Wh-Quantifier Interactions in Malagasy

Joachim Sabel

University Catholique de Louvian

In English-type languages, the syntactic relation corresponding to grammatical functions such as syntactic subject is very restricted and usually borne out by the external argument, whereas in languages such as Malagasy the syntactic subject can be borne out by a wide range of arguments. This difference correlates with distinctive properties of subjects and objects in these language types, for example, in the realm of overt A'-movement (topicalization, wh-question formation, relativization). Extraction patterns show a subject/object asymmetry in English-type languages and a subjects-only restriction in Malagasy-type languages. Concerning argument extraction, objects are more easily extractable than subjects in English (although both can be extracted), whereas in the Western Austronesian languages, the syntactic subject, but not the syntactic object may be extracted.

        In this talk, I will investigate argument asymmetries in Malagasy and English with respect to wh-quantifier interactions. Wh-quantifier interactions represent an area where English-type languages show subject/object asymmetries. However, the construction has not been examined in Malagasy and related languages (but see Saddy (1991) and Agüero-Bautista (2001) for some remarks on wh-quantifier interactions in Indonesian). I will argue that the data provide evidence for the following conclusions:

(i) Scope ambiguities with respect to wh-quantifier interactions only arise if the quantifier c-commands the base position of the wh-element (within a certain local domain).

(ii) In contrast to English, the respective derived subject in Malagasy is base-generated in the complement position of V.

        In his analysis of scope ambiguities, May (1977, 1985) has noticed that questions in which a wh-object is extracted across a subject quantifier, such as What did everyone [_VP buy t]?, are ambiguous in English. The question can be answered with an individual answer, such as An apple, or with a pair-list answer, such as John, apples; Mary, milk; Bill, bread... In contrast, if the subject is questioned and the quantifier occupies the object position, as in Who bought everything for Max?, the question is unambiguous. It can be answered only with a single answer, for example John. The phenomenon has been analyzed differently in May (1985), Aoun and Li (1993), and Chierchia (1991, 1993), but all these analyses are based on the view that wh-quantifier interactions display a subject/object asymmetry in languages such as English: Given that object wh-extraction proceeds from a structurally deeper position in the sentence than subject wh-extraction, a scope ambiguity arises only with respect to object wh-extraction.

        However, turning to Malagasy, we observe that this subject/object asymmetry is absent. The questions in (1) and (2) allow both for an individual answer and a pair-list answer:

(1)                                Inona                            no                                sasan'                                                                                                                                                                           [            ny                                             reny                                                                              tsirairay]                       amin'                ny                                 savony?         

what                               foc            pres-TT-wash                          the                  mother            every                                                with                            the             soap

'What did every woman wash with the soap?'

(2)                                Iza                                            no                                manasa                                                                                                                                     [                       ny                                             lamba               tsirairay]                                                                       amin'            ny                                 savony?

who               foc            pres-AT-wash                              the                  shirt                                 every                                                                                  with                the             soap

'Who washed every shirt with the soap?'

        I will argue that scope ambiguities with respect to wh-quantifier interactions arise if the quantifier c-commands the base position of the wh-element and that, for independent reasons, the logical object inona 'what'  in (1) and the logical subject iza 'who' in (2) are base-generated in the complement position of the verb (before extraction to Spec CP via Spec TP takes place). A similar variety of base-generated structures does not exist in English-type languages, where logical subjects cannot be merged as complements of V. Therefore, (2) is unambiguous in English. As will be shown, further empirical evidence for this analysis comes from wh-in situ questions in Malagasy, and from examples with long wh-extraction.