下载掌阅APP，畅读海量书库
立即打开
The theoretical framework that will be mainly utilized in this study is the Columbia School Linguistic theory, which holds the idea of one-form, one-meaning (Diver, 1969, 1995). Distinct from traditional grammar, the Columbia School theory (a framework within functional linguistics, which is opposed to formal linguistics), is sign-based and postulates that there are different meanings if the messages are different (Huffman, 1997, 2006; Reid, 2006). In the Columbia School framework, the main purpose of language is to communicate and there are extra-linguistic (i.e. psychological, strategic etc.) factors. That these factors may affect linguistic forms and should be taken into consideration is a tenet of Columbia School and not necessarily other linguistic schools. Take ego-centricity, for example, which means human beings generally consider their own experiences the most interesting ones (James, 1950; Kirsner, 1979). As people strive to be the center of attention, the words urge the hearer to put more attention on the speaker. Also, human beings tend to give more information when speaking about things in which they are interested (Reid, 1977; Diver, 1969, 1995).
Instead of language structure, this study emphasizes language use (i.e. how speakers/writers exploit the resources of their language). Hence, it is reasonable to study language use in real-world texts (i.e. in corpora), rather than in artificially created sentences (non-natural sentences) for qualitative and quantitative study. Therefore, corpus linguistics will play a significant role in this study. That is to say, this is a study of real texts, as opposed to Chomsky's splitting of performance from competence (Biber, Conrad, & Reppen, 2002; Chomsky, 1965). The goal of using corpora is to obtain the probability of certain language use in specific situations, in order to give a more explicit meaning of the demonstratives in both languages.
There are written languages and spoken languages. Spoken language, the informal-unplanned discourse in communicative situations, bears the following characteristics, which written language, a.k.a. literatures, does not have (Givón, 1979): first, COMMUNICATIVE STRESS: in a conversation, the speaker may be under communicative stress, where the speaker does not have time to plan before talking, and thus may not have the best communicative strategy and express him-/herself perfectly; second, TIME PRESSURE: in case of emergency, the speaker may not be able to take the best word choice under time pressure; third, DEGREE OF PLANNING: in a real-time conversation, a speaker does not have time to plan the utterance to a higher degree as a writer does (and of course, a writer can also go back and polish his/her writing before final submission); fourth, FACE-TO-FACE MONITORING: in a conversation, the speaker and the hearer have eye contact and gestures can be involved, which make the information sharing less dependent on the words; fifth, SHARED GENERAL PRAGMATIC BACKGROUND: in a conversation, there are several factors that are obvious and do not need to be signaled by utterance, such as topic familiarity, mood, context and so forth (p.105).
Based on the aforementioned characteristics, only written data will be analyzed in this study. There are two reasons why spoken data are not examined. First, unlike written data, spoken data does not provide background information of the speaker and the hearer. The relationship between the speaker and the hearer is not clear. We do not know their communication strategies and goals. In short, the lack of contextual information makes it difficult to account for extra-linguistic factors. Second, although speakers do not have the privilege of planning and editing their words during the discourse, they can use their intonation, gestures, and complexions to express themselves. Therefore, it is very likely, that not all the information is contained in the language. Consequently, there is a less precise description in the spoken data, which will inevitably affect the result of this study.
Due to the aims of this study, there are three main considerations when selecting the data sample. First, there should be two major texts. One should be in German and the other, in Chinese. It is necessary that both texts possess adequate instances of deictic and qualitative features of demonstratives in the target language. Second, both texts should involve discourse, culture, and social settings. Since language is for the purpose of communication, language content should not be taken away from reality; thus there is no ideal speaker or hearer, and there are special language uses, which may only be applied under this circumstance. In other words, language in both texts has to sound natural to native speakers and hence possesses a high degree of “communicative fidelity,” “a one-to-one correlation” between the sign and the message (Givón, 1979, p.108). Third, for the sake of a comparative study, it would be preferable to have parallel samples from both target languages with an official translation in the other language (Chafe & Danielewicz, 1987; Christensen, 1994, 2000; Li & Thompson, 1987).
Qualitative and quantitative studies will be conducted on both texts. Qualitative studies are supposed to explore the nature of demonstratives in the target languages. Note that, even in reproducing Chafe's pear stories (1980), a movie in which a set of consequent and simultaneous things happen while a farmer is harvesting pears, which contains strikingly small cultural differences among east and western countries, German speakers tend to use only one sentence to summarize the story. On the other hand, Chinese speakers tend to include moral judgments in the narration. Therefore, it is important to take cultural factors into consideration, because culture influences speakers' attitude towards facts and thus affect their word choice. Hence, in this book, for each example, the context will be given first and then the sentences will be analyzed based on the culture. In addition, the goal of the communication will be considered in order to uncover the meaning conveyed by the message. Moreover, examples will be compared to their counterpart sentences, which do not have a special communicative goal and are plain declarative sentences.
Quantitative analysis is necessary to this study as well. We would expect that human beings are rational and would make decisions based on their best interests. But in reality, we should not suppose that human beings are all ideally rational and well-behaved; first, human beings may not behave in the most rational or ideal manner; and second, human beings may be affected by various factors, either physical or emotional, and might not behave in the way they claim they would. Therefore, in this study, we need to “supplement subjective with objective data, specifically statistical observations” (Kirsner, 1979, p.361; cf. Contini-Morava, 1976). Therefore, in the quantitative part, this study will not declare which demonstrative will be used under which circumstance, without exception, but rather will identify general tendencies. A prediction is not expected regarding which one is used more commonly in real language, since language itself is creative and thus impossible to predict.
It is, however, not favorable to use translations (i.e. parallel texts, in cross-linguistic studies). Generally, using translation is recognized as biased, because translations from the original language into the target language cannot avoid being influenced by the translator's understanding. A translator must either be a native speaker of the original language or of the target language, as it is not possible to be a native speaker of two languages. Even if the translator is truly bilingual, the more frequently used language will override the one used less often and so the less-used language will be below the level of a native speaker. If a translator's understanding is inevitably added to the translation, the original meaning may be misunderstood or not be signaled in the translation.
However, using parallel samples (here the word parallel should be taken in the sense that the genre of text would be the same in the two languages, such as a German novel and a Chinese novel or a German essay on evolution and a Chinese essay on evolution) is a good way of having control of the usage of demonstratives in both the German and Chinese discourse. By the definition we have just given, parallel texts will likely contain the same or similar propositional contents. This enables a comparison of how demonstratives in the target languages operate within similar or the same propositional contents. In this way, differences between the target languages can be easily located and extra-linguistic factors largely controlled (Wu, 2004, pp.25—26).