Disagreeing in english and vietnamese

1.2. Empirical Study

1.2.1. Aims and methodology

1.2.1.1. Aims

This empirical study aims at getting the sufficient proof for the following hypotheses:

1. 
   Native speakers of English and Vietnamese may differ in assessing such socio-cultural parameters as age, length of time (familiarity), manner, occupation, setting, gender, and social status, which affect interactive talk.
   
2. 
   Speakers of English and Vietnamese may have different assessments of social situations.
   

1.2.1.2. Data collection methods and respondents

Sampling

To guarantee the validity and reliability of the research results, it is of great importance to strictly follow the sampling procedures and select an appropriate sample. According to Dörnyei (2003), each sample is actually a subset of the population and the sampling should be carried out in such a way so as to ensure this representativity. After giving a detailed description of sampling procedures that consist of (1) defining the sampling universe, (2) constructing sample stratification, and (3) fixing the sample size, Sankoff (1974: 22) clarifies Hymes’ notion of ‘speech community’ as regards the sample size:

A speech community sample need not include the large number of individuals usually required for other kinds of behavioral surveys. If people within a speech community indeed understand each other with a high degree of efficiency, this tends to place a limit on the extent of possible variation, and imposes a regularity (necessary for effective communication) not found to the same extent in other kinds of social behavior…. even for quite complex speech communities, samples of more than about 150 individuals tend to be redundant, bringing increasing data handling problems with diminishing analytical returns.

In the same vein, Dörnyei (2003: 74) suggests:

From a purely statistical point of view, a basic requirement is that the sample should have a normal distribution … the sample should include 30 or more people…. From the perspective of statistical significance… certain multivariate statistical procedures require more than 50 participants; for factor analysis, for example, we need a minimum of 100 but preferably more subjects.

Selection of regions

The empirical data were selected in two places: North America and Vietnam. These two places were deliberately chosen due to the present researcher’s ability to access the respondents, which saved her a considerable amount of time, energy and expenses.

North America is internationally known as a region with rapid development in the domains of economy, science and technology. It includes Canada and Northern States of the USA. Although English and French are legitimated official languages of Canada, the Canadian respondents are native speakers of English, either living in Toronto or coming there for their research or study. Toronto is the biggest city in Canada and belongs to Anglophone cultures. Some respondents of English are from other cities of Canada and US Northern states. North America is assumed a complex speech community that belongs to ‘less hierarchical societies’ where ‘status is allegedly far less marked in verbal and non-verbal interaction’ (Bargiela-Chiappini 2003: 1463) and ‘where individualism is assumed to be the basis of all interaction’ (Ide 1989: 241).

Hanoi, the political, economic and cultural center of Vietnam, situated in the North of Vietnam, has been famous for its historical and cultural traditions. From the socio-cultural point of view, Hanoi with its one thousand years of history, feudalist dynasties followed by communist regime, and the present ‘open-door’ policy offers abundant corpora for any researcher who attempts to go further afield the pure linguistic boundary to examine linguistic issues in relation to the wider socio-cultural context. Having undergone great socio-cultural and economic changes, the social structure of the Vietnamese has still been vertically hierarchical with emphasis on moral conduct and community-oriented solidarity (Nguyen D. H. 1995, Vu T. T. H. 1997, among others). As the home for many people coming from different parts of the country, Hanoi can be considered a conjunction of or meeting place for cultures and sub-cultures, and an ideal place to obtain data in terms of socio-linguistic representativeness.

Selection of respondents

The respondents, native speakers of English and Vietnamese, either indigenous inhabitants or permanent/life-time residents, are assumed to share the same sets of local socio-cultural norms, values and beliefs. They all have tertiary education and are ranked as middle-class citizens in regards to socio-economic conditions. This study takes as its chief objectives the perception and realization of disagreeing by native speakers of English and Vietnamese in their speech communities, therefore, the sample should be balanced and comparable in terms of age, gender and level of education, as Bauman & Sherzer (1974: 17) insist,

Linguistic descriptions must achieve both psychological and sociological validity. They must reflect the perspective not only of single individuals but also of social groups, networks or communities.

Individualism and privacy seem fundamental in Anglophone cultures, and native speakers are in favor of freedom from imposition and of actions. Thus, gaining assess to them appears to be problematic, let alone asking them to complete written questionnaires or recording their mundane casual conversation. To collect sufficient data for this study, the researcher spent ample time wandering around Toronto, especially the spacious campus of University of Toronto, its libraries and athletic centers meeting people. Among the respondents were ELI (English Language Institute) teachers coming from some US Northern States to Vietnam for their teaching. There was less complexity and trouble in administering the questionnaires among the Vietnamese respondents, interviewing them and recording their real-life interaction thanks to friendly or/and kinship networks.

Sample size

The elicited data were obtained from 100 English respondents (40 male and 60 female) and 100 Vietnamese respondents (50 male and 50 female). The respondents are supposed to belong to two homogeneous speech communities and possess the same sets of shared social norms, beliefs and values (Hymes 1974a-b, 1995), and be comparable and balanced in terms of education, economic condition, age, and gender. All of them have tertiary education or higher, MA or PhD, and they are ranked as middle-class citizens, aged from 18 to 72.

	GENDER		Total
First Language	Male	Female
English Count	40	60	100
English %	40.0%	60.0%	100.0%
Vietnamese Count	50	50	100
Vietnamese %	50.0%	50.0%	100.0%

Table 1 2: Gender correlation between English and Vietnamese respondents

The age groups were divided into two: Less than 30 and 30 or above, with the former consisted of 62 English and 50 Vietnamese, and the latter contained 38 English and 50 Vietnamese. The audio-taping data were gained from 8 tapes by 30 speakers of Vietnamese, and 6 tapes by 16 speakers of English. In addition, a number of English excerpts of mundane everyday speech used in this study are taken from the second source available in the literature owing to their availability and convenience.

	AGE GROUP		Total
First Language	Less than 30	30 or above
English Count	62	38	100
English %	62.0%	38.0%	100.0%
Vietnamese Count	50	50	100
Vietnamese %	50.0%	50.0%	100.0%

Table 1 3: Age group correlation between English and Vietnamese respondents

Even though the researcher is aware of the ideal balance as for gender and age group among the two groups of respondents, she should be satisfied with the present sample as after all it has offered a set of valid and reliable corpora for the empirical studies.

Written questionnaires

Written questionnaires have effectively been used in linguistic research by researchers like Ervin-Tripp (1969), Blum-Kulka (1982, 1989), Bayraktaroglu & Sifianou (2001), Bharuthram (2003) etc. as they help to collect a significant amount of data of controlled manner in rather a short time. The questionnaire in the form of Discourse Completion Task (DCT) used in the Cross-Cultural Speech Act Realization Project (CCSARP) has been chosen and modified (called Semi-Discourse Completion Task - SDCT) to adjust the purpose and limitation of the paper. The present author also makes use of the written questionnaire developed by Nguyen Q. in his 1998 Ph.D. dissertation (in Vietnamese). The English elicited data were obtained in Toronto, Canada in 2003, and the Vietnamese data were collected in Hanoi, Vietnam in 2003-4.

Like any other kind of written questionnaires, DCTs have the disadvantage of being unable to capture variables like hesitation, pauses, fillers, etc., which are typical features of spoken discourse responses. In addition, the respondents have more time to consider and reconsider their replies in writing than in spontaneous speaking, and they may provide more elaborate responses than those made in natural speech. Last but not least, it does not always seem to be an easy task to check the accuracy of elicited data, for people may wish to be seen and judged in a good light (Ackroyd & Hughes 1981: 83). Audio-taping data are used to make up for the drawbacks of written questionnaires.

Audio-taping

Some researchers tend to overestimate the advantages and underestimate the disadvantages of approaches or methods to explain their preference for the one, which they believe the most appropriate. However, the present researcher proposes to deploy more than one approach, i.e., to combine several approaches and methods so as to make good use of the advantages and minimize the disadvantages (Nguyen D. H. 1995, Cohen 1996, Gass 1996, Vu T. T. H. 1997, Lee-Wong 2000). By so doing the database used in this research can be considered quite sufficient for the empirical studies. It consists of the data collected from SDCTs, audio-taping of natural interactions by native speakers of English and Vietnamese, and excerpts from the recorded data deployed in the works by Sacks, Schegloff & Jefferson (1974, 1978); Schegloff, Jefferson, and Sacks (1977); Sacks & Schegloff (1979), Heritage & Drew (1979), Pomerantz (1975, 1978, 1984a-b), Levinson (1983), Goodwin & Goodwin (1987), Heritage (1997, 2002, forth.), Maynard (2003) and others. When necessary, the field notes taken by the researcher on the spot are referred to in order to make up for the inability to use videotaping.

One of the advantages of CA is that conversation analysts can use data from any available source to describe and analyze social actions provided that they are naturally occurring data. Analysts can deploy other researchers’ recordings together with their original for specific purposes of study. Psathas (1995: 53) asserts:

Because the researchers must make available, in transcripts and published extracts, the data on which their studies are based, other researchers may then examine the same, as well as additional, materials, and either replicate or extend the analyses first presented.

Confidentiality

The database used in CA is recordings, whether audio or video, and may be collected from any available source, provided that they should be mundane everyday talks occurring in natural settings. Protection of privacy and participants identities is essential, and as a result, permissions for recording should be obtained, and conversationalists anonymized (Psathas 1995, Dörnyei 2003). Aware of confidentiality as an important criterion in audio-taping and questionnaire administration, the researcher has used a different name or a letter to replace the S’s real name to anonymize all the respondents and coded the excerpts chose from the recorded data.

Data analysis

Central to this research is the speech act of disagreeing, its perception and linguistic realization. Hence, the data analysis methods are chosen to highlight issues involving disagreeing and its construct of form, function and meaning. The data analysis focuses on: (i) the structure of disagreeing and the use of honorifics,