Joanna Wall
Utrecht University & Meertens Institute

Perfect doubling in West Germanic: Not the sandwich it first seems

  1. Introduction. Perfect doubling constructions consist of either two forms of the verb have combined with the participle of a transitive or unergative verb, like (1), or two forms of the verb be with the participle of an unaccusative verb. Whilst this paper will focus solely on have perfect doubling constructions, its conclusions should be equally applicable to the be variants.

(1)                   Ik         heb                  vandaag          nog                  niet      gerookt                                               gehad.

I               have    today              still                  not       smoked.PTCP    had.PTCP

‘I have not yet smoked today.’                                                              (South-eastern Dutch; Koeneman et al. 2011: 37)

Have perfect doubling is found in modern German (e.g. Rödel 2011; Zybatow & Weskott 2018) and Dutch dialects (Koeneman et al. 2011), but not attested in modern English or Dutch. These constructions have proven an analytical puzzle, none the least due to their semantic proximity to present perfects. In this talk, I present evidence for perfect doubling in the previously understudied historical varieties of Dutch (HVDs), and propose an analysis linking the possibility of perfect doubling to the varying properties of present perfects, modals and passives in Dutch, English and German.

  1. Corpus study of HVDs. 512 instances of have-doubling were found in a large-scale corpus of approx. 83,000,000 word corpus covering 1050 to 1649. Firstly, based on syntactic characteristics, I show that these instances of have-doubling robustly include instances of perfect doubling, as opposed to other have-doubling variants found in, for instance, modern English. Secondly, based on an analysis of geographical distribution, I show that perfect doubling is crucially attested in Hollandic HVDs which formed the basis for modern Standard Dutch but where the construction is no longer attested.
  2. Analysis. Adopting a Minimalist framework, my departure point is Brandner and Larsson’s (2014) proposal that perfect doubling constructions are a combination of two semantically distinct present perfects. According to the standard typological classifications, I assume that one of these present perfects functions as a true perfect requiring current relevance whilst the other functions as a temporal past, lacking current relevance. I argue that the true perfect is found in English, German, Dutch dialects and modern Standard Dutch, whilst the temporal past is found in all varieties but English. Distancing the present analysis from Brandner et al.’s, I then propose a formal analysis based on Wurmbrand’s (2001) restructuring account. Whilst primarily focusing on German infinitives, Wurmbrand does propose that present perfect have can merge in two distinct functional projections (ModP, AuxP) without any semantic distinction. In contrast, I argue that have merging in the lower projection (ModP) results in a true perfect, whilst have merging in a higher position, which I will argue to be TP rather than AuxP, results in a past. Supporting evidence for this proposal includes the differing interactions between modal verbs and present perfects in West Germanic varieties.

Whilst this analysis correctly rules out perfect doubling in English, it makes the seemingly problematic prediction that perfect doubling should be possible in modern Standard Dutch, a variety with both present perfects, i.e. the structural means. However, I make the novel claim that the lack of perfect doubling constructions in that variety is only apparent and results from a PF operation, blocking the spell out of the embedded participial form of have. This proposal is supported by parallel verbal constructions in modern Standard Dutch where a covert auxiliary has also been posited, like perfect passives (e.g. het boek is verkocht (*geworden); van Bart et al. 1998) which feature only one overt auxiliary. In sum, I argue that perfect doubling is not the sandwich it first seems: whilst on the surface Dutch seems to parallel with English, it ultimately sides with German and Dutch dialects, namely in its structural means.

  1. References. 1. Van Bart, P., J. Kerstens & A. Sturm. 1998. Grammatica van het Nederlands. Een inleiding. Amsterdam: Amsterdam University Press. 2. Brandner, E. & I. Larsson. 2014. Perfect doubling and the grammaticalization of auxiliaries. Abstract. DiGS16, 3–5 July 2014 Budapest. 3. Koeneman, O. N. C. J., M. Lekakou & S. Barbiers. 2011. “Perfect doubling.” Linguistic Variation 11(1): 35-75. 4. Rödel, M. 2011. “New perspectives on double perfect constructions in German.” Tense across languages, ed. by M. Rathert & R. Musan, 127-146. Tübingen: Niemeyer. 5. Wurmbrand, S. 2001. Infinitives: restructuring and clause structure. Berlin/New York: Mouton de Gruyter. 6. Zybatow, T. & T. Weskott. 2018. “Das Doppelperfekt: Theorie und Empirie.” Zeitschrift für Sprachwissenschaft 37(1): 83–124.