Author:

Abstract:

Travelling features: Multiple sources, multiple destinations The formal features that are traditionally taken to define grammatical categories can be exchanged between categories. This implies that grammatical categories at best represent clusters of recurrently, but neither exclusively nor consistently, co-occurring features (in line with Croft 2001; Bickel 2010; De Smet 2010). The reason behind the promiscuous behaviour of formal features is that diachronically, the categories characterized by these features often derive from multiple source constructions (see Van de Velde et al. 2013). In this paper, we present two case studies that demonstrate feature exchange between historically distinct, but related categories. The first case study is on Dutch auxiliaries. Like in English, Dutch auxiliaries fall into two categories, depending on whether they take the ‘long’ infinitive (with te, cognate of English to) (1) or the bare infinitive (2). (1) hij lijkt *(te) werken (‘he seems to work’) (2) hij mag (*te) werken (‘he may work’) However, both categories contaminated each other. These contamination effects are strongest in constructions that distinguish both categories of auxiliaries from ordinary verbs, such as the infinitivus-pro-participium (IPP) (Schmid 2005). Thus, the bare infinitive construction has extended to auxiliaries traditionally patterning with the te-infinitive, starting in the IPP construction and later extending to new contexts, more specifically those contexts that superficially resemble IPP. The process can be shown to display lexical diffusion effects. The extension is illustrated in (3)-(4) with durative-aspectual auxiliary zitten (Haeseryn et al. 1997:970). (3) Hij zit de hele dag *(te) werken (‘he is working all day’) [no IPP] (4) Hij heeft de hele dag zitten (te) werken (‘he has been working all day’) [IPP] (5) omdat ze de hele dag zitten (te) werken (‘they have been working all day’) [no IPP, but superficially similar] The second case study deals with English quantifiers. As a lot grammaticalized and entered the category of quantifiers (Brems 2003; Traugott 2008), the behaviour of core members like some, many or a few also changed. Specifically, as a quantifier, a lot retained the possibility of combining with the degree modifier quite – a feature carried over from its historical origins as a noun phrase. Subsequently, quite also spread to a few and (marginally) some and many. In other words, a distributional feature travelled from nouns to quantifiers. This is illustrated in (6)-(8). (6) The fugitive had run quite a distance (1862, COHA) (7) We saw quite a lot of folks from East Wellmouth (1869, COHA) (8) so I figures he's been there quite some time. (1914) Bickel, B. 2010. ‘Grammatical relations typology’. In: J.J. Song (ed.), The Oxford handbook of language typology. Oxford: OUP. 399-444. Brems, L. 2003. ‘Measure Noun constructions: an instance of semantically-driven grammaticalization’. International Journal of Corpus Linguistics 8: 283-312. Croft, W. 2001. Radical construction grammar: syntactic theory in typological perspective. Oxford: Oxford University Press. De Smet, H. 2010. ‘English -ing-clauses and their problems: the structure of grammatical categories’. Linguistics 48: 1153-1193. Haeseryn, W. et al. 1997. Algemene Nederlandse spraakkunst. 2nd edn. Groningen: Martinus Nijhoff. Schmid, T. 2005. Infinitival syntax: infinitivus pro participio as a repair strategy. Amsterdam: Benjamins. Traugott, E.C. 2008. ‘The grammaticalization of NP of NP patterns’. In Bergs, A., & Diewald, G. (eds.). Constructions and language change. Berlin: Mouton de Gruyter. 23-45. Van de Velde, F., De Smet, H., & Ghesquière, L. 2013. ‘On multiple source constructions in language change’. Studies in Language 37: 473-489.