Text
E-book A Featural Typology of Bantu Agreement
How can there be so many different languages in the world while the abilityto use language seems to be fundamentally the same for every human being?This is without doubt one of the core questions of linguistics, and it is thisBig Question that is hidden behind the smaller questions in this book. Tostudy the variation and the underlying linguistic system in a scientific way,we ideally want to have a laboratory in which we can just change one param-eter and observe the effect. For a natural phenomenon like language, this isof course impossible. However, the Bantu language family comes close to anatural language variation lab: this language family consists of an estimated555 languages (Hammarstro ?m 2019), which are spoken in the area betweenCameroon, Kenya, and South Africa. Larry M. Hyman once said ‘If you’ve seenone Bantu language, you’ve seen them all – except they are all different!’ andthis is precisely what makes them so fascinating and suitable for comparativeresearch.In this book I therefore investigate a subset of the Bantu languages with re-spect to how arguments in the clause are licensed, and how this is reflectedin agreement marking on the verb. Precisely because there is microvariationin this area, we can unravel which features are involved in agreement andlicensing and pinpoint the featural parameters that give rise to the variation.The empirical basis covers Bantu object marking, in monotransitives andditransitives, as well as subject marking and subject inversion constructions.To account for the patterns found, I propose a new analysis that involvesAgree and Case licensing, and importantly takes the influence of informa-tion structure into account (building onMorimoto 2000;Halpert 2015; andmany others). The proposed analysis captures parameters as variation in for-mal features, and as a result, the featural analysis presented in this book shines a light on what is needed in the grammar to cover the attested typologicalvariation while taking into account learnability. Specifically, the data andanalysis in this book argue against both the Strong Uniformity Hypothesis(Chomsky 2001;Miyagawa 2010,2017) and the Strong Modularity Hypoth-esis (Chomsky 2008;Berwick & Chomsky 2011;Fanselow 2006;Fanselow &Lenertová 2011;Horvath 2010). The former claims that all languages share thesame set of grammatical features, and the latter that narrow syntax cannot beinfluenced by information-structural factors. Considering the clear influenceof information structure on the syntax of (some) Bantu languages, and con-sidering the amount of variation in sensitivity to these (and other) features, IfollowBiberauer (2011,2017b,2018ab,2019) in arguing that a model is to bepreferred in which features are not just assumed to be present or absent, butpostulated on the basis of their effects on the output.What is at stake, then, is more than just the analysis of subject or objectmarking in Bantu languages. But we have to start somewhere, and agreement isa useful starting point for our further discoveries. The questions for the currentbook, then, are which precise features play a role in agreement phenomena,and how these features can explain crosslinguistic variation. In this chapter,I start by introducing agreement and the notion of Agree (Section1.2) andthen discuss the features of the goal of agreement, specifically the notion of adefective goal (Section1.3). Section1.4presents a new way of thinking aboutfeatures (Section1.4.1), as well as Case as a feature responsible for licensing(Section1.4.2) and information structure (Section1.4.3). I then introduce theBantu languages (Section1.5), and I finish the chapter by briefly touching uponthe challenges and scope of the book (Sections1.5.3and1.6)
Tidak tersedia versi lain