Semantic stretch in the marketplace of ideas: A tragedy of the verbal commons?

We argue that language is a shared social resource that people may be tempted to stretch in pursuit of their own persuasive goals (e.g., describing something that is unusual as “unique”). We demonstrate this tendency toward semantic stretch in the laboratory and present four field studies that document the process at the cultural level--involving emotionally laden terms such as synonyms of good, scarcity terms such as unique, and high-status scientific terms such as relativity and negative reinforcement. We provide evidence that semantic stretch may eventually undermine the meaning of the stretched terms, leading to a kind of commons dilemma for semantic meaning. Semantic stretch in the marketplace of ideas: A tragedy of the verbal commons? Research on cross-cultural psychology has emphasized how culture alters the psychological processes of individuals (Markus & Kitiyama, 1991). It has shown that people in different cultures reason differently about such basic issues as how to assign causal priority to personality versus situational causes of behavior (Morris & Peng, 1994), deal with apparent contradictions (Peng & Nisbett, 1999), perceive individuals versus collectives as the source of agency (Morris, Menon, Ames, 2001), respond to the presence of choice (Iyengar & Lepper, 2000), or react to perceived threats to their honor (Cohen, 1998; Cohen & Nisbett, 1994), But while psychologists have studied how culture influences psychology they haven’t studied how psychology influences culture, ceding to other social sciences-history, sociology, anthropology--most questions about cultural content. Some researchers have insisted that psychologists have ceded too much. In the words of Dan Sperber (1990, p. 42), “Culture is the precipitate of cognition and communication in a human population.” In a recent book, The Psychological Foundations of Culture (edited by Schaller & Crandall, 2003) researchers explain how the tools of social psychology may help explain how norms, beliefs, and traditions emerge, evolve, and change. “If we are to fully understand culture, we must know the essentially psychological ‘natural forces’ through which individuals unintentionally create, sustain, and change the cultures that they comprise” (Schaller, Conway, and Crandall, 2003, p. 4). In this paper, we describe how social psychological processes affect one of the defining features of culture: language. While evidence for the extreme version of the Sapir-Whorf hypothesis is lacking (language doesn’t affect what can be thought), it does alter how easily things can be thought and, certainly, how easily they can be communicated (Lau, Lee, Chiu, 2003). Thus language constitutes one of the most important cultural resources shared by social groups. We argue that as a shared social resource, language is subject to the same commons dilemma that has been widely studied for other kinds of social resources (e.g., Kerr, 1989, 1995; Messick & Brewer, 1983; DeCremer & van Dijk, 2002). Individuals, in their attempts to persuade others, may be tempted to try to stretch the meaning of shared terms, particularly those that have some emotional resonance. Consider, for example, the following two (unsuccessful) persuasive appeals: Rashomon can be seen as a cinematic extension of Einstein's theory of relativity The Dallas Morning News, September 7, 1998 "We look at getting rid of the Confederacy as cultural genocide. They're attacking my ancestors, they're attacking my culture, they're attacking my heritage." –Mr. Robert Banks, Member of Sons of the Confederate Veterans Atlanta Journal, December 14, 1992 It is a stretch to claim that Rashomon extends Einstein’s theory of relativity, or that modern verbal attacks on the Confederacy constitute cultural genocide. But it is clear what the speakers intended in these examples, though they failed--they were trying to advance their own agenda by “borrowing” the emotional significance of the words relativity and genocide. They failed because the focal terms are too heavily embedded in other shared associations for listeners to allow them to use the terms in those ways. Below, however, we describe why similar attempts to persuade by many individuals may combine to create a kind of commons dilemma for terms and concepts that are consistently evoked in persuasion attempts. The psychology of cultural content and cultural change While it is unusual to take a psychological approach to cultural content, it is not unprecedented. David Rubin (1995) describes how the cognitive structure of memory affects the content of oral traditions such as sagas, ballads, and children’s counting-out out rhymes. He argues, for example, that sagas tend to focus on concrete, visualizeable actions because people find it easier to remember events that are concrete and visualizable. Homer is filled with concrete action, not because the Greeks had trouble with abstraction, but because the structure of human memory leads concrete actions to be more likely to survive generation after generation of oral transmission. Norenzayan and Atran (2003) argue that nonnatural beliefs (e.g. supernatural agents or talking animals) may be common in folklore because they make it more memorable. While mundane details help a story make sense, a certain proportion of minimally counterintuitive features add spice to a narrative and make it more memorable as a whole. Schaller, Conway, and Tanchuk (2002) studied the content of cultural stereotypes of ethnic groups, and show that they are more likely to persist unchanged--in one sample they analyzed cultural stereotypes in the United States across four decades--when individuals find them easier to communicate about. Most definitions of culture identify particular kinds of cultural content--norms, principle, symbols, values, practices--that are shared (Mead, 1955; Geertz, 1973, Schaller, Conway & Crandall, 2003). A few areas of psychology have studied how social practices and perceptions come to be shared at the level of groups (e.g., Thompson, Levine, & Messick, 1999; cite Moreland & Levine) or society (Farr & Moscovici, 1984). While psychologists have a long history of disciplinary skepticism about whether it is meaningful to talk about social groups at a level above the individuals who compose them (e.g., see the review of the history of the “group mind” concept in Wegner, Giuliano, & Hertel, 1985), it is clear that cultural content can persist even when the individuals who compose a group have experienced complete turnover (Sherif, 19xx? , Jacobs & Campbell, 1961; Weick & Gilfallan, 1971). Social psychologists have recently studied how the “culture of honor” in the American South has persisted over long periods of time despite great economic, demographic and social changes (Cohen, 1998; Cohen & Nisbett, 1994). While psychological studies of cultural content are rare, even rarer are studies of cultural change. Researchers have criticized the tendency to treat culture as static (McIntyre, Lyons, Clark, & Kashima, 2003; Adams & Markus, 2003), but few studies have studied why and how cultural practices change (for exceptions, see Kashima, 2000; McIntyre, Lyons, Clark, & Kashima, 2003). In part, this may be because it is often difficult to describe how a process of culture that requires meaning to be shared among many people can simultaneously allow for change. Below we do this by pointing out how individuals, pursuing their own desires to influence, mobilize, and persuade, may undermine a shared cultural resource thereby causing it to change over time. A commons dilemma for language Psychologists have studied how people handle conflict between individual and group interests in the literature on social dilemmas (e.g., see collections by Schroeder, 1995; Liebrand, Messick, & Wilke, 1990). Most of this research has studied actual, tangible resources (operationalized as concrete pools of money in experiment) but a few researchers have noted that logic of commons dilemma applies to less tangible resources (information dilemma of Sherry Schneider, Rutte, 1990). Rutte (1990) asks us to imagine a shared pathology lab at a major hospital that generally conducts lab tests in the order they arrive, but will perform them more rapidly if they are labeled “urgent.” Rutte notes that individual doctors and nurses have an incentive to stretch and overapply the “urgent” label to get their own tests done more quickly. This kind of stretch may erode the meaning of the urgent label over time--when everything becomes urgent, no one can easily signal that a test is particularly time-sensitive. Rutte’s example shows that shared meanings can set the stage for a commons dilemma in which some collective good is eroded through the instrumental overuse a large number of individuals (Hardin, 1968). We argue that people, when they are trying to persuade others, have an incentive to stretch--to overapply, to inappropriately enlarge or extend--the class of situations where emotion-laden terms are used. Many of us, for example, in an attempt to persuade a friend to try a restaurant we enjoyed, might trot out our favorite superlatives and describe it as “fantastic” or “the best,” even if it wasn’t actually the single best meal of our lives. We might be tempted to describe a scarce resource as “unique” because this makes it seem more important and valuable. We might be tempted to overapply a scientific or academic concept to establish our credentials as thoughtful people. By doing these things, we evoke the shared cognitive and emotional associations of a term to persuade others. The process of semantic stretch corresponds to the first half of the classic commons dilemma situation, in which individual actors have an incentive to overuse the collective good because they benefit in the short term from doing so. The second half of the classic commons dilemma occurs when the combined actions of individuals erode the collective good. Over time, individuals’ tendencies to engage in semantic stretch may undermine the meanings we can convey with our shared vocabulary. If many experiences are “fantastic,” then calling something “very good” may damn it with faint praise. There is anecdotal evidence that semantic stretch does create a commons dilemma. Language mavens often complain that meaning is trickling away from improperly used terms (Safire 1980, 1990). Within the academy, by the 1920s, Einstein and other physicists felt that the term “theory of relativity” had become stretched so that it was not quite as useful as a scientific term as it had once been; indeed, there was a (failed) collective movement in the 1920s to repair the original meaning by proposing a more restrictive term for the theory—e.g., not the “theory of relativity” but the “theory of invariants” (Bodanis, 2000, p. 260). Indeed, semantic stretch is likely to create a particularly subtle and insidious form of the commons dilemma. Many of the mechanisms that exacerbate commons dilemmas are present-e.g., people are less likely to worry about contributing to a commons dilemma when it doesn’t seem that their efforts could make a difference (Kerr & Kaufman-Gilliland, 1997) and because the benefits of cooperative behavior may take a long time to emerge... When people are uncertain about the characteristics of the dilemma they encounter (e.g., endowments, status of public good, impact of individual actions on the collective good), then their actions tend to be determined by the aspects of the dilemma about which they are certain (van Dijk, Wilke, Wilke, & Metman, 1999). And many of the solutions proposed for commons dilemmas are not easily to evoke--e.g., equality norms (Alison, etc..). Why does semantic stretch undermine meaning? Research on human memory has shown that when people learn more associations to a particular concept, they are noticeably slower at recalling any single association (e.g., Anderson & Bower, 1973). When semantic entrepreneurs use a particular concept in novel ways, individuals who are exposed to these novel uses may have more difficulty activating the original associations. If the novel associations are less emotionally intense than the original uses, then semantic stretch will reduce the emotional impact of the term. Shared meaning may also be undermined by simple habituation. In general, when people experience a stimulus that initially produces an extreme emotional reaction, they tend to adapt and react less intensely to it over time (Frederick & Loewenstein, 1999). If semantic entrepreneurs, repeat over and over, various emotionally loaded terms such as “excellent” or “unique” or “genocide,” then those words, once able to provoke a strong emotional reaction, are likely to lose at least some of their emotional force. Semantic change in linguistics. Linguists have, for many years, debated whether semantic change is regular and predictable (see Traugott & Dasher [2001] for the most recent and thorough review), but to our knowledge they have not discussed the process of semantic stretch we have described. Linguists have recognized the possibility that terms may lose semantic content--they label this effect as “fading and discoloration” or “bleaching” (Traugott & Dasher 2001, p. 59)--but they haven’t described specific mechanisms that produce fading or bleaching. One reason they may not have highlighted the particular mechanism we describe is that they typically concentrate on the motivations individuals have to communicate efficiently: to “be quick and easy” and to “be clear” (Traugott & Dasher 2001, p.59), rather than the motivations we highlight--to draw attention, to persuade, or to mobilize others. Study 1: Semantic stretch in the lab We wanted to assess whether semantic stretch would take place in social interactions, so we asked people to imagine a fairly minimal social situation in which they might use their powers of persuasion--a male classmate who they don’t know well asks about pizza place they ate at recently. We manipulated people’s desire to persuade the classmate to go to the restaurant by indicating that they either liked the classmate “a lot” or “don’t particularly like him.” This situation ranks on the low end of social situations that might encourage people to stretch-participants are not imagining that their jobs, grades, or relationships are hanging in the balance, they simply want to encourage an acquaintance to share a pleasant experience. We suspected that people who were trying to influence a classmate they liked would tend to use higher degrees of superlatives (e.g., fantastic rather than very good). We also suspected that people would stretch terms of scarcity (e.g., labeling an unusual menu as “unique”), so we gave them an opportunity to engage in this kind of stretch as well. Half the subjects read about a pizza place that served standard fare (e.g., pepperoni pizza) and half read about a pizza place that served more unusual entrees (e.g., Shrimp Scampi pizza). The question is whether people would be tempted to stretch their description of the unusual menu and label it as “unique.”