C ATEGORIZATION S CHEME

informants were unevenly distributed in terms of social factors, including competence, education, gender, and age. As a corollary, it is prone to yield a biased analysis.

Studies on language attitudes toward English as a field of inquiry were usually targeted at either English learners at various levels or students in schools. Therefore, conceptualizations of language by citizens from all walks of life, for example, those in service industry, are less explored. Relying on interview data for generalizations is inclined to result in biased analyses due to the lack of references. These concerns show that the interview data are insufficient to develop a sound analysis in this study.

Consequently, the study also selectively collected interesting editorials, news articles and messages from online social networks such as Facebook and BBS. These materials were approached with the theoretical frameworks of critical discourse analysis.

3.3CATEGORIZATION SCHEME

The section outlines a description of how data are analyzed in the study. Yet, the analysis needs to start with identifying what interview data are. Interview data, similar to naturally occurring data, are also conjectures between appropriate language competence and social positioning. Interview data are therefore interactional even though they are usually treated as artificial. Interview data are interactional in that they reveal the social positioning of both interviewers and interviewees. Contents and flows of interviews are manipulated by interviews, so interviewers’ control affect degrees of how interactive interviews are. Consequently, the influences of interactions and interviews as one type of social events are acknowledged in the study.

34 3.3.1 Interview data as interactional

Interview data are taken as interactional in the study for two reasons. First, though interview data have been recognized as artificial, an interview is a speech event which involves at least 2 parties talking to each other. Thus, interaction takes place. However, it also has to be noted that the interaction in interviews differs from the one taken place in natural conversation. Since interview data are interactional, the interaction levels are also possible sites where language ideology is located.

Interviews as a method in data elicitation have been criticized for the decontextualization of data, as argued by Briggs (1986). However, De Fina (2011) suggests that interviews generate valuable interactional data. Worthem et al. (2011) also point out that ways in which informants phrase their responses reveal their social positioning in relation to interviewers and other interviewees. Interviews can also generate interactional texts in which informants implicitly reveal their positioning through interaction.

Language ideology is argued to locate at different layers. Woolard (1998) discusses siting of language ideology (1998:06) while Kristiansen (2010) suggests different levels of language ideology (2010:60). Despite the differences in terminology, the attention centralizes the degrees of consciousness of language ideology. Woolard (1998) proposes that language ideology can be observed at three different levels, in linguistic practice itself, metalinguistic/metapragmatic discourse, that is, direct talk about language itself, and implicit metapragmatics. Kristiansen (2010) also identifies overt language ideology and covert language ideology.

Kristiansen argues that individuals perform linguistic behaviors different from what they self-report, denying a straightforward correspondence between linguistic behaviors and social values. Overt language ideology which individuals acquired

35

through learning prevails and reproduces itself in public domains. Another layer of covert language ideology refers to conceptualizations of variations deviant from official values. The distinction between overt and covert language ideology is assumed to attribute to the degree of awareness. Kristiansen (2010) claims that the distinction of the two levels of language ideology is a matter of “paying attention to”

(2010:60). These different approaches to the siting of language ideology argue for the viewpoint that language ideology is situated in different levels of consciousness.

3.3.2 Data categorization

For these two concerns, interview data cannot be treated as linear. Neither could interview data be considered a speech event which involves simple and impersonal question-and-answer exchanges. Therefore, the study divides data into two major categories based on informants’ responses.

The first category regards language ideology verbally described in direct responses to interview questions. In this type of language ideology, interviewees explicitly expressed their perceptions of English. Language ideology which was explicitly described by the interviewees in direct responses to the interview questions is discussed in Chapter 4.

The other category concerns semi-direct responses. That is, the informants initiated revealing their further thoughts invoked by interview questions but not directed at answering the questions. For example, when interviewees were asked about their daily use of English, some interviewees first directly answered the interview questions and related his/her stance to a particular event he/she had experienced. The semi-directed responses are therefore mostly narrations. How interviewees narrated and described the incidents reveal the social positioning and

36

their ideology of English. Besides, the interviewer’s control to interviews was significantly diminished. As a result, the reliability of data remarkably enhanced.

Semi-directed responses also allow an examination of underlying and context-specific conceptualizations of English which may differ from the established norms discussed in Chapter 4. Language ideology in semi-directed responses is presented in detail in Chapter 5.

Excerpts from both English users and non-English users are presented and analyzed. For the privacy concern, each informant was given a pseudo name. To reduce confusion possibly caused by numerous interviewees and excerpts, names of English users are squared in the excerpts shown in the following three chapters. Ann refers to the interviewer. Each line in each excerpt is numbered. Underlined are the parts in emphasis. Arrows point to the significant lines. The excerpts are transcribed in Chinese characters with English translations provided below each line. If interviewees codeswitch between Chinese and English, codeswitched parts are in italics for clarification.

37

C

4

L

I

D

R

Through interviews, it is found that the majority of the informants verbalized the necessity of English. This prevailing ideology of English as a necessity is apparent in the informants’ positive evaluations of proficient English users and deprecation toward incompetence in English. The prominence of English in Taiwan is still mostly attributed to symbolic and instrumental values which English has been thought to promise. As already discussed in the past literature (e.g. Chen 2006; Chen 2010; Tsai 2010; Wang 2000), English is recognized as strongly facilitative and instrumental career-wise. The instrumental motivations of English further accelerate and reinforce the positive images that English projects or is thought to project.

Commonly viewed as cultural and linguistic capital by general public, English has indexical values that stratify individuals into different social ranks based on English ownership, the sanction of legitimate English user (Widdowson 1994; Norton 1997; Higgins 2003). Hence, English is redefined as a commodity (Heller 2003, 2010).

In addition to exchange values, there is an inclination toward the emphasis on use values. With the number of English speakers increasing, competence in English begins to be recognized as basic in workplace. Hence, informants were gradually more attentive to prioritize use-value of English over its exchange value.

This chapter discusses language ideology of English explicitly described by informants in direct responses to the interview questions. Though both English users and non-English users demonstrate seemingly similar conceptualizations of English necessitation, the access to privileges brought by English influences to a great extent how both parties position themselves in relation to the other party in the increasingly

38

globalized context.