Reassessing the evolution of West Germanic preterite inflection

Abstract

This article takes a quantitative approach to the long-term dynamics of the preterite inflection in West Germanic, with a special focus on Dutch. In a first step, we replicate two often-cited studies on English and German (Lieberman et al. 2007 and Carroll et al. 2012, respectively) by looking at Dutch. This part also tackles some methodological shortcomings in the previous studies. In a second step, we delve deeper into the evolution of the preterite morphology in Dutch in the last 1200 years, by looking at several factors which have been previously only investigated in isolation or on limited time slices. Using multiple binomial regression analysis, the various factors are studied under multifactorial control.

Semantic differences between strong and weak verb forms in Dutch

“…This variable was coded at the text level, which means that we distinguished between texts with rhyme and prose, but we did not look at whether the verb form actually occurs in rhyme position in each specific attestation. We also added a numeric variable for century, because we know that (at least) the weakening of the originally strong preterites increases through time (Carroll et al 2012;De Smet and Van de Velde 2019;Lieberman et al 2007). Because this is less clear for the past participles (which become more conservative through time, as an effect of the Präteritumschwund, see section Research questions), we again added an interaction with CENTURY and PRINCIPAL PART.…”

A corpus-based quantitative analysis of twelve centuries of preterite and past participle morphology in Dutch

“…We excluded some hyperfrequent verbs, which are known to almost categorically resist weakening, and often show some irregularity in their forms that suggest they are verbs sui generis , rather than members of an analogically connected model of strong verbs. These were all verbs of the two highest frequency bins in De Smet and Van de Velde (2019) in which no verbs had weakened in over a 1000 years’ time, though only if their frequencies had remained stable throughout time (also based on De Smet & Van de Velde [2019:147]). Following those criteria, attestations of zijn ‘be,’ doen ‘do,’ see ‘zien,’ staan ‘stand,’ gaan ‘go,’ worden ‘become,’ and komen ‘come’ were not included 8 .…”

“…3.Although Präteritumschwund is more progressed in German than in Dutch, German shows less asymmetry between the amount of weakening in preterites and participles, likely because German is overall morphologically more conservative and shows overall less weakening (De Smet & Van de Velde, 2019). …”

“…The strong-to-weak evolution has been studied the most on a range of Germanic languages (Bybee [1985:120], Cuskley, Pugliese, Castellano, Colaiori, Loreto, & Tria [2014], Lieberman, Michel, Jackson, Tang, & Nowak [2007] for English; Carroll, Svare, & Salmons [2012] and Nowak [2015] for German, Dutch, and Luxembourgish; Dammel, Nowak, & Schmuck [2010] for German, Dutch, English, and Swedish; Strik [2015] for Frisian and Swedish; Dammel & Nowak [2015] for Frisian; and De Smet & Van de Velde [2019], Pijpops, Beuls, & Van de Velde [2015], Van Santen [1997], and Vosters [2012] for Dutch, inter alia ). However, most studies suffer from one of four shortcomings: (i) The studies are confined to a specific period in time, for example, by only taking into account younger data or synchronic slices; and (ii) They focus on one or two specific factors at a time and study them in isolation from each other.…”

“…However, most studies suffer from one of four shortcomings: (i) The studies are confined to a specific period in time, for example, by only taking into account younger data or synchronic slices; and (ii) They focus on one or two specific factors at a time and study them in isolation from each other. A first step toward a multifactorial approach was taken by De Smet and Van de Velde (2019), who replicated the often-cited studies of Lieberman et al (2007) and Carroll et al (2012) for Dutch. However, (iii) these studies were carried out using reference grammars to define whether a verb was weak or strong in a given period, which is a reasonable but linguistically quite a crude measurement, given that verbs are often not consistently strong or weak, especially when they are in transition to entirely weakening.…”