Taal en Tongval - Volume 67, Issue 2, 2015

Variation and change in constructions: At the intersection of construction grammar and variational linguistics

This introductory paper outlines the different trends and movements that have in recent years led to a marked increase in the number of linguistic studies that combine a construction-based theoretical outlook on grammar with a dedicated interest in issues of synchronic and/or diachronic language variation. In addition, it gives an overview of the papers included in this thematic issue and links them with broader tendencies in the fields of cognitive linguistics and construction grammar.

This articles investigates the rise of double modal constructions in Dutch. Double modal constructions combine two modal auxiliaries with one lexical verb (e.g. zal moeten gaan ‘shall have to go’). Little is known on why and how such complex verb constructions came into being. This article presents historical data for the earliest double modal constructions in Middle Dutch (13th century) drawing on both previous corpus studies and on new empirical observations. The historical data is then analyzed from a diachronic construction grammar perspective. The main theoretical point of this article is that the emergence of double modal constructions is a case of ‘constructional complexification’, i.e. the process in which constructions become increasingly larger in size.

This contribution focuses on grammaticalisation pathways of German resultative and modal constructions using the auxiliaries kriegen ‘to get, to receive’ and bekommen ‘to get, to receive’. In order to illustrate the complex grammaticalisation pathways of the two GET verbs and the subtypes of resultative-modal constructions in which they are used, first, historical and second, (more) contemporary data from varieties of German are analysed. Third, additional evidence is provided by means of already existing findings on Dutch and Afrikaans to complement the hypotheses derived from the analyses of German varieties.

The contribution provides answers to the following questions: Where do the resultative-modal constructions with kriegen and bekommen come from and what about their future potential? What are the main differences and parallels between the kriegen and the bekommen constructions which both use an auxiliary from the semantic network of German GET verbs? What do the synchronic data tell us about the present-day variation of the constructions, but also – on the basis of an “apparent time” hypothesis – about their diachronic grammaticalisation pathways? How do “real time” data correspond to synchronic variation and vice versa? Which parallels or differences can a language comparison reveal? To what extent do the kriegen and bekommen results mirror what we find with regard to Dutch krijgen (‘to get, to receive’) and Afrikaans kry (‘to get, to receive’)?

The aan-construction in 17th Century Dutch: A semasiological investigation

The Dutch aan-construction (e.g. Hij gaf een bos bloemen aan zijn vrouw ‘He gave a bouquet to his wife’), the prepositional alternative for the double object construction (Hij gaf zijn vrouw een bos bloemen ‘He gave his wife a bouquet’), is a post Middle Dutch innovation (i.e. after 1500 AD). The precise details of the rise of the aan-construction remain, however, understudied. It is for example unclear whether the construction really breaks through in the 17th century, as Weijnen and Gordijn (1970) argue on the basis of a small corpus of farces, and what its semantic range was in those early days. In this paper we try to shed more light on these issues. On the basis of a self-compiled corpus of literary Dutch, we firstly show that the construction was not only already frequently attested in the language use in the 17th century, but also covered a remarkably wide semantic range at that time. Next, via a detailed comparison with data for the 20th century, we show that there have been interesting changes concerning the semantic evolution of the aan-construction. The structural weight of a cluster of ‘do’- and ‘send’-verbs for example declines over time and at more general level there seems to be a trend towards more abstract uses of the aan-construction. A diachronic collostructional analysis (Hilpert, 2006) and Configural Frequency Analysis (von Eye, 2002) lends a statistical underpinning to our observations.

This paper analyses English multi-word insertions found in a Dutch corpus of naturally occurring spoken conversation. Specifically, the corpus is based on manual morphophonological transcriptions of three seasons of the reality TV show Expeditie Robinson (known as ‘Survivor’ in the English-speaking world) and contains 10,000 utterances produced by 52 Belgian Dutch and Netherlandic Dutch speakers. Our analysis zooms in on all English insertions found in the data that contain more than one word, such as alive and kicking and you’re almost there (300 instances for 187 types). Three different diagnostic tests are presented that measure the degree of fixedness and conventionality of these multi-word insertions, namely lexicographical treatment, raw frequency and paradigmatic modifiability. Results reveal that the English multi-word inclusions are typically highly conventional fixed expressions, copied as a whole from English and inserted in Dutch much like traditional loanwords. Constructions containing open slots (e.g. NOUN of the day) and prototypical codeswitches (e.g. someone’s got to do all the work) are rare. These findings will be linked to the nature of the English-Dutch contact setting. Overall, our study contributes to research on contact-induced variation and change by paying attention to the use of foreign (semi-)fixed phrases and constructions, which form an interesting but often neglected grey-zone on the continuum from borrowing to codeswitching. Additionally, it opens up the restricted focus on monolingual contexts of early research on phraseology and Construction Grammar.

In the late sixties, Weinreich, Labov & Herzog (1968) discussed five empirical problems that together defined the subject matter of sociolinguistics. The research program they initiated has been very influential in sociolinguistics, but is not well known outside of it. In the branches of linguistics primarily focused on describing and accounting for the linguistic knowledge possessed by the individual speaker, branches often referred to as ‘general linguistics’ or ‘theoretical linguistics’, issues of variation and change have not attracted much theoretical attention up until recently. This has changed now that the tradition of ‘usage-based linguistics’ has appeared on the scene. This article examines the commonalities between the concerns of usage-based linguistics and those of Weinreich, Labov & Herzog (1968). It concludes that combining the cognitive orientation of the usage-based approach and the social-interactional orientation of sociolinguistics holds promises for the development of an improved understanding of language change. This is illustrated with examples from language contact, more specifically using data from Dutch-Turkish contact, including the results of recent studies that have implemented some methodological innovations necessitated by the confrontation of usage-based with sociolinguistic frameworks.

What to wear today: een hemd met blazer (‘a dress shirt with suit jacket’) or een shirt met jasje (‘a dress shirt with suit jacket’)? Convergence and divergence in Dutch clothing terminology

This paper reports on a corpus-based investigation into naming preferences in Belgian Dutch and Netherlandic Dutch for fourteen clothing terms. The study is a follow-up of Geeraerts, Grondelaers and Speelman (1999), in which soccer and clothing terminology from 1950, 1970 and 1990 was analysed as an indicator of standardisation in Dutch. This study extends the clothing corpus with new, comparable data from 2012 collected from magazines and shop windows. A profile-based measure of linguistic uniformity quantifies the differences in naming preferences across the 14 concepts between different varieties of Dutch. The results shed new light on the current linguistic situation in the Low Countries. The diachronic convergence between Belgian Dutch and Netherlandic Dutch found in Geeraerts, Grondelaers and Speelman (1999) seems to have come to a halt in present-day Dutch. On the other hand, the recent data confirm that the distance between the language in the lower register shop windows and the standard language in magazines remains largest in Belgian Dutch.

This study applies the methodology described by Gries & Deshors (2014) within the framework of the Contrastive Interlanguage Analysis (Granger, 1996) to the partitive genitive inflection in post-quantifier adjectives in the Moroccan Dutch ethnolect. This implies fitting a logistic regression model on data from the complementary ConDiv and Moroccorp corpora to investigate the differences between the L1 variety and the (early L2/2L1) ethnolect variety. It was found that the Moroccan Dutch language users do not differ from ‘ordinary’ Dutch language users in the realisation of the partitive genitive -s suffix, neither through an outspoken preference for one of the inflectional variants, nor in the factors determining the alternation. This is considered a rather surprising result, as such differences do exist for a number of other grammatical phenomena (Cornips and Rooij, 2003; Van de Velde and Weerman, 2014). This finding can tell us something about the inflectional status of the partitive genitive. It appears that it is less non-transparent than other quirks in adjectival inflection.