Identity Implications of Influence Goals A Revised Analysis of Face-Threatening Acts and Application...

33
Identity Implications of Influence Goals A Revised Analysis of Face-Threatening Acts and Application to Seeking Compliance With Same-Sex Friends STEVEN R. WILSON Northwestern Universify CARLOS G. ALEMAN University of lllinois-Chicago GEOFF B. LEATHAM Southern Oregon State College Although politeness theoy offers one explanation for how threats to face arise during compliance-gaining episodes, it neither predicts the conditions under which seeking compli- ance will create multiple face threats nor explains how such threats arise within specific contexts. The authors challenge and revise politeness the0 y by analyzing potential implica- tions for both parties’face when the logical preconditions for seeking compliance areframed by spec@ influencegoals. In a test of a revised analysis offace threats,young adults imagined askingfavors, giving advice, and enforcing obligations with same-sexfriPnds. Perceivedface threats, interaction goals, and message qualities varied substantially a m s s compliance-gain- ing situations defined by these goals. Directions for exploring the identity implications of influence goals across relationshipsand cultures are discussed. ince the publication of Miller, Boster, Roloff, and Seibold’s (1977) ground-breaking study, communication scholars have had an S enduringinterest in how individualsseek compliance (see Seibold, Cantrill, & Meyers, 1994). Compliance-gainingepisodes are interesting, in part, because they can be complex (OKeefe, 1988), especially when issues pertaining to each party’s rights and responsibilities are raised, and both people feel that their identities are on the line. Given this potential, individuals often pursue multiple, conflicting goals as they seek compli- Steven R. Wilson (Ph.D., Purdue University, 1989) is an associateprofessor in the Department of CommunicationStudies at Northwestern University. Carlos G. Aleman (Ph.D., University of Iowa, 1996) is an assistantprofessor in the Department of Communication at the University of Illiois-Chicago. GeoflB. Leafham (Ph.D., University of Iowa, 1996) is an assistantprofessor in the Department of Communicationat Southern Oregon State College. We wish to thank Cindy Gallois and two anonymous reviewers for their helpful feedback on earlier drafts of this article.Please direct correspondenceto Steven R. Wilson, Department of Communication Studies, Northwestern University, 1881 Sheridan Road, Evanston, IL 60208-0818. Human Communication Research, Vol. 25 No. 1, September 1998 64-96 0 1998 International Communication Association 64

Transcript of Identity Implications of Influence Goals A Revised Analysis of Face-Threatening Acts and Application...

Identity Implications of Influence Goals A Revised Analysis of Face-Threatening Acts and Application to Seeking Compliance With Same-Sex Friends

STEVEN R. WILSON Northwestern Universify CARLOS G. ALEMAN University of lllinois-Chicago GEOFF B. LEATHAM Southern Oregon State College

Although politeness theoy offers one explanation for how threats to face arise during compliance-gaining episodes, it neither predicts the conditions under which seeking compli- ance will create multiple face threats nor explains how such threats arise within specific contexts. The authors challenge and revise politeness the0 y by analyzing potential implica- tions for both parties’face when the logical preconditions for seeking compliance areframed by spec@ influencegoals. In a test of a revised analysis offace threats, young adults imagined askingfavors, giving advice, and enforcing obligations with same-sexfriPnds. Perceivedface threats, interaction goals, and message qualities varied substantially a m s s compliance-gain- ing situations defined by these goals. Directions for exploring the identity implications of influence goals across relationships and cultures are discussed.

ince the publication of Miller, Boster, Roloff, and Seibold’s (1977) ground-breaking study, communication scholars have had an S enduring interest in how individuals seek compliance (see Seibold,

Cantrill, & Meyers, 1994). Compliance-gaining episodes are interesting, in part, because they can be complex (OKeefe, 1988), especially when issues pertaining to each party’s rights and responsibilities are raised, and both people feel that their identities are on the line. Given this potential, individuals often pursue multiple, conflicting goals as they seek compli-

Steven R. Wilson (Ph.D., Purdue University, 1989) is an associate professor in the Department of Communication Studies at Northwestern University. Carlos G. Aleman (Ph.D., University of Iowa, 1996) is an assistant professor in the Department of Communication at the University of Illiois-Chicago. GeoflB. Leafham (Ph.D., University of Iowa, 1996) is an assistant professor in the Department of Communication at Southern Oregon State College. We wish to thank Cindy Gallois and two anonymous reviewers for their helpful feedback on earlier drafts of this article. Please direct correspondence to Steven R. Wilson, Department of Communication Studies, Northwestern University, 1881 Sheridan Road, Evanston, IL 60208-0818. Human Communication Research, Vol. 25 No. 1, September 1998 64-96 0 1998 International Communication Association

64

Wilson et al. / IDEN’ITIY IMPLICATIONS 65

ance (Baxter, 1984; Dillard, Segrin, & Harden, 1989; Kim & Wilson, 1994; Leichty & Applegate, 1991; Lim & Bowers, 1991; Tracy, Craig, Smith, & Spisak, 1984).

Scholars frequently draw on Brown and Levinson’s (1978,1987) polite- ness theory to explain the potential complexity of compliance gaining. Politeness theory maintains that attempts to alter another individual’s behavior are inherently face threatening. Because of this, message sources often use politeness to balance their competing desires to be clear about what they want and to support the other party‘s face. Although politeness theory offers one account for how threats to face arise as individuals seek compliance, the theory contains problematic assumptions. In this article, we challenge and revise Brown and Levinson’s (1987) analysis of face threatening acts. Specifically, we (a) incorporate the concept of influence goals or commonly understood reasons why people seek compliance (e.g., to give advice, to ask favors) and (b) treat the defining conditions for speech acts as tacit knowledge that individuals use to define situations. Different threats to face are associated with different influence goals in a manner that can be predicted when the defining features for seeking compliance are overlaid onto specific influence goals. By analyzing the identity implications of influence goals, we provide a more precise ration- ale for how face threats arise as individuals seek compliance.

Our manuscript contains two major sections. The first section chal- lenges Brown and Levinson’s (1987) analysis of why seeking compliance is face threatening and presents our revised analysis of what threats to identity may arise as individuals seek compliance. The second section reports a study that tested predictions about how perceived face threats, interaction goals, and message qualities vary as young adults pursue various influence goals with same-sex friends.

BROWN AND LEVINSONS ANALYSIS OF FACE-THREATENING ACTS

Brown and Levinson (1978,1987) propose politeness theory to account for surprising regularity in the ways that speakers formulate requests (and other speech acts) across culture. According to the theory, individuals are emotionally attached to their face (see Goffman, 1967). Brown and Levinson (1987) subdivide face into two wants labeled negativeface and positiveface. Negative face is the desire to maintain one’s own autonomy. Individuals in any culture want to be shown proper deference and respect and not have their privacy and space invaded, their resources spent, and their actions restricted without just cause. Positive face is the desire to have one’s attributes and actions approved by sigruficant others. Professors

66 HLJMAN COMMUNICATION RESEARCH / September 1998

may want students to perceive them as competent, organized, and inter- esting, whereas parents may want their children to perceive them as loving, supportive, and fair. Desired attributes vary across roles and cultural definitions of roles, but all persons want approval and ratification from significant others.

Aside from maintaining face, Brown and Levinson (1987, p. 61) argue that relational interdependence creates motives for supporting other peo- ple‘s face. Conversationalists who threaten another party’s face risk retali- atory attacks on their own. Despite the interdependent nature of social relations, however, individuals often perform actions that threaten face.

Brown and Levinson (1987, p. 65) argue that many speech acts intrin- sically are face-threatening acts (FTAs), meaning that by definition they run contrary to the face wants of the speaker or hearer. As one example, consider directives or the class of speech acts designed to get a hearer to perform an action he or she otherwise would not have performed, such as requests, recommendations, and commands (see Searle, 1976). Direc- tives are defined by several constitutive rules or logical preconditions, including that the speaker must perceive and/or feel (a) need for action (i.e., there must be a reason why the desired actionneeds to be performed); (b) need for directive (i.e., it should not be obvious that the hearer already was about to perform the desired action); (c) ability (i.e., it should be plausible that the hearer is able to perform the desired action); (d) willing- ness/ obligation (i.e., it should be plausible that the hearer might be willing or obligated to perform the desired action); (e) rights (i.e., the speaker should possess the rights to issue the directive in this situation); and (f) sincerity (i.e., the speaker sincerely should want that the hearer perform the specified action [Labov & Fanshel, 1977; Searle, 19691). Based on these defining conditions, Brown and Levinson (1987, pp. 65-66) claim that directives intrinsically threaten the target’s negative face. A college student who asks a friend for a ride to class, for example, presumes that the friend (a) currently did not plan to offer a ride (otherwise there is no need for a request) and (b) might be willing to provide a ride (otherwise there is no sense in requesting). By making assumptions about what the friend was planning and what the friend may be willing to do, the student constrains her friend’s autonomy (Brown & Levinson, 1987, p. 144).

Based on constitutive rules, Brown and Levinson (1987) distinguish ”between acts that threaten negative face and those that threaten positive face’’ (p. 65) as well as “between acts that primarily threaten H‘s [the hearer’s] face and those that threaten primarily S’s [the speaker’s own] face” (p. 67). Table 1 presents examples that intrinsically threaten each of these four face wants.

Brown and Levinson’s (1987) classification of requests and other direc- tives as intrinsic FTAs has important implications for the study of compli-

Wilson et al. / IDENTI” IMPLICATIONS 67

TABLE 1 Examples of Brown and Levinson’s Four Types of Intrinsic Face-Threatening Acts

Acts That Primarily Threaten Sppnker‘s Face Hearer‘s Face

Positive face Apologies Criticisms

Negative face Promises Requests Confessions Insults

Offers Recornmenda tions

SOURCE: Brown and Levinson (1987, pp. 65-68). Copyright 1994 by Cambridge University Press. Adapted with permission.

ance gaining, which involves attempts to induce a target to perform a desired behavior that otherwise would not have occurred (Seibold et al., 1994, p. 546). Compliance-gaining messages often contain speech acts in addition to a directive, but those additional acts frequently support pre- conditions for the directive (see Tracy et al., 1984, pp. 532-533). When message sources hint, the directive only is implied. But sources always adhere to the rules for directives when seeking compliance by encourag- ing the target to perform a behavior that he or she otherwise did not plan to perform. Thus, sources plausibly might worry about constraining the target’s autonomy in any compliance-gaining situation.

Intrinsic FTAs are not all equally face threatening. Brown and Levinson (1987, pp. 74-84) claim that the amount of face threat created by any speech act is a function of three variables: relational distance, power, and a culturally defined ranking of the degree of imposition. Thus, requesting that one’s boss take over a time-consuming project is more imposing on the target’s negative face than is requesting that one’s best friend open a door.

Brown and Levinson (1987, p. 73) argue that as the size of an FTA increases, speakers are more likely to use polite forms of address. Accord- ing to their theory, politeness falls into two main categories: (a) indirect- ness, in which the speaker uses hints to communicate his or her intent implicitly and (b) redressive actions, in which the speaker uses accounts, compliments, downgrades, and other linguistic forms to minimize threats to the hearer’s approval and/or autonomy.

Politeness theory has been both extremely heuristic and widely criti- cized (for xwiews, see Brown & Levinson, 1987; Lim & Bowers, 1991; Tracy & Baratz, 1994; Wilson, 1992). We now examine two problems with Brown and Levinson’s (1987) account of whose face is threatened during compliance- gaining interactions.

One difficulty is that politeness theory does not predict when seeking compliance will create multiple face threats, because the theory assumes

68 HUMAN COMMUNICATION RESEARCH / September 1998

that any speech act intrinsically threatens one and only one type of face. Aside from constraining the target's negative face, directives also can threaten the target's positive face as well as the speaker's own desires for autonomy and approval (Craig, Tracy, & Spisak, 1986; Penman, 1990; Wilson, Kim, & Meischke, 1991-1992). As an example, imagine a student who said "You need to start putting forth some effort" to another group member who had failed to do his or her fair share of a class project. By uttering this directive, the student simultaneously attempts to constrain the target's future behavior (negative face) and communicates disap- proval of the target's current effort (positive face). Unlike constraint on the target's autonomy, however, other face threats are not intrinsic to direc- tives. A speaker uttering Searle's (1976) well-known directive, "Please pass the salt," places minor constraint on the target's autonomy but, in most cases, does not incur a debt or communicate any disapproval. Under what conditions do directives only constrain the target's autonomy, and under what conditions do they create multiple face threats? By focusing only on intrinsic FTAs, Brown and Levinson (1978) have little to say about other potential face threats arising from directives.

A second difficulty is that politeness theory presents an acontextual view of face threats. Although they analyze the broad parameters of relational distance and power, Brown and Levinson (1987) offer little insight about face concerns that arise within specific institutions and relationships. To develop a more contextualized view, scholars have ex- plored how threats to face are associated with task goals, roles, and procedures that define talk contexts, such as intellectual discussion (Tracy & Baratz, 1994; Tracy & Naughton, 1994), courtroom discourse (Penman, 1990), and labor/management negotiation (Wilson, 1992; Wilson, Meis- chke, & Kim, 1991). Although useful, work to date has been conducted primarily within formally structured contexts such as colloquia, court- rooms, and bargaining tables, which are recognizable due to well-defined roles and procedures as well as shared expectations about appropriate times and topics for talk.

To complement prior research, we explore how face threats arise and get managed during more informal and less structured interactions, such as compliance-gaining episodes between friends. Although an episode in which a college student asks a friend for a ride is not governed by strict task procedures and global rules for turn allocation, both participants nevertheless recognize what type of episode is unfolding if the student begins with the prerequest "Can I ask you a favor?" This episode is recognizable due to shared expectations associated with the influence goal that defines the interaction. By analyzing what is implied when the defining conditions for seeking compliance are framed by specific influ-

Wilson et al. / IDENTITY IMPLICATIONS 69

ence goals, we provide a more contextualized view of the range of face threats that potentially may arise during compliance-gaining interactions.

A REVISED ANALYSIS OF FACE-THREATENING ACTS

According to our revised analysis, message sources idenhfy likely threats to their own faces as well as the target's face based on their tacit knowledge of two factors: (a) specific influence goals (e.g., giving advice vs. asking favors) and (b) the defining conditions for seeking compliance (i.e., the constitutive rules for directives). Specific influence goals function as primary goals (Dillard, 1990) to enable participants to recognize what type of compliance-gaining episode is unfolding. We treat the rules for directives as a template that can be overlaid onto specific influence goals to predict possible threats to face. In this section, we explicate the concept of primary goal, describe three specific influence goals, and discuss the role of constitutive rules in anticipating face threats.

Primary Goals and Defining the Situation

Primary goals drive planning and action (Dillard, 1990). During a compliance-gaining episode, the primary goal is a d e s k to influence the target's behavior. This is the goal that motivates the message source to speak and, hence, explains why the interaction is taking place. The pri- mary goal "brackets the situation. It helps segment the flow of behavior into a meaningful unit; it says what the interaction is about'' (Dillard et al., 1989, p. 21). The primary goal helps both parties recognize what is going on and signals expectations about each party's identity, rights, and obli- gations (see Goffman, 1959). Secondary goals are cross-situational con- cerns that shape and constrain how individuals pursue the primary goal. Wanting to maintain one's own face and to support the target's face are examples of secondary goals.'

Everyday actors recognize that distinct types of influence goals can motivate them to seek compliance; hence, scholars have proposed typolo- gies of influence goals (e.g., Cody, Canary, & Smith, 1994; Dillard, 1989; Kipnis, Schmidt, & Wilkinson, 1980; Rule, Bisanz, & Kohn, 1985). Cody et al. (1994) view each influence goal as a different point of intersection between multiple situational dimensions. For example, compliance-gaining situations defined by the influence goal of enforcing an unfulfilled obliga- tion typically involve high rights to persuade, low target dominance, and anticipated resistance (Cody et al., 1994; Dillard & Burgoon, 1985). Al- though individuals distinguish compliance-gaining situations in terms of influence goals, the face concerns associated with particular influence

70 HUMAN COMMUMCATION RESEARCH / September 1998

goals have not been explored. Prior to undertaking such an analysis, we describe three specific influence goals in greater detail.

Description of Three Influence Goals

This section examines the influence goals of giving advice, asking favors, and enforcing obligations. These goals (a) are present in most typologies of influence goals and hence appear meaningful to everyday actors and @) differ along several dimensions that have implications for both parties' face (Cody et al., 1994; Dillard, 1989; Kipnis et al., 1980; Rule et al., 1985). Our characterization of these goals is based on speech act analyses of directives as well as empirical investigations of situational dimensions associated with these goals. Keep in mind that we characterize giving advice, asking favors, and enforcing obligations as three different influence goals and not as three forms of directives. Message sources use a directive explicitly or implicitly to pursue any influence goal; however, sources can pursue each of these influence goals via more than one form of directive (e.g., commands, recommendations, requests, suggestions).

Giving advice is a distinct influence goal because the message source (a) perceives that the target is choosing a less-than-optimal course of action, @) perceives that his or her recommendation benefits primarily the target as opposed to him- or herself, and (c) feels motivated to advise out of concern for the target's well-being (Searle, 1969; Wardaugh, 1985). An individual who exclaims "I thought my friend was giving me genuine advice, but he was just manipulating me" implies that these conditions have not been met. The term advice also implies that the target ultimately has choice about whether to follow the recommendation (Cody et al., 1994; Dillard, 1989).

Asking favors is a distinct goal because (a) the request benefits primar- ily the message source (self) rather than the target (other), and (b) the target is not strictly obligated to comply (Dillard, 1989; Tracy et al., 1984). By asking "Could you do me a favor?" a source signals that he or she is about to make a personal request that goes beyond what the target normally is expected to do. Favors include requests to borrow objects as well as to obtain assistance. Because the request benefits themselves, sources typi- cally seek favors from at least moderately intimate targets (Cody et al., 1994; Roloff, Janiszewski, McGrath, Bums, & Manrai, 1988).

Enforcing an unfulfilled obligation is a distinct influence goal in that the message source perceives (a) the target has a duty or responsibility to perform the desired action, @) the target already should have complied prior to the request, and hence (c) the source has high rights to persuade (Cody et al., 1994; Dillard & Burgoon, 1985). The target's prior obligation may stem from a promise, a contractual agreement, a collective responsi-

Wilson et al. / IDENTITY IMPLICATIONS 71

bility, or a relational or social norm (Roloff et al., 1988; Wilson & Kang, 1991). Message sources typically enforce obligations with targets who possess equal or less power than themselves (Cody et al., 1994).

Although individuals distinguish advice, favor, and obligation situ- ations, message sources in all three encourage the target to perform an action that otherwise might not occur. Given that this occurs during any compliance-gaining episode, our analysis must take account of what the preconditions for directives imply when sources pursue each of these three influence goals.

Constitutive Rules and Potential Threats to Face

Brown and Levinson (1987) analyze constitutive rules to idenhfy face threats intrinsic to speech acts in any context. In contrast, we treat consti- tutive rules as tacit knowledge that individuals themselves use to define and create context. When seeking compliance, a message source implicitly asserts that the rules for directives are in effect and accurately characterize the current reality. By giving advice, for example, a source implies that the target needs advice and that the source has the right to provide it. The target may not agree with these assumptions. Indeed, defining the situ- ation in this fashion could threaten either party’s positive face: Why does the target need advice? Is the source butting in where he or she has no right? We treat the constitutive rules for directives as stases or possible areas of contention about how to define influence situations (see Jacobs & Jackson, 1989; Kline, 1979; Streeck, 1980).

By treating constitutive rules as tacit knowledge, we can identify po- tential as well as intrinsic threats to face. Potential face threats refer to the entire range of threats to both parties’ identities that are relevant within a particular type of compliance-gaining situation. Potential threats plausi- bly could but not necessarily will arise in any specific situation of that type. As we will show, message sources do not always appear nosy when giving advice, but concerns about appearing nosy are pertinent to advice situ- ations in a way that they are not within favor situations.

To idenhfy potential face concerns associated with giving advice, ask- ing favors, and enforcing obligations, we analyze what is implied when a message source asserts that the rules for directive are in effect while pursuing each influence goal. Table 2 presents a list of concerns for each party‘s face and shows how they can be tied to the rules for directives.

Similar to Brown and Levinson (1987, we assume that pursuing any influence goal intrinsically places some constraint on the target’s auton- omy. However, degree of threat to the target’s negative face should vary depending on several factors, including the specific influence goal. For example, presuming that a target is willing to help may constrain the

72 HUMAN COMMUNICATION RESEARCH / September 1998

TABLE 2 Constitutive Rules for Directives and Potential

Face Concerns Associated With Three Influence Goals

Influence Goal

Rules for Directives Advice Favor Obligation

Need for action Does the target understand what needs to be done?

Why does the target Need for directive need advice?

Ability Is the target capable of carrying out the advice?

Willingness Is the target willing to listen to the advice?

Rights Does the source have the right to give this advice?

Sincerity

Why does the source need help in this situation?

Is the target willing to help?

Does the source have the right to ask this favor?

Is the source willing to help in return?

Does the target understand what needs to be done?

Why does the target have to be asked to comply?

Is the target capable of hlfilling the obligation?

fulfill his or her obligations?

Is the target willing to

target’s behavior and resources more than presuming that he or she simply is willing to listen (see Table 2). Aside from intrinsic constraint on the target’s autonomy we predict that other potential face threats also will be associated with each influence goal.

Besides constraining the target’s negative face, message sources should perceive potential threats to the target’s and to their own positive face when giving advice. By giving advice, message sources presume that the target otherwise would have chosen a less-than-optimal course of action (or there is no need for advice) and thereby risk implying that the target lacks competence, foresight, or motivation (Goldsmith, 1992). By giving advice, message sources also presume that they know the optimal course of action for the target, that the target is willing to listen, and that they have the right to advise (see Table 2). A message source who makes these presumptions too liberally risks appearing to be a know-it-all and/or to be someone who is nosy and who butts in to other people’s affairs (Wardaugh, 1985). In contrast to these face threats, message sources should not perceive that they have constrained their own negative face by giving advice, because they are motivated primarily out of concern for the target rather than for themselves and because the target can choose whether to follow their recommendation (see Table 2).

When asking favors, message sources should perceive strong constraint on the target’s negative face. Asking a favor presumes that a target may

Wilson et al. / IDENTITY IMPLICATIONS 73

be willing to comply with a request even though it primarily benefits the source. Message sources also should perceive potential threats to their own positive and negative face when asking favors. By asking a favor, message sources imply that they cannot handle the situation alone (or there would be no need for a request) and thereby risk appearing lazy rather than self-reliant (Craig et al., 1986). Indeed, people at times refrain from seeking social support out of fear of appearing as though they ask for help even when they should be able to handle their own problems (Goldsmith, 1992). Assuming that they are sincere, message sources who ask favors also may incur a debt to reciprocate help at a later time and hence constrain their future autonomy (Goldsmith, 1992; Roloff et al., 1988). In contrast to these potential threats, message sources should not perceive threats to the target’s positive face when asking favors. Sources ask favors of targets who possess needed abilities or resources; moreover, criticizing the target would only reduce the possibility that he or she would grant the favor. After the initial request for a favor, the target’s positive face may be threatened if he or she refuses to help, especially if the target clearly has the ability to comply and shares a close relationship with the message source (see Table 2). Under these conditions, the target may appear to be someone who cannot be counted on for help in times of need (Roloff et al., 1988). The initial request for a favor itself, however, should not threaten the target’s positive face.

Finally, message sources should perceive strong constraint on the tar- get’s negative face when enforcing obligations, because they are implying that the target has a duty to comply. When enforcing unfulfilled obliga- tions, message sources also should perceive potential threats to the target’s positive face. Sources are asking the target to perform a behavior that, by definition, the target already was obligated to perform, and hence, there should be no need for a request (see Table 2). Under these circumstances, even indirect requests can raise questions about why the target has not yet complied and imply that he or she is incompetent or lazy (Labov & Fanshel, 1977; Wilson et al., 1991/1992). In contrast to the target’s positive face, message sources should not perceive threats to their own positive or negative face when enforcing unfulfilled obligations. Sources typically believe that they have strong rights to persuade, and attention is focused on why the target has yet to fulfill his or her obligation (Cody et al., 1994; Wilson, 1990; Wilson & Kang, 1991). Sources also should not feel indebted to the target when enforcing obligations because the target already should have complied without the request.

In sum, message sources, when seeking compliance, identify potential threats to the target’s as well as to their own face based on tacit knowledge of the specific influence goal and the defining conditions for directives. We have shown how multiple yet distinct face concerns should be associ-

74 HUMAN COMMUNICATION RESEARCH / September 1998

ated with giving advice, asking favors, and enforcing obligations by analyzing what it means to seek compliance within situations defined by each of these three influence goals.

No single study could redeem all of the claims made in this section. As an initial step in evaluating our revised analysis, we report a study that tests predictions about how face threats will arise and be managed within compliance-gaining episodes involving a same-sex friend.

DERIVING HYPOTHESES FROM THE REVISED ANALYSIS OF FTAs

This section offers predictions about how perceived face threats, inter- action goals, and message qualities will differ as young adults seek compliance from same-sex friends. Our revised analysis speaks directly to how face threats arise as message sources give advice, ask favors, or enforce obligations and, hence, offers straightforward hypotheses about how perceived face threats will differ across situations defined by these influence goals. Similar to Brown and Levinson (1987), we assume that message sources typically have motives to maintain their own face as well as to support the target’s face and design their compliance-gaining mes- sages in part to pursue these goals. Thus, our revised analysis suggests how interaction goals and message qualities also will differ across advice, favor, and obligation situations. Before deriving hypotheses, we explain our decision to examine compliance-gaining episodes involving same-sex friends in this initial study

Although we have stressed that face threats are contextually bound, our analysis of the identity implications of influence goals has been framed quite broadly with regard to relationships. In the study reported below, we examine how young adults seek compliance from same-sex friends for three reasons. First, friendships are especially important for young adults as they make transitions, such as attending college, working full-time, and/or moving from their family home (Rawlins, 1992). Second, friends are frequent targets for young adults’ compliance-gaining appeals, with same-sex friends being more common targets than opposite-sex friends (Cody et al., 1994; Rule et al., 1985; Wilson & Kang, 1991). Third, face concerns should be salient during compliance-gaining episodes in- volving friends. Within middle-class American culture, friends expect each other to say what they really think rather than what the other wants to hear (Katriel & Philipsen, 1981), but such openness simultaneously creates the need to protect friends from disapproval (Rawlins, 1992). Ideally, friends can rely on each other for help, but this simultaneously creates a need to avoid imposing too much on friends (Rawlins, 1992).

Wilson et al. / IDENTITY IMPLICATIONS 75

Hypotheses Concerning Perceived Face Threats

As young adults seek compliance from friends, our revised analysis predicts that they will perceive different face threats in situations defined by different influence goals. We present five hypotheses about differences in perceived face threats across advice, favor, and obligation situations, each of which is tied to potential face threats summarized in Table 2.

Seeking compliance intrinsically constrains the target’s autonomy; hence, message sources should perceive moderate risk of this face threat in situations defined by all three goals. Giving advice, however, presumes only that the target is willing to listen.

H1: Sources perceive greater constraint on a friend’s negative face when asking favors or enforcing obligations than when giving advice.

Regarding the target’s positive face, giving advice can imply that the target lacks foresight, whereas enforcing an obligation implies that the target already should have complied.

H2 Sources perceive p a t e r risk of hatening their friend’s desire for approval when giving advice or enforcing obligations than when asking favors.

Given that sincerely wanting a favor implies a willingness to help in return, a third hypothesis is proposed.

H3: Sources perceive greater risk of constraining their own future autonomy when asking favors than when giving advice or enforcing obligations.

The next two hypotheses pertain to different aspects of the message source’s own positive face.

H4: Sources perceive greater risk of appearing lazy when asking favors than

H5: Sources perceive greater risk of appearing nosy when giving advice than when giving advice or enforcing obligations.

when asking favors or enforcing obligations.

H4 and H5 assume that persons desire approval from friends, but the undesirable attribute that they wish to avoid (i.e., being seen as lazy vs. nosy) depends on the influence goal.

Given our reasoning so far, we do not expect that H1 through H5 will be moderated by biological sex. According to our revised account, people identify likely threats to face based on their implicit knowledge of situ- ational dimensions that distinguish advice, favors, and obligations as well as of the constitutive rules for directives. Men and women should share similar conceptions of what counts as advice or favors and possess similar

76 HUMAN COMMUNICATION RESEARCH / September 1998

understanding of the logical preconditions for performing directives. If women tend to perceive greater levels of status discrepancy or social distance within their same-sex friendships than do men, then it is possible that women on average will perceive higher levels of face threat than men across all three influence goals (see Brown & Levinson, 1987; Holtgraves & Yang, 1992). Even if women on average perceive that seeking compliance is more face threatening than do men, we still anticipate that both sexes will perceive different types of face threats as they pursue different influ- ence goals. Influence goals and participant sex should not interact in their effects on perceived face threats.

Hypotheses Concerning Interaction Goals

Message sources often pursue primary and secondary goals when seeking compliance. Our revised analysis of face threats has implications for the importance attributed to secondary goals. According to Brown and Levinson (1987, pp. 68-73), as face threat increases so does a speaker's motivation to protect face. Based on this reasoning, if message sources perceive that different face threats vary across advice, favor, and obliga- tion situations, then their importance ratings for secondary goals associ- ated with protecting face also should vary. Because sources should per- ceive greater threat to their own autonomy when asking favors than when giving advice or enforcing obligations (H3), for example, we also predict that they will place greater importance on maintaining their own auton- omy in favor than in advice or obligation situations (H9).

Despite such similarities, there are two reasons why our predictions for secondary goals do not exactly parallel those for perceived face threats. First, young adults may not wish to redress every ETA when seeking compliance and at times may even want to attack their friend's face (Craig et al., 1986; Wilson, 1990). When enforcing obligations, individuals may have limited desire to protect their friend's face because the target is violating responsibilities and may be taking advantage of them. Second, young adults sometimes may want to support their friend's face even though the friend's face is not threatened. When asking favors, individuals may want to support the friend's positive face to ingratiate him or her into doing the favor rather than to mitigate against any perceived face threat (Craig et al., 1986). In sum, face threats do not always lead to secondary goals, and sources can have reasons for fonning secondary goals other than face threats. These two caveats are reflected in the following predictions.

Favors and obligations impose heavily on the target, but sources are less concerned about imposing if the target already should have complied.

H6 Message sources express greater concern about supporting their friend's auton- omy when asking favors than when giving advice or enforcing obligations.

Wilson et al. / IDENTITY IMPLICATIONS 77

Advice can imply that a target lacks foresight, whereas favors may be gained through ingratiation.

H7: Sources express greater concern about supporting their friend's desire for approval when giving advice or asking favors than when enforcing obligations.

Unfulfilled obligations can create anger toward the target for taking advantage.

H8: Sources want to attack their friend's face more when enforcing obligations than when giving advice or asking favors.

Due to the norm of reciprocity,

H9: Sources express greater concern about maintaining their own autonomy when asking favors than when giving advice or enforcing obligations.

Because they may appear nosy when giving advice or lazy when asking favors,

H10: Sources express greater concern about maintaining their own positive face when giving advice or asking favors than when enforcing obligations.

Finally given that they perceive high rights to enforce obligations,

H11: Sources place greater importance on the primary goal of gaining compliance when enforcing obligations than when giving advice or asking favors.

Once again, we do not expect biological sex to qualify H6 through H11. European Americans stereotypically associate masculine with task concerns/ autonomy and feminine with relational concems/connection (Bern, 1993; Buzzanell, 1994). Reminiscent of these gender stereotypes, Dallinger and Hample (1994) found small but significant biological sex differences in a large-scale study (N = 1471)-such that college-aged men reported greater concem than women about accomplishing influence goals effectively, whereas women reported greater concern about the secondary goals of maintaining face and relationships. Despite any sex differences across influence goals, however, we expect both women and men to vary their importance ratings for primary and secondary goals in similar fashion between advice, favors, and obligation situations.

78 HUMAN COMMUNICAnON RESEARCH / September 1998

Hypotheses Concerning Message Qualities

If interaction goals drive planning and action (Dillard, 1990), then young adults should alter how they seek compliance depending on whether they are giving advice, asking favors, or enforcing obligations. We expect that five message qualities will vary across situations defined by different influence goals: (a) do/do not do the FTA, (b) reason giving, (c) approval, (d) pressure, and (e) persistence. Do/do not do the FTA asks whether message sources actually would talk to their friend about the issue or i f they would say nothing. Persistence asks whether message sources would continue seeking compliance if the friend resisted their initial request. Assuming that they do seek compliance, we expect sources to vary the degree to which they give explicit reasons to justify their initial request (reason giving), make the target feel positive or supported as opposed to negative or attacked (approval), and make the target feel he or she must do what they ask as opposed to giving the target a choice about whether to comply (pressure).

Specific predictions about how each of these five message qualities will differ across advice, favor, and obligation situations follow largely from our predictions for interaction goals. Given that message sources have greater concern about the primary goal and less about secondary goals when enforcing obligations, we predict the following:

Sources actually do the ETA more often (H12), exert stronger pressure (H13), and persist more after encountering resistance (H14) when enforcing obli- gations than when giving advice or asking favors.

To justlfy imposing on the target and/or to avoid looking lazy or nosy, we also expect the following:

H15: Sources give more reasons when asking favors or giving advice than when enforcing obligations.

Favors can be gained through ingratiation, advice can imply that the target’s current actions are less than optimal, and unfulfilled obligations can create a desire to attack face.

H16: Sources express more approval when asking favors than when giving advice and when giving advice than when enforcing obligations.

We do not expect biological sex to quahfy H12 through H16. Although men and women may differ in their overall use of reasons, approval, and pressure (e.g., Baxter, 1984; Holtgraves & Yang, 1992), we still expect both

Wilson et al. / IDENTITY IMPLICATIONS 79

sexes to vary these message qualities in similar fashion as they pursue different influence goals.

METHOD

Pretests: Development of Hypothetical Scenarios

This study evaluated whether perceived face threats, interaction goals, and message qualities differed as young adults imagined giving advice, asking favors, or enforcing obligations with same-sex friends. To enhance generalizability, we developed four hypothetical compliance-gaining sce- narios that instantiated each of the three influence goals (advice, favors, obligations) in two pretests. In the first pretest, 30 students from an introductory communication course at a large midwestern university (hereafter referred to as University A) recalled real-life examples of one advice, one favor, and one obligation situation in which they had been the message source. Participants received extra credit for describing situations that had occurred recently, involved a same-sex friend whom they had known for at least 6 months, and involved attempts to change the friend’s behavior rather than only his or her opinions. We also searched prior compliance-gaining studies that employed a large number of scenarios (e.g., Leichty & Applegate, 1991; Tracy et al., 1984) for appropriate situ- ations. From this corpus of situations, we initially selected six scenarios that appeared to be clear examples for each of the three influence goals (for a total of 18 scenarios).

To assure that these 18 scenarios were perceived as intended, Pretest 2 had 146 undergraduates from introductory communication courses at a second, large midwestem university (referred to here as University B) each evaluate one of the 18 scenarios (i.e., between 7 and 9 participants evalu- ated each scenario). After reading their situation, participants rated it along eight dimensions identified by Cody, Greene, Marston, OHair, Baaske, and Schneider (1985): intimacy, power, source benefits, target benefits, rights to request, target resistance, relational consequences, and source anxiety. Each dimension was assessed via three-item, 7-point Likert scales adapted from Cody et al. (1985). Participants also rated whether their primary motivation for seeking compliance was concern for the target (e.g., “The major reason I am making this request is that I’m worried about Pat’s [the target’s] well-being”) and whether the target already should have complied prior to the source’s request (e.g., ”Pat already should have done what I asked”) on three-item, 7-point Likert scales (1 = strongly disagree, 7 = strongly agree). Finally, participants completed a three-item semantic differential measure of whether the situation was

80 HUMAN COMMUMCATTON RESEARCH / September 1998

TABLE 3 Perceived Dimensions for Situations

Defined by Three Influence Goals (Pretest 2)

Influence Goal Situational Dimension Advice Favor Obligution F p Eta2 Alpha

Intimacy Source benefits Target benefits Rights to request Source power Target resistance Relational

consequences Source anxiety Concern for target Prior obligation Realism

5.2p (1.12) 3.Ua (1.21) 5.55’ (1.18) 4.69a (1.20) 2Ma (0.81) 4.56a (0.85)

3.50’ (1.73) 3.24 (1.54) 6.3Sa (0.82) 2.86a (1.14) 6.02 (0.71)

5.81a (0.96) 5.50b (1.34) 3.29b (1.13) 4.86’ (1.09) 3.04a (1.15) 3.20b (1.08)

4.14a (1.74) 3.92 (1.84) 1.60b (1.06) 1.74b (1.15) 5.84 (0.97)

4.52b (1.56) 9.45 5.52b (0.81) 43.34 4.03‘ (1.35) 29.22 6.43b (0.65) 29.92 3.Sb (1.14) 4.12 3.37b (1.18) 15.72

5.29b (1.31) 10.24 3.81 (1.73) 1.48 1.94b (0.98) 241.99 6.08‘ (1.12) 131.32 6.29 (0.60) 2.79

- .001 .001 .001 .001 .02 .001

.001

,001 .001 ns

ns

-

- .17 .48 .38 .38 .08 .25

.18

.83

.73

-

- -

- .85 .85 .78 .80 .67 .82

.90

.93

.96

.97

.69 - NOTE: N = 99 (33 advice, 31 favor, 35 obligation). Numbers outside parentheses are mean scores; numbers inside parentheses are standard deviations. Higher mean scores = greater intimacy, source benefits, and so forth. Mean scores with different superscripts are sigruh- cantly different by the Student-Newman-Keuls test at p < .05.

realistic (1 = implausible, 7 = plausible). Reliabilities for these measures ranged from a = .69 to .97 (see Table 3).

Based on ratings from Pretest 2, we selected four scenarios per influence goal for use in the experiment (total = 12 scenarios). For giving advice, these situations involved recommending that a same-sex friend break off a romantic relationship, stop smoking, talk to his or her boss about last minute changes in the friend’s work schedule, and stay in college as opposed to dropping out. For asking favors, the situations involved requesting that a friend loan the source $50, loan the source his or her car for an evening, give the source a 45-minute ride to the airport on two occasions, and find an efficiency apartment in which only the source would live. For enforcing unfulfilled obligations, these situations in- volved requesting that a friend clean the bathroom when it was his or her turn (the friends also were roommates), repay an overdue loan of $50, carry out his or her part of a bargain to exchange class notes, and carry out his or her overdue promise to have the oil changed in a car he or she had sold to the source. Appendix A displays the advice (stop smoking), favor (ride to airport), and obligation (exchange notes) scenarios.’

Perceptions of advice, favor, and obligation situations differed signifi- cantly in ways consistent with our earlier descriptions of each influence goal. Table 3 reports descriptive statistics, one-way ANOVAs, and post hoc analyses testing differences in perceived situational dimensions across the three influence goals. Data from the 99 participants in Pretest 2 who

Wilson et al. / IDENTITY IMPLICATIONS 81

responded to one of the 12 scenarios were collapsed across the four scenarios instantiating each influence goal. Advice, favor, and obligation situations varied primarily in terms of source and target benefits, rights to request, target resistance, concern for the target, and prior obligation (all eta’s > .20; see Table 3). Sources who gave advice were perceived as being motivated primarily by concern for the target. These scenarios possessed low source and high targetbenefits, fairly high target resistance, and low prior obligation to comply Favor scenarios had high source and low target benefits and low obligation to comply before the request was made. Unfulfilled obligation scenarios possessed high source benefits, high rights to request, and high prior obligation to comply. Scenarios instantiating all three influence goals were perceived as realistic, and realism ratings did not differ across influence goals. In sum, scenarios instantiating all three influence goals differed along several dimensions as intended.

Experimental Participants and Procedures

Experimental participants were 303 students (178 females, 123 males, 2 no response; M age = 21 years) enrolled in introductory communication courses at one of two large midwestern universities who received extra credit in return for completing a questionnaire about persuasive commu- nication. Students from University A completed the questionnaire outside of class. Of the 170 students who took questionnaires, 143 (83%) returned them to their instructors within a week. Students at University B ( n = 160) completed the questionnaire during class.

Participants each read one of the 12 hypothetical compliance-gaining scenarios, wrote exactly what they would say to the target, and reported whether they actually would confront the target and persist if the target resisted. They then completed closed-ended measures of their interaction goals as well as of perceived threats to both parties’ face. Participants from University A also completed an additional closed-ended measure not pertinent to the current report. Compliance-gaining messages were coded for degree of reason giving, approval, and pressure.

Dependent Variables

Interaction gods. After finishing the message construction task, partici- pants completed atwo-item, 7-point Likert scale assessing how important it would be for them to accomplish the primary goal of gaining compliance (e.g., ”I want to get Chris [the target] to comply with my request”). They also completed five two-item, 7-point Likert scales asking how important it would be for them to accomplish the secondary goals of (a) supporting

82 HUMAN COMMUNICATION RESEARCH / September 1998

the target's negative face (e.g., "I want to keep from imposing too much on Chris"); (b) supporting the target's positive face (e.g., "I want to make sure that Chris feels good about himself"); (c) attacking the target's face (e.g., "I want to make Chris feel a little guilty"); (d) protecting their own negative face (e.g., "I want to keep myself from feeling like I am in Chris's debt"); and (e) protecting their own positive face (e.g., ''I want to keep a positive image in Chris's eyes"). All scales were bounded by the endpoints not at all important and very important. Reliability coefficients for these measures ranged from a = .65 to .89 (see Table 4).

Perceived face threats. After completing their goal ratings, participants responded to two-item, 7-point Likert scales assessing potential threats to both parties' face. Specifically they evaluated the degree to which seeking compliance would threaten (a) the target's negative face (e.g., "Chris could feel pressured to do what I asked in this situation"); @) the target's positive face (e.g., "Making a request in this situation could embarrass Chris"); (c) their own negative face (e.g., "If he complied with my request, I would feel like I owed Chris something in return"); (d) their own positive face by making them look lazy (e.g., "By saying something in this situation, I could look like the type of person who expects my friends to do what I could do for myself"); and (e) their own positive face by making them look nosy (e.g., "I could look like I was overstepping my bounds or putting my nose where it didn't belong"). All scales were bounded by the endpoints strongly disagree and strongly agree. Reliability coefficients ranged from a = .40 to .92; several were low (see Table 4). It merits note that the primary consequence of low reliability is to attenuate the relationship between that measure and other variables. For example, finding that perceived threat to the target's negative face did not differ across influence goals would be relatively uninformative because it might be due to lack of an effect, unreliability of the measure for the target's negative face, or both. But if perceived threat to the target's negative face does differ across influence goals, then we can infer that the effect is strong indeed.

Message qualities. Written compliance-gaining messages were coded for degree of reason giving, approval, and pressure. Reasons included any statement in which the source (a) explained or justified why he or she was making the request/recommendation and/or (b) explained or justified why the target should comply or feel comfortable complying with the request/recommendation. Three coders who were masked to the research hypotheses independently analyzed the number of reasons in each of 30% (90/303) of the messages. For the three pairs of coders, Guetzkow's U (which assesses disagreement about the total number of reasons across the 90 messages) ranged from .05 to .12 and averaged .08. After this, the three

Wilson et al. / IDENTITY IMPLICATIONS 83

TABLE 4 Descriptive Statistics and A Priori Contrasts for Perceived

Face Threats, Interaction Goals, and Message Qualities Across Situations Defined by Three Influence Goals

Influence Goal A Priori Contrasts Dependent Variable Advice Favor Obligation t p 9 Alpha

Perceived face threats (Hl) Target’s negative face

Predicted contrast +2 -1 -1 Mean (SD) 3.45 (1.05) 4.43 (1.05) 4.94 (1.00) 7.16 .001 .16 .40

Predicted contrast +1 -2 +1 Mean (SD) 3.85 (1.26) 3.15 (1.41) 3.50 (1.40) 3.02 .003 .03 .63

Predicted contrast -1 +2 -1 Mean (SD) 2.09 (1.12) 5.27 (1.36) 1.81 (1.28) 20.47 .001 .58 .92

(H2) Target’s positive face

(H3) Source’s negative face

(H4) Source’s positive face (lazy)

Predicted contrast -1 +2 -1 Mean (SD) 2.65 (1.21) 3.26 (1.51) 2.08 (1.21) 5.48 ,001 .G9 .55

(H5) Source’s positive face (nosy)

Predicted contrast +2 -1 -1 Mean (SD) 3.80 (1.67) 2.86 (1.58) 1.84 (1.15) 8.09 .001 .18 .90

Interaction goals (H6) Support target’s negative face

Predicted contrast -1 +2 -1 Mean (SD) 4.66 (1.33) 5.26 (1.30) 2.82 (1.29) 9.40 .001 .23 .65

(H7) Support target’s positive face

Pwdicted contrast +1 +1 -2 Mean (SD) 5.66 (1.04) 5.51 (1.25) 3.99 (1.63) 9.45 .001 -23 .78

Predicted contrast -1 -1 +2 Mean (SD) 2.52 (1.41) 1.98 (1.11) 3.26 (1.54) 5.98 .001 .ll .69

(H8) Attack target’s face

(H9) Maintain own negative face

Predicted contrast -1 +2 -1 Mean (SD) 3.96 (1.94) 3.76 (1.87) 5.48 (1.41) -3.11 .002 NA .89

(H10) Protect own positive face

Predicted contrast +1 +1 -2 Mean (SD) 5.35 (1.24) 5.89 (1.21) 4.83 (1.37) 5.06 .001 .08 .77

(H11) Gain target’s compliance

Predicted contrast -1 -1 +2 Mean (SD) 4.08 (1.41) 4.80 (1.54) 6.10 (1.08) 9.90 .001 .25 .78

Message qualities (H12) Do the ETA

Predicted contrast -1 -1 +2 Mean (SO) 0.85 (0.35) 0.90 (0.31) 0.95 (0.22) 1.91 .057 .01 NA

84 HUMAN COMMUNICATION RESEARCH / September 1998

TABLE 4 Continued ~~~

Influence Goal A Priori Contrasts Dependent Variable Advice Favor Obligation t p 1.2 Alpha

Message qualities (H13) Exert pressure to comply

Predicted contrast -1 -1 +2 Mean (SD) 3.58 (0.72) 3.44 (0.78) 3.76 (0.70) 2.81 ,005 .03 NA

(H14) Persist after resistance

Predicted contrast -1 -1 +2 Mean (SD) 0.47 (0.50) 0.32 (0.49) 0.79 (0.41) 6.81 .001 .14 NA

Predicted contrast +1 +1 -2 Mean (SD) 3.19 (1.41) 2.44 (1.21) 2.10 (0.92) 4.72 .001 .07 NA

Predicted contrast 0 +1 -1 Mean (SD) 3.32 (0.89) 3.93 (0.64) 3.09 (0.82) 7.33 .001 .15 NA

(H16) Number of reasons

(H16) Express approval

~~ ~

NOTE: For Perceived Face Threats and Interaction Goals, N = 303 (108 Advice, 97 Favor, 98 Obligation). For Message Qualities, N = 295 (105 Advice, 94 Favor, 96 Obligation). Numbers outside of parentheses are mean scores; numbers within parentheses are standard deviations. For Interaction Goals, higher scores = greater importance. For do theface-threatening acts and persist afier resistance, 0 = no and 1 = yes.

coders divided up the remaining 213 messages and coded the number of reasons in each.

Approval referred to whether the message source tried to make the target feel positive/supported as opposed to negative/attacked. Sources communicated approval by highlighting the target’s positive qualities, emphasizing that they valued the source-target relationship, making ex- cuses for the target’s shortcomings, expressing exaggerated gratitude for compliance, and expressing concern for the target’s well-being. They communicated disapproval by emphasizing the target’s failures/short- comings, making the target feel embarrassed/guilty, threatening the tar- get with aggressive retaliation, and expressing strong negative emotions toward the target. Three different coders who were masked to the research hypotheses independently rated the overall degree of approval in each of 26% (80/303) of the compliance-gaining messages on a 1 to 5 scale (1 = strong/explicit disapproval, 3 = neutral or balanced, 5 = strong/explicit approval). Treating each of the three coders as items in a scale, Cronbach’s a = .72.

Pressure referred to whether the message source made the target feel he or she must comply as opposed to giving the target a choice. Sources exerted negative pressure by issuing commands; using threats, warnings, guilt, or debt to compel compliance; arguing the target had no choice but

Wilson et al. / IDENTITY IMPLICATIONS 85

to comply; demanding immediate compliance; and assuming that the target automatically would comply. Although more socially appropriate, they also exerted positive pressure by begging the target to comply, labeling the request as a large favor, explicitly incurring a debt to the target, and offering to repay the target (bargain). In contrast, sources reduced pressure by stressing that the target could choose not to comply; by being hesitant about the target’s ability/willingness to comply; by using miti- gated request forms (e.g., “Is there any way that you possibly might be able to”), disclaimers (e.g., “I‘m not trying to be pushy”), or qualifiers (e.g., “maybe you should do X ) ; and by indicating that they had other options if the target could not comply. Three different coders who were masked to the research hypotheses independently rated the overall degree of pres- sure in each of 36% (110/303) of the compliancegaining messages on a 1 to 5 scale (1 = very weak pressure, 5 = very strong pressure); Cronbach’s a= .83.

Immediately after writing their message, Participants completed single- item measures of do/do not do the FTA and persistence. For do/do not do the ETA, participants read ”If this were a real situation, would you actually confront Chris [the target] and talk to him about the situation?” Those answering yes were scored 1, whereas those answering no were scored 0. For persistence, participants answered “drop it” (0) or “persist” (1) to the question ”If Chris initially refused your request, would you drop the issue or persist in trying to persuade him/her?”

RESULTS

Perceived Face Threats

H1 through H5 predicted that young adults would perceive different threats to both parties’ face depending on whether they gave advice, asked favors, or enforced obligations with a same-sex friend. Means and stan- dard deviations for levels of five types of perceived face threat (target’s negative and positive face; source’s negative face; source’s positive face- lazy and nosy) appear in Table 4. To evaluate H1 through H5, a 3 x 2 factorial MANOVAcrossing influence goals (advice, favor, obligation) and biological sex (male, female) was conducted. The five measures of per- ceived face threat served as dependent variables. The MANOVA revealed a sigruficant effect for influence goals, Wilks’s lambda = .28, approximate F(10,574) = 67.72, p < .001, indicating that level of perceived face threats as a group differed significantly across situations defined by different influence goals. In contrast, neither the main effect for participant sex, Wilks‘s lambda = .99, approximate F(4, 287) = .15, ns, nor the Influence Goal x Sex interaction were significant, Wilks‘s lambda = .95, approximate

86 HUMAN COMMUNICATION RESEARCH / September 1998

F(10,574) = 1.55, ns, indicating that men and women did not differ in their perceptions of the five face threats.

To interpret the significant multivariate effect for influence goals, a priori contrasts designed to test the specific predictions made in H1 through H5 were conducted (see Table 4). Results for these planned comparisons ( t ) and indices of effect size (Y ’) are shown in Table 4. Two conclusions are apparent. First, perceived level for all five types of face threat differed significantly across influence goals. Second, these per- ceived differences are consistent with H1 through H5. As one example, we reasoned that young adults would perceive less threat to the target’s autonomy (Hl) when giving advice than when asking favors or enforcing obligations. Although participants perceived moderate levels of threat to the target’s autonomy in situations defined by all three influence goals (as politeness theory would predict), they also perceived sigruficantly less constraint on the target’s autonomy in advice than in favor and obligation situations. Indeed, the planned contrast testing our reasoning (i.e., +2, -1, -1) accounted for 16% of the variance in perceived threat to the target’s negative face (see the Y * in Table 4).

In general, the a priori contrasts in Table 4 fit the observed data well: All five contrasts are positive, and they account for 3% to 58% of the variance in perceived face threats. In several cases, indices of effect size are constrained by unreliable measurement. After correction for attenu- ation, for example, explained variance for the target’s negative face (Hl) jumps from Y‘ = .16 to yZc = .41. In sum, it appears that our revised analysis offers an accurate description of the multiple face threats that participants view as relevant to situations defined by these three influence goals.

Interaction Goals

H6 through H11 predicted that young adults would rate the importance of supporting both parties’ face differently depending on whether they gave advice, asked favors, or enforced unfulfilled obligations with a same-sex friend. Means and standard deviations for the primary goal of gaining compliance and for five secondary goals (i.e., supporting the target’s negative and positive face, attacking the target’s face, maintaining the source’s own negative and positive face) appear in Table 4. To evaluate H6 to H11, a 3 x 2 factorial MANOVA crossing influence goals (advice, favor, obligation) and participant sex was conducted in which importance ratings for the six interaction goals were dependent variables. The MA- NOVA revealed a significant effect for influence goals, Wilks’s lambda = .49, approximate F(12, 580) = 24.70, p c .001, revealing that importance ratings for the six interaction goals as a set varied across situations defined by different influence goals. The main effect for participant sex also was

Wilson et al. / IDENTITY IMPLICATIONS 87

significant, Wilks’s lambda = .95, approximate F(6, 290) = 2.90, p < .02? Consistent with our expectations, however, the Influence Goal x Sex interaction was not sigruficant, Wilks‘s lambda = .96, approximate F(12, 580) = 1.32, ns, reflecting that males and females varied their importance ratings for interaction goals similarly across advice, favor, and obligation situations.

We again conducted a priori contrasts (see Table 4) for each of the six interaction goals to interpret the sigruficant multivariate effect. As with perceived face threats, importance ratings for all six interaction goals varied sigruficantly across advice, favor, and obligation situations. In five of six cases, mean ratings were consistent with the contrasts specified in Table 4. For example, we reasoned that young adults would display more concern about supporting their friend’s positive face (H7) when giving advice (because the friend might appear to lack foresight) or asking favors (to ingratiate the friend) than when enforcing obligations. The a priori contrast testing our reasoning explained 25% of the variance in impor- tance ratings for this secondary goal.

For the goal of maintaining the message source’s own negative face (H9), data were inconsistent with our prediction. Given that participants perceived the most threat to their own autonomy when asking favors, we expected that they also would display more concern about maintaining autonomy when asking favors than when giving advice or enforcing obligations. Inspection of mean importance ratings, however, suggests that participants were most concerned about maintaining autonomy when enforcing obligations (see Table 4). Indeed, a post hoc contrast testing this suggestion (i.e., -1, -1, +2 for advice, favor, and obligation situations) fits the data well, t(298) = 6.02, p < .001, ? = .11. Why were participants concerned about protecting their own autonomy when en- forcing unfulfilled obligations? Given that their friend already should have complied, perhaps participants were unwilling to strike any further bargains or compromises. The data, but for this one exception, were consistent with our predictions, and the planned contrasts accounted for 8% to 25% of the variance in goal importance ratings.

Message Qualities

Although participants perceived face threats and prioritized goals differently depending on the influence goal, the question remained whether these differences were evident in their compliance-gaining mes- sages. We predicted that young adults would vary their likelihood of doing the ETA, using reasons, approval, and pressure, and persisting after encountering resistance across advice, favor, and obligation situations. To evaluate H12 to H16, a 3 x 2 factorial MANOVA crossing influence goals

88 HUMAN COMMUNICATION RESEARCH / September 1998

(advice, favor, obligation) and participant sex was conducted in which the five message qualities were dependent variables. The MANOVA revealed a sigruficant effect for influence goals, Wilks’s lambda = .59, approximate F(12,516) = 15.70, p c .001, indicating that the five message qualities as a set varied across situations. The main effect for participant sex also was sigxuficant, Wilks’s lambda = .93, approximate F(6,258) = 3.88, p < .002.4 Consistent with our expectations, however, the Influence Goal x Sex interaction was not significant, Wilks’s lambda = .97, approximate F(12, 516) = 0.91, ns, indicating that differences in message qualities across influence goals were not moderated by participant sex.

To interpret the signhcant multivariate effect for influence goals, a priori contrasts for each of the five message qualities were conducted (see Table4). The planned contrast for “do the FTA” (H12) was only marginally significant, reflecting that most participants said they would confront their friend. Planned contrasts for reason giving, approval, pressure, and per- sistence were all sigruficant, and mean scores fit the contrasts specified in Table 4. For example, we reasoned that young adults would express more approval of their friend (H16) when asking favors (because the friend would be helping without personal gain) than when giving advice (be- cause they implied the friend was choosing a less-than-optimal course of action) and, in turn, when giving advice than when enforcing unfulfilled obligations (because their friend had violated a mutual agreement). The a priori contrast testing our reasoning explained 15% of the variance in the degree of approval within participants’ compliance-gaining appeals. Re- garding other message qualities, participants gave more reasons (H15) to justify their request when giving advice and asking favors than when enforcing obligations. Although they provided fewer reasons, participants exerted the most pressure to comply (H13) and rated themselves as most likely to persist (H14) when enforcing obligations. In sum, participants varied features of their compliance-gaining messages as expected based on our revised analysis of face threats.

DISCUSSION

In this article, we outline a revised theoretical analysis of how threats to face arise during compliance-gaining episodes. According to politeness theory (Brown & Levinson, 1978,1987), seeking compliance inherently is face threatening. Brown and Levinson argue that most speech acts by definition threaten one (and only one) type of face regardless of context. By presuming that a target might be willing to alter his or her actions, for example, directives (e.g., requests, recommendations) intrinsically threaten the target’s desire for autonomy. Although attempts to gain

Wilson et al. / IDENTITY IMPLICATIONS 89

compliance do constrain the target's negative face, in some cases they also create threats to the target's positive face as well as the message source's own face. Politeness theory has little to say about when these additional, potential threats to face are likely to occur.

We challenge and revise politeness theory by exploring potential face threats within particular contexts. Brown and Levinson (1978,1987) focus only on the broad relational parameters of intimacy and power, whereas we examine specific reasons why message sources seek compliance. Dur- ing compliance-gaining interactions, the influence goal defines the con- text: It brackets the interaction, says what the episode is about, and signals expectations for both parties. Giving advice, asking favors, and enforcing obligations are three meaningful influence goals that differ in terms of rights to request, benefits to both parties, anticipated resistance, and prior obligation to comply. Phrases such as "Let me give you some advice" signal that recognizable episodes with predictable face threats are about to unfold.

A second difference from politeness theory is that we treat the consti- tutive rules for directives as tacit knowledge that individuals use to define situations. "Speech acts create and at the same time presuppose relational and other aspects of context. There is no unilateral relation of causation nor a one-to-one correspondence between acts and contexts but rather a flexible relation of mutual elaboration" (Streeck, 1980, p. 151). By asking a favor, for instance, message sources imply that they need help and that the target may be willing to provide it (see Table 2). Targets and third parties may not agree. Thus, we treat the constitutive rules for directives as a template that can be overlaid onto particular contexts and analyze what these preconditions imply when framed by specific influence goals. By treating constitutive rules as tactic knowledge used by participants themselves, we are able to idenhfy potential as well as intrinsic threats to face.

Aside from presenting our revised analysis, we report an initial study that tests predictions about how perceived face threats, interaction goals, and message qualities would vary as young adults gave advice, asked favors, or enforced unfulfilled obligations with a same-sex friend. Results for all three sets of dependent variables were largely consistent with our predictions. When giving advice, participants anticipated potential threats to both their friend's and their own positive face (because they might appear to be butting into their friend's affairs). Participants also displayed strong concern about maintaining both parties' positive face when giving advice. Given these secondary goals, participants were more likely to provide reasons to justify their requests when giving advice than when pursuing other influence goals.

90 HUMAN COMMUNICATION RESEARCH / September 1998

Participants who asked favors anticipated potential threats to both parties’ autonomy and to their own positive face (because they might appear too lazy to handle their own problems). When asking favors, participants displayed strong concern for not imposing too much on their friend’s autonomy and for protecting both parties’ positive face. Given these concerns, participants expressed greater approval of their friend and rated themselves as less likely to persist if their friend resisted when asking favors than when pursuing other influence goals.

When enforcing unfulfilled obligations, participants anticipated poten- tial threats to their friend’s negative and positive face. Because their friend already should have complied, however, participants displayed more concern about the primary goal of seeking compliance and less concern about secondary goals when enforcing obligations (with the unexpected exception of wanting to maintain their own autonomy). Given these concerns, participants expressed less approval and exerted greater pres- sure for their friend to comply when enforcing obligations than when pursuing other influence goals. Participants did not provide numerous reasons why their friend should fulfill their obligation, but most partici- pants indicated that they would persist at seeking compliance if the friend resisted their initial appeal (see Table 4).

We also expected that men’s and women’s perceptions of face threats as well as their interaction goals and message qualities would vary simi- larly across advice, favor, and obligation situations. Although biological sex exerted a few small but statistically sigxuficant main effects (see Notes 3 and 4), none of our research hypotheses were qualified by biological sex.

Although consistent with our predictions, this study provides only tentative support for our revised analysis of face threat. To state our position strongly, we believe that giving advice, asking favors, and enforc- ing obligations each are associated with different potential threats to face across age group, biological sex/gender, relationship type, institutional context, and culture. According to our revised analysis, persons rely on two sources of shared knowledge to identify potential face threats: (a) situational dimensions that define specific influence goals (i.e., what counts as advice or favors) and (b) preconditions that define what always is presumed by seeking compliance (i.e., constitutive rules for directives). Neither type of knowledge should vary across age, gender, relationship, or culture. Studies with U.S. participants indicate that giving advice, asking favors, and enforcing obligations are common, recognizable activi- ties within family, friendship, romantic, and work relationships (Cody et al., 1994; Dillard, 1989; Kipnis et al., 1980; Rule et al., 1985; Wilson & Kang, 1991). Advice, favors, and obligations also appear to be meaningful influ- ence goals across cultures (e.g., Fitch, 1994; Holtgraves & Yang, 1992; Kim, Hunter, Miyahara, Horvath, Bresnahan, & Yoon, 1996; Kim & Wilson, 1994). Constitutive rules also should be understood similarly across cul-

Wilson et al. / IDENTITY IMPLICATIONS 91

ture if they specify logical preconditions for performing speech acts (for debate on this point, see Brown & Levinson, 1987, pp. 25-27).

We recognize that our claims in the preceding paragraph go well beyond the data from this study, which were drawn from predominantly European American college students who imagined compliance-gaining situations with same-sex-friends. Although college-aged men and women share similar perceptions of which face threats are associated with specific influence goals, future research should evaluate whether these percep- tions remain stable across age groups, relationships, and cultures. Regard- ing relationships, we predict that persons on average will perceive greater threat to the target’s positive face when giving advice than when asking favors, regardless of whether the target is their work supervisor or roman- tic partner. Giving advice in either case implies that the target has em- barked on a less-than-optimal course of action; asking favors does not. Similarly, we predict that persons on average will perceive greater threat to the target’s negative face when asking favors than when giving advice, both with supervisors and romantic partners. Asking favors in either case presumes that the target may be willing to perform an action without direct benefit; giving advice presumes only that the target will listen. Similar to Brown and Levinson (1987), we also expect that persons in general will perceive larger total threat to the target’s face when pursuing either influence goal (advice or favors) with work supervisors than with romantic partners, because organizational superiors/subordinates are characterized by greater social distance and status difference than are romantic partners. Despite this, specific potential face threats still should be associated with advice and favors in either relationship.

Regarding culture, we predict that persons will perceive greater threat to their own face when asking favors than when enforcing unfulfilled obligations with a close friend, regardless of whether the friends live in an individualistic culture such as the United States (which emphasizes inde- pendent self-construals) or in collectivistic cultures such as Japan, Mexico, or South Korea (which emphasize interdependent self-construals; see Kim et al., 1996). Asking favors in either culture may imply that the message source lacks motivation, whereas enforcing obligations should divert attention away from the source to the issue of why the target has yet to comply. Similar to Kim et al. (1996) and Kim and Wilson (1994), we also expect that members of individualistic cultures will rate being clear and efficient in accomplishing either influence goal (favors or obligations) as more important and will rate protecting both parties‘ face as less impor- tant than will members of collectivistic cultures. In addition to this cultural difference, however, particular face threats associated with these two influence goals should be similar in either culture.

This example allows us to compare our own analysis of cross-cultural similarity with that from politeness theory. According to Brown and

92 HUMAN COMMUNICATION RESEARCH / September 1998

Levinson (1987, pp. 13-15, 33-35, 78-80), culture may influence the ”weightiness” but not the type of face threat associated with directives. Individuals from different cultures may disagree about the degree to which the same directive imposes on a target, perhaps because they hold different perceptions of the participants’ relational distance or status. Despite cultural differences in degree, directives should be seen as an ETA for only the target’s negative face in any culture.

Our revised analysis focuses attention on the context in which a direc- tive is performed. The degree of intrinsic threat to the target’s autonomy as well as a variety of other potential face threats depend on the specific influence goal. Giving advice creates one set of potential face threats for the message source and target; asking favors creates another. Cultural values and expectations may influence the likelihood that any particular face threat actually will arise in a type of situation, but culture should not affect which threats are associated with an influence goal. Persons from individualistic cultures that prioritize self-determination may be more likely to infer that a source is lazy for asking a favor than persons from collectivist cultures that prioritize interdependence and fate. But individu- als from either culture should not infer that a source is nosy for asking a favor. Being lazy is pertinent to favors in a way that being nosy is not.

Aside from examining the identity implications of influence goals across relationships and cultures, future research should explore how threats to face arise and get managed in ambiguous situations where participants disagree about what is going on (Goffman, 1959,1967). In this study, we developed and pretested hypothetical compliance-gaining sce- narios that were clear (i.e., consensually agreed on) examples of advice, favor, and obligation situations. We did this purposefully to show that differences in perceived face threat, interaction goals, and message quali- ties could be attributed to the manipulation of influence goals.

In reality, participants can disagree about which influence goal defines or should define a compliance-gaining episode. Acollege student (source) who asks a fellow group member (target) to complete work on a class project may perceive that she is enforcing an unfulfilled obligation, but the target may perceive that he is doing the group a favor by completing extra work. Using our terminology, the message source and target place the template of constitutive rules onto different influence goals (see Table 2). Given this, the message source may perceive that the target is irresponsible for not fulfilling his obligation, whereas the target may perceive that the source is pushy for pressuring him to comply with a nonobligatory request. When participants misunderstand each other’s face concerns, ambiguous situations may degenerate into heated arguments about needs, rights, and obligations with explicit attacks on face. By viewing the constitutive rules for directives as stases, our revised analysis highlights how compliance-gaining episodes can be ambiguous and suggests how

Wilson et al. / IDENTITY IMPLICATIONS 93

arguments about what a source did or did not imply in seeking compliance could be analyzed across conversational turns (see Jacobs &Jackson, 1989; Kline, 1979; Wilson et al., 1991). In sum, our revised analysis of face threats suggests several directions for future research on the identity implications of influence goals.

APPENDIX A Examples of Hypothetical Advice, Favor, and Obligation Scenarios

Giving Advice: Stop Smoking Scenario

You have known your friend, Kelly for 5 years. Kelly started smoking a few years ago and seem to be smoking a lot recently In your opinion, smoking has serious consequences. You believe that if Kelly quit smoking for a short time, Kelly would be in a better position to stop altogether. You want Kelly to quit smoking completely for a week. You speak to Kelly.

Asking a Favor: Ride to Airport Scenario

You have known your friend, Chris, for 3 years. You are going to Florida and do not want to leave your car at the airport for a week while you are gone. You want Chris to take you to the [name of city] airport for your flight, which takes off at 8 a.m. this Saturday You'd also like Chris to pick you up when you return next week. You speak to Chris.

Enforcing an Obligation: Exchange Notes Scenario

You have known your friend, Pat, for a few semesters. You both are taking classes this semester. Pat has a set of notes and old exams for a class you are taking. You also have a set of notes and old exams for a class Pat is taking. The two of you decided to exchange notes and old exams. As agreed, you gave Pat your notes and old exams 2 weeks ago. You have yet to receive anything from Pat. Now, your teacher just announced the first test. You want Pat to give you the notes and old exam for the class as agreed so you can study and get a good grade. You speak to Pat.

NOTES

1. The term primly i s not meant to imply that gaining compliance always is more important than supporting face. According to Dillard (1990, p. 47), message sources may adopt Brown and Levinson's (1987) "do not do the RA" strategy when secondary goals are more important and decide not to seek compliance at all.

94 HUMAN COMMUNICATION RESEARCH / September 1998

2. All 12 compliancegaining scenarios are available from the first author on request. 3. Across situations, small but statistically sigruhcant sex differences occurred for two of

the six interaction goals situations. Women (M = 5.48, SD = 1.26) as a group rated protecting their own positive face as more important than did men (M = 5.16, S D = 1.43), t(299) = -2.01, p < .05,J = .01; whereas men (M = 2.83, SD = 1.50) rated attacking the target’s face as more important than did women (M = 2.42, SD = 1.41), t(299) = 2 .37 ,~ < ,02, r2 = .02.

4. Across situations, small but statistically significant sex differences occurred for two of the five message qualities. Women as a group included more reasons (M = 2.72, SD = 1.32) and expressed more approval of the target (M = 3.59, SD = 0.84) than did men (argument M = 2.43, S D = 1.24; approval M = 3.22, SD = 0.87), t(282) = -1.93, p = .055, ? = .01, and t(291) = -3.62, p < .001, # = .04, respectively for reason giving and approval.

REFERENCES

Baxter, L. A. (1984). An investigation of compliancegaining as politeness. Human Communi-

Bem, S. L. (1993). The lenses ofgender. New Haven, CT: Yale University Press. Brown, l?, & Levinson, S. C. (1978). Universals in language use: Politeness phenomena. In E. N.

Goody (Ed.), Questions and politeness (pp. 56-289). Cambridge,UK: Cambridge University Press.

Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge, L K Cambridge University Press.

Buzzanell, P. M. (1994). Gaining a voice: Feminist perspectives in organizational communi- cation. Management Communication Quarterly, 7,339-383.

Cody, M. J., Canary, D. J., & Smith, S. W. (1994). Compliance-gaining goals: An inductive analysis of actors’ goal types, strategies, and success. In J. Daly & J. Weimann (Eds.), Strategic interpersonal communication (pp. 33-90). Hillsdale, NJ: Lawrence Erlbaum.

Cody, M. J., Greene, J. O., Marston, I? J., OHair, H. D., Baaske, K. T., &Schneider, M. J. (1985). Situation perception and strategy selection. In M. L. McLaughlin (Ed.), Communication yearbook 9 (pp. 391-420). Beverly Hills, C A Sage.

Craig, R. T., Tracy, K., &Spisak, F. (1986). The discourse of requests: Assessment of a politeness approach. Human Communication Research, 12,437-468.

Dallinger, J. M., & Hample, D. (1994). The effects of gender on compliancegaining strategy endorsement and suppression. Communication Reports, 7’43-49.

Dillard, J. P. (1989). Types of influence goals in personal relationships. loumal of Social and Personal Relationships, 6,293-308.

Dillard, J. P. (1990). A goaldriven model of interpersonal influence. In J. F! Dillard (Ed.), Seeking compliance: The production of interpersonal influence messages (pp. 41-56). Scottsdale, A Z Gorsuch Scarisbrick.

Dillard, J. P., & Burgoon, M. (1985). Situational influences on the selection of compliance- gaining messages: Two tests of the predictive utility of the Cody-McLaughlin typology. Communication Monographs, 52,289-304.

Dillard, J. I?, Se-, C., & Harden, J. M. (1989). Primary and secondary goals in the production of interpersonal influence messages. Communication Monographs, 56,19-38.

Fitch, K. L. (1994). A cross-cultural study of directive sequences and some implications for compliance-gaining research. Communication Monographs, 61,185-209.

Goffman, E. (1959). The presentation of selfin pveryday life. New York Doubleday. Goffman, E. (1967). Interaction ritual: Essays in face-to-face behavior. Chicago: Adrine. Goldsmith, D. (1992). Managing conflicting goals in supportive interaction: An integrative

cation Research, 10,427-456.

theoretical framework. Communication Research, 19,264-286.

Wilson et al. / IDENTlTY IMPLICATIONS 95

Holtgraves, T., & Yang, J. N. (1992). Interpersonal underpinnings of request strategies: General principles and differences due to culture and gender. Journal of Personality and

Jacobs, S., &Jackson, S. (1989). Building a model of conversational argument. In B. Dervine, L. Grossberg, 8. J. OKeefe, & E. Wartella (Eds.), Rethinking communication: Vol. 2. Paradigm exemplars (pp. 172-187). Newbury Park, CA: Sage.

Katriel, T., & Philipsen, G. (1981). “What we need is communication”: “Communication” as a cultural category in some American speech. Communication Monographs, 48,302-317.

Kim, M. S., Hunter, J. E., Miyahara, A., Horvath, A. M., Bresnahan, M., & Yoon, H. J. (1996). Individual- vs. cultural-level dimensions of individualism and collectivism: Effects on preferred conversational style. Communication Monographs, 63,28-49.

Kim, M. S., & Wilson, S. R. (1994). A cross-cultural comparison of implicit theories of requesting. Communication Monographs, 61,210-235.

Kipnis, D., Schmidt, S. W., & Wilkinson, I. (1980). Intraorganizational influence tactics: Explorations in getting one’s way. Journal of Applied Psychology, 65,440-452.

Kline, S. L. (1979). Towards a contemporary linguistic interpretation of the concept of stasis. Journal of the American Forensic Association, 16,95-103.

Labov, W., &Fanshe], D. (1977). Therapeuticdiscourse: Psychot~a~asconversation. New York Academic Press.

Leichty, G., & Applegate, J. L. (1991). Social-cognitive and situational influences on the use of face-saving persuasive strategies. Human Communication Research, 17,451-483.

Lim, T. S., & Bowers, J. W. (1991). Facework Solidarity, approbation, and tact. Human Communication Research, 17,415-449.

Miller, G. R., Boster, F. J., Roloff, M. E., & Seibold, D. (1977). Compliance-gaining message strategies: A typology and some findings concerning effects of situational differences. Communication Monographs, 44,37-51.

OKeefe, 8. J. (1988). The logic of message design: Individual differences in reasoning about communication. Communication Monographs, 55,80-104.

Penman, R. (1990). Facework and politeness: Multiple goals in courtroom discourse. Journal of Language and Social Psychology, 7,101-121.

Rawlins, W. W. (1992). Friendship matters: Communication, dialectics, and the life course. New York Aldine de Gruyter.

Roloff, M. E., Janiszewski, C. A., McGrath, M. A., Bums, C. S., & Manrai, L. A. (1988). Acquiring resources from intimates: When obligation substitutes for persuasion. Human Communication Research, 14,364-396.

Rule, 8. G., Bisanz, G. L., & Kohn, M. (1985). Anatomy of a persuasion schema: Targets, goals, and strategies. Journal of Personality and Social Psychology, 48,1127-1140.

Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge, U K Cambridge University Press.

Searle, J. R. (1976). A classification of illocutionary acts. Language in Society, 5,1-23. Seibold, D. R., Cantrill, J. G., & Meyers, R. A. (1994). Communication and interpersonal

influence. In M. L. Knapp & G. R. Miller (Eds.), Handbook of Interpersonal Communication (Vol. 2, pp. 542-588). Thousand Oaks, C A Sage.

Streeck, J. (1980). Speech acts in interaction: Acritique of Searle. Discourse Processes, 3,133-154. Tracy, K., & Baratz, S. (1994). The case for case studies of facework. In S. Tig-Toomey (Ed.),

Tracy, K., Craig, R. T., Smith, M., & Spisak, F. (1984). The discourse of requests: Assessment

Tracy, K., & Naughton, J. (1994). The identity work of questioning in intellectual discussion.

Wardaugh, R. (1985). How conversation works. Oxford, UK Basil Blackwell.

Social P ~ y ~ h o l ~ g y , 62,246-256.

The challenge of facework (pp. 287-306). Albany, NY: SUNY Press.

of a compliance-gaining approach. Human Communication Research, 4,513-538.

Communication Monographs, 61,281-302.

96 HUMAN COMMUNICATION RESEARCH / September 1998

Wilson, S. R. (1990). Development and test of a cognitive rules model of interaction goals. Communication Monographs, 57,81-103.

Wilson, S. R. (1992). Face and facework in negotiation. In L. L. Putnam & M. E. Roloff (Eds.), Communicntwn and negotiation (pp. 176-205). Newbury Park, C A Sage.

Wilson, S. R., & Kang, K. H. (1991). Communication and unfulfilled obligations: Individual differences in causal judgments. Communication Research, 18,799-824.

Wilson, S. R., Kim, M. S., & Meischke, H. (1991). Evaluating Brown and Levinson’s politeness theory: Arevised analysis of directives and face. Research on languageand Social Interaction, 25,215-252.

Wilson, S. R., Meischke, H., & Kim, M. S. (1991). A revised analysis of directives and face: Implications for argument and negotiation. In F. H. Eemem, R. Grootendorst, J. A. Blair, & C. A. Wdlard (Eds.), Proceedings ofthe second international conference on argumentation (pp. 470- 480). Amsterdam: SICSAT.