Chapter 3 Methodology
3.3 Data Analysis
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
paying several previous visits. On these previous visits, the observer spent some time with the children and the parents in order to get familiar with the children and also make the recording task clear to the parents. With such previous visits, the children would thus get used to the presence of the observer and the camcorder. During data collection, it was inevitable that children might occasionally look at the camcorder or observer, but this did not affect the children’s performance of speech.
During each visit, the observer would not start the recording until the children had been used to the presence of the observer and the camcorder. Each session of the recording lasted for one hour or so. Normally, the recording was not suspended unless it was necessary to do so — when the children needed to use the bathroom, for
example.
The overall length of the data examined in the study was about nine hours long.
All the recorded and observed conversations were further transcribed into Chinese characters.4 All the data were transcribed according to the CHAT format suggested by the CHILDES project (MacWhinney, 2000). The transcribing conventions are shown in Appendix A.
3.3 Data Analysis
As an endeavor to examine children’s requests and politeness in their requests, cases of requests produced by children in the data were identified first. Then the
4 The transcribing was conducted by the assistants of the Child Language Acquisition Lab of the Graduate Institute of Linguistics affiliated with NCCU, Taipei, and the researcher of this study is one of the assistants.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
politeness involved in each case of requests were examined.
3.2.1 Cases of requests
Given the discussion and review on speech acts and requests in the previous chapter, cases of requests in the data were identified according to the principles in the following. All utterances produced by children were first functionally determined whether they convey an illocutionary act of request in the immediate context. As discussed above, an utterance was identified as a case of requests according to the addressee’s response or reaction to the utterance, i.e., the perlocutionary act of the utterance. For example, such utterance as The water is boiling can convey a request if the addressee of the utterance does something to the boiling water as a response. The utterance may only be considered as a statement, if the addressee does not do any act in reply, otherwise, for example, when talking about pictures or illustrations in a book.
After identified with the illocutionary act, a request case was then to be analyzed according to its linguistic form. As discussed in the literature on speech acts and children’s request repertoire (e.g., Ervin-Tripp, 1976; 1977, Garvey, 1975; Gordon &
Ervin-Tripp, 1984; Searle, 1975), requests may be conveyed via various linguistic forms. The formal identification was done with regard to the syntactic structure as well as other linguistic elements with which a request was conveyed. A request can be encoded with imperative forms, such as gei wo ‘give me’, dakai ‘open it’, and he shuei ‘drink the water’; interrogative forms, such as ke-bu-keyi gei wo ‘Can you give
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
it to me?’; and declarative forms, e.g., Wo xiang he shuei ‘I want to drink water’ and Dianhua xiang-le ‘The phone is ringing’. With the formal identification, a request case was then coded accordingly.
In addition to the simple syntactic analysis, a request was also considered in terms of complex linguistic forms. Complex linguistic forms include a tag question attached to a declarative or an imperative, for example:
(1) Bang wo dakai, haoma/hao-bu-hao?
Help me open all right?
‘Help me open it, all right?’
(2) Wo xiang he shuei, keyi ma?
I want drink water, can PRT
‘I want to drink water, can I?’
The tag questions, hao-bu-hao or haoma, attached to the request utterance can in turn serve as a syntactic test for imperative utterances: an imperative utterance can be appended to with such tag questions while other non-imperative utterance cannot, e.g.,
*Wo bu hui nong, hamao ‘I can’t do it, all right?’.5 Other complex linguistic forms could be embedded sentences. This type of complex forms may include interrogatives embedded in a declarative as an indirect question, for example:
5 I am thankful to Prof. Miao-ling Hsieh, one of the committee members, for her suggestion to provide this syntactic test for imperative utterances.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(3) Wo xiang-shuo ni shi-bu-shi keyi dakai chuanghu.
I think you is-not-is can open window
‘I wonder whether you can open the window.’
The above complex linguistic forms were analyzed respectively as tag questions ((1)
& (2)) or embedded questions (3). A last type of complex forms could be reduced imperatives. Some cases of requests could be conveyed without encoding the action linguistically, but with the desired objects instead. For example, in order to get a book, the speaker could request the addressee to get the book by simply saying Shu ‘the book’, instead of Shu nalai ‘Bring me the book’. Such case was thus considered as reduced imperatives, if observed in the data.
Moreover, request cases were further considered in terms of lexical forms utilized to encode a request intent. As pointed out in the literature (e.g., Mei, 1994;
Nakamura, 1996), certain lexical forms may serve to mark a request intent as well as to defer to politeness. In Mandarin Chinese such lexical items may include qing
‘please’, bang ‘to help’, and mafan ‘please’, for example:6
(4) Bang wo dao shui.
Help me pour water
‘Help me pour water.’
6 The lexical form bang was considered as a request marker on the basis of Sealey (1999). She points out that children, when requesting, may take advantage of their ‘being a child’ to ask the addressee perform an act. By so doing, children are indeed showing their addressee their inability to do an act. In Mandarin, the lexical form bang serves a similar function to indicate one’s inability.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(5) Qing guan men.
Please close door
‘Please close the door.’
Thus, such lexical forms were particularly identified, if found in the data. With the above formal identification, this study may then collect the linguistic elements, both syntactic structures and lexical items, which are utilized by Mandarin-speaking children to convey their requests, and hence to amass their request repertoire.
Apart from the formal identification of request instances, in response to the distinction between direct and indirect requests proposed in the previous chapter, a request was further considered in terms of logical inference. Despite the fact that a request intent, direct or indirect one, can be conveyed with various linguistic forms, such as I want…, Can you…, May I…, I am wondering if…, etc., the intended illocutionary act can be explicitly or implicitly encoded in the locutionary act or the proposition of the utterance. As discussed both in Searle (1975) and Gordon and Ervin-Tripp (1984), a request intent can be conveyed via conventionalized indirect means. By utilizing conventionalized linguistic structures to encode his/her intended act, the speaker can convey a request intent via manipulation of the felicity conditions;
s/he can query the preparatory condition or express the sincerity condition. Whichever means the speaker chooses to convey a request, the locutionary act or proposition of the utterance may indicate the intended act to be performed by the addressee, and hence the addressee can effortlessly perceive the illocutionary act without any further inference. On the other hand, a request intent may be conveyed implicitly in the locutionary act, without specifying the intended act clearly in the utterance. Requests
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
conveyed by this means, according to Searle and Gordon and Ervin-Tripp, may be considered as non-conventionalized indirect requests. When receiving such requests, the addressee needs to undertake a series of inference in order to procure the intended illocutionary act. As pointed out by Gordon and Ervin-Tripp, children may be able to perform requests via such non-conventionalized means as well as conventionalized means, and thus this study also considered both means to requests. Therefore, a request intent, once identified, was further coded with respect to the explicitness of the illocutionary act. Those requests with explicit illocutionary act in the proposition were considered as requests with explicature and those without explicit act were considered as requests via implicature.
As mentioned above, an utterance was identified as a request with respect to its surrounding context, particularly the addressee’s response, namely the perlocutionary act. The addressee’s response to a request utterance was thus considered. Each response was analyzed according to its proposition content as positive compliance, temporization, alternative, or non-compliance. Positive compliance referred to the cases in which the addressee either acknowledges the request verbally with
affirmative responses or complies with the request non-verbally by carrying out the intended act. When the addressee did not immediately comply with the request in the adjacent turn, but in a further following turn, the response was thus considered as temporization. In addition, when receiving a request, the addressee may offer an alternative to the intended act, instead of performing the intended act immediately.
Cases like this were considered as alternatives. At last, a request may be rejected directly or indirectly. The children’s requests may not be complied by the addressee, neither in the contiguous turn nor in a further following turn before another request or
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
any reformulated request is performed. Such cases were thus analyzed as non-compliance.
As a recap, an utterance performed by the children in the study was analyzed with respect to its locutionary act or proposition of the utterance, illocutionary act, and perlocutionary act. Cases with the illocutionary act of requesting or directive were identified as the target cases of the present study, namely, request, judged by the illocutionary and perlocutionary acts the utterance may convey. Then, each request case was further analyzed according to its locutionary and perlocutionary acts. In terms of the locutionary act, a request case was formally analyzed with respect to imperatives, declaratives, and interrogatives. Also, the locutionary act of a request was analyzed as to whether the intended act was explicitly conveyed or implied in the proposition of a request. Finally, the addressee’s responses to a request, i.e., the perlocutionary act, were considered as to whether the request was immediately complied, temporized, altered, or rejected.
3.2.2 Cases of politeness
Given instances of requests identified, each request instance would then be examined according to the politeness associated with it. The politeness was considered under the framework of Ervin-Tripp et al.’s (1990). In their study on English-speaking children’s politeness, Ervin-Tripp et al. analyzed the politeness in children’s control acts in terms of social indices, social tactics and persuasive tactics.
According to Ervin-Tripp et al., social indices refer to linguistic markers that reflect
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
or maintain or alter the social relations and status between interlocutors (Ervin-Tripp, p.c.; Ervin-Tripp et al., 1990; Levinson, 1983). Such indices or deixis include address terms, honorifics, or other indices.7 Social tactics refer to strategic uses of social indices and linguistic elements to mitigate or hedge a request so as to adhere to the requirement of politeness. Persuasive tactics refer to reasons or justifications that children provide after or before the request at the first attempts or retries (Ervin-Tripp, p.c.; Ervin-Tripp et al., 1990). Based on the distinction among social indices, social tactics and persuasive tactics, the politeness associated with all the request instances would be considered and coded accordingly.
In Mandarin Chinese, there are a number of address terms, particularly the terms used to address family members. These address terms and other honorific titles are good example of social indices or deixis (Levinson, 1983). Normally, it is obligatory for the speaker in Chinese culture to address the addressee with proper address terms when there is an obvious social distance or difference in status between them. For example, in child-mother conversation, children should address their mothers with the address term mama ‘mother’. Thus, a request utterance would be encoded with
respect to whether an address term was used, when observed in the data.
As to the consideration of social tactics, the focus of this study would be on children’s strategic uses of different request forms with different directness of illocutionary force with respect to the requirement of politeness in the immediate context. The strategic uses of linguistic elements to mitigate or aggravate a request would considered in terms of the coding used for request forms. As mentioned above,
7 According to Levinson (1983), it is conventional for linguists or pragmaticians to refer to such terms as social deixis while for philosophers and sociologists to refer to them as social indices. Following the linguistic convention, the study will use social deixis.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
the syntactic structure(s) utilized to encode a request intent would be considered as to the syntactic constructions used to encode requests and the utilization of accompanied linguistic elements, such as sentence-final particles or tag questions . Given such syntactic coding, the politeness of a request would not be coded again separately so as to avoid redundancy.
As pointed out by Garvey (1975) and Gordon and Ervin-Tripp (1984), a speech act such as request may not solely consist of only one utterance conveying the illocutionary intent. A request can be conveyed via a structured sequence wherein persuasion is a part.8 As mentioned in Gordon and Ervin-Tripp (1984), a persuasion or persuasive adjuncts can precede or follow the intended request, and it can be
‘reasons, promises, threats, and so forth, which justifies the request or persuade the hearer’ (Gordon & Ervin-Tripp, 1984, p. 301). With such persuasion, the speaker not only takes into consideration the addressee’s negative face, but also enhances the effectiveness of his/her requests, as reported by Ervin-Tripp et al. (1990). On the basis of Ervin-Tripp et al.’s report, persuasion preceding or following a request would be considered as well, for its effect on the success of a request. Therefore, a request was coded with or without a persuasion and then further judged its politeness.
In summary, the study followed Ervin-Tripp et al.’s (1990) study, and also
analyzed politeness in terms of social deixis or indices, strategic uses of request forms, and persuasive tactics. A request instance was coded whether and which social deixis was utilized, whether linguistic forms were manipulated to hedge or aggravate a
8 According to Gordon and Ervin-Tripp (1984), the sequence of a request or an instrumental act contains the following components, including attention-getters, framing moves, persuasive adjuncts, instrumental moves, responses, and remedies (Gordon & Ervin-Tripp, 1984, p. 301), and each of such components is optional. A request can be conducted with any one of these elements; it does not require all the elements to prevail.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
request, and whether a persuasion was provided. With such coding, the politeness of a request would be further considered and discussed.
3.2 The Coding System
Given the above identifying methodology of requests and politeness above, the conversations to be investigated in the study were coded accordingly. The coding format is in the CLAN format, following the one suggested in the CHILDES project (MacWhinney, 2000). The codes in the coding system are explicated as follows:
I. Requests A. Form:
Theoretically speaking, the default form used by speakers to encode a request is not imperatives. Imperatives are the default form of commands (Levinson, 1983; Searle, 1975). Empirical evidence provided in studies on children’s pragmatic development, however, suggested that children, particularly when requesting their mothers to do an act, tend to use imperatives. Therefore, imperatives in the data were also judged with respect to the context and
determined if they served the function of request. The formal categories included in the coding system here were based on Ervin-Tripp’s (1977) repertoire of children’s requests and the data observed in this study. The coding used for formal elements is as follows:
a. Imperatives: imperatives refer to the syntactic structure with a
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
covert second person subject. They can also be further modified or aggravated with other linguistic forms, such as hedgers,
sentence-final particles or tag questions. Based on the overt linguistic elements appended to an imperative utterance, imperatives could be further identifies as (i)simple imperatives (PIP)9: referring to the cases of typical imperatives without any lexical or syntactic modification or aggravation in any forms; (ii) imperatives with sentence-final particle (IPP): referring to the cases of modified imperatives, particularly with sentential particles, such as o ‘oh’, ma,
‘a question particle’ a ‘ah’ and la ‘an interjective particle’.10 (iii) imperative with a tag (IPT): referring to the cases of imperatives modified with tag questions, such as A-not-A tag and question markers, including haoma ‘all right’, keyima ‘OK’, xingma ‘OK’, etc.
b. Interrogatives: this category refers to any interrogative forms, including interrogatives involving wh-words, interrogatives answered with simple ‘yes’ or ‘no’, and
interrogatives embedded in other syntactic structures, such as another interrogative or a declarative. Theoretically speaking, interrogatives seem to be the basic form used for requests (Levinson, 1983; Searle, 1975). Hence, interrogatives were
9 Letters in the parentheses hereafter refer to the codes used in the coding system of this study.
10 These particles has been discussed in Hsu (1996; 2000).
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
also taken into consideration here. The cases of interrogatives observed in the current data and the codes of them includes:
(i) WH interrogative (WHI): referring to the WH
interrogative used to convey a request intent; (ii) A-NOT-A interrogative (ANA): referring to the requests expressed with such A-not-A forms as ke-bu-keyi or neng-bu-neng ‘can;
could’ in the matrix clause; (iii) Embedded interrogative (EMI): referring to any interrogative forms that are embedded in other structures, such as an interrogative or a declarative;
and (iv) yes-no interrogative (YNQ): referring to an
interrogative form that requires either ‘yes’ or ‘no’ as reply.11
c. Declarative (DEC): requests may also be conveyed with a declarative. Any request cases that were neither conveyed with interrogative nor imperative were subsumed in this category. Request forms in this category cannot be tagged with such tag questions as hao-bu-hao or haoma. This category excluded WANT/NEED structure, which was considered separately as a different category, since some studies has pointed out a specialized function of such structure in child discourse (e.g., Gordon & Ervin-Tripp, 1990;
Ervin-Tripp, 1977); children were found to rely heavily on the structure to perform their requests to their mothers.
11 The coding system here only included yes-no interrogative, since this was the only type of interrogative forms observed in the current data. Other interrogative forms may include WH-interrogatives and embedded interrogatives, but these were not found in the data at hand.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
d. WANT/NEED structure (WAN): it has been discussed in the literature (e.g. Blum-Kulka, 1990; Ervin-Tripp, 1977; Gordon &
d. WANT/NEED structure (WAN): it has been discussed in the literature (e.g. Blum-Kulka, 1990; Ervin-Tripp, 1977; Gordon &