norms of politeness, seem to be acquired even later (Axia and Baroni, 1985; Baroni and Axia, 1989). One can thus probably think that the endowment effect is part of a developmental trajectory and would take two forms. The first one, “primitive,” in primates and young human primates, can be explained by an evolutionary justification (Bruner et al., 2020). The second, more sophisticated, depends on the pragmatic abilities and the capacity toward a theory of mind of the individual and manifests itself by behaving accordingly to the specific social norms of the society in which one lives. This hypothesis makes it essential to reproduce our study with Kanak children of several ages. A very recent study (Prou, 2021) indicates an endowment effect in Kanak children aged 4-5. However, this study does not allow any conclusions to be drawn because it was conducted in a French context (nursery school with a French teacher/experimenter and in French). The study carried out among older Kanak children aged 6–7, conducted in a tribe, by a Kanak experimenter and in the Kanak language, indicates a reverse endowment effect (the vast majority of children accept the exchange (Jamet et al., 2017a,b; Jamet and Baratgin, 2018). This result can be explained by the identification of children with their cultural group, which leads them to imitate in excess the behavior observed in adults. The next studies should be conducted under a comparative life span approach with the two different contexts.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Mr. Maudinet Marc, PhD in anthropology, Former pedagogical director of the master Gestion et Politiques du Handicap Sciences Politique of Paris, freelance councilor expert in the European Council, president of the scientific council of the FISAF. Mr. Deberge Dominique, retired professor of economy and management, retired education psycho-sociologist, Noumea, New Caledonia. Ms. Wanguene Marie-Louise, pedagogic councilor FELP, deputy mayor of the commune of Hienghène, Haut-Coulna tribe, New Caledonia. Mr. Lionel Zannier, manager of the formation of specialized teachers in the institute of formation of teachers of New Caledonia, Noumea, New Caledonia. The patients/participants provided their written informed consent to participate in this study.

JB and FJ: conceptual elaboration and design of the study. FJ and PG: data collection. JB: data analysis and draft of the manuscript. JB, FJ, and PG: critical revision of the manuscript. All authors contributed to the article and approved the submitted version.

We thank the P-A-R-I-S Association for the technical and financial help we received as well as the CHArt laboratory which participated in financing the publication of the article in open access.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.