Message Production

SAGE Books

Seeking and Resisting Compliance: Why People Say

What They Do When Trying to Influence Others

Author: Steven R. Wilson

Pub. Date: 2012

Product: SAGE Books

DOI: https://dx.doi.org/10.4135/9781452233185

Keywords: politeness theory, speech acts, adjacency pairs, compliance, targeting, speech act theory, logic

Disciplines: Persuasion, Communication Studies, Attitudes & Persuasion, Social Psychology, Media,

Communication & Cultural Studies

Access Date: November 28, 2022

Publishing Company: SAGE Publications, Inc.

City: Thousand Oaks

Online ISBN: 9781452233185

© 2012 SAGE Publications, Inc. All Rights Reserved.

Discourse Perspectives on Persuasive Message Production

Has a volunteer from the American Red Cross ever telephoned and asked you to donate blood? In the United

States today, someone requires a blood transfusion every 12 seconds. Aside from lifesaving transfusions,

blood products such as plasma, platelets, and immune serum globulin are used to treat patients suffering

from cancer, leukemia, and other life-threatening diseases (American Red Cross, 1989). The Red Cross col-

lects, processes, and distributes one-half of the nation's blood supply. Although the organization is constantly

attempting to recruit new donors, the vast majority of individuals donating blood (78–91%) are repeat donors

(Piliavin, 1990). Thus the Red Cross keeps careful records of prior donors and telephones them periodically

to encourage repeat donations. Both professional and volunteer solicitors are used in this process.

Several years ago, three colleagues and I collaborated with the American Red Cross's Great Lakes Regional

Blood Center in Lansing, Michigan, to study how volunteer telephone solicitors attempted to persuade prior

donors to give blood again (Wilson et al., 1994). Participants in our study were college undergraduates who

volunteered to work one evening as telephone solicitors on behalf of the American Red Cross. Participants

initially attended a 1-hour training session conducted by the regional center's staff, at which they learned

about procedures for donating blood, uses of blood products, steps for maintaining donor confidentiality, re-

quirements for medical eligibility, and suggestions for overcoming common obstacles to donating.

After completing this training, participants returned 2 to 4 weeks later to telephone prior blood donors. Upon

arrival at the communication laboratory, each volunteer received a calling list of 20 prior donors and a sheet

of instructions. For each call, participants were instructed to follow a three-part sequence in which they asked

for permission to audiotape the call, verified background information, and asked for an appointment. Instruc-

tions for the last step suggested the following wording: “I'm calling to see if we could schedule you to donate

blood again. Can we count on you again?” Following this initial request, participants used their own judgment

about whether and how to persist in seeking to convince the prior donor to comply. Our primary interests in

the study were learning how college student volunteers decided whether to persist after prior donors resisted

their initial requests and whether volunteers persisted differently depending on the ways in which prior donors

resisted. The conversation in Example 6.1 between Mike and Mr. Jenkins provides a convenient starting point

for illustrating the concerns of discourse scholars. Looking Ahead …

This chapter reviews discourse perspectives on persuasive message production. Although diverse,

these perspectives focus on how people rely on knowledge of linguistic forms and conventions when

generating influence messages, and how they design messages in light of unfolding talk. The chap-

Example 6.1

((Phone rings; another party answers, the caller, “Mike”, asks for “Mr. Jenkins;” the other party then

goes to get Mr. Jenkins.))

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 2 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Questions Posed by Discourse Scholars

Discourse scholars would raise a number of very specific questions about audiotaped interactions such as

that displayed in Example 6.1. These perspectives focus our attention on why Mike and Mr. Jenkins say what

they do, turn by turn, as their conversation unfolds. To get a feel for this type of analysis, start by examining

the form of Mike's initial request for Mr. Jenkins to donate blood again (see lines 14–15). Why does Mike

phrase this initial request as he does? Couldn't he get to the point more quickly rather than being so wordy?

Why include the adverb just, “I'm just calling”, rather than say only “I'm calling”? Why use the phrase “to see if

I can get you uh to” rather than the less tentative “to get you to”? Or for that matter, why doesn't Mike just ask,

“Will you donate blood?” Do Mike's word choices have anything to do with the nature of his request? After

all, he is asking Mr. Jenkins to take about 90 minutes out of his schedule, drive to the regional blood center,

and donate bodily fluid. Do Mike's word choices also reflect that he is making this request of a stranger who

is older than himself?

Turning to Mr. Jenkins, why does he initially say “OK” rather than “Sure” or “No, I can't” in response to Mike's

initial request (see line 16)? If Mr. Jenkins isn't saying yes or no to Mike's request at this point, what is he

saying? And don't people usually respond to requests by saying yes (complying) or no (refusing) in the very

next turn at talk? If not, what other types of actions often follow requests?

When Mr. Jenkins, in lines 24–26, finally does give an answer to Mike's request, why doesn't he just say

“Yeah” or “Sure”? In other words, why does he mention that he's been “a little sick” in the past? Is he sick at

this time? And why does Mr. Jenkins tell Mike, a complete stranger, that he tries to donate blood twice a year?

Similarly, why does Mike ask exactly what Mr. Jenkins means by the phrase “a little sick” in lines 27–28? Does

Mike have a personal interest in Mr. Jenkins's health during the past winter? Are Mike's questions somehow

related to his initial request for Mr. Jenkins to donate blood?

Finally, why does Mr. Jenkins not seem surprised to learn that an individual can donate blood only 24 hours

after feeling better from a cold? Notice that he says “Yeah right” rather than “Oh really?” in line 47. But if Mr.

Jenkins already understands that a cold is not a reason for failing to donate blood over a long period of time,

why did he mention having been “a little sick” back in line 25?

These questions about the “Mr. Jenkins-Mike” conversation illustrate the types of issues posed by discourse

perspectives on persuasive message production. The theories reviewed in this chapter come from many

fields, including communication, philosophy, psycholinguistics, sociolinguistics, sociology, and social psychol-

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 3 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

ogy.1 Although diverse, these perspectives clarify how people use knowledge of linguistic forms and conven-

tions when generating influence messages. Discourse scholars explore how people's attempts to seek and

resist compliance are shaped and constrained by what each party already has said. They often examine how

people attempt to create and sustain desired identities for themselves, as well as their interactional partners,

through talk. Discourse scholars highlight patterns that occur at minute levels of talk, such as precise regular-

ities in the ways that people word requests and responses to requests.

The remainder of this chapter is divided into three major sections. In the first, I review the basic concepts

and assumptions with which readers need to be familiar in order to understand discourse perspectives. In this

section I cover briefly some relevant writings on conversational maxims, speech acts, and the “conversation

analysis” tradition. In the second and third main sections, I explore in more depth two concepts with special

relevance for an understanding of persuasive message production. First, message targets often state “obsta-

cles” when resisting message sources' influence attempt, and sources and targets may negotiate the feasibili-

ty of overcoming obstacles to compliance. “Obstacles” are analyzed using both speech act (Searle, 1969) and

attribution (Weiner, 1986) theories. Second, message sources, by seeking compliance, often raise threats to

face for both participants, and the forms of requests and responses are shaped in part by desires to maintain

face. Threats to face are analyzed using Brown and Levinson's (1987) politeness theory and O'Keefe's (1988)

theory of message design logics.

Background Concepts and Assumptions

Conversational Maxims

People often mean more than what they say. Consider the following example:

B's words in Example 6.2, taken literally, do not answer A's question. A asks about the location of a male

named Bill, whereas B says that a specific automobile is located outside the house of a female named Sue.

Despite this, we assume that B is attempting to answer A's question. In order to hear line 2 as an answer,

we must infer (a) that Bill owns a yellow VW, and hence (b) that Bill may be inside, or near, Sue's house. B

appears to assume that A can make these inferences. Example 6.2

1 A: Where's Bill?

2 B: There's a yellow VW outside Sue's house

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 4 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Example 6.2 shows how we make inferences that go beyond the literal meaning of words. In some cases,

however, what a speaker means has virtually no connection to what he or she says. Consider a fictional ex-

ample, which could have occurred at many points as I wrote this book:

My words, taken literally, do not seem even remotely related to my friend's question. What does the weather

have to do with this book? Did I simply fail to hear my friend's question? Possibly, but you probably already

have inferred a different meaning: I am having difficulty with the book, and I do not wish to talk about the topic

at this time. How did you arrive at this inference? How is it possible for me to imply that I do not wish to talk

about writing a book by talking about rain?

Paul Grice (1975), a British philosopher of language, has proposed a theory of conversational implicature to

explain how speakers can infer things distinct from the conventional meaning of their words. He begins with

the observation that conversation is a “cooperative” activity, meaning that it requires minimal levels of collab-

oration and coordination between participants. As Grice puts it:

Our talk exchanges do not normally consist of a succession of disconnected remarks.… They are

characteristically, to some degree at least, cooperative efforts; and each participant recognizes in

them, to some extent, a common purpose or set of purposes, or at least a mutually accepted direc-

tion. This purpose or direction may be fixed from the start … or it may evolve during the exchange.…

But at each stage, SOME possible conversational moves would be excluded as conversationally un-

suitable. (p. 45)

Given this observation, Grice argues that conversationalists share a set of overarching assumptions about

talk. Participants are expected to follow what Grice labels the “cooperative principle”, namely: “Make your

conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or

direction of the talk exchange in which you are engaged” (p. 45). Grice has formulated four principles, or

“maxims”, the following of which will lead to behavior consistent with the cooperative principle. Grice's max-

ims of quantity, quality, relation, and manner are summarized briefly in Table 6.1.

The quantity maxim pertains to the expected amount of talk. It is violated when speakers are overinformative,

as B is in Example 6.4, or underinformative, as B is in Example 6.5:

Example 6.3

1 Friend: So, how's your book coming?

2 Me: It sure is starting to rain hard outside.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 5 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Table 6.1 Grice's (1975) Four Conversational Maxims

The quality maxim pertains to the sincerity of talk, such as expectations that speakers will avoid dishonesty,

deception, hearsay, or gossip. The maxim of relation states that speakers should make relevant contributions

given the current topic and purpose(s) of talk. This maxim seems to be violated in Example 6.6, where a psy-

chiatrist tries to interview a patient who is being admitted to a hospital.2

The maxim of manner pertains to the clarity of talk, such as expectations that speakers will avoid obscurity,

ambiguity, and other factors that may hinder understanding. To sum, these four maxims “specify what partic-

ipants must do in order to converse in a maximally efficient, rational, co-operative way” (Levinson, 1983, p.

102).

Now many readers will object that these maxims, although they may describe some idealized form of talk, Example 6.4

1 A: What's up?

2 B: Well, I got up this morning, and I was feeling pretty good, but then I went downstairs to eat breakfast, and I was out of or-

ange juice—I hate it when that happens, and then …

Example 6.5

1 A: Do you know what time it is?

2 B: Yes.

Example 6.6

1 Dr.: What is your name?

2 Patient: Well, let's say you might have thought you had

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 6 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

do not accurately characterize “real” conversation. But Grice's (1975) point is different from, and more subtle

than, this. Grice will readily admit that people often do not follow the four maxims, to the letter, during every-

day conversation. His point is that when a speaker does not follow the maxims at a literal level, our initial

impulse still is to assume that, contrary to appearances, the speaker is adhering to them at some deeper lev-

el. For instance, we assume that speaker B in Example 6.2 is being cooperative even though his response in

line 2, taken literally, seems to violate the maxims of quantity and relation. Our expectations that conversation

is cooperative are so strong that we look for a way of hearing B's response as an “answer”, and we find it by

inferring that Bill must own a yellow VW.

The strength of our expectations about cooperation creates an interesting possibility, which Grice (1975) calls

“flouting” the maxims. A speaker flouts or exploits a maxim when he or she blatantly violates the principle in

order to achieve some communicative purpose. In Example 6.3, for instance, I flout the maxim of relevance

in line 2 by abruptly shifting the topic. In doing so, I assume that my friend will infer that my comment is “rel-

evant” to her question in line 1 in the deeper sense that it signals that I do not wish to discuss the topic of

writing a book.

Each of Grice's four maxims may be flouted strategically (see Grice, 1975, pp. 52–56; Levinson, 1983, pp.

109–113). Tautologies such as “Boys will be boys” and “If she does it, she does it” are totally noninformative

at a literal level and hence violate the quantity maxim; nonetheless, they communicate meaning in terms of

what they implicate. Figures of speech such as irony and metaphor often violate the quality maxim. If some-

one with whom we have been on close terms until now betrays an important confidence, we might say, “X is

a fine friend” and mean just the opposite. Transparent questions violate the relation maxim to emphasize a

point:

Finally, speakers may use excessive technical jargon intentionally, and thereby violate the maxim of manner,

to create a humorous effect (e.g., Buzzanell, Burrell, Stafford, & Berkowitz, 1996).

Grice's theory of conversational implicature offers several insights about persuasive message production. The

maxims are guidelines to which participants orient during compliance-gaining episodes. Consider the quantity

maxim. Participants hold expectations about how many reasons a message source should provide to justify a

request. Message sources often offer little or no justification when making small requests of people they know

well, such as asking to borrow a pencil from a friend (Brown & Levinson, 1987; Roloff, Janiszewski, McGrath, Example 6.7

1 A: Are you going to the bar tonight?

2 B: Is the Pope Catholic?

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 7 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Burns, & Manrai, 1988). Of course, a target may believe that a request needs more justification than the mes-

sage source initially provides, as in the following example, which involves my stepdaughter Ashlee and me:

Put differently, message sources and targets jointly negotiate how much reasoning is sufficient to warrant

compliance (Jackson & Jacobs, 1980).

Although participants orient to the maxims during compliance-gaining episodes, at times they also deviate

from them in order to accomplish other purposes. As the discussion in the upcoming section on politeness

theory shows, message sources and targets at times say more than literally needs to be said (quantity), say

things with more than one plausible interpretation (manner), and so forth in order to protect the face of both

parties. Grice specifies how speakers should converse in order to communicate in a maximally efficient fash-

ion, but communicating efficiently is only one of several goals that participants typically pursue when seek-

ing and resisting compliance (see Chapter 5). Deviations from Grice's maxims during compliance-gaining

episodes often can be understood as attempts to manage multiple conflicting goals.

Speech Acts

People not only “say” things with words, they also “do” things with words. People perform a variety of actions

through talk, such as requesting, promising, warning, describing, and asserting. Speech act theorists have

grappled with how speakers are able to make their intentions apparent through talk and how hearers are able

to recognize speakers' intentions. Like the work on conversational maxims, speech act theories can be traced

to the writings of two British philosophers of language, John Austin and John Searle.

According to Austin (1962), a speaker simultaneously does three things whenever she or he says something.

Consider lines 14–15 from the conversation between Mike and Mr. Jenkins in Example 6.1, in which Mike

says, “I'm just calling to see if I can get you uh to set up an appointment to donate blood.” At one level, Mike

is referring to objects and saying something about those objects. Mike talks about himself (I'm, I), Mr. Jenkins

(you), appointments, donating, and blood. Austin refers to talking about topics as the locutionary level of a

speech act.

At a second level, Mike is doing more than talking about topics. Surely we recognize that Mike is “requesting”

that Mr. Jenkins set up an appointment with the Red Cross. Mike is saying something that he intends Mr. Example 6.8

1 Ashlee: Can you drive me to school tomorrow?

2 Me: Why can't you take the bus?

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 8 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Jenkins to recognize as a request. Austin (1962) refers to the type of action performed by a statement as the

illocutionary level of a speech act; thus Mike's phrase has the “illocutionary force” of requesting.

At yet a third level, Mike brings about certain effects in Mr. Jenkins by saying what he does. He may make Mr.

Jenkins feel annoyed, pressured, or pleased to help. He may persuade, or fail to persuade, Mr. Jenkins to set

up an appointment. Some of these effects may be intended by Mike, whereas others may be unforeseen and

unintended. Austin refers to the range of effects brought about by a statement as the perlocutionary level of a

speech act.

Speech act theorists focus primarily on the illocutionary level of speech acts, posing questions such as, How

are we able to recognize a speaker's intentions from what he or she says? How do we know, for example, that

Mike is “requesting” in lines 14–15, rather than “promising” or “warning” Mr. Jenkins? To address this issue,

Searle (1969) distinguishes between regulative and constitutive rules. Regulative rules are social conventions

that govern preexisting forms of behavior. Rules associated with table manners, such as “Don't talk with your

mouth full” and “Start with the silverware on the outside”, are regulative rules. These rules specify “proper”

ways of eating; however, the act of eating existed prior to, and exists apart from, these cultural conventions.

A regulative rule takes the form “If Y, then X”, and someone violating such a rule is perceived as acting “incor-

rectly.”

Constitutive rules are social conventions that create and define forms of behavior, such that the behavior

does not exist apart from the rules. The rules for a game such as American football are constitutive rules.

The concept of “touchdown” is created by the rules of football and does not exist independent of those rules;

after all, could someone have scored a “touchdown” before the game of football was invented? Moreover,

one can score a touchdown only by following the rules. If I run out onto the field during halftime at a Purdue

University football game, pick up the ball, and cross into the opponent's end zone, have I scored six points for

the Boilermakers? One would not say that I have scored a touchdown “improperly”, but rather that I have not

scored a touchdown at all. A constitutive rule takes the form “Doing X counts as Y.” Someone who violates a

constitutive rule is perceived as acting “incoherently”, in the sense that it is not clear what he or she is doing

(e.g., my behavior at the Purdue football game).

According to Searle (1969), any speech act is governed by a set of constitutive rules that specify the neces-

sary and sufficient conditions for performing that speech act. Put differently, these rules specify the conditions

that must exist for a behavior to “count as” performing a speech act. Searle analyzes what conditions must

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 9 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

be assumed to exist in order for a speaker to be recognized as performing speech acts, such as requesting,

promising, warning, and advising. A set of constitutive rules for requesting, adapted from Searle (1969) as

well as Labov and Fanshel (1977), appears in Table 6.2.3

The propositional content rule sets restrictions on the content or locution of a speech act. For requests, this

rule states that the source's message must refer to a present or future action to be performed by the target

rather than by the source. It sounds sensible for me to ask my stepson, “Brendan, would you take out the

garbage?” In contrast, it would sound odd for me to say to myself, “Steve, would you take out the garbage?”

Similarly, the request “Brendan, could you take out the garbage tomorrow morning?” sounds plausible. “Bren-

dan, could you take out the garbage last week?” makes little sense. My stepson initially might interpret this

latter message as a joke, or perhaps as an indirect rebuke for his failure to take the garbage out the week

before. If neither of these interpretations is plausible, however, and I repeat my request that Brendan “take

out the garbage last week”, Brendan might infer that I have “lost touch with reality” (i.e., I am talking without

any point to my talk).

SOURCE: Adapted from Searle (1969) and Labov and Fanshel (1977).

Table 6.2 Searle's (1969) Constitutive Rules for the Speech Act of Requesting

Type of Rule Definition

Propositional content rule The message source refers to a present or future action to be performed by the target.

Preparatory rules

Need for action The message source perceives a reason the requested action needs to be performed.

Need for request It should not be obvious that the message target already was planning to perform the requested behavior.

Ability The message source believes that the target plausibly may be able to perform the requested action.

Willingness The message source believes that the target plausibly may be willing to perform the requested action.

Sincerity rule The message source sincerely wants the requested action to be performed.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 10 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Preparatory rules, sometimes also called “felicity conditions”, state restrictions about the situations that must

exist in order for an individual to perform a speech act. Table 6.2 lists four preparatory rules. First, the source

must perceive some reason the action needs to be performed. I would be violating this “need for action” rule

if the garage door at our home was open, Brendan and I both knew this, and yet I still said, “Brendan, would

you open the garage door?” Second, it should not be apparent that the target already was about to perform

the requested action. I would be violating this “need for request” rule if Brendan had gone out to the garage,

reached up, and put his finger on the garage door opener's button before I asked, “Brendan, would you open

the garage door?”

A third preparatory rule states that the source must believe that the target plausibly can perform the requested

action. I would be violating this “ability” rule if some of my papers blew under our Chevy Suburban truck and I

said, “Brendan, pick up the Suburban so that I can get my papers out from underneath it without getting dirty.”

A fourth preparatory rule is that the source also must believe that the target plausibly is willing to perform the

requested action. A source may perceive willingness because he or she believes that the target will enjoy

performing the requested action, will appreciate the need to perform the requested action, or will feel a sense

of obligation to help the source (Roloff et al., 1988). My students occasionally violate this rule, as a joke, by

“requesting” that I give them all A's (i.e., they are joking rather than requesting).

The sincerity rule states restrictions on the psychological state of the message source. For requests, this rule

states that the message source sincerely must want the requested action to be performed. This rule helps

explain why, under most conditions, we hear phrases such as “Go jump in a lake” as insults rather than as

sincere requests.

So, to return to an earlier question, how do we know that Mike is “requesting” in lines 14–15 of the Example

6.1 rather than “promising” or “warning” Mr. Jenkins? How are we able to identify the illocutionary force of

Mike's utterance? The answer, in part, lies in the constitutive rules listed in Table 6.2. Consistent with the

propositional content rule, Mike refers to a future action, “setting up an appointment”, that he hopes will be

performed by Mr. Jenkins rather than himself. Mike also inquires explicitly about Mr. Jenkins's “willingness” to

perform the desired action in line 14, when he says, “to see if I can get you”, as well as in line 23, when he

says, “Would you, uh, help us out.” Mike's utterance appears to enact the necessary and sufficient conditions

for performing a request.

One additional complication in interpreting the illocutionary force of Mike's utterance in lines 14–15 is that,

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 11 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

taken literally, it does not “request” that Mr. Jenkins set up an appointment. Notice that Mike literally does not

say, “I request that you set up an appointment to donate blood”, or even “Please set up an appointment to

donate blood.” Mike literally describes his purpose as to see what he, himself, is capable of doing, as is ap-

parent in the following fictional dialogue:

Put differently, Mike's utterance in line 12 is an indirect speech act. Indirect speech acts are messages that,

given the circumstances, imply additional or different purposes from those conventionally associated with their

linguistic form (Searle, 1975). Thus the utterance “Can you shut the door?” usually is heard as an indirect

request to shut the door, rather than, or in addition to, a literal question about the target's ability to do so. We

hear Mike as making an indirect request in lines 14–15 for several reasons, including that Mike inquires about

Mr. Jenkins's willingness (a rule for requesting) and that Mike's call would seem odd if his sole purpose were

to see what he, himself, is able to get a complete stranger to do.

Having said all this, I should add that the illocutionary force of Mike's utterance in line 12 also is apparent, in

good measure, because of the larger institutional and conversational context. Mike identifies himself as “call-

ing from the American Red Cross” in line 04, and Mr. Jenkins has donated blood with the Red Cross in the

past. Indeed, Mike confirms information about the date of Mr. Jenkins's last donation in the turns immediately

preceding line 12. Given this, we clearly expect a request for repeat donation to occur at some point in the

call. Critics have charged that Searle's (1969) version of speech act theory pays too little attention to how the

illocutionary force of an utterance is shaped by institutional roles and purposes, as well as by preceding and

projected future conversational turns (Levinson, 1983; Streeck, 1980). Searle often seems to treat “context”

as something “given” or already established prior to the utterance of a speech act, rather than as something

created, in part, through the performance of speech acts (Streeck, 1980).

Despite these limitations, speech act theory, with its emphasis on constitutive rules, provides a useful con-

ceptual framework for analyzing persuasive message production. By specifying the necessary and sufficient

Mike: I'm just calling to see if I can get you uh to set up an appointment to donate blood. Mr. Jenkins: Go ahead, see if you can get me to do

that.

Mike: OK, will you set up an appointment?

Mr. Jenkins: Yes.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 12 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

conditions for requesting, constitutive rules highlight what a speaker assumes by saying something intended

to count as a request. As Searle (1969) puts it: “The preparatory conditions tell us (at least part of) what a

speaker implies in the performance of the act. To put it generally, in the performance of any illocutionary act,

the speaker implies that the preparatory conditions of the act are satisfied” (p. 65). For example, if I request

that my stepson, Brendan, clean out the garage, I imply that the garage needs to be cleaned, that Brendan

wasn't already going to clean the garage, and so forth (see the preparatory rules in Table 6.2). Brendan may

resist my request by arguing that one of the preparatory conditions for requesting, in fact, is not satisfied. If

Brendan exclaims, “It can't be dirty, I just cleaned it last week”, he is saying, in effect, that there is no “need for

action.” As will become apparent below in the section on obstacles to compliance, speech act theory offers

one framework for analyzing how targets resist message sources' requests.

In addition to highlighting obstacles to compliance, speech act theory clarifies how seeking compliance can

threaten identities. By making his request in line 12, Mike presumes that Mr. Jenkins plausibly may be “willing”

to donate blood (see Table 6.2). Mr. Jenkins may fear how he will appear if he says that he is not willing to

do so, as giving blood generally is regarded as helping with a “good cause.” Indeed, Mr. Jenkins emphasizes

that he usually is willing to help and justifies why he has not been in to donate “all winter” (see lines 22–24).

Alternatively, Mike may fear how he himself will appear if he assumes that Mr. Jenkins is willing to donate

blood too rapidly, because he is making a large request of a complete stranger. By specifying what a speak-

er “implies” in requesting, the constitutive rules clarify how either party's identity can be threatened during a

compliance-gaining episode (see Wilson, Aleman, & Leatham, 1998). As noted below, Brown and Levinson's

(1987) politeness theory draws heavily on concepts from speech act theory.

The Conversation Analysis Perspective

Conversation analysis (CA), as a perspective on analyzing situated talk, emerged during the late 1960s and

early 1970s. It is associated most closely with Harvey Sacks, Emanuel Schegloff, Gail Jefferson, and their

colleagues in sociology. In the past two decades, CA has gained adherents in many disciplines (see Psathas,

1995). CA is best characterized as a set of shared commitments or “habitual ways of thinking” rather than as

a set of “theoretical precepts” (Zimmerman, 1988, p. 429). These shared commitments are explored in detail

elsewhere (e.g., Levinson, 1983; Nofsinger, 1991; Psathas, 1995; Zimmerman, 1988). Here I briefly explain

some goals and assumptions of CA and then introduce some of the concepts associated with this perspec-

tive, such as adjacency pairs, conditional relevance, and the preference for agreement.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 13 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Goals and Assumptions of CA

CA scholars search for meaningful regularities in the structure of conversation. They attempt to identify pat-

terns and sequences in the most minute details of everyday talk. CA researchers have documented regular-

ities in the location of overlaps and pauses (Sacks, Schegloff, & Jefferson, 1974), the use of the interjection

“oh” as opposed to “uh” (Heritage, 1984), and the placement of laughter within talk (Jefferson, 1979). We al-

ready have encountered another meaningful sequence, the “prerequest”, in which a speaker inquires about

the status of the other party or the situation prior to making a request (see Chapter 3, Examples 3.3, 3.4, 3.5,

3.6). Aside from identifying meaningful regularities, CA scholars investigate how participants “do” conversa-

tion. By orienting to the details of talk, CA scholars seek to understand how participants interpret what others

are doing and how they negotiate the meanings of their own actions.

CA scholars share several assumptions about studying talk (Levinson, 1983; Psathas, 1995; Zimmerman,

1988). First, they argue that conversations exhibit stable, orderly properties that are the achievements of the

interactants themselves. To identify such regularities, CA scholars typically (a) videotape or audiotape a cor-

pus of conversations; (b) transcribe those conversations in detail, using notations that mark pauses, extended

vowels, rising and falling pitch, and so on; and (c) listen repeatedly to the conversations, searching for in-

terpretable phenomena that occur across multiple conversations (see Hopper, Koch, & Mandelbaum, 1986).

Second, CA scholars assume that conversation is managed locally, meaning that participants negotiate, turn

by turn, whether and how much to talk, what to say and do, and so forth. Third, CA scholars advocate an

inductive, data-to-theory approach. They approach conversation not with the goal of testing preexisting theo-

ries, but with an openness to discovering how participants themselves “do” talk. Because of this, CA scholars

often resist theoretically grounded coding schemes and quantitative methods of interaction analysis that place

talk into preformulated categories. They also seek to explain conversational regularities while resorting only

to details of the talk itself and often resist discussing concepts such as interaction goals, personality traits, or

“master” identities (e.g., gender, cultural, or ethnic identities) that go beyond the talk.

As this description indicates, the assumptions of conversation analysis differ in important respects from those

associated with other theoretical perspectives reviewed in this book. CA scholars hesitate to make infer-

ences about the goals underlying a speaker's actions, whereas most communication scholars take for grant-

ed that persuasive message production is goal oriented (Mandelbaum & Pomerantz, 1991; Tracy, 1991a; see

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 14 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Chapter 5). One can appreciate the aims and insights of CA, however, without taking all CA assumptions as

one's own. Unlike traditional compliance-gaining researchers, who relied heavily on hypothetical scenarios

(see Chapter 4), CA scholars analyze naturally occurring compliance-gaining episodes (e.g., Mandelbaum &

Pomerantz, 1991; Sanders & Fitch, 2001). Several analytic concepts developed by CA scholars have also

turned out to be useful for understanding persuasive message production. It is to these concepts that I now

turn.

Analytic Concepts in CA

Schegloff and Sacks (1973) argue that the adjacency pair is a fundamental unit of conversational organiza-

tion. An adjacency pair is a two-part action sequence, such as question-answer, greeting-greeting, offer-ac-

ceptance/refusal, and request-grant/refusal. Jacobs (1994) provides a succinct summary of the adjacency

pair concept:

Adjacency pairs are two turns long, having two parts said by different speakers in adjacent turns at

talk. By using the first part of an adjacency pair (a “first pair part”, or FPP) a speaker establishes

a “next turn position” that casts the recipient into the role of respondent and structures the rank of

appropriate next moves to those second pair parts (SPPs) that are congruent with the FPP. (p. 213)

By creating an expectation about what types of SPPs will follow an FPP, the concept of adjacency pairs gives

a sense of coherence to the sequencing of conversation. Thus conversation might be thought of as a series

of adjacency pairs: FPP1-SPP1, FPP2-SPP2, FPP3-SPP3, and so forth.

A quick peek at everyday conversation, however, reveals that the two parts of an adjacency pair do not always

occur adjacently. Consider the following two examples:

Both examples contain an insertion sequence, in which one adjacency pair (FPP2-SPP2) is embedded within

another (FPP1-SPP1). But if adjacency pairs do not always occur adjacently, can they still provide a sense of Example 6.9

FPP1 A: May I have a bottle of Mich?

FPP2 B: Are you twenty-one?

Example 6.10

FPP1 A: U:hm (.) what's the price now eh with V.A.T.

[value added tax] do you know eh

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 15 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

coherence to the sequencing of conversation?

According to Schegloff, the glue that binds the two parts of an adjacency pair together is not their adjacent

location, but the concept of conditional relevance: “By conditional relevance of one item on another we mean:

Given the first, the second is expectable; upon its occurrence it can be seen to be a second item to the first;

upon its nonoccurrence it can be seen to be officially absent” (as quoted in Jacobs, 1994, p. 214). That is,

when an SPP does not occur in the turn immediately following the FPP, both parties still anticipate that the

SPP will occur at some future point, and both interpret subsequent utterances as relevant to the eventual per-

formance of the SPP. This is evident in both of the prior examples. In Example 6.9, B asks a question imme-

diately after FPP1 to determine whether the customer, A, is of legal drinking age before addressing A's initial

request for a beer. In Example 6.10, B asks for time immediately after FPP1 in order to work out calculations

needed to answer the customer's initial question. In both cases, the target (B) seeks more information before

deciding whether, or how, to comply with the source's (A) initial request. Insertion sequences are not limited

to two turns; indeed, a target may perform many actions before deciding whether to comply with a source's

request (see Jacobs, 1994, pp. 214–215; Levinson, 1983, p. 305). Even in these cases, however, both parties

expect that the target eventually will address the source's initial request. Given the notion of conditional rele-

vance, the dominant adjacency pair (FPP1- SPP1) can organize extended sequences of talk.

Insertion sequences occur regularly during compliance-gaining episodes. Recall the study described at the

beginning of this chapter, in which college students volunteering on behalf of the American Red Cross tele-

phoned actual blood donors (Wilson et al., 1994). After transcribing these telephone calls, we analyzed what

prior blood donors (targets) said in the conversational turn immediately following the participant's initial re-

quest to donate blood again. Slightly more than one-half (56%) of the prior donors immediately said no,

whereas another 16% immediately responded with an “unconditional yes” to the participant's request. Another

12% of the prior donors immediately responded with a “conditional yes”, indicating that they would be willing

to comply if a stated obstacle (e.g., illness, busy schedule) could be overcome (see Example 6.1). The final

16% of the prior donors did not respond immediately with either yes or no, but instead asked questions about

where and/or when they would donate (i.e., insertion sequences). In some cases, these prior donors used the

participants' answers to their questions as subsequent grounds for compliance-resistance strategies, but in

other cases they subsequently agreed to donate.

Aside from insertion sequences, multiple adjacency pairs can be linked in the form of presequences. A pre-

sequence is an adjacency pair that sets up, and is interpreted in light of, another adjacency pair yet to come

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 16 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

(Beach & Dunning, 1982; Jacobs & Jackson, 1983). Examples include prerequests, preinvitations, and pre-

announcements. During a prerequest, the FPP typically checks on preconditions (i.e., the preparatory rules)

for performing an upcoming request, such as the target's ability to comply. As I have noted in Chapter 3, both

parties implicitly understand the functions of prerequests, and hence several types of sequences may unfold

following their occurrence (see Chapter 3, Examples 3.3, 3.4, 3.5, 3.6).

A final analytic concept for CA is the notion of preference organization. Of the SPPs that can coherently follow

an FPP, there is usually one SPP that is preferred and one that is dispreferred. For example, the preferred

SPP for a request is a grant (i.e., compliance), whereas the dispreferred SPP for a request is a refusal. CA

scholars go to great lengths to emphasize that preference in this context is not intended to describe a psy-

chological concept, and does not refer to the wishes of the message source or target (e.g., Levinson, 1983,

p. 307; Nofsinger, 1991, pp. 71–72; but see Holtgraves, 1992, pp. 148–149). Rather, preference refers to

structural features of the talk, in that dispreferred SPPs typically are linguistically “marked”, whereas preferred

SPPs typically are not. In comparison to preferred SPPs, dispreferred SPPs are more likely to (a) be preced-

ed by pauses, (b) begin with prefaces (e.g., “Uh” or “Well”), (c) begin with token agreements before disagree-

ments (e.g., “Yes, but …”), (d) include qualifiers (e.g., “I don't know for sure, but …”), (e) include apologies

and expressions of appreciation, and (f) include accounts for why the dispreferred SPPs are being performed

(see Levinson, 1983, pp. 334–335).

Research outside of the CA tradition also provides support for the claim that dispreferred SPPs tend to be

linguistically marked. In the blood donation study described at the start of this chapter, for example, we found

that 80% of the prior donors (i.e., targets) who said no also gave reasons why they were refusing to donate

again in the turn adjacent to the participant's request (Wilson et al., 1994). In contrast, very few donors who

immediately said yes gave reasons why they were willing to donate again. Similarly, in a study of responses

to date requests, Folkes (1982) found that 84% of college students who recalled episodes in which they had

turned down dates indicated that they gave reasons why they were saying no, whereas only 24% of students

of who recalled episodes in which they said yes gave reasons why they complied.

The concept of preference organization has important implications for persuasive message production. In

particular, it appears that people are expected to provide reasons, as well as other linguistic markers, when

refusing requests (a dispreferred SPP). This expectation is not idiosyncratic or arbitrary; rather, it reflects an

organizational principle that functions across a wide range of adjacency pairs. As a consequence, people may

feel awkward when faced with admonitions to “just say no”, because doing so violates a pervasive conversa-

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 17 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

tional pattern.

To this point, I have reviewed basic concepts from work on conversational maxims, speech acts, and the CA

perspective. With this background completed, I turn to work on two concepts with special relevance to per-

suasive message production: obstacles and face.

Obstacles to Compliance: Attribution and Speech Act Theories

The CA concept of preference organization suggests that when message targets initially refuse requests, they

often will mention obstacles that hinder them from complying. Obstacles are “the message recipients' beliefs

that cause them to be unwilling or unable to respond immediately in the way the communicator desires” (Clark

& Delia, 1979, p. 200).

Obstacles are central to any compliance-gaining episode. Message sources may anticipate potential obsta-

cles from targets as they formulate their initial requests. Targets often will disclose obstacles as the basis for

their resisting sources' requests. In Example 6.1, Mr. Jenkins states that he has been “a little sick there all

winter there”, which could function as an obstacle to his donating blood now. Which obstacles targets disclose

are constrained, to some degree, by the form and content of sources' initial requests. Targets sometimes dis-

close particular obstacles even though the obstacles named are not the “real” reasons for their resistance.

Both sources and targets know that this can occur, and occasionally the “actual” reason for a target's resis-

tance itself may become a topic of discussion. After encountering obstacles, sources must choose whether

or not to persist in seeking compliance, and, if they decide to persist, they must decide whether or not to ad-

dress the disclosed obstacles directly (as Mike does in Example 6.1, lines 27–46). Speech act and attribution

theorists have explored obstacles. In this section I examine five questions about the nature of obstacles and

their roles at various points in compliance-gaining interactions.

How Have Obstacles Been Described?

Speech act (Austin, 1962; Searle, 1969) and attribution (Heider, 1958; Weiner, 1986) theories offer different,

albeit complementary, descriptions of the obstacle construct. For speech act theorists, the constitutive rules

for directives offer a framework for analyzing potential obstacles to compliance. Directives, such as requests,

recommendations, and commands, are speech acts designed to get the hearer to perform an action that he

or she otherwise would not have performed (Searle, 1976). Directives are at the heart of any attempt to gain

compliance, given that people seek compliance only when they presume that their targets otherwise would

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 18 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

not have performed the desired actions (see the definition of compliance gaining in Chapter 1). It is true that

compliance-seeking messages frequently contain speech acts in addition to directives, such as assertions

or promises. Often, however, these additional acts support the logical preconditions for using directives; for

example, these acts might explain why the desired action needs to be performed, why the message source

cannot perform the action him- or herself, or why the target should be willing to comply (Tracy, Craig, Smith,

& Spisak, 1984; see Table 6.2). It also is true that when message sources hint, they do not state directives

explicitly. Even when hinting, however, message sources still adhere to the rules for directives by encouraging

targets to perform behaviors that the targets otherwise did not plan to perform.

Because directives are at the heart of attempts to gain compliance, message sources adhere to the rules for

directives whenever they seek compliance. Put differently, a message source, by seeking compliance, explic-

itly or implicitly asserts that the rules for directives are “in effect” and provide an accurate characterization of

the current reality. As an example, when I ask my 9-year-old stepdaughter, Annie, to make her bed, I assert

that the preconditions for making this request are met (see Table 6.2). By making my request, in other words,

I assume the following: that Annie's bed currently is not made, that it needs to be made, that Annie was not

about to make the bed, that Annie has the ability to make the bed at this time, that Annie is willing and/or

obligated to make her bed, that I have the right to remind Annie to make her bed, and that I sincerely want

the bed made.

As is apparent from this example, the message target does not always passively accept the source's “defini-

tion of the situation” (Goffman, 1959). From the perspective of speech act theory, resisting compliance means

challenging the message source's definition of the situation—that is, challenging the source's assertion that

the rules for directives provide an accurate characterization of the current situation. Resisting compliance

means asserting that the source's request is “defective.” In the “make the bed” situation, Annie might object

to my request because (a) she thinks that the task of making her bed is unnecessary (there is no need for

the action), (b) she already had planned to make the bed sometime later (there is no need for my request),

(c) she is too busy to make the bed right now (she lacks ability), (d) she does not want to make her bed (she

lacks willingness), (e) she does not have to make her bed (she is not obligated), (f) she can keep her own

room any way that she wants (I have no right to make this request), or (g) she believes that I don't really care

about the bed, but just want to hassle her (I lack sincerity).

During compliance-gaining episodes, the constitutive rules for directives function as stases, or possible points

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 19 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

of clash in arguments about how the current situation should be defined (Jackson & Jacobs, 1980; Kline,

1979; Wilson, Kim, & Meischke, 1991/1992). As a consequence, the rules for directives should offer a com-

prehensive framework for analyzing obstacles to compliance, because resisting compliance means asserting

that the conditions specified in one of these rules is not met in the current situation. Each of Annie's refusals

in the prior example functions in this way. Indeed, a variety of specific obstacles can be analyzed using a

speech act theory framework (e.g., Gibbs, 1986; Ifert & Roloff, 1994).

Attribution theory offers a second framework for describing obstacles. Attributions are personal judgments

about the causes for actions or events; thus attribution theorists study how everyday actors answer “why”

questions (Heider, 1958). You may make attributions about a wide variety of events, such as why you did

well on a test, why a stranger bumped into you at a concert, why your romantic partner got upset during a

conversation, why you received (or did not receive) a raise at work, or why a close friend or family member

died at a young age. Attributions play an important role in how you interpret the meanings of events. Imagine

that a stranger bumps into you as you are waiting in line to enter a movie theater. Does what is “going on”

seem different if the stranger bumps into you accidentally because the theater entrance is overcrowded as

opposed to if the stranger bumps into you because he or she is impatient about waiting in line? Attributions

also influence your emotional reactions to events. You are likely to feel happy about having received an A on

an exam regardless of whether you studied hard or whether the exam was so easy that everyone did well.

Only in the former case, however, are you likely to take “pride” in how you performed on the exam (Weiner,

1986).

During a compliance-gaining interaction, a salient question can be why the target is refusing to comply imme-

diately with the message source's request. From an attributional perspective, obstacles are causes or reasons

for the target's noncompliance. Attributions are personal judgments, and thus message sources and targets

may make different attributions about why the target is refusing to comply (Jones & Nisbett, 1972; Wilson &

Kang, 1991; Wilson, Levine, Cruz, & Rao, 1997).

When we try to think about the possible causes for any specific event, the number of candidate possibilities

may seem overwhelming. Consider why a student, Joan, performed poorly in delivering her first speech in

a public speaking class. Perhaps Joan did not practice her speech in advance. Perhaps she picked a topic

with a great deal of technical information that was difficult to explain. Perhaps Joan lacks skill at organizing

her ideas clearly. Or perhaps she is very nervous about speaking in public. Perhaps Joan felt ill on the day

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 20 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

of her speech, or was distracted because she had gotten into a fight with her boyfriend the evening before.

Perhaps Joan's instructor has very high expectations. Perhaps Joan was unlucky, because the best student

in the class happened to be selected to speak just before she gave her own speech. Given all the possible

reasons for Joan's performance, how do we organize and make sense of different attributions?

In his attributional theory of motivation and emotion, Bernard Weiner (1986) argues that people distinguish

causes along three basic dimensions. The first dimension, labeled locus, is concerned with whether the cause

resides within the actor (internal) or outside of actor, in the external environment (external). Joan's anxi-

ety about public speaking and her decision not to practice in advance are examples of internal causes; her

teacher's high expectations and the fight with her boyfriend are examples of external causes. The second

dimension, labeled stability, concerns whether the cause fluctuates over time (unstable) or remains constant

over time (stable). Joan's anxiety about public speaking and her teacher's high expectations are relatively sta-

ble causes; her fight with her boyfriend and her being unlucky at having the best student randomly selected to

speak just before her are unstable causes (because they do not happen each time she gives a speech).4 The

third dimension, controllability, concerns whether the cause can be influenced by the actor and/or other par-

ties (controllable) or whether it is beyond anyone's control (uncontrollable). Joan's decision not to practice and

her instructor's high expectations are examples of controllable causes; her feeling ill and her being unlucky

at having the best student randomly selected to speak just before her are uncontrollable causes. According

to Weiner, locus, stability, and controllability are distinguishable on logical grounds; for example, people may

succeed at a task due to internal causes that are either stable (e.g., ability) or unstable (e.g., effort). Weiner

also reviews evidence that everyday actors distinguish perceived causes for interpersonal events, such as

marital conflict, along these three dimensions.

By crossing the attributional dimensions of locus, stability, and controllability, we can classify eight different

types of obstacles to compliance in a variety of situations. Table 6.3 lists examples of eight different types

of reasons individuals may give for saying no to a date request. These data are from a study conducted by

Folkes (1982), who asked each of 64 female and 64 male undergraduates to describe a recent episode in

which she or he had turned down, and/or been turned down for, a date request made by someone the par-

ticipant did not know well. Table 6.4 lists eight different types of reasons individuals might give for refusing to

participate, after promising to do so, in a research study. My colleagues and I generated these data by asking

undergraduates to list reasons a student might sign up for an out-of-class experiment but then later back out

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 21 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

on participating (see Wilson, Cruz, Marshall, & Rao, 1993). As you will see below, message targets are not

equally likely to disclose these eight types of obstacles, and message sources make choices about whether

and how to persist based on the locus, stability, and controllability of the disclosed obstacles.

Do Message Sources Anticipate Obstacles in Their Initial Requests?

Imagine that you want to know the time and do not happen to be wearing a watch. You approach a stranger,

say, “Excuse me”, and then ask for the time. There are several ways you might phrase your request, such as

“Do you happen to know what time it is?” “Can you tell me the time?” or “Would you mind telling me the time?”

Which of these requests would you be most likely to say in this situation? Would your choice of phrasing be

completely random?

SOURCE: From “Communicating the Reasons for Social Rejection”, by V. S. Folkes, 1982, Journal of Exper-

imental Social Psychology, 18, p. 242. Copyright 1982 by Academic Press. Reprinted with permission.

NOTE: Personal causes referred to the rejected person (i.e., internal locus), whereas impersonal qualities re-

Table 6.3 Eight Reasons for Turning Down a Date Request, Varying Along Three Attributional Dimensions

Attributional Dimensions Target's Reason for Saying No

1. Impersonal, unstable, uncontrollable The rejecter had to study for finals.

2. Impersonal, unstable, controllable The rejecter would rather go to a dance that night than go to the movies.

3. Impersonal, stable, uncontrollable The rejecter was seriously involved with someone.

4. Impersonal, stable, controllable The rejecter did not want to mess up a relationship with someone else she was dating.

5. Personal, unstable, uncontrollable The rejected person was in a bad mood.

6. Personal, unstable, controllable The rejected person had a lot of nerve calling up the rejecter the same night to ask her out.

7. Personal, stable, uncontrollable The rejected person was too old for the rejecter.

8. Personal, stable, controllable The rejecter did not agree with the rejected person's religious beliefs.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 22 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

lated to the rejecter or the larger situation (i.e., external locus).

According to Ellen Francik and Herbert Clark (1985), speakers word requests in a predictable fashion, based

on how they perceive obstacles in the situation. Francik and Clark propose an “obstacle hypothesis”, which

states that “speakers design requests to overcome the greatest potential obstacle they see to getting the

information they want” (p. 560). To illustrate this hypothesis: Imagine that you want to buy a popular CD that

has just been released. You drive to the store, enter, and approach a clerk. Are you likely to say, “Would you

be willing to sell me the new ____ CD?” Probably not. The clerk's willingness is not at issue, given that the

store's purpose is to make a profit by selling CDs. Availability is a more plausible obstacle that could prevent

you from obtaining the new CD; hence you might start with a prerequest such as “Do you have the new ____

CD?” According to Francik and Clark, a speaker typically mentions what he or she perceives to be the most

likely obstacle because this (a) helps the target realize what he or she needs to do in order to comply (e.g.,

check whether the new CD has arrived) and (b) provides the target with a plausible excuse if he or she is

unable or unwilling to comply (e.g., “Sorry, we haven't received it yet”).

Table 6.4 Eight Reasons for Refusing to Participate in a Research Study as Promised, Varying Along Three

Attributional Dimensions

Attributional Dimensions Target's Reason for Saying No

1. Internal, stable, control-

lable

I always seem to volunteer whenever anyone needs help. I agreed to be a big brother/ sister and will be spending

that time with my little brother/sister.

2. Internal, stable, uncon-

trollable I just can't do it. No matter what I do, every time I participate in one of these experiments I make a fool of myself.

3. Internal, unstable, con-

trollable I invited friends out to happy hour at that time, so I'm not going to be at the experiment.

4. Internal, unstable, un-

controllable My boyfriend/girlfriend and I just had a huge fight. I'm in a bad mood and don't feel up to coming to the experiment.

5. External, stable, con-

trollable

It's important to me to graduate in three years, so I'm taking 19 credits every term. I just don't have the time to do

any extra credit.

6. External, stable, uncon- I won't be there. I'm in the Marketing Club, and we meet at that time every week.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 23 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

SOURCE: From “An Attributional Analysis of Compliance-Gaining Interactions”, by S. R. Wilson, M. G. Cruz,

L. J. Marshall, & N. Rao, 1993, Communication Monographs, 60, pp. 370–371. Copyright 1993 by the Speech

Communication Association. Adapted with permission.

The obstacle hypothesis has received support in several studies (Francik & Clark, 1985; Gibbs, 1986; Gibbs

& Mueller, 1988; Roloff & Janiszewski, 1989). For example, Francik and Clark (1985, Experiments 1 and 2)

had undergraduates listen to short tape-recorded descriptions of 24 hypothetical scenarios that contained one

of three types of obstacles: (a) ability (e.g., asking someone who is not visibly wearing a watch if he or she

knows the time), (b) willingness (e.g., asking someone about sensitive family information that the person may

not want to reveal), or (c) memory (e.g., asking someone a question when you cannot recall whether you

already have asked that person the same question at a prior time). Other versions of the scenarios did not

appear to contain any obstacles (e.g., asking someone who is visibly wearing a watch if he or she knows the

time). After hearing each scenario, participants stated aloud what they would say in the situation.

Francik and Clark (1985) report two main findings from this research. First, participants used more indirect

requests (e.g., “Do you have any idea what time it is?”) than direct requests (e.g., “What time is it?”) in scenar-

ios where obstacles were perceived as present rather than absent. Second, participants varied the forms of

their indirect requests depending on the likely obstacles. Participants used the form “Do you know” when they

were uncertain whether the target knew some desired information, “Did you see/hear/notice” when they were

uncertain whether the target noticed and remembered the desired information, “Would you” or “Could you”

when it was unclear whether the target was willing to disclose the desired information, and “Have I asked you”

when they were uncertain whether they had asked about the same information at an earlier time. Consistent

with the obstacle hypothesis, the participants also comprehended requests for information more quickly, and

evaluated them as more appropriate, when they addressed plausible rather than implausible obstacles (Fran-

cik & Clark, 1985; Gibbs, 1986).

trollable

7. External, unstable, con-

trollable I just decided not to do it because the professor offered so little extra credit that it just isn't worth my time.

8. External, unstable, un-

controllable

I can't do it then because I have to meet with one of my professors during that time. It's the only time he can meet

this week.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 24 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Francik and Clark (1985) examined only requests for information, but Michael Roloff and Chris Janiszewski

(1989) have shown how the obstacle hypothesis can apply to requests for assistance as well. They distinguish

two types of assistance. A request to “borrow” occurs when the message source asks the target to lend him or

her a material object temporarily, promising to return that object in the same condition (e.g., you miss a class

and ask to borrow another student's notes). For requests to borrow, likely obstacles include that the target

may want to use the object him- or herself, may fear that the object will not be returned in a timely fashion,

and/or may fear that the message source will damage the object. A request for a “favor” occurs when the

source asks the target to perform some service that primarily will help the source and that will consume the

target's time and/or energy (e.g., you need a ride to a location 25 minutes away and ask someone you know

to drive you). For requests for favors, likely obstacles include that the target may lack time and/or energy to

provide the service, may lack willingness to help, and/or may prefer to do other things with his or her time.

According to Roloff and Janiszewski (1989), the obstacle hypothesis implies that “speakers will adapt their

requests to overcome potential sources of resistance to providing each type of assistance” (p. 36). In a test of

this reasoning, 120 undergraduate students wrote exactly what they would say in response to a hypothetical

“request to borrow” scenario (i.e., asking to borrow either the target's class notes or $25) and a “request for

a favor” scenario (i.e., asking the target either to type a paper or to provide a ride). Some students imagined

that the target in both situations was another student in their class whom they considered a friend, whereas

others imagined the target was an acquaintance or a stranger in their class. For each message, coders ini-

tially analyzed whether it contained one or more grammatical clauses. Clauses then were classified into one

of the seven categories shown in Table 6.5. Roloff and Janiszewski developed these categories based on

Brown and Levinson's (1987) types of positive and negative politeness (see below). Several categories also

pertain to the constitutive rules for requests (e.g., inquiries, explanations).

Table 6.5 Roloff and Janiszewski's (1989) Seven Types of Clauses Within Request Messages

Type of Clauses Definition/Example

Contractual clauses Those clauses containing information about the amount, duration, or contingencies of the requested action

Example: I will return your notes next class.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 25 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

SOURCE: From “Overcoming Obstacles to Interpersonal Compliance: A Principle of Message Construction”,

by M. E. Roloff & C. A. Janiszewski, 1989, Human Communication Research, 16, 1989, p. 45. Copyright 1989

by the International Communication Association. Adapted with permission.

Consistent with the obstacle hypothesis, participants in Roloff and Janiszewski's (1989) study varied whether

they included specific types of clauses in their request messages depending on the type of assistance being

sought. For example, contractual clauses occurred more frequently in “borrow” than in “favor” messages,

whereas inquiries about ability and inducements occurred more often in “favor” than in “borrow” messages.

These differences make sense in light of the obstacles relevant to these two types of requests: Contractual

clauses assure the target that the borrowed object will be cared for and returned in a timely fashion, whereas

inquiries and inducements address the target's ability and willingness to provide a favor.

Inquiries Those clauses in which the requester asks about the ability of the target to provide the needed assistance

Example: Are you busy this afternoon?

Inducements Those clauses in which the requester offers a compliment, exchange, or gratitude to the target

Example: I will pay you for the gas.

Explanations Those clauses in which the requester indicates why he or she needs the requested action

Example: I have been sick.

Apologies Those clauses in which the requester acknowledges the inappropriateness of the requested assistance

Example: I know this is an imposition.

Introductions Those clauses in which the requester introduces or identifies him- or herself

Example: I live in your dorm.

Requests Those clauses in which the requester explicitly asks for the assistance

Example: Can you give me a ride downtown?

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 26 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

To summarize, message sources do anticipate potential obstacles from targets as they design their compli-

ance-seeking messages. Research on the obstacle hypothesis illustrates two important points about persua-

sive message production. First, these studies show that people often pursue interaction goals with little con-

scious awareness. When asking someone for the time, do you think carefully about whether to say, “Do you

happen to know what time it is?” or “Can you tell me the time?” or “Would you mind telling me the time?” De-

spite the fact that we seem to make such requests in a mindless fashion, we alter the forms of our requests

to address the most plausible obstacles that might prevent us from attaining our goals. In other words, per-

suasive message production can be purposive without being highly mindful (Kellermann, 1992).

A second point is that the obstacle hypothesis has received support precisely because persuasive message

production is goal oriented. Message sources anticipate obstacles to compliance in their requests because

doing so helps them accomplish their primary and secondary goals (Francik & Clark, 1985). In some cases,

however, sources may be more likely to accomplish their goals by ignoring obstacles. As Gibbs (1986) ex-

plains:

In other contexts, speakers … will design their requests without any consideration of the potential

obstacles for addressees. A general making a request of a private is just such a case. Failure to

mention an appropriate obstacle helps the general maintain his superior social status relative to the

private. It is clear that speakers may satisfy many communicative goals by either highlighting or ig-

noring certain obstacles when formulating indirect speech acts. (p. 194)

In other words, we should not always expect speakers to address obstacles in their compliance-seeking mes-

sages. Future research might explore boundary conditions on the obstacle hypothesis.

Do Targets Always Disclose Their “Real” Reasons for Resisting Compliance?

Imagine that a person from your workplace asks you for a date (if you are in a long-term, exclusive romantic

relationship, then remember a time when you dated other people). The person who has asked you out is

someone you do not find physically or socially attractive. How would you go about refusing the person's date

request? Would you tell the person that you simply do not feel attracted to him or her? Or would you search

for another excuse, such as that you already are dating someone else. Would you use this latter excuse even

if you and your current dating partner both have agreed to see other people?

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 27 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Bernard Weiner, Valerie Folkes, and their colleagues have compared public and private reasons for refusing

requests from an attributional perspective (Folkes, 1982; Weiner, Amirkhan, Folkes, & Verette, 1987; Weiner,

Figueroa- Munoz, & Kakihara, 1991). In the first investigation, Folkes (1982, Study 1) asked each of 64 female

and 64 male undergraduates to recall a recent episode in which she or he had been rejected for a date and/

or had been the rejecter of a date. Each participant was asked whether the rejecter had explained why he or

she was saying no and, if so, what the rejecter's public explanation had been. As Folkes notes, this publicly

stated reason may have been the real reason, may have been different from the real reason, or may have

been only part of the reason the date rejecter said no. Each participant then was asked to list five “private”

reasons the rejecter in his or her episode had turned down the date. Coders categorized both the public and

the private reasons as (a) personal (related to the rejected person) or impersonal (related to the rejecter or

situation), (b) stable or unstable, and (c) controllable or uncontrollable (see Table 6.3 for examples). Nearly

two-thirds (64%) of the public reasons for refusing a date were impersonal, uncontrollable, and unstable (e.g.,

the rejecter had to study for final exams), whereas less than one-third (30%) of the private reasons were of

this type.

Message targets at times withhold or alter their “real reasons” in a variety of situations aside from date re-

quests. Weiner et al. (1991, Study 1) asked each of 97 undergraduate students to describe an occasion when

he or she “gave an excuse to someone that was not the truth. That is, there was a real reason why something

happened or failed to occur, or why you responded positively or negatively, and you withheld that reason and

gave a different one” (p. 6). Each participant was asked to describe the situation, including what he or she

communicated to the other person and what the real reason was that the participant withheld. Each partici-

pant also described his or her goal(s), or what he or she wanted to accomplish by giving the false excuse or

reason.

Several findings from Weiner et al.'s (1991) study are of interest. First, participants had no trouble recalling

recent incidents in which they had communicated false excuses. A second finding is that disclosed (false) and

withheld (true) reasons differed along all three attributional dimensions. As in Folkes's (1982) earlier study,

the vast majority of public reasons were external, uncontrollable, and unstable. In contrast, more than 90% of

the privately withheld reasons were internal and controllable. A third finding is that participants reported that

they withheld their “real” reasons to accomplish four goals; specifically, they wanted to avoid (a) complying

with the other party's request, (b) hurting the other party's feelings, (c) making the other party angry, and (d)

making themselves look bad. In Dillard's (1990b) terminology, targets may give false reasons for resisting

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 28 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

compliance in order to accomplish both primary and secondary goals.

Of course, message sources realize that, at times, targets' publicly stated reasons for resisting compliance

are different from their real reasons. In a study exploring this issue, Danette Ifert and Michael Roloff (1994)

had undergraduate students imagine that they were going to ask a target person to loan them notes from a

class as well as to type an eight-page paper for them. Students imagined that the target in both situations was

either a friend or a stranger in their class. Participants initially listed all of the obstacles that would prevent

the target from complying with their request or “private” reasons why the target might say no. Participants

then wrote out exactly what they expected the target would say to deny their request (i.e., the target's public

reasons). Coders placed both private and public reasons into one of seven categories, which were further

grouped into those obstacles that primarily reflected the target's inability to comply (i.e., uncontrollable caus-

es) versus those that reflected the target's unwillingness to comply (i.e., controllable causes). Coders also

analyzed whether participants expected the target to include each private reason in his or her refusal mes-

sage.

Participants from Ifert and Roloff's (1994) study expected that a target would be much more likely to disclose

an “inability” obstacle than an “unwillingness” obstacle. Participants, on average, believed that message tar-

gets would mention 42% of the inability obstacles that they perceived as present in the situation, but would

mention only 12% of the unwillingness obstacles. Consistent with what targets actually tend to do, message

sources expected that targets would disclose uncontrollable reasons for saying no while withholding control-

lable reasons.

What do message sources do when they are skeptical about whether targets' public reasons for resisting

compliance are the “real” reasons for the targets' resistance? In some cases, a message source simply may

accept the target's public reason for saying no. If a target refuses a date request by saying that he or she “has

to study for finals”, the message source may accept this public explanation because pressing for the “real”

reason could embarrass both parties. In other cases, however, a message source may risk pressing for the

“real” reason, especially when the request is perceived to be in the target's best interest (Boster, 1995). Fu-

ture research might examine when message sources and targets are likely to debate the legitimacy of public

obstacles to compliance.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 29 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Does the Nature of an Initial Request Constrain How the Target Resists Com-

pliance?

Imagine once again that a person from your workplace asks you for a date, but you are not physically or so-

cially attracted to the person. If the person asks, “Would you like to go see a movie Saturday night?” you might

respond, “I'm sorry, but I'm busy Saturday.” If the person instead asks, “Would you like to go see a movie

sometime?” would you still respond, “I'm sorry, but I'm busy Saturday”? Probably not, because the person has

not mentioned a specific day. Being busy on Saturday is a “relevant” obstacle to the first, but not the second,

date request. To some extent, message sources can constrain which obstacles to compliance are, and are

not, relevant by the phrasing of their initial requests.

According to Gaylen Paulson and Michael Roloff (1997), the form and the content of a date request both can

influence the nature of a target's refusal. Table 6.6 presents six date requests that vary in form and content.

Regarding request form, Paulson and Roloff argue that when a person asks a virtual stranger for a date, the

source typically perceives the target's willingness to be the most plausible obstacle to compliance. In light

of the obstacle hypothesis, discussed above, this means that date requests with strangers that address the

target's “willingness” to comply should be perceived as more conventional, and as more appropriate or po-

lite, than requests that address other obstacles, such as the target's “ability” to comply. In turn, both of these

request forms should be perceived as more polite than requests that the source presumes will raise no obsta-

cles, because such “obstacle-free” forms place greater constraint on the target's autonomy (see Table 6.6).

Given this reasoning, Paulson and Roloff predict that date requests with strangers that address willingness

obstacles will prompt more polite refusals than will requests that address ability obstacles, and that obstacle-

free forms will prompt the least polite refusals.

Regarding request content, Paulson and Roloff (1997) argue that date requests can be specific or ambiguous

about the proposed joint activity, participants, setting, and/or time. In particular, they predict that requests that

are ambiguous with regard to time should prompt more stable refusals (“Sorry, but I'm seriously involved with

someone”) than requests that propose specific times, because unstable obstacles such as being “busy Sat-

urday” are not relevant for refusing the former type of date request (see Table 6.6).

Table 6.6 Paulson and Roloff's (1997) Six Examples of Date Requests

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 30 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

To test these hypotheses, Paulson and Roloff asked each of 65 male and 63 female undergraduates to imag-

ine that another student in the participant's communication class, whom the participant did not know, asked

the participant for a date using one of the six requests shown in Table 6.9. After reading the request, each

participant was asked to imagine that for some reason he or she wanted to turn down the date. Paulson and

Roloff had each participant write out exactly what he or she would say to the other person to refuse the date

request and then rate both the request and his or her own refusal message in terms of perceived convention-

ality and politeness. Coders also analyzed whether the obstacles to compliance in participants' refusal mes-

sages were stable or unstable. The results largely supported the researchers' predictions, in that participants

(a) perceived “willingness”-based date requests as more polite than “ability”-based requests, which they in

turn saw as more conventional and polite than “obstacle-free” requests; (b) perceived their own refusals to

be most polite when the date requester used a “willingness” form and least polite when the date requester

posed no obstacles; and (c) included more stable obstacles when the date requester was ambiguous rather

than specific about the proposed time for the date.

Paulson and Roloff's (1997) findings suggest that message sources should consider how they phrase im-

portant requests, because “rather small changes in wording [can] reliably influence the content of refusals”

(p. 284). Asking a virtual stranger for a date without addressing potential obstacles, for example, typically is

perceived as overbearing and is likely to prompt a less polite refusal than other date request forms. Accord-

ing to Paulson and Roloff, message sources may benefit from being either ambiguous or specific about the

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 31 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

proposed content of their date requests, depending on the relative importance of their own primary and sec-

ondary goals.

Do Message Sources Vary in Whether, and how, They Persist in Seeking Com-

pliance Depending on the Types of Obstacles Disclosed by Targets?

As the examples presented throughout this text show, message targets often respond to compliance-seeking

messages by asking questions or posing obstacles. Despite this, research on compliance gaining typically

has examined only people's initial responses to hypothetical scenarios (see Chapter 4). Among the smaller

subset of studies that have analyzed subsequent compliance-seeking strategies, many have instructed par-

ticipants simply to imagine that the target said no to their initial request and then asked what participants

would say or do at that point (e.g., Cody, Canary, & Smith, 1994; deTurck, 1985, 1987; Hample & Dallinger,

1998; Wilson, Whipple, & Grau, 1996). In other words, many studies have provided no information about why

the target was saying no. To my knowledge, only two studies have investigated whether message sources

vary in terms of whether, and how, they persist in seeking compliance when targets disclose different types of

obstacles.

In the first study, several colleagues and I applied Weiner's (1986) attributional theory to predict whether and

how message sources would persist after hearing targets disclose different reasons for their failing to honor

an earlier commitment (Wilson et al., 1993). I have described the procedures for this study in detail at the

beginning of Chapter 2. As you may recall, each participant was given a list of 20 students (actually confed-

erates) who supposedly had signed up to earn extra credit by completing a study being conducted by the

fictitious “Professor Jones.” Participants in our study were instructed to telephone these 20 students and re-

mind them about the time and place of their own study. If any of the 20 students had changed their minds

about coming, participants were supposed to try to persuade them to honor their promise to complete their

study. We told our participants to use their own judgment about what to say and how long to persist in trying

to persuade Professor Jones's students to show up.

In reality, the participants telephoned 20 confederates who were working with us. Of the 20 phone calls, 10

were answered by confederates who were instructed simply to say yes, they would show up for Professor

Jones's study as promised. One phone number always went to a wrong number, and another always went to

someone we knew was not there to receive the call. The remaining 8 calls were answered by confederates

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 32 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

who were trained to say no, they would not show up. When prompted, these 8 confederates all disclosed

different obstacles, or reasons they had changed their minds about participating in Professor Jones's study.

Developed in a pilot study, these obstacles varied in terms of Weiner's (1986) three causal dimensions of lo-

cus, stability, and controllability. Examples of all of the types of obstacles appear in Table 6.3.

Based on Weiner's (1986) theory, we reasoned that message sources would vary the degree to which they

persisted, as well as their specific choices of compliance-seeking strategies, depending on the reasons the

students (confederates) gave for refusing to participate in the research project as promised. According to

Weiner, individuals set higher expectations and hence are more motivated to persist when they perceive that

their initial failure to accomplish a goal is due to a cause that might change over time (unstable) rather than a

cause that will remain constant (stable). In addition to stability, attributional theorists suggest that individuals

set higher expectations and persist longer when they believe that their initial failure is due to something that

can be controlled rather than something about which they are helpless (e.g., Anderson & Jennings, 1980;

Dweck, 1999). Following this reasoning, we predicted that participants would persist longer with targets (con-

federates) who disclosed unstable, internal, and controllable reasons for refusing to participate in the study as

promised than with targets who disclosed stable, external, and controllable reasons for saying no. Depending

on the locus, stability, and controllability of the disclosed obstacle, we also predicted that participants would

vary in terms of whether they (a) perceived the target's reason for saying no as sincere, (b) addressed that

obstacle directly, and (c) used antisocial compliance-seeking strategies.

Results from the study were consistent with, albeit more complicated than, these predictions. First, partici-

pants perceived most internal causes as controllable and external causes as uncontrollable; hence we col-

lapsed locus and controllability into a single attributional dimension. Second, the effects of locus/controllability

depended on stability. In most cases, predictions regarding locus/controllability were confirmed for unstable

but not for stable causes. When participants encountered targets (confederates) who disclosed unstable rea-

sons for refusing to honor their commitment, the participants persisted longer, denied the validity of the obsta-

cles more often, used guilt more frequently, and perceived the targets as more sincere if their reasons were

“internal/ controllable” rather than “external/uncontrollable.” When participants encountered targets who dis-

closed stable reasons for refusing to honor their commitment, however, the locus/controllability of the reasons

did not matter. Put differently, participants appeared to evaluate obstacles using a two-step process. Initially,

they attended to stability, and they gave up seeking compliance quickly with targets who disclosed obstacles

that were unlikely to change over time. In cases where the obstacles were unstable, the participants then

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 33 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

attended to locus/controllability, choosing to persist longer at seeking compliance with targets who disclosed

obstacles that were within rather than beyond the targets' control.

In the second study, Ifert and Roloff (1996) examined the joint effects of types of expressed obstacles and

relational intimacy to predict whether, and how, message sources would persist in the face of resistance from

targets. Participants in their study, 257 undergraduate students, wrote out exactly how they would make two

requests (asking to borrow the target's class notes for two days and asking the target to type an eight-page

paper on behalf of the source) of either a stranger or a friend. After writing the initial request, each participant

encountered one of three types of obstacles supposedly expressed by the target: (a) unwillingness obstacles

(e.g., “There's no way I would do that”), (b) inability obstacles (e.g., “I don't have notes for that day”), or (c)

imposition obstacles (e.g., “I don't have time to do it”). Unwillingness and inability obstacles embody two of

the preparatory rules for requests (see Table 6.2), whereas imposition obstacles express a mixture of inability

and unwillingness. After reading an obstacle statement, each participant completed a measure of how much

he or she would want to persist in seeking compliance, after which the participant wrote out exactly what he

or she would say after encountering that obstacle. These subsequent messages were coded for the presence

of (a) inquiries about why the target was refusing the request; (b) persuasion cues, or phrases that explained

or justified the initial request; and (c) forgiving cues, or phrases that accepted the target's refusal gracefully.

Ifert and Roloff (1996) predicted that participants in general would show greater desire to persist after en-

countering resistance from intimate targets (i.e., friends) as opposed to nonintimate targets (i.e., strangers),

because friends often believe that they should be able to count on each other in times of need (see Roloff

et al., 1988). They also predicted that participants' desire to persist, as well as their use of specific forms of

persistence, would depend on the types of obstacles disclosed by the targets. Specifically, participants were

predicted to show especially high desire to persist, as well as to use more inquiries and persuasion cues

and fewer forgiving cues, when intimate targets disclosed unwillingness obstacles (which are controllable) as

opposed to inability obstacles (which are uncontrollable). The researchers reasoned that unwillingness obsta-

cles are perceived as particularly unexpected and inappropriate within intimate relationships, and hence pre-

dicted that participants would feel most justified in confronting these obstacles when the target was a friend.

The results were largely consistent with these predictions: Intimacy was a stronger predictor of participants'

desire to persist, as well as their use of the three specific forms of persistence, when targets disclosed unwill-

ingness as opposed to inability obstacles. Responses to imposition obstacles, in most cases, were similar to

responses to unwillingness rather than inability obstacles. In sum, Ifert and Roloff's (1996) as well as Wilson

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 34 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

et al.'s (1993) research shows that message sources vary their likelihood and means of persisting depending,

in part, on the nature of the obstacles disclosed by the targets.

In this section I have addressed five questions about the role of obstacles during compliance-gaining

episodes. Both speech act and attribution theories have informed research on obstacles. One research find-

ing I have described above is that message targets sometimes withhold their “real” reasons for saying no

in order to spare the sources, as well as themselves, awkwardness and embarrassment. This suggests that

people design their influence messages, in part, in light of both parties' public identities, or “face.” Erving Goff-

man (1967), the well-known symbolic interactionist, defines face as the “social value that a person effectively

claims for himself by the line others assume he has taken during a particular contact. Face is an image of self

delineated in terms of approved social attributes” (p. 5).

Requests and refusals project ongoing definitions of situations, in the sense that they imply that certain states

of affairs exist, or should exist (Wilson et al., 1991/1992, 1998). Thus seeking and resisting compliance can

raise threats to face for both parties. Perhaps not surprisingly, then, people commonly report that the desire

to maintain face is one of their goals during compliance-gaining episodes (Dillard, Segrin, & Harden, 1989;

Hample & Dallinger, 1987a; Kim et al., 1996; see Chapter 5). Face concerns shape the form of request and

refusal messages, and individuals differ systematically in their views about possible means for balancing com-

peting desires to seek compliance and sustain face. Face can be explored through the lenses of politeness

and message design logic theories.

Threats to Face: Politeness and Message Design Logic Theories

Brown and Levinson's Politeness Theory

Basic Assumptions of Politeness Theory

Penelope Brown, an anthropologist, and Stephen C. Levinson, a linguist, developed politeness theory based

on their extensive fieldwork in three diverse languages and cultures: English as spoken in Great Britain; Tamil,

a dialect spoken in southern India; and Tzetal, the language of the Mayan Indians in Mexico.5 They present

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 35 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

politeness theory to account for “the extraordinary parallelism in the linguistic minutia of the utterances with

which persons choose to express themselves in quite unrelated languages and cultures” (Brown & Levinson,

1987, p. 55). One example of this parallelism is that speakers from all three cultures often use indirect speech

acts when seeking another party's compliance. Brown and Levinson show that indirect requests such as “Can

you shut the door?” “Are you planning on shutting the door?” and “Did you shut the door?” occur regularly in

English, Tamil, and Tzetal (p. 136).

To account for linguistic parallels across cultures, politeness theory posits that speakers in all cultures want

to maintain “face.” Building on Goffman (1967), Brown and Levinson divide face into two basic wants, labeled

negative and positive face. Negative face is the desire to maintain one's own autonomy. Individuals in any

culture want to be shown respect and not have their privacy and space invaded, their resources spent, and

their actions restricted without cause. Positive face is the desire to have one's own attributes and actions ap-

proved of by significant others. Which attributes are desired and which persons are significant depend on the

roles being enacted. College professors typically want students to perceive them as competent, organized,

and interesting. Most parents hope that their children will view them as loving, supportive, and fair. Although

desired attributes vary across roles, and cultural definitions of roles, individuals in all cultures want approval

and ratification from the people they love, admire, and/or value.

Aside from maintaining one's own face, Brown and Levinson (1987) argue, relational interdependence cre-

ates motives for supporting other people's face. Face is a social commodity. Although we want to maintain

face, it can be granted only by others (Goffman, 1967). A conversationalist who threatens another party's face

risks retaliatory attacks on his or her own. Politeness theory thus assumes that “it is generally in everyone's

best interest to maintain each other's face” (Brown & Levinson, 1987, p. 61).

Despite the interdependent nature of social relations, individuals often perform actions that threaten face. In

fact, Brown and Levinson (1987, p. 65) argue that many speech acts intrinsically are face threatening, mean-

ing that by their very nature they run contrary to the face wants of the message source or target. Politeness

theory's concept of an intrinsic face-threatening act, or FTA, draws on assumptions from speech act theory

discussed earlier in this chapter. To see how, consider “directives” such as requests, recommendations, and

commands (Searle, 1976). Directives, like any speech act, are defined by constitutive rules or logical pre-

conditions (see Table 6.2). Based on these defining conditions, Brown and Levinson (1987, pp. 65–66) claim

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 36 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

that directives intrinsically threaten the target's negative face. When Mike asks Mr. Jenkins to donate blood in

Example 6.1, for instance, he assumes that Mr. Jenkins (a) is unlikely to set up an appointment on his own in

the near future (otherwise, Mike does not need to call Mr. Jenkins) and (b) might be willing to donate again

(otherwise, Mike is wasting his time calling Mr. Jenkins). By making assumptions about what Mr. Jenkins is

likely to do and what he may be willing to do, Mike constrains Mr. Jenkins's autonomy (Brown & Levinson,

1987, p. 144).

Based on the constitutive rules for various speech acts, politeness theory distinguishes between acts that

threaten negative face (i.e., constrain autonomy) and those that threaten positive face (i.e., communicate dis-

approval). It also distinguishes between acts that primarily threaten the message target's face and those that

threaten the message source's own face. Table 6.7 presents examples of speech acts that intrinsically threat-

en each of these four face wants. As is apparent in the table, each speech act is assumed to threaten one,

and only one, type of face.6

Brown and Levinson's (1987) classification of directives as intrinsic FTAs has important implications for the

study of compliance gaining. As I noted above in the section on obstacles to compliance, directives are at the

heart of any attempt to gain compliance. Given the centrality of directives to seeking compliance, it follows

that any attempt to gain compliance, by definition, places some degree of constraint on the target's negative

face. Thus the message source plausibly might worry about constraining the target's autonomy within any

compliance- gaining situation.

Seeking compliance is intrinsically face threatening, but not all attempts are equally face threatening. Accord-

ing to politeness theory, the amount of face threat created by any speech act is a linear function of three

variables: (a) distance, (b) power, and (c) ranking of the degree of imposition within a culture (Brown & Levin-

son, 1987, p. 74). Relational distance is the degree of familiarity between the message source and target.

Politeness theory assumes that as social distance increases, so does the magnitude or “weightiness” of face

threat. Imagine that you lost your purse and/or wallet and had to ask to borrow a quarter from a complete

stranger in order to make a telephone call. You might say something such as, “Excuse me, I'm sorry to bother

you, but I've lost my wallet. I need to make a phone call, and I have absolutely no money. Would you happen

to have a quarter?” Now imagine how you would ask your best friend for a quarter in the same situation. Can

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 37 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

you imagine saying “Excuse me”, or apologizing, or explaining in this much detail to your best friend? Accord-

ing to politeness theory, the same request for a quarter is more face threatening when the target is a stranger

than when the target is a close friend.

Power refers to the degree of status or control the message source has in relation to the target. Politeness

theory assumes that as the source's status increases relative to the target, the weightiness of any face threat

decreases. All other things being equal, a sergeant in the U.S. armed forces creates a larger face threat by

asking for assistance from a lieutenant than by asking for assistance from a private, because the private is

obligated formally to follow most directions from the sergeant.

SOURCE: From Politeness: Some Universals in Language Usage, by P. Brown & S. C. Levinson, 1987 (pp.

65–68). Copyright 1987 by Cambridge University Press. Adapted with permission.

Ranking refers to the extent to which a speech act, within a particular situation and cultural context, interferes

with the target's desire to maintain face. Asking to borrow $2,500 from a close friend places greater constraint

on the friend's autonomy than does asking to borrow a quarter, because the friend, by lending the larger sum

of money, may limit his or her own ability to make desired purchases or even to pay monthly bills. Politeness

theory assumes that, in any culture, distance, power, and rank are the crucial factors determining the weight-

iness of an FTA.7

When contemplating performing an FTA, politeness theory proposes, a speaker chooses from five options.

Table 6.7 Examples of Brown and Levinson's Four Types of Intrinsic Face-Threatening Acts

Acts That Primarily Threaten Speaker's Face Hearer's Face

Positive face Apologies Criticisms

Confessions Insults

Negative face Promises Requests

Offers Recommendations

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 38 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

These five “superstrategies” are shown in Table 6.8, along with examples from a scenario that assumes you

would like to buy a new house but are $3,000 short on the down payment. Should you ask your parent(s) to

loan you the money, realizing that it may be several years before you can repay the loan in its entirety? After

thinking about the situation, you might decide that asking to borrow $3,000 imposes too much on your parent.

In this case, you probably would choose not to do the FTA (see superstrategy 5).

If you do decide to perform the FTA, you must choose between going on- or off-record. Performing an FTA

“on-record” means saying something that makes your intention to ask for a loan clear, such as “Mom, can you

loan me $3,000 for my down payment on the house?” Performing an FTA “off-record” means saying some-

thing more ambiguous, that might or might not be heard as asking for a loan. Going off-record can be seen

as a strategic violation of Grice's (1975) maxim of manner. The example of superstrategy 4 in Table 6.8 is

off-record because, as the following examples show, it plausibly might or might not be a request for a loan:

Table 6.8 Brown and Levinson's (1987) Five Superstrategies for Doing FTAs, With Examples From a Hypo-

thetical Loan Scenario

Superstrategy Definition/Example

1. Do the FTA baldly,

without redress. State the FTA in the most direct, clear, and concise way possible.

Example: Mom, loan me $3,000.

2. Do the FTA using

positive politeness. Give assurances that the message source values the target and wants what the target wants.

Example: Mom, we're family, and we always help each other out. I know that you want me to live in a safer neighbor-

hood. I can do that if you'll lend me $3,000 for the down payment on my house.

3. Do the FTA using

negative politeness. Give assurances that the message source respects the target's freedom and will interfere with it as little as possible.

Example: Mom, I hate to ask you this, but is there any way that you might be able to lend me $3,000 for the down pay-

ment on my house? I'll pay you back as soon as I can.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 39 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

SOURCE: From Politeness: Some Universals in Language Usage, by P. Brown & S. C. Levinson, 1987 (pp.

65–68). Copyright 1987 by Cambridge University Press. Adapted with permission.

The parents in lines 2A and 2B respond to line 1 as implied request, whereas the parent in line 2C responds

to it only as an expression of frustration. Offrecord requests are polite in the sense that they provide the mes-

sage target with greater options. Rather than being forced to say either yes or no, the parent can respond to

line 1 as an implied request or instead choose to hear only the literal meaning of the statement (as in line 2C).

Going off-record, however, runs the risk that the parent may not even realize that a request is being made.

If you decide to perform the FTA on-record, you still must choose to do so with or without redress. Performing

the FTA baldly, without redress (superstrategy 1), means doing it “in the most direct, clear, unambiguous, and

concise way possible” (Brown & Levinson, 1987, p. 69). This means following Grice's (1975) maxims quite

literally, by making the request as clear as possible (manner), using no more words than absolutely necessary

(quantity), and so forth. Although most of us probably would not use a bald, on-record request such as the

one in Table 6.8 when asking to borrow a large sum of money, we might if the request were much smaller

(e.g., “Mom, loan me a dollar”) or if there was an urgent need to speak efficiently (e.g., “Mom, get out!” after

discovering that her house is on fire; Kellermann & Park, 2001).

Performing the FTA with redressive action means adding language to the request that attempts to minimize

or counteract potential face damage. Politeness forms, in this sense, are “principled deviations” from Grice's

(1975) quantity maxim, meaning that people who use them are saying more than literally is necessary in order

4. Do the FTA off-

record.

Do the FTA so that the message source's intent is ambiguous and so that the source can deny having performed the

FTA if necessary.

Example: I don't know how I'm going to come up with all the money for the down payment.

5. Don't do the FTA. Choose not to perform the FTA.

Example: Say nothing.

Example 6.11

1 You: I don't know how I'm going to come up with the money for the down payment.

2A Parent: How much are you short? Maybe I could help you with part of it.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 40 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

to be polite. Two types of politeness are available. Negative politeness (superstrategy 3) offers assurances

that the message source recognizes and respects the target's autonomy. The example in Table 6.8 communi-

cates negative politeness by (a) apologizing for the intrusion (“I hate to ask you”), (b) expressing doubt about

the target's ability to comply (“any way that you might be able to”), and (c) minimizing the length of the impo-

sition (“as soon as I can”). Positive politeness (superstrategy 2) offers assurances that the message source

likes and shares similar wants with the target. The example in Table 6.8 communicates positive politeness

by (a) using in-group markers (“we're family”), (b) asserting common ground (“you want me to live”), and (c)

providing reasons (“I can do that if”).

Politeness theory assumes that each of the five superstrategies can be carried out in multiple ways. Brown

and Levinson devote the bulk of the discussion in their book Politeness: Some Universals in Language Usage

(1987) to documenting concrete linguistic strategies by which speakers of English, Tamil, and Tzetal com-

municate negative or positive politeness. Table 6.9 summarizes 10 linguistic strategies that speakers use to

enact negative politeness. Mike uses several of these strategies in the phone call shown in Example 6.1, in-

cluding giving deference (line 02), being pessimistic (line 14), minimizing imposition (line 14), and using con-

ventionally indirect requests (line 23). Brown and Levinson also document 15 linguistic strategies by which

speakers in all three languages enact positive politeness. Politeness theory also assumes that negative and

positive politeness are alternative ways of redressing an FTA: by using avoidance-based language to show

deference and respect (negative politeness), or by using approach-based language to communicate approval

(positive politeness; see Brown & Levinson, 1987, pp. 68–70).

Table 6.9 Ten Linguistic Strategies for Enacting the Superstrategy of Negative Politeness

Linguistic Strategy Definition/Example

1. Be conventionally indirect. Use indirect request forms.

Example: Can you do X? (versus “Do X”).

2. Question, hedge. Avoid committing to assumptions underlying the request. Example: I was wondering whether you might…

3. Be pessimistic. Express doubt about whether the target can/will comply.

Example: Is there any way that you could … ?

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 41 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

SOURCE: From Politeness: Some Universals in Language Usage, by P. Brown & S. C. Levinson, 1987 (pp.

129–211). Copyright 1987 by Cambridge University Press. Adapted with permission.

When contemplating performing an FTA, how does a speaker choose between the five superstrategies? Po-

liteness theory makes two important assumptions to answer this question. The first is that the five super-

strategies are rank ordered from the least polite option (do the FTA baldly, without redress) to the most polite

option (don't do the FTA; see Table 6.8). Brown and Levinson (1987, p. 18) argue that negative politeness

(superstrategy 3) is more polite than positive politeness (superstrategy 2) because it is avoidance rather than

4. Minimize the imposition. Downplay the size/length of the request.

Example: I was just calling to see if you could …

5. Give deference. Humble oneself and/or emphasize the target's status.

Example: Excuse me, sir, but I wondered …

6. Apologize. Indicate reluctance to infringe on the target.

Example: I hate to ask this, but…

7. Impersonalize both parties. Avoid explicitly mentioning the source and/or target.

Example: It needs to be done (versus “You need to do it”).

8. State a general rule. State the request as an instance of some general rule.

Example: Employees will wash their hands before returning to work (versus “You need to wash your

hands …”).

9. Nominalize. Remove the actor from “doing” the request.

Example: Your cooperation is urgently requested (versus “We urgently request your cooperation”).

10. Go on record as incurring a

debt. Explicitly claim indebtedness to the message target.

Example: I'd be eternally grateful if you could …

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 42 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

approach based. The second assumption is that a speaker selects a superstrategy based on the perceived

weightiness of the FTA. When seeking compliance, the more a potential request threatens the message tar-

get's autonomy, the more the source will want to choose a more polite (i.e., a higher numbered) superstrat-

egy (see Table 6.8). As noted above, the message source should choose a less polite superstrategy when

concerns about efficiency outweigh face concerns. A source also is likely to choose a less polite superstrat-

egy when the degree of face threat from seeking compliance is low (e.g., the request is small, the message

source has greater relative power). A speaker who uses more politeness than is necessary risks implying, by

accident, that his or her request is very large. If I were to say to my stepdaughter, “Sheridan, I hate to ask

you this, but is there any way you possibly might be able help your mom?” she might well think, “If Steve's

being this nice about it, then Mom must want me to do something really awful.” In sum, politeness theory

does not assume that message sources always will be polite when seeking compliance. Rather, the theory

predicts that sources will vary their levels of politeness based on the degree to which their attempts to gain

compliance threaten the message targets' face. Politeness theory assumes that people are “rational” in the

sense of being able to reason from goals to linguistic means for accomplishing their goals.8

Evaluation of Politeness Theory

Useful theories are heuristic; that is, they stimulate research and debate about issues of importance. Brown

and Levinson's (1987) politeness theory fares well on this criterion. Scholars have drawn upon, extended,

and modified politeness theory to pose questions about people's choice of compliance-seeking and -resist-

ing messages across cultures, genders, and relationship types (e.g., Baxter, 1984; Holtgraves & Yang, 1992;

Lambert, 1996a; Leichty & Applegate, 1991; Lim & Bowers, 1991; Metts, Cupach, & Imahori, 1992; Wilson

& Kunkel, 2000). Others have tested whether persuaders who enact specific superstrategies, such as going

off-record, are perceived as polite, appropriate, competent, and/or effective (e.g., Blum-Kulka, 1987; Clark &

Schunk, 1980; Fairhurst, Green, & Snavely, 1984; Holtgraves & Yang, 1990; Kline & Floyd, 1990; Metts et

al., 1992). In addition to its application in the study of persuasive message production, politeness theory has

stimulated research on topics such as children's language acquisition, second-language learning, and cultural

universals in the use of honorifics (for a review, see Brown & Levinson, 1987, preface).

Politeness theory has been heuristic for at least two reasons. One is that the theory shows how people's

social concerns (i.e., the macro level) both reflect and constrain the linguistic details of their talk (i.e., the

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 43 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

micro level). Brown and Levinson (1987) are not the first to suggest that people in all cultures share basic

wants (e.g., Maslow, 1943; Schutz, 1971). Politeness theory also is not the first theory to describe and explain

linguistic patterns; indeed, CA scholars take this as their primary task. What is unique about politeness theory

is that it links the macro and micro levels. Brown and Levinson show how regularities in the wording of speech

acts can be explained by the assumptions that conversational participants (a) desire to maintain face, (b) are

mutually aware of this desire, and (c) are rational agents (see note 8).

A second reason politeness theory has heuristic value is that it highlights both cross-cultural similarities and

differences in influence messages. According to the theory, people in all cultures (a) desire approval and au-

tonomy; (b) have motives to support other people's face; (c) use distance, power, and rank to determine the

weightiness of face threat; and (d) have options such as going “on-record” versus “off-record.” On the other

hand, the theory recognizes that cultures vary in (a) how distance and power are perceived within the “same”

relationships (e.g., friendship), (b) how culture members rank the imposition of specific FTAs, (c) which per-

sons culture members view as “significant” and hence as persons whose approval is desirable, and (d) which

persons culture members view as having power or status and hence as persons who deserve face protection.

Rather than privileging commonality or difference, politeness theory attempts to account for both.

Despite these strengths, politeness theory has been widely criticized (e.g., Coupland, Grainger, & Coupland,

1988; Craig, Tracy, & Spisak, 1986; Goldsmith, 2000; Lim & Bowers, 1991; Tracy & Baratz, 1994; Wilson,

1992; Wilson et al., 1991/1992, 1998; for a response to several criticisms, see Brown & Levinson, 1987, pref-

ace). Here I describe four common criticisms. First, Brown and Levinson's (1987) two types of face have been

questioned. Lim and Bowers (1991) argue that the concept of positive face compounds two basic human

wants: the need to be included (which they term “fellowship face”) versus the need to have one's abilities

respected (which they term “competence face”). They propose three types of face and three corresponding

types of politeness. Fellowship face is supported by statements of solidarity. Competence face is supported by

statements of approbation. Finally, autonomy face is supported by the demonstration of tact. After analyzing

written responses to hypothetical compliance-gaining scenarios, Lim and Bowers conclude that (a) message

sources match types of politeness to types of face threat (e.g., requests that threaten the target's fellowship

face are redressed with solidarity statements), and (b) sources use all three types of politeness independent-

ly (e.g., making solidarity statements neither guarantees nor precludes that a source will make statements

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 44 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

of approbation). Distinguishing fellowship and competence face seems especially important for cross-cultural

research on persuasive message production, because these two types of positive face may vary in salience

within individualist versus collectivist cultures (for a discussion of individualism and collectivism, see Chapter

5).

A second criticism is that politeness theory's analysis of how face threats arise is grounded too heavily in the

form of speech acts, without sufficient analysis of the context in which acts are performed. Consider a mother

who tells her son, “You need to start trying in your math class.” This directive constrains the son's options and

thus threatens his negative face, but it simultaneously threatens the son's positive face by asserting that, up

to this point, he has not been “trying.” In contrast, a mother who asks her son to “please pass the salt” places

minor constraint on the son's autonomy but, in most cases, communicates no disapproval. When do directives

threaten only the message target's negative face, and when do they create multiple threats to the target's face

as well as to the source's own face? Brown and Levinson (1987) classify FTAs based only on the constitu-

tive rules for speech acts, irrespective of the contexts in which those acts are performed; thus they have little

to say about this issue (Wilson et al., 1991/1992). Aside from the form of speech acts, threats to face also

depend on the influence goal underlying an attempt to gain compliance (Wilson et al., 1998). For example,

women and men from diverse cultural backgrounds at times worry that they might appear to be “butting into”

the affairs of others when they give advice, but they rarely have this concern when asking favors. In contrast,

individuals at times worry that they might “owe” their targets big debts in the future when they ask favors, but

they rarely have this concern when giving advice (Cai & Wilson, 2000; Wilson, Anastasiou, Kim, Aleman, &

Oetzel, 2000; Wilson & Kunkel, 2000). Aside from influence goals that define an interaction, threats to face

also vary depending on the unfolding sequence of talk. Thus the degree to which giving advice threatens face

varies depending on whether the target has explicitly asked for advice, has disclosed a problem, or has said

nothing at all yet about the topic of advice (Goldsmith, 2000).

Third, Brown and Levinson's (1987) claim that positive politeness, negative politeness, and off-record super-

strategies are three mutually exclusive options has been criticized. Several studies have shown that, rather

than using only one superstrategy, message sources often mix two or three within the same compliance-seek-

ing message (e.g., Craig et al., 1986; Holtgraves & Yang, 1992; Lim & Bowers, 1991). Consider Example

6.12, in which a college student speaks aloud her response to a hypothetical situation that involves making a

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 45 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

large request of a friend:

Example 6.12 contains numerous positive politeness strategies, including using in-group markers (“we'll” and

“friends”), giving reasons (“I don't have a place to live”), assuming that the target is interested in the source's

wants (“I've got some great news”), and being optimistic (“I know we'll all get along well”). But the message

also contains negative politeness strategies, such as hedging and indirectness (“I was wondering if it would

be possible”), minimizing the imposition (“at the outside I would guess a couple of weeks”), and indicating re-

luctance to impose (“I hesitate”). Rather than choosing between mutually exclusive options, message sources

often mix different superstrategies, sequentially or even simultaneously, to redress an FTA.

A fourth criticism targets politeness theory's claim that, regardless of context, the five superstrategies can be

rank ordered from least to most polite. Several researchers have found exceptions to Brown and Levinson's

(1987) rank ordering, such as on-record strategies being perceived as more polite than off-record strategies

and positive politeness being rated as more polite than negative politeness (e.g., Baxter, 1984; Blum-Kulka,

1987; Dillard, Wilson, Tusing, & Kinney, 1997; Holtgraves & Yang, 1990). These exceptions suggest that the

perceived politeness of any superstrategy depends heavily on the specific relational, institutional, and cultural

context of the situation (Coupland et al., 1988; Dillard et al., 1997; Tracy & Baratz, 1994). Consider the follow-

ing example, in which a nurse (N1) attempts to convince an elderly hospital patient (M) to take her medication:

Although N1 in Example 6.13 issues many bald on-record requests, she also uses several positive politeness

strategies (e.g., in-group markers such as “love” and “darling”) and negative politeness strategies (e.g., mini-

mizing the FTA in “not much left” and giving reasons such as “bad stomach”). Despite these politeness forms,

the nurse's attempts to gain compliance may not be heard as polite. A nurse's addressing an elderly patient

as “love” and “darling” may be seen more as her treating the patient like a child than as communicating affec-

tion or respect. This impression is reinforced when the nurse (N1) talks about the patient to a colleague (N2)

in the patient's presence. Politeness judgments are contextually bound; hence any attempt to rank order the

five superstrategies across contexts seems doomed to failure. Example 6.12

Hey Milly, uh, I got some great news for you. Uh, I'll be moving to Philadelphia to start a new job and,

huh, they want me there on, uh, Monday of next week, uh, I don't have a place to live and, uh, need

to take some time to locate at a place. I was wondering, uh, if it would be possible for me to stay with

Example 6.13

1 N1: Milly (1.5) Milly (.) come on love (.) wake up (2.0)

2 ((4/5 syllables)) (1.0) that's

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 46 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

In light of these criticisms of politeness theory, scholars typically have taken one of two tacks. Some advocate

maintaining the theory's basic concepts (e.g., the intrinsic FTA, the three situational variables that determine

weightiness, the five superstrategies) but elaborating or modifying particular assumptions that have proven

problematic (e.g., Lim & Bowers, 1991; Wilson et al., 1991/1992, 1998). The challenge for this “conservative”

approach is whether it can address the criticisms noted above while maintaining the basic core of politeness

theory. A second group of scholars advocate scrapping politeness theory in favor of context-specific analyses

of politeness (e.g., Coupland et al., 1988; Tracy & Baratz, 1994). The danger of this more “radical” approach

is that context-specific analyses may lose sight of those cross-cultural linguistic regularities that motivated the

development of politeness theory in the first place. Regardless of the ultimate outcome, Brown and Levinson's

(1987) theory has stimulated a great deal of research about how detailed regularities in influence messages

can be understood as attempts to maintain face.

O'Keefe's theory of Message Design Logics

Barbara O'Keefe's theory of message design logics emerged from her earlier “goal and behavioral complexity”

analysis (see Chapter 5). After sketching the genesis of message design logics below, I describe the theory's

key concepts and assumptions, and then evaluate its utility.

Background of the Theory

O'Keefe's (1988, 1990, 1991) theory of message design logics responds to limitations in the earlier “goal and

behavioral complexity” account of message production. As you will recall from Chapter 5, O'Keefe and Delia

(1982) suggest that a communication situation is complex when its constituent features create multiple “sit-

uationally relevant objectives” that appear contradictory and difficult to reconcile. For example, interpersonal

arguments in which two parties disagree about a controversial issue are complex because both parties can

be expected to present and defend their own positions clearly, but also to show respect for the other party's

ideas (i.e., support the other's face). O'Keefe and Delia propose three strategies for managing such goal con-

flict: (a) selection, or giving priority to one goal (primary or secondary) while ignoring others; (b) separation, or

addressing multiple goals one at a time or through different channels; and (c) integration, or reframing the sit-

uation so as to pursue multiple goals simultaneously. To investigate these strategies, O'Keefe and Shepherd

(1987) paired college students with divergent opinions on a controversial issue, instructed each student to try

to convince the other to accept his or her own position on the issue, and then examined the students' use of

selection, separation, and integration strategies, turn by turn, throughout the discussions.

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 47 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Although O'Keefe initially viewed “selection”, “separation”, and “integration” simply as three different strate-

gies for managing goal conflict, she eventually has come to see them as reflecting three fundamentally differ-

ent ideas about the nature and possibilities of communication. As O'Keefe (1988) puts it, “These three ‘goal

management strategies’ reflect three different ways of reasoning from goals to messages rather than three

alternative message forms derived from the same way of reasoning about communication” (p. 83). To under-

stand her thinking, first consider the strategy of “separation.” During an interpersonal argument, a separation

strategy might involve directly criticizing the other person's views while also including apologies, disclaimers,

qualifiers, and other attempts to offset the criticism. Separation strategies can be interpreted easily within

Brown and Levinson's (1987) politeness theory. Criticizing another person's views can be heard as express-

ing disapproval of that person him- or herself. Given the mutual benefit of supporting face, a speaker who

criticizes another person's views should be motivated to redress threats to that person's positive face with just

the sort of politeness strategies (e.g., apologies, qualifiers) that O'Keefe and Shepherd (1987) observed in

separation strategies.

Both selection and integration strategies, however, make less sense from the perspective of politeness theory.

During an interpersonal argument, selection strategies might involve directly criticizing the other person's

views with no attempt to qualify or tone down the disagreement (i.e., pursue only the primary goal). Alterna-

tively, selection could involve lying about one's own position and claiming to hold the same view as the other

person in order to avoid disagreement (i.e., pursue only the secondary goal). Some of O'Keefe and Shep-

herd's (1987) participants told exactly such lies, even though both parties were explicitly informed at the start

that they were paired together precisely because they held highly divergent opinions. From the perspective of

politeness theory, why would an individual choose either to disagree “baldly, on-record” or to tell a detectable

lie about his or her own position when other options are available?

Integration strategies also are difficult to comprehend within politeness theory. During an interpersonal argu-

ment, integration strategies might involve directly criticizing the other person's views, but doing so within a

larger discussion of how considering the strengths and weaknesses of all positions—one's own, the other

person's, and other relevant views—allows everyone to make informed choices. Although the speaker still

advocates his or her own views, the speaker's actions are reframed from “criticizing the other's view” (an FTA)

to “mutually becoming better informed about the issue” (not an FTA). Which of Brown and Levinson's (1987)

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 48 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

five superstrategies is this (see Table 6.8)? As O'Keefe (1991) explains, reframing is not an instance of the

off-record superstrategy: “In indirectness, the FTA is performed but off-record; whereas in this approach, the

FTA is not performed at all” (p. 140). It also is not an instance of the “don't do the FTA” strategy, because the

speaker does attempt to influence the other party. To summarize, politeness theory presumes that a speaker

is limited to using linguistic strategies, within the current context, to redress FTAs, whereas integration strate-

gies presuppose the possibility of redefining the context so as to pursue the primary goal without threatening

face at all. Hence O'Keefe (1988) concludes that the three goal management strategies “result from variation

in the systems of principles used in reasoning from ends to communicative means, differences in the very

definitions of communication that individuals construct and employ” (p. 84).

Key Concepts and Assumptions

Based on the points noted above, O'Keefe (1988) argues for the existence of three message design logics, or

systems of means-ends reasoning about communication. A message design logic is “a constellation of related

beliefs: a communication constituting concept, a conception of the functional possibilities of communication,

unit formation procedures, and principles of coherence” (p. 84). The design logics represent three fundamen-

tally different ideas about what talk is relevant in social situations (Grice, 1975). Key characteristics of each of

these three design logics, which O'Keefe labels expressive, conventional, and rhetorical, are shown in Table

6.10.

The fundamental premise of the expressive logic is that communication “is a medium for expressing thoughts

and feelings” (O'Keefe, 1988, p. 84). Individuals using this logic view “saying what's on one's mind” as the

primary function of communication. Expressive individuals communicate directly and literally, and they inter-

pret other people's messages in the same way. They view communication as successful to the degree that

it is clear and honest. Messages generated by the expressive logic tend to focus temporally on the past. If

an individual using the expressive logic is asked, “Why did you say X now?” the answer typically is “Because

some immediately prior event made me think or feel X, and so I said X.” Expressive messages often contain

“pragmatically pointless” content, in the sense that the speaker says what he or she is thinking even if these

thoughts do not help accomplish situationally relevant objectives. When faced with conflicting goals, such as

seeking compliance but also supporting the target's face, persons employing the expressive logic see two

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 49 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

(and only two) options: Say it or don't say it.

Consider how someone using the expressive design logic might respond to the following “group leader” hy-

pothetical situation:

You have been assigned to a group project in one of your classes…. it is important to you to get a

good grade in this class…. the instructor … designated you to be the leader of your group…. Your

duties as group leader will include telling the instructor what grade you think each individual in the

group deserves based on their individual contributions. One group member (whose name is Ron) …

seldom makes it to group meetings on time and entirely skipped one meeting without calling anyone

in advance…. At the next meeting, Ron arrived late but apologized for missing the previous meet-

ing and mentioned something about family problems. Ron did volunteer to do all the background

research on one important aspect of the group's topic…. the group project is due next week. The

group planned to put together the final draft of its report at the meeting scheduled for tomorrow af-

ternoon. Ron calls you up today and says that he doesn't have his library research done and can't

get it finished before the meeting. He says he just needs more time. (O'Keefe, 1988, p. 93)

Table 6.10 Characteristics of Three Message Design Logics

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 50 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

What would you say to Ron? O'Keefe asked 92 undergraduate students to write out what they would say in

this situation. One student wrote the following:

Ron, I can't believe you haven't finished your research. You've been inconsiderate to the group all

along. Several members even suggested that you be taken out of the group but we decided to give

you a chance. Now what are we supposed to do? It was your responsibility and you backed out. I'm

afraid I'm going to have to tell the T.A. that you haven't done your share. I will be so mad if we get a

bad grade on this—I need an A in this course. (quoted in O'Keefe, 1990, p. 95).

Several characteristics of the expressive logic are evident in this message. Much of the message simply de-

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 51 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

scribes and evaluates Ron's past behavior. The message also contains a good deal of pragmatically pointless

content. There is still a week before the project is due, and 24 hours before the group's next meeting. Despite

this, the student does not attempt to persuade Ron to make progress on his research. Most of the message

simply vents the student's frustration with Ron. Being frustrated is an understandable reaction, but that alone

does not forward the group's task. Even the threat is noncontingent. The student simply says that he or she is

going to tell the T.A. about Ron's failure to contribute, rather than that he or she will do this unless Ron starts

doing his part.9

The fundamental premise of the conventional message design logic is that communication “is a game played

cooperatively, according to socially conventional rules and procedures” (O'Keefe, 1988, p. 86). Communica-

tion still is viewed as a means of expressing ideas, but these ideas reflect the social effects that the speaker

wants to achieve rather than the thoughts that happen to be on the speaker's mind. For example, if an indi-

vidual using this logic needed assistance, he or she might say something that, given the context, would count

as a “request”, because making a request is a way of obtaining assistance by calling on the conventional

obligations that exist between persons. Individuals applying this logic view communication as successful to

the degree to which speakers can achieve their ends while adhering to the norms of appropriateness. They

view communicative contexts as “predefined” or “given”, meaning that people's roles, rights, and relations are

established by the nature of a situation and hence are relatively inflexible.

Messages generated by the conventional logic tend to be focused temporally on the present. If an individual

using the conventional logic is asked, “Why did you say X now?” the answer typically is “Because I wanted

to accomplish Y, and X is the typical way of doing so”, or “Because you just said Z, and X is the relevant

response to Z.” Conventional messages often include explicit references to constitutive rules that underlie

speech acts, rights and obligations that give rise to speech acts, and mitigating conditions that bear on those

rights and obligations. When faced with conflicting goals, such as seeking compliance but also supporting

face, persons using this logic will employ “conventional” solutions such as politeness strategies. Consider a

second student's response to the Ron scenario:

Well, Ron, I'm sorry you don't have your part of the project done. We have given you several breaks

thus far and I don't see how we can give you any more. The whole group is depending on you so I

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 52 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

would suggest to you to get it done or at the most bring in what you have got done. If you don't get

this done I'm going to have to give you an F for the project. If you can't hold up your responsibility

with this group even under these adverse conditions (family problems), how are you going to make

it in life? (quoted in O'Keefe, 1990, p. 96)

Several characteristics of the conventional logic are evident in this message. This student addresses the pri-

mary situationally relevant objective explicitly, by “suggesting” that Ron get his research done by the meeting

and by following through with a conditional threat. He or she also mentions Ron's obligations: Ron hasn't done

his “part” of the project, and he has a “responsibility” to the group. Compared with the expressive message

above, this message spends less time berating Ron and more time emphasizing what Ron should be doing

at this point in time. This message also briefly acknowledges the mitigating circumstances (family problems)

cited by Ron.

The fundamental premise of the rhetorical logic is that “communication is the creation and negotiation of social

selves and situations” (O'Keefe, 1988, p. 87). Within this logic, selves and social situations are not viewed

as predefined by a conventional system of rules; rather, selves are enacted in dramas, and social situations

are constantly “in flux.” From the rhetorical view, communication is the process by which context is created.

Rather than being “given” and “fixed”, context is interactionally negotiated; speakers can redefine it strategi-

cally to accomplish their goals. Individuals applying this logic view communication as successful when par-

ticipants reach a consensus about the reality in which they want to be engaged (i.e., agreement about “what

is going on”). Rhetorical communicators shy away from relying on formal authority based on their roles or

positions. Instead, they seek consensus by advocating the benefits of the reality they enact and by listening

carefully to determine the drama and perspective being enacted by others.

Messages generated by the rhetorical logic tend to focus temporally on the future. If an individual using the

rhetorical logic is asked, “Why did you say X now?” the answer typically is “Because I am trying to accomplish

goal Y.” Rhetorical messages frequently contain explicit context-describing phrases. Rather than citing rules

or sanctions, rhetorical messages often contain “rationale” arguments about why the speaker's definition of

the situation is advantageous and makes sense. When faced with conflicting goals, such as seeking compli-

ance and supporting face, persons using this logic look for ways to redefine the situation so as to induce the

desired action without threatening face. Consider yet another student's response to the Ron scenario:

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 53 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Well, Ron, it's due next week, and we have to get it all to the typist. OK, if it's not done it's not. Tell

you what. Why don't you jot down your main ideas so that we can include them in the introduction

and the conclusion. Also, tell me when you think your section should come in the whole project.

Then get it to my apartment by 10:00 the next day because I have to get it to the typist by 2:00. Is

this okay? I'll just explain to the group that you'll have it done but not by meeting time. We all want

a good grade, so if you need the time to make your part better, go ahead. But if I can't get it to the

typist in time, you'll have to type it. Alright, take it easy. (quoted in O'Keefe, 1990, p. 97)

This message illustrates several characteristics of the rhetorical logic. Most of the message is future oriented,

proposing a step-by-step plan for getting Ron's work done. The student lays out what Ron needs to do at

each point in time and outlines his or her own role as well. He or she avoids threatening Ron and seeks rather

than presumes Ron's agreement with the plan (“Is this okay?”). The speaker also redefines what is going on

in a face-saving way: Ron is taking time to make his part better so as to assure a good group grade, rather

than just completing his part belatedly.

O'Keefe (1988, 1990, 1991) argues that these three message design logics form a natural developmental

progression. The first logic an individual must acquire is the expressive, so that he or she understands that

words can be used to express internal thoughts and feelings. Should the individual learn that words can be

used to perform actions in given contexts, and not just to express thoughts directly, then the conventional

logic has subsumed the expressive. Should the individual come to understand that communication conven-

tions and contexts can be redefined, then the rhetorical logic has subsumed the conventional. According to

O'Keefe, an individual does not discard one design logic when he or she acquires another: the conventional

logic user still employs words to express thoughts as well as to perform other actions, and the rhetorical logic

user may choose to play within the situation as currently defined rather than to propose a new situation.

In sum, O'Keefe's (1988) theory presumes that messages “are produced by two conceptually and empirically

separable features of the message design process” (p. 96): (a) the message source's goals and (b) the design

logic that a message source uses to reason from goals to messages. The people who produced different mes-

sages in response to the Ron scenario, for instance, could do so because they formed and pursued different

goals: One individual might pursue none of the potential situationally relevant objectives (e.g., convincing Ron

to do his part, saving Ron's face), a second might pursue only one of these goals, and a third might pursue

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 54 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

multiple goals. Distinct from the number of goals that motivate a message, however, people also might pro-

duce different messages in response to the Ron scenario because they rely on different systems of reasoning

to move from goals to messages. Although expressive, conventional, and rhetorical communicators all may

form multiple goals, each sees fundamentally different options for managing goal conflicts so as to maintain

face (see the bottom row in Table 6.10).

Given this conceptualization, O'Keefe (1988, 1990; O'Keefe, Lambert, & Lambert, 1997) argues that people

relying on different design logics are most likely to produce different messages, as well as to misunderstand

each other, within “complex” communication situations. Simple situations tend to call out the same message

from all communicators. In contrast, differences in design logics are most likely to affect the production and

interpretation of messages in complex situations where message sources must manage multiple, contradic-

tory situational objectives. In the responses to the Ron scenario presented above, for example, an expressive

person might interpret the rhetorical individual's statement that Ron is taking more time to “make [his] part

better” as an inaccurate and dishonest description of what is “really” going on. Similarly, the rhetorical person

might view the expressive individual as simply wanting to “chew out” Ron, even though this may create more

work for other group members who then have to complete Ron's part.

Evaluation of O'Keefe's Theory

O'Keefe's (1988) theory of message design logics incorporates the work of discourse scholars such as Grice

(1975) and Searle (1969), as well as the symbolic interactionism of Goffman (1959, 1967), to describe and

explain individual differences in the message production process. Parts of the theory have received empir-

ical support. For example, O'Keefe assumes that the expressive, conventional, and rhetorical design logics

are developmentally ordered and reflect increasing levels of communicative competence. If this is indeed

the case, then people should rate compliance-seeking/resisting messages that display the rhetorical logic

as more competent and appropriate than expressive or conventional messages. Data from several studies

are consistent with this claim (Bingham & Burleson, 1989; O'Keefe, Lambert, & Lambert, 1993; O'Keefe &

McCornack, 1987). O'Keefe and McCornack (1987) had 213 undergraduate students each read a different

compliance-seeking message that had been generated in a previous study in response to the Ron scenario.

These messages already had been coded for the number of goals pursued by the message source (none,

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 55 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

one, multiple) as well as for the design logic evident in the message (scored 1 = expressive, 2 = conven-

tional, 3 = rhetorical). After reading the message, each participant rated the likelihood that it would get Ron

to complete his part, the degree to which the message supported Ron's face, and the degree to which they

perceived the source of the message as competent and attractive. Participants who evaluated influence mes-

sages with more “advanced” design logics, as opposed to messages that displayed the expressive logic, on

average perceived the message as more likely to get Ron to do his part and more supportive of Ron's face,

and the message source as more competent and attractive.

Although the rhetorical design logic may be more competent, in the sense of providing a broader range of

options for managing complicated compliance- gaining situations, O'Keefe et al. (1997) do emphasize that

“every design logic provides a logically consistent and potentially satisfactory way for an individual to use lan-

guage” (p. 49). Rather than training expressives and conventional to use the rhetorical logic, O'Keefe et al.

recommend training everyone “in more general strategies to help them recognize and accommodate to di-

versity in design logics” (p. 50). For example, people might be taught to appreciate the system of reasoning

behind all three sample responses to the Ron scenario (see note 8). This recommendation is consistent with

recent evidence indicating that how individuals interpret and evaluate messages displaying the rhetorical de-

sign logic depends on the individuals' own particular design logics (O'Keefe et al., 1997; Peterson & Albrecht,

1996).

Although some data support assumptions from O'Keefe's theory, several challenges for the theory remain.

First, one might argue that O'Keefe's procedures for measuring design logics and testing the theory have

been tautological: A person's design logic is inferred from qualities of his or her compliance- seeking mes-

sages, but then design logics are used to “explain” why that person produced that type of message. One

might hope for a measure of design logic independent of messages, one that taps more directly into design

logics as a “constellation of related beliefs” (O'Keefe, 1988, p. 84) and “type of knowledge organization” (p.

97). Measuring design logics is complicated, however, because design logics are systems of implicit and tac-

it beliefs; one cannot simply ask a person, “What's your design logic?” O'Keefe and Lambert (1989) have

made initial strides in developing a “thought checklist” procedure that measures the types of thoughts typically

salient to expressive, conventional, and rhetorical communicators. Much more research is needed to estab-

lish the reliability and validity of this newer measure (see Waldron & Cegala, 1992).

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 56 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

Second, one might argue that O'Keefe (1988) should provide more detail about exactly how individuals, using

any design logic, go about “reasoning from goals to messages” (p. 83). On what types of prior knowledge do

individuals rely when they are deciding what to say as they seek/resist compliance? How is the knowledge

composing a design logic integrated with knowledge about particular message targets and influence goals?

How are people able to “reason” from goals to messages within the actual time constraints of face-to- face

interaction? These questions call for more detail about the psychological processes that underlie message

production. I review cognitive theories of message production, including O'Keefe and Lambert's (1995) ideas

about psychological processes, in the next chapter. But as O'Keefe (1992) has argued, the theory of design

logics provides important insights into persuasive message production, even without a detailed analysis of

psychological process. The theory helps explain why people see different ideas as relevant to express in the

“same” compliance-gaining situation, as well as why people draw different interpretations from the “same”

compliance-seeking message. By emphasizing how situations are constantly defined through interaction, the

theory also shows how individuals, when seeking and/or resisting compliance, have means beyond formal

politeness strategies for addressing threats to face.

Summary

Discourse theories of persuasive message production perspectives focus on how people rely on knowledge

of linguistic forms and conventions when generating influence messages, and how they design messages

in light of the unfolding talk. After using an extended compliance-gaining sequence to illustrate the types of

questions raised by discourse scholars, I have reviewed in this chapter some basic concepts from work on

conversational maxims, speech acts, and the conversation analysis perspective. Following this background

material, I have explored in greater detail two concepts of central relevance to persuasive message produc-

tion: (a) obstacles, analyzed from the view of speech act and attribution theories; and (b) face, analyzed from

the view of politeness and message design logic theories.

Notes

1. As Schiffrin (1994) notes, the field of discourse analysis is “widely recognized as one of the most vast, but

also one of the least defined, areas in linguistics” (p. 1). It incorporates works from disciplines in which mod-

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 57 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

els and methods for understanding discourse first were developed (e.g., anthropology, linguistics, philosophy,

sociology) as well as disciplines that have applied and extended such models given their own concerns (e.g.,

communication, cognitive science, social psychology). To describe this breadth, Schiffrin compares a “formal-

ist paradigm” of discourse, which looks at the internal structure of language units “above the sentence”, with a

“functionalist paradigm”, which looks at the interdependence of discourse and social life by exploring functions

of language external to the linguistic system itself. Some of the theory and research reviewed in this chap-

ter falls outside the bounds of even a broad treatment of discourse such as Schiffrin's. Each work reviewed

here, however, speaks to what Schiffrin identifies as the two central goals of discourse analysis: (a) sequential

goals, or discovering principles underlying the ordering of utterances; and (b) semantic or pragmatic goals, or

discovering how the structure and use of language allow people to convey and interpret meaning (p. 41).

2. The patient's contribution in lines 2–3 still could be interpreted sensibly if the patient was signaling, indi-

rectly, his refusal to cooperate with the physician. If the patient simply is not orienting to the doctor's question,

however, then lines 2–3 violate the relevance maxim.

3. Searle (1969, p. 66) also includes an “essential” rule for requests, which states that the speaker's message

counts as an attempt to get the target to perform the desired action. For any speech act, the essential rule

spells out the speaker's illocutionary force. The essential rule assumes that the speaker has adhered to each

of the other rules listed in Table 6.2.

4. Different people might perceive the same cause as falling at different points on one of the three dimensions;

for example, some might perceive Joan's fight with her boyfriend as a stable cause if they fight frequently. The

general point, however, is that people use the dimensions of locus, stability, and controllability to describe and

compare different causes.

5. Brown and Levinson (1978) originally published their politeness theory in a 250-page book chapter that

appeared as part of a collection of essays in volume 8 of the book series Cambridge Papers in Social Anthro-

pology. Almost 10 years later, Cambridge University Press reissued that original chapter, accompanied by a

new 50-page introduction and an expanded bibliography, as a book (Brown & Levinson, 1987). Because of

its wider availability, I have used the reissued 1987 book for all quotations and materials presented here.

6. Brown and Levinson (1987, pp. 67, 286) do qualify that some acts, such as requests for personal infor-

mation, intrinsically threaten the target's positive and negative face. Unfortunately, they do not provide any

guidelines for determining when an act intrinsically threatens multiple types of face, nor do they consider the

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 58 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others

possibility that some acts may create either one or multiple face threats depending on the context. Brown and

Levinson's qualification also is inconsistent with their own formula for determining the magnitude of an FTA.

For more on this issue, see Wilson et al. (1991/1992, 1998).

7. Politeness theory recognizes that distance (D), power (P), and rank (R) are context dependent. A bank

manager, for instance, typically will have greater power than a bank teller, but perhaps not if the teller rep-

resents his or her union in negotiations or sits on a jury that is trying the manager for embezzlement. Thus

“situational factors enter into the values of D, P, and R, so that the values assessed hold only for [the speaker]

and [the hearer] in a particular context, and for a particular FTA” (Brown & Levinson, 1987, p. 79). Politeness

theory also recognizes that values of D, P, and R are socially negotiated. Speakers may use greater or less-

er politeness than normally is expected in a situation precisely to affect the perceived values of D, P, and R

(Brown & Levinson, 1987, pp. 81–83).

8. Brown and Levinson's (1987, pp. 58–59) claim that people are “rational” often is misunderstood. It is not in-

tended to mean that participants always think a great deal about what to say during influence interactions, or

that participants always choose the most effective means for accomplishing their influence and interperson-

al goals. Rationality simply implies that people possess knowledge about linguistic means for accomplishing

goals, and that their talk can be understood as an attempt to balance their competing desires to avoid threat-

ening face versus to be clear and say no more than is necessary.

9. An expressive communicator might defend this response to the Ron situation by arguing that it is obvious,

given the events, that Ron cannot be counted on to complete his part of the project, and that the group would

be better off excluding Ron and finishing the work on their own. In this sense, the goal of persuading Ron to

do his part may no longer be a sensible “situationally relevant objective.”

• politeness theory

• speech acts

• adjacency pairs

• compliance

• targeting

• speech act theory

• logic

https://dx.doi.org/10.4135/9781452233185

SAGE

© 2002 by Sage Publications, Inc.

SAGE Books

Page 59 of 59 Seeking and Resisting Compliance: Why People Say What They Do When

Trying to Influence Others