This thesis investigates, first of all, universal connective patterns among inter-related notions coded by repetitive, additive, and restrictive grams from a cross-linguistic typological perspective, and then it applies the insights obtained from the cross-linguistic comparison to solve disputed issues in the literature on Chinese repetitive adverbs, additive adverbs and restrictive adverbs respectively.

The Chinese repetitive adverbs hai (還), zai (再) and you (又), the additive adverbs dou (都) and ye (也), and the restrictive adverbs jiu (就) and cai (才) have long been an area of fascination due to their complex and distinctive syntactico-semantic properties. However, existing studies on such multifunctional adverbs are largely conducted from an intra-lingual point of view, which...[ Read more ]

This thesis investigates, first of all, universal connective patterns among inter-related notions coded by repetitive, additive, and restrictive grams from a cross-linguistic typological perspective, and then it applies the insights obtained from the cross-linguistic comparison to solve disputed issues in the literature on Chinese repetitive adverbs, additive adverbs and restrictive adverbs respectively.

The Chinese repetitive adverbs hai (還), zai (再) and you (又), the additive adverbs dou (都) and ye (也), and the restrictive adverbs jiu (就) and cai (才) have long been an area of fascination due to their complex and distinctive syntactico-semantic properties. However, existing studies on such multifunctional adverbs are largely conducted from an intra-lingual point of view, which makes the semantic analysis on these adverbs substantially depend on the introspective inquiry by individual researchers, leading to a natural consequence that different researchers may come up with different analyses on the same issue about the same adverb. To diminish this randomness, we need to take recourse to other analytical tools that may help us rid of this undesirable subjectivity. Cross-linguistic comparison is an ideal approach to allow us to transform the traditional introspective contemplation into a purely empirical investigation.

Among all available methods, the Semantic Map Model (SMM) is arguably the most powerful tool to present the outcome of cross-linguistic comparison. On one hand, it can provide a universal connective pattern of the related notions that is applicable to all languages; on the other hand, it can also provide a representation of the semantic network of a particular multifunctional gram in a particular language. Therefore, the SMM is adopted as the primary method to cater for the need of presenting both language-universals and language-particulars.

Based on first-hand data collected from 40 languages and second-hand data from reference grammars of more than 80 languages, we set up conceptual spaces centering on “repetition”, “additive focus” and “restriction” respectively. With the instantiations obtained through the constructing process, we discuss some important technical issues pertaining to the construction of conceptual spaces. These issues include how to combine the merits of the meaning-driven approach and the data-driven approach, how to dynamicize a conceptual space into a diachronic map, and how to integrate different conceptual spaces into one coherent chunk.

Enlightened by observations acquired from the conceptual space for repetitive grams, several controversies surrounding the semantics of Chinese repetitive adverbs, particularly hai and zai, are addressed. The first one is the semantic characterization of hai. After reviewing two extreme approaches which take the “abstractionist” and “polysemist” stands respectively, we argue that neither could objectively help determine whether to combine or to separate two related functions of hai. The cross-linguistic comparison, however, could maximally avoid this disadvantage as the decision to combine or separate completely depends on how the cross-linguistic data presents the related functions. Another issue is the high degree of polysemy of this adverb. The cross-linguistic approach guides us again in finding the regularities of the “high-degree-polysemous” grams and the “low-degree-polysemous” grams. It is found that hai and its counterparts which belong to the “high-degree” group could expand the “repetition” use to “continuation” or “supplement”, while other grams that belong to the “low-degree” group could only extend “repetition” to “increment” or “inverted-sequence”. The last issue concerns the differentiation of the repetitive meanings expressed by hai and zai. By employing the semantic features used in the construction of the conceptual space, we propose that the feature of “increasing” is more essential for zai to deliver the repetition function while that of “similar” is more essential for hai to deliver the same function.

The conceptual space for additive grams also provides insight into some disputed points regarding the Chinese additive adverbs dou and ye. In constructing the conceptual space, we find that the conceptual relations among “additive focus” (ADD), “universal quantification” (UQ) and “scalar trigger” (SCA) are more complex than seen among the other links in this domain. By delving into the relation between ADD and UQ, we uncover a prevalent mechanism that can contribute to the reading of UQ: merging the arguments of different propositions into one subject position. This observation greatly supports the sum operator account of the nature of dou. By examining the cross-linguistic data involving UQ and SCA, we point out that the behavior of expressing both UQ and SCA by one identical form is an exclusive feature of Mandarin dou in our sample. This synchronic observation and the diachronic investigation of dou both demonstrate that there are two distinct dou’s in Chinese. The other pair of closely-linked concepts, namely ADD-SCA, also assists us in pursuing an account for special behaviors of Chinese additive adverbs. By comparing ye with its cross-linguistic counterparts, we identify ye as a pseudo scalar trigger: it cannot trigger “likelihood” as an authentic scalar trigger does. However, NPI provides the context where ye can regain the same irrealis feature of “likelihood”. Therefore, compared with dou, ye shows a sharp contrast with respect to its distributions in affirmative and negative sentences.

In the same vein, we have applied the insight gained from the conceptual space for restrictive grams to solve some long-standing issues involving Chinese adverbs jiu and cai. Apart from suggesting a more objective categorization of jiu and cai, the cross-linguistic perspective also assists us in identifying two universal mechanisms for coding the “restriction” function. One of the mechanisms presents a non-scalar and non-time-related focus, whereas the other mechanism presents a scalar and time-related focus. Chinese jiu and cai are proved to be one pair of grams among many others across languages that adopt these two mechanisms respectively.

It is also suggested in this research that the use of the conceptual space is not restricted to explaining existing and known issues. The semantic map model as presented in this research is a promising tool by which many more problems could be identified and solved, be it in Chinese or in other languages.