mummxumummmm.
TOAVOIDFINESmnonorbdonddou.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
Walla-9.1

 

 

 

 

MSU ’ RETURNING MATERIALS:
Place in book drop to
uaamuas remove this checkout from

‘—_ your record. .FINES will
v r be charged if book is

 

returned after the date
stamped be10w.

 

”’5’; f

This is to certify that the

dissertation entitled

A NEW MODEL OF POLITENESS IN DISCOURSE

presented by V '

Tae-Seop Lim -

has been accepted towards fulfillment
' of the requirements for

 

 

 

 

Ph . D. degree in Communlcatlon
' I .
/ Major professor
MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

“mu“...— l. -. .-.

A NEW MODEL OF POLITENESS IN DISCOUBSE

By

Tae-Seop Lil

A DISSERTATION
Sub-itted to
Michigan State University
in partial fulfill-cut of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Depart-ent of Con-unication

1988

 

 

 

 

ABSTRACT
A NEW MODEL OF POLITENESS IN DISOOURSE
By
Tae-Seop Lim

This study explored Brown and Levinson’s model for its adequacy in
explaining politeness phenomena. Since Brown and Levinson’s model
revealed substantial limitations, a new theoretical nodel of politeness
that extended Brown and Levinson’s fralework was proposed. This new
Iodel conceptualizes face wants and politeness more comprehensively and
accounts for both positive and negative politeness manifested in
performances of both face-threatening acts and non-face-threatening
acts. Positive and negative politeness expressed in a message are
postulated to be a function of speakers’ verbal aggressiveness,
empathy, and the obligations to save positive and negative face,
respectively. The obligations to save positive and negative face are
postulated to be a sum of the relational obligations to save positive
and negative face, respectively, and the threats to positive and
negative face, respectively. The relational obligation to save
positive face is postulated to be positively related to the power the
bearer has over the speaker, and negatively related to the relational
distance between the speaker and the hearer; the relational obligation
to save negative face is postulated to be positively related to both
power disparity and relational distance.

Two experiments were conducted to test the adequacy of the present
Iodel in explaining politeness in discourse. Experiment I examined
whether the present model explained adequately politeness behavior in
phatic communication where no intrinsically face-threatening acts

ii

(non-FTAs)
performing
politeness
Experiment
politeness
performing

politeness

were involved. This experiment found that even when
non-FTAs, subjects manifested both positive and negative

as a function of relational distance and power disparity.

2 tested the adequacy of the present model in explaining
phenomena related to FTAs. This experiment found that when
FTAs, subjects employed both positive and negative

as a function of relational distance, power disparity, the

levels of positive and negative face-threats, and verbal

aggressiveness. These results generally supported the present model

and rejected Brown and Levinson’s model. In addition, based on

utterances

generated in Experilents l and 2, a content coding system

for politeness was developed.

iii

Dedicated to my wife

Myung-Hee

iv

ACKNOWLEDGEMENTS

Throughout my graduate education, I have been fortunate enough to
have several mentors and endless support fron my family. The members
of my dissertation comnittee, Drs. Kathy Kellermann, Gerald Miller,
Franklin Boater, and Paul Munsell, are respected scholars. Each did
hours of reading, commenting, and constructively criticizing so that
this dissertation could be completed. The quality of the dissertation
was enhanced by their guidance and expertise. I am grateful for their
contributions to not only this dissertation but also to the general
quality of my graduate education.

Dr. Kathy Kellermann, the chair of the dissertation committee,
deserves additional mention. For several years she has taken an active
role in my professional development. As my research team advisor she
heightened ny intellectual growth and productivity. She showed me a
great deal of understanding that helped me adjust myself to new
environments. Her tutelage and friendship have been, and will
continue to be, valued by Is.

Back in my country, Korea, I have four great supporters: sy
parents and parents—in-law. They shared my goal of conpleting a
doctoral degree. They prayed for me at every possible moment and
sacrificed their welfare to support me financially. Without their
moral and financial support, this dissertation could not have been

conpleted.

For more than five years, Iy wife Myung-Hee gave me endless
support. When I was depressed, she cheered me up. When I needed
assistance, she gave Ie her hand unconditionally. When I was busy
working, she showed unconnon patience. Especially, she took care of
our new son Jay Phil allost by herself so that I could have tips to
work on this dissertation. If she had been less supportive, I could

not have made it.

vi

TABLE OF CONTENTS

LI ST OF TABLES O O O O O O O O O O O O O O O O O O O

LI ST OF F1 GURES O O O O O O O O O O O O O O O O O . O

LIST OF SYMBOLS . . . . . . . . . . . . . . . . . .

CHAPTER

I.

II.

A MODEL OF POLITENESS BEHAVIOR . . . . . .

INTRODUCTION . . . . . . . . . . . . . .
BROWN AND LEVINSON’S MODEL OF POLITENESS
Face and Face-Threat . . . . . . .
Selection of Politeness Strategy .
Tests of Brown and Levinson’s Model
CRITIQUE OF BROWN AND LEVINSON’S MODEL AND
DIRECTIONS FOR A NEW MODEL . . . . . . . .
Limited to One Kind of Face-Threat at
Lisited to Politeness Related to PTA
Limited to One Side of Positive and

a

Tile

Negative Politeness . . . . . . . . . . . . .
Proposition Based on Inconsistent Definitions .

Conclusion of Critique of Brown and
Levinson’s Model . . . . . . . . . .
A NEW MODEL OF POLITENESS . . . . . . . .
Face and Face-Saving Obligation . . .
Face-Threat and Face-Saving Obligation
Politeness and Face-Saving Obligation
Politeness of a Message . . . . . . .
Explanation of Previous Findings . .

EXPERIMENT 1 O O O O O O O O O O O O O O O

HymTHESES O O O O O O O O O O
METHODS O O O O O O O O O O O
Situation and Manipulation of

Variables . . . . .
Pilot Study . . . .

vii

O O O O O O
O O O O O O O
O O O O O O O

Page

xi

xii

H

HQCOOOH

l4
l4
17

22
27

30
32
32
38
40
42
46

51

51
53

54
56

CHAPTER Page

Participants and Procedure . . . . . . . . . . 59
Instrumentation . . . . . . . . . . . . . . . . 59
Coding . . . . . . . . . . . . . . . . . . . . 63
RESULTS . . . . . . . . . . . . . . . . . . . . . . 68
Manipulation Check . . . . . . . . . . . . . . 68
Politeness in Performing a
Non-Face-Threatening Act . . . . . . . . . . . 70
The Effects of Personality Variables . . . . . 71
The Effects of Relational Variables . . . . . . 74
The Effect of Gender . . . . . . . . . . . . . 78
DISCUSSION . . . . . . . . . . . . . . . . . . . . . 79
III. EXPERIMENT 2 . . . . . . . . . . . . . . . . . . . . 83
HYPOTHESES . . . . . . . . . . . . . . . . . . . . . 83
METHODS . . . . . . . . . . . . . . . . . . . . . . 85
Situation and Manipulation of Independent
Variables . . . . . . . . . . . . . . . . . . . 85
Pilot Study . . . . . . . . . . . . . . . . . . 89
Participants and Procedure . . . . . . . . . . 94
Instrumentation . . . . . . . . . . . . . . . . 95
Coding . . . . . . . . . . . . . . . . . . . 96
RESULTS . . . . . . . . . . . . . . . . . . . . . . 99
Manipulation Check . . . . . . . . . . . . . . 99

Co-occurrence of Positive and

Negative Politeness . . . . . . . . . . . . . . 105

The Effects of Personality Variables . . . . . 105

The Effects of Face-Threats and

Relational Variables . . . . . . . . . . . . . 112

The Effect of Gender . . . . .
DISCUSSION . . . . . . . . . . . . . . . . . . . . . 124

IV. DEVELOPMENT OF CONTENT CODING SYSTEM . . . . . . . . 131

STATEMENT OF PROBLEM . . . . . . . . . . . . . . . . 131
Review of Literature . . . . . . . . . . . . . 133
METHODS . . . . . . . . . . . . . . . . . . . . . . 135
Sample Utterances . . . . . . . . . . . . . . . 135
Identification of Unique Politeness Devices . . 136
Rating of Politeness . . . . . . . . . . . . . 136
RESULTS . . . . . . . . . . . . . . . . . . . . . . 137
Confidence of Approval . . . . . . . . . . . . 137
Exaggeration of Approval . . . . . . . . . . . I38
Tentativeness of Disapproval . . . . . . . . . 141
Understatedness of Disapproval . . . . . . . . 142
Hesitance of Imposition . . . . . . . . . . . . 145
Indirectness of Imposition . . . . . . . . . . 148
Deference Shown in Use of Address Terls . . . . 151
DISCUSSION . . . . . . . . . . . . . . . . . . . . . 151

viii

CHAPTER Page
V. GENERAL DISCUSSION AND CONCLUSION . . . . . . . . . . 157

GENERAL DISCUSSION . . . . . . . . . . . . . . . . . 157
The Present Model vs. Brown and Levinson
Model . . . . . . . . . . . . . . . . . . . . . 157
FTA vs. Non-PTA . . . . . . . . . . . . . . . . 159
Relative Ilportance of Deter-inants of

Politeness . . . . . . . . . . . . . . . . . . 161
The Role of Gender . . . . . . . . . . . . . . 162
The Role of Personality Variables . . . . . . . 164
Dilensionality of Politeness . . . . . . . . . 165
Modification of the Present Model . . . . . . . 167
The Influence of Different Goals . . . . . . . 170
SUMMARY AND CONCLUSION . . . . . . . . . . . . . . . 172

APPENDI C.ES O O O O O O O O O O O O O O O O O O O O O O O O O O 1 78

BI BLIWRAPHY O O O O O O O O O O O O O O O O O O O O O O O O O 192

ix

11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

23.
24.

25.
26.
27.
28.
29.
30.
31.

32.
33.
34.

LIST OF TABLES

Chart of Tactics . . . . . . . . . . . . .
Descriptive Statistics for Rated Distance
Descriptive Statistics for Rated Power . .
Verbal Aggressiveness Measurement Scale .
Enpathy Measurement Scale . . . . . . . .
Politeness Rating Scale . . . . . . . . .
Descriptive Statistics for Rated Distance .
Descriptive Statistics for Rated Power . . .
Correlations Between Positive Politeness and
Predictors . . . . . . . . . . . . . . . . . . . . . .
Correlations Between Negative Politeness and
Predictors . . . . . . . . . . . . . . . . . . .
Descriptive Statistics for Positive Politeness .
Descriptive Statistics for Negative Politeness .
Descriptive Statistics for Rated Distance . . .
Descriptive Statistics for Rated Power Disparity . .
Descriptive Statistics for Rated Positive Face-Threat
Descriptive Statistics for Rated Negative Face-Threat
Confirmatory Factor Analysis of Positive Politeness .
Descriptive Statistics for Rated Distance . . . . . .
Descriptive Statistics for Rated Power Disparity . . .
Descriptive Statistics for Rated Positive Face-Threat
Descriptive Statistics for Rated Negative Face-Threat
Correlation Analysis for Approach-Based

Positive Politeness . . . . . . . . . . . . . . . . .
Correlation Analysis for Avoidance-Based

Positive Politeness . . . . . . . . . . . . . . . . .
Correlation Analysis for Avoidance-Based

Negative Politeness . . . . . . . . . . . . . . . . .
Means and $03 for Approach-Based Positive Politeness .
Means and SDs for Avoidance-Based Positive Politeness
Means and SDs for Avoidance-Based Negative Politeness
Content Coding Scheme for Confidence of Approval . . .
Content Coding Scheme for Exaggeratedness of Approval
Content Coding Scheme for Tentativeness of Disapproval
Content Coding Scheme for Understatedness

of Disapproval . . . . . . . . . . . . . . . . . . . .
Content Coding Schele for Resitance of Imposition . .
Content Coding Scheme for Indirectness of Imposition .
Equations in the Final Model . . . . . . . . . . . . .

O O O O O O
O O O O O O O O
O O O O O O O O
O O O O O O O O
O O O O O O O O
O O O O O O O O

X

Page

10
58

58
60
62
64
69
69

72

72
75
75
91
91
93
93
97
100
101
102
103

108
109

110
113
114
115
139
140
143

144
146
147
169

LIST OF FIGURES

FIGURE Page

1. Diagralsatic Representation of Brown and Levinson
Made 1 O O O O O O O O O O O O O O O O O O O O O O O O 7

xi

SYMBOLS
RO .
P
ROn .
O .
P

O .
n

P .
P

P .
n
ap °
P .
vp

P .
an

P .
vn

T .
P

T .
n

A .
E .
P .

D .
W .
x

R .
x
D[S,H]
P[H,S]

LIST OF SYMBOLS
MEANING
Relational obligation to save positive face
Relational obligation to save negative face
Total obligation to save positive face
Total obligation to save negative face
positive politeness
negative politeness
Approach-based positive politeness
Avoidance-based positive politeness
Approach-based negative politeness
Avoidance-based negative politeness
Positive face-threat
Negative face-threat
Verbal Aggressiveness
Espathy
Power disparity
Relational distance
Relative imposition
Absolute iaposition
Social distance between com-unicators

Power the bearer has over the speaker

xii

CHAPTER I

A MODEL OF POLITENESS BEHAVIOR

INTRODUCTION

Being polite is a social value promoted by many societies, even
though the social norms relating to what is and what is not considered
polite behavior may vary across cultures (House a Kasper, 1981).
Goffman (1967) explains the importance of politeness in social
interactions in terms of face. Salient in face-to-face interactions is
the social recognition of an individual’s face, i.e., "the positive
social value a person effectively claims for himself by the line others
assume he has taken during a particular contact" (Goffman, 1967, p. 5).
Social actors who have internalized a social norm of considerateness
constantly pay attention to each other’s need for face-support, since
failure to support each other’s face results in mutual embarrassment
and deterioration of their relationship. Since politeness is a means
of satisfying others’ need for face-support, it is crucial for
preserving and developing relationships. Thus, being polite is
considered a very important social value in a civilized society.

Many researchers have demonstrated an interest in the linguistic
phenomenon of politeness (e.g., Brend, 1978; Brown, 1976; Clark a
Schunk, 1980, 1981; Ferguson, 1976; Lakoff, 1973; Leech, 1977).

Despite this extensive effort, however, the literature on politeness

typically lacks any theoretical framework that can systematically and

comprehensively explain how politeness manifests itself in social

interactions. Most studies of politeness have focused on specific

features of politeness such as formulaic expressions (Ferguson, 1964,
1

2
1976; Tannen a Oztek, 1981), acguisition of pglite expressions (Axia &
Baroni, 1985; Bates, 1976; Becker, 1982; Camras, Pristo, a Brown, 1985;
Mitchell-Kernan a Kernan, 1977; Nippold, Leonard, a Anastopoulos,
1982), pplitengss markeps (Brend, 1978; Ervin-Tripp, 1977; Gibbs, 1981,
1986; House a Kasper, 1981; Kemper & Thissen, 1981), gender differences
(Baroni a D’Urso, 1984; Baxter, 1984; Brown, 1976, 1980; Kemper, 1984;
Lakoff, 1975; Shimanoff, 1977), cultural differences (Bowman & Okuda,
1985; Carrell & Konneker, 1981), and pituational differences in
politeness (Baxter, 1984; Becker, 1982; James, 1978; Lakoff, 1972,
Shimanoff, 1977). Even the effort to account for general politeness
phenomena has not progressed much beyond com-on sense platitudes such
as ”don’t impose," "give options," and "be friendly" (Lakoff, 1974).

An exception to this tendency is the work of Brown and Levinson
(1978). In their seminal work in which they analyze polite expressions
in various languages such as English, Tzeltal (a Mayan language), South
Indian Tamil, Malagasy, and Japanese, Brown and Levinson propose a
rational model of politeness that is alleged to be universally
applicable. The alleged universal applicability of Brown and
Levinson’s model of politeness and the fact that it is one of the few
available theoretical approaches to politeness have made their work the
central explanation of politeness phenomena. Consequently, many
researchers (e.g., Baxter, 1984; Craig, Tracy, a Spisak, 1986;
Shimanoff, 1977; Tracy, Craig, Smith, Spisak, 1984) adopted Brown and
Levinson’s model to explain politeness in communicative interactions.
However, these researchers found that the model did not explain

politeness phenomena well, emphasizing the necessity to develop a new

3
theoretical model that can explain politeness more accurately and
comprehensively.
This study will explore Brown and Levinson’s approach for its
adequacy in explaining politeness phenomena. As this exploration will
reveal substantial limitations in the theoretical model, a new model

that eliminates these limitations will be proposed.

BROWN AND LEVINSON’S MODEL OF POLITENESS
Face and Face-Threat

Brown and Levinson (1978) establish a model explaining how
politeness manifests itself in social interactions by extending
Goffman’s (1967) concepts of face and face—saving. Brown and Levinson
first distinguish between two kinds of face: positive and negative
face. Positive face is the want to be thought of as a desirable human
being, while negative face is the want not to be imposed on by others.
In other words, positive face is the desire for approval while negative
face is the desire for autonomy or self-determination. When engaged in
social interactions, social actors are expected to save both positive
and negative face of each other. One’s failure to preserve any of the
other’s face wants will make the other embarrassed, which eventually
prevents one from achieving his/her conversational goals. Thus, people
strive to preserve others’ face wants.

However, many communicative acts are inherently face-threatening,
since those acts (e.g, orders, requests, disagreements, criticisms,
apologies, threats, etc.) by nature run contrary to the face wants of
either a bearer or a speaker. Positive face is threatened when

speakers ignore hearers’ desire to be approved and appreciated. For

4
example, when speakers disagree with or criticize hearers, hearers may
feel that they are disapproved. Negative face is threatened when
speakers violate hearers’ desire to be autonomous. For example, when
speakers order or command hearers to do something, hearers may feel
that they are controlled by other persons.

The amount of face threatened by an FTA (face threatening act) is
relative to the relationship between the social actors. In other
words, the amount of face-threat carried by an FTA is perceived
differently depending on the relationship the speaker has with the
bearer. When the bearer has higher power than the speaker, or when the
hearer is not intimate with the speaker, the speaker may feel that the
amount of face-threat carried by the FTA is higher than the absolute
(or objective) face-threat of the PTA, and vice versa. For example,
asking a person to close the door is less face-threatening when the
person is a co-worker than when the person is a superior. Brown and
Levinson, therefore, propose that the amount of face-threat carried by
a particular speech act in a particular situation (Wx) is a function of
the social distance between the speaker and the bearer (D[S,H]); the
power the hearer has over the speaker (P[H,S]); and the absolute (or
objective) imposition inherent to the speech act (Rx); that is, the
relative (or subjective) imposition, Wx, can be expressed as the

following function:
Wx = D[S,H] + P[H,S] + Rx' (1)

Brown and Levinson suggest that function (1) applies to both kinds

of face-threat, i.e., positive and negative face-threats. Regardless

5
of whether the face threatened is positive or negative, the relative
face-threat (Wx) is computed in the same manner.

Of the three factors determining the relative face-threat (i.e.,
social distance, power difference, and absolute face-threat), the only
one influenced by the kind of face threatened is the absolute
face-threat. Brown and Levinson, therefore, suggest that by filling in
the slot for the absolute face-threat with the ranking of the relevant
face-threat, one can apply the model to both kinds of face-threat.
Brown and Levinson explain how one can fill in the absolute face-threat
(Rx) as follows:

R is a culturally and situationally defined ranking of

impositions by the degree to which they are considered to

interfere with an agent’s wants of self-determination pg of
approval (his negative- and positive-face wants). In general,
there are probably two such scales or ranks that are emically
identifiable for negative-face ETAs: a ranking of impositions in
proportion to the expenditure (a) of services and (b) of
goods....£op £158 a ainst ositive face, the ranking involves an
assessment of the amount of ’pain’ given to H’s (Hearer’s) face,
based on the discrepancy between H’s own desired self-image and

that presented in the FTA. (1978, pp. 82—83)

In this conceptualization of the absolute face-threat (Rx), Brown and
Levinson clearly suggest that Rx of a communicative act should be
filled with either the ranking of positive face-threat 9; that of
negative face-threat, depending on the kind of face threatened by the
act. In other words, when the face-threatened is positive face, Rx is
replaced by the ranking of positive face—threat (i.e., the amount of
pain caused by the FTA); thus, the relative face-threat of positive
FTAs is computed by summing the social distance, the power difference,
and the ranking of positive face-threat. When the face threatened is

negative face, Rx is replaced by the ranking of negative face-threat

(i.e., expenditure of services and goods caused by the FTA); thus, the

6

relative face-threat of negative FTAs is computed by summing the social
distance, the power difference, and the ranking of negative
face-threat.
Selection of Poiiteness Strategy

Given the concepts of face and face-threat, Brown and Levinson
conceptualize politeness as the expression of the intention to mitigate
face-threats carried by certain communicative acts, i.e., efforts to
save the face wants of hearer or of self. Since politeness is to save
others’ face wants, as the threat to others’ face wants increases,
persons are expected to employ devices that show a higher degree of
politeness. Consequently, politeness is defined as the enactment of
behaviors that are referred to as "strategies" by Brown and Levinson.
As multiple politeness strategies are available, Brown and Levinson
propose people select a more "redressive" (or polite) strategy as the
relative face-threat increases. A redressive strategy is one that
"gives face" to the bearer, that is, a strategy indicating that no face
threat is intended or desired (Brown & Levinson, 1978).

Brown and Levinson propose five superstrategies of politeness.
The superstrategies, when presented in order from most redressive to
least redressive, are (1) don’t do the FTA (face-threatening act), (2)
do it off-record, (3) negative politeness, (4) positive politeness, and
(5) do it baldly on-record. Figure 1 provides a schematic diagram of
strategy selection as a function of relative face-threat. As can be
seen in Figure 1, "don’t do the FTA" is the most redressive strategy,
since the speaker gives up performing the FTA. This strategy is
usually employed when Wx (subjective imposition of an act) is extremely

high, that is, the risk of face-loss is great. With the "off—record"

Lesser

Risk /

/
Do the FTA
of

\

face

loss
5. Don’t do

 

Greater

Figure 1.

1. Without redressive action, baldly

/
/
On record
/ \ 2. Positive
\ / Politeness
With redressive action
3. Negative
Politeness
\

4. Off record

the FTA

Diagrammatic Representation of Brown and Levinson model

 

8
strategy, which is employed when Wx is very high, a speaker performs
the act in a vague manner that could be interpreted by a hearer as some
other act. The "baldly on-record" strategy makes no attempt to
acknowledge a hearer’s face wants. This strategy is used when there is
no face-threat due to emergency or other acceptable reasons (e.g.,
military training).

While the above three superstrategies, i.e., "don’t do the FTA,"
"off-record," and "baldly on-record," are used when the face-threat is
abnormally high or low, positive politeness and negative politeness are
employed when the face-threat falls in-between. Positive politeness is
employed when the face-threat is moderately low, and negative
politeness is employed when the face-threat is moderately high. The
main function of a positive politeness strategy is to satisfy hearers’
positive face (i.e, hearers’ desire for approval). Brown and Levinson
argue that since a positive politeness strategy is oriented toward a
hearer’s positive face, it is an approach-based strategy.

"’anoints’ the face of the addressee

Specifically, positive politeness
by indicating that in some respects, 8 (or the speaker) wants H’s (or
the addressee’s) wants" (Brown a Levinson, 1978, p. 75). For example,
by treating an addressee as a member of an in-group, a friend, or a
person whose wants and personality traits are known and liked, a
speaker can show that he/she wants what the addressee wants.

Negative politeness is purported to acknowledge hearers’ negative
face wants (i.e. the desire not to be imposed on). Since a negative
politeness strategy is oriented toward a hearer’s negative face, Brown

and Levinson argue it is an avoidance-based strategy. Specifically, a

negative politeness strategy "consists in assurances that the speaker

9
recognizes and respects the addressee’s negative face wants and will
not (or will only minimally) interfere with the addressee’s freedom of
action" (Brown a Levinson, 1978, p. 75). In short, as tactics listed
in Table 1 show, a positive politeness strategy is purported to
ectively premoge hearers’ desire for approval, whereas a negative
politeness strategy is purported to paseiveiy presepve hearers’ desire
for autonomy.

Brown and Levinson argue that as the relative face-threat
increases, speakers would select a more redressive (or face-saving)
strategy. Since a negative politeness strategy is more redressive than
a positive politeness strategy, a speaker is likely to select a
negative politeness strategy over a positive politeness strategy when
the relative face-threat is fairly high, and a positive politeness
strategy over a negative politeness strategy when the relative
face-threat is low. In other words, regardless of the kind of face
threatened, a speaker will select a politeness strategy that actively
promotes the hearer’s positive face want when the relative face-threat
is low; when the relative face-threat is high, a speaker will select a
politeness strategy that passively preserves the hearer’s negative face
want, again regardless of the kind of face threatened. Compare, for

example, the following two requests:

(1) Relationship: Mutual friend with equal power
Goodness, yep cut your hair!(...) By the way, I came to

borrow some flour. (Brown a Levinson, 1978, p. 108)
(2) Relationship: Mutual stranger; hearer has more power

I just want to ask you if on can d lend me a in it of
paper. (Brown & Levinson, 1978, p. 182)

10

Table 1. Chart of Tactics

Besitive peliteness

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.

Notice, attend to H (his interests, wants, needs, goods)
Exaggerate (interest, approval, sympathy with H)

Intensify interest to H

Use in-group identity markers

Seek agreement

Avoid disagreement

Presuppose/raise/assert common ground

Joke

Assert or presuppose S’s knowledge of or concern for H’s wants
Offer, promise

Be optimistic

Include both 8 and H in the activity

Give (or ask for) reasons

Assume or assert reciprocity

Give gifts to H (goods, sympathy, understanding, cooperation)

Negative peliteness

O

1. Be conventionally indirect

2. Question, hedge

3. Be pessimistic

4. Minimize the imposition

5. Give deference

6. Apologize

7. Impersonalize S and H: Avoid the pronouns ’I’ and 'You’
8. State the FTA as a general rule
9. Nominalize

10. Go on record as incurring a debt, or as not indebting H
ecord

1. Give hints

2. Give association clues

3. Presuppose

4. Understate

5. Overstate

6. Use tautologies

7. Use contradictions

8. Be ironic

9. Use metaphors .

10. Use rhetorical questions

11. Be ambiguous

12. Be vague
13. Over-generalize

14. Displace H

15. Be incomplete, use ellipsis

11
Since request (1) is made to a friend with equal power, the face-threat
relative to the given relationship is low, although the absolute
face-threat carried by an act of requesting flour may not be very low.
Thus, the speaker would employ a positive politeness strategy; that is,
the speaker shows interest in the bearer by noticing that the bearer
had his/her hair cut. In contrast, although the absolute face-threat
is very low in request (2), since the bearer is a stranger who is
higher in power than the speaker, the relative face-threat is high.
Thus, the speaker would select a negative politeness strategy; that is,
the speaker minimizes the imposition by using such phrases as "you
could lend me" and "tiny bit of".
Tests oi Bpewp egg Levinson’e Megel

Since Brown and Levinson proposed their model of politeness, a
number of researchers have adopted the model to explain various
communicative interactions (e.g., Applegate, 1982; Baxter, 1984; Craig,
Tracy, a Spisak, 1986; McLaughlin, Cody, a O’Hair, 1983; McLaughlin,
Cody, & Rosenstein, 1983; Shimanoff, 1977; Tracy, 1983; Tracy et a1.,
1984). However, these researchers found that the model did not explain
politeness phenomena accurately (Baxter, 1984; Craig, Tracy, a Spisak,
1986; Shimanoff, 1977), and that the model was too simplistic to
explain effectively the way people select a politeness strategy
'(Baxter, 1984).

First, Craig et a1. (1986) and Shimanoff (1977) found that
speakers often did not restrict themselves to a single superstrategy;
instead complex combinations of positive and negative politeness
strategies occurred in many conversations analyzed by these

researchers. Brown and Levinson’s framework only permits the mixture

 

12
of superstrategies under two conditions. First, an auxiliary
superstrategy can be employed to soften the tone of a main
superstrategy. For example, hedges (e.g., "like," "sort of") can be
used to render more vague expression of an extreme positive politeness
opinion. Second, a speaker can alternate using positive politeness and
negative politeness, moving back and forth between approaching and
distancing in the interaction. Even in these two cases, however, Brown
and Levinson permit only one main superstrategy at a time that is
selected based on the relative face threat. However, the examples of
mixed strategies found in Craig et al.’s (1986) and Shimanoff’s (1977)
studies were not as simple as Brown and Levinson suggested. Positive
and negative politeness strategies were juxtaposed in the mixture
rather than one used to soften the tone of the other. In other words,
two main superstrategies, especially positive and negative politeness,
co-occurred in many examples, which is contradictory to what Brown and
Levinson delineated.

In addition, Baxter (1984) and Craig et a1. (1986) found
"multifunctionality" of an expression. In other words, different
superstrategies were realized simultaneously in the same language. For
example, "the phrase, 'Would you mind doing a favor for me?’,
acknowledges an imposition while suggesting that a ’favor’ should be
granted in this relationship unless the bearer clearly ‘minds’ doing so
for some reason" (Craig et al., 1986, p. 452). By acknowledging an
imposition, the speaker shows his/her intention to preserve the
hearer’s negative face; and by presuming the bearer will willingly
help, the speaker promotes the hearer’s positive face. These findings

together suggest that Brown and Levinson’s model, which conceptualizes

13
positive and negative politeness strategies as mutually exclusive
categories, does not accurately represent the way people enact behavior
in conversation.

Second, an empirical test of Brown and Levinson’s model suggests
that a large portion of the variance in politeness can be explained by
factors other than the three predictors proposed by Brown and Levinson.
Baxter (1984), using a self-report, likelihood-of-use analysis of
politeness strategies, tested the effects of relational distance, power
disparity, the magnitude of face-threat, and gender on use of
politeness strategies. Baxter found that gender of an actor was as
good a predictor of politeness as relational distance, and a better
predictor of politeness than power disparity and the magnitude of
face-threat. This finding suggests that Brown and Levinson’s model is
too simplistic to explain effectively use of politeness in social
interactions.

The evidence discussed above is not enough to conclude that Brown
and Levinson’s model is misleading, but sufficient to establish that it
has several limitations to overcome. The purpose of this study is to
examine limitations of Brown and Levinson’s model in detail, and
develop a new model that would eliminate these limitations. This study
will first discuss four major conceptual problems that limit the
explanatory power of the model. Specifically, it will be shown that:
(1) Brown and Levinson’s model accounts for only one kind of face want
at a time despite both positive and negative face being supported at
the same time; (2) The model limits the scope of politeness by
conceptualizing it only as a means to perform FTAs effectively; (3) The

model focuses on a few nonrepresentative elements of positive and

14
negative face wants when conceptualizing positive and negative
politeness; and (4) This nonrepresentative conceptualization of
positive and negative politeness leads Brown and Levinson to the
inaccurate claim that negative politeness should be used when the
face-threat is high, and positive politeness should be employed when
the face-threat is low. The paper then will propose a new framework
that will extend Brown and Levinson’s model by solving these four

problems.

CRITIQUE OF BROWN AND LEVINSON’S MODEL AND DIRECTIONS FOR A NEW MODEL

Lipited to One Kind pf Face-Threat et a Iipe

Brown and Levinson’s model is purported to explain only one kind

 

of face-threat, either positive or negative, for a given act. In the
discussion of how to calculate the immediate predictor of politeness,
i.e., the relative face-threat (see equation 1), Brown and Levinson
clearly suggest that their model is intended to account for one kind of
face-threat at a time. Specifically, they propose that the absolute
face-threat (Rx) of an act should be filled with eiehep the ranking of
positive face-threat 9; that of negative face-threat, depending on the
kind of face threatened by the act. When the face threatened is
positive face, the relative face-threat is obtained by summing the
social distance, the power difference, and the ranking of positive
face-threat; and when the face threatened is negative face, the
relative face-threat is computed by summing the social distance, the
power difference, and the ranking of negative face-threat.

This one-face-at-a-time approach of Brown and Levinson’s model is

based on the assumption that one communicative act threatens only one

15
kind of face, either positive or negative face. When they discuss the
nature of face threatening acts, Brown and Levinson make a clear
"distinction between acts that threaten negative face and those that
threaten positive face" (1978, p. 70). Furthermore, Brown and Levinson
present two groups of face threatening acts, one of which consists of
acts threatening only positive face and the other of which consists of
acts threatening only negative face. If we can assume that one act
threatens only one kind of face, we can fully account for the
face-threat carried by an act by explaining either positive or negative
face-threat, depending on the kind of face threatened.

Many acts ppreeten poeitive and negative face a; the eepe pipe.

It is highly improbable that one speech act threatens only one type of

 

face want. In fact, many speech acts threaten both types of face want
at the same time. Consider, for example, the utterance "Can you do it
again, please?" Brown and Levinson would argue that this utterance
threatens only negative face of the bearer, since the illocutionary act
(or speech act) of this utterance is a request that contradicts the
hearer’s desire not to be imposed on. This request, however, clearly
violates the hearer’s desire to be approved by implying that the work
done is not satisfactory; that is, the act also threatens positive face
of the bearer. Also consider the utterance "I don’t think it works
that way." To Brown and Levinson, this utterance should threaten only
'positive face of the bearer, because the illocutionary act of this
utterance is a disagreement that contradicts the hearer’s desire to be
approved. However, this disagreement not only disapproves what the

bearer said but also imposes on the bearer that he/she should adopt

16
what the speaker is about to say. In other words, a disagreement can
threaten both face wants at the same time.

Necessity to eecount fop both pesitive and pegegive threate at
epe_. Since most communicative acts threaten both positive and
negative face at the same time (Craig et al., 1986), the
one-face-at-a-time approach is not adequate to fully account for the
face-threat carried by those acts; that is, it is not realistic for
Brown and Levinson to claim that, depending on the kind of face
threatened by an act, the relative face-threat (Rx) should be
calculated from eithep the ranking of positive face-threat pp that of
negative face-threat of the act. Instead, the relative face-threat of
a communicative act should comprise both the threat to positive face
and the threat to negative face carried by the act.

Given that the relative face-threat should take into account both
positive and negative face-threat, advocates of Brown and Levinson’s
model may suggest that we should calculate the overall amount of face
threat by combining the impositions caused by positive face threat and
those caused by negative face threat, and then fill in the slot for the
absolute face-threat with this total amount of face-threat. This
suggestion, however, is not likely to work. For two different factors
to be combined into a single index, it is required that these two
factors represent the same underlying construct. Positive and negative
face, however, represent two distinct human needs (Craig et al., 1986).
As mentioned earlier, negative face is the desire for
self-determination, while positive face is the desire for approval.
Thus, threats to positive face refer to lack of approval while threats

to negative face refer to interferences with self-determination.

17
Positive and negative face-threat clearly represent two different
constructs. Summing positive and negative face-threat of an act, then,
is as unreasonable as summing weight and height of a person.

In summary, since many communicative acts threaten both positive
and negative face at the same time, we should account for both positive
and negative face-threat carried by a communicative act to explain
politeness properly. However, summing positive and negative
face-threat to make a single index of face-threat is not reasonable.
Since positive and negative face-threat do not represent the same
underlying dimension, we should maintain separate indices for different
kinds of face-threat when calculating the absolute face-threat or the
relative face-threat of an act. Specifically, the absolute threat to
positive and negative face should be measured separately, and the'
relative threat to positive and negative face should be computed
separately from their respective absolute face-threat.

Lieited to Politeness Related te FTAs

Brown and Levinson’s nodel centers around acts that intrinsically
threaten hearers’ face wants, i.e., face threatening acts or FTAs.
Politeness and politeness strategies are conceptualized as necessary
tools to perform these FTAs successfully. Brown and Levinson explain
the mechanism underlying their main proposition that a rational agent
would tend to choose a more redressive superstrategy as the relative
face-threat increases, as follows:

Our MP (or model person) would not do all files with the strategy

of least risk because it costs more in effort and loss of clarity,

because he may wish to satisfy the other perennial desire of H’s

-- for positive face -- but most importantly because choice of the

least risky strategy may indicate to H (or hearer) that the

FTA is more threatening than it actually is, since it would imply
an excessively high rating of P (or power disparity) or D (or

18
relational distance) or R (or absolute face-threat), or some
combination. In short, our original assumptions that define our
MP as a 'rational agent with face’ predict that rational
face-bearing agents will choose wa s of doin ace-t r t nin
eepe that minimize those threats, hence will choose a
higher-numbered (i.e., more redressive) strategy as the threat
increases. (1978, p. 88)
In this explanation, Brown and Levinson clearly show that their
politeness model is primarily for those acts that threaten hearers’
face wants by stating that politeness strategies are ways of doing
face-threatening acts. Of the FTAs, Brown and Levinson are
particularly intrigued by the speech act of a request; thus, when they
need to illustrate their model, they often use requests as examples.
Peiipepess ean be top ecps ether than FTAs. First, politeness
does not seem to be limited to those acts that intrinsically threaten
others’ face, but can be found in many other social interactions.
Suppose, for example, that Person A says only "see you later" to Person
B who has just helped out Person A with something. Person B might
think that what he/she has done for Person A is considered undesirable
by Person A, since Person A did not say "thank you." Or, Person B
might even think that Person A evaluates him/her as an undesirable
person. In any case, Person B will lose his/her positive face and
decide that Person A is an impolite person. For another example,
suppose that your acquaintance ignores you and just passes by you when
you try to greet him/her by saying hello. You might feel that you are
disapproved, and thus will lose your face. In both of the above
examples, no intrinsically face threatening act was involved; the
approval expected by one party simply is not given by the other party.

Many ritualized (or formulaic) exchanges of mutual approval and/or

power—giving (Ferguson, 1964, 1976; Laver, 1981; Tannen a Oztek, 1981)

19
such as greeting, parting, appreciating, apologizing, congratulating,
and other phatic communication, share this characteristic; that is,
these politeness formulas are not to perform some FTAs effectively but
to support or save ritualistically others’ desire for approval or
power. Consequently, politeness is not limited only to the performance
of acts that intrinsically threaten others’ face.

Second, people should be polite not only verbally but nonverbally
(Lakoff, 1973). We are told, for example, to cover our mouth when we
cough and stand aside as someone enters a door we are in front of. Our
parents also taught us that taking an excessively relaxed posture when
talking with superiors is considered impolite. This nonverbal
politeness is observed in order to support or save others’ desire for
approval or autonomy regardless of whether we perform a face
threatening act or not. For example, if we take an extremely relaxed
posture in front of a person who is by far superior to us (regardless
of what we are doing), the person should feel that his/her power is not
properly respected and that he/she lost his/her face; that is, use of
nonverbal behavior that does not provide the other with the expected
level of approval or autonomy causes loss of the other’s face. It
seems to be that, for a given relationship, there is an expected level
of approval and autonomy, and that only a certain range of nonverbal
behavior, which can properly satisfy this expected level of approval or
autonomy, is likely to be used (Goffman, 1967). In short, non-FTAs as
well as FTAs require politeness, and nonverbal behavior as well as

verbal behavior is regulated by politeness considerations.

20

Neeessity to account for the face-eeving obligation. Clearly many

 

politeness phenomena occur not as means to perform an FTA but as means
to support the expected level of approval or autonomy that any
communicator establishes when he/she is engaged in a social
interaction. Thus, politeness conceptualized as an effort to mitigate
face-threat carried by FTAs does not seem to reflect the rich concept
of politeness fully. Moreover, since politeness is not limited to the
occasions in which threats to face are involved, it is not likely that
one can comprehensively explain politeness based on the concept of the
relative face-threat. In order to explain politeness more
comprehensively by accommodating politeness not related to FTAs, we
need to extend the relative face-threat to a more general concept.
Goffman’s (1967) approach provides the vehicle to extend politeness to
all behavior.

Goffman (1967) argued that when engaged in a social interaction,
people establish an expected level of face support from the encounter.
If the encounter sustains this expectation, people would have few
feelings about the matter, since the sustenance has been long taken for
granted. If events establish a face for people that is better than
they might have expected, they are likely to "feel good;" if their
ordinary expectations are not fulfilled, they will "feel bad" or "feel
hurt" (Goffman, 1967). Since people recognize that others also hold
this desire for approval and autonomy, they feel obliged to help others
satisfy their desire. In other words, it is a mutually shared belief
among people that it is their obligation to support each other’s face,
i.e., desire for approval and autonomy. This obligation to save face

exists regardless of whether one performs a face threatening act or

21
not. Simply by having a relationship with another person, people incur
obligations to maintain the person’s desire for approval and autonomy
(Goffman, 1967).

According to Goffman (1967), politeness is fulfillment of one’s
obligations to support the other’s face. When one fails to fulfill
this obligation, he/she is perceived to be inconsiderate, which will
lead the other to find ways to punish him/her. Since people want to
avoid being punished, they tend to conduct themselves during an
interaction so as to fulfill the obligation to save others’ face. In
short, the obligation to save face leads people to be polite. Since
the obligation to support others’ desire for approval and autonomy is
not limited to occasions when intrinsically face threatening acts are
performed, it can explain politeness phenomena more comprehensively.
When no FTA is performed, the degree of politeness to be used is a
function of the obligation to maintain the hearer’s face incurred by
having a relationship with the hearer. For example, when we greet
someone, all that matters in determining the proper level of politeness
are the characteristics of the relationship. If the person has more
power than us, we should maintain his/her desire for autonomy by
showing deference. If the person is a friend, we may want to pay more
attention to his/her desire for approval than his/her desire for
autonomy. When an FTA is performed, the obligation to maintain the
hearer’s face incurred by doing this FTA is added to the obligation to
maintain the hearer’s face incurred by being engaged in the
relationship. The obligation to save face in this case is the same as

the concept of the relative face-threat.

22
In short, viewing politeness as an obligation to save face

regardless of the behavior involved enables us to explain politeness
phenomena more comprehensively. Earlier, it was argued that people
attempt to satisfy both positive and negative face wants of their
partners at the same time. In other words, people have obligations to
save both positive and negative face wants of their partner in the same
message. Thus, in order to account for the politeness of an act
adequately, we need to assess the obligation to save positive face and

that to save negative face incurred by performing the act in the given

relationship.
Lipited te Qpe Side pf Peeipive epg Negetive Polipepeee

Brown and Levinson argue that positive politeness strategies are
those which save a hearer’s positive face and that negative politeness
strategies are those which preserve a hearer’s negative face; that is,
positive politeness should satisfy a hearer’s desire for approval,
whereas negative politeness should satisfy a hearer’s desire for
autonomy. However, in the process of conceptualizing the politeness
strategies in detail, Brown and Levinson define positive politeness as
an approach-based strategy and negative politeness as an
avoidance-based strategy. In other words, positive politeness is a
strategy of "minimizing social distance" and negative politeness is a
strategy of "social distancing" (Brown & Levinson, 1978, p. 135).

’v ' e s o v ' r al . It is
generally acknowledged that people use informal language when they want
to decrease the social distance between them and the people they

encounter (Scotton, 1983). Since positive politeness is intended to

decrease social distance, it is characterized by informality. When

23
conceptualizing positive politeness in detail, Brown and Levinson argue
that positive politeness is realized in forms of informal language as
follows:
...the linguistic realizations of positive politeness are in many

respects simply representative of the normel lingpietic behavior
between intimates, where interest and approval of each other’s

personality, presuppositions indicating shared wants and shared
knowledge, implicit claims to reciprocity of obligations or to
reflexivity of wants, etc. are routinely exchanged. Perhaps the
only feature that distinguishes positive politeness redress from
normal everyday intimate language behavior is an element of
exaggeration... (1978, p. 106)
In this description of positive politeness, Brown and Levinson argue
that positive politeness is expressed in the pepeei linguistic behavior
between intimates. When we take into account that normal or unmarked
linguistic behavior between intimates is informal language (Scotton,
1983), Brown and Levinson suggest that positive politeness is realized
in informal language. The specific tactics of positive politeness that
are proposed by Brown and Levinson reflect this identification of
positive politeness with informality. Under the superstrategy of
positive politeness, Brown and Levinson propose three major strategies:
(1) claim 'common ground’, (2) convey that the speaker and the bearer
are cooperators, and (3) fulfill the hearer’s want by giving goods,
sympathy, understanding, and cooperation. These strategies all suggest
that a speaker should act informally toward a bearer to be positively
polite.
It is certain that some intimate linguistic behavior expresses
approval of the partner. However, it is not true that ell kinds of
informal linguistic behavior show approval of the partner, or that epiy

informal linguistic behavior conveys approval of the partner. First,

many types of informal language suggested by Brown and Levinson as

24
positive politeness tactics do not necessarily approve a hearer’s
performances or possessions, which a genuine positive politeness
strategy is supposed to do. For example, Brown and Levinson propose
’making jokes’ and ’being optimistic’ as positive politeness tactics.
These tactics may make the speaker appear friendly; however, they do
not always approve a hearer. Suppose, for illustration, that Person A
makes a joke when conversing with Person B. It is likely that Person B
thinks that Person A is a friendly person, but it is highly unlikely
that Person B believes Person A approves any of his/her possessions or
performances. Brown and Levinson also propose that use of in-group
identity markers (such as mate, pal, buddy, honey, and Johnny) are a
positive politeness tactic. Use of these identity markers may affirm
that the speaker wants to be on friendly terms with the hearer;
however, it does not necessarily entail that the bearer is evaluated as
a valuable person by the speaker.

Second, and more importantly, there are many other ways to save
others’ positive face. We can directly approve others’ possessions and
performances by using formal language. For example, a student can give
a compliment to one of his/her professors at the end of a term by
saying "Professor Brown, I enjoyed your class very much." In this
compliment, the student approves the professor’s performance without
using informal language. As a matter of fact, approval has no
relationship with formality; people can compliment, praise, or
appreciate others using formal language as well as informal language.
Furthermore, we can satisfy others’ positive face using avoidance-based
tactics; that is, we can preserve others’ desire for approval by

showing that we do not have any intention to disapprove them. Suppose

 

25
that we want to disagree with others very politely. Then, we may not
want to outright disagree with them; instead, we will use all kinds of
devices to convey our intention that we do not want to attack their
self-concept. Thus, we may want to be tentative, saying "I am not very
sure, but I don’t think I quite agree with you" or "Isn’t there any
chance that ....". We may also understate the problem, saying "I think
slightly differently." Consequently, formality and approval have no
relationship with each other, making Brown and Levinson’s definition of
positive politeness a problem.

Negative peliteness is pot always avoidance-beeed. It seems to be
true that, at least in Western society, most negative politeness
strategies perform "the function of minimizing the particular
imposition that the FTA unavoidably effects" (Brown & Levinson, 1978,
p. 134). However, not all negative politeness strategies are purported
to prevent leee of others’ negative face. Sometimes, people use
negative politeness to actively peeeete others’ negative face. For
example, giving deference is a way to promote others’ desire for
autonomy, since deference conveys that a bearer is of higher social
status than a speaker.

As a matter of fact, Brown and Levinson propose giving deference
as a negative politeness tactic; however, they do not think that giving
deference actively promotes hearers’ negative face, but argue that
"deference serves to defuse potential face-threatening acts by
indicating that the addressee’s rights to relative immunity from
impositions are recognized -- and moreover that the speaker is not in a
position to coerce the addressee’s compliance in any way" (1978, p.

183). In other words, Brown and Levinson conceptualize deference as a

26
way to prevent loss of hearers’ negative face. This conceptualization
may be adequate if deference occurs only when an FTA is performed.
However, people often give deference to superiors when they do not
perform an FTA. For example, when people greet, talk with, or part
their superiors, they may give deference using the "vous" form pronouns
(Brown & Gilman, 1960) such as title plus last name (e.g., Professor
Brown), sir, madam, or lady, simply to show that they recognize the
superiors have more power. In this case, the function of this
deference is not to protect the superiors from the threat to negative
face caused by an FTA, but to promote actively the superiors’ desire
for peyep.

The main reason Brown and Levinson do not conceptualize deference
as a strategy that actively promotes others’ negative face seems to be
that whereas negative face is the desire for eptepepy, deference
promotes the desire for peye_. Logically, however, the desire for
autonomy is a part of the desire for power; that is, autonomy
represents the middle of the continuum of power, the extremes of which
represent dominance and submission. Submission is the state in which
one’s behavior is interfered with by the other due to lack of power.
Autonomy refers to the state in which one is not controlled by the
other because both persons are equally powerful. Dominance is the
state in which one interferes with the other’s behavior since he/she
has more power than the other. When conceptualized more generally,
therefore, negative face is the desire for power. Given this
conceptualization, since deference promotes hearers’ desire for power,

giving deference promotes hearers’ negative face.

27
Necessity to recopeeptualize pesitive epdenegetive politeness. In

summary, Brown and Levinson conceptualize positive politeness as

 

linguistic behavior promoting informality between communicators, and
negative politeness as linguistic behavior protecting hearers’ freedom
of action, which does not fully account for their earlier
conceptualization of face wants and politeness. Positive face want
refers to the desire that one’s possessions and performances are
approved, while negative face want, when conceptualized more generally,
refers to the desire that one has enough power to determine his/her own
destiny. In other words, positive face is the want for approval and
negative face is the want for power. Brown and Levinson define
politeness as an effort to promote (or support) or protect (or save)
these two face wants. Thus, positive politeness should refer to the
behavior that promotes or protects hearers’ wants for approval and
negative politeness should refer to the behavior that promotes or
protects hearers’ wants for power.
Propesition Based op inconsistent Defipitiope

As discussed earlier, Brown and Levinson propose that when the
relative face-threat is high, speakers would choose negative politeness
strategies over positive politeness strategies; and when the relative
face-threat is low, speakers would select positive politeness
strategies over negative politeness strategies. Notice that this
proposition is not influenced by the kind of face threatened. Speakers
would use strategies satisfying hearers’ desire for approval when there
is a small amount of negative face-threat as well as when there is a
small amount of positive face-threat, and speakers would employ

strategies satisfying hearers’ desire for autonomy when there is a

28
large amount of positive face-threat as well as when there is a large
amount of negative face-threat. In other words, the proposition
implies that satisfying positive face nullifies a small threat to
negative face, and satisfying negative face nullifies a large threat to
positive face. However, it is highly unlikely that satisfying one face
want nullifies threats to the other face want.

No compepsation pechanism between peeitive and negative tees.
Although they do not explain how a large threat to positive face is
nullified by satisfying negative face, Brown and Levinson explain how a
small threat to negative face is nullified by satisfying positive face.
As mentioned earlier, the positive politeness tactics proposed by Brown
and Levinson are actually devices to make a relationship informal;
thus, the tactics contribute to decreasing social distance. Because
decreases in relational distance lead to decreases in the relative
threat to negative face, the previously existing low degree of negative
threat will be decreased to the level of nonexistence when the
relational distance is decreased by the speaker’s use of positive
politeness tactics defined as informality. Thus, the speaker has no
more obligation to save the hearer’s negative face, and will use only
positive politeness strategies in his/her request.

However, as argued earlier, defining positive politeness as
informal language is misleading. When we define positive politeness
considering that it involves formal approval as well as informal
approval, and passive avoidance of disapproval as well as active
approval, it is not likely that positive politeness decreases social
distance. Thus, use of positive politeness is not expected to nullify

a low level of negative face-threat.

29

It is more obvious that satisfying one kind of face does not
nullify threats to the other kind of face when we consider the case in
which a speaker performs a communicative act threatening a hearer’s
positive face to a large extent. Suppose, for example, that a student
wants to challenge what a professor says in class. Since the bearer
(i.e., the professor) has more power than the speaker (i.e., the
student) and disagreement as an act carries a relatively high degree of
positive face threat, the challenging behavior in the given setting
should reveal a very high level of positive face-threat. Thus,
following Brown and Levinson, the student would use a negative
politeness strategy, saying, "I am terribly sorry (apology). but I just
want to ask you (minimize the imposition) if you would allow (make
minimal assumptions) me to say that what you are saying is really'
stupid." Notice, in this example, that the student does not make any
effort to protect the professor’s positive face, since Brown and
Levinson propose negative politeness satisfying only negative face
should be used when the face-threat is high. Although the student here
tries very hard to be negatively polite, the professor must feel that
he/she is disapproved by the student. The reason why the professor
feels disapproved is that positive and negative face are distinct needs
(Craig et al., 1986), and peth of the face wants ppet be satisfied to
avoid embarrassment during social interactions (Brown & Levinson, 1978;
Craig et al., 1986). Thus, even if one face want is fully satisfied,
if the other face want is violated, a hearer is likely to be

embarrassed.

30

Neceesity te gave the fees being threatened. Since saving one
face want does not nullify the threat to the other, the threat to a
given face want should be mitigated by redressing the given face want.
Specifically, once there exists any threat to positive face, speakers
should employ tactics that show they do not intend to disapprove
hearers; and once there exists any threat to negative face, speakers
should use tactics that show they do not intend to interfere with
hearers’ freedom of action.

Earlier, it was shown that many communicative acts threaten both
positive and negative face at the same time. Given that an act
threatens both positive and negative face and that threats to a given
face want can be mitigated only by redressing the given face want, we
can expect speakers would try to preserve both positive and negative
face in the same message. While conducting a study to test Brown and
Levinson’s model, Shimanoff (1977) and Craig et al. (1986) found that
politeness tactics employed by communicators are not restricted to a
single superstrategy, but involve complex combinations of tactics from
positive and negative politeness strategies. In other words, in a
single message, speakers pay attention to both positive and negative
face (Craig et al., 1986). Thus, we need to look at both kinds of
face-saving efforts, i.e., positive and negative politeness, as well as
account for both kinds of face-saving obligations.

nc usion Cr't' ue 0 Br and via ’ de

Brown and Levinson’s model of politeness reveals four major
limitations that should be addressed to explain politeness behavior
more accurately. First, the model limits its explanation to one kind

of face at a time, based on the assumption that one act threatens only

31
one kind of face. Since social interactions threaten both positive and
negative face, a new model should account for threats to both kinds of
face. Second, Brown and Levinson’s model centers around those acts
which threaten hearers’ face wants, proposing relative face-threat as
the immediate predictor of politeness output. Other researchers (e.g.,
Ferguson, 1976; Lakoff, 1973) have shown that politeness behavior can
be observed when no FTA is involved, urging us to use a more
comprehensive predictor. Goffman (1967) suggests that a social
relationship is characterized by the obligation to save each other’s
face that exists whether an FTA is performed or not. Since the
obligation to save face subsumes the concept of relative face-threat as
well as accounts for non-FTA related obligations, a new model should
employ the concept of face-saving obligation to explain politeness
output.

Third, Brown and Levinson limit positive politeness to informality
devices (or approach-based tactics) and negative politeness to
formality devices (or avoidance-based tactics). However, positive face
can be supported by avoidance-based tactics and negative face can be
supported by approach-based tactics. In order to conceptualize
politeness more comprehensively, therefore, we need to redefine
positive and negative politeness considering various aspects of
positive and negative face wants and various ways to satisfy these face
wants. Fourth, since Brown and Levinson conceptualize positive
politeness as informality and negative politeness as formality, they
propose that satisfying one kind of face nullifies the threat to the
other kind of face. However, when we define positive and negative

politeness considering various aspects of positive and negative face, a

32
threat to a given face want can be mitigated only by redressing the
given face want. Thus, social actors need to fulfill their obligation
to save both positive and negative face of each other by being both
positively and negatively polite. A new model, therefore, should
explain how the obligations to save positive and negative face
influence the level of positive and negative politeness, respectively.
Given the limitations of Brown and Levinson’s model and directions for

extending these limitations, the following model is proposed.

A NEW MODEL OF POLITENESS

This section will identify and define basic elements of politeness
behavior and specify relationships between these elements. Certain
distinctions identified by Brown and Levinson that are useful and have
found supporting evidence will be included in the new model, and the
four major problems of Brown and Levinson’s model identified and
explained in the critique will be resolved.
Eece end Fece-Seving Obligation

Face is defined as the positive social value persons claim for
themselves. Simply speaking, therefore, face is the same concept as
self-respect. Face or self-respect has at least two different
dimensions: positive and negative face (Brown a Levinson, 1978).
Positive face refers to the want that one is considered to be desirable
by others, that is, desire for approval. The most important aspects of
self that should be approved during social interactions are one’s
possessions and performances (Brown & Levinson, 1978). Thus, positive
face is defined as a desire that one’s possessions and performances are

approved by others. The performances here include everything one

33
publicly performs or achieves. The possessions include personal
traits, physical appearance, personal artifacts, personal associations
(i.e., significant others), knowledge, ideas and opinions, and goals
and intentions; that is, the possessions refer to literally everything
one publicly possesses. Possessions and performances that are not
known to others do not need to be approved, since face is publicly
claimed self-image. In short, positive face is the desire that one’s
possessions and performances be approved, or at least not be
disapproved.

Negative face is defined as the desire for power. Brown and
Levinson (1978) define negative face as the want that one’s actions be
unimpeded by others; that is, it is the desire for autonomy or
self-determination. As mentioned earlier, however, the desire for
autonomy is a part of the desire for power; that is, the desire for
autonomy, or the desire not to be controlled by others, is the desire
to be equally powerful. When defined more generally, therefore,
negative face is the desire for power. Power here does not refer to
absolute social rank or social power, but refers to relational power,
i.e., ability to control the partner in a given relationship. Thus,
negative face is the desire to control the relational partner, or at
least not to be controlled by the partner.

When engaged in social interaction, people are expected to
preserve others’ face and expect others to save their face (Goffman,
1967). In other words, it is a mutually shared belief among people
that it is their obligation to support each other’s desire for approval
and power. Thus, simply by having a relationship with another person,

people incur obligations to maintain the person’s desire for approval

34
and power (Goffman, 1967). Since obligations to maintain the other’s
face originate from the relationship one is having with the other, the
degree to which one is obliged to maintain the other’s face depends on
the characteristics of the relationship.

Burgoon and Hale (1984), based on a synthesis of diverse bodies of
literature including anthropological and psychotherapeutic analyses of
human behavior, measurement of meaning, emotional expression,
interpersonal behavior, relational development, and dyadic

interactions, propose seven dimensions of relationgl_communication:

 

pggegfidisparity (or dominance), relational distance (or intimacy),
_si!ilarity, emotional arousal, composure, formality of the_situation,
and task:§ggieleorientationeof,thescontext. Of these seven dimensions,
however, only the first three dimensions (i.e., power disparity,
relational distance, and similarity) are properties of a relationship;
the other four dimensions (i.e., emotional arousal, composure,
formality, and task-social orientation) represent either an emotional
state of a source or a characteristic of a setting. In short, power
disparity, relational distance, and similarity are the most commonly
explored dimensions of relationships.

Similarity, defined as the degree to which people share like
attitudes, beliefs, personal characteristics, experiences, and so
forth, plays an important role at an early stage of a relationship;
that is, similarity promotes more communication and escalates
relationships to more intimate levels (Burgoon a Hale, 1984).
Similarity, however, becomes less important, as a relationship becomes
more established. In established relationships, similarity is absorbed

into relational intimacy; that is, similarity is considered as a

 

35
subtheme of relational intimacy (Burgoon, Pfau, Parrott, Birk, Coker, &
Burgoon, 1987). Many researchers (Brown a Ford, 1961; Brown a Gilman,
1960; Brown & Levinson, 1978; Ervin-Tripp, 1972), therefore, argue that
power disparity and relational distance are two of the most important
factors representing a relationship. In short, the extent to which a
speaker is obliged to save the partner’s face is expected to be
influenced by power disparity and’relgtignelwdistanee.

When a hearer has more relational power, a speaker is expected to
show more considerateness for both positive and negative face of the
bearer (Brown a Levinson, 1978; Goffman, 1967); that is, one’s
obligation to sustain the other’s desire for approval and power
increases as the power the other has over one increases. A hearer with
more relational power has higher expectations for approval and respect,
and has more capabilities to punish a speaker in instances when the
speaker does not fulfill his/her obligations. Thus, a speaker is more
obliged to satisfy positive and negative face of a bearer who has more
power than him/her.

While for both positive and negative face there is a positive
relationship between the relational power a bearer has over a speaker
and a speaker’s obligation to save a hearer’s face, the relationship
between relational distance and the obligation to save face is
different for positive face and negative face. For negative face, it
is expected that as relational distance decreases, the less obligation
a speaker has to save a hearer’s desire for power (Brown a Levinson,
1978). As a relationship becomes closer, partners become increasingly
interdependent; thus, they become more aware of the necessity for

mutual assistance (Kelley et al., 1983). In other words, as a

36
relationship becomes closer, partners grant each other more rights to
ask for assistance; thus, they tend to lower the desire not to be
controlled by the other. In short, in a closer relationship, partners
have less obligation to preserve each other’s negative face.

As for positive face, Brown and Levinson (1978), without offering
any rationale, predict a positive relationship between the relational
distance and the obligation to save positive face. However, literature
on interpersonal behavior suggests the opposite relationship between
these two factors; that is, as_a relationship becomes closer, partners
are expected to have more obligations to maintain each other’s positive
face. Kelley (1983) argues that people expect more rewards from a
close relationship than a distant relationship, since they invest more
efforts in a closer relationship. When we consider that one major way
to be rewarded in a social interaction is to be approved by others
(Huston, 1983), it is likely that people expect more approval from
those who are closer to them. Thus, as relational distance decreases,
the obligation to support each other’s desire for approval increases.

In summary, the obligation to save a hearer’s positive face or
desire for approval is positively related to the power a hearer has
over a speaker and negatively related to the social distance between a
bearer and a speaker. In other words, as the power a hearer has over a
speaker increases and as the relational distance decreases, a speaker’s
obligation to save a hearer’s positive face increases. The obligation
”EQHBQVC a hearer’s negative face or desire for power is positively
related to both the power a bearer has over a speaker and the social
distance. Thus, as the power a bearer has over a speaker increases and

as the relational distance increases, a speaker’s obligation to save a

37
hearer’s negative face increases. In short, the obligation to save a
hearer’s face that is incurred by having a relationship with the bearer

can be expressed as the following vector function:
RO
[D]
RD
n

where ROI) and R0n are the relational obligation to save the hearer’s

(2)

ll

'0
F—‘I
>4 9
w

+

U
r——-v
I
'4 U‘
L-—J

+
f—‘1
N O
I..___l

positive and negative face, respectively; P is the relational power the
hearer has over the speaker; D is the relational distance between the
hearer and the speaker; and a, b, c, x, y, and z are positive
constants. The relational obligation to save face is always greater
than or equal to zero; that is, no relationship makes people obliged to
be impolite. The power disparity can be either positive or negative;
when the speaker has more power than the hearer, P is smaller than
zero; when the bearer has more power than the speaker, P is greater
than zero; when both participants have equal power, P is zero. The
relational distance by definition cannot be negative. D is zero or
nearly zero when the relationship is extremely intimate. Constants a
and b represent the culturally defined importance of P and D
respectively in determining the relational obligation to save positive
face; constants x and y represent the culturally defined importance of
P and D respectively in determining the relational obligation to save
negative face. Constant c represents the level of R0p when P and D are
zero; that is, c is the level of positive-face saving obligation when
the speaker and the hearer have equal power and the relationship is

extremely intimate (e.g., between best friends or married couples).

38
Constant z represents the level of ROn when P and D are zero; that is,
z is the level of negative-face saving obligation when the speaker and
the bearer have equal power and the relationship is extremely intimate.

Face-Threat and Face-Saving Obligation

As Brown and Levinson (1978) argue, many cemmunicative acts are
inherently face-threatening, since those acts by nature run contrary to
the desire for approval and power. Thus, when performing an FTA, a
speaker is obliged to mitigate the face-threat caused by the FTA as
well as fulfill the face-saving obligation incurred by having a
relationship. Specifically, as the face-threat increases, peeple would
feel more obliged to mitigate the face-threat.

Many communicative acts threaten both positive and negative face
of a bearer, even though they differ in the extent to which they
threaten each face. Some acts threaten one kind of face more than the
other kind of face and some acts threaten both kinds of face equally.
For example, disagreements threaten hearers’ positive face highly by
disapproving what hearers say, and threaten hearers’ negative face
mildly by indirectly imposing speakers’ opinions on hearers. In
contrast, suggestions and advice threaten hearers’ negative face highly
by revealing speakers’ intention to control hearers, and threaten
hearers’ positive face mildly by indirectly disapproving hearers.
\Threats, warnings, and asking for corrections threaten both positive
and negative face very highly by directly disapproving hearers and
overtly diminishing hearers’ power.

Since many communicative acts threaten both positive and negative
face at the same time, when performing these acts, speakers are obliged

to mitigate the threat to both kinds of face in addition to fulfilling

39
the obligation to preserve both kinds of face incurred by having a
relationship. In other words, the total amount of obligation to save
positive and negative face in performing an FTA should be a sum of the
relational obligation to save positive and negative face and the
obligation to save positive and negative face caused by performing the
FTA. Considering that the obligation to save a face want is
proportional to the extent to which the face want is threatened (Brown
a Levinson, 1978; Goffman, 1967), the total face-saving obligation in
performing an FTA is then a sum of the relational face-saving
obligation and the amount of face-threat.

The face-threat here is a set of situationally as well as
culturally defined rankings of imposition (negative face-threat) and
disapproval (positive face-threat). In any culture there is a
generally agreed ranking of impositions and disapproval for a given act
(Brown & Levinson, 1978). For example, in the United States, everyone
knows that asking others to lend ten dollars is more imposing than
asking them to pass the salt. These intra-culturally defined rankings
of face-threat, however, should be modified in some situations.
Particularly, these rankings are influenced by the role relationship a
speaker has with a bearer. When performing the FTA is the speaker’s
right due to the role relationship, the FTA will have a much lower
level of face-threat than the normally expected face-threat (Brown a
Levinson, 1978; Craig et al., 1986). For example, asking others to
type a 30-page report may be highly face-threatening (at least in terms
of negative face). However, a professor asking his/her secretary to

type a 30-page report may not be highly face-threatening, because it is

40
his/her role right. In short, the total obligation to save positive
(Op) and negative face (Op) in performing an FTA can be expressed as

the following vector function:

RO T
l p ] + l p ] (3)
R0 T
n

where ROI) and Ron are the relational obligations to save positive and

r-—-|

O O
'6

L—J
ll

negative face, respectively; TD and Tn are the amount of positive
face-threat and the amount of negative face-threat, respectively, that
are relative to the given role relationship. 0p and On are always
greater than or equal to ROp and Ron, respectively, since TD and TD
range from zero to infinity. When an act is not intrinsically
face-threatening, both TD and Tn are zero and the total obligations to
save positive and negative face are equal to the relational obligations
to save positive and negative face.
Politeness and Face-Saving Obligation

Just as people are expected to have desire for approval and power,
so also they are expected to sustain a standard of considerateness
(Goffman, 1967). They are expected to be considerate enough to fulfill
the obligation to promote and preserve others’ desire for approval and
power willingly and spontaneously. Politeness is this spontaneous
fulfillment of one’s obligation to promote or preserve the other’s face
(Goffman, 1967; Brown A Levinson, 1978). The obligation to promote or
preserve positive face, or the desire for approval, is fulfilled by
being positively polite; and the obligation to promote or preserve
negative face, or the desire for power, is realized by being negatively

polite. Thus, the more persons are obliged to promote or preserve

 

41
others’ face, the more polite expressions they would employ; that is,
the degree of being positively and negatively polite is a function of
the obligation to save positive and negative face, respectively.

Since the fulfillment of the face-saving obligation is based on
people’s considerateness, politeness is also expected to be influenced
by a speaker’s considerateness. When speakers are highly considerate,
they will fully realize their obligation to save face, becoming highly
polite; however, when speakers are highly inconsiderate, they will not
fulfill their face-saving obligation at all, becoming impolite. Many
factors can affect awspeaker’s considerateness and politeness. Most
importantly, speakers’ personal traits may affect their
considerateness. Especially, speakers’ verbal aggressiveness is
expected to be negatively related to their considerateness, since
verbal aggressiveness represents the tendency to attack others’
self-concept (Infante a Wigley, 1986), and considerateness represents
the tendency to support gthers’_selffconcept (Goffman, 1967). Thus,
more verbally aggressive persons are expected to be less polite, since
they have a stronger tendency to ignore their obligation to save
others’ face. Empathy, which is the ability to perceive from the
standpoint of the other (Brown a Keller, 1973), seems to be positively
related to considerateness, since both considerateness and empathy take
the standpoint of the other into consideration. Thus, pore empathetic
persons are expected to be more polite, since they have a stronger
tendency to fulfill their obligation to save others’ face. In short,
politeness is influenced by personal traits such as verbal

aggressiveness and empathy as well as by the obligation to save face.

42
Specifically, the degree of positive and negative politeness can be
expressed as a function of the obligation to save positive and negative

face, respectively, and verbal aggressiveness and empathy as follows:

[Z’l= [2"] - A[:] + all]

where PD and Pn are the level of positive politeness and negative
politeness, respectively; 0p and On are the total obligation to save
positive and negative face, respectively; A represents verbal
aggressiveness; E represents empathy; and, m, n, o, and p are positive
constants. A and E here have defined range of -1 to +1. PD and PD can
be negative. When A is very high (near to +1) and E is very low (near
to -1) and when 0p and On are relatively low, PD and Pn are expected to
be negative. In other words, speakers with high verbal aggressiveness
and low empathy would use impolite expressions when the obligation to
save face is low. Constants m and n represent the importance of A and E
respectively in determining the level of positive politeness; constants
c and p represent the importance of A and E respectively in determining
the level of negative politeness.
Politenese of a Message

Politeness of a message is defined as the extent to which the
message promotes or preserves a receiver’s face (Goffman, 1967).
Specifically, positive politeness is the extent to which a message
promotes or preserves a receiver’s desire for approval, while negative
politeness is the extent to which a message promotes or preserves a

receiver’s desire for power (Brown a Levinson, 1978).

43

To date, many researchers (e.g., Brown a Levinson, 1978; Clark &
Schunk, 1980, 1981; Ervin-Tripp, 1977; Gibbs, 1981, 1986; Kemper &
Thissen, 1981) have proposed schemes for coding or quantifying
politeness. However, none of these schemes are systematic and
comprehensive enough to enable us to make consistent judgments on the
degree of positive and negative politeness of a message. For example,
the politeness strategy system proposed by Brown and Levinson, which is
one of the most comprehensive systems proposed to date, simply lists
output tactics which do not vary in the extent to which they support
positive or negative face (see Table 1). In other words, the tactics
listed under Brown and Levinson’s positive and negative superstrategies
are not systematically linked to an underlying continuum of politeness;
thus, they do not form any scale on which we can quantify positive and
negative politeness (Craig et al., 1986). Furthermore, as discussed
earlier, the tactics represent only a small portion of positive and
negative politeness. Specifically, tactics for positive politeness are
concerned only with promoting others’ desire for approval, while those
for negative politeness are concerned only with avoiding interference
with others’ desire for power. A better scheme of quantification
should include standards that can differentiate output tactics in terms
of the degree of politeness, and that consider both promotion and
preservation of face for each type of politeness.

People’s politeness behavior seems to be governed by such
principles as "if possible, perform only those acts that promote
others’ face" (Brown a Levinson, 1978; Goffman, 1967; Lakoff, 1972) and
"if it is necessary to perform those acts that threaten others’ face,

then be indirect" (Brown a Levinson, 1978; Ervin-Tripp, 1977; Lakoff,

44
1972). To be positively polite, therefore, speakers directly approve
hearers using such tactics as compliments, praises, and admirations; or
when disapproval is necessary, they disapprove hearers as indirectly as
possible. Not all tactics that approve others’ possessions and
performances have the same degree of positive politeness; rather, the
degree of positive politeness varies depending on he! giyeetly a tactic
approves a bearer. The most important standards for the directness of
approval are exaggeration and confidence (Brown & Levinson, 1978).
When an evaluation is positive, the more confident and/or the more
exaggerated the evaluation is, the more approval it shows to a bearer.
For example, "You sure did an excellent job" is more approving than
"You did a good job, I guess," since the former is more confident and
more exaggerating.

While approval is more direct when it is expressed in a confident
and exaggerated manner, disapproval is more indirect when it is
expressed in a tentative and understated manner; that is, tentativeness
and understatement are two of the most important criteria for
indirectness of disapproval. When an evaluation is negative, the more
tentative and/or the more understated the evaluation is, the less
disapproval it shows to a bearer. For example, "I’m not sure, but I
guess it isn’t so bad" is less disapproving than "It is terrible,"
since the former is more tentative (or hesitating) and understated.

To be negatively polite, people directly give power to hearers; or
when interference (or taking power from hearers) is necessary, they
interfere with others’ freedom of action as indirectly as possible.

One main way to give power to others is to use honorifics (Brown and

Levinson, 1978). However, not all honorifics give the same level of

45
power to hearers. The level of negative politeness (or power-giving)
carried by an honorific is determined by the level of deference and
self~abasement associated with the honorific. In many Asian societies,
people systematically vary the amount of deference and self-abasement
by using different sets of honorifics; that is, there are several sets
of honorifics that differ in the degree of deference, and speakers
choose one that is the most appropriate to the intended level of
deference and self-abasement (Magier, 1984). In these cultures, the
more deference the honorific set one is using shows, the more power the
other is given. In Western society, especially in the United States,
the most commonly used honorific system is the pronoun system.
Specifically, people use the "vous" form pronouns such as sir, madam,
lady, Mr., Mrs., and Dr., when they need to show deference; and they
use the "tu" form pronouns such as a first name when they do not need
to show deference (Brown a Gilman, 1960). Thus, the "vous" form
pronouns express more deference and give more power than "tu" form
pronouns.

When exerting control over hearers is inevitable, speakers "give
options" and do "not impose" in order to show that it is not their
intention to interfere with hearers’ freedom of action. A way to give
options is to be conventionally indirect, asking questions or
permissions instead of requesting or ordering (Brown a Levinson, 1978).
For example, "May I have the salt?" or "Would you mind passing the
salt?" gives options while "Pass the salt, please" does not. In order
not to impose, speakers tend to be tentative or hesitant (Brown &
Levinson, 1978); that is, they show that they know they are not

supposed to exert influence. For example, "I was wondering if you can

46
by any chance lend me your notes" lowers the imposition by being highly
hesitant. In other words, two of the most important criteria for
indirectness of interference are conventional indirectness and
hesitance.

In short, literature suggests that the degree of positive
politeness can be determined by confidence and exaggeration shown in
approval, and tentativeness and understatement shown in disapproval.
The degree of negative politeness depends on the degree of deference
shown in use of pronoun, and conventional indirectness and hesitance of
imposition.

In summary, this paper proposes a model of politeness that
explains both positive and negative politeness manifested in the same
message. Positive and negative politeness expressed in a message are a
function of speakers’ verbal aggressiveness, empathy and the
obligations to save positive and negative face, respectively. The
obligations to save positive and negative face are a sum of the
relational obligations to save positive and negative face,
respectively, and the threats to positive and negative face,
respectively. The relational obligation to save positive face is
positively related to the power the bearer has over the speaker, and
negatively related to the relational distance between the speaker and
the hearer; the relational obligation to save negative face is
positively related to both the power disparity and the relational
distance.

x lanation of rev ous Fi 'n s
The present model can explain many empirical findings that Brown

and Levinson’s model has problems with. First, Craig et al. (1986) and

47
Shimanoff (1977) found that speakers used complex combinations of
positive and negative politeness strategies in requesting messages, and
Craig et al. (1986) and Baxter (1984) found that different
superstrategies (mainly positive and negative politeness) were realized
simultaneously in the same language. In other words, positive and
negative politeness strategies co-occurred in the same message (Craig
et al, 1986). Brown and Levinson’s model, which accounts for only one
kind of face-threat at a time by arguing that positive and negative
politeness are mutually exclusive, has difficulties in explaining these
findings.

The present model, which argues people support or preserve both
positive and negative face in the same message, can explain these
findings without any problem. Consider, for illustration, that you
want to request a friend of yours to let you stay at his/her apartment
until you can get your own place to live (Craig et al., 1986). Since
the person you are asking a favor of is your friend, the social
distance and the power disparity are both low; that is, your obligation
to promote the friend’s positive face is high, and to support negative
face is low. Furthermore, since requesting a friend to let you stay in
his/her apartment for a couple of days threatens the friend’s negative
face highly, the obligation to save negative face increases very much.
Overall, you may be highly obliged to support both positive and
negative face of the friend. Since only positive politeness can
fulfill the obligation to save positive face and only negative
politeness can realize the obligation to save negative face, your
message will be characterized by combinations of positive and negative

politeness.

48

Second, Ferguson (1976), analyzing ethnographic data, found that
people use ritualized verbal and nonverbal politeness formulas to
support each other’s face, and that use of politeness formulas varies
in correlation with relational distance and power disparity. In other
words, even in phatic communication where no face-threatening act is
involved, people vary the degree of politeness in accordance with
relational distance and power disparity. Since Brown and Levinson’s
model conceptualizes politeness as a means to mitigate face-threat
carried by FTAs, it cannot explain politeness manifested in greetings,
thank yous, wishes for health, condolences, topical blessings, etc.

The present model, however, can explain varying degrees of
politeness expressed in phatic communication. Since the relational
obligation to save positive face is high among intimates, the model
predicts that people would be positively polite to friends more than to
strangers; thus, they would tend to give more compliments and make
warmer comments to those who are closer to them. Since the obligation
to save negative face is high when people greet or part superiors, they
should be negatively polite to superiors more than to subordinates;
thus, the model predicts people would tend to show more deference to
superiors. In Asian societies, for example, the degree to which one
lowers his/her head when bowing is proportional to the degree of
deference he/she is giving. When greeting or parting superiors, the
model predicts, Asians would lower their head more than when greeting
or parting subordinates. In short, the present model can explain the
influences of relational distance and power disparity on the degree of

politeness manifested in phatic communication.

49

Finally, some researchers (Baxter, 1984; Applegate, 1982) found
that gender is as good a predictor of politeness as relational
distance, power disparity, or relative face-threat. Specifically,
females were significantly more polite than males. Since Brown and
Levinson’s model does not account for individual differences, it cannot
explain this finding. The present model, however, can explain this
gender difference in terms of verbal aggressiveness. In this society,
females are expected to show more empathy and less verbal aggression in
their speech than males. In other words, people believe use of
verbally aggressive expressions is not appropriate for females (Lakoff,
1975). Females, therefore, are significantly lower in verbal
aggressiveness than males (Infante a Wigley, 1986). Since females are
less verbally aggressive, the present model predicts that females would
be more polite. In short, by incorporating such personality variables
as verbal aggressiveness and empathy, the present model explains gender
differences in politeness.

In summary, Brown and Levinson’s model reveals many conceptual
problems that prevent the model from explaining politeness phenomena
adequately. A new theoretical model that solves the problems of Brown
and Levinson’s model is proposed. The present model argues that people
have obligations to preserve both positive and negative face of others
regardless of the behavior involved. Thus, the model accounts for both
positive and negative politeness manifested in the same message.
Positive and negative politeness expressed in a message are a function
of speakers’ verbal aggressiveness, empathy, and the obligations to
save positive and negative face, respectively. The obligations to save

positive and negative face are a sum of the relational obligations to

50
save positive and negative face, respectively, and the threats to
positive and negative face, respectively. The relational obligation to
save positive face is positively related to the power the bearer has
over the speaker, and negatively related to the relational distance
between the speaker and the bearer; the relational obligation to save
negative face is positively related to both the power disparity and the
relational distance.

The present model explains well the empirical findings which Brown
and Levinson’s model has problems with. However, the ability to
provide an eg pee explanation does not indicate that the present model
can explain politeness adequately. In order to determine the adequacy
of the model, we need to test main propositions of the model directly.
Thus, two experiments will be designed to test the model. Experinent
one will examine whether the present model explains adequately
politeness phenomena manifested in phatic communication where no
intrinsically face-threatening acts are involved. Experiment two will
test the adequacy of the present model in explaining politeness

phenomena related to face-threatening acts.

52
respectively, when the interaction does not involve any
face-threatening act. Furthermore, as equation 2 states, the
relational obligations to save positive and negative face are a
function of the relational power the bearer has over the speaker and
the relational distance. Thus, when we substitute the power disparity
and the relational distance for the total obligations to save positive
and negative face in equation 4, we obtain the following equation for

the degree of positive and negative politeness:

where PD and Pn are the level of positive politeness and the level of
negative politeness, respectively; E represents empathy; A represents
verbal aggressiveness; P is the power the bearer has over the speaker;
D is the relational distance between the bearer and the speaker; and a,
b, c, m, n, o, p, x, y, and z are positive constants.

Equation 5 states that positive and negative politeness manifested
in a message are a function of the speaker’s empathy and verbal
aggressiveness, the power disparity, and the relational distance.
Specifically, the speaker’s empathy is positively related to the level
of both positive and negative politeness; the speaker’s verbal
aggressiveness is negatively related to the level of both positive and
negative politeness; the power disparity is positively related to the
level of both positive and negative politeness; and the relational
distance is negatively related to the level of positive politeness, and

positively related to the level of negative politeness. Based on the

53
predictions of the present model, the following hypotheses can be
proposed:

HYPOTHESIS 1: Actors’ empathy will be positively correlated with
the level of positive politeness and the level of
negative politeness manifested in performing a
non-FTA (non-face-threatening act).

HYPOTHESIS 2: Actors’ verbal aggressiveness will be negatively
correlated with the level of positive politeness
and the level of negative politeness manifested in
performing a non-FTA.

HYPOTHESIS 3: As the power the bearer has over the actor
increases, the level of positive politeness and the
level of negative politeness the actor manifests in
performing a non-FTA will increase.

HYPOTHESIS 4: As the social distance between the bearer and the
actor increases, the level of positive politeness
the actor manifests in performing a non-FTA will
decrease and the level of negative politeness the
actor manifests in performing the non-FTA will
increase.

Different from Brown and Levinson’s model that conceptualizes
positive and negative politeness as mutually exclusive categories, the
present model claims that people have obligations to save both positive
and negative face in the same message, and fulfill these obligations by
behaving in both a positively and negatively polite manner. Thus, the
above hypotheses predict both the level of positive politeness and the

level of negative politeness to be manifested in performing a non-FTA.

METHODS
This experiment is concerned with the effects of verbal
aggression, empathy, power disparity, and relational distance on
positive and negative politeness manifested in performing a
non-face-threatening act. This eXperiment employed a message

generation technique. Subjects were asked to write verbatim what they

54

typically say when they greet and part people. The level of positive
and negative politeness manifested in these generated messages were
rated by a group of judges.
Situation and Manipulation of Relational Variables

The non-face—threatening acts to be studied in this experiment are
greeting and parting. The most crucial standard for distinguishing
between a face-threatening act and a non-face-threatening act is
intrinsicality (Brown a Levinson, 1978). An act is intrinsically
face-threatening if the face-threat is a part of the act and there is
no way to detach this threat from the act. A request is an
intrinsically face-threatening act, because by definition exertion of
control is a part of a request. There is no request that does not
exert control on a bearer (Searle, 1969). An act is not intrinsically
face-threatening when the face-threat is not a part of the act but
caused by the speaker’s abnormal conduct. When one does not greet or
part the other properly, the face of the other would be highly
threatened. However, this threat is not intrinsic to greeting or
parting but to the user. In fact, greeting and parting actually
promote people’s desire for approval and power. In short, two of the
most representative non-face-threatening acts/{re greeting and parting.

A situation in which people exchange both greeting and parting
without performing an FTA is a short encounter composed of a greeting,
an exchange of two or three lines of small talk, and a parting. A
hypothetical situation of a short encounter was created withifour
different variations (see Appendix A). Participants were asked to
imagine that they ran into a person they knew in a hallway, and write

verbatim what they typically say and/or do to the person and what the

 

 

55
person typically says and/or does to them in the form of a script.
Relational distance was manipulated by informing participants that the
person they ran into was one they know very well (low distance) or just
an acquaintance (high distance). Power disparity was manipulated by
informing participants that the person they ran into was a professor
(high power) or a fellow student (equal power). Specifically, the
following instructions were given to subjects in the low distance/equal
power condition:
Imagine that after a two week term break, a new term begins today.
Also imagine that on your way to a class, you run into a classmate
(Mike Miller) whom you have not seen for a while. Both Mike
Miller and you took several courses together, and worked in the
same group for one class requiring a group research project.
Thus, Mike Miller and you know each other very well.
You notice that Mike Miller had his hair cut during the break and
wears a nice spring coat. Thus, he looks very refreshing. Since
your class begins in ten minutes, you do not have a lot of time to
talk with this classmate. Thus, you just want to exchange
greetings and partings.
What do you typically say to greet and part a classmate you know
very well? What will this classmate say to greet and part you?
Write verbatim what you and this classmate will say to greet and
then part each other. In other words, write a script in which you
and this classmate exchange greetings and then partings. You can
write as many turns as you think is necessary.
For the high distance/equal power condition, the person encountered was
described as an acquaintance with whom participants took a class
together and were in the same group when they did group exercises in
the class. Specifically, the last two sentences of the first paragraph
in the above instructions were replaced with "Both Mike Miller and you
took a course together last term, and were in the same group when you
did group exercises in the class. Other than that, you do not know

.Mike Miller very well." The phrase "you know very well" in the

subsequent sentences was altered into "you do not know very well." For

56

the low distance/high power condition, the person encountered was
described as a professor from whom participants took several classes
including an independent study. Specifically, the first paragraph of
the above instruction was substituted with "Imagine that after a two
week term break, a new term begins today. Also imagine that on your
way to a class, you run into a professor (Dr. Mike Miller) whom you
have not seen for a while. You know Mike Miller since you took several
classes from him; especially, you took an independent study under him
last term. Thus, Mike Miller and you know each other very well."; and
"classmate" in the subsequent sentences was altered into "professor."
For the high distance/high power condition, the person encountered was
described as "a professor (Dr. Mike Miller) whom you have not seen for
a while. You know Mike Miller since you took a class from him last
term. Because you were very active in this class, the professor also
learned your name. Other than that, you do not know Mike Miller very
well." The phrase "you know very well" in the subsequent sentences was
also altered into "you do not know very well."
Pilot Study

In order to check whether these variations of the situation could
manipulate the relational variables as designed, a pilot study was
performed. One hundred forty seven subjects were asked to read one of
the four situational variations (i.e., high distance/high power, high
distance/equal power, low distance/high power, and low distance/equal
power) and rate the relational distance between them and Mike Miller on
a 7-point scale (6 being very distant and 0 being very close) and also
rate the power Mike Miller has over them (also on a 7-point scale: 6

being Mike Miller has much more power and 0 being both parties have

57
equal power). If the manipulation of relational variables was done
successfully, readers of the "low distance" instructions should rate
the relational distance lower than readers of the "high distance"
instructions; readers of the "equal power" instructions should rate
Mike Miller less powerful than readers of the "high power"
instructions. Since relational variables were manipulated by the first
paragraph of each hypothetical situation, the subjects were given only
this part of the instructions.

Descriptive statistics for the rated social distance and the rated
power disparity are shown in Table 2 and 3, respectively. Tests of
homogeneity of variance for the rated variables showed that variances
of the rated distance were homogeneous across the four different
conditions (Cochran’s C=.26, p<.999; Bartlett-Box E=.06, p<.980);
however, variances of the rated power disparity were heterogeneous
(Cochran’s C=.39, p<.020; Bartlett-Box E=6.04, p<.001). Two-way
analyses of variance yielded a significant effect of relational
distance on the rated social distance (2:58.60, gt=1/143, p<.001,
etaz=.29) and a significant effect of power disparity on the rated
power disparity (2:32.67, d_f=1/143, p<.001, eta2=.36). Specifically,
subjects in the distant condition rated the relationship more distant
(M=3.92, SD=.98) than subjects in the close condition (M=2.70, SD=.95);
subjects in the hearer-high condition rated Mike Miller more powerful
(M=2.82, SD=1.58) than subjects in the equal power condition (M=.77,
SD=l.l3). Other main effects and interactions effects were not
statistically significant.

Since the rated power disparity showed heterogeneity of variance,

two Kruskal-Wallis one-way analyses of variance were performed to test

58

Table 2. Descriptive Statistics for Rated Distance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Relational Power
Equal Hearer High
Close M = 2.77 M = 2.62 M = 2.70
(Low) SD = .99 SD = .92 SD = .95
Relational n = 39 n = 37 n = 76
Distance Distant M = 4.08 M = 3.74 M = 3.92
(High) SD = .97 SD = .98 SD = .98
n = 36 n = 35 n = 71
M = 3.40 M = 3.17 M = 3.29
SD = 1.17 SD = 1.10 SD = 1.14
n = 75 n = 72 N = 147
Table 3. Descriptive Statistics for Rated Power
Relational Power
Equal Hearer High
Close M = 1.03 M = 2.76 M = 1.87
(Low) SD = 1.33 SD = 1.71 SD = 1.75
Relational n = 39 n = 37 n = 76
Distance Distant M = .50 M = 2.89 M = 1.68
(High) SD = .81 SD = 1.45 SD = 1.67
n = 36 n = 35 n = 71
M = .77 M = 2.82 M = 1.78
SD = 1.13 SD = 1.58 SD = 1.71
n = 75 n = 72 N = 147

 

 

 

 

59

the effects of relational distance and power disparity on the rated
power disparity. The results from these analyses were not different
from the results from the two-way ANOVA for the rated power. These
results generally indicated that the instructions could manipulate the
relational variables very effectively.
Participants and Procedure

Participants were 100 volunteers from various undergraduate
courses at a large Midwestern university. Participants received extra
credit points for participating in the experiment. Participants were
randomly assigned to one of the four experimental conditions in a 2 X 2
factorial design crossing two levels of power disparity (a bearer
having more power vs. equal power) with two levels of relational
distance (close vs. distant), making the sample size of each condition
25. Subjects were asked to write a script for the given short
encounter situation and to complete measurement scales for verbal
aggressiveness and empathy. All questionnaires were administered in
class.
lnstrumentatipn

Verbal aggressiveness was measured by Infante and Wigley’s (1986)
20-item measurement scale shown in Table 4. A confirmatory factor
analysis using Package (Hunter R Cohen, 1969) showed that the items in
this measurement scale were internally consistent. First, the
deviations between observed correlations and reproduced correlations
were minute. Only two out of 190 observed correlations significantly
deviated from the expected correlations at p<.01; no deviation exceeded
3 times the standard error (.092) of the mean correlation (.287); and

no item had more than one observed correlation that deviated

 

 

 

60

Table 4. Verbal Aggressiveness Measurement Scale

Indicate how often each statement is true for you in general when you
try to influence other persons.

IOO

11.

12.

13.

14.

Never Always
true true
0 1 2 3 4 5 6

I am extremely careful to avoid attacking individuals’
intelligence when 1 attack their ideas.

I try very hard to avoid having other people feel bad about
themselves when I try to influence them.

When people refuse to do a task I know is important without good
reason, I tell them they are unreasonable.

When others do things I regard as stupid I try to be extremely
gentle with them.

If individuals I am trying to influence really deserve it I attack
their character. '

When people behave in ways that are in very poor taste, I insult
them in order to shock them into proper behavior.

I try to make people feel good about themselves even when their
ideas are stupid.

When people simply will not budge on a matter of importance I lose
my temper and say rather strong things to them.

When people criticize my shortcomings, I take it in good humor and
do not try to get back at them.

When individuals insult me, I get a lot of pleasure out of really
telling them off.

When I dislike individuals greatly, I try not to show it in what I
say or how I say it.

I like poking fun at peeple who do things which are very stupid in
order to stimulate their intelligence.

When I attack persons’ ideas, I try not to damage their
self-concepts.

When an argument shifts to personal attacks, I try very hard to
change the subject.

15.

16.

17.

18.

19O

20.

61

Table 4 (Cont’d)

When I try to influence people, I make a great effort not to
offend them.

When people do things which are mean or cruel, I attack their
character in order to help correct their behavior.

I refuse to participate in arguments when they involve personal
attacks.

When nothing seems to work in trying to influence others, I yell
and scream in order to get some movement from them.

When I am not able to refute others’ positions, I try to make them
feel defensive in order to weaken their positions.

When individuals are very stubborn, I use insults to soften the
stubbornness.

 

 

62

Table 5. Empathy Measurement Scale
Indicate how often each statement is true for you in general.
Never Always

true true
0 1 2 3 4 5 6

1. Before criticizing somebody, I try to imagine how I would feel if
I were in their place.

*2. If I’m sure I’m right about something, I don’t waste much time
listening to other people’s arguments.

3. I sometimes try to understand my friends better by imagining how
things look from their perspective.

4. I sometimes find it difficult to see thing from the other person’s
point of view.

5. I try to look at everyone’s side of a disagreement before I make a
decision.

6. When I’m upset at someone, I usually try to "put myself in his or
her shoes" for a while.

7. I believe there are two sides to every question and I try to look
at them both.

* Dropped out of analysis

63
significantly from its expected correlation. Second, the deviations of
individual correlations from the mean correlation were also quite small
(i.e., the matrix was flat). Only two out of 190 correlations deviated
from the mean correlation at p<.01; and no item had more than one
correlation that deviated significantly from the mean correlation. The
reliability coefficient elppe for this 20-item verbal aggressiveness
measure was .89.

Empathy was measured by Davis’s (1983) 7-item measurement scale of
empathy as perspective-taking presented in Table 5. A confirmatory
factor analysis using Package showed that one of these seven items
(item #2 in Table 5) did not represent the same underlying dimension.
The correlations between this one item and other items were
considerably lower than the mean correlation. The average correlation
between this one item and other items was .14, while the mean
correlation of the overall matrix was .37. Reliability analysis for
this scale showed that the item-total correlation for this item (.20)
was much lower than other item-total correlations (the next lowest one
was .50). Thus, this item was excluded from further analyses. The
reliability coefficient elppe for the 6-item empathy measurement scale
was .83.

Gains

While defining positive and negative politeness, the present model
suggests (1) confidence shown in approval, (2) exaggeratedness of
approval, (3) tentativeness of disapproval, and (4) understatedness of
disapproval as criteria for determining the level of positive
politeness, and (5) frequency of "vous" pronouns or address terms, (6)

conventional indirectness in influence attempts, and (7) hesitance of

64

Table 6. Politeness Rating Scale

Positive Politeness

I. Approval

1.

3.

How much does the speaker try to approve or appreciate the

hearer’s performance, possessions, or abilities? (General
approval)

How confidently does the speaker approve or appreciate the
hearer’s performance, possessions, or abilities? (Confidence
of approval)

How exaggeratedly does the speaker approve or appreciate the
hearer’s performance, possessions, or abilities?
(Exaggeratedness of approval)

II. Avoidance of Disapproval

4.

How much does the speaker try to avoid criticizing or
disapproving of the hearer’s performance, possessions, or
abilities? (General avoidance of disapproval)

How hesitantly does the speaker criticize or disapprove of
the hearer’s performance, possessions, or abilities?
(Tentativeness of disapproval)

How understatedly does the speaker criticize or disapprove of
the hearer’s performance, possessions, or abilities?
(Understatedness of disapproval)

Negative Politeness

III.

7.

8.

Deference

How respectful is the speaker to the hearer? (General
deference)

How frequently does the speaker use "vous" form pronouns or
address terms? (Frequency of using "vous" pronouns)

65

Table 6 (Cont’d)
IV. Avoidance of Imposition

9. How much does the speaker try to avoid imposing on the bearer
or interfering with the hearer’s freedom of action? (General
avoidance of imposition)

10. How hesistantly does the speaker impose on the bearer or
interfere with the hearer’s freedom of action? (Tentativeness
of imposition)

11. How indirectly does the speaker impose on the bearer or
interfere with the hearer’s freedom of action? (Indirectness
of imposition)

Global Politeness

12. How much does the speaker try to express his/her meaning
politely?

H's—.- a

66

imposition as criteria for determining the level of negative
politeness. In order to code positive and negative politeness
manifested in messages an instrument (see Table 6) was developed. This
instrument included seven items that represented the above seven
criteria for positive and negative politeness, four items that measured
the extent to which the speaker generally (1) approves the bearer, (2)
tries to avoid disapproving of the hearer, (3) shows respect to the
hearer, and (4) tries to avoid imposing on the bearer, and one item
that measured global politeness. In other words, positive politeness
was measured by six items: three items representing the degree of
approval or approach-based positive politeness (i.e., general approval,
confidence of approval, and exaggeratedness of approval) and three
items representing the degree of avoiding disapproval or
avoidance-based positive politeness (i.e., general avoidance of
disapproval, tentativeness of disapproval, and understatedness of
disapproval). Negative politeness was measured by five items: two
items representing the degree of power-giving or approach-based
negative politeness (i.e., general respect and frequency of "vous"
pronouns) and three items representing the degree of avoiding
imposition or avoidance-based negative politeness (i.e., general
avoidance of imposition, conventional indirectness, and hesitance of
imposition). In addition global politeness asking "how polite is the
speaker?" was measured.

Three judges read scripts generated by participants and rated
positive and negative politeness manifested in the participants’
utterances in each script based on this instrument. Of the twelve

items in the instrument, three items measuring avoidance of disapproval

-39.“!

67
(i.e., general avoidance of disapproval, tentativeness of disapproval,
and understatedness of disapproval; items 4, 5, and 6 in Table 6) and
three items measuring avoidance of imposition (general avoidance of
imposition, conventional indirectness, and hesitance of exerted
control; items 9, 10, and 11 in Table 6) were excluded from further
analyses since all the subjects scored zero for these categories. In
other words, only approach-based positive and negative politeness were
analyzed in this study. Inter-rater reliability (Cureton’s reliability
of average ratings) was .91 for "general approval," .88 for "confidence
of approval," .86 for "exaggeratedness of approval," .88 for "general
respect," and .96 for "frequency of vous pronouns or address terms."
The reliability coefficient elppe was .92 for the three-item
measurement of approval and the correlation between the two items in
the measurement of deference was .96.

Inter-rater reliability for global politeness was very low
(Cureton’s reliability of average ratings was .61). Different judges
seemed to focus on different dimensions of politeness. For one judge,
the correlation between global politeness and positive politeness
(r=.37) was higher than that between global politeness and negative
politeness (r=.21), whereas the other two judges yielded higher
correlations between global politeness and negative politeness (r=.42
and .33) than those between global politeness and positive politeness
(r=.21 and .23). Because of low reliability, global politeness was not

included in further analyses.

68
RESULTS

Man' u ation Check

In order to check whether variations in the hypothetical
situations properly manipulated the relational variables, subjects were
asked to rate the relational distance and power disparity between them
and Mike Miller before they wrote out the script. Five-point
Likert-type scales were employed to rate these relational variables.
For the measure of relational distance, "1" represented "very close"
and "5" represented "very distant"; for the measure of power disparity,
"1" represented "Mike Miller (i.e., the bearer) has much more power

over you," "3"

represented "Both of you have equal power," and "5"
represented "You have much more power over Mike Miller." If the
manipulation of relational variables was done successfully, subjects in
the "low distance" conditions should rate the relational distance lower
than subjects in the "high distance" conditions; subjects in the "equal
power" conditions should rate Mike Miller less powerful (i.e., higher
score on the rated power) than subjects in the "high power" conditions.
Descriptive statistics for the rated relational distance and the
rated power disparity are presented in Table 7 and Table 8,
respectively. A two-way analysis of variance for the rated relational
distance showed a significant main effect of relational distance
(2:92.40, gt=1l96, p<.001, eta2=.48). Specifically, subjects in the
distant condition rated the relationship more distant (M=3.46, SD=.76)
than subjects in the close condition (M=2.12, SD=.63). The main effect
of power disparity (F=1.67, g1=1l96, p<.200, eta2=.01) and the
interaction effect between relational distance and power disparity

(2:.52, gt=1l96, p<.475, eta2=.003) were insignificant. A test of

69

Table 7. Descriptive Statistics for Rated Distance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Relational Power
Equal Hearer High

Close M = 2.16 M = 2.08 M = 2.12
(Low) SD = .69 SD = .57 SD = .63

Relational n = 39 n = 37 n = 76
Distance Distant M = 3.60 M = 3.32 M = 3.46
(High) SD = .65 SD = .85 SD = .76

n = 25 n = 25 n = 50
M = 2.88 M = 2.70 M = 2.79
SD = .98 SD = .95 SD = .97
n = 50 n = 50 N = 100

Table 8. Descriptive Statistics for Rated Power
Relational Power
Equal Hearer High

Close M = 3.40 M = 2.40 M = 2.90
(Low) SD = .50 SD = .87 SD = .86

Relational n = 39 n = 37 n = 76
Distance Distant M = 3.20 M = 2.52 M = 2.86
(High) SD = .41 SD = .65 SD = .64

n = 25 n = 25 n = 50
M = 3.30 M = 2.46 M = 2.88
SD = .46 SD = .76 SD = .76
n = 50 n = 50 N = 100

 

 

 

 

70
homogeneity of variance showed that variances of the rated distance
were homogeneous across the four different conditions (Cochran’s C=.37,
p<.122; Bartlett-Box E=1.37, p<.251).

A two-way (relational distance x power disparity) analysis of
variance for the rated power yielded a significant main effect of power
disparity (2:44.29, gt=1l96, p<.001, eta2=.3l). Subjects in the
hearer-high condition rated Mike Miller more powerful (M=3.30, SD=.46)
than subjects in the equal power condition (M=2.46, SD=.76). The main
effect of relational distance (E=.10, gt=ll96, p<.752, eta2=.001) and
the interaction effect between relational distance and power disparity
(g=1.61, gt=1/96, p<.208, eta2=.01) were insignificant. A test of
homogeneity of variance revealed that variances of the rated power
disparity were not homogeneous (Cochran’s C=.47, p<.003; Bartlett-Box
£=5.00, p<.002). Two Kruskal-Wallis one-way analyses of variance were
performed to test the effects of relational distance and power
disparity on the rated power disparity. The results were not different
from the results yielded by the two-way analysis of variance for the
rated power. In short, the manipulation of the relational variables
seemed to be successful.

Politeness in Performing a Non-Face-Threatening Act

One of the most basic claims of the present model is that people
would try to be polite even when performing a non-face-threatening act.
In order to test this claim, two one—sample t-tests were performed, one
each for positive and negative politeness. Both positive (M=9.24,
80:10.24, t=9.03, g_=99, p<.001) and negative politeness (M=1.03,
SD=1.06, t=9.73, gt=99, p<.001) were significantly different from zero

or non-occurrence of politeness. In other words, even when the act

71

involved did not threaten the hearer’s face, subjects tried to maintain
positive and negative politeness, supporting the present model.
The Effects pt Personality Variables

It was hypothesized that empathy would be positively correlated
with positive and negative politeness (Hypothesis 1) and that verbal
aggressiveness would be negatively correlated with positive and
negative politeness in performing a non-FTA (Hypothesis 2).

The Effecte of Personality Variables on Eoeitive Eplitenes . A

zero-order correlation analysis yielded statistically insignificant

 

relationships between positive politeness and verbal aggressiveness
(r=-.13, p<.206) and between positive politeness and empathy (r=.11,
p<.289). Table 9 presents Pearson product-moment correlation
coefficients between positive politeness and verbal aggression,
empathy, relational distance, and power disparity. A regression
analysis of positive politeness on relational distance, power
disparity, verbal aggression, and empathy revealed that none of the
personality variables had a significant effect on the dependent
variable. The standardized regression coefficient pete was -.07
(t=-.69, gt=99, p<.493) for verbal aggression and was .11 for empathy
(t=l.00, gt=99, p<.319). In contrast, the two relational variables,
i.e., relational distance (beta=-.21, t=-2.19, gis99, p<.031) and power
disparity (beta=.24, t=2.44, g§=99, p<.017) showed considerably larger
effect sizes. The squared multiple regression coefficient (R2) for
these two relational variables was .10 (E=5.09, gt=2/97, p<.008). The
change in R2 caused by adding the two personality variables was .02

(2:1.33, gt=2/97, p<.270). In short, both verbal aggressiveness and

Table 9.

Empathy
Gender

Distance

Power

Positive
Politeness

Table 10.

Empathy
Gender
Distance
Power

Negative
Politeness

Verbal
Aggression

-.4499
-.3316
.0347
.0163

-01275

Verbal
Aggression

-.4499
-.3316
.0347
.0163

.0144

72

Empathy

.2301
-00380
- s 1866

.1071

Empathy

.2301
-.0380
-.1866
-.1980

Gender

”00210
-.0629

.2046

Gender

-0 0210
-.0629
-.0388

Correlations between Positive Politeness and Predictors

Relational Power

Distance

.0000

-.2180

Disparity

.2180

Correlations between Negative Politeness and Predictors

Relational Power

Distance

.0000
.1424

Disparity

.8261

73
empathy did not show any significant relationship with positive
politeness. The results did not support either Hypothesis 1 or
Hypothesis 2.

The Effects of Personality Variables on Negative Pelitenees. A
zero-order correlation analysis for negative politeness showed almost
no covarying relationship (r=.01, p<.887) between verbal aggressiveness
and negative politeness and a moderate negative relationship (r=-.20,
p<.048) between empathy and negative politeness as presented in Table
10. However, a regression analysis of negative politeness on
relational distance, power disparity, verbal aggression, and empathy
revealed that the relationship between empathy and negative politeness
was accounted for by the other predictors. The standardized regression
coefficient pete for empathy was -.05 (t=-.83, d1=99, p<.411). Verbal
aggressiveness (beta=-.03, t=-.44, dt=99, p<.662) also did not show any
significant effect on negative politeness. In contrast, relational
distance (beta=.14, t=2.53, dt=99, p<.013) and power disparity
(beta=.82, t=14.35, gf=99, p<.001) yielded large effect sizes. The
squared multiple regression coefficient for these two relational
variables was .70 (F=114.65, gt=2/97, p<.001). Addition of verbal
aggression and empathy increased R2 by less than .01 (£=.34, §t=2/97,
p<.710). In short, both verbal aggressiveness and empathy did not show
any significant relationship with negative politeness. The results
rejected Hypothesis 1 and Hypothesis 2.

,x’IjEEEEEEEFE} the results altogether indicated that verbal
aggression andJempathy of an actor were not uniquely correlated with
the level of positive and negative politeness in the message generated

by the actor. Thus, Hypothesis 1, that predicted actors’ empathy will

74
be positively correlated with the level of positive politeness and the
level of negative politeness manifested in performing a
non-face-threatening act, and Hypothesis 2, that predicted actors’
verbal aggressiveness will be negatively correlated with the level of
positive politeness and the level of negative politeness manifested in

performing a non-face-threatening act, were rejected.

- ”a...” _ _
"V-—I--~._-

The Sffecte of Relational Variables

It was hypothesized that as the power the bearer has over the
actor increases, both positive and negative politeness would increase
(Hypothesis 3) and that as the social distance between the bearer and
the actor increases, the level of positive politeness would decrease
and the level of negative politeness would increase (Hypothesis 4).

The Effects pf Relational Variables on Positive Eelitenee . In

order to test the effects of relational distance and power disparity on

 

positive politeness, the occurrence of positive politeness was assessed
across the four conditions of the 2 x 2 design (i.e., high
distance/high power, high distance/low power, low distance/high power,
and low distance/low power). Since the effects of verbal
aggressiveness and empathy on positive politeness were not
statistically significant, the effects of these personality variables
were not removed from the dependent variable of positive politeness.
Descriptive statistics for positive politeness are shown in Table 11.

A test for homogeneity of variance showed that the variances of
the four conditions were homogeneous for positive politeness (Cochran’s
C=.31, p<.672; Bartlett-Box E:1.00, p<.391). The results indicate that

the data for positive politeness did not violate the homogeneity of

Table 11.

Descriptive Statistics for Positive Politeness

 

Relational Power

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Equal Hearer High

Close M = 8.28 M =14.64 M =11.46
(Low) SD =10.31 SD =10.95 SD =11.01

Relational n = 39 n = 37 n = 76
Distance Distant M = 5.76 M = 8.28 M = 7.02
(High) SD = 7.75 SD =10.05 SD = 8.97

n = 25 n = 25 n = 50
M = 7.02 M =11.46 M = 9.24
SD = 9.12 SD =10.89 SD =10.24

n = 50 n = 50 N = 100

Table 12. Descriptive Statistics for Negative Politeness

Relational Power
Equal Hearer High

Close M = .08 M = 1.68 M = .88
(Low) SD = .28 SD = .90 SD = 1.04

Relational n a 39 n = 37 n = 76
Distance Distant M = .24 M = 2.12 M = 1.18
(High) SD = .44 SD = .53 SD = 1.06

n = 25 n = 25 n = 50
M = .16 M = 1.90 M = 1.03
SD = .37 SD = .76 SD = 1.06

n = 50 n = 50 N = 100

 

 

 

 

76
variance assumption of ANOVA. Thus, ANOVA was employed to examine the
influences of the relational variables on positive politeness.

A two-way ANOVA for positive politeness yielded significant main
effects of relational distance (2:5.10, gt=1/96, p<.026, eta2=.05) and
power disparity (i=5.1o, et=1/96, p<.026, eta2=.05) and an
insignificant interaction effect between these two relational variables
(E=1.00, gt=1l96, p<.332, eta2=.01). Specifically, actors tended to
show more positive politeness to close persons (M=11.46, SD=11.01) than
distant persons (M=7.02, SD=8.97), and to powerful persons (M=11.46,
SD=10.89) than to persons with equal power (M=7.02, SD=9.12). These
results support Hypothesis 3 predicting a positive effect of power
disparity on positive politeness and Hypothesis 4 predicting a negative
effect of relational distance on positive politeness.

he Ef ect Rel i nal V riab es on e at'v P ' ess. In
order to test the effects of relational distance and power disparity on
negative politeness, the occurrence of negative politeness was assessed
across the four conditions of the 2 X 2 design (i.e., high
distance/high power, high distance/low power, low distance/high power,
and low distance/low power). Since the effects of verbal
aggressiveness and empathy on negative politeness were not
statistically significant, the effects of these personality variables
were not removed from the dependent variable of negative politeness.
Descriptive statistics for negative politeness are shown in Table 12.

A test for homogeneity of variance showed that variances of the
four conditions were not homogeneous for negative politeness (Cochran’s
C=.60, p<.001; Bartlett-Box 2:10.95, p<.001). In other words, the data

for negative politeness violated the homogeneity of variance assumption

77
of ANOVA. Thus, Kruskal-Wallis one-way analysis of variance for
ordinal scales was used to analyze the effects of the relational
variables on negative politeness.

A Kruskal-Wallis one-way ANOVA of negative politeness yielded a
significant effect for power disparity (chi-square=68.25, gt=1,
p<.001). Specifically, subjects showed more deference to the hearers
with more power (mean rank=72.78) than the hearers with less power
(mean rank=28.22). The effect of relational distance on negative
politeness was statistically insignificant (chi-square=3.32, eg=1,
p<.072).

Since analysis of variance (i.e., E-test for interval scales) is
robust to the violation of the equality of variances, especially when
all groups have the same cell size, a two-way analysis of variance was
performed for negative politeness. This analysis yielded very similar
results to the Kruskal-Wallis one-way analysis of variance. The
influence of power disparity was highly significant (S=223.71, gi=l/96,
p<.001) and its effect size was remarkably large (eta2=.68).
Specifically, subjects showed more deference to more powerful persons
(M=1.90, SD=.76) than less powerful persons (M=.16, SD=.37). In
contrast, even though the effect of relational distance (£=6.65,
gi=1l96, p<.011) was statistically significant, the effect size was
relatively small (eta2=.02). The interaction between relational
distance and power disparity was not statistically significant (2:1.45,
,gf=1/96, p<.232, eta2=.01). In short, power disparity showed a strong
influence on negative politeness, supporting Hypothesis 3, but
relational distance showed a small effect on negative politeness, not

supporting Hypothesis 4.

78

In sum, the results indicated that as the power the bearer had
over the speaker increased, the speaker tended to show more positive
wand negative politeness. These findings support Hypothesis 3 that
predicted the level of positive and negative politeness an actor
manifests in performing a non-FTA will increase with the power the
hearer has over the actor. The results also indicate that as the
relational distance increased, people tended to show less positive
politeness; however, the effect of relational distance on negative
politeness was minute. These results partially support Hypothesis 4
that predicted the level of positive politeness the actor manifests in
performing a non-FTA will decrease and the level of negative politeness
the actor manifests in performing the non-FTA will increase with the
social distance between the bearer and the actor.
Theegffect of Gender

As Table 9 shows, gender yielded a moderate negative relationship
with verbal aggression (r=-.33, p<.001) and a moderate positive
relationship with empathy (r=.23, p<.021). A regression analysis of
gender on verbal aggression and empathy yielded a squared multiple
regression coefficient of .12 (E=6.50, ei=2/97, p<.002). Verbal
aggression was a better predictor of gender than empathy, its

standardized regression coefficient beta being -.29 (t=-2.70,.gt=99,

 

p<.009). The beta for empathy was .10 (t=.95, gt=99, p<.345). A
one-way analysis of variance also showed that verbal aggressiveness was
significantly different between different genders (2:12.11, gtsl/SB,
p<.001, etaz=.11). Males (M=46.23, SD=12.90) were more aggressive than
females (M=36.42, SD=13.73). In short, verbal aggressiveness could

predict gender quite well as argued earlier.

79

A one-way ANOVA for positive politeness showed that males and
females were different in positive politeness (£=4.28, df=1l98, p<.04l,
eta2=.04). Since verbal aggressiveness was not equivalently
distributed between different genders, a one-way analysis of covariance
eliminating the effect of verbal aggression was performed for positive
politeness. The results showed that gender was not a significant
predictor of positive politeness (2:3.01, gt:1/97, p<.086) any more.
Negative politeness was not different between males and females.
One-way analysis of variance yielded an insignificant effect of gender
on negative politeness (E=.513, gt=1/98, p<.476). In short, gender was
not a very good predictor of politeness, and most of its effect on

politeness could be explained by verbal aggressiveness.

DISCUSSION

_....1......._

 

 

—~

One of the major findings of this experiment is that people make .
use of politeness devices even when no intrinsic face threat is
involved in the intended act. Subjects in this experiment actively
made use of such positive politeness devices as compliment and such
negative politeness devices as "vous" form pronouns or address terms.
This finding supports Ferguson’s (1976) claim that people employ
politeness expressions in a ritualized communication interaction as

well as the present model’s claim that people share obligations to be

M. ._ ._ the?!“ pa

polite regardless of the kind of act they are performing.

This experiment partially confirmed the politeness model proposed .
x/“\
in Chapter 1. First, the power the bearer has over the speaker was a (\l#)

\.

good predictor of both positive and negative politeness. Speakers show

more positive and negative politeness to hearers with more relational

80

power. In other words, people appreciate their superior’s performances
or possessions more than their peer’s performances or possessions;
people show more deference to their superior than to their peer.

Second, even though the effect size was small, relational distance
between two actors turned out to influence positive politeness.
Speakers show more positive politeness to friends than acquaintances.
In other words, people appreciate their friend’s performances or
possessions more than their acquaintance’s performances or possessions.

While a two-way analysis of variance yielded a small but
significant effect of relational distance on negative politeness,
Kruskal-Wallis one-way analysis of variance showed the effect of
relational distance on negative politeness was statistically
insignificant. One possible explanation for this result is that the
manipulation of relational distance in this experiment is not enough to
trigger differences in negative politeness. In this experiment,
relational distance was manipulated by describing the hypothetical
person as a classmate with whom subjects took several courses together
for the close condition and a classmate with whom subjects took only
one class together for the distant condition, or by describing the
hypothetical person as a professor from whom subjects took several
courses for the close condition and a professor from whom subjects took
only one class. This distinction certainly made differences in
subjects’ perception of relational distance. However, it seems that
because use of pronouns or address terms (which is one major criterion
for negative politeness) is highly conventionalized, people may not
vary the way they address others unless there is a wide difference in

relational distance. In short, the manipulation of relational distance

81
may not be enough to cause subjects to change their behavior concerning
negative politeness.

Although this experiment failed to find a significant effect of
relational distance on negative politeness, the findings concerning the
effect of relational distance on positive politeness and the effect of
power disparity on both positive and negative politeness confirmed the
predictions made by the present model. By confirming these predictions
of the present model, the findings also support the claim of the model
that people incur obligations to promote or save their partner’s
positive and negative face simply by having a relationship with the
partner.

The present model’s predictions concerning personality variables
such as verbal aggression and empathy were not supported. The results
indicated that verbal aggressiveness and empathy did not influence
positive and negative politeness in non-face-threatening situations.
One possible explanation for this finding is that the particular
instruments measuring verbal aggressiveness and empathy employed in
this experiment are devised mainly to investigate people’s selection of
message contents with different degrees of verbal aggressiveness and
empathy. Thus, these instruments might not be adequate to measure
differences in verbal aggressiveness and empathy in the peppep message
contents are expressed. Another possible explanation is that the
relationship between personality variables and politeness is contingent

on the nature of the act being performed. Greeting and parting are so

 

.—.. fl. ,

ritualized that actors may not have enough opportunities to reflect

their idiosyncrasies in the interaction.

82

Different from previous studies (Baxter, 1984; Brown, 1976, 1980;
Kemper, 1984; Shimanoff, 1981), this experiment found that gengepflwase-
not a very powerful predictor of politeness. Gender showed a small
effect on positive politeness; however, when differences in verbal
aggressiveness were removed from gender differences, the effect became
statistically insignificant. Moreover, gender differences in negative
politeness were insignificant.

In summary, this experiment proved that people maintain both
positive and negative politeness when they perform a non-FTA. Both
relational distance and power disparity were found to have significant
impacts on positive and negative politeness. However, personality
variables such as verbal aggressiveness and empathy did not show any

significant influence on politeness.

CHAPTER III

EXPERIMENT 2

HYPOTHESES

The model proposed in Chapter 1 offers an explanation for
politeness manifested in performing a face-threatening act as well as
politeness manifested in performing a non-face-threatening act. While
experiment I examined the adequacy of the present model in explaining
politeness manifested in performing a non-FTA, this experiment tests
the adequacy of the present model in accounting for politeness
manifested in performing an FTA.

Equation 4 states that the degrees of positive and negative
kpoliteness are a function of the total obligations to save positive and
negative face, respectively, and the speakeris verbal aggressiveness

,W t‘ M
,‘Wfifl. rte-fl Au

and empathy. According to equation 3, the total obligation to save

WW...“ .---- W-fl

positive face is a sum of the relational obligation to save positive

" — —‘~M~¢_—.—a- m- , .2 .-. Mao-I‘-

face and the threat to positive face carried by the act; the total

obligation to save negative face is a sum of the relational obligation

to save negative face and the threat to negative face carried by the

_,-1arr'l' ‘—”~a-q-

act. Furthermore, as equation 2 states, the relational obligations to
save positive and negative face are a function of the relational power
the bearer has over the speaker and the relational distance. Thus,
when we substitute the threats to positive and negative face, the power
disparity, and the relational distance for the total obligations to

save positive and negative face into equation 4, we obtain the

following equation for the degree of positive and negative politeness:

83

84

where PD and Pn are the level of positive politeness and the level of
negative politeness, respectively; TD and Tn are the threat to positive
face and the threat to negative face, respectively; E represents
empathy; A represents verbal aggressiveness; P is the power the bearer
has over the speaker; D is the relational distance between the bearer
and the speaker; and a, b, c, m, n, o, p, x, y, and z are positive
constants.

A.

Equation §_§tates that positive and negative politeness manifested
W ,-... __A -._ _ ._ ._ ._ _ - H __ ___7____“ —-.., __ --_ * -- .

...u _-.. a- .

in a message are a function of the threats to positive and negative

face, respectively, the speaker’s empathy and verbal aggressiveness,
the power disparity, and the relational distance. Specifically, the
threat to positive face and the threat to negative face are positively
related to the level of positive and negative politeness, respectively;
the speaker’s empathy is positively related to the level of both
positive and negative politeness; the speaker’s verbal aggressiveness
is negatively related to the level of both positive and negative
politeness; the power disparity is positively related to the level of
both positive and negative politeness; and the relational distance is
negatively related to the level of positive politeness, and positively
related to the level of negative politeness. Based on the predictions
of the present model for politeness related to FTA, the following

hypotheses can be made:

85

HYPOTHESIS 1: The threat to positive face carried by an FTA will
increase the level of positive politeness
manifested in perforning the FTA, and the threat to
negative face carried by the FTA will increase the
level of negative politeness manifested in
performing the FTA.

HYPOTHESIS 2: Empathy will be positively correlated with the
level of positive politeness and the level of
negative politeness manifested in performing an
FTA.

HYPOTHESIS 3: Verbal aggressiveness will be negatively correlated
with the level of positive politeness and the level
of negative politeness manifested in performing an
FTA.

HYPOTHESIS 4: The power the bearer has over the actor will
increase the level of positive politeness and the
level of negative politeness manifested in
performing an FTA.

HYPOTHESIS 5: The social distance between the bearer and the
actor will decrease the level of positive
politeness manifested in performing an FTA and

increase the level of negative politeness
manifested in performing the FTA.

METHODS

This experiment was concerned with the effects of face-threat.u,
carried by an act, verbal aggressiveness, empathy, power disparity, and
relational distance on positive and negative politeness manifested in
performing a face-threatening act. This experiment employed a message
generation technique. Subjects were asked tgiwrite verbatim what they
would say when they needed to perform a given act in a given situation.
The level of positive and negative politeness manifested in these
generated messages was rated by a group of judges.
Situations and Manipulation of Independent Variables

This experiment employed four different face-threatening acts to

manipulate the degree of positive and negative face-threat. Each of

86
these four FTAs represented one of the four conditions of face-threat
obtained by crossing two levels of positive face threat (high vs. low)
with two levels of negative face-threat (high vs. low). The
face-threatening act used to create the high positive/high negative
face-threat condition was one that requests hearers to re-do their work
due to the poor quality of their first work (i.e., "request of
re-writing"). This request is high in positive face-threat because it
disapproves of a hearer’s performance by presupposing that the
performance is of poor quality. This request is high in negative
face-threat if the work to be re-done takes a lot of time and energy.
Thus, the FTA used in the high positive/high negative face-threat
condition was a request to re-do a difficult job.

The face-threatening act employed to create the high positive/low
negative face-threat condition was one that criticizes hearers’
performance (i.e., "negative evaluation"). This criticism is high in
positive face-threat since it discredits hearers by expressing that
their performance is of poor quality; negative face-threat of this
criticism is low since hearers are not controlled by speakers.

The face-threatening act used in the low positive/high negative
face-threat condition was one that requests a bearer to do (not re-do)
work that requires a lot of time and energy (i.e., "request of taking
over another’s responsibility"). This request threatens negative face
highly, but does not threaten positive face to any significant extent
because it does not imply any disapproval of the person one is talking
to.

The face-threatening act used in the low positive/low negative

face-threat condition was one that asks others’ opinions (i.e., "asking

87
opinion"). Asking hearers’ opinions is low in both positive and
negative face-threat when hearers are ready to make comments (thus, no
extra effort is needed), and the issue is not directed toward some
negative aspects of hearers (thus, no disapproval is intended).
In sum, four speech acts were used to create four different
conditions of face-threat. Specifically, the high positive face-threat

condition included "negative evaluation" and "request of re-writing";

the low positive face-threat condition included "asking epinion" and

up.»
Hr..-

"request of taking over another’s responsibility"; the high negative
face-threat condition included "request of re-writing" and "request of
taking over another’s responsibility"; the low negative face-threat
condition included "asking opinion" and "negative evaluation."

A hypothetical situation was created with sixteen possible
variations, depending on the degree of positive face-threat (high vs.
low), the degree of negative face-threat (high vs. low), power
disparity (equal vs. hearer low), and relational distance (high vs.
low) (see Appendix B). Specifically, the hypothetical scenario for the
high positive/high negative face-threat, equal power, and low
relational distance condition was as follows:

Imagine that you are involved in a group research project that
will solely decide your course grade in a class where you very
much want to get a high grade. Naturally, you are one of the more
enthusiastic members of the group. In your view, one of the group
members (John Brown) has done his part of the project very poorly.
John was supposed to write the opening chapter of your research
report. The opening chapter was expected to make three major
arguments. In order to do this properly, John should have found
enough evidence to prove these three important points of the
chapter. However, evidence provided in this member’s writing is
mostly irrelevant to the issues, thus weakening the validity of
the arguments.

So, you want to ask John to re-do his part of the project in the
remaining time before the final project is due. John will

88

basically have to start over again. Suppose that John ie seeeone
you regard as a good friend.

For the high positive/low negative face-threat condition, the second
paragraph was rewritten as: "One day, John, who yop pegepg pp p good
teiee_, asks you what you think about his work. Since this matter is
related to the group’s grade, you want to tell him that he did a poor
job, hoping that he/she volunteers to re-do the work." For the low
positive/high negative face-threat condition, the second paragraph was
rewritten as: "Since you want a better group grade and since you think
John Brown does not want to do his part of the project again, you want
to ask another member of the group (Bill Jones), who you think is very
actively involved in this group project, to take over and re-do John’s
part of the project in the remaining time before the final project is
due. If Bill decides to take over the job, he will basically have to
start over again. Suppose that Bill is someone yep pegepd 5; e geee
tpieeg." For the low positive/low negative face-threat condition, the
second paragraph was rewritten as: "One day you run into another
member of your group (Bill Jones), whoeyou regerd as a good friep .
You want to know what Bill thinks of John’s work. So, you decide to
ask Bill’s opinion on John’s work. Suppose that Bill already read the
opening chapter done by John."

For the high relational distance condition, the person in question
(i.e., the hearer) was described as one participants do not know very
well, except for project group meetings as compared to a "friend" in
the low relational distance condition. For the hearer—low power
disparity condition (i.e., the power a bearer has over a speaker is

low), the agent (i.e., the role to be assumed by participants) was

89
described as an undergraduate teacher’s assistant (UTA) who was in
charge of a group as compared to a "group member" in the equal power
condition. Specifically, the first paragraph of the scenario was
replaced by the following:

Imagine that you enrolled in an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took
this class last year, you are allowed to enroll in the independent
study to be a UTA of this class. As one of five UTAs in this
class, your duty involves taking charge of seven students who are
working together on a group research project. Your grade for the
independent study as well as the course grade of these students
will be decided solely based on the group’s performance in the
project. You very much want to get a high grade, because you want
to be a UTA for this class again next term. In your view, one of
the group members (John Brown) has done his part of the project
very poorly. John was supposed to write the opening chapter of
the group’s research report. The opening chapter was expected to
make three major arguments. In order to do this properly, John
should have found enough evidence to prove these three important
points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity
of the arguments.

Pilot Study

In order to check whether these variations of the situation could
manipulate the relational variables and face-threats as designed, a
pilot study was performed. Two separate 2 X 2 factorial designs (i.e.,
one 2 X 2 design for relational distance and power disparity and one 2
X 2 design for positive and negative face-threat) were employed instead
of one 2 X 2 X 2 X 2 factorial design to eliminate unnecessary
interaction effects between relational variables and face-threats.
Perception of relational distance and power disparity might be
influenced by the type of act the speaker is performing and perception
of face-threat might be influenced by the relationship between the
speaker and the bearer. For example, people might perceive the same

relationship more distant when they criticize their partner than when

90
they ask opinions of their partner; and perceived level of face-threat
of criticizing the partner may be lower when the power has less power.
Thus, in order to control the effects of positive and negative
face-threats on perceived levels of relational distance and power
disparity, subjects rated four relational variations of the same act
(i.e., the levels of positive and negative face threat were set
constant); and in order to control the effects of the nature of the
relationship on perceived levels of positive and negative threats,
subjects rated four act-type variations of the same relationship (i.e.,
the levels of relational distance and power disparity were set
constant).

One hundred seventy six subjects volunteered to participate in
this pilot study. Subjects were asked to read one of the four
situational variations (i.e., high distance/low power, high
distance/equal power, low distance/low power, and low distance/equal
power) of the act asking the opinion of another member and rate the
relational distance between them and the hypothetical actor on a
7-point scale (6 being "very distant" and 0 "very close") and also rate
the power the bearer has over them also on a 7-point scale (0 being
"John Brown has much less power than you" and 6 being "both of you have
equal power). If the manipulation of relational variables was done
successfully, readers of the "low distance" instructions should rate
the relational distance lower than readers of the "high distance"
instructions; readers of the "equal power" instructions should rate
their hearer less powerful than readers of the "low power"

instructions.

Table 13.

Descriptive Statistics for Rated Distance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Relational Power
Equal Hearer Low

Close M = 2.39 M = 2.36 M = 2.38
(Low) SD = 1.17 SD = 1.14 SD = 1.15

Relational n = 39 n = 37 n = 76
Distance Distant M = 4.77 M = 4.32 M = 4.55
(High) SD = .89 SD = 1.18 SD = 1.06

n = 44 n = 44 n = 88
M = 3.58 M = 3.34 M = 3.46
SD = 1.58 SD = 1.52 SD = 1.55

n = 88 n = 88 N = 176
Table 14. Descriptive Statistics for Rated Power Disparity

Relational Power
Equal Hearer Low

Close M = 3.82 M = 3.61 M = 3.72
(Low) SD = 2.00 SD = 1.50 SD = 1.76

Relational n = 39 n = 37 n = 76
Distance Distant M = 4.00 M = 3.02 M = 3.51
(High) SD = 1.74 SD = 1.85 SD = 1.85

n = 44 n = 44 n = 88
M = 3.91 M = 3.32 M = 3.61
SD = 1.87 SD = 1.70 SD = 1.80

n = 88 n = 88 N = 176

 

 

 

 

92

Descriptive statistics for the rated social distance and the rated
power disparity are shown in Tables 13 and 14, respectively. Two
two-way analyses of variance yielded a significant effect of relational
distance on the rated relational distance (E=171.50, et=lll72, p<.001,
eta2=.50) and a significant effect of power disparity on the rated
power disparity (2:4.84, gt=1/172, p<.029, eta2=.03). Specifically,
subjects in the distant condition rated the relationship more distant
(M=4.55, SD=1.06) than subjects in the close condition (M=2.38,
SD=1.15); subjects in the hearer-low condition rated the bearer less
powerful (M=3.32, SD=1.70) than subjects in the equal power condition
(M=3.91, SD=1.87). Other main effects and interactions effects were
not statistically significant. Tests of homogeneity of variance showed
that variances of both the rated distance (Cochran’s C=.29, p<.851;
Bartlett-Box E=l.44, p<.230) and the rated power disparity (Cochran’s
C=.32, p<.333; Bartlett-Box 2:1.24, p<.292) were homogeneous across the
four different conditions.

Subjects also read the four act-type variations (i.e., asking
opinion, negative evaluation, request of re-writing, and request of
taking over another’s responsibility) of the low distance/equal power
condition and rated positive and negative face-threats carried by the
acts on seven-point scales (6 representing extremely high face-threat
and 0 no face-threat). If the manipulation of face-threat was done
effectively, "request of re-writing" and "negative evaluation," which
represent the high positive face-threat condition, should be rated
higher for positive face-threat than the other two acts and "request of

re-writing" and "request of taking over another’s responsibility,"

93

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 15. Descriptive Statistics for Rated Positive Face-Threat
Positive Face-Threat
Low High
M = .86 M = 2.55 M = 1.70
Low SD = 1.27 SD = 1.42 SD = 1.58
Relational n = 39 n = 37 n = 76
Face-Threat M = .89 M = 3.11 M = 2.00
High SD = 1.32 SD = 1.71 SD = 1.89
n = 44 n = 44 n = 88
M = .88 M = 2.83 M = 1.85
SD = 1.28 SD = 1.59 SD = 1.74
n = 88 n = 88 N = 176
Table 16. Descriptive Statistics for Rated Negative Face-Threat
Positive Face-Threat
Low High
M = 2.82 M = 3.05 M = 2.93
Low SD = 1.66 SD = 1.68 SD = 1.67
Relational n = 39 n = 37 n = 76
Face-Threat M = 4.84 M = 4.64 M = 4.74
High SD = 1.31 SD = 1.26 SD = 1.28
n = 44 n = 44 n = 88
M = 3.83 M = 3.84 M = 3.84
SD = 1.80 SD = 1.68 SD = 1.74
n = 88 n = 88 N = 176

 

 

 

 

94

which represent the high negative face-threat condition, should be
rated higher for negative face-threat than the other two acts.

Descriptive statistics for the rated positive face-threat and the
rated negative face-threat are presented in Tables 15 and 16,
respectively. Two two-way analyses of variance showed a significant
effect of positive face-threat on the rated positive face-threat
(F=81.00, gi=1/172. p<.001, eta2=.32) and a significant effect of
negative face-threat on the rated negative face-threat (£=64.59,
ei=lll72, p<.001, eta2=.27). Subjects in the low positive face-threat
condition rated the positive face-threat lower (M=.88, SD=1.28) than
subjects in the high positive face-threat condition (M=2.83, SD=1.59);
subjects in the low negative face-threat condition rated the negative
face-threat lower (M=2.93, SD=1.67) than subjects in the high negative
face-threat condition (M=4.74, SD=1.28). No other main effects or
interaction effects were statistically significant. Tests of
homogeneity of variance showed that variances of both the rated
positive face-threat (Cochran’s C=.35, p<.069; Bartlett-Box 2:1.62,
p<.183) and the rated negative face-threat (Cochran’s C=.32, p<.303;
Bartlett-Box E=1.97, p<.117) were homogeneous across the four different
conditions. These results altogether indicated that the instructions
could manipulate the relational variables successfully.
Participants and Procedure

Participants were 400 volunteers from several undergraduate
courses at a large Midwestern university. The participants received
extra credit points for completing the experiment. Participants were

randomly assigned to one of the sixteen experimental conditions in a 2

95

(high vs. low positive face-threat) X 2 (high vs. low negative
face-threat) X 2 (equal vs. low power disparity) X 2 (high vs. low
relational distance) factorial design, and asked to write verbatim what
they would say to perform the given act in the given relationship.
Participants were also asked to complete measurement scales for verbal
aggressiveness and empathy. All questionnaires were administered in
class.
lpstrpmentation

The same measures as in Experiment 1 were used to measure verbal
aggressiveness and empathy. A confirmatory factor analysis (Package by
Hunter R Cohen, 1969) for Infante and Wigley’s (1986) 20-item verbal
aggressiveness measure showed that the items in this measurement scale
were internally consistent. The deviations between observed
correlations and reproduced correlations were minute. Only one out of
190 observed correlations significantly deviated from its expected
correlation at p<.01; no deviation exceeded 3 times the standard error
(.095) of the mean correlation (.242); and no item had more than one
observed correlation that deviated significantly from its expected
correlation. The deviations of individual correlations from the mean
correlation were also quite small (i.e., the matrix was flat). Only
four out of 190 correlations deviated from the mean correlation at
p<.01; and no item had more than one correlation that significantly
deviated from the mean correlation. The reliability coefficient elpee
for this 20-item verbal aggressiveness measure was .86. These results
were highly consistent with the results from the validity and

reliability tests of Experiment 1.

96

A confirmatory factor analysis (Package) for Davis’ (1983) 7—item
empathy measure showed that one of these seven items (item #2 in Table
5) did not represent the same underlying dimension. The correlations
between this one item and other items (average correlation was .374)
were considerably lower than the mean correlation (.433). A
reliability analysis for this measure showed that the item-total
correlation for this item (.31) was also much lower than other
item-total correlations (the next lowest correlation was .45). Thus,
this item was excluded from further analyses. The reliability
coefficient elpee for the 6-item empathy measurement scale was .86.
These results also were highly consistent with the results from
Experiment 1.
M

The same measures for positive and negative politeness as in
Experiment 1 were employed. Three judges read utterances generated by
participants and rated positive and negative politeness. Of the twelve
items in the instrument (see Table 6), two items measuring the degree
of power-giving (items 7 and 8 in Table 6, i.e., "general deference"
and "frequency of using ’vous’ pronouns") were excluded from further
analyses since all the subjects scored zero for these categories.
Inter-rater reliability (Cureton’s reliability of average ratings) was
.82 for "general approval," .89 for "confidence of approval," .86 for
"exaggeratedness of approval," .85 for "general avoidance of
disapproval," .93 for "tentativeness of disapproval," .90 for
"understatedness of disapproval," .85 for "general avoidance of
imposition," .88 for "conventional indirectness," and .79 for

"hesitance of imposition."

97

Table 17. Confirmatory Factor Analysis of Positive Politeness
Apppoval Avoidance of Disepproval
General Confi- Exagger- General Tenta- Understa-

dence atedness tiveness tedness

General .972

Approval

Conf. of .933 .893

Approval

Exag. of .937 .888 .899

Approval

Avoid. of .249 .260 .151 .999

Disapproval

Tent. of .148 .147 .079 .772 .575

Disapproval

Under. of .284 .298 .202 .878 .610 .723

Disapproval

Factor 1 .988 .943 .946 .229 .130 .272

Factor 2 .261 .270 .166 .999 .749 .845

 

 

 

98

A confirmatory factor analysis using PACKAGE showed that the six
items measuring positive politeness (i.e., three items for approval and
three items for avoidance of disapproval) were not unidimensional. It
showed that items measuring approval and those measuring avoidance of
disapproval were forming two different clusters. Correlations between
items in the same cluster were very high (average r=.92 for items
measuring approval; average r=.75 for items measuring avoidance of
disapproval), whereas those between items in the different clusters
were very low (average r=.20) as Table 17 presents. The three items
for approval and the three items for avoidance of disapproval as well
as the three items for avoidance of imposition each seemed to be
unidimensional. The reliability coefficient elphe was .97 for the
three-item measurement of approval (or approach-based positive
politeness), .90 for the three-item measurement of avoidance of
disapproval (or avoidance-based positive politeness), and .82 for the
three-item measurement of avoidance of imposition (or avoidance-based
negative politeness).

Inter-rater reliability for global politeness was very low
(Cureton’s reliability of average ratings was .59). Similar to
Experiment 1, raters focused on different dimensions of politeness to
determine the degree of global politeness. One judge rated global
politeness more like approach-based positive politeness (r=.37) than
avoidance-based positive politeness (r=.23) or avoidance-based negative
politeness (r=.29); the other two judges rated global politeness more
like avoidance-based negative politeness (r=.40 and .39) than

approach-based (r=.34 and .23) or avoidance-based positive politeness

99
(r=.21 and .28). Because of low reliability, global politeness was

excluded from further analysis.

RESULTS
Manipulation Check
In order to check whether variations in the hypothetical

situations properly manipulated the relational variables and

face-threats, subjects were asked to rate the relational distance and
power disparity between them and the hypothetical actor and positive
and negative face-threats carried by the act before they wrote out the
message. Five-point Likert-type scales were employed to rate these
relational variables. For the measure of relational distance, "1"
represented "very close" and "5" represented "very distant." For the
measure of power disparity, "1" represented "John Brown has much more
power over you," "3" represented "Both of you have equal power," and "5
represented "You have much more power over John Brown." For the
measure of positive and negative face-threat, "1" represented "Not at
all" and "5" represented "Greatly." If the manipulation of relational
variables was done successfully, subjects in the "low distance"
conditions should rate the relational distance lower than subjects in
the "high distance" conditions; subjects in the "bearer-low power"
conditions should rate their hearer less powerful than subjects in the
"equal power" conditions; subjects in the "high positive face-threat"
conditions should rate the positive face-threat higher than subjects in
the "low positive face-threat" conditions; subjects in the "high
negative face-threat" conditions should rate the negative face-threat

higher than subjects in the "low negative face-threat" conditions.

Table 18.

100

Descriptive Statistics for Rated Distance

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N = 400 (n = 25 for each cell )
Negative Threat
Low High
Positive threat Positive threat
Low High Low High
E
q M= 2.00 M: 2.16 M: 2.04 M: 2.20 M=2.10
P u
C o a SD= .50 SD= .47 SD= .61 SD= .65 SD= .56
l w l
o e
s r H
D e M: 2.33 M: 2.08 M: 2.19 M: 2.17 M=2.19
i L
s o SD= .56 SD= .63 SD= .49 SD= .48 SD= .54
t w
a
n E
c D q M: 3.64 M: 3.76 M: 3.48 M= 3.76 M=3.66
e i P u
s o a SD= .57 SD= .60 SD= .71 SD= .72 SD= .65
t w l
a e
n r H
t M: 3.56 M: 3.92 M: 3.56 M= 3.64 M=3.67
L
o SD= .65 SD= .40 SD= .58 SD= .64 SD= .59
w
M: 2.89 M: 2.97 M: 2.81 M: 2.95 M=2.91
SD= .92 SD= 1.01 SD= .92 SD= .98 SD= .96

 

101

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 19. Descriptive Statistics for Rated Power Disparity
N = 400 (n = 25 for each cell)
Negative Threat
Low High
Positive threat Positive threat
Low High Low High
E
q = 2.96 M: 2.92 M: 2.92 M: 3.00 M=2.95
P u
C o a SD= .54 SD= .64 SD= .57 SD= .65 SD= .59
l w 1
o e
s r H
D e M: 3.54 M= 3.65 M: 3.42 M: 3.71 M=2.96
i L
s o SD= .51 SD= .69 SD= .86 SD= .62 SD= .60
t w
a
n E
c D q M: 3.00 M: 3.04 M: 2.92 M: 3.08 M=3.01
e i P u
s o a SD= .41 SD= .68 SD= .70 SD= .76 SD= .64
t w l
a e
n r H
t M: 3.72 M: 3.60 M: 3.48 M: 3.88 M=3.67
L
o SD= .61 SD= .82 SD= .82 SD= .53 SD= .71
W
M: 3.30 M: 3.31 M: 3.19 M: 3.41 M=3.30
SD= .61 SD= .77 SD= .78 SD= .74 SD= .73

 

102

 

 

 

 

 

 

 

 

 

 

 

 

Table 20. Descriptive Statistics for Rated Positive Face-Threat
N = 400 (n = 25 for each cell)
Negative Threat
Low High
Positive threat Positive threat
Low High Low High
E
q M: 1.16 M= 3.52 M: 1.12 M: 3.44 M=2.3l
P u
C o a SD= .37 SD= .77 SD= .33 SD= .87 SD=1.33
l w l
o e
s r H
D e M: 1.17 M: 3.53 M: 1.27 M: 3.54 M=2.38
i L
s o SD= .38 SD= .86 SD= .53 SD= .59 SD=1.32
t w
a
n E
c D q M: 1.40 M: 3.44 M: 1.08 M: 3.16 M=2.27
e i P u
s o a SD= .50 SD= .87 SD= .28 SD= .62 SD=1.20
t w l
a e
n r H
t M: 1.04 M: 3.60 M: 1.04 M: 3.40 =2.27
L
o SD= .20 SD= .87 SD= .20 SD= .65 SD=1.35
w
M: 1.19 M: 3.52 M= 1.13 M= 3.38 M=3.31
SD= .40 SD= .83 SD= .37 SD= .70 SD=1.30

 

 

 

 

 

 

Table 21.

103

Descriptive Statistics for Rated Negative Face-Threat

N = 400 (n = 25 for each cell)

 

Negative Threat

 

Low

High

 

Positive threat

Positive threat

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Low High Low High
E
q M: 2.08 M: 2.12 M: 3.80 M: 3.76 M=2.94
P u
C o a SD= .97 SD= .91 SD= .87 SD= 1.05 SD=1.28
l w l
o e
s r H
D e M: 2.25 M: 2.58 M: 4.08 M: 4.00 M=3.23
i L
s o SD= .85 SD= .95 SD= .89 SD= .88 SD=1.20
t w
a
n E
c D q M: 2.24 M: 2.52 M= 3.76 M: 3.76 M=3.07
e i P u
s o a SD= .78 SD= .92 SD= .83 SD= .97 SD=1.11
t w l
a e
n r H
t M: 2.28 M: 2.28 M: 3.92 M: 3.96 M=3.11
L
o SD= .98 SD= .84 SD= .76 SD= .79 SD=1.18
w
M: 2.21 M: 2.38 M: 3.89 M: 3.87 M=3.09
SD= .87 SD= .95 SD= .84 SD= .92 SD=1.20

 

104

Descriptive statistics for the rated relational distance, the
rated power disparity, the rated positive face-threat, and the rated
negative face-threat are presented in Tables l8, 19, 20, and 21,
respectively. A series of four-way (relational distance X power
disparity X positive face-threat X negative face-threat) analyses of
variance for the rated variables yielded significant main effects of
power disparity on the rated power disparity (2:94.90, et=ll384,
p<.001, eta2=.19), relational distance on the rated relational distance
(fi=672.57, gt=1/383, p<.001, eta2=.63), positive face-threat on the
rated positive face-threat (S=1437.61, et=ll384, p<.001, eta2=.78), and
negative face-threat on the rated negative face-threat (£:310.905,
et=1l384, p<.001, eta2=.46). Specifically, subjects in the bearer-low
condition rated the bearer less powerful (M=3.63, SD=.70) than subjects
in the equal power condition (M=2.98, SD=.62); subjects in the distant
condition rated the relationship more distant (M=3.67, SD=.62) than
subjects in the close condition (M=2.15, SD=.55); subjects in the low
positive face-threat condition rated the positive face-threat lower
(M=l.16, SD=.38) than subjects in the high positive face-threat
condition (M=3.46, SD=.77); subjects in the low negative face-threat
condition rated the negative face-threat lower (M=2.30, SD=.92) than
subjects in the high negative face-threat condition (M=3.88, SD=.88).
No other main effects or interaction effects turned out to be
significant.

A series of tests of homogeneity revealed that the variances of
the sixteen conditions were homogeneous for the rated distance
(Cochran’s C=.10, p<.679; Bartlett-Box S=1.18, p<.279) and the rated

negative face-threat (Cochran’s C=.09, p<.999; Bartlett-Box 2:.47,

105
p<.958); however, the variances were heterogeneous for the rated power
(Bartlett-Box £=l.76, p<.034; Cochran’s C=.ll, p<.258) and the rated
positive face-threat (Cochran’s C=.13, p<.014; Bartlett-Box E=9.73,
p<.001). Eight Kruskal-Wallis one-way analyses of variance were
performed to test the effects of relational distance, power disparity,

and positive and negative face-threats on the rated power disparity and
the rated positive face-threat. The results from these analyses were

not different from the results from the four-way ANOVAs for the rated
power and the rated positive face-threat. These results indicate that
the manipulation of the independent variables was done successfully.

Co-occurrence of Poeitive epd Negative Eolitenese

One of the basic claims of the present model is that speakers

 

would try to maintain both positive and negative politeness. In order
to test this claim, the frequencies of co-occurrence of positive and
negative politeness, only positive politeness, only negative
politeness, and non-occurrence of any type of politeness were counted.
Two hundred eighty nine out of 400 subjects showed a certain degree of
both positive and negative politeness, 11 subjects used only positive
politeness; 87 subjects employed only negative politeness; and 13
subjects did not show any type of politeness. A chi-square test
revealed that the observed probability of co-occurrence of positive and
negative politeness was significantly different from chance probability
(chi-square=9.99, ei=1, p<.002). supporting the present model.
The_§ffecte of Personality Variables

It was hypothesized that empathy will be positively correlated

with positive and negative politeness (Hypothesis 2) and that verbal

106
aggressiveness will be negatively correlated with positive and negative
politeness (Hypothesis 3).

Apppeech-Beeed Poeitive Politeneee. A zero-order correlation
analysis for approach-based positive politeness (i.e., approval) and
independent variables showed that verbal aggressiveness and empathy had
a strong negative relationship (r=-.54, p<.001). However, as shown in
Table 22, neither verbal aggressiveness (r=-.10, p<.051) nor empathy
(r=.04, p<.330) yielded any significant relationship with approval.
Regression of approval on the independent variables (including
relational distance, power disparity, positive and negative
face-threats, verbal aggressiveness, and empathy) showed that
personality variables were not good predictors of approach-based
positive politeness. The standardized regression coefficient eete was
-.10 (t=-1.78, ei=399, p<.075) for verbal aggressiveness and .001
(t=.02, et=399, p<.987) for empathy. These results were consistent
with the findings of Experiment 1 that personality variables did not
influence the level of approval shown in phatic communication.

Negative face-threat (beta=.006, t=-.l3, gt=399, p<.896) and power
disparity (beta=-.02, t=-.32, et=399, p<.750) also did not predict
approach-based positive politeness well. However, positive face-threat
(beta=.15, t=3.12, et=399, p<.002) and relational distance (beta=-.21,
t=-4.30, et=399, p<.001) showed stronger effects on approach-based
positive politeness. Multiple R2 for these two variables was .07
(2:13.91, gi=2/397, p<.001), and change in R2 by the other four
variables (i.e., empathy, verbal aggressiveness, negative face-threat,
and power disparity) was .01 (2:1.15, et=4l395, p<.334). In short,

personality variables did not show any significant influence on

107
approach-based positive politeness, which did not support Hypothesis 2
and Hypothesis 3.

Avoidence:Seeed Poeitive Politeneee. Empathy did not have any
significant impact on avoidance-based positive politeness (i.e.,
avoidance of disapproval). The Pearson product-moment correlation
coefficient for the relationship between empathy and avoidance of
disapproval was .08 (p<.109) as shown in Table 23. However, verbal
aggressiveness showed a statistically significant negative relationship
(r=-.17, p<.001) with avoidance of disapproval even though the effect
size was small. Regression of avoidance of disapproval on the
independent variables yielded an insignificant effect of empathy and a
significant effect of verbal aggressiveness on avoidance-based positive
politeness. The standardized regression coefficient eete was -.02
(t=-.44, ei=399, p<.659) for empathy and -.20 (t=-4.34, et=399, p<.001)
for verbal aggressiveness.

Positive face-threat (beta=.37, t=9.42, et=399, p<.001), power
disparity (beta=-.29, t=-7.37, et=399, p<.001), and relational distance
(beta=-.38, t=-9.68, et=399, p<.001) showed very strong effects on
avoidance-based positive politeness. Multiple R2 for these three
variables was .36 (2:72.78, Q£=3/396. p<.001), and change in multiple
R2 by verbal aggressiveness was .04 (2:23.67, et=ll398, p<.001).
Negative face-threat did not show any significant effect on
avoidance-based positive politeness (beta=.06, t=1.54, et=399, p<.124).
In short, empathy did not influence avoidance-based positive
politeness; that is, Hypothesis 2 was not supported. Verbal
aggressiveness showed a moderate effect on avoidance-based positive

politeness, partially supporting Hypothesis 3.

Table 22.

Correlation Analysis

Relational Power
Distance Disp.

Power .000
Disparity

Negative .000 .000
Threat

Positive .000 .000
Threat

Verbal -.023 -.046
Aggressiveness

Empathy .036 -.021
Gender -.125 .021
Approach-based

Positive -.206 -.011

Politeness

108

for Approach—Based Positive Politeness

Negative Positive Verbal Empathy Gender
Threat Threat Aggre.

O 000

.000 .000

0000 0000 '0 536

.000 .000 -.278 .175

-.008 .151 -.098 .045 -.007

109

Table 23. Correlation Analysis for Avoidance-Based Positive Politeness

Relational Power Negative Positive Verbal Empathy Gender
Distance Disp. Threat Threat Aggre.

Power .000

Disparity

Negative .000 .000

Threat

Positive .000 .000 .000

Threat

verbal -0023 '0046 0000 0000

Aggressiveness

Gender -.125 .021 .000 .000 -.278 .175

Avoidance-Based
Positive -.376 -.280 .057 .369 -.169 .080 .095
Politeness

110

Correlation Analysis for Avoidance-Based Negative Politeness

Relational Power
Disp.

Table 24.

Distance
Power .000
Disparity
Negative .000
Threat
Positive .000
Threat
Verbal -.023
Aggressiveness
Empathy .036
Gender -.125
Avoidance-Based
Negative .145

Politeness

.000

.000

-so46

-0021

.021

-.040

Negative Positive Verbal Empathy Gender
Threat Threat Aggre.

.000

.000 .000

.000 .000 -.536

.000 .000 -.278 .175

.521 -.080 -.l40 .112 .097

111

Avoidance-Beeed Negative Politeness. A zero-order correlation

 

analysis for avoidance of imposition (or avoidance-based negative
politeness) yielded small but statistically significant effects of
empathy (r=.11, p<.025) and verbal aggressiveness (r=-.14, p<.005) as
Table 24 shows. However, the relationship between empathy and negative
politeness turned out to be accounted for by the other predictors.
Regression of avoidance of imposition on relational distance, power
disparity, positive and negative face-threats, verbal aggressiveness,
and empathy yielded no significant effect of empathy and a
statistically significant but small effect of verbal aggressiveness.

The standardized regression coefficient beta for empathy was .04

 

(t=.90, et=399, p<.369), while that for verbal aggressiveness was -.11
(t=-2.32, et=399, p<.021).

Positive face-threat (beta=-.07, t=-1.80, et=399, p<.072) and
power disparity (beta=.-.04, t=-l.06, et=399, p<.29l) also did not show
any statistically significant effect on avoidance-based negative
politeness. In contrast, negative face-threat (beta=.52, t=12.50,
et=399, p<.001) and relational distance (beta=.14, t=3.39, et=399,
p<.001) turned out to be very good predictors of avoidance-based
negative politeness. R2 for these two variables was .29 (2:81.98,
.Q£=2/397. p<.001), and change in R2 by the other four variables was .03
(E=4.00, et=4l395, p<.004). In short, while empathy did not influence
avoidance-based negative politeness, verbal aggressiveness showed a
moderate influence on avoidance-based negative politeness. Thus,
Hypothesis 2 was not supported, and Hypothesis 3 was partially

supported.

112

In conclusion, empathy did not show any significant relationship
with any of the two types of positive politeness or negative
politeness. Thus, Hypothesis 2 predicting a positive relationship
between empathy and the levels of positive and negative politeness
manifested in performing an FTA was rejected. Verbal aggressiveness
showed statistically significant effects on both types of
avoidance-based politeness (i.e., avoidance of disapproval and
avoidance of imposition). However, the effect size was very small in
both cases. Furthermore, approach-based positive politeness, i.e.,
approval, was not influenced by the speaker’s verbal aggressiveness.
Thus, Hypothesis 3, that predicted verbal aggressiveness would be
negatively correlated with the level of positive politeness and the
level of negative politeness manifested in performing an FTA, was only
minimally supported.
The Effects of Fees-Threats and Relational Variables

Test of Homogeneity of Variance. Since verbal aggressiveness had
a significant effect on avoidance-based positive and negative
politeness (i.e., avoidance of disapproval and avoidance of
imposition), the effect was statistically removed from these dependent
variables by using residuals from the regression of them on verbal
aggressiveness instead of original scores for further analysis.
However, because no personality variable had any significant influence
on approach-based positive politeness (i.e., approval), original scores
were used for this particular variable. The effect of empathy was not
removed from any dependent variable since this personality variable did
not have any statistically significant influence on the dependent

variables.

113

Table 25. Means and SDs for Approach-Based Positive Politeness

N = 400 (n = 25 for each cell)

 

Negative Threat

 

 

 

 

 

 

OOSOfi‘OH-U

 

 

 

 

 

Low High
Positive threat Positive threat
Low High Low High

E

q M: 1.64 M: 6.60 M: 6.04 M: 4.20 M=4.62
P u

C o a SD= 1.94 SD= 7.24 SD= 6.72 SD= 6.74 SD=6.27
l w l

o e

s r H

e M: 1.71 M: 5.96 M: 6.42 M: 2.58 M=4.25
L

o SD= 1.58 SD= 6.40 SD= 6.39 SD= 4.75 SD=5.55
w
E

D q M: .32 M= 5.16 M: 1.96 M: 1.40 M=2.21
i P u

s o a SD= .90 SD= 3.60 SD= 2.96 SD= 2.92 SD=3.29
t w l

a e

n r H

t M: .32 M: 5.52 M= 2.00 M: 1.68 M=2.38
L

o SD= .90 SD= 6.26 SD= 2.72 SD= 4.11 SD=4.40
w

M: .99 M: 5.81 M: 4.21 M: 2.46 M=3.39

SD= 1.54 SD= 5.97 SD= 5.39 SD= 4.89 SD=5.10

 

 

 

 

 

 

114

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 26. Means and SDs for Avoidance-Based Positive Politeness *
N = 400 (n = 25 for each cell)
Negative Threat
Low High
Positive threat Positive threat
Low High Low High
E
q M: 1.17 M: 4.56 M= .96 M: 4.61 M=2.83
P u
C o a SD= 4.45 SD= 5.03 SD= 4.13 SD= 4.92 SD=4.91
l w l
o e
s r H
D e "=’2018 M: 1081 "3-1.04 M: 2045 M: 027
i L
s o SD= 1.25 SD= 3.47 SD= 1.92 SD= 3.69 SD=3.34
t w
a
n E
c D q M=-2.13 M: .38 M=-2.02 M: 1.83 =-.49
e i P u
s o a SD= 1.30 SD= 3.60 SD= 1.28 SD= 3.85 SD=3.22
t w l
a e
n r H
t =-3.27 M=-2.31 M=-3.47 M=-1.31 =-2.59
L
o SD= 1.05 SD= 2.18 SD= .57 SD= 2.53 SD=1.95
w
”3-1.60 M: 1012 n=-1039 "3 1089 M: 000
SD= 2.95 SD= 4.42 SD= 2.85 SD= 4.35 SD=4.00

 

* Scores are residuals from the regression of avoidance of
disapproval on verbal aggressiveness

115

Table 27. Means and SDs for Avoidance-Based Negative Politeness *

N = 400 (n = 25 for each cell)

 

Negativ

e Threat

 

Low

High

 

Positive threat

Positive threat

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Low High Low High
E
q “=-2026 "=-3o42 M: 2046 M: 090 "=-058
P u
C o a SD= 4.65 SD= 1.71 SD= 3.68 SD= 3.50 SD=4.22
l w l
0 e
s r H
D e M=-2030 "=-3018 M: 1035 M: 1058 M=-065
i L
s o SD= 4.25 SD= 1.54 SD= 4.00 SD= 5.58 SD=4.55
t w
a
n E
C D q M=-1006 ”=-1045 M: 3099 M: 2062 "=1002
e i P u
8 0 a SD= 4.46 SD= 1.28 SD= 2.93 SD= 4.65 SD=4.25
t w l
8 e
n r H
t =‘2057 “=-2012 M: 3028 N: 2059 M: 029
L
o SD= 4.63 SD= 1.07 SD= 4.27 SD= 3.72 SD=4.51
W
”=-2005 "=T'2055 M: 2076 M: 1092 M: .00
SD= 4.47 SD= 1.61 SD= 3.83 SD= 4.42 SD=4.42

 

* Scores are residuals from the regre
imposition on verbal aggressiveness

ssion of avoidance of

116

In order to examine the effects of face-threats and relational
variables on approach-based and avoidance-based positive and
avoidance-based negative politeness, four-way ANOVAs were conducted on
the resultant 16 cells obtained when crossing 2 levels of positive
face-threat, 2 levels of negative politeness, 2 levels of relational
distance, and 2 levels of power disparity. Descriptive statistics for
approach-based and avoidance-based positive politeness and
avoidance-based negative politeness are presented in Tables 25, 26, and
27, respectively.

Tests for homogeneity of variance showed that the variances of the
sixteen conditions were not homogeneous for all three dependent
variables. Cochran’s C was .15 (p<.001) for approach-based positive
politeness, .16 (p<.001) for avoidance-based positive politeness, and
.14 (p<.004) for avoidance-based negative politeness; Bartlett-Box E
was 15.85 (p<.001) for approach-based positive politeness, 14.48
(p<.001) for avoidance-based positive politeness, and 8.87 (p<.001) for
avoidance-based negative politeness. These results indicate that the
data for both positive and negative politeness violated the homogeneity
of variance assumption of ANOVA. Thus, a Kruskal-Wallis one-way
analysis of variance for ordinal scales was used as a main technique to
analyze the effects of face-threats and the relational variables on
politeness.

It was hypothesized that positive face-threat will increase the
level of positive politeness and negative face-threat will increase the
level of negative politeness manifested in performing the FTA
(Hypothesis 1). It was also hypothesized that as the power the bearer

has over the actor increases the level of both positive and negative

117
politeness will increase (Hypothesis 4). As for the influence of
social distance, it was hypothesized that as the social distance
increases the level of positive politeness will decrease and the level
of negative politeness will increase.

Approach-Based Positive Politeness. A series of Kruskal-Wallis
one-way analyses of variance yielded a significant effect of relational
distance (chi-square=9.40, et=1, p<.003) and positive face-threat
(chi-square=3.92, e1=1, p<.048) on approach-based positive politeness.
Specifically, subjects tended to give more approval when the bearer was
more intimate to them (mean rank=216.39) than when the hearer was less
intimate to them (mean rank=184.61) and when the act they were
performing carried more positive face-threat (mean rank=210.35) than
when the act carried less positive face-threat (mean rank=190.65). The
effects of negative face-threat (chi-square=1.30, et=1, p<.258) and
power disparity (chi-square=.025, gtsl, p<.875) on approach-based
positive politeness were statistically insignificant.

Since ANOVA is relatively robust to the violation of the
assumption of homogeneity of variance, a four-way analysis of variance
was performed for approach-based positive politeness. The four-way
ANOVA for approach-based positive politeness yielded significant main
effects of relational distance (2:20.10, d£=1/3B4, p<.001, eta2=.04)
and positive face-threat (2:10.80, gi=1/384, p<.001, eta2=.02). In
other words, people tend to give more approval to those who are closer
to them (M=4.44, SD=5.91) than those who are more distant (M=2.34,
SD=3.86) or when positive face-threat is higher (M=4.16, SD=5.70) than
when positive face-threat is lower (M=2.62, SD=4.29). The main effects

of negative face-threat (2:.02, et=1/384, p<.890) and power disparity

118
(§=.06, et=1/384, p<.814) on approach-based positive politeness were
insignificant. These results are very consistent with the results from
the Kruskal-Wallis one-way analyses of variance for approach-based
positive politeness.

One unexpected two-way interaction did occur; that is, the
interaction between positive face-threat and negative face-threat was
significant (2:48.61, et=1/384, p<.001, eta2=.10). In other words, the
effect of positive face-threat on approval varied depending on the
level of negative face-threat. Specifically, when negative face-threat
was low, approval was higher when positive face-threat was high
(M=5.81, SD=5.97) than when positive face-threat was low (M=.99,
SD=1.54); when negative face-threat was high, approval was higher when
positive face-threat was low (M=4.21, SD=5.39) than when positive
face-threat was high (M=2.46, SD=4.89).

The effect of positive face-threat on approval was consistent with
the predictions of the present model when negative face-threat was low,
that is, when subjects did not need to ask a big favor of the hearers
(when performing "negative evaluation" or "asking opinion").
Specifically, subjects tended to give more positive comments on the
hearer’s work when they needed to criticize the work (i.e., when
performing "negative evaluation") than when they did not need to
criticize the work (i.e., when performing "asking opinion"). This
seems to be due to the fact that the behavior of subjects was guided
mainly by the motivation to be polite. Other than this politeness
goal, they did not seem to have any particular goal.

However, inconsistent with the predictions of the present model,

when negative face-threat was high, that is, when subjects wanted to

119
ask a big favor of hearers (i.e., when performing "request of
re-writing" or "request of taking another’s responsibility"), they
tended to give many more compliments to the hearers when hearers did
not do anything wrong (i.e., when performing "request of taking
another’s responsibility") than when hearers did something wrong (i.e.,
when performing "request of re-writing"). Specifically, subjects
tended to emphasize that the bearer had abilities to do a good job or
that the bearer performed very well with his/her earlier job when they
do not have to criticize the hearer’s performances. However, when they
needed to criticize the hearer and ask a big favor of the hearer at the
same time, they did not use a lot of positive comments. Such a result
might stem from subjects being guided by the motivation to achieve
their request goal as well as the motivation to be polite. Subjects
might have thought that when the bearer did not do anything wrong,
encouraging the bearer to accept their request by approving the hearer
was a better way to achieve their request goal, while when the bearer
made some obvious mistakes, obligating the bearer to correct the
mistake by not approving the hearer very highly was a better way to
achieve their request goal. Other interactions were not statistically
significant.

In short, the results supported Hypothesis 1 that predicted
positive face-threat will increase the level of positive politeness and
Hypothesis 5 predicting that as the social distance between actors
increases the level of positive politeness will decrease. However,
Hypothesis 4, which predicted that as the power the hearer has over the
actor increases the level of positive politeness will increase was not

supported.

120

Avoidence:§eeed Positive Politeness. A series of Kruskal-Wallis
one-way analyses of variance yielded significant effects of relational
distance (chi-square=63.68, et=1, p<.001), power disparity
(chi-square=31.91, et=1, p<.001), and positive face-threat
(chi-square=53.72, et=1, p<.001) on avoidance-based positive
politeness. Subjects made more effort to avoid disapproving of the
bearer when the bearer was more intimate to them (mean rank=246.62)
than when the bearer was less intimate to them (mean rank=154.38), when
the bearer had more power over them (mean rank=233.15) then when the
bearer had less power (mean rank=167.85), and when the act they were
performing carried more threat to the hearer’s desire to be approved
(mean rank=242.86) than when the act carried less threat to the
hearer’s desire to be approved (mean rank=158.14). The effect of
negative face-threat (chi-square=2.13, et=1, p<.144) was statistically
insignificant.

A four-way ANOVA (for interval variables) for avoidance-based
positive politeness showed significant main effects of relational
distance (F=94.50, gt=1l384, p<.001, eta2=.15), power disparity
(E=53.95, et=1l384, p<.001, eta2=.08), and positive face-threat
(3:89.05, gt=1/3a4, p<.001, eta2=.14) and an insignificant effect of
negative face-threat (2:2.38, et=ll384, p<.123). Specifically
speaking, the speaker tended to employ more avoidance-based positive
politeness when the relational distance was low (M=1.55, SD=4.38) than
when the relational distance was high (M=-1.54, SD=2.86), when the
hearer’s power was high (M=1.17, SD=4.46) than when the hearer’s power
was low (M=-1.16, SD=3.08), and when positive face-threat was high

(M=1.50, SD=4.39) than when positive face-threat was low (M=-1.49,

121
SD=2.89). The level of avoidance-based positive politeness did not
vary as a function of negative face-threat. In other words, Brown and
Levinson’s claim that people use positive politeness to mitigate
negative face-threat was not supported. These results from the
four-way ANOVA were very consistent with the results from
Kruskal-Wallis one-way analysis of variance. No interaction effect was
statistically significant.

In short, the results supported Hypothesis 1 that predicted as
positive face-threat increases the level of positive politeness will
increase and Hypothesis 5 predicting that as the social distance
between actors increases the level of positive politeness will
decrease. In addition, different from approach-based positive
politeness, avoidance-based positive politeness was influenced by power
disparity. Thus, Hypothesis 4, which predicted that as the power the
bearer has over the actor increases the level of positive politeness
will increase, was supported.

Negetive Politeneee. A series of Kruskal-Wallis one-way analyses
of variance showed significant effects of relational distance
(chi-square=9.58, et=1, p<.002) and negative face-threat
(chi-square=130.35, et=1, p<.001). Specifically, subjects made more
effort to avoid imposing on the hearer when the hearer was less
intimate to them (mean rank=218.39) than when the hearer was more
intimate to them (mean rank=182.61) and when the act they were
performing carried more threat to the hearer’s desire for power (mean
rank=266.49) than when the act carried less threat to the hearer’s

desire for power (mean rank=l34.51). The effects of power disparity

122
(chi-square=1.61, et=1, p<.204) and positive face-threat
(chi-square=2.42, ei=1, p<.120) were insignificant.

A four-way ANOVA for avoidance-based negative politeness yielded
significant main effects of relational distance (2:11.61, gt=1l384,
p<.001, eta2=.02) and negative face-threat ($154.32, eta/384, _13<.001,
eta2=.28). In other words, actors tended to show more negative
politeness to distant persons (M=.66, SD=4.39) than close persons
(M=-.61, SD=4.38) and when negative face-threat is high (M=2.35,
SD=4.14) than when negative face-threat is low (M=-2.30, SD=3.35). In
contrast, the main effect of positive face-threat was not significant
(E=3.21, Q£=1/334. p<.074). In other words, the level of
avoidance-based negative politeness did not vary as a function of
positive face-threat; that is, Brown and Levinson’s claim that negative
politeness is employed when positive face-threat is high was not
supported. The main effect of power disparity was also statistically
insignificant (E=1.16, et=ll384, p<.282). These results were the same
as the results yielded by Kruskal-Wallis one-way analysis of variance.
All interaction effects were not statistically significant. In short,
the results supported Hypothesis 1 that predicted negative face-threat
will increase the level of negative politeness and Hypothesis 5
predicting that as the social distance between actors increases the
level of negative politeness will increase. However, Hypothesis 4,
which predicted that as the power the hearer has over the actor
increases the level of negative politeness will increase, was not
supported.

In sum, the results generally indicated that as the level of

positive face-threat increased, the level of both approach-based

123

positive politeness and avoidance-based positive politeness increased
and that as the level of negative face-threat increased, the level of
avoidance-based politeness increased. These findings altogether
supported Hypothesis 1 stating that the threat to positive face carried
by an FTA will increase the level of positive politeness manifested in
performing the FTA, and the threat to negative face carried by the FTA
will increase the level of negative politeness manifested in performing
the FTA. The results also indicated that as social distance increased,
both approach—based and avoidance-based positive politeness decreased
and avoidance-based negative politeness increased. The findings
support Hypothesis 5 predicting a negative influence of social distance
on the level of positive politeness manifested in performing an FTA and
a positive influence of social distance on the level of negative
politeness manifested in performing the FTA. The findings concerning
the influence of power disparity on positive and negative politeness is
not conclusive. Power disparity influenced avoidance-based positive
politeness significantly, but did not show any impact on approach-based
positive politeness and avoidance-based negative politeness. Thus,
Hypothesis 4 stating the power the bearer has over the actor will
increase the level of positive politeness and the level of negative
politeness manifested in performing an FTA was not supported.
TheeSffect of Gender

As Table 23 shows, gender yielded a moderate negative relationship
with verbal aggression (r=-.28, p<.001) and a moderate positive
relationship with empathy (r=.18, p<.001). A regression analysis of
gender on verbal aggression and empathy yielded a squared multiple

regression coefficient of .08 (F=16.80, Q£=2/397. p<.001). Verbal

124
aggression was a better predictor of gender (beta=-.26, t=-4.52,
ef=399, p<.001) than empathy (beta=.04, t=.65, ef=399, p<.518). A
one-way analysis of variance also showed that verbal aggressiveness was
significantly different between different genders (2:33.23, d£=1/398,
p<.001, eta2=.08). Males (M=40.53, SD=12.46) were more verbally
aggressive than females (M=33.33, SD=12.01).

A one-way ANOVA for positive politeness showed that males and
females were not different in approach-based positive politeness
(3:.018, et=ll398, p<.894). The effects of gender on avoidance-based
positive politeness (£=3.62, et=ll398, p<.058) and on avoidance-based
negative politeness (F=3.75, di=1l398, p<.053) were very close to the
level of being statistically significant. However, since verbal
aggressiveness was not equivalently distributed between different
genders, a one-way analysis of covariance eliminating the effect of
verbal aggression was performed for positive politeness. The results
showed that gender was not a significant predictor of avoidance-based
positive politeness (£=1.02, et=ll397, p<.312) or avoidance-based
negative politeness (2:1.48, et=ll397, p<.225) any more. These
findings concerning gender were highly consistent with the findings of
Experiment 1. In short, gender was not a very good predictor of
politeness, and most of its effect on politeness could be accounted for

by verbal aggressiveness.

DISCUSSION
One of the major findings of this experiment is that peopleflwept__

to save or proeote both positive and negative face of their.hearers.

.r
..-

——._‘ -..o

In other words, speakers employ both positive and negative politeness

125
in the same message although the degrees of positive and negative
politeness are contingent on the relational variables and the levels of
face-threats. This finding supports one of the main claims of the
present model and indicates that Brown and Levinson’s model, which
conceptualizes positive and negative politenessas mutually exclusive
strategies, has a crucial limitation in explaining politeness

phenomena.

Another significant finding of this experiment is that positive
politeness is not a unidimensional but bidimensional construct.
Specifically, approach-based positive politeness (or approval) and
avoidance-based positive politeness (or avoidance of disapproval) form
different dimensions from each other. Even though the relationship
between approach-based negative politeness (or deference) and
avoidance-based negative politeness (avoidance of imposition) was not
examined, there is a high probability that these two types of negative
politeness will represent two different dimensions. Brown and Levinson
conceptualize politeness as a bidimensional construct; however, the
findings from this experiment indicate that politeness is at least a
three-dimensional, and possibly four-dimensional, construct.

In chapter one, while critiquing Brown and Levinson’s
conceptualization of politeness strategies that positive politeness is
an approach-based and negative politeness is an avoidance-based
strategy, the present model argues that both positive and negative
politeness can be either approach-based or avoidance-based, proposing
four different types of politeness. However, the present model did not

predict that these four types of politeness would form separate

126
dimensions. Thus, some minor changes in the present model seem to be
needed to accommodate these findings.

This experiment mostly confirmed the politeness model proposed in
Chapter 1. First, as predicted by the model, the relational distance
between a speaker and a bearer was a good predictor of approach-based
and avoidance-based positive politeness and negative politeness.
Speakers show more approach-based and avoidance-based positive
politeness and less negative politeness to persons who are closer to
them. This finding is consistent with the prediction made by the
present model and inconsistent with Brown and Levinson’s prediction.
Brown and Levinson predict that speakers will show less positive
politeness to those who are closer to them, which is exactly opposite
to what this experiment found.

Second, the level of positive face-threat was a good indicator of
the level of positive politeness and the level of negative face-threat
was a good indicator of negative politeness. For both approach-based
and avoidance-based positive politeness, as positive face-threat
increased, positive politeness increased; for avoidance-based negative
politeness, as negative face-threat increased, negative politeness
increased. These findings are consistent with the predictions of the
present model and inconsistent with Brown and Levinson’s model. Brown
and Levinson predict that regardless of the type of face-threat, as
face-threat increases, the level of negative politeness increases; as
face-threat decreases, the level of positive politeness increases.
Thus, according to Brown and Levinson, high positive face-threat can be
mitigated by negative politeness and low negative face-threat can be

mitigated by positive politeness. However, this experiment revealed

127
that in a normal situation, only positive face-threat affected the
level of positive politeness and only negative face-threat influenced
the level of negative politeness.

This study found that when negative face-threat was high (i.e.,

_e A Awn.‘

when asking a big favor) subjects used more approach-based positive
politeness (i.e., approval) when positive face-threat was low (i.e., 3
when no criticism is intended) than when positive face-threat was high
(i.e., when criticism is intended). In other words, when subjects

wanted to ask a big favor of a hearer, they approved the hearer more

when they did not need to criticize the hearer than when they needed to
criticize the bearer. As discussed earlier, one possible explanation

for this finding is that when people want to ask others to do them a

big favor their behavior is guided by the motivation to achieve their
request goal in addition to the obligation to be polite. One strategy
people often use when they need to gain compliance of others is the
"liking principle" (Cialdini, 1988) that expresses their approval of

the receivers.

In the high negative-face condition (i.e., when asking a big
favor), the subjects’ behavior was guided by a compliance-gaining goal;
and in order to increase the probability to gain compliance, the
subjects may have been motivated to utilize this liking principle.
However, when the subjects needed to criticize a bearer (i.e., when
positive face-threat was high) as well as gain compliance of the
bearer, they could not use the liking principle very often, since their
intention to criticize the bearer could not be clearly expressed if
they used the liking principle excessively. Consequently, the subject

who needed to criticize the bearer used the liking principle (i.e.,

128
expressed approval) less frequently than the subject who did not have
to criticize the hearer. In other words, when compliance-gaining is
the primary goal, subjects used more approach-based positive politeness
(i.e., liking) when positive face-threat was low than when positive
face-threat was high.

This finding that subjects showed a high degree of approval (i.e.,
approach-based positive politeness) when negative face-threat was high
and positive face-threat was low is inconsistent with Brown and
Levinson’s prediction. Brown and Levinson predict that speakers would
use more approval when negative face-threat is lower (i.e., when making
a relative small request). In contrast, this experiment found that
speakers used more approval when negative face-threat was high‘and
positive face-threat was low (i.e., asking a big favor) than when both
negative and positive face-threats were low (i.e., asking opinions).

The results concerning the effect of power disparity are not
conclusive. Power disparity was a good predictor of only
avoidance-based positive politeness. Speakers made more effort to
avoid disapproving of hearers with more relational power. However,
power disparity was not a good indicator of approach-based positive
politeness and avoidance-based negative politeness. One possible
explanation for these results is that power disparity was not
manipulated well. In this experiment, power disparity was manipulated
by describing both the speaker and the bearer as students taking the
same course (the equal power condition) or by describing the speaker as
a undergraduate teacher’s assistant of a class and the bearer as a
student taking the class (the bearer low condition). Although there

was a statistically significant difference in perceived power disparity

129
between these two conditions, it might not have been enough to make the
speaker to change their politeness behavior.

Another possible explanation is that power differences in this
country are not a deciding factor for people’s language use any more.
Brown and Gilman (1960) argue that power disparity once was a powerful
factor underlying people’s use of language. However, as the
egalitarian ideology expanded, power disparity began to lose its
influence and relational distance became a more powerful factor. Thus,
power disparity may not be a powerful predictor of people’s language
use in a highly egalitarian society such as the United States. In
other words, in this country, speakers do not tend to be concerned
about the power disparity unless the disparity is too huge to be
ignored. Given this, the power disparity between a student and an
undergraduate teacher’s assistant was not large enough to make subjects
take the disparity into consideration when they generated messages.

Consistent with Experiment 1, the effect of empathy on politeness
and the effect of verbal aggressiveness on approach-based positive
politeness were insignificant. Verbal aggressiveness, however, showed
significant effects on both avoidance-based politeness, i.e., avoidance
of disapproval and avoidance of imposition. In other words, verbal
aggressiveness seems to decrease the level of effort to avoid
threatening hearers’ positive and negative face, while it did not
influence use of approach-based tactics such as approval and deference.

Also consistent with Experiment 1, the effect of gender on
positive and negative politeness was trivial. While gender did not
influence use of approach-based positive politeness in performing an

FTA, it showed a marginally significant effect on avoidance-based

130
positive and negative politeness. However, as argued in Chapter 1,
this effect could be well accounted for by the actor’s verbal
aggressiveness. After removing the differences in verbal
aggressiveness from differences in gender, gender did not have any
significant influence on politeness.

In summary, this experiment found that people employed both
positive and negative politeness in the same message. Approach-based
positive politeness showed a positive relationship with positive
face-threat and a negative relationship with relational distance.
Avoidance-based positive politeness revealed positive relationships
with positive face-threat and power disparity and negative
relationships with relational distance and verbal aggressiveness.
Avoidance-based negative politeness had positive relationships with
negative face-threat and relational distance and a negative

relationship with verbal aggressiveness.

W

CHAPTER IV

DEVELOPMENT OF CONTENT CODING SYSTEM

STATEMENT OF PROBLEM
To date, several researchers (e.g., Clark a Schunk, 1980;
Ervin-Tripp, 1977; Francik a Clark, 1985; Gibbs, 1981, 1986; House a

Kasper, 1981; Kempter & Thissen, 1981) have proposed content egding

 

systems for politeness.‘ However, none of these coding systems are

-,nmmfl,, w-O"

 

comprehensive enough to code politeness behavior manifested in various
social interactions (Baxter, 1984; Craig et al., 1986; Shimanoff,
1977). Two major reasons exist for this absence of a comprehensive
categorical coding scheme. First, no theoretical framework
conceptualizing politeness adequately has guided the development of
past coding systems. In order to develop an adequate coding system for
a certain construct, the construct should be defined properly. To
date, however, most researchers defined politeness as "indirectness of

imposition,’ which is only a part of negative politeness, missing many

 

other important aspects of politeness such as approach-based and
avoidance-based positive politeness and approach-based negative
politeness. Thus, the existing coding systems include only those
linguistic devices that express certain degrees of "indirectness of
imposition."

Second, most researchers based their coding scheme on the
observation of one homogeneous group’s (usually, college students)
performance of a single act (typically, simple request such as asking a

piece of information). Different social groups may have different

131

--.-—.n—-.—

132
repertoires of politeness expressions and different types of acts might
be expressed in different sets of politeness devices. Thus, a coding
system composed of politeness devices identified from the observation
of one group/act would not be applied to a study of another group/act.
In short, to generate a comprehensive coding scheme, many different

types of acts performed by various social groups should be observed.

Since the present model provides a comprehensive definition of
politeness, the first problem discussed above (i.e., lack of proper
definition of politeness) can be easily resolved. However, it is very
difficult, if not impossible, to resolve the second problem (i.e., lack
of observations of multiple acts performed by multiple groups) in a few
studies, because it requires a series of observations to examine
politeness devices employed by different social groups in performing
different types of acts. The most practical solution for this problem
might be accumulating single studies that identify politeness devices
used by a few groups in performing a few acts and generate a content
coding system for the given data, and then developing a more general
content coding system by synthesizing these linguistic devices
identified in different studies.

Given this, the purpose of this chapter is to conduct a study that
will become a building block for developing a general content coding
system for politeness. Specifically, this chapter will identify
politeness devices that are used by the participants in Experiments 1
and 2 in performing five different acts (i.e., greeting-and-parting,
asking opinions of others, negative evaluation, request of re-writing,

and request of taking over another’s responsibility) and examine the

133
degree to which each device represents the relevant dimension of
politeness.
Review of Literature

As mentioned earlier, none of the existing content coding systems
for politeness include those linguistic devices that vary in positive
politeness or approach-based negative politeness; all of these coding
systems focus on a single dimension of negative politeness, namely,
avoidance-based negative politeness. Furthermore, these coding systems
do not cover all aspects of avoidance-based negative politeness; they
are concerned with only one aspect of avoidance-based negative
politeness, that is, "avoidance of imposition" and ignore the other
aspect of avoidance-based negative politeness, that is, "hesitance of
imposition."

Of these coding systems for "avoidance of imposition," perhaps the
most systematic one is a six-category coding system proposed by
Ervin-Tripp (1977). When listed from the most imposing (i.e., the
least polite) to the least imposing (i.e., the most polite) form, the
categories are (1) personal need or desire statement (i.e., stating
what the speaker needs; e.g., I need a match), (2) imperatives (i.e.,
direct order; e.g., Give me a match), (3) imbedded imperatives (i.e.,
question-asking with desired acts specified; e.g., Can you give me a
match?), (4) permission directives (i.e., asking permission; e.g., May
I have a match?), (5) question directives (i.e., question-asking with
desired acts omitted; e.g., Have you got a match?), and (6) hints
(i.e., requiring inferences; e.g., The matches are all gone).

Clark and Schunk (1980) elaborate on Ervin-Tripp’s "imbedded

imperatives" and subdivide this category into (1) obligation (i.e.,

“#-
‘hfi—‘fl

134

reminding the hearers of the obligation to grant the request; e.g.,
Shouldn’t you give me a match?), (2) commitment (i.e., asking the
hearers whether or not they will commit themselves to granting the
request; e.g., Will you tell me...?), (3) memory (asking the hearers
whether or not they can remember whether the speaker asked them earlier
for the object of the request; e.g., Have I already asked you...?), (4)
ability (giving the bearer the opportunity to deny the questioned
ability; e.g., Can you tell me...?), and (5) imposition (admitting that
the speaker is imposing on the bearer e.g., Would you mind...?).

"Personal need statement" is primarily directed downward to
subordinates and "imperative" is usually directed to subordinates or
familiar equals. "Ability" statements allow the hearers to avoid the
embarrassment of being asked a request they could not comply with,’
while "imposition" statements offer the bearer the authority to say
that the request imposes too much. A "question directive" is not in
the form of a request; rather, it is often identical with an
information question and misunderstanding is possible. A hint is
employed when the speakers can rely on shared rules in structured
situations in offices and classroooms, and on shared understanding of
habits and motives in living groups and families.

In short, the literature on politeness shows that there are ten

-‘——.

distinct linguistic devices that vary in the degree of "indirectness of

imposition." However, the literature on politeness does not provide

.. can...

any categorical coding scheme for other dimensions of politeness. In
order to develop content coding systems for various dimensions of
politeness, this chapter will design a study that identifies linguistic

devices that vary in approach-based and avoidance-based positive and

135
negative politeness and examines the extent to which these devices
represent the given dimension of politeness.

As argued in Chapter 1, approach-based positive, avoidance-based
positive, and avoidance-based negative politeness are each measured by
two different criteria and approach-based negative politeness is
measured by a single criterion. Specifically, the level of
approach-based positive politeness is determined by (1) confidence of
approval and (2) exaggeratedness of approval; the degree of
avoidance-based positive politeness is decided by (3) tentativeness of
disapproval and (4) understatedness of disapproval; the level of
approach-based negative politeness is decided by (5) the degree of
deference expressed in the use of address terms; and the level of
avoidance-based negative politeness is determined by (6) hesitance of
imposition and (7) indirectness of imposition. Thus, this study will
identify a set of politeness devices for each of these seven criteria

and examine the extent to which these devices satisfy each criterion.

METHODS

Coders read messages generated by participants in Experiments 1
and 2 and identified unique linguistic devices that expressed varying
degrees of the given politeness criteria. Through a collapsing
process, a set of politeness devices was generated for each of the
seven criteria of politeness. A group of judges rated the extent to
which each device satisfied the given criterion.
Sample Utterances

Messages analyzed in this study were 100 greeting-and-parting

scripts generated by participants in Experiment 1 and 400 utterances

136
generated by participants in Experiment 2. Included in the 400
utterances generated in Experiment 2 were 100 utterances asking
opinions of others, 100 negative evaluations, 100 requests of
re-writing, and 100 requests of taking over another’s responsibility.

ldentificatiop oi Unique Politeness Devices

For each of the seven criteria of politeness (i.e., confidence of

approval, exaggeratedness of approval, tentativeness of disapproval,
understatedness of disapproval, deference shown in the use of address
terms; hesistance of imposition, and indirectness of imposition), three
coders independently identified all politeness devices that more or
less satisfied the given politeness criterion, and made a list of
unique politeness expressions for the criterion by eliminating all
duplicates. Three different lists of unique expressions made by three
coders were combined; and again, duplicates were eliminated, which
generated a list of politeness devices for each of the seven criteria
of politeness. Then, the coders together read the list of devices for
each criterion and collapsed similar devices into broader categories.
When collapsing categories, coders used a very strict categorization
process in which only very similar politeness devices were collapsed
into broader categories and no collapsing was done when coders failed
to reach a consensus. Through these procedures, a set of unique
linguistic devices for each of the seven politeness criteria were
generated.
Betieg et Peliteness

Judges employed in this study were 110 volunteers from several
undergraduate courses at a large Midwestern university. The

participants received extra credit points for participating in the

137
experiment. For each criteria of politeness, judges were asked to read
the list of linguistic devices and rate the degrees to which the
devices satisfy the criteria on a ll-point scale ranging from 0 to 10.
The lower extreme on this scale (i.e., 0) denotated that the device did
not satisfy the criteria at all, whereas the higher extreme (i.e., 10)

signified that the device satisfied the criteria very highly.

RESULTS
Sonfidence of Apppoval
Six categories of politeness devices emerged for "confidence of
approval" as presented in Table 28. When listed from the most
confident to the least confident approval, these categories were (1)
absolute confidence, (2) realization, (3) belief, (4) inference, (5)

supposition, and (6) appearance. Linguistic devices categorized into

,__. _____.- .—

“rm-bx“.

"Absolute confidence" (e.g., It is evident...) showed an extremely high
degree of confidence, not allowing any chance that what was said might
be wrong. _Devices in "realization" (e.g., I realize) expressed that
the speakers recently got to know what they were saying. A "belief"
statement specified that what the speaker thought was generally true.
While an "inference" statement implied that what was said was one’s
personal judgment, a "supposition" implied what was said was simply a
guess. Devices in the "appearance" category expressed what was said
was based on a superficial observation.

Judges’ ratings for the six categories in this category system
revealed that the categories varied widely in the degrees to which they
manifested confidence of approval. The category "absolute confidence,"

which was rated the most confident, was very close to the upper limit

138

of the scale (i.e., "10"); the category "appearance,' which was judged
the least confident, was not very far from the lower limit of the scale
(i.e., "0"). This category system also showed fine gradations of the
means of the categories, implying people can make fine distinctions in
confidence of approval based on linguistic devices. Except for
"absolute confidence," the difference between any two neighboring means
was approximately 1.
Exaggeration of Approval

Eleven different categories were obtained for "exaggeration of
approval" as shown in Table 29. When listed from the most exaggerating
to the least exaggerating form, these categories were (1) fantastic,
(2) exceptional, (3) outstanding, (4) lovable (e.g., I love it), (5)
excellent, (6) great, (7) good, (8) impressive, (9) fine, (10) likable

(e.g., I like it), and (11) okay.

\‘qm

---1u-.. ‘1“

Most linguistic devices showing more or less exaggeration of

_ approval were adjectives modifying the nouns or pronouns that

"'" -w... ...—

represented the hearer’s performance or possession. Judges’ ratings of
exaggeration of approval revealed that these categories were not evenly
distributed. Except for the category "okay," all categories were rated
higher than 5 (i.e., the midpoint of the scale). In addition, the
means of the first six categories from the highest exaggeratedness of
approval (i.e., fantastic, exceptional, outstanding, lovable,
excellent, and great) did not vary a lot. These results suggest that
persons have few linguistic means of expressing mild levels of
approval; consequently, the use of "okay" seems to be a clear marker of

barely acceptable approval.

Table 28.
Mean SD Sategory
8.87/0.14 1) Absolute
Confidence
6.12/2.4l 2) Realization
5.45/2.64 3) Belief
4.23/3.7l 4) Inference
3.11/2.04 5) Supposition
2.03/2.26 6) Appearance

139

Content Coding Scheme for Confidence of Approval

ELIE-.2138.

It is evidence

I know

It is clear

I am sure

I have faith/confidence

I (can) see
I realize

I trust
believe
can tell

HH

think
find
notice
feel

HHHH

I suppose
I guess

It seems/looks/sounds
It seems/looks/sounds like
It seems/looks/sounds as though

Table 29.

Mean D

9.45/0.56
9.33/1.01
9.12/1.79
8.72/3.01
8.45/2.06

8.12/2.45

6.86/2.23

6.23/6.23

5.32/2.14

5.27/4.40

3.32/2.74

140

Content Coding Scheme for Exaggeratedness of Approval

Category

1) Fantastic
2) Exceptional

3) Outstanding
4) Lovable

5) Excellent

6) Great

7) Good
8) Impressive

9) Fine

10) Likable

11) Okay

Exam le

It is fantastic
It is exceptional
It is outstanding
I love it

It is excellent
It is the best

It is great

It is good
It is done well

I am impressed by
It is impressive

It is fine

It is better

It is clear

It is interesting
I like it

It is okay

141

Tentativeness of Disapproval

The collapsing process for the devices in "tentativeness of
disapproval" yielded ten different categories presented in Table 30.
When listed in the decreasing order of tentativeness of disapproval,
these categories were (1) probability, (2) appearance, (3)
understanding, (4) supposition, (5) apology, (6) inference, (7)
retreat, (8) clarification, (9) no offense, and (10) face defense. A
"probability" device implied that what was said has a chance to be
true. An "understanding" statement expressed that the speaker
understood why the hearer did not perform very well. An "apology"
showed that the speakers were regretful about their criticism of the
bearer, while a "retreat" admitted that the speakers were too critical.

With a "clarification,"

a speaker specified which part of the hearer’s
performance they did not want to criticize. A "no offense" device
expressed the speaker’s hope that the hearer would not be offended by
the criticism, and a "face defense" implied that the speaker was not an
undesirable person. As in the category system for "confidence of
approval," devices in "appearance" expressed what was said was based on
a superficial observation; a "supposition" implied what was said was
not based on facts; and an "inference" statement implied that what was
said was a personal judgment.

Judges’ ratings revealed that the ten categories varied widely in
the degrees to which they manifested tentativeness of disapproval. The
category "probability" was rated extremely tentative (M=9.42, SD=3.04)
and the category "face defense" was judged very low in tentativeness

(M=1.72, SD=2.12). In addition, the ten categories were distributed

relatively evenly over the continuum of tentativeness of disapproval,

142
implying that persons can use linguistic devices to state fairly
exactly their tentativeness of disapproval. Roughly, three different
types of linguistic devices appeared in this category: hedges (i.e.,
devices implying that what is said is true only in certain respects
such as devices in the category "probability"), disclaimers (i.e.,

devices implying that the speaker does not have an intention to harm

N N N N

the bearer such as the categories "retreat, clarification, no

offense,' and "face defense"), and subject + verbs (such as devices in

0' fl 0

"appearance, supposition,’ and "inference"). Among these three types
of devices, hedges were rated the most tentative and disclaimers were
rated the least tentative. It appeared that hedges were rated more
tentative because they implied the disapproval might not be true and
that disclaimers were rated less tentative since they were employed to

indicate that disapproval was coming up.

Understatedness ef Dieapproval

 

 

Seven categories of linguistic devices emerged for the criterion
of "understatedness of disapproval" as shown in Table 31. When listed
from the most understated to the least understated, the categories were
(1) off-topic, (2) not good enough, (3) unsteady, (4) not good, (5)
irrelevant, (6) not right, and (7) poor. Devices in the "off-topic"
category expressed what the bearer had done was different from what the
speaker had in mind. The "not good enough" category showed that what
the bearer had done might or might not be good but that it was not good
enough for the standard the speaker had in mind. Devices in the
"unsteady" category implied that the hearer’s work showed some
inconsistency in it, while the "not good" category expressed the

hearer’s performance was not good in an absolute sense. An

Table 30.

e SD

9.42/3.04

8.13/1.94

7.65/3.12

6.55/2.99

5.33/3.99
3.12/2.04

3.04/3.03

2.50/3.78

1.91/2.54

l.72/2.12

143

Content Coding Scheme for Tentativeness of Disapproval

Cetegopy
1) Probability

2) Appearance

3) Understanding

4) Supposition

5) Apology

6) Inference

7) Retreat

8) Clarification

9) No offense

10) Face Defense

Hassles

maybe

perhaps

might/may
possibly/by chance

It seems/looks/sounds
It seems/looks/sounds like
It seems/looks/sounds as though

I know you have been busy with
other things

I suppose
I guess

I am sorry to say this

think
feel
find
notice

HHHH

Maybe I am too critical

I am not saying you are a bad
writer

I hope you don’t take offense at
this

I don’t want to sound like a
real jerk

144

Table 31. Content Coding Scheme for Understatedness of Disapproval

MeenZSD Sategopy Examples
7.12/5.34 1) Off-Topic It is not what we need

It is not what we had in mind
It is not what the teacher wants

5.78/6.04 2) Not good It is not good enough
enough It is not strong enough
It is not sufficient enough

5.12/5.17 3) Unsteady It is shaky
It is off-base

3.12/4.26 4) Not good It is not clear
It is not good
It is not convincing
It is not strong
It is not well-researched

1.85/1.64 5) Irrelevant It is irrelevant to ...
It is unsupportive of ...

1.23/0.53 6) Not right It is not done right
It is not addressed right

0.12/0.14 7) Poor It is poor
It is of poor quality

145

"irrelevant" device showed that a part or all of the hearer’s work is
irrelevant to the hearer’s goal. A "not right" statement implied that
the hearer did the assignment wrong, while a "poor" statement bluntly
expressed that the quality of the hearer’s performance was poor.

Judges’ ratings of understatedness of disapproval showed that
these categories were not evenly distributed over the continuum of
understatedness of disapproval. The means of the categories ranged
from 0.12 to 7.12. In other words, the category system did not include
any category that showed a very high degree of understatedness,
implying it is very difficult to truly understate one’s disapproval.
Generally speaking, devices stating that the hearer’s performance or
possession was pet desirable (e.g., It is not good) were rated more
understated than devices saying that the hearer’s performance or
possession was endesipeele (e.g., It is poor). This finding seems to
be due to the fact that saying something is not good allows for it
being not bad, either.
He ' ance osit n

As can be seen in Table 32, nine categories of devices were
obtained for "hesitance of imposition." When listed in decreasing
order of hesitance of imposition, these categories were (1)
uncertainty, (2) probability, (3) appearance, (4) empathy, (5)
supposition, (6) favor, (7) willingness, (8) inference, and (9)
awareness. An "uncertainty" statement showed that the speakers were
uncertain about the hearer’s willingness or ability to do what they
would ask the bearer to do. An "empathy" statement showed that the
speaker knew the feelings of the hearer. With a "favor" statement, the

speakers indicated they would ask a favor of the hearer. A

Table 32.

Mean SD

8.89/3.71

8.58/3.43

7.98/2.98

6.38/4.44

6.12/3.14

5.92/2.89

3.81/4.02

3.12/2.04

2.77/3.75

146

Content Coding Scheme for Hesitance of Imposition

Category

1) Uncertainty

2) Probability

3) Appearance

4) Empathy

5) Supposition

6) Favor

7) Willingness

8) Inference

9) Awareness

Examples

I wonder
I was wondering
I was curious

maybe

perhaps

might/may
possibly/by chance

It seems/looks/sounds
It seems/looks/sounds like
It seems/looks/sounds as though

I know it’s asking a lot

I know this may be inconvenient

I know this will take a great part of
your time

You may be mad and feel it’s a lot of
work

I suppose
I guess

I have a real important/big favor to
ask of you

take/have time to do ...
try to do ...

consider doing ...

be willing to do ...

be interested in doing ...

I think
I feel

I find

I notice

Do you know that we need to re-do
John’s part

Table 33.
ean SD Category
9.34/0.14 1) Hint
9.03/1.08 2) Possibility
8.87/2.19 3) Permission
Directives
8.12/3.19 4) Imposition
7.78/2.33 5) Ability/
Commitment +
please
6.78/2.36 6) Self-
Assessment
6.23/5.03 7) Leading
Hesitance
5.78/2.19 8) Wish
5.34/4.27 9) Conditional
4.98/2.50 10) Ability
4.48/3.09 11) Commitment
4.12/2.76 12) Please +

Imperatives

147

Content Coding Scheme for Indirectness of Imposition

Examples

If you would like, I am more than
happy to help

Is there any way/possibility you
can/could do ...?

Can I ask you to do ...?
Could I ask you to do ...?
May I ask you to do...?

Would/do you mind doing ...?
Would it be too much to ask you ...?

Could you please do ...?
Would you please do ...?

Do you think you can do ...?
Do you think you could/would do ...?

Perhaps/maybe/I guess you can do ...
Perhaps/maybe/I guess you could do .

I hope/wish you can do ...
I hope/wish you could/would do ...

If you could/would do ..., I will
appreciate it

If you could/would do ..., it will
give you a better grade.

Can you do ...?
Could you do ...?

Will you do ...?
Would you do ...?
Would you like to do ...?

Please do ...

148

Table 33 (Cont’d)

2.33/2.00 13) Suggestion I suggest that you do ...
It would be a good idea to do ...
I would like ask/invite you to do ..
Why don’t you do ...?

1.21/1.69 l4) Imperatives Do it
0.99/1.27 15) Need Statement I need/would like/want you to do .
0.87/O.14 16) Obligation You should/have to/go it over

You need to go it over

It should/has to be ...

You owe it to us to do ...

It’s only fair if you do ...
"willingness" device asked the hearers’ willingness to do what the
speaker requested them to do, while an "awareness" device asked whether
or not the hearers were aware of the situation which the speaker was
in. The functions of the categories of "probability," "appearance,"

' and "inference" were the same as in "tentativeness of

"supposition,'
disapproval."

Judges’ ratings revealed that these categories were distributed
quite evenly on the continuum of hesitance of imposition; however, no
category represented either extreme of the continuum. Except for the
relatively large gap between the categories "favor" and "willingness,"
the means of the categories showed fine gradations. The results also
indicate that people have many ways of being hesistant so as not to
impose on others through inclusion of modifiers (e.g., "probability")
and disclaimers (e.g., "empathy").

Ipdirectnees of lppositien

As presented in Table 33, sixteen categories of politeness devices

emerged for "indirectness of imposition.’ This category system

149
included eight out of ten categories proposed by Ervin-Tripp and Clark
and Schunk (i.e., hint, permission directives, imposition, ability,
commitment, imperatives, need statement, and obligation). The category
of "memory" proposed by Clark and Schunk and "question directives"
proposed by Ervin-Tripp were not used in any of the 100
greeting-and-parting scripts generated in Experiment 1 or the 400
utterances generated in Experiment 2.

In addition to the eight categories proposed by Ervin-Tripp and
Clark and Schunk, the present study identified eight more frequently
used categories: suggestions, please + imperatives, conditionals,
wishes, leading hesitance, self-assessment, ability/commitment +
please, and possibilties. A "suggestion" (e.g., It would be a good
idea to do...) suggests or recommends the bearer to do the action
desired by the speaker. It was rated more indirect than imperative and
less indirect than "please + imperatives." A "please + imperative"
(e.g., please do...) was rated more indirect than suggestion and less
indirect than a "commitment." A "conditional" (e.g., If you could
do..., I will appreciate it) states that if the hearers could/would do
what the speakers want them to do, some positive outcomes will be
obtained. This form of request was rated more indirect than an
"ability" statement and less indirect than a "wish" statement.

A "wish" statement (e.g., I wish you can do...) simply states the
speaker’s wish. It was rated more indirect than a "conditional" and
less indirect than a "leading hesitance." A "leading hesitance" (e.g.,
Perhaps you can do...) shows a certain degree of hesitance, but it

tends to lead the hearers to doing what the speaker wants them to do.

150
It was rated more indirect than a "wish" statement and less indirect
than a "self—assessment." A "self-assessment" (e.g., Do you think you
can do...?) asks whether hearers think that they have abilities to do
what the speaker asks them to do. It was rated more indirect than a
"leading hesitance" and less indirect than an "ability/commitment +
please." An "ability or commitment + please" (e.g., Could/would you
please do...?) adds "please" to "ability" or "commitment." This form
of request was rated more indirect than a "self-assessment" and less
indirect than an "imposition." Finally, a "possibility" statement (Is
there any possibility you can do...?) asks whether or not there is any
chance that the hearers can do what the speaker asks them to do. It
was rated more indirect than a "permission directive" and less indirect
than a "hint."

When compared to other criteria of politeness, indirectedness of
imposition yielded many more categories. In other words, it appeared
that people had more politeness devices available when they want to
impose on others indirectly than when they want to approve or
disapprove of others. One reason for this finding might be that people
impose on others much more frequently than they approve or disapprove
of others. Thus, people have more chances to develop devices to impose
on others subtlely. Judges’ ratings showed that the categories in the
coding system were distributed very evenly from one extreme to the
other extreme of the continuum of indirectness of imposition. This
coding system also showed fine gradations of the means of the

categories.

151

Seteeence Shown in Uee of Address Terms

In both experiments, the types of relationships analyzed were
limited in number. In both Experiments 1 and 2, relational distance
and power disparity had only two different levels; that is, both
experiments employed four different types of relationships.
Consequently, the participants in the experiments did not use a variety
of address terms. Only four different types of address terms emerged
in Experiment 1: Mr. + last name, Dr. + last name, Professor + last
name, and first name. In Experiment 2, participants used only one type
of address term, i.e., first name. Because of the limited number of
different types of address terms, a categorical coding system was not

developed for this particular criterion.

DISCUSSION

This study developed a categorical coding system for each of six
criteria for positive and negative politeness (i.e., confidence of
approval and exaggeratedness of approval for approach-based positive
politeness; tentativeness of disapproval and understatedness of
disapproval for avoidance-based positive politeness; and hesistance of
imposition and indirectness of imposition for avoidance-based negative
politeness). The category system for "confidence of approval" showed
that the devices in the "absolute confidence" category (e.g., it is
evident) manifested the highest degree of confidence in approval; the
category system for "exaggeratedness of approval" revealed that the
category of "fantastic" approved the bearer the most exaggeratedly.
Thus, a combination of these devices, e.g., "It is evident that you did

a fantastic job," carries the highest degree of approach-based positive

152
politeness. In contrast, the "appearance" category (e.g., it seems
like) manifested the lowest degree of confidence of approval; the
"okay" category showed the lowest degree of exaggeratedness of
approval. Thus, a combination of these devices, e.g., "It seems like
you did O.K.," carries the lowest degree of approach-based positive
politeness.

For avoidance-based positive politeness, the category
"probability" (e.g., maybe) expressed the highest degree of
tentativeness of disapproval, while the category "off-topic" (e.g., it
is not what we need) manifested the highest degree of understatedness
of disapproval. In contrast, the category "face defense" (e.g., I
don’t want to sound like a real jerk) showed the lowest degree of
tentativeness of disapproval, while the category "poor" manifested the
lowest degree of understatedness of disapproval. Thus, a combination
of the categories of "probability" and "off-topic," e.g., "Maybe it is
not what we need," is much more polite than a combination of the
categories of "face defense" and "poor," e.g., "I don’t want to sound
like a real jerk, but you did a poor job."

For avoidance-based negative politeness, "uncertainty" devices
(e.g., I was wondering) expressed the highest degree of hesitance of
imposition and "possibility" devices (e.g., Is there any possibility
you could do it?) showed the highest degree of indirectness of
imposition except for a "hint." Thus, a combination of these two types
of devices, e.g., "I was wondering if there is any possibility you
could do it," expresses a very high degree of avoidance-based negative
politeness. The category "awareness" (e.g., Do you know that we need

to rewrite the first chapter?) showed the least hesitance in

153
imposition; the category "obligation" (e.g., You should do it) was the
most direct in imposition. Thus, a combination of these two
categories, e.g., "Do you know that we need to rewrite the first
chapter? You should do it," is very low in avoidance-based negative
politeness.

Three categories, i.e., "appearance," "supposition," and
"inference," appeared in three different coding systems, i.e.,
confidence of approval, tentativeness of disapproval, and hesitance of
imposition. Because the category "appearance" showed the least
confidence of the three categories, its mean score was the lowest of
the three for confidence of approval and the highest for tentativeness
of disapproval and hesitance of imposition; because the category
"inference" showed the most confidence of the three categories, its
mean score was the highest of the three for confidence of approval and
the lowest for tentativeness of disapproval and hesitance of
imposition. Consequently, "appearance" showed more intention to avoid
disapproval or imposition, while "inference" expressed more approval.
In other words, "appearance" was more polite when used with disapproval
or imposition than "inference"; and "inference" was more polite when
used with approval than "appearance."

When we compare the content coding system for exaggeration of
approval (Table 29) with that for understatedness of disapproval (Table
31), it is noticeable that all categories in the coding system for
understatedness of disapproval express less approval than any category
in the coding system for exaggeratedness of approval. "Off-topic"
(e.g., It’s not what the teacher wants), which is the most approving

category for understatedness of disapproval, does not show as much

154
approval as "okay" (e.g., It’s okay), which is the least approving
category for exaggeratedness of approval.
However, the fact that categories for exaggeratedness of approval
show more approval does not imply that they are more polite. The
devices in the category system of exaggeration of approval are employed

to approve hearers when they performed their job successfully, whereas

the devices in the category system of understatedness of disapproval
are used to avoid disapproving of hearers when they failed to perform
their job successfully. The politeness of a device in the category
system of understatedness of disapproval depends on the extent to which
it avoids disapproving of a hearer, whereas the politeness of a device
in the category system of exaggeratedness of approval depends on the
degree to which it approves a bearer. Thus, some devices in the
category system of understatedness of disapproval are more polite than
some devices in the category system of exaggeratedness of approval.

As mentioned earlier, the only criterion for politeness that has
any existing content coding system is "indirectness of imposition."
Thus, it is impossible to compare the coding systems developed by the
present study with other coding systems for such criteria as

N N

"confidence of approval, exaggeratedness of approval," "tentativeness
of disapproval," "understatedness of disapproval," and "hesitance of
imposition." For the criterion "indirectness of imposition," the
present study generated 16 categories of linguistic devices, whereas
Ervin-Tripp (1978) and Clark and Schunk (1980) together proposed 10
categories of linguistic devices. Two out of these 10 categories

(i.e., "memory" of Clark and Schunk and "question directive" of

Ervin-Tripp) did not emerge as unique categories and 8 unique

155
categories that are not included in either of Ervin-Tripp or Clark and
Schunk emerged in this study.
The reason for these differences between the present study and
previous studies seems to be that the acts analyzed in the present
study were different from the acts analyzed by Ervin-Tripp or Clark and

Schunk. Both Ervin-Tripp’s and Clark and Schunk’s work focused on a

simple request that asks for some information or an easily obtainable
object. Thus, a "memory" statement (e.g., Did I already ask you about
it?) or a "question directive" (e.g., Do you gotta match) turns out to
be frequently used by speakers. However, since the present study
focused on requests that asked the hearers to perform certain acts,
these categories were not employed at all; rather, some more active
categories such as "possibility" (e.g., Is there any way you could do
it?) or "self-assessment" (e.g., Do you think you can do it?) emerged
as major categories.

As mentioned earlier, the content coding systems developed in this
study are not exhaustive. The number of acts analyzed and the number
of social groups observed in this study were limited. Different social
groups might use different sets of politeness devices and the same
group of people use different sets of politeness devices when they
perform different types of acts. Thus, in order to make the coding
systems more comprehensive, categories developed based on some other
groups’ (e.g., older or younger generations than college students or
non-college students) performance of other important communicative acts
(e.g., disagreement or apology) should be added.

Although the category systems developed here are not exhaustive,

they are more comprehensive than any other category system. This study

156
analyzed more acts and observed more actors than any other study.
Specifically, this study analyzed five different acts (i.e.,
greeting-and-parting, request of re-writing, request of taking over
another’s responsibility, negative evaluation, and asking opinions of
others) performed by 500 participants. Thus, as shown in the
comparison between the present study and Ervin-Tripp’s and Clark and
Schunk’s studies, the present study pr0poses a much more comprehensive
category system than previous studies.

In conclusion, this study showed that people used various
linguistic devices to express politeness and were able to recognize the
differences in the degree of politeness among these devices. Thus, it
is clear that we can use category coding systems composed of linguistic
devices with varying degrees of politeness to analyze politeness
behavior. It is also clear that a relationship exists between
linguistic devices and politeness; that how we use language determines

how polite we will be judged to be.

CHAPTER V

GENERAL DISCUSSION AND CONCLUSION

GENERAL DISCUSSION
The Preeent Model ve. Brown and Levinson’s Model
There are five major differences between the present model and
Brown and Levinson’s model.I First, Brown and Levinson’s model accounts
for only one kind of face want at a time, whereas the present model
/flexplains both positive and negative face-threats at the same time.
{Eidecond, Brown and Levinson’s model limits politeness to the phenomena
in which speakers want to mitigate face-threat carried by
face-threatening acts, while the present model claims speakers strive
to be polite even when they perform non-face-threatening acts. Third,
Brown and Levinson predict that regardless of the type of face-threat,
speakers will employ positive politeness when face-threat is low and
negative politeness when face-threat is high. In contrast, the present
model predicts that speakers will increase positive politeness when
positive face-threat increases and increase negative politeness when
negative face-threat increases. Fourth, Brown and Levinson predict
that as relational distance increases, both positive and negative
politeness will decrease. The present model predicts that as
relational distance increases, positive politeness will increase and
negative politeness will decrease. Finally, Brown and Levinson’s model
does not include personality variables, whereas the present model

includes two personality variables, i.e., verbal aggressiveness and

empathy.

157

158

The results from the two experiments support the present model in
four out of these five contrasts. First, both experiments proved that
speakers attempted to be positively and negatively polite in the same
message. Experiment 1 revealed that both positive and negative
politeness were significantly different from zero, and Experiment 2
showed that most messages manifested both positive and negative
politeness. Second, Experiment 1 showed that speakers attempted to be
polite even though they did not perform any intrinsically
face-threatening act. Even in greeting and parting others, people were
compelled to promote others’ face wants.

Third, Experiment 2 showed that as negative face-threat increased,
the level of negative politeness increased; as positive face-threat
increased, the levels of approach-based and avoidance-based positive
politeness increased. However, except when a compliance-gaining goal
was more important than the politeness goal, negative face-threat did
not influence the level of positive politeness and positive face-threat
did not affect the level of negative politeness. Even when a
compliance-gaining goal was the primary goal, the way positive
politeness compensated for negative face-threat was different from what
Brown and Levinson predict. Brown and Levinson predict that more
positive politeness would be employed when negative face-threat is low.
In contrast, Experiment 2 found that subjects employed more positive
politeness when negative face-threat was high. This finding can be
explained in terms of a "solidarity move." When people think that
their relationship with hearers is too distant to make a highly
imposing request (i.e., act with high negative face-threat), they want

to make the relationship closer by behaving in the manner in which

159
persons in a close relationship behave. Consequently, people show more
positive politeness when they ask a big favor of an acquaintance than
when they ask a small request of an acquaintance.

Finally, both experiments demonstrated that as relational distance
increased, approach-based and avoidance-based positive politeness
decreased and negative politeness increased, which is opposite of Brown
and Levinson’s claim. For these four crucial reasons, the present
study rejects Brown and Levinson’s model and adopts the present model’s
perspective.

FTA ve. Non-FTA

Experiment I examined people’s politeness behavior in performing a
non-FTA and Experiment 2 examined people’s politeness behavior in
performing FTAs. A comparison of the results from these two
experiments indicates that people’s politeness behavior is different in
two respects when they perform a non-FTA and when they perform an FTA.
First, non-FTAs and FTAs can be distinguished by performances that
exhibit invariance along different dimensions of politeness. When
performing a non-FTA, subjects across different conditions manifested
the same degree of avoidance-based politeness. The reason that
avoidance-based politeness was invariant across conditions in
Experiment 1 seems to be that when the acts performed (i.e., greeting
and parting) did not threaten the hearer’s face, the speaker did not
have to employ politeness devices expressing that no offense was
intended. However, when performing an FTA, subjects showed varying
degrees of avoidance-based politeness.

In addition, subjects in Experiment 2 manifested the same degree

of approach-based negative politeness. This invariance of

160
approach-based negative politeness is due to the way power disparity
was manipulated in Experiment 2. Different from Experiment 1, where
the power of the bearer was either higher than or equal to the speaker,
the power of the bearer was either lower than or equal to the speaker
in Experiment 2. In this society, people generally use the same set of

address terms ("tu" pronouns) for both peers and subordinates

(Ervin-Tripp, 1972). Thus, the use of address terms in Experiment 2
was invariant across different power conditions.

A second difference in the performance of FTAs vs. non-FTAs is
that relational distance is a more important predictor of politeness in
performing FTAs, whereas power disparity is a better predictor of
politeness in performing non-FTAs. In a modern egalitarian society,
power disparity is not as powerful a predictor of people’s linguistic
behavior as it once was; in contrast, relational distance is a more
powerful predictor of linguistic behavior than it once was (Brown a
Gilman, 1960). Thus, relational distance is expected to be a better
predictor of politeness than power disparity as was found in politeness
behavior in performing FTAs. However, the way the relational variables
affect politeness behavior in greeting and parting others (i.e., acts
used in Experiment 1) seems to be different from the way they influence
politeness behavior in requesting, criticizing, or question-asking,
because the nature of a greeting or a parting is different from the
nature of a request, criticism, or question-asking. Greetings and
partings are highly ritualized acts, while requests, criticisms, and
question-askings are not. Usually ritualized acts (e.g., ceremony and

religious rituals) seem to emphasize giving proper respect to

161
authority. Power disparity, therefore, appears to remain a good
predictor of people’s linguistic behavior in ritualized acts.
Selative Ipportance f eterminants of Politeness

Experiments 1 and 2 found that some factors are more important
than others in determining the degree of politeness. First, except for
approach-based negative politeness in Experiment 1, relational distance
was a better predictor of politeness than power disparity. As
mentioned in the discussion of differences in politeness behavior
between an FTA and non-FTA, the egalitarian movement in Western society
in the eighteenth and nineteenth centuries influenced people’s
linguistic behavior. Before the movement, "power semantic" (Brown a
Gilman, 1960) was so strong that when people were engaged in a
conversation with a superior, they had to reflect in their utterances
their respect for the superior. After the movement, however,
"solidarity semantic" (Brown & Gilman, 1960) became so important that
people in a close relationship began to ignore the "power semantic."
Consequently, relational distance became a more important
characteristic of a relationship, in an egalitarian society.

Second, negative face-threat was a more important determinant of
negative politeness and positive face-threat was a more important
determinant of positive politeness. As argued in Chapter 1, because
positive and negative face-wants are distinct human desires, saving
others’ positive face (i.e., being positively polite) does not
compensate for their loss of negative face; and saving their negative
face (i.e., being negatively polite) does not compensate for their loss
of positive face. People should save positive face to mitigate a

threat to positive face and save negative face to mitigate a threat to

162
negative face. Consequently, positive face-threat was a better
predictor of positive politeness and negative face-threat was a better
predictor of negative politeness.

Finally, situational variables such as relational distance, power
disparity, positive face-threat, and negative face-threat were found to
be more important determinants of politeness than personality
variables. These findings are inconsistant with the findings of
compliance-gaining research that personality variables were better
predictors of compliance-gaining message selection. One possible
explanation for this inconsistency is that the present study did not
test the effects of all possible personality variables on politeness.
Only verbal aggressiveness and empathy were examined in this study.
Thus, there is a possibility that some personality variables that are
good predictors of politeness may emerge. Another possible explanation
is that personality variables may be good predictors of selection of
the message ceptent, but are not good predictors of the yey in which
the content is expressed. In any society, being polite is considered
as a rule to be observed and individuals may not be allowed to freely
select their own way of expressing the content. Thus, the influence of
personality variables on politeness behavior seems to be limited.

The Sole of Gender

Another important finding of this study is that gender did not
influence any dimension of politeness. In Chapter 1, it was argued
that the effect of gender could be explained in terms of verbal
aggressiveness. Both experiments showed that verbal aggressiveness is
a good predictor of gender. The correlation between verbal

aggressiveness and gender was -.33 for Experiment 1 and —.28 for

163
Experiment 2. Specifically, males were more verbally aggressive than
females. In Experiment 1, males were found to be less polite in terms
of approach-based positive politeness. However, when the effect of
verbal aggressiveness was removed, the effect of gender on
approach-based positive politeness became insignificant.

Except for approach-based positive politeness in greeting and
parting, the effect of gender failed to reach any significant level.
Gender did not influence approach-based negative politeness in
Experiment 1 and did not show any significant influence on any type of
politeness in Experiment 2. These findings are highly inconsistent
with Baxter’s (1984) finding that gender was one of the best predictors
of politeness.

Baxter employed a single speech act to analyze politeness. The
act Baxter used in her study was "request of re-writing," which was one
of the five acts used in the present study. Thus, the difference in
findings between these two studies does not seem to be due to the
difference in the type of act analyzed. One possible explanation for
inconsistency is that social desirability might have influenced
Baxter’s study. Baxter employed a message selection technique that
presented a set of politeness tactics to subjects and asked them to
rate their likelihood to use each of the tactics. Since using impolite
tactics is not "acting like a lady," females might have tended to rate
tactics with higher degrees of politeness as more likely to use.

Another possible explanation is that the present study and
Baxter’s study employed different means of measuring politeness.

Baxter generated a list of politeness tactics based on Brown and

Levinson’s list of politeness tactics. Baxter grouped these tactics

164
into three basic categories based on an exploratory factor analysis:
functional face threat, positive politeness, and negative politeness.
By adopting Brown and Levinson’s proposition, Baxter asspeed that
negative politeness was more polite than positive politeness which in
turn is more polite than functional face threat. Baxter concluded that
females were more polite than males since they tended to use negative
politeness more often than males.

In contrast, the present study does not assume that all positive
politeness or all negative politeness devices have the same degree of
politeness or that negative politeness is more polite than positive
politeness. The present study conceptualized positive and negative
politeness as mutually independent dimensions and examined the effect
of gender on these dimensions separately. The categorical coding
systems developed in Chapter 4 support the perspective of the present
study. Positive politeness was just as polite as negative politeness
and not all tactics for each type of politeness were judged to be the
same in politeness. Thus, the findings of the present study appear to
be more reliable.

Theegole of Personality Variables

The only area in which the present model failed to be supported is
in terms of the role personality factors play in determining positive
and negative politeness of utterances. Both experiments showed that
empathy was a poor predictor of both positive and negative politeness.
The reason that empathy did not show any effect on politeness seems to
be that "putting oneself in others’ shoes" or understanding others’
feelings does not necessarily mean that one recognizes others’ desire

to maintain face. In other words, understanding of others’ face wants

165
is not a crucial outcome of empathy. Thus, a highly empathetic person
may understand others’ emotional states very well but may not be highly
aware of others’ face wants.

The experiments also revealed that verbal aggressiveness did not
influence the levels of approach-based positive and negative
politeness. The finding that approach-based politeness was not
influenced by verbal aggressiveness seems very reasonable. The verbal
aggressiveness measure employed in this study measures one’s tendency
to attack others’ self-concept (Infante a Wigley, 1986) and is not
concerned with one’s tendency to promote others’ self-concept.
Approach-based politeness focuses on the way one is promoting others’
face and is not concerned with the way one is attacking others’ face.
People’s tendency to attack others’ self-concept may influence the way
they criticize or impose on others; however, it may not influence the
way they promote others’ self-concept. Consistent with this
explanation, Experiment 2 found that avoidance-based politeness was
influenced by verbal aggressiveness.

Dimensionality of Politenese

Both Brown and Levinson’s model and the present model
conceptualized politeness as a two dimensional construct (i.e., the
positive politeness and negative politeness dimensions). However,
Experiment 2 showed that approach-based and avoidance-based positive
politeness formed different dimensions, resulting in three different
dimensions of politeness, i.e., negative politeness, approach-based
positive politeness, and avoidance-based positive politeness.

Even though there is no direct evidence, it is likely that

approach-based and avoidance-based negative politeness would form

166
different dimensions. Approach-based negative politeness refers to
deference shown in the use of address terms, while avoidance-based
negative politeness refers to avoidance of imposition. As these two
different types of negative politeness are very different in nature,
there is no theoretical reason for us to assume that they covary. In
fact, Experiment 2 revealed that approach-based negative politeness was
invariant across different situations, while avoidance-based negative
politeness varied significantly across different situations. In short,
Experiment 2 suggested that politeness is a four dimensional contruct
comprising approach-based positive, avoidance-based positive,
approach-based negative, and avoidance-based negative politeness.

One possible counter-argument against the claim that these four
different types of politeness form different dimensions is that the
four types of politeness represent the four quadrants of two different
dimensions, i.e., positive-negative dimension and approach-avoidance
dimension. However, it is not likely that negative politeness and
positive politeness together form one dimension and that approach-based
politeness and avoidance—based politeness together form another
dimension. In order for positive and negative politeness to form a
single dimension, as positive politeness increases, negative politeness
should decrease. But, there is no reason that we should assume that
the more people satisfy others’ desire for approval (i.e., positively
polite), the less they satisfy others’ desire for power (i.e.,
negatively polite). Furthermore, a confirmatory factor analysis among
items in positive and negative politeness in Experiment 2 showed that
the correlations between the factors of approach-based positive

politeness and negative politeness (r=-.101, p<.290) and

167

avoidance-based positive politeness and negative politeness (r=.003,
p<.999) were not significantly different from zero.

In order for approach-based and avoidance-based politeness to form
a single dimension, as approach-based politeness increases,
avoidance-based politeness should decrease. However, it appears to be
unreasonable to assume that the more people approve others (i.e., more
approach-based positive politeness), the less they try to avoid
disapproving of others (i.e., less avoidance-based positive politeness)
or that the more deference people show to others (i.e., more
approach-based negative politeness), the less they try to avoid
imposing on others (i.e., less avoidance-based negative politeness).
Moreover, a confirmatory factor analysis of positive politeness in
Experiment 2 showed that the correlation between the factors of
approach-based positive politeness and avoidance-based positive
politeness was not negative but positive (r=.24, p<.025). Thus, the
claim that the four types of politeness represent the four quadrants of
the positive-negative dimension and the approach-avoidance dimension is
not plausible.
Modification oi the Beesent Medel

In summary, both Experiments 1 and 2 proved that the present model
predicted politeness behavior more accurately than Brown and Levinson’s
model. In addition, these experiments indicated that the present model
had three problems that need to be resolved. First, politeness is a
four dimensional construct rather than a two dimensional construct.
Second, empathy is not a factor determining the level of politeness.

Finally, verbal aggressiveness influences only avoidance-based

168
politeness behavior. Consequently, the present model needs to be
modified to resolve these three problems.

To accommodate the findings of this study, the model proposed in
Chapter 1 will be modified. First, politeness is conceptualized as a
four dimensional construct. In other words, politeness is not a vector
of positive politeness and negative politeness but a vector of
approach-based positive politeness, avoidance-based positive
politeness, approach-based negative politeness, and avoidance-based
negative politeness. Second, empathy is eliminated from the model.
Finally, the influence of verbal aggressiveness on both approach-based
positive and negative politeness was eliminated.

These changes do not affect such variables as relational distance,
power disparity, and the obligations to save positive and negative
face, since the influences of these variables are expected to be the
same for both approach-based and avoidance-based politeness. Thus,
Equations 2 and 3 in Chapter 1 (see Table 34) do not need to be
modified. The only equation that includes such variables as
politeness, verbal aggressiveness, and empathy is Equation 4. Thus,

Equation 4 is modified as follows:

 

 

 

 

 

 

apq 'Opfi r0!

v 0 m

p = P - A (7)
Pan On 0
LP .. L04 LnJ

VII II

where Pap’ Pvp’ Pan’ and Pvn are approach-based pos1t1ve,
avoidance-based positive, approach-based negative, and avoidance-based

negative politeness, respectively; A represents verbal aggressiveness

169

Table 34. Equations in the Final Model

Fina odel

[:33] = 14:] + »[':1 + [:1

(3)

r-—-1
O
0'6
l———l
ll
r——-|
row
0
”0'6
h—J
+
r-—I
e'd
L—J

 

 

 

 

 

 

n 11
PP 1 '0! r01

ap P

v 0 m

p = p _ A (7)
P O 0

an 11
LP .1 root Ln.

VII I1

RO : Relational obligation to save positive face
R0: : Relational obligation to save negative face
0 : Total obligation to save positive face
Op : Total obligation to save negative face

Pa : Approach-based positive politeness
Pvp : Avoidance-based positive politeness
Pa: ° Approach-based negative politeness
Pvn Avoidance-based negative politeness

T . Positive face-threat
T: : Negative face-threat
A : Verbal Aggressiveness
E : Empathy

P : Power disparity

D ° Relational distance

170
of the speaker and has a defined range of +1 to -1; and m and n are
positive constants. The differences between Equation 4 and Equation 7
are (1) empathy is dropped out of Equation 4; (2) both positive and
negative politeness are subdivided into approach-based and
avoidance-based politeness in Equation 7; and (3) in Equation 7, the

constants for verbal aggressiveness (i.e., A) are 0 when calculating

the level of approach-based politeness.

The Influences of Different Goals

 

The predictions of the present model tested in Experiments 1 and 2
were based on the assumption that people are striving to be polite.
Typically, one of the primary goals people have in social interaction
is to behave politely. Sometimes, however, people have a more
important goal than the politeness goal such as a relational goal or a
compliance-gaining goal. In these cases, people might act differently
from what the present model would normally predict. For example, when
people want to show off their power, they might show less positive and
negative politeness than the present model would normally predict; when
people want to act friendly to someone, they might show more positive
politeness and less negative politeness than the present model would
normally predict.

The present model can predict the level of positive and negative
politeness when the speaker’s behavior is guided by a relational goal.
The relational distance and power disparity in these cases should be
the ones that the speaker has in mind, not the ones assumed by both the
speaker and the bearer. Suppose, for example, that a speaker and a
bearer have regarded each other as casual friends and that since the

speaker does not like the bearer, the speaker wants to make the

 

171 F\
relationship more distant (i.e., wants to consider the hearer an
acquaintance). In this case, the speaker’s politeness behavior can be
predicted in the same way in which politeness behavior of a typical
acquaintanceship is predicted. In short, the relational distance and
power disparity are based on the relationship that is projected by the
speaker.

Compliance-gaining goals seem to influence people’s politeness
behavior. As discussed earlier, for example, when speakers want to
gain compliance from an acquaintance, they use a high degree of
positive politeness to make the relationship closer. In contrast, when
gaining compliance is taken for granted due to an emergency situation,
speakers are expected to use less polite expressions (such as orders).
The present model, however, does not explain the influence of
compliance-gaining goals very well.

The present model also has a problem in explaining the influences
of situationally induced aggressiveness. Lim (1988) suggest when
speakers realize that their compliance-goal is blocked, they become
verbally aggressive out of frustration. Thus, it is expected that when
speakers are frustrated because of a blocked goal, they become less
polite. The present model can explain the influence of verbal
aggressiveness as a trait; however, it cannot explain the effect of
situationally induced verbal aggressiveness, or more generally, the
effect of the speaker’s emotional state. Thus, future research should
be directed so that these limitations of the present model can be
extended. This future research should focus on "goal-frustration" and

"goal-presumption" as possible additional determinants of politeness

172
under the thesis that frustration and presumption decrease exhibited

politeness.

SUMMARY AND CONCLUSION
This study examined the adequacy of Brown and Levinson’s model of
politeness. The examination showed that Brown and Levinson’s model had

at least four crucial conceptual problems. First, Brown and Levinson’s

model accounts for only one kind of face want at a time despite both
positive and negative face being supported at the same time. Second,
the model limits the scope of politeness by conceptualizing it only as
a means to perform FTAs effectively. Third, the model focuses on a few
nonrepresentative elements of positive and negative face wants when
conceptualizing positive and negative politeness. Finally, this
nonrepresentative conceptualization of positive and negative politeness
leads Brown and Levinson to the inaccurate claim that negative
politeness should be used when the face-threat is high, and positive
politeness should be employed when the face-threat is low.

To eliminate the shortcomings of Brown and Levinson’s model, this
study proposed a new politeness model that accounted for both positive
and negative politeness manifested in the same message, explained
politeness phenomena which are not related to performing a
face-threatening act, and conceptualized politeness more systematically
by examining approach-based positive politeness (approval),
avoidance-based positive politeness (avoidance of disapproval),
approach-based negative politeness (deference), and avoidance-based
negative politeness (avoidance of imposition). This model argued that

the immediate predictors of politeness were the total obligations to be

173
positively and negatively polite; the total obligations to be polite
are a sum of the relational obligations to be positively and negatively
polite and the positive and negative face-threats carried by the act;
and the relational obligations to be polite are a function of the
relational distance between the bearer and the speaker and the power
the bearer has over the speaker.

Two experiments were conducted to test the adequacy of the model
proposed in this study. Experiment 1 tested whether the present model
made adequate predictions concerning the politeness manifested in
performing a non-face-threatening act. This experiment showed that
people tended to be polite by using approach-based positive and
negative politeness even when they performed an intrinsically
non-face-threatening act such as greeting and parting. Experiment 1
also revealed that the levels of positive and negative politeness
manifested in performing a non-FTA were influenced by relational
distance and power disparity.

Experiment 2 tested whether or not the present model made adequate
predictions concerning the politeness manifested in performing a
face-threatening act. This experiment showed that people attempted to
be polite both positively and negatively in the same message. Three
factors turned out to be good predictors of positive and negative
politeness. First, relational distance had a positive impact on
avoidance-based negative politeness and a negative influence on both
approach-based and avoidance-based positive politeness. Second, the
level of positive face-threat had a positive effect on the level of
both types of positive politeness. Finally, the level of negative

face-threat had a positive impact on the level of avoidance-based

174
negative politeness. Power disparity was not a good predictor of
approach-based positive politeness and avoidance—based negative
politeness, but a good predictor of avoidance-based positive
politeness.
Different from expectations, verbal aggressiveness and empathy of

a speaker did not have any substantive influence on any of the

politeness types. However, the results from these two experiments
generally supported the model proposed by the present study and yielded
considerable evidence againt Brown and Levinson’s model.

In addition, to develop categorical coding systems for the
criteria of politeness, this study identified unique politeness devices
employed by participants in Experiments 1 and 2 and examined the degree
to which each device represented the given criteria. Although the
categorical coding systems proposed in this study are not complete
because of the limited number of acts examined, they appeared to be
highly usable at least for the acts observed in these experiments. In
order to develop a more comprehensive coding system, however, more
categories exhibited in performance of different acts needed to be
added to the present coding system.

The results from this study have many implications. First, a
comprehensive theoretical framework that can explain a variety of
politeness phenomena is established. While Brown and Levinson limit
politeness phemonena to some special situations in which speakers
attempt to perform a face-threatening act effectively, the present
model includes all the phenomena that show any emblem of "facework."
Second, this study establishes a systematic explanatory scheme for

politeness phenomena. Most of the previous studies on politeness,

175
including Brown and Levinson’s work, focus on some particular aspect of
politeness behavior such as politeness strategies or some particular
factors influencing politeness such as gender or power differences. In
contrast, this study attempts to explain politeness systematically by
including all the crucial factors influencing politeness and by looking
at all the aspects of politeness phenomena.

Two major aspects of the present model are yet to be tested.
First, each equation should be tested individually. The present model
proposes three equations: Equation 7 proposes positive and negative
politeness are a function of the total obligations to save positive and
negative face, respectively: Equation 3 proposes the total obligations
to save positive and negative face are a sum of the relational
obligations to save positive and negative face, respectively, and
positive and negative face-threat, respectively; Equation 2 proposes
the relational obligations to save positive and negative face are a
function of the relational distance and the power disparity. Since
this study tested a combination of the three equations, two important
elements of the model, i.e., total face-saving obligation and
relational face-saving obligation, were not included in any of the
hypotheses. In order to know the effects of these face-saving
obligations, each equation needs to be tested individually. Second,
more studies need to be done to examine whether this model can explain
the influences of relational goals (e.g., to make the relationship
closer) and task goals (e.g., compliance-gaining goals). If necessary,
new factors should be added to explain politeness behavior guided by

such interaction goals as relational goals or compliance-gaining goals.

176
In short, more studies are need to test the adequacy of the model
completely.

This study has two limitations. First, the utterances analyzed in
this study were not sampled from natural conversations, but generated
based on hypothetical situations. The underlying assumption of this
method was that people’s linguistic behavior is guided by higher level
knowledge structures such as schemata. Thus, it was assumed that
people can realistically simulate their linguistic behavior in
laboratory situations. However, there is a possibility that people may
generate a more polite message in a laboratory situation than in an
actual interaction situation, because the laboratory situation allows
them enough time to monitor their message. There also is a possibility
that people act less polite in a laboratory than in an actual
interaction, since they do not have to be concerned with the hearer’s
response. This message generation technique also ignores politeness
expressed through nonverbal channels (e.g., pitch, intonation,
gestures, etc).

Second, this study measured politeness by means of multiple
ratings. Generally, a judge’s rating can be more subjective than
coding unitized content based on a categorical coding system. One main
reason that the present study did not use categorical coding systems is
that there was no established categorical coding system that has been
proven valid and reliable. However, this study developed relatively
comprehensive coding schemes for six different criteria of politeness.
Although the reliability and validity of these coding schemes were not

tested, they can potentially be useful tools for future research.

177
In conclusion, this study proposed a systematic and comprehensive
explanatory scheme for politeness and most of the predictions made by
the model were supported. However, the results are not conclusive by
any means. More studies analyzing natural conversations using

categorical coding schemes seem to be needed.

APPENDI CES

 

 

APPENDIX A

INSTRUCTIONS USED IN EXPERIMENT 1

Low Distance/Equal Power

Imagine that after a two week term break, a new term begins today.

Also imagine that on your way to a class, you run into a classmate
(Mike Miller) whom you have not seen for a while. Both Mike Miller and

you took several courses together, and worked in the same group for one
class requiring a group research project. Thus, ' e i 0

know each ether very well.

You notice that Mike Miller had his hair cut during the break and wears
a nice spring coat. Thus, he looks very refreshing. Since your class

begins in ten minutes, you do not have a lot of time to talk with this

classmate. Thus, you just want to exchange greetings and partings.

What do you typieally say to greet and part a classmate you know very
well? What will this classmate say to greet and part you? Weite
vepbatim eet what you and this classmate will say to greet and then
part each other. In other words, write a script in which you and this
classmate exchange greetings and them partings. You can write as many
turns as you think is necessary.

High Distance/Equal Power

Imagine that after a two week term break, a new term begins today.

Also imagine that on your way to a class, you run into an acquaintance
(Mike Miller) whom you have not seen for a while. Both Mike Miller and
you took a course together last term, and were in the same group when
you did group exercises in the class. Other than that, you fie pet kpow

Mike Miller very well.

You notice that Mike Miller had his hair cut during the break and wears
a nice spring coat. Thus, he looks very refreshing. Since your class
begins in ten minutes, you do not have a lot of time to talk with this
acquaintance. Thus, you just want to exchange greetings and partings.

What do you typically say to greet and part a fellow student you do not
know very well? What will this acquaintance say to greet and part you?
Write verbatim ept what you and this acquaintance will say to greet and
then part each other. In other words, write a script in which you and

this acquaintance exchange greetings and then partings. You can write

as many turns as you think is necessary.

 

178

179
High Distance/Hearer-High Power

Imagine that after a two week term break, a new term begins today.
Also imagine that on your way to a class, you run into a professor (Dr.
Mike Miller) whom you have not seen for a while. You know Mike Miller
since you took a class from him last term. Because you were very
active in this class, the professor also learned your name. Other than

that, you do not know Mike Miller very well.

You notice that Mike Miller had his hair cut during the break and wears
a nice spring coat. Thus, he looks very refreshing. Since your class
begins in ten minutes, you do not have a lot of time to talk with this
professor. Thus, you just want to exchange greetings and partings.

What do you typically say to greet and part a professor you do not know
very well? What will this professor say to greet and part you? Weite
verbatim eet what you and this professor will say to greet and then
part each other. In other words, write a script in which you and this
professor exchange greetings and then partings. You can write as many
turns as you think is necessary.

Low Distance/Hearer-High Power

Imagine that after a two week term break, a new term begins today.

Also imagine that on your way to a class, you run into a professor (Dr.
Mike Miller) whom you have not seen for a while. You know Mike Miller
since you took several classes from him; especially, you took an

independent study under him last term. Thus, Hike Miller epg you know
eech other very well.

You notice that Mike Miller had his hair cut during the break and wears
a nice spring coat. Thus, he looks very refreshing. Since your class
begins in ten minutes, you do not have a lot of time to talk with this
professor. Thus, you just want to exchange greetings and partings.

What do you typically say to greet and part a professor you know very
well? What will this professor say to greet and part you? Wpite
verbatim ept what you and this professor will say to greet and then
part each other. In other words, write a script in which you and this
professor exchange greetings and then partings. You can write as many
turns as you think is necessary.

180

APPENDIX B

INSTRUCTIONS USED IN EXPERIMENT 2

Low Distance/Equal Power/Low Positive Threat/Low Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

One day you run into another member of your group (Bill Jones), who yeu
pegard as a good lpiend. You want to know what Bill thinks of John’s
work. So, you decide to ask Bill’s opinion on John’s work. Suppose
that Bill already read the opening chapter done by John.

Wpite vepbatim the request or utterance you will make to ask Bill’s
opinion on John’s work. In other words, write out what you will say to
Bill to know what he thinks of John’s work.

Low Distance/Equal Power/High Positive Threat/High Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

So, you want to ask John to re-do his part of the project in the
remaining time before the final project is due. John will basically

have to start over again. Suppose that John is eomeeee yep pegepe as e
good friend.

 

Write verbatim the request or utterance you will make to ask John to do
his work again. In other words, write out what you will say to John
to motivate him to do his part of the project again.

__._1

1

 

181
Low Distance/Equal Power/High Positive Threat/Low Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

One day, John, who yeu pegard pp p geee triep , asks you what you think

about his work. Since this matter is related to the group’s grade, you
want to tell him that he did a poor job, hoping that he volunteers to
improve his work.

Write vepbatim the comment you will make on John’s work. In other
words, write out what you will say to John in order to let him know
that he did a poor job.

Low Distance/Equal Power/Low Positive Threat/High Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

Now, since you want a better group grade and since you think John Brown
does not want to do his part of the project again, you want to ask
another menber of the group (Bill Jones), who you think is very
actively involved in this group project, to take over and re-do John’s
part of the project in the remaining time before the final project is
due. If Bill decides to take over the job, he will basically have to
start over again. Suppose that Bill is seeeope yeu eegerd pp p geed
m-

Write verbatim the request or utterance you will make to ask Bill to
take over John’s part of the project. In other words, write out what
you will say to Bill.

182
High Distance/Equal Power/Low Positive Threat/Low Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

One day you run into another member of your group (Bill Jones), who you
do not know very well except for project group meetieg . You want to
know what Bill thinks of John’s work. So, you decide to ask Bill’s
opinion on John’s work. Suppose that Bill already read the opening
chapter done by John.

Write verbatim the request or utterance you will make to ask Bill’s
opinion on John’s work. In other words, write out what you will say to
Bill to know what he thinks of John’s work. ' '

 

High Distance/Equal Power/High Positive Threat/High Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

So, you want to ask John to re-do his part of the project in the
remaining time before the final project is due. John will basically
have to start over again. Suppose that 0 do not ow oh own ve

well except for project group meetipgs.

Write verbatim the request or utterance you will make to ask John to do
his work again. In other words, write out what you will say to John
to motivate him to do his part of the project again.

 

183
High Distance/Equal Power/High Positive Threat/Low Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

One day, John, who you do not know very well eycept foe prejeet gpopp

meetin s, asks you what you think about his work. Since this matter is
related to the group’s grade, you want to tell him that he did a poor
job, hoping that he volunteers to improve his work.

 

Write verbatim the comment you will make on John’s work. In other
words, write out what you will say to John in order to let him know
that he did a poor job.

 

High Distance/Equal Power/Low Positive Threat/High Negative Threat

Imagine that you are involved in a group research project that will
solely decide your course grade in a class where you very much want to
get a high grade. Naturally, you are one of the more enthusiastic
members of the group. In your view, one of the group members (John
Brown) has done his part of the project very poorly. John was supposed
to write the opening chapter of your research report. The opening
chapter was expected to make three major arguments. In order to do
this properly, John should have found enough evidence to prove these
three important points of the chapter. However, evidence provided in
this member’s writing is mostly irrelevant to the issues, thus
weakening the validity of the arguments.

Now, since you want a better group grade and since you think John Brown
does not want to do his part of the project again, you want to ask
another member of the group (Bill Jones), who you think is very
actively involved in this group project, to take over and re-do John’s
part of the project in the remaining time before the final project is
due. If Bill decides to take over the job, he will basically have to

start over again. Suppose that you do not know Bill vepy well except
for project group meetings.

Write verbatim the request or utterance you will make to ask Bill to
take over John’s part of the project. In other words, write out what
you will say to Bill.

 

184
Low Distance/Hearer-Low Power/Low Positive Threat/Low Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a

high grade, because you want to be a UTA for this class a sin next
term. In your view, one of the group members (John Brown has done his

part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

One day you run into another member of your group (Bill Jones), whe you
regard as e good friend. You want to know what Bill thinks of John’s
work. So, you decide to ask Bill’s opinion on John’s work. Suppose
that Bill already read the Opening chapter done by John.

Write verbatim the request or utterance you will make to ask Bill’s
opinion on John’s work. In other words, write out what you will say to
Bill to know what he thinks of John’s work.

 

185
Low Distance/Hearer-Low Power/High Positive Threat/High Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

So, you want to ask John to re-do his part of the project in the
remaining time before the final project is due. John will basically

have to start over again. Suppose that Johp is eemeepe yep pegppg pp p
good friend.

Write verbatim the request or utterance you will make to ask John to do
his work again. In other words, write out what you will say to John
to motivate him to do his part of the project again.

 

186
Low Distance/Hearsr-Low Power/High Positive Threat/Low Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

One day, John, who on e ard as a cod fr' nd, asks you what you think
about his work. Since this matter is related to the group’s grade, you
want to tell him that he did a poor job, hoping that he volunteers to
improve his work.

Write verbatim the comment you will make on John’s work. In other
words, write out what you will say to John in order to let him know
that he did a poor job.

187
Low Distance/Hearer-Low Power/Low Positive Threat/High Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

Now, since you want a better group grade and since you think John Brown
does not want to do his part of the project again, you want to ask
another member of the group (Bill Jones), who you think is very
actively involved in this group project, to take over and re-do John’s
part of the project in the remaining time before the final project is
due. If Bill decides to take over the job, he will basically have to

start over again. Suppose that Bill is someone you pegard as p good
frigid-

Write verbatim the request or utterance you will make to ask Bill to
take over John’s part of the project. In other words, write out what
you will say to Bill.

 

188
High Distance/Hearsr-Low Power/Low Positive Threat/Low Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

One day you run into another member of your group (Bill Jones), who you
do notphnow very well except for project gropp peetipge. You want to
know what Bill thinks of John’s work. So, you decide to ask Billis
opinion on John’s work. Suppose that Bill already read the opening
chapter done by John.

Write verbatim the request or utterance you will make to ask Bill’s
opinion on John’s work. In other words, write out what you will say to
Bill to know what he thinks of John’s work.

189

High Distance/Hearer-Low Power/
High Positive Threat/High Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
Opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

So, you want to ask John to re-do his part of the project in the
remaining time before the final project is due. John will basically
have to start over again. Suppose that yop fie pet hpew Jehp Spowp vepy
well except tor ppoject group peetings.

Write verbatim the request or utterance you will make to ask John to do
his work again. In other words, write out what you will say to John
to motivate him to do his part of the project again.

190
High Distance/Hearer-Low Power/High Positive Threat/Low Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class a ain next
term. In your view, one of the group members (John Brown has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

One day, John, who yeu do pot know yerv w l e t or

eetin , asks you what you think about his work. Since this matter is
related to the group’s grade, you want to tell him that he did a poor
job, hoping that he volunteers to improve his work.

Write verbatim the comment you will make on John’s work. In other
words, write out what you will say to John in order to let him know
that he did a poor job.

191
High Distance/Hearer-Low Power/Low Positive Threat/High Negative Threat

Imagine that you enrolled an independent study to work as an
undergraduate teacher’s assistant (UTA) for a class taught by a
professor you respect very much. Since you did well when you took this
class last year, you are allowed to enroll the independent study to be
a UTA of this class. As one of five UTAs in this class, your duty
involves taking charge of seven students who are working together on a
group research project. Your grade for the independent study as well
as the course grade of these students will be decided solely based on
the group’s performance in the project. You very much want to get a
high grade, because you want to be a UTA for this class again next
term. In your view, one of the group members (John Brown) has done his
part of the project very poorly. John was supposed to write the
opening chapter of the group’s research report. The opening chapter
was expected to make three major arguments. In order to do this
properly, John should have found enough evidence to prove these three
important points of the chapter. However, evidence provided in John’s
writing is mostly irrelevant to the issues, weakening the validity of
the arguments.

Now, since you want a better group grade and since you think John Brown
does not want to do his part of the project again, you want to ask
another member of the group (Bill Jones), who you think is very
actively involved in this group project, to take over and re-do John’s
part of the project in the remaining time before the final project is
due. If Bill decides to take over the job, he will basically have to
start over again. Suppose that you do not know Bill vepy well eyeept
for project group meetipgs.

Write verbatim the request or utterance you will make to ask Bill to
take over John’s part of the project. In other words, write out what
you will say to Bill.

BI BL I OGRAPHY

BIBLIOGRAPHY

Applegate, J. (1982). Construct system development and
identity-management skills in persuasive contexts. Paper
presented at the meeting of the Western Speech Communication
Association.

Axia, G., & Baroni, M.R. (1985). Linguistic politeness at different
age levels. Child Sevelepnent, SS, 918-927.

Baroni, M. R., a D’Urso, V. (1984). Some experimental findings about
the question of politeness and women’s speech. hengpege_in
Societ , 13, 67-72.

Bates. E- (1976). WWW
pppgpetiee. New York: Academic Press.

Baxter, L. A. (1984). An investigation of compliance-gaining as

politeness. Watch. 19. 427-456-

Becker, J. A. (1982). Children’s strategic use of requests to mark and
manipulate social status. In S. A. Kuczaj, II (Ed.), Lengpege

developpent (Vol. 2): Langpage, thought, end eplture (pp. 1-35).
Hillsdale, NJ: Erlbaum.

 

Boster, F. J. (1986). Determinants of the persistence and content of
compliance-gaining behavior. Paper presented at the annual
convention of the International Communication Association,
Chicago, Illinois.

Bowman, J., a Okuda, T. (1985). Japanese-American communication:

Mysteries, enigmas, and Possibilities. Splletin el the
Association fen Spsineee Senpunicatien, AS, 18-21.
Brend. R. M. (1978). Politeness. WW

Linguistics in hengpage Teaching, 1S, 253-256.

Brown, P. (1976). Women and politeness: A new perspective on language

and society. Reviews in Anthpepolegy, S, 240-249.

Brown, R., a Ford, M. (1961). Address in American English. Jeupnel of
Abnormal and Social Psychology, SS, 375-385.

Brown, R., a Gilman, A. (1960). The pronouns of power and solidarity.

In T. A. Sebeok (Ed.), Style_in_lengpege (pp. 253-276).
Cambridge: MIT Press.

Brown. c. To: I: Keller. P. W. (1973). MW”
explepetion e1 interpersonal eopmunication. Englewood Cliffs:

Prentice-Hall.

192

193

Brown, P. & Levinson, S. (1978). Universals in language usage:
Politeness phenomena. In E. Goody (Ed.), Qpestiens Ann
Politeness: Strategies in seeial interectien (pp. 56-289).

Cambridge: Cambridge University Press.

Burgoon, J. K., & Hale, J. L. (1984). The fundamental topoi of
relational communication. Communication Monogpephs, Sl, 193-214.

Burgoon, J. K., Pfau, M., Parrott, R., Birk, T., Coker, R., a Burgoon,
M. (1987). Relational communication, satisfaction,
compliance-gaining strategies, and compliance in communication

between physicians and patients. Communication Monogrephs, SA,
307-324.

Camras, L. A., Pristo, T. M., a Brown, M. J. K. (1985). Directive
choice by children and adults: Affect, situation, and linguistic

politeness. herrill-Palmer Qparterly, Si, 19-31.

Carrell, P. L., a Konneker, B. H. (1981). Politeness: Comparing
native and nonnative judgments. Lengpege_heppning, S1, 17-30.

Cialdini, R. B. (1988). Influence: Scienee end Eypetiee. Glenview,

IL.: Scott, Foresman and Co.

Clark, H. H., & Schunk, D. H. (1980). Polite responses to polite
requests. Co nition, S, 111-143.

Clark, H. H., a Schunk, D. H. (1981). Politeness in requests: A
rejoinder to Kemper and Thissen. Segnitien, S, 311-315.

Cody, M. J., & McLaughlin, M. L. (1985). The situation as a construct
in interpersonal communication research. In M. L. Knapp a G. R.
Miller (Eds.), book f ' er erso al 0 ice 10 (DP.
263-312). Beverly Hills, California: Sage Publications.

Craig, R. T., Tracy, C., & Spisak, F. (1986). The discourse of
requests: Assessment of a politeness approach. flpnen
Compunication Research, l2, 437-468.

Davis, M. H. (1983). Measuring individual differences in empathy:
Evidence for a multidimensional approach. Journal of persenality
and oc’al s cholo , AA, 113-126.

Ervin-Tripp, S. (1972). On sociolinguistic rules: Alternation and
concurrence. In J. J. Gumperz, & D. Hymes (Eds.), Sipections in
eocioligpistice (pp. 213-250). NY: Holt, Rinehart a Winston.

Ervin-Tripp, S. (1977). Wait for me, roller skate. In S. Ervin-Tripp
a C. Mitchell-Kernan (Eds.), Chile discouree (pp. 165-188). NY:
Academic Press.

 

194

Ferguson, C. A. (1964). Root-echo responses in Syrian Arabic

politeness formulas. In D. 8. Stuart (Ed.), hingpietie etudy in
memory of R. Slade harpell (pp. 37-45). Georgetown University
Press.

Ferguson, C. A. (1976). The structure and use of politeness formulas.
Langpage in Society, S, 137-151.

Gibbs, R. W. (1981). Your wish is my command: Convention and context

in interpreting indirect requests. Journal of Verbal Learning and
Verbal Behavier, 29, 431-444.

Gibbs, R. W. (1986). What makes some indirect speech acts
conventional? Journal of Memopy end Lengpege, SS, 181-196.

Goffman, E. (1967). Interpction pitual: Ssseys on feee-to-feee
behavior. Garden City, NY: Doubleday Anchor Books.

 

House, J., & Kasper, G. (1981). Politeness markers in English and

German. In F. Coulmas (Ed.), genvereational reutine:

Explorations in standardized eommunicatien situations and
prepatterned_epeech (pp. 157-185). The Hague: Mouton.

Hunter, J. E., a Boster, F. J. (1987). A model of compliance-gaining
message selection. Communicetien honogrpph, SS, 61-69.

Hunter, J. E., R Cohen, S. H. (1969). Package: A system of computer
routines for the anlysis of correlational data. Sdpcational

Psychological heasuremen , 2S, 697-700.

Huston, T. L. (1983). Power. In Kelley, H. H., Berscheid, E.,
Christensen, A., Harvey, J. H., Huston, T. L., Levinger, G.,
McClintock, E., Peplau, L. A., a Peterson, D. R. (Eds.), Slose
pelationships (pp. 169-219). NY: W. H. Freeman And Company.

Infante, D. A., a Wigley, C. J. (1986). Verbal aggressiveness: An

interpersonal model and measure. Communication honogpaph, SS,
61-69.

James, S. L. (1978). Effect of listener age and situation on the

politeness of children’s directives. Journal of Psycholingpistic
Research, 7, 307—317.

Kelley, H. H. (1983). Love and commitment. In Kelley, H. H.,
Berscheid, E., Christensen, A., Harvey, J. H., Huston, T. L.,
Levinger, G., McClintock, E., Peplau, L. A., 6 Peterson, D. R.

(Eds.), Slose pelationships (pp. 169-219). NY: W. H. Freeman And
Company.

 

195

Kelley, H. H., Berscheid, E., Christensen, A., Harvey, J. H., Huston,
T. L., Levinger, G., McClintock, E., Peplau, L. A., a Peterson, D.

R. (1983). Close pelationshipe (pp. 169-219). NY: W. H. Freeman
And Company.

Kemper, S. (1984). When to speak like a lady. Sex Rolee, 19, 435-443.

Kemper, S., a Thissen, D. (1981). Memory for the dimensions of

requests. Joupnal of Verbal Learning and Verbal Behavior, 2S,
552-563.

Lakoff, R. T. (1972). Language in context. Lan a e, AS, 907-927.

Lakoff, R. T. (1973). The logic of politeness; or minding your p’s and

q’s. In C. Colum et al. (Eds.), Papers from the flinth Segional

Meeting of the Chicego Lingpistic Seciet! (Pp. 292-305). Chicago:
Chicago Linguistic Society.

Lakoff, R. T. (1975). Lengpege end women’e pleee. NY: Harper and Row

Laver, J. D. M. H. (1981). Linguistic routines and politeness in
greeting and parting. In F. Coulmas (Ed.), genvepsational

routine: Exploratione in etendapdinee eepnpnieatien eitpetions
and prepatterned epeech (pp. 289-304). The Hague: Mouton.

Leech. G- N. (1977). Language and tact. W
University of Trier, Series A, Paper No. 46.

Lim, T. (1988). The influences of receivers’ resistance on persuaders’
verbal aggressiveness. A paper presented at the annunal convention
of the International Communication Association, New Orleans.

Magier, M. (1984). Systems of deference in Rajasthani. Paper
presented at the 36th annual meeting of the Association for Asian
Studies, Washington, D.C.

Marwell, G., & Schmitt, D. R. (1967). Dimension of compliance-gaining
behavior: An empirical analysis. Seeiepetpy, S_, 350-364.

McLaughlin, M. L., Cody, M. J., & O’Hair, H. D. (1983). The management
of failure events: Some contextual determinants of accounting
behavior. Human Coppunication esear h, 9, 208-224.

 

McLaughlin, M. L., Cody, M. J., a Rosenstein, N. E. (1983). Account
sequences in conversations between strangers. Cepmunieation
Monographs, SS, 102-125.

Mitchell-Kernan, C., a Kernan, K. (1977). Pragmatics of directive
choice among children. In S. Ervin-Tripp & C. Mitchell-Kernan
(Eds.), Child discouree (pp. 198-208). NY: Academic Press.

196

Nippold, M. A., Leonard, L. B., Anastopoulos, A. (1982). Development
in the use and understanding of polite forms in children. ournal

of Speech and hearing Researc , SS, 193-202.

Osgood, C. E., Suci, G. J., a Tannenbaum, P. H. (1957). The
measurepent of meani g. Urbana, Illinois: University of Illinois
Press.

Scotton, C. M. (1983). The negotiation of identities in conversation:
A theory of markedness and code choice. international Journal of
Sociology of Langpage, AA, 115-136.

Searle, J. R. (1969). Speeeh acte. Cambridge: Cambridge University
Press.

Shimanoff, S. B. (1977). Investigating politeness. In E. O. Keenan a

T. L. Bennet (Eds.), Siecourse ecress time and space (pp.
213-241). Los Angeles: University of Southern California.

Tannen, D., & Oztek, P. C. (1981). Health to our mouth: Formulaic
expressions in Turkish and Greek. In F. Coulmas (Ed.),
Conversational routine: Explanations in etendardieed

communicetion situetions and prepatterned speech (pp. 37-53). The
Hague: Mouton.

 

 

Tracy, K. (1983). The issue-event distinction: A rule of conversation

and its scope condition. human Communicetion Research, S,
320-324.

Tracy, K., Craig, R. T., Smith, M., a Spisak, F. (1984). The discourse
of requests: Assessment of a compliance-gaining approach. Human

Communication Reseerch, l_, 513-538.

Williams, D. L., a Boster, F. J. (1981). The effects of beneficial
situational characteristics, negativism, and dogmatism on
compliance-gaining message selection. Paper presented at the
annual meeting of the International Communication
Association, Minneapolis, Minnesota.