Towards a Typology of Locative Inversion - Bantu, Perhaps Chinese and English - But Beyond?

Locative inversion (LI) is a construction that is very prominent in Bantu languages. It involves inversion of a locative with the logical ⁄ thematic subject. The inversion is accompanied by a reversal of grammatical functions whereby the locative becomes the subject. LI is associated with a special discourse function, that of presentational focus. Within Bantu there is quite some variation, especially concerning thematic restrictions and the robustness of locative morphology. While the subjecthood of locatives is relatively easy to diagnose in languages where they agree with the verb, an alternative analysis in terms of topicalization suggests itself when there is no agreement. LI has been described for very few other languages including English and Chinese. In both languages, the evidence for the subjecthood of the locative is limited or equivocal. LI thus appears to be a very rare phenomenon and this rarity may be the key to a fundamental understanding of the construction. 1. In Place of an Introduction: Locative Inversion in Bantu Locative inversion (LI) is a term that is predominantly used for a construction found in Bantu languages where a subject and a locative are linearly inverted. We will first introduce what is often taken to be the prime example of LI, i.e. LI in Chichewa. In this language LI involves a reversal of grammatical functions, and to make meaningful crosslinguistic comparison possible, we will take this to be a definitional property of LI (i.e. LI in the narrow sense). Thereafter we will turn to Bantu-internal variation, discussing thematic restrictions, locative morphology as well as cases of LI that are arguably not accompanied by a reversal of grammatical functions (and therefore do not represent instances of LI in the strict sense). Finally, we address possible instances of LI in other languages. Since only limited information is available on languages other than Bantu, this article will not be a balanced typology of LI but rather an overview over the available literature. The apparent rarity of LI in the languages of the world will be the topic of the conclusion. 1.1. THE NOUN CLASS SYSTEM Before we can address LI in Bantu, it is important to understand the nature of the Bantu noun class system (Katamba 2003): Most Bantu languages have a number of different noun classes which are similar to grammatical genders although the classification follows certain semantic principles. Each noun in Bantu inherently belongs to one class. What is special is that Bantu languages have a large number of such classes, sometimes more than 20. The noun classes usually appear as prefixes to the noun and trigger (often phonologically similar) agreement marking on demonstratives, pronouns, adjectives and verbs L N C 3 2 7 0 B D is pa tc h: 20 .1 .1 1 N o. of pa ge s: 21 C E :G ay at hr i Jo u r n a l N a m e M a n u s c r ip t N o . To c he ad : Y P PE :S ha ra ny a Language and Linguistics Compass (2011): 1–21, 10.1111/j.1749-818x.2011.00270.x a 2011 The Author Language and Linguistics Compass a 2011 Blackwell Publishing Ltd 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 (as subject or object agreement). Here is an example from Swahili where the subject toto agrees with the demonstrative le as well as the verb ona. Additionally, the object chungwa agrees with the demonstrative ha and is also cross-referenced on the verb by ya(Krifka 1995: 1399; the unmarked order is SVO, the subject appears immediately preverbally; noun classes are indicated by Arabic numerals): (1) Wa-toto wa-le wa-na-ya-ona ma-chungwa ha-ya. 2-child 2-that 2-PRS-6-see 6-orange this-6 ‘That child sees this orange.’ (Swahili) We are now ready to approach LI.