Transcript
Page 1: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

Research on Gossip: Taxonomy, Methods, and Future Directions

Eric K. FosterUniversity of Pennsylvania

A half century of gossip research from multiple disciplines is reviewed. Discussed aredefinitions of the construct; social, evolutionary, and personal functions of the practice;and data collection methods. Though people engage in the practice frequently, there hasbeen relatively little psychological research on gossip. The layperson’s understandingof the term is included in, but insufficient to encompass, definitions used by researchers.Most data are ethnographic and discursive, and few parametric data exist. The areacould benefit from better experimental methods and instruments. Neurobiological andsocial network analysis methods are promising foundations for further study. There arereal-world implications for understanding gossip. Strengthening gossip theory andresearch methods will beneficially inform the way we view the practice in context.

Virtually all of us frequently find ourselvesproducing, hearing, or otherwise participatingin evaluative comments about someone who isnot present in the conversation. It is often valu-able (and sometimes unavoidable) to be part ofsuch communications. To function efficiently ina complex social environment, humans requireinformation about those around them. But socialinterconnections are complex, and it is impos-sible to be present at many primary exchangesto absorb this kind of information directly.Thus, many people are eager to pick it upthrough an intermediary, whether or not theyhave the luxury and patience to confirm it latereither directly or indirectly. This phenomenon,of course, is called gossip. It is an importantsocial behavior that nearly everyone experi-ences, contributes to, and presumably intu-itively understands. The purpose of this articleis to review and summarize research on thisphenomenon and point to some promising waysto study it going forward.

A paradox of gossip is that it is ubiquitous,though there are numerous social sanctionsagainst it. Anthropologists and others have doc-umented its practice the world over (Besnier,1989; Gluckman, 1963; Haviland, 1977; Levin& Arluke, 1985; Loudon, 1961; Stirling, 1956).

Indeed, social anthropology is sometimes saidto be the social science of gossip. Yet, gossip(and rumor, which differs primarily by alwaysbeing speculative and sometimes pertaining toevents rather than people) has been denouncedfrom antiquity to the present. Most societieshave explicit sanctions against gossip, and nu-merous cautionary narratives demonstrate itsunwanted outcomes. “Whoever repeats gossiplacks understanding” is the admonition inEcclesiastes. “Be not a tale bearer,” Leviticuscautions. “Don’t gossip,” parents and teacherssimply warn children. Nevertheless, unless aconversationalist specifically draws attention tothe fact that gossiping is occurring, it is likely toproceed relatively unhindered (Yerkovich,1977). Obviously, for gossiping and the sanc-tions against it to coexist, there must be value inthe generation and consumption of gossip thatoutweighs the counterforce of the sanctions.

There is no denying that gossip, like rumor,“can steal illusions, wreck relationships, and stirup a cauldron of trouble” (Rosnow, 2001, p.203). Targets may be hurt by seeing how othersperceive their affairs, by distortion or manipu-lation of information, or by the violation ofprivate matters. Many ethical condemnations ofgossip revolve around presumed rules of pri-vacy. Bok (1983), for instance, sees gossip un-equivocally as morally indefensible because ofthis violation. A relevant historical trivium: Thedumbwaiter was invented so that servants couldnot overhear the most private familial and busi-ness affairs of their employers over dinner andthen “retail their masters’ business” (Hecht,

Eric K. Foster, Marketing Department, The WhartonSchool of Business, University of Pennsylvania.

Correspondence concerning this article should be ad-dressed to Eric K. Foster, 191 Presidential Boulevard PH6,Bala, PA 19004. E-mail: [email protected]

Review of General Psychology Copyright 2004 by the Educational Publishing Foundation2004, Vol. 8, No. 2, 78–99 1089-2680/04/$12.00 DOI: 10.1037/1089-2680.8.2.78

78

Page 2: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

1956, cited in Goffman, 1959, p. 213). Misin-formation is also at the heart of condemnationsof gossip. Harrington and Bielby (1995), forinstance, in their study of the effects of elec-tronic bulletin boards (precursors of Internetchat rooms) on facets of gossip, discovered thatonline discussions contain a greater demand forexplicit referencing of verifiable sources. Thesesources would compensate for the lack of otherconfirmatory indicators, such as status, exper-tise, interpersonal connections, direct access toinformation, and so forth, that are generallymore available in face-to-face conversationswith known players. Ayim (1994) also por-trayed gossipers as intent on establishing theveracity of gossip, although Blumberg (1972)remarked that extensive monitoring of gossipover time is unlikely and impractical. In eithercase, the inclination to gossip covertly, anony-mously, or vicariously betrays an awareness ofthe violation of privacy norms. Generally, peo-ple seek to guard themselves against the chargeof indiscretion when gossiping.

An interesting counterpoint to the gossip-violates-privacy charge was articulated bySchoeman (1994). He argued that gossip, farfrom violating privacy, is in fact consistent withprivacy norms because it attenuates direct andpublic conflict. People usually know they arebeing (or will be) talked about when deviatingfrom norms: “We all fully expect to be dis-cussed by others who know us, with no sense ofimpropriety” (Schoeman, 1994, p. 80), even ifwe prefer not to be. Gossip provides the of-fended with a subtle way to pressure the of-fender. A public figure, for instance, may capit-ulate to gossip for a transgression, yet be re-lieved that the affair did not “get into thepapers.” (In the case of positive gossip about aperson, the gossiper is spared being obsequiouswhile buoying the target’s reputation with a“third-party endorsement.”) Blumberg (1972),finding strong evidence that evaluations travel-ing through the network tend to stop short oftheir targets, concluded that a norm seems toexist “to keep people from learning too muchabout what others think of them” (p. 161). Wertand Salovey (2004), in this issue, also acknowl-edge that gossip affords one the benefits ofvarious veins of social comparison while avoid-ing the risks of embarrassment or confrontation.Thus, gossip, far from violating privacy, may beconstrued as a protective social norm. It is less

bruising when originating from people moresocially distant from the target and less likely tobe practiced by those closer to the target (who,as in the case of family members, might beexpected to be more direct in theircommunications).

How much do people gossip? Apart fromobvious individual differences, the answer de-pends greatly on how the term is defined. Dun-bar, Marriott, and Duncan (1997) sampled 45conversations every 30 s while eavesdroppingin public. They found that “social topics” con-cerning either those persons present or thirdparties were the purpose of conversation abouttwo thirds of the time among both men andwomen. Only a small proportion (less than 5%)of this conversational time was spent in criticalremarks about third parties (i.e., malicious gos-sip in the colloquial sense). Similarly, defininggossip as Dunbar et al. (1997) did, Emler (1994)reported that about 70% of conversation timeinvolved gossip.

Both of these estimates would certainly di-minish, however, if only social remarks aboutpeople not present were the criterion. In anearlier article, Emler (1990), limiting his defi-nition to gossip about absent persons, found thatnamed acquaintances were still the most com-mon topic of conversation after self-disclosure.Slade (1997) distinguished five types of infor-mal conversation and reported that about oneseventh (14%) of the time spent in workplacecoffee-break conversation consisted of gossip.However, because she limited her definition tonegative gossip, the percentage would presum-ably be higher with positive remarks included.There is little empirical evidence that womengossip more frequently than men (cf. Dunbar,1993a, 1994; Eckert, 1990; Foster, 2003; Havi-land, 1977; Levin & Arluke, 1985; Loudon,1961; Nevo, Nevo, & Derech-Zehavi, 1993); ingeneral, any reported differences between thesexes are small.

Thus, with current data, it is not easy to get aprecise fix on the amount of time that peoplespend gossiping. Nevertheless, it is reasonableto conclude that most people spend a nontrivialportion of their interpersonal time gossiping.Indeed, several writers have remarked that notto gossip (or not to respond to gossipy overtureswith at least a modicum of interest) is to bequickly marginalized from the local social fab-ric (Bergmann, 1993; Eggins & Slade, 1997;

79SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 3: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

Gluckman, 1963). On the other hand, a reputa-tion for too much gossip may also marginalize aperson (Bergmann, 1993; Gilmore, 1978; Yerk-ovich, 1977), perhaps because the security ofinformation passed to such a person is not as-sured, and the value of information learnedfrom him or her is questionable. Confining thedomain to the workplace, Kurland and Pelled(2000) postulated a curvilinear relationshipwhereby too much or too little gossiping mayadversely affect one’s referent (attractional)power. None of these authors presented empir-ical evidence for these relationships, however.

Psychology researchers have largely over-looked gossip. The volume of work on the topicis scant both in journals and, particularly, intextbooks. I have located only a single psychol-ogy textbook with gossip in the index (twopages on the topic, in passing). Nor is it listed inthe cumulative subject index of recent editionsof the venerable Handbook of Social Psychol-ogy (Gilbert, Fiske, & Lindzey, 1998; Lindzey& Aronson, 1970, 1985). The number of journalarticles has remained thin over the last 40 yearsas well. I gleaned references to the topic inPsycINFO, ERIC, and JSTOR since 1970 usingthe single keyword gossip; results are shown inFigure 1. The frequencies observed are quitesmall relative to any other activity that takes upso much of our daily interpersonal communica-tions. Nor, according to a Web search of “gossip

and syllabus,” is gossip a topic typically in-cluded in classroom materials (Table 1), al-though psychology students certainly seem fas-cinated when it is broached in class.

Defining the Gossip Construct

This special issue is testament to the fact thatgossip does not lend itself to simple formulaicdefinitions or uniform explanations. We all“know” what gossip is, but defining, identify-ing, and measuring it is a complex enterprise forpractical investigation. At the very least, theeveryday understanding of the term gossip isincluded in, but insufficient to encompass, theconstruct as used by researchers. Defining thecontent, circumstances, and functions of gossipwill help to put the research methods used tostudy it into perspective.

At the most general level, gossip behaviorincludes “idle talk” or “chit chat” about dailylife. Dunbar (2004), in this issue, defines itbroadly as conversation about social and per-sonal topics. In some feminist criticism, gossipis nearly synonymous with “women’s talk” ingeneral (E. B. Brown, 1990; Coates, 1988; Eg-gins & Slade, 1997; Jones, 1980; Spacks, 1982)or “girl talk” (Eckert, 1990). Rysman (1977)traced the etymology of the term as it refersspecifically to women. The parallels in men’sconversation—“shop talk,” “shooting the

Figure 1. Electronic searches for articles related to gossip, 1971–2000. JSTOR subject areas available were anthropology,education, and sociology; 37 journals were searched in JSTOR, including Social Psychology, Social Psychology Quarterly,Annual Review of Sociology, and Annual Review of Anthropology.

80 FOSTER

Page 4: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

breeze” (Fine & Rosnow, 1978), or “killingsome time together”—are essentially inter-changeable with what is being referred to asgossip when it is defined so generally as idle orsocial conversation.

Third Parties Not Present

More typically, most people would agree thatwhatever is included in the content of gossip,the label is justified primarily by the exchangeof information about absent third parties, inaccord with the popular meaning of the term.Most of us know what happens to a gossipyconversation when the target (or a relative orclose associate of the target as his or her proxy)enters within earshot (Bergmann, 1993; Yer-kovich, 1977). Besnier (1989) wrote that talkabout absent persons is enough to justify thelabel gossip. Hannerz (1967) used as his work-ing definition of gossip the private transmissionbetween “A and B talking about C” (p. 36). Sothe absence of the third party seems to be aminimal and, for many, sufficient requirement.

There are, to be sure, exceptions even to thissimple rule. Gluckman (1963) noted that gossipis sometimes (but only rarely) face to face withthe target. Children have been found to gossipin the presence of their target (Goodwin, 1982),and both Roy (1958) and Handelman (1973)

recorded instances of the same phenomenon inthe small workplace groups they respectivelystudied. The situations they described are usu-ally oppositional and require special social ne-gotiation strategies by the participants. For in-stance, the gossiper may need to avoid eyecontact or otherwise simulate the nonpresenceof the target. The clinicians Medini and Rosen-berg (1976) believed that when clients or psy-chotherapists disclose relevant truths aboutthemselves, it may be called gossip: gossipabout the self. For these authors, the definingfeature of gossip was that it contrasts the dis-crepancy between the public and the private life(cf. Goffman, 1959). It is the instrument “forthose who wish to know how life is lived behindthe social mask” (Medini & Rosenberg, 1976, p.462). Other instances in which the targets arepresent include certain culturally sanctionedpublic events. Gilmore (1978) reported on theCarnaval (or “fiesta of gossip”) in a small Span-ish community, where people take to the streetsin a giant masquerade and hurl “slander, vilifi-cation, and innuendo” openly at each other (p.93). Hollywood and private dinner club“roasts,” good-natured ribbings honoring peo-ple of notable achievement, seem to performessentially the same function: a public gossip-ing, but one made safe because of the ritualizednature and the obvious contrast with the re-

Table 1Internet Searches for Syllabi on Topics Related to Gossip

Syllabus type

Approximate frequency

November 2001 February 2004 Increase (%)

Social learning 181,000 699,000 286Networking 40,800 122,000 199Interpersonal process 20,000 44,400 122Conversation style 16,000 30,600 91Social reputation 10,700 23,800 122Interpersonal relations 8,600 19,900 131Interpersonal communications 8,300 18,500 123Personality style theory 6,800 27,000 297Gossip or gossiping 3,000 5,770 92Rumor 1,800 3,240 80Grapevine 1,300 2,760 112

Note. These searches consisted of the topic words listed here plus the word syllabus. Thecontent of many of the hits is far off base, so the absolute numbers shown are no doubt greatlyinflated from a research point of view. Relatively, however, the numbers are reliable, becausethey are roughly replicable from different sources, and their ranks changed little over the27-month period. The point here is to show that syllabi referring to gossip consistently occuron the low end in a list of related social exchange and interpersonal communication topics andcontinue to grow relatively slowly.

81SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 5: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

doubtable achievements of the target. Ritualizedgossip seems to come to the aid of practically allsocieties when the need arises to blow off somecollective steam (cf. Stirling, 1956).

Nevertheless, circumstances that break therule of an absent gossip target are relativelyunusual and occupy a unique position in thesocial matrix. It may be that gossip about theself would better be labeled self-disclosure(Kuttler, Parker, & La Greca, 2002), and whenconsequently passed to a third party, it becomesgossip. Gossip about others who are present isprobably better labeled public disclosure or rid-icule (Kuttler et al., 2002; which may or maynot be made innocuous by other facets of thesituation). For the most part, authors agree thatday-to-day gossip refers to talk about absentthird parties.

Evaluative Content

When we consider the valence of gossipyremarks—that is, positive or negative evalua-tions being made by the gossiper—considerablecomplexity and variation are introduced into theconstruct. Social conversation without valenceis essentially the dissemination of human news:who landed a job or got admitted to college,who is having a baby, what products or servicesso-and-so uses, who moved in down the street,and the like. Assuming that, alongside theseapparently bald facts, there are no implied no-tions that carry significant evaluative connota-tions, these items may be considered gossip of abenign and nonevaluative nature for generalinformational purposes. Tannen (1990) definedgossip in this broad way, requiring only therepetition of news about a third party, the pass-ing along of everyday details. Bergmann (1993)relied on this same general definition of gossip,the passing of news about the personal affairs ofothers. Besnier (1989) and Hannerz (1967) ap-peared likewise content with a gossip lacking anevaluative component.

Nevertheless, for fruitful research purposes,limiting the definition of gossip to mere newsdissemination is unsatisfactory for two reasons.First, popular understanding of gossip clearlyincludes a negative evaluative component that,in part, forms the basis for the social sanctionsagainst its practice. Second, most exchanges ofpersonal news carry with them some evaluativemeaning implicit in the shared tacit knowledge,

histories, and cultural norms of the conversa-tionalists. Consequently, some theoretical socialfunctions of gossip (described in more detailshortly) are such that they would be effectivelyeviscerated if the evaluative component werenot an ingredient of the genre.

Thus, positive and negative evaluations (ortone) directed at a third party contribute to thesocial and research characterization of gossip.Such judgments carried along on the gossipchannel have, indeed, great implications forhow we choose to exist in concert with ourfellow humans. Although the surface of a gos-sipy conversation may appear casual, idle, ortrivial (Rosnow, 2001), the meta-communica-tive value is of quite a different kind when weinclude the evaluative components, implied orexplicit, that usually accompany suchcommunications.

When we consider the social and psycholog-ical functions of gossip, it is sensible to includepositive evaluations in the definition of gossip,although the everyday idea of gossip rarelydoes. Sabini and Silver (1978); Rosnow (2001);Baumeister, Zhang, and Vohs (2004; this issue);and others adamantly include positive remarksunder the general rubric of gossip. Mettetal(1982) observed positive and negative gossipamong adolescents, although exchanges amongyounger peers were likely to be more negative,suggesting that maturity may play a part indeveloping a more subtle and complex form(Gottman & Mettetal, 1986). Leaper and Holli-day (1995) coded for both kinds of gossip intheir recorded conversations and found genderdifferences along these lines. Noon and Del-bridge (1993) acknowledged the possible posi-tive consequences of gossiping behavior andlimited their definition to “value-laden informa-tion” (p. 25), citing a distinction made by Eliasand Scotson (1965) between “praise gossip”and “blame gossip.” Dunbar (1993b) assertedthat people should not hold gossip only to itspejorative sense (and doing so “says more aboutwhat these individuals talk about than anythingelse” [p. 729]). As noted earlier, Dunbar and hisassociates (1997) reported data showing thatmalicious gossip takes up only a small—though, no doubt, important—part of conversa-tional time.

Conceiving of gossip as having either posi-tive or negative valence is hardly a novel idea:Machiavelli (1516/1995) maintained that “all

82 FOSTER

Page 6: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

men, when they are talked about, . . . are re-marked upon for various qualities which bringthem either praise or blame” (p. 89). Gossipcertainly influences reputations; yet, there is nological reason to suppose that this is solelyaccomplished with negative remarks. Consoli-dation of power in the workplace, for instance,would require not only the denigration of rivals’programs but the extolling of preferred players’programs. Third-party information of both thepositive and negative kind has been shown toaugment people’s opinions of their coworkersin either respective direction (Burt & Knez,1995).

Situational Factors

Content is not the only thing that distin-guishes the gossip genre. Situation plays animportant role, too. The Oxford English Dictio-nary includes little about the evaluative compo-nents of gossip just discussed and only glancesat the situational aspects: As a noun, gossip is(1) in relation to the person baptized, a godfa-ther or godmother, a sponsor; (2a) a familiaracquaintance, friend, chum, formerly applied toboth sexes, now only (somewhat archaic) towomen; (2b) applied to a woman’s femalefriends invited to be present at a birth; (3) aperson, mostly a woman, of light and triflingcharacter, especially one who delights in idletalk, a newsmonger, a tattler; (4) the conversa-tion of such a person, idle talk, trifling orgroundless rumor, tittle-tattle; or (5) easy, un-restrained talk or writing, especially about per-sons or social incidents. But Hannerz (1967)and Rosnow (2001) have observed that contextis a necessary consideration in establishing thatgossip is present. Abrahams (1970) referred tothe need for “the right setting and . . . the prop-erly licensed conditions” (p. 292) for gossip tooccur; Yerkovich (1977) cited as a condition forgossip the congeniality of the situation; andSpacks (1982) asserted that “it’s a certain atmo-sphere, most of all, that makes gossip recogniz-able: of intimacy, of gusto, often of surprise andrevelation” (p. 30). The situational aspect ofgossip cannot be entirely separated from thecontent of the gossip, any more than the func-tions of gossip can be separated from the form;the inherent meaning of the content depends onthese other factors.

There are still other ways of defining gossipas it plays out in the human drama. Gilmore(1978) listed 11 different words used in a smallSpanish community for varieties of gossipingbehavior. Relevant variables he identified werepurpose, agent, information value, intensity,status of target, and context. On some occasionsin that society, the target of the gossip wasintended by the gossiper to catch wind of it.Cross-status gossip had its own definition, asdid gossip lacking deliberate instrumentality(regardless of the sensitivity of the target to thecontent). Some gossip would be discounted de-pending on the source and presumed to be ex-aggerated or untrue; this gossip, too, had a spe-cific name. Gilmore extracted a four-part typol-ogy from these varieties of gossip, dependingon number of gossipers, status of gossipee, in-strumentality, and legitimacy (or credibility).

These examples serve to show that settlingupon a final characterization of gossip for re-search purposes is not a simple matter. Yet,consistent with many researchers’ observations,the most frequently encountered definition maybe summarized as follows: In a context of con-geniality, gossip is the exchange of personalinformation (positive or negative) in an evalu-ative way (positive or negative) about absentthird parties. Definitions of the phenomenonmay be more or less restrictive than this sum-mary, as seen by the many typifications cited inthis review (and this special issue), which dem-onstrate that one must be clear about the con-struct before embarking on data collection.

Social Functions of Gossip

It is not surprising that the social functions ofgossip vary considerably from person to person,situation to situation, and (from the evidenceoffered in this issue) author to author. It may besaid, however, that the literature has coalescedaround four major social functions of gossiprooted in social exchange theories (viz., Blau,1964; Foa, 1971; Homans, 1950; Moreno, 1993;Parsons, 1960). These functions were essen-tially foreshadowed in an article written byStirling in 1956. She remarked upon gossip associally beneficial in that it facilitates informa-tion flow, provides recreation, and strengthenscontrol sanctions, thereby creating group soli-darity. Yet, it also can be “an outlet for hostileaggression” (Stirling, 1956, p. 263). Stirling

83SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 7: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

thus implied the four social functions of gossipencountered repeatedly in gossip literature inthe years since her article: information, enter-tainment, friendship (or intimacy), andinfluence.

Information

As a mechanism of information exchange,gossip is frequently described as an efficientand, at times, exclusive means of gathering ordisseminating information. From gossip, “theindividual gets a map of his social environment”(Hannerz, 1967, p. 57), particularly in low-ac-cessibility networks spread thin by high mobil-ity. At the group level, gossip has been aptlycalled “a slow scanning of the total informa-tional resource of the group” (Roberts, 1964, p.441) or “a sort of tally sheet for public opinion”(Szwed, 1966, p. 435). The “official line,” inthis age of sophisticated and instantaneouslyinformed publics, is often dismissed in favor of“the inside scoop” that only gossip can provide(Ayim, 1994; Crampton, Hodge, & Mishra,1998; Levin & Arluke, 1987; Rosnow, 2001).Suls (1977) observed that, although it may bepossible to communicate directly with otherpeople regarding needed social comparison in-formation, gossip may be the better means tosuch knowledge if the information sought is ofan unfavorable kind. From the tradition of so-cial exchange in psychology, gossip is oftenportrayed as a kind of currency, traded like anyother, and assessed for its value by the taker onthe basis of timeliness, usefulness, and, espe-cially, rarity. Rosnow and Fine (1976) observedthat the transactional nature of gossip seemed toparallel traditional patterns of economicexchange.

Bergmann (1993), in a related vein, believedthat it is the unequal distribution of knowledgethat makes the information spread through gos-sip valuable. The gossip producer’s “reputationand position within the gossip triad is [sic]essentially determined by the potential and fac-tual access he has” to information about others’private lives (p. 67). Baumeister et al. (2004), inthis issue, agree that social status may be ele-vated by gossiping: Frequently, listeners inferthat the gossiper is in possession of specialknowledge or understanding of social rules andstandards. For Rosnow and Fine (1976), theideal exemplar of this stereotype is the profes-

sional gossip columnist, a tradition, inciden-tally, that in the United States goes back to1730, when Benjamin Franklin wrote a columnfor the Pennsylvania Gazette. Before then, inEngland, Addison and Steele’s Tatler and Spec-tator, which were penny periodicals for the bur-geoning middle class, were early prototypes oftoday’s gossip columns. People looked to thesegossipy writers, as people do today, as sourcesof social knowledge and the proper ways to liveand behave.

The fact that informality and privacy are im-portant conditions for the transmission of gossip(Hannerz, 1967; Schoeman, 1994) necessarilymakes gossip “scarce,” because people natu-rally guard the flow of information about them-selves (Derlega & Chaikin, 1977; Hannerz,1967; Haviland, 1977; Szwed, 1966). Yerko-vich (1977) pointed out that “information, nomatter how salient or scandalous, isn’t gossipunless the participants know enough about thepeople involved to experience the thrill of rev-elation” (p. 196). Bergmann (1993) wrote thatinformational exchanges between spouses at theend of the day might not be gossip for that veryreason: “The ‘thrill’ and commitment does [sic]not nearly reach the normal dimensions of gos-sip conversations” (p. 68), although preciselywhere along this continuum gossip begins orends is certainly hard to say.

People sometimes go to extraordinary lengthsto gather personal information about their fel-low humans. Haviland (1977) reported peoplehabitually peeking in their neighbors’ door-ways, and children have sometimes been re-cruited as proxies (Goffman, 1959; Haviland,1977; Hotchkiss, 1962). The latter, Rosnow(2001) pointed out, has the double advantage ofeasy access and impunity from sanctions forlater spreading the information. Priests and pol-iticians have resorted to making liquor availableto facilitate information flow between theclasses (Szwed, 1966). Social workers in remoteareas have recognized that the pains of extremeisolation can be ameliorated by passing some“constructive gossip” to their clients, because“gossip at its worst is far better than no attentionat all” for these socially isolated people (G. W.Brown, 1985, p. 387). In the same vein, Mediniand Rosenberg (1976) believed that the infor-mational value in passing gossip can be restor-ative and therapeutic by way of “the same boatphenomenon” (p. 454)—not an especially so-

84 FOSTER

Page 8: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

phisticated therapy approach, perhaps, but pre-sumably useful.

Entertainment

Gossip as entertainment can be readily in-ferred by observing conversationalists passingthe time gossiping. Although the gossipee mightcertainly be sensitive about the information be-ing passed, this does not obviate the fact thatgossip can exist solely for the entertainment orrecreational value of the gossipers. It is “thesheer fun which for most gossipers explainstheir involvement” (Spacks, 1982, p. 31). Ben-Ze’ev (1994) and Stirling (1956) remarked onthe obvious pleasure derived from gossiping, asdid Rosnow (1977), who maintained that thereare times when gossip serves no external pur-pose of exploitation or influence but merely theimmediacy of amusement. Gelles (1989) ob-served that gossip, in its use of storytelling,satisfies the emotions in the same way thatliterature can; Spacks (1982) concurred. In theSpanish rural community he studied, Gilmore(1978) explained that gossip provided the pri-mary source of entertainment: “Nothing is en-joyed so thoroughly or treasured so preciously”(p. 92). In certain work environments, gossipingcan provide relief from monotony, as Roy(1958) experienced in his 2-month stint as afactory worker. And although the mass mediaproduction of gossip concerning public figuressuch as actors, politicians, and sports figures isconstructively different in a number of waysfrom privately conducted gossiping about tar-gets the participants know directly, the enter-tainment value of gossip is clearly the basis ofthis enormous cultural and economic enterprise.

To some degree, it may be said that theentertainment value of gossip occurs outside theactual exchange. As with buying a lottery ticketor waiting for the next at bat while watching abaseball game, there is enjoyment in the inter-vening moments, a carryover of interest as towhat change and new elements future tidbits ofinformation or action might bear. In any event,gossip—before, during, or after the ex-change—is a bulwark against life’s monotony,providing considerable stimulation for very lit-tle cost. Perhaps, as Spacks (1982) maintained,we cannot fully account for the entertainmentvalue of gossip: To explain what is rewarding in

gossip is “like efforts to elucidate what’s funnyin a joke” (p. 21).

Friendship

The friendship or intimacy function of gos-siping refers both to dyadic interchanges and tothe way in which gossip brings groups togetherthrough the sharing of norms, thereby establish-ing boundaries to distinguish insiders from out-siders. What begins as a trusted exchange inprivate becomes at the group level the knowl-edge, norm, and trust boundaries of tribes,clans, and cultures.

Sharing gossip is a way to telegraph to thelistener the gossiper’s confidence in the recipi-ent (Hannerz, 1967). If the gossiper is relativelysure of the security of the “dyadic boundary,”he or she will feel safe in disclosing to thelistener (Derlega & Chaikin, 1977) and so ce-ment the relationship. The relationship betweenMonica Lewinsky and Linda Tripp during theWilliam J. Clinton presidency was an examplein high Washingtonian circles of this expecta-tion, although, of course, it was breached inhistoric fashion.

In other contexts, a number of authors haveobserved the likelihood of gossiping with and,to a lesser extent, about friends (Blumberg,1972). “People in this town can’t keep theirmouths shut, and the worst of all are yourfriends,” complains a villager (Gilmore, 1978,p. 94). Notwithstanding the “strangers on atrain” phenomenon (Bergmann, 1993; Derlega& Chaikin, 1977), gossip is less likely to takeplace between casual acquaintances or strangersthan between friends (Blumberg, 1972), proba-bly because shared social meanings and historyare essential to understanding the subtleties ofthe gossip (Abrahams, 1970; Noon & Del-bridge, 1993).

At the group level, Gluckman (1963) notedhow outsiders simply cannot understand gossip,and at times it is deliberately used by insiders toexclude outsiders (Dunbar, 2004; Eckert, 1990;Loudon, 1961; Noon & Delbridge, 1993). New-comers find themselves struggling to stay up tospeed in casual conversations, wherein mean-ings are firmly rooted in long and complicatedhistories and, in the case of professional groups,frequently expressed in arcane jargon (Gluck-man, 1963). Yerkovich (1977) tracked howevaluations of discrete events gradually became

85SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 9: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

categorized abstractions comprising “the storeof shared knowledge that familiar individualsuse when they interact with one another” (p.194). Outsiders risk infringing on group valuesif they do not wait for the group to induct them(Abrahams, 1970) in a process that sometimesdevelops into a ritual or rule-bound ceremony(Noon & Delbridge, 1993; Roy, 1958) not of-ficially sanctioned and formalized but nonethe-less tacitly necessary to joining.

Influence

Establishing friendship at the dyadic or grouplevel is closely related to boundary enforcementand gossip’s influence function, widely dis-cussed by gossip writers. As a means of corral-ling (or expelling) the wayward and eccentric,gossip is acknowledged to be an efficient socialmechanism. The aim of gossip could be eitherto reform or to stigmatize the sinner, as Cox(1970) put it. Enquist and Leimar (1993) andDunbar (2004), in this volume, maintain thatgossip is a kind of informal policing device forcontrolling free riders and social cheats. In fact,these authors posit that, evolutionarily, this isthe most important function of language in gen-eral and gossip in particular.

It is not much of a deductive leap to realizethat what one hears about others can just aseasily be said to others about oneself; in thisway, we can learn how to behave—what to doand what not to do—from listening to gossip.Children no doubt learn the norms of their cul-ture, neighborhood, and various professions bylistening to frequent object lessons concoctedad hoc by parents. Coworkers learn what isexpected of them by hearing stories holdinghigh performers up to praise and low performersto shame; the “corporate culture” in an organi-zation is commonly expressed this way in gos-sipy stories (Kurland & Pelled, 2000; Noon &Delbridge, 1993). Indeed, culture in general de-pends on repetition of norms and mores in manyforms, both formal and informal, to maintain itshold on members. Gossip is arguably the mostcommon form because it requires no specialskill to produce, as do storytelling and singing,for instance (Abrahams, 1970). Many forms ofsocial comparison—and, therefore, social un-derstanding—may be expressed in the form ofgossip (Wert & Salovey, 2004). Baumeister etal. (2004), in this issue, present evidence that a

primary function of gossip is cultural learning ina general form.

Much of Gluckman’s (1963) research fo-cused on the coercive aspect of gossip from theperspective of the group. Paine (1967) coun-tered that it is “the individual who should betaken into account in forwarding and protectingindividual interests” (p. 278). Building on thework of both Paine (1967, 1970) and Goffman(1959), Cox (1970) wrote that gossip occurswhen a person “directly interferes in another’simpression-management, hence forcing the au-dience to redefine his victim’s role” (p. 88).

One of the conditions for gossip to be influ-ential is that people must agree on the norms forbehavior and what constitutes acceptability;gossipers typically articulate these things. Eck-ert (1990), for instance, in her study of adoles-cent girls’ gossip, referred to a “good person”(p. 95) as having acquired symbolic capital inthe eyes of his or her age group and, therefore,having relatively more power to influence oth-ers. Such people are repositories of groupnorms, and their opinions therefore have moreweight in shepherding conformity. Conformityis essential for the survival of the group as awhole, which may account for the particularlyvitriolic form of gossip observed in groups un-der pressure to survive and in open competitionwith one another (Cox, 1970; Gluckman, 1963).

Gossip’s potential to restrict freedom moti-vates people not only to minimize their eccen-tricities but also to minimize gossip about them-selves whenever possible. People might try tobe present, for example, when they sense theymay be being talked about (Gilmore, 1978), orthey may try to ferret out the sources of gossipabout themselves. Haviland (1977) observedthat although people are intensely and oftenindiscreetly curious about their neighbors, theygo to considerable lengths to hide details oftheir own daily lives. Szwed’s (1966) New-foundlanders also guarded themselves by vari-ous means against “overexposure.” Residents inGilmore’s (1978) Spanish town did not eveninvite neighbors into their homes, for fear ofgiving them the opportunity to talk about “allthat is ‘wrong’ ” (p. 94) there. Although vul-nerability to the influence of gossip will varyfrom person to person, most of us choose toprotect our membership in our respectivegroups by conforming at least outwardly to

86 FOSTER

Page 10: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

many norms, the breach of which is likely to becirculated in the form of gossip.

The influence function of gossip probablycontributes significantly to the popular knee-jerk denigration of the practice. However, asthis catalog has shown, gossip can also providean uncomplicated, benign pastime for many;hold communities together against the forces ofsocial entropy; cement dyads and groups withtrust and intimacy through private disclosuresand reiteration of norms; and relieve profoundisolation for some in remote locations orthrough therapy. Just as gossip divides, it mayalso bring together and provide harbor; itsboundary-making property both includes on theone hand and excludes on the other. For thesereasons, gossip should be understood to includenegative and positive valences, in both contentand effect, if the social functions of gossipregularly discussed are to make the fullest no-mothetical sense.

Other Functions of Gossip

Evolutionary Utility

The field of evolutionary biology provides aunique perspective on the social functions ofgossip, predictably reducing them to one criticalfunction: survival. Barkow (1992) suggestedthat gossip was selected for among our ances-tors because it provided information necessaryfor survival. News and evaluations about “rela-tives, rivals, mates and potential mates, off-spring, partners in social exchange, and the veryhigh-ranking” would be of particular interest, aswould “control over resources, sexual activities,births and deaths, current alliances/friendshipsand political involvements, health, and reputa-tion about reliability as a partner in social ex-change” (Barkow, 1992, p. 628). With such alarge part of everyday life taken up with socialinformation exchanges, it was inevitable thatthe search by theorists and researchers for anadaptive, evolutionary mechanism for gossipwould arise.

Dunbar (1992, 1993a, 1993b, 1994) has pre-sented a cogent argument for the developmentof gossip (and human language in general)based on data about primate grooming behav-iors, socially coherent group size, and neocor-tex/rest-of-brain volume ratios. The thrust of hisproposition is that, as primate groups become

larger, they need a more efficient way to main-tain social coherence than that afforded bygrooming. (Nonhuman primates have complexsocial structures demonstrably based on dyadicgrooming relationships.) By what means didhuman animals maintain the social coherence ofgroups much larger than nonhuman primategroups without encroaching on essential forag-ing time? Language is Dunbar’s answer, be-cause it affords efficient exchange of socialinformation.

Language thus becomes “grooming at a dis-tance.” “In effect, humans were now exploitingthe greater efficiency of language as a bondingmechanism to allow themselves to live in largergroups for the same investment of social time”(Dunbar, 1994, pp. 115–116). Language notonly allows members of the group to interactwith a wider set of individuals at any one timeby exchanging socially relevant information, italso greatly speeds up the process of acquiringthe information (Dunbar, 1993b). To keep trackof the social world, the social animal “needs toknow who is in and who is out, who is friendswith whom, and who is the best ally of theday. . . . The animal has to keep track of all this,constantly updating its social map with eachday’s new observations” (Dunbar, 1994, p. 66).As groups increase in size, the number of third-party social relationships needing to be kepttrack of increases exponentially. This, con-cluded Dunbar, is the mechanism for the evo-lution of gossip. Gossip—and the computinghardware required to process it—evolved tosolve the major adaptive problem of largergroup sizes.

Dynamic Utility and Guilt

A number of writers have used the phrase“letting off steam” as a purpose of gossip (Gil-more, 1978; Levin & Arluke, 1985, 1987;Rosenbaum & Subrin, 1963; Stirling, 1956).This function implies a cathartic release fromanger, guilt, anxiety, or some other unpleasantinternal state and a return to a balanced state ofrepose. Inevitably, any such mechanism in theextreme can go awry. In a case related byRosenbaum and Subrin (1963), a man is so-cially debilitated owing to his compulsive habitof passing obviously made-up gossip and hisdeliberate attempts to stir up mayhem by settingpeople against each other. Another case in-

87SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 11: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

volves a woman who gossips incessantly tofriends about others who are ill in hospitals.Perhaps gossiping has psychological utility fordealing with hostility and fear. In Freudianterms, in such cases, it may be a discharge ofOedipal longing, the cathecting of latent fanta-sies, or trace expressions of sibling rivalry thatmotivates gossip in a bid to curry authority’sfavor (Rosenbaum & Subrin, 1963). Some so-cial insights can be found in the dynamic viewthat gossip is a means of displacing inherentlyirresolvable hostilities. Substitution of the ob-ject of fear, hostility, or envy with a gossiptarget can sometimes result in social bondingand group formation. Gossip about newly hiredemployees in the workplace, for instance, prob-ably as common as gossip about authority fig-ures, serves to relieve an intractable uneasefrom unknown forces at the same time it vali-dates the status quo.

The observation of gossip alloyed in somemanner with guilt is a frequent one. Eckert(1990) noticed the dilemma adolescent girlsface if they are to uphold norms by gossipingabout those who break them: The upholdersmust themselves break a norm by gossiping.Awareness of the cost to others (even if theyremain unaware of it) ineluctably accompaniesthe gossiper. Spacks (1982), generally an apol-ogist for gossip, recognized its harmful poten-tial as an “indigestible problem” (p. 29) andacknowledged the irreducible complexity of thephenomenon: “Few gossip without guilt, fewdefend without ambivalence” (p. 21). It seemsat times that gossip generates as much guilt asits production simultaneously relieves.

Bergmann (1993) referred to the pervasiveways in which people assuage their guilt as“neutralizing behaviors” (p. 77) designed to di-vert attention from the compromising ethics ofspreading or even hearing news about absentothers. One reason the workplace is often ahotbed of gossip, he asserted, is that legitimatework activities provide a convenient cover “sothat [gossip] appears as the unintended, acci-dental, and thereby excusable activity of anoccasional gossiper” (Bergmann, 1993, pp. 77–78). Thus, people camouflage their motives by“accidentally” and passively acquiring privateinformation. Though gossip may not be the cen-tral purpose, but rather a byproduct of the largersituation, it is only apparently superfluous (Ros-now, 2001), as we pretend to be absorbed in

“serious” thought or activity while attendingclosely to gossip around us. This is both theexternal embodiment and denial of guilt.

The desire to defend against the guilt accom-panying gossip is pervasive but sometimesmore subtle than retreating to covert discussion.A number of writers have noted a conversa-tional tick of some kind that loosens the mantleof guilt and deflects culpability for the gossiper.Haviland (1977), for instance, noted that “. . . orso I hear” tacked onto a gossipy remark couldprotect the gossiper from harsh judgments.Bernstein (1984) observed the phrase “it’s justtalk” as serving a similar purpose. Anothermanifestation of this reluctance to face the na-ture of gossip squarely is the plethora of euphe-misms employed to name it: shooting thebreeze, shop talk, chit-chat, kibitzing, chewingthe fat, emotional speculation, being in the loop,hock a tchainik, “nothing really” (in answer to“what did you talk about?”), the scoop, theinside story, small talk, soap, dope, killing timetogether, socializing. In addition, complete lin-guistic substitutions are frequently made forwhat are essentially gossipy gatherings: “let’sgo for coffee,” “let’s have lunch sometimesoon,” “a cocktail after work,” “bridge on Fri-days,” “meet me for a drink?” The list of suchpretexts is, of course, endless. The evasivenessof these locutions is evidence that gossiping isnot universally approved of, and their sheernumber brings us back to the paradox noted atthe opening of this article: Predictable and per-sistent disapproval has made nary a dent in theubiquity of the practice.

Seen through the lenses of social function,survival, dynamic utility, and guilt, a motiva-tional view of gossip emerges. That gossip blos-soms in the presence of many messages andcues to curtail it supports the view that theremay be strong motivations to gossip. It is gen-erated to resolve numerous internal tensions(Wert & Salovey, 2004) or achieve externalgoals. As with many behaviors and emotions,however, people are not necessarily phenome-nologically aware of their gossiping motiva-tions. Baumeister et al. (2004), in this issue,make the analogy that, just as sexual desire maybe independent of reproduction, gossip maycontribute to cultural learning even if people arenot aware of their desire to promote such learn-ing by gossiping. More than once in my ownexperience, a research participant has spontane-

88 FOSTER

Page 12: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

ously denied ever having gossiped, yet in thenext sentence conveyed something patentlygossipy. Identifying the motives behind andconditions moderating such behavior warrants asolid place for gossip research in the generalpsychological literature, ripe for new methodsof investigation.

Methodologies Used in Gossip Research

Gossip research has roots in many disci-plines, which contributes to a wide variety ofresearch methods. This section categorizessome of these approaches, not as an exhaustivereview but, rather, as a brief description of themost common methods and their benefits anddrawbacks.

Participant Observers

Important early contributions to gossip the-ory came from social anthropology and its sub-specialty, ethnography. Numerous investiga-tions consisted of traditional field studies, withthe researchers more or less immersing them-selves in the culture and recording events asthey happened. Participant-observer field stud-ies require a large time commitment, but theyreturn a rich store of primary data, typically inthe form of field notes jotted down in closetemporal and physical proximity to actualevents. Table 2 lists most of the major partici-pant-observer studies in gossip research, alongwith the time commitments involved (as dis-closed in the articles).

Table 2Participant-Observer Gossip Studies: Time to Completion, Location, and Notes

Study Time to completion Location Notes

Besnier (1989) Time during 1980–1982 andin 1985

Pacific isle Recorded 10 hr of gossip in 15settings, capturing 8.6% ofthe inhabitants

Eder & Enke (1991) “On a regular basis for threeyears” (pp. 495–496)

Middle school in a medium-sized midwesterncommunity

Audiotaped or videotaped 4 to 8conversations of 11 differentgroups each, 79 studentstotal; ethnographic notes alsoused “to help interpret andanalyze recorded data” (p.496)

Gilmore (1978) Time during 1972–1973 anda month in 1977

Rural Spanish community Used “the standard observation-participation methods” (p. 99)

Handelman (1973) Not specified, recordedmany conversations of 16full-time workers

Israel

Hannerz (1967) Not specified, but longenough to “getestablished in personalrelationships with [theneighborhood] inhabitantsby way of frequent,seemingly casualencounters” (p. 45)

An “American ghetto” inWashington, DC

Haviland (1977) “From time to time over thepast ten years” (p. 186)

Zinacatan, Mexico Author and his family set up ahousehold and lived there

Loudon (1961) Between 1957 and 1960 WalesRoy (1958) Two months, full-time

factory workFactory, location not

specifiedBecame coworker of his

participantsSzwed (1966) Time during 1962–1964 Small town in NewfoundlandYerkovich (1977) One year “Sharing the life of an upper

middle-class white Anglo-Saxon urban community”(p. 192)

The approach was “influencedby. . .the ethnography of

speaking” (p. 192)

89SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 13: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

There are a number of benefits to this kind ofresearch. Close observation of communicationsas they unfold is an unparalleled way to gathersocial data. Ideally, patient adaptation to thesurroundings on the part of the researcher even-tually makes his or her presence relatively un-obtrusive, and genuine behavior or details thatare often captured poorly after the fact by self-report can be observed with immediacy andsensory continuity. Much situational informa-tion, including status relationships, personal andinterpersonal histories, and the routine daily ac-tivities of those observed, may also take onimportance and lend evaluative meaning to theinterpretation of gossipy exchanges. Roy(1958), for instance, realized that it took sometime after beginning a factory job before “thedisconnected became connected, the nonsensemade sense, the obscure became clear, and thesilly actually funny . . . the interaction began toreveal structure” (p. 161). Had he not taken thetime to acquire such knowledge, the subtletiesof the repartee among his coworkers wouldhave remained obscure and subject to signifi-cant misinterpretation.

Attempts by researchers to overcome bound-aries are not always successful, however.Gluckman (1963), cited heavily in gossip liter-ature, noted how he could not penetrate certaingroups “because I did not know enough gos-sip. . . . The outsider cannot always detect theslight personal knockdown” (p. 309). Gluckmanspent a significant portion of his classic 1963article discussing the work and observations ofElizabeth Colson, who spent many hours inter-viewing the Makah Indians of the Northwest.She came to see that Makah gossip is far morethan in-group aggression; it is a means of reas-serting central values of the culture. However,like Gluckman, she sensed her own position asan outsider among them: “Only others of theirown community have the technical knowledgeto compete in the game, or to appreciate theskill with which a point is scored” (Colson,1953, p. 228). Such limits of field study werebrought front and center by P. Wilson (1974),who questioned the presumptions and conclu-sions of Colson, citing other evidence that theMakah probably mistrusted her and misled herwith their self-reports. Thus, to be just an ob-server and not a participant means risking “notgetting it.”

Nevertheless, a good deal of our current the-oretical knowledge about gossip comes fromfield studies such as the ones just described.Hannerz (1967) remarked that “probably thereis no other way of acquiring knowledge aboutgossip” (p. 45). This is an overstatement per-haps, but it reflects the value of the method, tobe sure.

Video and Audio Recording

Other researchers, because of demographicdistance from their subjects, do not become“participants” but still collect information di-rectly, usually by ingratiating themselves withtheir subjects and then recording conversations.Eder and Enke (1991), for instance, studiedadolescent boys and girls (10 to 14 years of age)in a middle school setting. Over a period of 3years, these researchers ate lunch with studentsin the cafeteria or attended other activities. Intime, the observers felt they had achieved a highlevel of rapport. They then recorded 16 gossip-ing episodes and, by scrutinizing transcripts,found that gossipers’ remarks tend to be sensi-tive to what immediately precedes as well as towhat is expected to follow, because the discus-sants are aware of a predetermined, surprisinglypredictable (although certainly complex andflexible) conversation structure. In particular,the first response to an opening remark about agossip target would decisively determine thedirection the remainder of the episode wouldtake: “Responses to [target] evaluations aremore important than the evaluations themselvesin determining the nature of subsequent re-sponses” (Eder & Enke, 1991, p. 505). In addi-tion, expressions of agreement with remarksabout an unknown target (and later support forthe view the group expressed as a whole) dem-onstrate the group-norm-maintaining functionof gossip in action.

Mettetal (1982) transcribed audiotapes andvideotapes of girls 6–17 years of age and codedthe frequency and valence of gossiping behav-ior, finding that within this particular group, talkabout third parties was negative about half ofthe time and that, unlike some instances withEder and Enke’s subjects, the targets were al-ways known to both parties doing the gossiping.Planalp (1993), in her judges’ analyses of re-corded conversations, also found that mutualknowledge of third parties was a discriminant

90 FOSTER

Page 14: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

factor in predicting whether interactants werefriends or acquaintances. Nicolopoulou (1997)traced how the thematic content of boys’ andgirls’ gossip and conversations became moredifferentiated over the course of a school year.Besnier (1989) recorded and transcribedsome 50,000 words of conversation among in-habitants of a small Pacific island, yielding fine-grained data on gossip productions and howgossipers co-create their gossip. In particular, heshowed how leaders in the conversation manip-ulated their listeners to contribute gossip. Manyauthors, in fact, have remarked on this collabo-rative facet of gossiping exchanges. Baumeisteret al. (2004), in this issue, also point out that,cognitively, collaboration in the development ofa gossipy conversation necessarily augmentsthe social learning value of the exchange be-cause the content is better remembered andmore likely to be repeated.

Eavesdropping

The solution of a number of researchers to theproblems of exclusion or altering the socialrelations of those observed has been to eaves-drop in public places. This approach preservesthe interactional aspects and spontaneity of gos-siping. By spending hours in a campus meetingplace, for instance, Levin and Arluke (1985)cataloged features of gossip they overheard,which presumably was unaffected by their pres-ence. The authors found virtually no differencesin degree of gossip between men and womenbut some differences in content (women talkedmore about relationships; men talked moreabout sports figures and other public figures).Dunbar et al. (1997) also eavesdropped in pub-lic places, and McCormick and McCormick(1992) lurked online to find a high concentra-tion (52%) of “social topics” in an analysis ofe-mail messages.

Although unobtrusive, there are a number ofinherent disadvantages to the eavesdroppingmethod. People may be responsible for whatthey say in a public place, but we must askwhether they have a reasonable expectation ofprivacy. Should they know when they are beingresearch participants, even if the method used iscompletely noninteractive and anonymous?Thus, ethical issues arise. Further, the research-ers have no way of collecting information aboutthe gossipers, such as the length of time the

parties have known one another, the status re-lationships involved, or any other relevant his-tory. Information on the targets and their rela-tionship to the gossipers, too, is necessarilysketchy at best. The fact that these conversa-tions occurred publicly may have put some con-straints on the types of information transferred(although, to be sure, most of us have had someexperience with overhearing astonishingly in-discreet exchanges in public). Selection biasmust be taken into account as well, becausethese are typically convenience samples.

Questionnaire Studies

Participant observation, recordings, andeavesdropping all provide descriptive data onthe phenomenon of gossip and can capture someof the richness and complexity of the behavioras it occurs. There are fewer correlational stud-ies involving the use of questionnaires. Onesuch study was conducted by Jaeger, Skleder,Rind, and Rosnow (1994) with sorority mem-bers as participants. Acknowledging the diffi-culty of gathering empirical information onwhat is essentially a secretive or, at least, pri-vate behavior, these researchers gathered psy-chometric and network data on members of asorority in an effort to capture personality in-formation about both the gossipers and the tar-gets of gossip. The authors were not present atactual gossiping events or privy to the historiesof the sisters; however, the data neverthelessshed quantitative light on a number of gossip-related issues. On the basis of self-reports madeover several weeks, correlations were calculatedon the participants’ degree of gossiping behav-ior (or tendency to be a target of gossip) andother factors such as need for approval, self-esteem, general anxiety, and a rough measure ofsocial network position. Among other results,the authors found that moderate gossiperstended toward the social middle (i.e., high andlow gossipers have fewer close contacts). Highgossipers tend to be more anxious, yet lower inneed for social approval. And those less likelyto be gossiped about are higher in need forsocial approval, yet have fewer friends; thosemore likely to be targets report more friends.

Nevo et al. (1993) conducted a quantitativestudy after developing their own “tendency togossip questionnaire” (TGQ). They also admin-istered a social desirability scale to correlate

91SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 15: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

with their participants’ tendency to gossip; asexpected, the correlation was negative (r ��.33). Also reported from the TGQ was a ten-dency of women to gossip more than men (r �.20). However, the authors conceded that “so-cial pressure may lead men to under-report theirtendency to gossip. . . . Men may engage ingossip almost to the same extent as women”(Nevo et al., 1993, p. 980). This study wasperhaps the first in the gossip canon to use aspecifically gossip-related instrument. The de-velopment and application of such instrumentscarry the caveats that typically accompany self-report data: People may not be particularly at-tuned to the behavior being inquired about, orthey may have conscious or unconscious mo-tives for slanting reports about their ownbehaviors.

Experimental Research

Several innovative experimental studies havebeen done on gossip, all relying on hypotheticalvignettes. D. S. Wilson, Wilczynski, Wells, andWeiser (2000) presented participants with sys-tematically varied fictional gossip scenarios andobtained approval ratings of the targets, gossip-ers, and listeners in the scenarios. The resultsimplied that gossip (or fictional gossip, at least)that serves group norms reflects better on thegossipers and more harshly on the targets thangossip that is self-serving. Also, interesting gen-der interactions were found. After exposure totypes of gossip, such as true and false negativegossip, men and women responded aboutequally positively to truthful gossip, but womenresponded more negatively to false gossip. Menresponded more positively to speaking directlyto the supposed transgressor of a norm, and theyresponded much more negatively than womendid to a do-nothing approach. Men and womenequally disapproved of self-serving gossip.These results support a hypothesis that peoplegenerally approve of negative gossip in re-sponse to norm violations.

Kuttler et al. (2002) also used fictional sce-narios to gauge sensitivity to variables thatmight influence the perceived veracity of gos-sip. They found that preadolescents have a sur-prisingly keen sense of the restraints theyshould impose on third-hand information. Alarge number of third through sixth graderswere presented with vignettes involving differ-

ing cues. Conflicting information from the tar-gets themselves or from or between third partiescontributed ambiguity to the gossipy communi-cations in the stories. The reliability of speakersin the stories was manipulated as well, havingthem transmit untrue information either inten-tionally or unintentionally. Strong majorities ofthe participants did perceive the transmissionsas being gossip, and this perception wasstrongly associated with labeling the behavioras unacceptable. Simple, unambiguous gossip,however, had the lowest effect in this regard,�2(1, N � 384) � 32.33, p � .001, reffect size �.29, and both intentionally and unintentionallyfalse remarks had the highest, �2(1, N �384) � 72.41, p � .001, reffect size � .43 forboth. In reporting their confidence with the in-formation passed, the youngsters showed somesensitivity to the conflicting cues in the gossip-ing situations. For instance, the children weresurer of information when facts conflicted fromtwo third parties than when facts conflictedwhen heard from the actual source (target) anda third party, t(382) � 3.38, p � .001, reffectsize � .17. An interesting finding was that theyounger children tended to be somewhat moreskeptical of the gossip than the older children,t(382) � 3.30, p � .001, reffect size � .17.

Finally, Blumberg (1972), though never us-ing the word gossip, designed two studies toexplore the “communication of interpersonalevaluations.” His experiments showed a stronglikelihood that evaluations stop short of targets(r � .76); there was some interaction, however,with more positive evaluations reaching theirtargets than negative ones (r � .20). Positiveevaluations were also more likely to be sharedwith either close friends or more distant ac-quaintances. Blumberg’s first study involvedhypothetical statements about passing evalua-tions. His second inquired about actual eventsand circumstances. Although still open to self-reporting limitations, the results were consistentwith the first study. One drawback of the secondstudy was that it involved only femaleparticipants.

Future Directions

On the basis of existing literature, it is safe toconclude that social perception is often affectedby gossip, as humans spend a significant portion

92 FOSTER

Page 16: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

of their interactions participating in gossip inone form or another. I have reviewed the basicdefinitions of the construct and described themost common methods used in this area ofresearch, most of which has been conducted insocial sciences other than psychology. Thereremain numerous questions about the topic thatneed exploring to produce further insights intothis ubiquitous social activity.

Methods more sophisticated than eavesdrop-ping and more convenient than ethnographyneed to be developed and applied. Certainly,more experimental work with randomized as-signments is warranted. As mentioned earlier,D. S. Wilson et al. (2000), Kuttler et al. (2002),and Blumberg (1972) systematically varied gos-sip scenarios and then asked for ratings of var-ious kinds, with interesting results. S. Wert(personal communication, February 24, 2002),of the Psychology Department at Yale Univer-sity, has reported that spontaneous gossip canbe generated in a laboratory and affected byself-esteem manipulations. Independent vari-ables such as stimulus, source, mood, socialcontext, and group composition may be manip-ulated. Dependent variables of interest might bebelief in the veracity of the gossip, likelihood oftransmission across status boundaries, feelingswhile hearing or producing gossip, changes inattitudes toward sender or target, and other cog-nitive and affective elements.

Factors that determine people’s response togossip and how they act upon it could helpexplain how gossiping plays a role in socialcognition and relationships. Do people’s self-perceptions of gossip correspond to their as-sumptions about how others gossip, or is there aself-serving bias? How aware are individuals ofthe gossiping they do? Emler (1990), for in-stance, showed experimentally that people arelikely to exert considerable effort to protecttheir reputations, particularly with those closestto them and more so in “dense” social networks(i.e., networks with a relatively high percentageof dyadic contacts). This suggests that peoplehave a keen sensitivity to gossip and its effects.Two additional issues are relevant: Do those “inthe loop” have a competitive advantage in thesocial, economic, or political marketplace? Dogroups that gossip more on average than theirrival groups also have a competitive advantageor disadvantage? Longitudinal studies could ex-

plore whether amount of gossiping activity orskill contributes to social difficulties or adjust-ment, at both the individual and the group level.

Assuming that gossip was essential to ourprehistoric ancestors’ survival, as Dunbar andothers maintain, certain physiological and neu-rological processes might be measurably af-fected at the moment of gossip production andperhaps even by overtures to gossip. In a seriesof impressive experiments involving brain im-aging with magnetic resonance imaging (MRI)and functional MRI (fMRI), Damasio (2000,2001) and Adolphs and Damasio (2000)showed that simply asking respondents to imag-ine certain emotional events in their lives pro-duces activity in specific areas of the brain.Work of this kind could yield important data onhow gossip information is stored and retrievedand how much of the brain is involved in pro-cessing it. Manipulating type of gossip, targets,or partners may activate different parts of thebrain and neurohormonal systems. Analysis offMRIs and other data could thus lead to neuro-biologically based models of gossiping behav-ior designed to scrutinize and explain the natureof gossip as a social, biological, and even sur-vival mechanism.

Another avenue of inquiry is whether, duringgossip production, people believe it has, or in-tend it to have, consequences outside of theconversation. Degree of interest in heard gossipmay depend on how engaged individuals arewith the target and the source, as well as withthe content. H. Hom (personal communication,February 25, 2002), of the Psychology Depart-ment at the University of Virginia, has devel-oped a diary-type questionnaire to capture dataaimed at answering questions along these andother lines.

Further survey research could produce much-needed parametric data. The social functions ofgossiping behavior, as noted earlier in this arti-cle, are frequently identified by observers. Ihave developed and used an instrument thatdifferentiates these functions moderately welland reliably; it is appended to this article. Atheoretical objective in organizational settingsmight involve determining how the social func-tions of gossip are affected by the nature of theformal network (the organizational chart or pre-scribed network). Does gossip become more“informational” or more “influential” in astrongly hierarchical organization? Relative to

93SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 17: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

typical workplaces, groups without a nominaleconomic purpose (e.g., clubs, dormitories, con-gregations, and study groups) or groups withmore or fewer status ranks may shift the moti-vational functions of gossip in predictable ways.Answers to such questions could aid in buildingmodels of the social and organizational pro-cesses of gossip.

On the assumption that the perceived envi-ronment guides behavior to some degree, I havealso examined gossip with sociometric methodsto determine how gossiping behavior and thestructure of the social network interact. Onequestion that may be answered with networkdata is whether those who gossip more thanaverage in relation to their local peer group aremore, or less, connected to the members of thatgroup. That is, is there a linear relationshipbetween one’s connections to a network andhow much one gossips? Or, as some have pro-posed, is there a quadratic relationship whereby,to maximize their knowledge of the network,people gossip neither too much nor too little,but “just the right amount”? In a study of 46residence assistants (RAs) at a small college, Ifound evidence of a strong, positive linear rela-tionship between the number of times one isnamed as a gossip partner by one’s peers andthe amount of social connection one has inaddition to these partnerships (r � .67; Foster,2003). It is possible that the complete lack of aquadratic relationship was due to a restriction ofrange, but of what dimension(s) remains to bedetermined.

The value of social network analysis in gos-sip research is also hinted at by a typology ofgossip that emerged in the RA study. Throughthe use of a technique that distinguished socialgroup members according to the similarity oftheir gossip connections to and from one an-other, four identifiable groups emerged from thehighly complex gossiping interrelationships.First was a group whose tenure as RAs waslonger than that of others. These folks were wellestablished socially, as they had many gossippartners. However, they reported a lower fre-quency of contact and less RA-related adviceseeking. The second group differed from thefirst in that they tended to be newcomers. Theyalso had many gossip partners, but they had ahigh frequency of contact with others, soughtmuch RA-related advice, and reported manyclose friendships among other RAs. (This group

resembled the “freshman class” Wert andSalovey [2004] discuss elsewhere in this issue.)In other words, the second group was getting intouch with the network; the first group wasstaying in touch.

The third group, very unlike the first two,reported few gossip partners (and were namedvery infrequently by others as gossip partners,as well). They had a low frequency of contactwith others, generally did not seek out advicefrom their RA peers, and did not report closefriends among the RAs. They were social“black boxes.” Finally, a fourth group estab-lished close friendships, but not much in theway of gossip partners, advice seeking, or fre-quent contact with others. These people weresociable enough, but they did not frequentlyemploy gossip or advice seeking as part of theirsocial repertoire.

Figure 2 summarizes these groups in a 2 � 2typology, along with suggested dimensions onthe axes that seem to differentiate the groups.Members of the two groups composing the up-per half of the grid, relative to those in the lowerhalf, manifested interest in the network beyondtheir immediate neighborhood, as indicated byrelatively high numbers of gossip partners.Members in the left half of the grid appeared tobe less active than those in the right half in theirapproach to social information exchanges ingeneral, in that they had less frequent contactand fewer types of contact with other membersof the network. Certainly, there were “bridges”across the boundaries of these four groups, butthe in-group bias for gossiping within them wasstrong.

Additional network issues may be exploredwith social network analysis. Network density(i.e., “close knit” or “loose knit” on variouskinds of links), for instance, may have measur-able effects on both the frequency and type ofgossipy communications. Further, network lo-cation information is easily captured in so-ciomatrices, presented in sociograms, and ex-amined with network analysis. Network prox-imity and status equality of gossip targets andpartners may be important variables in both thespread of gossip and the predominance of gos-sip’s social functions within a given group. Anadvantage of the type of data collected in anetwork study is that the ties reported by indi-viduals, although subjective with regard to the

94 FOSTER

Page 18: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

respondent, are objective with regard to peers.Thus, certain social response bias problems as-sociated with gossip self-reports may beovercome.

Concluding Remarks

As with stereotypes, persuasion, conformity,group dynamics, and other important areas ofpsychology, there are wide real-world implica-tions for a better understanding of gossipingbehavior within both the research and the gen-eral communities. And yet, gossip is relativelyunfamiliar terrain in psychology. Few empiricaldata exist; only a limited number of experimen-tal studies have been brought to bear on thetopic. The general moral ambivalence towardgossip and the instrumental ambiguities in-volved call out for serious sociocognitive andpsychophysiological explanations of how it isgenerated, perceived, processed, and used.

This review has shown that gossip is a het-erogeneous phenomenon in content, forms, andfunctions. It may be manifested differently

across the life span. It may idly fill hours of theday, passing practically unnoticed, or it may bemotivationally generated or consumed (con-sciously or not) to alter internal states andachieve socially important external goals. As ageneral psychological strategy, people passivelyand actively use gossip to determine their viewof the world and to convey it meaningfully toothers. To no small extent, we learn how tobehave, think, and communicate from and withgossip. For these reasons, it has the potential tocreate collegiality and understanding as well asinequality and conflict along gossip lines be-tween and within social groups.

From the researcher’s perspective, the com-plexities of gossip may be unsatisfying andproblematic. But, of course, they are also chal-lenges and opportunities for fecund and excitinginquiry. Strengthening the theory, methods, andinstruments in this area will better inform theway we—managers, administrators, parents,teachers, peers, and colleagues—view this ubiq-uitous and fascinating behavior in context.

Figure 2. Four-group typology of gossipers.

95SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 19: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

References

Abrahams, R. D. (1970). A performance-centeredapproach to gossip. Man, 5, 290–301.

Adolphs, R., & Damasio, A. R. (2000). Neurobiologyof emotion at a systems level. In J. C. Borod (Ed.),The neuropsychology of emotion (pp. 194–213).New York: Oxford University Press.

Ayim, M. (1994). Knowledge through the grapevine:Gossip as inquiry. In R. F. Goodman & A. Ben-Ze’ev (Eds.), Good gossip (pp. 85–99). Lawrence:University Press of Kansas.

Barkow, J. H. (1992). Beneath new culture is oldpsychology: Gossip and social stratification. InJ. H. Barkow, L. Cosmides, & J. Tooby (Eds.), Theadapted mind (pp. 627–637). Oxford, England:Oxford University Press.

Baumeister, R. F., Zhang, L., & Vohs, K. D. (2004).Gossip as cultural learning. Review of GeneralPsychology, 8, 111–121.

Ben-Ze’ev, A. (1994). The vindication of gossip. InR. F. Goodman & A. Ben-Ze’ev (Eds.), Goodgossip (pp. 11–24). Lawrence: University Press ofKansas.

Bergmann, J. R. (1993). Discreet indiscretions: Thesocial organization of gossip. New York: Aldinede Gruyter.

Bernstein, L. M. (1984). Ko e lau pe (it’s just talk):Ambiguity and informal social control in a Tonganvillage. Dissertation Abstracts International,45(3A), 877–878.

Besnier, N. (1989). Information withholding as amanipulative and collusive strategy in Nukulaelaegossip. Language in Society, 18, 315–341.

Blau, P. M. (1964). Exchange and power in sociallife. New York: Wiley.

Blumberg, H. H. (1972). Communication of interper-sonal evaluations. Journal of Personality and So-cial Psychology, 23, 157–162.

Bok, S. (1983). Secrets: On the ethics of concealmentand revelation. New York: Pantheon Books.

Brown, E. B. (1990). Motley moments: Soap operas,carnival, gossip and the power of the utterance. InM. E. Brown (Ed.), Television and women’s cul-ture: The politics of the popular (pp. 183–198).London: Sage.

Brown, G. W. (1985). Joining two social institutionsto counter rural Alaskan child abuse. Child Abuseand Neglect, 9, 383–388.

Burt, R. S., & Knez, M. (1995). Kinds of third-partyeffects on trust. Rationality and Society, 7, 255–292.

Coates, J. (1988). Gossip revisited: Language in all-female groups. In J. Coates & D. Cameron (Eds.),Women in their speech communities: New perspec-tives on language and sex (pp. 94–122). London:Longman.

Colson, E. (1953). The Makah Indians. Manchester,England: Manchester University Press.

Cox, B. A. (1970). What is Hopi gossip about? In-formation management and Hopi factions. Man, 5,88–98.

Crampton, S. M., Hodge, J. W., & Mishra, J. M.(1998). The informal communication network:Factors influencing grapevine activity. Public Per-sonnel Management, 27, 569–584.

Damasio, A. R. (2000). A neurobiology for con-sciousness. In T. Metzinger (Ed.), Neural corre-lates of consciousness: Empirical and conceptualquestions (pp. 111–120). Cambridge, MA: MITPress.

Damasio, A. R. (2001). Emotion and the humanbrain. In A. R. Damasio, A. Harrington, J. Kagan,B. S. McEwen, H. Moss, & R. Shaikh (Eds.), Unityof knowledge: The convergence of natural andhuman science (Vol. 935, pp. 101–106). NewYork: New York Academy of Sciences.

Derlega, V. J., & Chaikin, A. L. (1977). Privacy andself-disclosure in social relationships. Journal ofSocial Issues, 33, 102–115.

Dunbar, R. I. M. (1992). Why gossip is good for you.New Scientist, 136, 28–31.

Dunbar, R. I. M. (1993a). Coevolution of neocorticalsize, group size and language in humans. Behav-ioral and Brain Sciences, 16, 681–694.

Dunbar, R. I. M. (1993b). On the origins of language:A history of constraints and windows of opportu-nity. Behavioral and Brain Sciences, 16, 721–735.

Dunbar, R. I. M. (1994). Grooming, gossip, and theevolution of language. London: Faber & Faber.

Dunbar, R. I. M. (2004). Gossip in evolutionaryperspective. Review of General Psychology, 8,100–110.

Dunbar, R. I. M., Marriott, A., & Duncan, N. D. C.(1997). Human conversational behavior. HumanNature, 8, 231–246.

Eckert, P. (1990). Cooperative competition in ado-lescent “girl talk.” Discourse Processes, 13, 91–122.

Eder, D., & Enke, J. L. (1991). The structure ofgossip: Opportunities and constraints on collectiveexpression among adolescents. American Socio-logical Review, 56, 494–508.

Eggins, S., & Slade, D. (1997). Analyzing casualconversation. London: Cassell.

Elias, N., & Scotson, J. L. (1965). The establishedand the outsiders. London: Frank Cass.

Emler, N. (1990). A social psychology of reputation.European Review of Social Psychology, 1, 171–193.

Emler, N. (1994). Gossip, reputation and social adap-tion. In R. F. Goodman & A. Ben-Ze’ev (Eds.),Good gossip (pp. 119–140). Lawrence: UniversityPress of Kansas.

96 FOSTER

Page 20: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

Enquist, M., & Leimar, O. (1993). The evolution ofcooperation in mobile organisms. Animal Behav-iour, 45, 747–757.

Fine, G. A., & Rosnow, R. L. (1978). Gossip, gos-sipers, gossiping. Personality and Social Psychol-ogy Bulletin, 4, 161–168.

Foa, U. G. (1971). Interpersonal and economic re-sources. Science, 171, 345–351.

Foster, E. K. (2003). Researching gossip with socialnetwork analysis. Unpublished doctoral disserta-tion, Temple University, Philadelphia, PA.

Gelles, E. B. (1989). Gossip: An eighteenth-centurycase. Journal of Social History, 22, 667–683.

Gilbert, D. T., Fiske, S. T., & Lindzey, G. (Eds.).(1998). The handbook of social psychology (4thed., Vols. 1 & 2). New York: McGraw-Hill.

Gilmore, D. (1978). Varieties of gossip in a Spanishrural community. Ethnology, 17, 89–99.

Gluckman, M. (1963). Gossip and scandal. CurrentAnthropology, 4, 307–316.

Goffman, E. (1959). The presentation of self in ev-eryday life. New York: Doubleday.

Goodwin, M. H. (1982). “Instigating”: Storytelling asa social process. American Ethnologist, 9, 799–819.

Gottman, J., & Mettetal, G. (1986). Speculationsabout social and affective development: Friendshipand acquaintanceship through adolescence. In J.Gottman & J. Parker (Eds.), Conversations offriends: Speculations on affective development(pp. 192–237). New York: Cambridge UniversityPress.

Handelman, D. (1973). Gossip in encounters: Thetransmission of information in a bounded socialsetting. Man, 8, 210–227.

Hannerz, U. (1967). Gossip networks and culture in aBlack American ghetto. Ethnos, 32, 35–59.

Harrington, C. L., & Bielby, D. D. (1995). Where didyou hear that? Technology and the social organi-zation of gossip. Sociological Quarterly, 36, 607–628.

Haviland, J. B. (1977). Gossip as competition inZinacantan. Journal of Communication, 27, 186–191.

Hecht, J. J. (1956). The domestic servant class ineighteenth-century England. London: Routledge &Kegan Paul.

Homans, G. C. (1950). The human group. New York:Harcourt, Brace.

Hotchkiss, J. C. (1962, November). Children errand-runners: Their roles in the social life of a smallMexican community. Paper presented at the annualmeeting of the American Anthropological Associ-ation, Chicago.

Jaeger, M. E., Skleder, A. A., Rind, B., & Rosnow,R. L. (1994). Gossip, gossipers, gossipees. In R. F.Goodman & A. Ben-Ze’ev (Eds.), Good gossip

(pp. 154–168). Lawrence: University Press ofKansas.

Jones, D. (1980). Gossip: Notes on women’s oralculture. In C. Kramarae (Ed.), The voices andwords of women and men (pp. 193–198). Oxford,England: Pergamon Press.

Kurland, N. B., & Pelled, L. H. (2000). Passing theword: Toward a model of gossip and power in theworkplace. Academy of Management Review, 25,428–438.

Kuttler, A. F., Parker, J. G., & La Greca, A. M.(2002). Developmental and gender differences inpreadolescents’ judgments of the veracity of gos-sip. Merrill-Palmer Quarterly, 48, 105–132.

Leaper, C., & Holliday, H. (1995). Gossip in same-gender and cross-gender friends’ conversations.Personal Relationships, 2, 237–246.

Levin, J., & Arluke, A. (1985). An exploratory anal-ysis of sex differences in gossip. Sex Roles, 12,281–286.

Levin, J., & Arluke, A. (1987). Gossip: The insidescoop. New York: Plenum.

Lindzey, G., & Aronson, E. (Eds.). (1970). The hand-book of social psychology (2nd ed., Vols. 1–5).Reading, MA: Addison-Wesley.

Lindzey, G., & Aronson, E. (Eds.). (1985). The hand-book of social psychology (3rd ed., Vols. 1 & 2).New York: Random House.

Loudon, J. (1961). Kinship and crisis in South Wales.British Journal of Sociology, 12, 333–350.

Machiavelli, N. (1995). The prince and other politi-cal writings. London: Everyman. (Original workpublished 1516)

McCormick, N. B., & McCormick, J. W. (1992).Computer friends and foes: Content of undergrad-uates’ electronic mail. Computers and Human Be-haviour, 8, 379–405.

Medini, G., & Rosenberg, E. H. (1976). Gossip andpsychotherapy. American Journal of Psychother-apy, 30, 452–462.

Mettetal, G. W. (1982). The conversations of femalefriends at three ages: The importance of fantasy,gossip, and self-disclosure. Unpublished doctoraldissertation, University of Illinois, Urbana.

Moreno, J. L. (1993). Who shall survive?: Founda-tions of sociometry, group psychotherapy and so-ciodrama. McLean, VA: American Society ofGroup Psychotherapy and Psychodrama.

Nevo, O., Nevo, B., & Derech-Zehavi, A. (1993).The development of the tendency to gossip ques-tionnaire: Construct and concurrent validation fora sample of Israeli college students. Educationaland Psychological Measurement, 53, 973–981.

Nicolopoulou, A. (1997). Worldmaking and identityformation in children’s narrative play-acting. In B.Cox & C. Lightfoot (Eds.), Sociogenic perspec-tives in internalization (pp. 157–187). Hillsdale,NJ: Erlbaum.

97SPECIAL ISSUE: RESEARCH ON GOSSIP

Page 21: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

Noon, M., & Delbridge, R. (1993). News from be-hind my hand: Gossip in organizations. Organiza-tion Studies, 14, 23–36.

Paine, R. (1967). What is gossip about? An alternatehypothesis. Man, 2, 278–285.

Paine, R. (1970). Informal communication and infor-mation-management. Canadian Review of Sociol-ogy and Anthropology, 7, 172–188.

Parsons, T. (1960). Structure and process in modernsocieties. Glencoe, IL: Free Press.

Planalp, S. (1993). Friends’ and acquaintances’ con-versations II: Coded differences. Journal of Socialand Personal Relationships, 10, 339–354.

Roberts, J. M. (1964). The self-management of cul-tures. In W. H. Goodenough (Ed.), Explorations incultural anthropology (pp. 433–454). New York:McGraw-Hill.

Rosenbaum, J. B., & Subrin, M. (1963). The psy-chology of gossip. Journal of the American Psy-choanalytic Association, 11, 817–831.

Rosnow, R. L. (1977). Gossip and marketplace psy-chology. Journal of Communication, 27, 158–163.

Rosnow, R. L. (2001). Rumor and gossip in interper-sonal interaction and beyond: A social exchangeperspective. In R. M. Kowalski (Ed.), Behavingbadly: Aversive behaviors in interpersonal rela-tionships (pp. 203–232). Washington, DC: Amer-ican Psychological Association.

Rosnow, R. L., & Fine, G. A. (1976). Rumor andgossip: The social psychology of hearsay. NewYork: Elsevier.

Roy, D. F. (1958). Banana time: Job satisfaction andinformal interaction. Human Organization, 18,158–168.

Rysman, A. (1977). How the “gossip” became awoman. Journal of Communication, 27, 176–180.

Sabini, J. P., & Silver, M. (1978). Moral reproach andmoral action. Journal for the Theory of SocialBehaviour, 8, 103–123.

Schoeman, F. (1994). Gossip and privacy. In R. F.Goodman & A. Ben-Ze’ev (Eds.), Good gossip(pp. 72– 82). Lawrence: University Press ofKansas.

Slade, D. (1997). Stories and gossip in English: Themacro-structure of casual talk. Prospect, 12, 43–71.

Spacks, P. M. (1982). In praise of gossip. HudsonReview, 35, 19–38.

Stirling, R. B. (1956). Some psychological mecha-nisms operative in gossip. Social Forces, 34, 262–267.

Suls, J. M. (1977). Gossip as social comparison.Journal of Communication, 27, 164–168.

Szwed, J. F. (1966). Gossip, drinking, and socialcontrol: Consensus and communication in a New-foundland parish. Ethnology, 5, 434–441.

Tannen, D. (1990). You just don’t understand:Women and men in conversation. New York:Morrow.

Wert, S. R., & Salovey, P. (2004). A social compar-ison account of gossip. Review of General Psy-chology, 8, 122–137.

Wilson, D. S., Wilczynski, C., Wells, A., & Weiser,L. (2000). Gossip and other aspects of language asgroup-level adaptations. In C. Heyes & L. Huber(Eds.), The evolution of cognition (pp. 347–365).Cambridge, MA: MIT Press.

Wilson, P. (1974). Filcher of good names: An en-quiry into anthropology and gossip. Man, 9, 93–102.

Yerkovich, S. (1977). Gossip as a way of speaking.Journal of Communication, 26, 192–196.

98 FOSTER

Page 22: Research on Gossip: Taxonomy, Methods, and …Research on Gossip: Taxonomy, Methods, and Future Directions Eric K. Foster University of Pennsylvania A half century of gossip research

Appendix

Gossip Functions Questionnaire

The Gossip Functions Questionnaire consists of 24statements, 6 for each subscale. They are rated from 1(strongly disagree) to 5 (strongly agree). Alpha in-ternal-consistency reliabilities over three administra-tions (N � 110) are as follows: information, .80;friendship, .81; influence, .64; and entertainment, .80.The test–retest reliability for the entire scale is .64(mean days to retest � 29, SD � 16.8). A formattedversion, along with a spreadsheet that totals thescores with reversals, is available from the author.

Information

1. Generally, I try to figure out what is going on inthe lives of people around me.

2. For me, informal chatting is unimportant forpurposes of general information gathering.

3. Listening to people’s opinions of others helpsme better judge aspects of my own life.

4. When something personal happens to other peo-ple I know, if it doesn’t directly affect me, I don’tcare to know about it.

5. I like to know what is going on with people, thatis, who is dating whom, who is getting a new job,what classes people have had, and such.

6. For me, gossiping with others is not a very goodway to get useful information.

Friendship

1. Talking about the personal lives of other peoplemakes me feel in touch with my social circle.

2. I believe that close friends can easily get per-sonal information from each other about otherpeople.

3. After I become friends with someone, I gener-ally start to hear more about the personal issues ofothers from them.

4. With good friends, I tend to share informationI’ve heard about others.

5. Some of my friendships were formed in part bytalking about some third person.

6. I’m bored by spending time with my friends justtalking about mutual friends.

Influence

1. When someone does something inappropriate, Ithink others should know so the person will be lesslikely to do it again.

2. Hearing stories about others could help meavoid saying or doing the wrong thing.

3. When I hear gossip, it can change my behaviortoward someone, for better or for worse.

4. It’s fair to say that gossip tells us what to do, thatis, how to behave in a lot of situations.

5. Though people will “talk” about others, this isnot effective at bringing people into line.

6. It is OK to tell a new person in my circle whatit was about someone who is gone that no one liked.

Entertainment

1. I don’t have to know whether talk about peopleis true or not to enjoy the activity.

2. I like to gossip at times.3. For me, there’s nothing enjoyable about passing

on personal information.4. I can’t understand why people get so much

pleasure out of gossiping.5. I could spend hours listening to stories about the

lives of the people I know.6. I can’t stand being around people who talk about

other people behind their backs.

Received June 2, 2003Accepted July 1, 2003 �

99SPECIAL ISSUE: RESEARCH ON GOSSIP