Research Questions

European languages form compound quantificational expressions by a semantically transparent and compositional process of word formation which combines a quantifier and restriction-denoting expression: someone, everyone, no-one, somewhere, everywhere, nowhere etc. In many Asian languages, including, Japanese, Korean, Chinese, and Malayalam, on the other hand, a different strategy is employed: compound expressions are constructed from interrogatives in combination with a morpheme whose normal denotation is concerned with conjunction or disjunction: e.g. Malayalam, aar-um (who-and: anybody), ent-um (what-and: anything), eppoozh-um (when-and: always), aar-oo (who-or: somebody), ent-oo (what-or: something), eppoozh-oo (when-or: sometime). This second strategy, which we call the Q-word quantification strategy, its relation to the more widely studied strategies found in most Indo-European languages, and its consequences for Universal Grammar and linguistic typology form the core research questions of this project.
From the work that we have carried out already, it is clear that the Q-word quantification strategy is itself subject to parametric variation and, for that reason, we propose to investigate a range of genetically unrelated languages which exhibit the Q-quantificational strategy, including Chinese, Japanese, Korean and Malayalam with a view to determining the range of parametric variation that exists and of providing a theoretical analysis which will provide a principled basis to account for the observed variation. As an example, we observe that in the Malayalam examples cited above, a morphological process of suffixation is involved in forming compound quantificational expressions, whereas, in Chinese, quantification is accomplished syntactically through the satisfaction of an association requirement between the question-word and the quantificational element, which takes the form of a VP initial adverbial element: Shei dou xihuan ta (Who ALL like him: Everybody likes him). This has the consequence that universally-quantified direct objects must move from post-verbal position to one preceding the quantificational element: Women shei dou xihuan (We who ALL like: We like everyone). In Chinese, again in contrast with the other languages cited, there is no overt marking of indefinites, which are, however, restricted to modal and polarity contexts: Ni maile shenme ma? (You bought what QM: Did you buy anything?). The relationship between quantification and polarity contexts will form one of the topics of the investigation.
Another striking characteristic of quantification in the Q-word languages is that numerical partitives are typically not realised by determiners, but by splitting the range of quantification and the numerical component: e.g. Chinese: Pinguo, ta maile san-ge (Apples he bought three-Cl: He bought three of the apples). Korean also permits this kind of separation with the DP itself; e.g. John -i 2-kwen-uy chayk -ul ilkessta (John-nom 2-Cl -Gen book-acc read) vs. John -i chayk-ul 2-kwen ilkessta (John-nom book 2-Cl read: John read two of the books). We propose to explore this phenomenon and its relationship to other strategies of quantification in the languages concerned. On a more general level the study of different types of quantification is most likely to provide evidence bearing on one of the most vexed and controversial recent questions in comparative linguistics, namely that of semantic variation and parametrisation. The question that we would like to address here is to what extent the differences in quantificational strategies warrant the establishment of a Semantic Parameter.