~trmihlI

ELSEVIER Journal of Pragmatics 27 (1997) 505-522

Polite requests in
English and Chinese business correspondence
in Hong Kong
Lorrita N.T. Y e u n g *
English Department, Lingnan College, Fu Yei, Yuen Man, Hong Kong

Received September 1994; revised version June 1995

Abstract
This paper examines the use of polite requests in both English and Chinese business corre-
spondence in Hong Kong. The analysis makes use of the three factors of imposition, social
distance and relative power in Brown and Levinson's (1987) theory of politeness. It is found
that only the factor of imposition has a statistically significant impact on linguistic choice in
the English data. There is also a significant effect when all three factors are combined. But
none of the factors, either alone or combined, shows any statistically significant effect in
the Chinese corpus. The results are attributed to the special characteristics of the type of
discourse investigated, the particular socio-cultural factors involved, as well as the way the
three factors are defined conceptually.

1. Rationale for the study

1.1. Scalar nature of linguistic politeness and the mechanism of choice

In all the major studies of politeness (Lakoff, 1973; Leach, 1983; Brown and
Levinson, 1987; Ervin-Tripp, 1976; Blum-Kulka, 1987; Ide, 1989; Fraser, 1990;
Kasper, 1990), there appears to be general agreement that there are different degrees
of politeness manifested in linguistic expressions. This certainly lends theoretical
support to the intuitive view that polite expressions can be put on a graduated scale
ranging from very polite to not very polite. While further empirical evidence would
illuminate the scalar nature of linguistic politeness in different languages, the ques-
tion remains what determines the appropriate degree of politeness and motivates the
choice of the corresponding linguistic expression. This is by no means a trivial ques-
tion in terms of both theory and pedagogy. Understanding the mechanism of choice
is the key to understanding the politeness phenomenon. It would also help the lan-

* Phone: +852 2616-7781; Fax: +852 2461-5270; E-mail: lntyeung@ln.edu.hk

0378-2166/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved

SSDI 0378-2166(95)00050-X
506 L.N.T. Yeung /Journal of Pragmatics 27 (1997) 505-522

guage learner to acquire mastery over the use of polite expressions. However, it is on
this question that there are diverse opinions and theories.

1.2. Theoretical background

I would like to group the different theoretical positions into three main categories,
using Fraser's terminology in his major review of politeness literature (1990): the
social-norm view (represented by the traditional social etiquette approach), the con-
versational-maxim view (Lakoff, 1973; Leech, 1983), and the face-saving view
(Brown and Levinson, 1987).
In understanding the mechanism of choice in actual language use, the social-norm
view would appear to be too narrowly prescriptive. This is especially so in a lan-
guage such as English, where linguistic choice cannot be easily explained with ref-
erence to prescriptive rules.
The conversational-maxim view, on the other hand, provides a set of principles to
account for linguistic politeness. However, the maxims are not formulated in such a
way that they can be readily tested empirically. Lakoff (1973), for example, posits
the maxim, 'be polite', as other than Grice's maxim 'be clear', to explain the use and
interpretation of polite language as distinct from the more practical use of language.
She also recognizes different levels of politeness, but never goes into the question of
how the choice is made. Leech (1983) provides a list of maxims, each with a set of
scales for determining the appropriate use of each maxim in given situations. But it
is a long list of maxims and scales posited by Leech. And as Fraser (1990) has
observed, the theory is difficult to evaluate in terms of how the maxims and scales
are to be applied, as they are not sufficiently clearly formulated for empirical testing.
For example, the tact maxim, which presumably is applicable to requests, says that
the speaker should minimize hearer costs and maximize hearer benefit. To determine
the degree of tact to be exercised, the speaker has to refer to a number of relevant
scales. They are those governing the relative cost and benefit to the speaker and
hearer (cost and benefit scale), the amount of choice the speaker permits the hearer
about the act requested (optionality scale), the extent to which the hearer has to infer
from the expression in order to recover the illocutionary intent (indirectness scale),
the relative right of the speaker to impose his or her will on the hearer (authority
scale), and the degree of familiarity between the speaker and hearer (social distance
scale). It appears that while some of the scales have to do with factors of the speech
context (such as the cost and benefit, authority scale and social distance scales), oth-
ers (like the optionality and the indirectness scales) have to do more directly with the
formulation of linguistic expressions. While admitting that there are different
degrees of politeness, Leech has not calibrated the scales in any precise way. It is not
easy to verify exactly how the scales inform the use of a particular maxim, which in
turn motivates the linguistic choice of politeness. It will be seen that the face-saving
view of Brown and Levinson (1987), on the other hand, while using similar con-
cepts, provides a more precise formulation.
The face-saving view (Brown and Levinson, 1987) uses Goffman's (1967) con-
cept of face to explain the politeness phenomenon. Face refers to the "public self-
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 507

image that every member [of a society] wants to claim for himself" (Brown and
Levinson, 1987: 66). To maintain the other's face is to recognize and respect the
claim members of society make with respect to each other in interaction. The act of
communicating such an acknowledgment is politeness. According to Brown and
Levinson (1987), face consists of both positive face and negative face. The former
refers to the positive self-image of the interactant, "including the desire that this self-
image be appreciated and approved of" by others. The latter represents the basic ter-
ritorial claim of the individual and his or her "freedom to action and freedom from
imposition" (1987: 66). Politeness is conceived of as an instrumental act in address-
ing the interactant's need to maintain or to remove threat to positive and negative
face. Certain acts of politeness, such as requests, are intrinsically face-threatening
and thus require strategic redress. The strategic acts of politeness can be classified as
bald-on-record, on-record with redress, and off-record, encompassing a range of dif-
ferent degrees of politeness. The choice of appropriate polite expressions in a given
context depends on a number of factors which Brown and Levinson have reduced to
a simple formula.
For their politeness formula, Brown and Levinson (1987) postulate three indepen-
dent variables that have a systematic effect on the choice of politeness strategies.
They are: the social distance (D) of the speaker and hearer, the relative power (P)
between them, and the absolute ranking (R) of the imposition in the particular cul-
ture. In Brown and Levinson's view, each of them has an independent effect on the
strategic choice of polite expressions. In addition, their effects can be specified in
terms of a numerical value, which in turn can be demonstrated as reflected in the
choice of polite expressions. Such a postulation naturally has enormous heuristic
value. Brown and Levinson's theory clearly surpasses the other two theoretical posi-
tions, mentioned above, in this regard.
But Brown and Levinson's theory is by no means free from challenges and criti-
cisms. For example, Brown and Levinson's range of politeness strategies from bald-
on-record through on-record to off-record suggests that politeness can be equated
with indirectness. Blum-Kulka's study (1987) finds that off-record strategies, being
the most indirect, are not actually considered the most polite by language users, at
least not in Hebrew, and (albeit to a lesser degree) not even in English. This, how-
ever, does not invalidate the notion that there is a gradation of politeness expressed
in on-record redress strategies. But such a scale needs further investigation and ver-
ification.
Regarding the three factors of P, D, and R, subsequent research has yielded mixed
findings. Brown and Gilman's study of Shakespeare's four major tragedies (cited in
Kasper, 1990) finds that while imposition and power account for the characters'
polite expressions, social distance does not. They have also suggested affect as
another determining factor. On the other hand, Cherry (1988) examines a set of let-
ters of petition written by academics of different ranks to the president of an Amer-
ican university. The study shows that relative power does not predict the relative
politeness of requests.
Brown and Levinson's claim to universal validity has also been under attack. For
instance, Ide (1989) argues that it is not universally true that politeness is only the
508 L.N.T. Yeung / Journal ~f Pragmatics 27 (1997) 505-522

making of a strategic choice, as in the English language. In Japanese, for example,

the use of polite expressions is more normative and prescriptive in nature. In an hon-
orific language such as Japanese, the latitude of choice is much narrower compared
to English. But basically, Ide endorses Brown and Levinson's notion of D, P and R.
She has investigated the effect of D and P on the choice of polite expressions by
holding R constant. As expected, her Japanese subjects yielded a much more pro-
nounced correlation pattern between the degree of politeness and the social variables
than her American subjects. Basically her study has not refuted, but rather recon-
firmed as well as extended, Brown and Levinson's framework.
Matsumoto (1988) has also queried Brown and Levinson's notion of deference as
necessarily only a negative strategy of removing imposition. By using examples of
formulaic greetings in Japanese, she points out how deference is given to the
addressee through the use of an imposition. She further explains that such imposition
or request-making is socially restrictive, according to the relative status of the inter-
actants and the context of the act. It would be interesting to see how deference and
imposition are viewed and dealt with in other languages and cultures. Nwoye (1992)
finds that imposition is not viewed unfavorably in native Algerian culture, as it is in
Western English-speaking culture. It is, therefore, not necessary to redress imposi-
tion in the same way. The question is whether such cultural variations invalidate
Brown and Levinson's politeness formula.
Despite recent criticisms of Brown and Levinson's politeness theory (Kasper,
1990; Fraser, 1990: Ide, 1989; Werkhofer, 1992), its conceptual framework has
proved to be the most germane in generating empirical research. Subsequent
research and theorizing have tended to further refine and extend the model (Ide,
1989; Matsumoto, 1988, 1989; Blum-Kulka and House, 1989; Nwoye, 1992), rather
than refute it completely. So far, Brown and Levinson's theory has offered the most
parsimonious explanation for the politeness phenomenon. But as mentioned earlier,
there are still questions unanswered by the theory.

1.3. Quantitative study of natural data

In spite of the empirically precise nature of Brown and Levinson's theory (1987),
surprisingly few studies have used a quantitative approach to verify it. Reviewing
the literature on linguistic research on politeness, Held (1992) recognizes the promis-
ing value of quantitative studies, yet observes that only initial steps have been taken
in the quantitative direction. Held attributes this to the predominance of the interpre-
tative paradigm in linguistics in general and in politeness studies in particular. A
more fundamental reason may be the basic objection to a simplistic, deterministic
view of social phenomena, as pointed out by Werkhofer (1992). Nevertheless, as
Werkhofer has mentioned, despite criticisms of such a nature, Brown and Levinson
still maintain that, as supported by research, "the three sociological factors are cru-
cial in determining the level of politeness which a speaker (S) will use to an
addressee (H)" (1987: 15). But to build a stronger case for the decisive influence of
sociological factors on complex social behavior such as language, more empirical
studies on naturally occurring social phenomena are clearly in order.
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 509

However, of the few existing empirical studies on the effect of social distance, rel-
ative power, and imposition on linguistic politeness, even fewer involve research on
naturally occurring data. Most of them make use of data collected by means of ques-
tionnaire elicitation (Hill et al., 1986; Blum-Kulka, 1987; Blum-Kulka and House,
1989; Ide, 1989) and role-play (Garcfa, 1993; Nash, 1983). Nevertheless, the impor-
tance of naturally occurring data has been well-recognized (House, 1989; Stubbs,
1986). As Stubbs points out, "intuitions about stylistic variation are simply not
accessible or reliable" (1986: 22). This obviously applies to the supposedly socially
calibrated phenomena of linguistic politeness.
While claiming an independent effect of each of the three variables, Brown and
Levinson (1987) also postulate an interactive effect among them. According to
Brown and Levinson, the three factors interact with each other in such a way that
together they trigger off a constant threshold effect for the choice of a particular
level of linguistic politeness. In other words, the total effect of the three factors taken
together has an impact which is quite independent of their values taken separately.
For example, a lower status (which, say, can be represented by a numeric of +2)
combined with a low social distance (say, 0) in the face of a request of medium
imposition (say, + 1) give a threshold level of 3, which generates a moderately polite
request. But the same moderately polite request (triggered by 3) can also be chosen
when the three factors are assessed to be of moderate value (each represented by + 1).
Such quantitative conceptions of the politeness theory particularly lend themselves
to statistical treatments.
Brown and Levinson (1987) further point out that the three factors of D, P and R
are context-dependent. The values that are assigned to them can only depend on the
particular situation, which is socio-culturally specific. But apart from the socio-cul-
tural milieu, the type of discourse or genre obviously has a decided impact on the
linguistic choice. Halliday and Hasan (1985) have clearly demonstrated how the
genre, which represents intertextual conventions arising out of typical socio-cultural
interactions, equally constrains the linguistic choice in a systematic way. But so far,
politeness research has concentrated on oral interactions almost exclusively. In view
of the situation, Kasper (1990) has called for more research on the impact of dis-
course type on politeness.

1.4. Studies on Chinese politeness versus English politeness

It is commonly agreed (Fraser, 1990; Kasper, 1990; Ide, 1989) that the English
language underlines much of Brown and Levinson's theory (1987); its applicability
to the English language can be assumed. But it is not clear whether it is equally
applicable to the Chinese language. So it is also appropriate at this point to go into
some of the recent studies on Chinese politeness as part of the background to this
study.
There are not too many studies on Chinese politeness. Among the few, there is
unanimous agreement that the origin of Chinese politeness is 'li', which traces its
roots to the Book of Rites in ancient China. 'Li' originally has to do with decorum
and propriety that regulates interpersonal behavior in order to ensure harmony and
510 L.N.T. Yeun~ / Journal of Pra~matics 27 (1997) 505-522

order in society. For example, both Shih (1988), a scholar in Taiwan, and Gu (1990),
a researcher in Mainland China, believe that the concept of 'li' still explains the phe-
nomenon of Chinese politeness in modern China. Both agree that in the tradition of
'li', Chinese politeness still emphasizes deference for the other and modesty for one-
self. As respectfulness is often shown through formality, polite language tends to be
formal, and informal forms are regarded as neutral.
Gu (1990) argues that contrary to Brown and Levinson's theory, Chinese polite-
ness is normative rather than strategic in nature. Appropriate display of politeness in
the proper context is obligatory, as lack of it will incur social sanction. Shih (1988)
also finds that appropriateness and moderation according to one's role and status in
society are important guiding principles in Chinese politeness. In contrast, Ameri-
cans would put far less emphasis on status differences. Instead, they would value
individualism and equality much more. In the same vein, Gu (1990) points out that
face in Chinese cannot be equated with the concept of negative face in Brown and
Levinson (1987), which stresses individual territorial rights and freedom from
impediments. It would seem that both are arguing for the use of the concept of social
discernment, which Ide (1989) finds more suited to describing and explaining Japan-
ese politeness. According to Ide's conceptual framework, polite language acts as a
kind of social index reflecting the relative positions of interactants in given social
situations. However, unlike Japanese, Chinese is not an honorific language. When it
comes to signaling politeness in social interaction, the choices are far less restrictive
and rigid in Chinese than those in the Japanese language. Furthermore, society has
evolved considerably from the old style of 'li' (Shih, 1988; Gu, 1990). Some of the
formal terms of politeness are now considered classical and no longer fashionable.
Basing himself on Brown and Levinson's formula of P, D, and R (1987) to
account for the choice of polite expressions, Shih (1988) adopts their conceptual
framework in his analysis of Chinese politeness. Although no spe6ific reference is
made to Brown and Levinson, Zhang ( 1991 ) uses the notions of status difference and
familiarity, as well as of imposition, to examine the effects of social and contextual
factors on the use of modality in polite expressions in modern standard Chinese.
Zhang sees modality as expressing varying degrees of tentativeness, which signals
politeness. It is obvious from his study that Zhang's notion of tentativeness can sub-
stitute for Brown and Levinson's redressive strategy of negative politeness.
With particular regard to the applicability of negative politeness in Chinese, the
opinion seems divided. We have seen that Zhang's analysis of Chinese (1991) basi-
cally shares a similar view to Brown and Levinson's (1987). Shih (1988) has con-
ducted a study on request-making strategies among Chinese and American subjects.
He finds that a much larger percentage of the Chinese (60%) make off-record
requests than the Americans (0%) to avoid imposition (1988: 149). This appears to
correspond to Nash's findings (1983). On the other hand, Gu (1990) finds that cer-
tain acts such as invitations and offers, which Brown and Levinson find face-threat-
ening, are not imposing in the eyes of the Chinese. Nevertheless, Gu has said noth-
ing about how the Chinese view the imposition of requests.
One can see from the above that the picture of Chinese politeness emerging from
recent studies is less than complete or precise. In addition, none of the studies has
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 511

used natural data or applied the empirical test of the formula of P, D and R, although
Brown and Levinson's framework (1987) is used openly or implicitly in analyzing
Chinese politeness.

2. Purpose of the study

The purpose of this study is to examine the formulation of requests in the English
and Chinese business correspondence in Hong Kong to see if the factors of P, D, and
R can predict linguistic choice, as postulated by Brown and Levinson (1987), and if
they operate in the same way in both English and Chinese. A comparison of the two
linguistic samples should shed light on the cultural variations, if any, and help fur-
ther clarify the universal applicability of Brown and Levinson's framework (1987) in
explaining politeness phenomena.

3. Method

3.1. Sample

In Hong Kong, both English and Chinese are used as the language of business.
However, English is clearly the more commonly used language in business corre-
spondence because of Hong Kong's recent history as a British colony and the inter-
national character of its business ventures. Although 98% of its population are Chi-
nese, there is a large number of expatriates from English-speaking countries such as
Britain, the United States, and Australia, working in the business community as well
as the public sector. Many local Chinese have also acquired native or near-native
competency in the English language. But Chinese, too, is used among local firms,
particularly those with trade relations to Mainland China and Taiwan. In communi-
cating with the local consumers and the local staff, major organizations are also in
the habit of having a Chinese version of their correspondence in English. Most of
these Chinese versions are not literal translations of their English counterparts.
Rather, they display distinct Chinese conventions quite different from the English
original. In Hong Kong, Classical Chinese is used in some business correspondence,
while Modem Standard Chinese is used in others. Modem Standard Chinese appears
to be more popular, but it is often interspersed with Classical Chinese in one and the
same correspondence, l
For the present study, a total of 360 English and 181 Chinese letters and memos
were collected over a two-year period from a spectrum of organizations, including
utility companies, educational institutes, government departments, banks, publishing
firms, hotels, and other commercial organizations in Hong Kong.

t The written style of Chinese used in business correspondence in Hong Kong is considered to be quite
distinct from that used by the Mainlanders and yet different again from that of the Taiwanese.
512 L.N.T. Yeung / Journal t~fPragmatics 27 (1997) 505-522

3.2. The dependent variables." Linguistic jbrmulations of varying degrees of

politeness

Two lists of expressions making polite requests - one in English and the other in
Chinese - were made up by selecting items which occur most frequently in the two
samples. The English list consists of eight items, and the Chinese list of four. To
establish the degree of politeness conveyed by the various items in the lists, they
were given to two groups of qualified informants for ranking respectively. In other
words, the informants were asked to indicate the relative politeness of the items by
ordering them from the most polite to the least polite. The 12 English informants
were native-speaker English teachers in the Hong Kong Polytechnic and the City
Polytechnic (now City University) in Hong Kong. The 5 Chinese informants selected
were teachers of Chinese and Translation in the universities referred to earlier. Their
views were sought because being native speakers and teachers of their own lan-
guage, they should have both intuitive and clearly-articulated, objectively-acquired
knowledge about their respective languages. Their views should also reflect the stan-
dards and norms held in their own culture, To arrive at an overall ranking by the
informants, the ranking of each item in the two lists was turned into a score and the
mean was taken of all the scores of each item.
The rankings of the different polite formulations of request are given below. The
larger the mean, the lower the ranking, the less polite the expression appeared to the
raters as a whole. The English list first:

1. It would be much appreciated if you would respond quickly. Mean: 2.4

2. 1 would appreciate it if you could respond quickly. Mean: 2.6
3. We would like to invite you to respond. Mean: 2.9
4. I hope you will respond quickly. Mean: 3.1
5. I should be grateful if you could respond quickly. Mean: 3.7
6. You are requested to respond quickly. Mean: 5.7
7. I should like you to respond quickly. Mean: 5.8
8. Please respond quickly. Mean: 6.0

From the above list, it can be seen that the ranking corresponds to the major existing
theories and research findings on linguistic politeness. For example, Blum-Kulka
(1987) finds that mood derivables or imperatives are the least polite. Item 8, "Please
respond quickly", is an imperative plus a request marker (House, 1989). Following
imperatives are first the performative in Item 6, "You are requested to respond
quickly", then the hedged performative in Item 3, "We would like to invite you to
respond quickly", and then the want statement in Item 7, "I would like you to
respond quickly". Item 3 ranks relatively high in politeness in our list, most proba-
bly because of the requestive word, 'invite', which is semantically more polite than
performatives like 'ask' in Blum-Kulka's list. Item 5, "I should be grateful if you
could respond quickly", expressing gratitude as a politeness strategy, is rated on the
average as less polite than the statement of hope in Item 6, "I hope you will respond
quickly". This would appear to be largely in keeping with Brown and Levinson's
 L.N.T. Yeung / Journal r?fPragmatics 27 (1997) 505-522 513

theory (1987) that giving thanks is a face-threatening act. Item 1, "It would be
appreciated if you could respond quickly", uses three strategies of distancing:
impersonalization, the conditional, and tense (Giv6n, 1989; Koike, 1989; Dancygier,
1993; Perkins, 1983; Kress and Hodge, 1979). Impersonalization removes the pres-
ence and identity of the maker of the request, thus making it less embarrassing to
turn down the request. The conditional and the subjunctive attenuate the immediacy
and weaken the manipulative force of the request. Overall, the ranking distribution
reflects the principle of negative politeness (Brown and Levinson, 1987), which aims
at not impeding the addressee's freedom of action. It also echoes the relationship
between uncertainty and politeness posited by Giv6n, (1989), which basically oper-
ates on the principle of how strongly the speaker thinks he can and wants to manip-
ulate the addressee.
The Chinese items show the following ranking (the romanization convention used
is based on the Yale system):

1. Yu moung gwai goungsi hipjo si chi wudoung, Mean: 1

If HON PASS HON company help this CLASS activity, 2
jaak baatsing mingaam.
then overwhelmingly grateful.
2. Gingkei gwai goungsi hipjo si chi wudoung. Mean: 2.3
Respectfully
beseech
3. Heimong lei naang hipjo ngomun je chi wudoung. Mean: 2.5
Hope you able our this
4. Ching lei hipjo ngomun je chi wudoung. Mean: 3
Please

The first item has two honorific markers, 'moung' and 'gwai', for the addressee and
an expression of gratitude, 'mingaam', modified by a strong boosting adjunct,
'baatsing'. The item ranking second in politeness contains a deferential requestive
verb 'gingkei' and an addressee honorific. The third is marked by a cognitive verb
of certainty, 'heimong' (Giv6n, 1989). The last one has a politeness marker, 'ching',
equivalent in meaning and usage to the English 'please'. As indicated by the mean
scores, the ranking apparently represents an order of deferential politeness. Never-
theless, there were three expert raters who ranked Items 2 and 3 as equal in polite-
ness and one who found no difference among Items 1, 2 and 3. The reason given was
that the variations are only stylistic differences between Classical and Modem Stan-
dard Chinese.

3.3. The independent variables: The contextual factors of D, P, and R

The independent variables of D, P, and R were operationalized according to the

definition of Brown and Levinson (1987). D was defined as the social distance or

2 HON stands for honorific. PASS is short for passivity. CLASS is classifier.
514 L.N.T. Yeung /Journal of Pragmatics 27 (1997) 505-522

degree of familiarity or closeness between the speaker (S) and the hearer (H). P was
taken to refer to the relative status or power difference between S and H, which acts
as a function of the ability of S to impose his or her will on H with regard to the
request in question. R was interpreted as the degree of imposition of the request on
H. The imposition is situationally determined, taking into consideration H's obliga-
tion to comply with the request, his or her ability to do it, and likely enjoyment of
doing it, as well as the difficulty of the task itself for H to carry out, and for S to
make the request. All three are to be understood as sociological factors in that they
are socio-culturally oriented.
D and R were given a scale of 0-1-2, representing low, medium and high respec-
tively. In other words, for D, a rating of zero means that S and H know each other
well. A rating of 1 means H and S are just acquaintances, whereas a rating of 3
means H and S are strangers. For R, a 0-rating means the task requested of H would
cause minimal or only very slight imposition. 1 means the imposition on H is mod-
erate rather than substantial, the latter represented by 3. P was given a five-point
scale, from +2, +1 of relatively higher status of S, through 0 of equal status of S and
H, to -1 and - 2 of lower status of S relative to H.
Some examples of ratings of the three factors are given as follows:
As an example of D, a sales manager sending direct mail shots to potential
clients would be given a rating of 2. A bank manager requesting further informa-
tion from his client concerning a loan would be rated as having a D-rating of I.
Co-workers writing to each other would have a D-rating of 0. Regarding P, a
manager asking his subordinate to perform a task within the latter's sphere of
duties would have a rating of +2. A staff manager writing to line managers ask-
ing them to help fill in a survey would be rated as having a P of +1. Colleagues
of the same rank would be rated as having a P of 0. An accountant writing to the
tax assessor asking for a reassessment would be given a P-rating of -1. As for R,
a request by the manager asking the staff to put in extra work on a holiday would
be rated as 2. A request by an outsider to an organization asking the manager
to make arrangements for a visit to the work units would be given a rating of 1.
A reminder asking H to call in case of any problems would be rated as 0 on the
R scale.
Request formulations using linguistic structures such as the items in the above two
ranking lists were identified in the English and Chinese samples respectively. They
were then assessed by two raters independently according to the D, P, and R factors,
using the above rating scales. Inter-rater agreement was over 95%.

3.4. Statistical computation

To find out whether the three sociological factors have an independent and com-
bined influence on the choice of linguistic politeness, the Abacus Superanova statis-
tical package was run to make multiple regression analysis. While the p-value of
each of the three independent variables indicates the statistical significance of their
independent effect, the p-value of the statistical model as a whole indicates the
strength of the three factors' combined impact.
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 515

To find out how much each of the factors contributes to linguistic variation, a
step-wise regression analysis was made by using the Statview II package of Abacus
Concepts.

4. Results

The results are given in Tables 1, 2, 3 and 4. As shown in Table 1, of the three
contextual factors, only imposition has the potential to predict English politeness for-
mulations at a highly statistically significant level (p-value 0.0001).

Table 1
Results of multiple regression analysis of the effects of contextual factors on linguistic politeness in the
English sample

Mean Std. Regression Std. p-value of

Dev. coefficient Error coefficient

Intercept 5.771
Imposition (R) 0.841 0.711 -0.64 0.123 0.0001 *
Social distance (D) 1.641 0.492 -0.075 0.179 0.6741
Power (P) -0.128 0.759 -0.209 0.114 0.0923

* Statistically significant

However, as can be seen in Table 2, the combined effect of the three factors also
has a significant impact, explaining about 16% of the variation in the politeness
expressions at a similarly high probability value of 0.0001. A step-wise regression
analysis shows that imposition accounts for about 13% of the linguistic variation.

Table 2
Model summary of regression analysis of the effects of contextual factors on linguistic politeness in the
English sample

Count of items 197

R 0.399
R-squared 0.159
p-value 0.0001 *

*Statistically significant

In the Chinese sample, however, none of the contextual factors has any significant
effect at a statistically satisfactory level, as is displayed in Tables 3 and 4.
516 L.N.T. Yeung /Journal ofPragmatics 27 (1997) 505-522

Table 3
Results of regression analysis of the effects of contextual factors on linguistic politeness in the Chinese
sample

Mean Std. Regression Std. p-value of

Dev. coefficient Error coefficient

Intercept 2.676
Imposition (R) 0.932 0.812 -0.095 0.05 0.0615
Social Distance (D) 1.924 0.293 0.03 0.146 0.8349
Power (P) -0.386 0.626 0.471 0.068 0.1432

Table 4
Model summary of regression analysis of the effects of contextual factors on linguistic politeness in the
Chinese sample

Count of items 132

R 0.216
R-squared 0.046
p-value 0.1061

5. Discussion

5.1. English sample

In the English sample, only the factor of imposition has an independent effect, as
postulated by Brown and Levinson (1987), at a statistically significant level. This
should not be surprising, especially if one bears in mind the fact that this factor actu-
ally takes into account the effects of both P and D. For example, whether H feels
obligated to perform a request is a function of H's relationship of power with S. Sim-
ilarly, but probably to a lesser extent, the social distance between S and H has a bear-
ing on how difficult it is for S to make the request.
As postulated by Brown and Levinson, all three factors taken together have a sig-
nificant impact. The statistical level is the same as, but not better than, that of the
factor of imposition alone. This does seem to indicate that the factor of imposition
can replace the other two as the single reliable predictor of the linguistic choice of
politeness. In fact, in another study done by Blum-Kulka and House (1989), the fac-
tors which were found to be significantly correlated with indirect politeness were
obligation, right, dominance, compliance, and difficulty, which can all be subsumed
under Brown and Levinson's ranking of imposition.
In the model of the present research, the portion of variation accounted for is at
the 15.9% level. It is obvious that there are other factors at play which have not been
taken into consideration. Brown and Levinson (1987) admit that there are other rel-
evant factors, but maintain that these are the ones "that have a principled effect" on
the assessment of the situation for making strategic politeness decisions (p. 85). The
 L.N.T. Yeung / Journal ~'Pragmatics 27 (1997) 505-522 517

present study indicates that they only explain a relatively small proportion of the
linguistic choices. However, Blum-Kulka and House (1989) did find higher por-
tions of the regression accounted for 3 by the contextual factors mentioned above,
which closely approximate Brown and Levinson's concept; however, their data
were from Hebrew, German, and Spanish, not from English. Obviously, a direct
comparison cannot be made. The situations were also based on oral interactions,
not business correspondence. In addition, as their data were elicited responses to
a questionnaire, the results may tend towards the ideal when compared to natural
data.
The factor of social distance does not appear to have any significance at all in the
present data. In fact, business correspondence is written discourse of a semi-formal
style. The expected conventional forms of courtesy tend to have a neutralizing effect
on the linguistic choice, so that expressions reserved for intimacy are typically not
used. On the other hand, it is a tenet of good business writing to be friendly and not
stand-offish or ceremonious. In light of the above, the range of choices based on
social distance would be much reduced at both ends of the politeness scale. It should
not come as a surprise that social distance does not have a calibrated effect on the
linguistic choice of politeness.
Power has a much higher significance level than social distance, but not at an
acceptable level. There are studies which have found that power is not directly and
neatly correlated with linguistic politeness. As mentioned earlier, Cherry (1988) has
demonstrated with naturally-occurring data that lower status writers use non-polite
strategies to make a point to an addressee of higher status. Cherry attributes this to a
rhetorical use of strategy to underline the urgency and importance of the requested
action. In a similar way, Drake and Moberg (1986) observe that organizational influ-
encers may deliberately violate power or social distance expectations to achieve cer-
tain aims. They recognize the importance of understanding the strategic use of
politeness in business situations. For example, superiors may use powerless linguis-
tic forms to disarm the subordinates into compliance, when other inducements are
insufficient. Alternatively, subordinates may adopt powerful forms to convey a sense
of new-found confidence.
From the above, it is obvious that natural data of linguistic formulations of polite-
ness do not conform exactly to expectations. But even data elicited with the ideal in
mind may not generate a neatly correlated pattern. Hill et al. (1986) designed a study
expressly to elicit the stereotypical forms of politeness in situations involving
addressees of typically different status. 4 They found no clear-cut correlations
between the degree of politeness and relative status in the American English corpus.
Unlike in the Japanese counterpart, there was only a loosely clustered pattern. The
explanation offered is that Americans, coming from an individualistic culture, have
a strategic freedom to choose linguistic expressions. The same may well apply to the
English-speaking world of business in Hong Kong, where there is a tendency to play

Regression analysis shows an R2 of 26.1, 36.1 and 51 for Hebrew, German and Spanish, respectively.
4 This is the same study referred to earlier in connection with Ide (1989).
518 L.N.T. Yeune / Journal of Pragmatics 27 (1997) 505-522

down hierarchical differences, even if only ostensibly. This is also evident in the first
name address terms used among business associates, and between superior and sub-
ordinates.

5.2. Chinese sample

The factor of imposition does not have a statistically significant impact on the
choice of politeness expressions (0.0615), although the p-value of 0.0615 is not too
far away from the acceptable level of 0.05. The other two factors do not show a sig-
nificant effect on linguistic variation either,
In interpreting these results, one must bear in mind the fact that most of the Chi-
nese informants found that the deferential expressions of classical origin are not gen-
erally different from the plain Modern Standard Chinese expressions, which appear
to be less deferential.
Furthermore, in the Chinese socio-cultural context, imposition may not be viewed
in the same way as in the English-speaking world. First of all, there is a difference
in degree. The wish not to impose on others may not be of the same overriding
importance as in a Western individualistic culture. As other researchers (Matsumoto,
1988; Blum-Kulka, 1989; Nwoye, 1992) have pointed out, in a more group-oriented
culture such as Japan, Israel, and the Igbo of Nigeria, imposition is not as important
a consideration as in the Western world. In the Chinese context, the principle regard-
ing on-record appropriateness and that of reciprocity must be considered. According
to the former, on-record requests are only made when one is in a position to do so.
Otherwise, one will choose to go off-record or use an intermediary to do the work on
one's behalf (Shih, 1988; Silin, 1976). Here, one can even argue that the same con-
textual factors of social distance, power, and imposition are equally at work in decid-
ing whether or not it is appropriate to go on record. But exactly how the factors trig-
ger off a threshold effect for on-record requests lies outside the premises of the
present research. Nevertheless, in the Chinese case, it is of vital importance to decide
what constitutes an appropriate on-record request, that is, one that can be broached
directly without loss of face. The importance of on-record appropriateness also
explains the indirect strategies of hinting and going through a third party used by
Chinese subjects when dealing with what is recognized as face-threatening requests
in role-plays (Nash, 1983) and questionnaire surveys (Shih, 1988). Such strategies
were adopted much more frequently than did their American counterparts in the
studies. But when requests are actually made as is appropriate, the principle of reci-
procity in Chinese society (Yang, 1987) reduces the burden of imposition. The
resulting strategy is to stress the sincerity of the gratitude felt as an indication of the
desire and commitment to reciprocate at some future date. This enables them to view
the imposition of a request with a more positive orientation than the Westerners. For
example, the item with the number one average score for politeness has such a mark-
ing (see above, p. 513, sentence 1:

Yu moung ........ baatsing gaamming

If granted ...... overwhelmingly grateful
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 519

Part of this strategy is to boost the strength of the expression of gratitude.

In contrast, there is a tendency in English to tone down the acknowledgment of
thanks with a modal and a conditional, such as in "I would be grateful if ...", so as
not to sound too imposing. According to Holmes (1984), the use of a booster as an
intensifier of positive affective meaning is a mark of a positive politeness strategy.
On the other hand, the use of a mitigating device attenuates negative affective feel-
ings. The latter strategy is characteristic of the negative politeness which predomi-
nates in the Anglophone cultures (Brown and Levinson, 1987; Blum-Kulka et al.,
1989).
Indeed, one can go on to argue that there may even be a difference in kind when
it comes to Chinese politeness. Some speech acts otherwise regarded as imposing in
the Anglo-Western culture are not viewed as such in the Chinese context. Gu (1990)
has shown that Chinese invitations are given with a lot of insistence, which appears
to be completely contrary to the negative strategies typically used for invitations in
English. The latter are characteristically marked with mitigating devices such as
modals, 5 as if for fear of imposing too much on the invitee. To the Chinese, the
expression of thanks after being granted a request (Gu, 1990), or in anticipation of it,
is intuitively a positive politeness strategy. Unlike in English, thanking is not a face-
threatening act to be redressed with mitigation before the request is granted. Nor is
it a tainted strategy to be avoided if possible. 6 As far as making requests is con-
cemed, Chinese people are debt-sensitive rather than debt-aversive. According to the
principle of reciprocity, by making a request, the requester assumes the obligation to
repay the requestee many times over, which would appear to be to the latter's bene-
fit. The fact that this is a commonly-accepted principle can be attested by the many
Chinese maxims about its importance. Two such examples are 'lei ging ngo yaat
chet, ngo ging lei yat cheung' (favor and respect must be returned many times over)
and 'loy yi baat wong fei laai yaa' (lack of reciprocation is improper decorum). In
the light of certain repayment, the imposition would be much less of a question. At
the same time, it would not be surprising that the eager acknowledgment of indebt-
edness is used as a positive strategy.
The factors of social distance and power are not at all statistically significant in
the Chinese sample. As mentioned earlier, Chinese is basically not an honorific lan-
guage like Japanese, where linguistic choice is strictly governed by the relative sta-
tus and social distance of the interactants. There are certainly honorific terms in clas-
sical Chinese, but many of these are absent in Standard M o d e m Chinese. It would
seem that if they are retained at all in writing, they have become stylized formulae,
acting as a mark of formal style rather than as a politeness strategy. Thus, contrary
to expectation, at least in the case of written requests, formal classical style is not
necessarily the equivalent of politeness. Classical Chinese and M o d e m Standard

5 An example would be "I was wondering whether you might be free this Saturday. My wife and I'd
love to have you come and have dinner with us".
6 Severalnative English speakers told me that they were surprised and somewhat offended when they
saw letters asking for requests end with a note of thanks. To them, it seems too presumptuous on the part
of the letter-writer who makes the request, implying that the request will be granted.
520 L.N.T. Yeung / Journal t~f Pragmatics 27 (1997) 505-522

Chinese are just two alternative written styles, the latter certainly being the more
popular one. In the present sample of business correspondence, both these different
styles are used. It appears that formal classical style is a matter of individual choice,
rather than social discernment in the sense postulated by Ide (1989). This is in fact
what has been suggested by some of the Chinese experts referred to earlier. The sta-
tistical results certainly lend support to this view.

6. Conclusion

Of the three factors Brown and Levinson (1987) have postulated as having an
independent effect on the choice of politeness expressions, only the factor of impo-
sition has a statistically significant impact in the English data of the present study.
The other two factors (social distance and relative power) do not show an indepen-
dent impact of statistical significance. The three taken as a whole have an effect, but
not much greater than the factor of imposition alone. On closer examination, it
appears that imposition reflects, at least to a certain extent, the other two factors of
relative power and social distance. The latter two certainly form part of the consid-
eration in assessing the imposition of the request.
The type of discourse obviously makes a difference as to how much influence the
three factors have on linguistic choice. The conventional expectations of business
correspondence to create an impression of being friendly, efficient, and not status-
conscious probably interfere with the linguistic choices which would otherwise have
been made in similar face-to-face and more status-oriented situations. Whether the
factors of social distance and relative power have a higher predictive power in other
types of discourse has yet to be verified by further research.
The Brown and Levinson framework (1987) does not seem to work for the Chi-
nese data in the study. The Chinese situation is complicated by the fact that there are
two styles: Classical Chinese and Modem Standard Chinese. There is no unanimous
agreement that one is inherently more polite than the other, although the former may
seem to be more deferential. Besides, the Chinese appear to have a somewhat differ-
ent system for the choice of politeness strategies, which is not accurately reflected by
the factors postulated by Brown and Levinson. For example, the importance of on-
record appropriateness means that the Chinese have to make a decision whether it is
a judicious proposition to make an on-record request. Once that is decided, they need
not redress it in the same way or to the same degree as their Anglophone counter-
parts. The strength and prevalence of the principle of reciprocity in the Chinese con-
text is another complementary factor explaining this difference in strategy. If one
wants to measure the marked effect of imposition on variation in the Chinese choice
of politeness strategies, off-record strategies should be included as well.

References

Blum-Kulka, Shoshana, 1987. Indirectness and politeness in requests: Same or different'? Journal of
Pragmatics 1 l: 131-146.
 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522 521

Blum-Kulka, Shoshana and Juliane House, 1989. Cross-cultural situational variation in requesting
behavior. In: S. Blum-Kulka, J. House and G. Kasper, eds., Cross-cultural pragmatics: Requests and
apologies. Norwood, NJ: Ablex.
Blum-Kulka, Shoshana, Juliane House and Gabriele Kasper, 1989. Investigating cross-cultural pragmat-
ics: An introductory overview. In: S. Blum-Kulka, J. House and G. Kasper, eds., Cross-cultural prag-
matics: Requests and apologies, 1-36. Norwood, NJ: Ablex.
Brown, Penelope and Steven Levinson, 1978. Universals in language usage: Politeness phenomena. In:
E. Goody, ed., Questions and politeness: Strategies in social interaction, 56-289. Cambridge: Cam-
bridge University Press. [Reissued 1987 as Politeness: Some universals in language usage. Cambridge
University Press.]
Cherry, Roger D., 1988. Politeness in written persuasion. Journal of Pragmatics 12: 63-81.
Dancygier. Barbara, 1993. Interpreting conditionals: Time, knowledge, and causation. Journal of Prag-
matics 19: 403434.
Drake, Bruce H. and Dennis J. Moberg, 1986. Communicating influence attempts in dyads: Linguistic
sedatives and palliatives. Academy of Management Review 11(3): 567-584.
Ervin-Tripp, Susan, 1976. Is Sybil there? the structure of some American English directives. Language
in Society 5: 25~56.
Fraser, Bruce, 1990. Perspectives on politeness. Journal of Pragmatics 14: 219-236.
Garcfa, Carmen, 1993. Making a request and responding to it: A case study of Peruvian Spanish speak-
ers. Journal of Pragmatics 19: 127-152.
Giv6n, Talmy, 1989. Mind, code and content: Essays in pragmatics. Hillsdale, NJ: Erlbaum.
Goffman, Erving, 1967. Interaction ritual: Essays on face-to-face behavior. New York: Anchor Books.
Gu, Yueguo, 1990. Politeness phenomena in Modern Chinese. Journal of Pragmatics 14: 237-257.
Halliday, M.A.K. and Ruqaiya Hasan, 1985. Language, context, and text: Aspects of language in a
social-semiotic perspective. Deakin, Victoria: Deakin University Press.
Held, Gudrun, 1992. Politeness in linguistic research. In: R.J. Watts, S. Ide and K. Ehlich, eds., Polite-
ness in language: Studies in its history, theory and practice, 131-154. Berlin and New York: Mouton
de Gruyter.
Hill. Beverly, Sachiko Ide, Shoko Ikuta, Akiko Kawasaki and Tsunao Ogino, 1986. Universals of
linguistic politeness: Some quantitative evidence from Japanese and American English. Journal of
Pragmatics 10: 347-371.
Holmes, Janet, 1984. Modifying illocutionary force. Journal of Pragmatics 8: 345-365.
House, Juliane, 1989. Politeness in English and German: The functions of Please and Bitte. In: S. Blum-
Kulka, J. House and G. Kasper, eds., Cross-cultural pragmatics: Requests and apologies, 96-122.
Norwood, N J: Ablex.
lde, Sachiko, 1989. Formal forms and discernment: Two neglected aspects of universals of linguistic
politeness. Multilingua 8(2/3): 223-248.
Kasper, Gabriele, 1990. Linguistic politeness: Current research issues. Journal of Pragmatics 14:
193-218.
Koike, Dale April, 1989. Requests and the role of deixis in politeness. Journal of Pragmatics 13:
187-202.
Kress, Gunter and R. Hodge, 1979. Language as ideology. London: Routledge.
Lakoff, Robin, 1973. The logic of politeness, or minding your p's and q's. Chicago Linguistics Society
9: 292-305.
Leech, Geoffrey, 1983. Principles of pragmatics. London and New York: Longman.
Matsumoto, Yoshiko, 1988. Reexamination of the universality of face: Politeness phenomena in Japan-
ese. Journal of Pragmatics 12: 403-426.
Matsumoto, Yoshiko, 1989. Politeness and conversational universals observations from Japanese. Mul-
tilingua 8(2/3): 207-221.
Nash, Thomas, 1983. An instance of American and Chinese politeness strategy. RELC Journal 14(2):
December 87-98.
Nwoye, Onuigbo G., 1992. Linguistic politeness and socio-cultural variations of the notion of face. Jour-
nal of Pragmatics 18: 309-328.
Perkins, R. Michael, 1983. Modal expressions in English. London: Frances Pinter.
522 L.N.T. Yeung / Journal of Pragmatics 27 (1997) 505-522

Shih, Yu-hwei, 1988. Conversational politeness and foreign language teaching. Taiwan: Crane Publish-
ing.
Silin, Robert H., 1976. Leadership and values: The organization of large-scale Taiwanese enterprises.
Cambridge, MA: East Asian Research Centre, Harvard University Press.
Stubbs, Michael, 1986. 'A matter of prolonged field work': Notes towards a modal grammar of English.
Applied Linguistics 7(1), Spring: 1-25.
Werkbofer, Konrad T., 1992. Traditional and modern view: The social constitution and the power of
politeness. In: R.J. Watts, S. Ide and K. Ehlich, eds., Politeness in language: Studies in its history,
theory and practice, 155-202. Berlin and New York: Mouton de Gruyter.
Yang, Lien-Sheng, 1987. The concept of pao as a basis for social relations in China. In: Chinese thought
and institutions (translated version). Hong Kong: Chinese University Press.
Zhang, Delu, 1991. Role relationships and their realization in mood and modality. Text 11(2): 289-318.