The choice of the perfect auxiliary in contemporary spoken Danish

Abstract

This article examines the distribution and the usage pattern of the two perfect auxiliaries – have and be – in contemporary spoken Danish in order to investigate whether the choice between the two auxiliaries is undergoing systematic changes. To do so the LANCHART corpus of contemporary spoken Danish is used. First, the macro-level distributional pattern is established through quantitative apparent and real-time analysis. In the subsequent qualitative micro-level analysis, particular attention is paid to cases where the selection of the auxiliary deviates from what could be expected based on Sorace’s Auxiliary Selection Hierarchy and grammars of Danish. The investigation focuses on the verbs gå ‘go’ and komme ‘come’ – the two most frequent verbs of motion in the corpus. The apparent time analysis of gå indicates an increased use of have across generations. In real time, no change is registered, indicating that any change is stemming from generational differences. With komme – although unexpected examples with have are attested – no statistically significant change is registered neither in apparent nor in real time. The micro-level analysis shows that the variation in auxiliary selection largely conforms to expectations; however, deviations are registered and discussed.

Keywords:

• Perfect auxiliary selection
• spoken language variation
• Danish
• corpus study

Acknowledgements

I would like to thank the participants of the symposium on perfect auxiliaries at Bernstorff Slot in June 2016 for the 0discussion and useful comments on my presentation. I am also grateful to Elisabet Engdahl, Frans Gregersen and the two anonymous referees for the valuable comments, which have improved this article. All remaining errors are, of course, mine.

Notes

¹ Throughout the article, the two main auxiliaries are referred to with have and be in small capitals, independently of the specific language. The corresponding Danish auxiliaries are have (present tense har, past tense havde) and være (present tense er, past tense var).

² In contemporary Scandinavian languages, a choice of perfect auxiliary is present in Danish and Faroese (Hansen and Heltoft 2011; Heycock and Petersen 2017; Petersen 2013), and to some degree in Norwegian (Faarlund, Lie, and Vannebo 1997; Larsson 2014, 2015; Lie 1972, 1989), while it has been lost in Icelandic and Swedish (Johannisson 1945; Larsson 2009).

³ Throughout the article, the supine form of the verb gå ‘go’, i.e., gået, is glossed as ‘go.sup’, regardless of the intended meaning. The actual meaning of the verb is given in the paraphrase. Likewise, the supine of the verb komme ‘come’, i.e., kommet, is glossed as ‘come.sup’.

⁴ Hansen and Heltoft (2011, 678) provide the example: så havde de dumpet ‘then they had failed the exam’. dumpe is literally ‘fall, drop, plump’.

⁵ Note that Mikkelsen, some 100 years before Hansen and Heltoft (2011), and Götzsche (1997), claimed that the “expansion” was going in the other direction, so that be would often, especially in everyday spoken language, be used instead of have (Mikkelsen [1911] 1975, 423). As our spoken language data, for obvious reasons, does not go back that far, it is hard to test these seemingly contradictory claims but the question could be clarified by a diachronic study based on written language data.

⁶ The numbers are based on the whole interviews, i.e., contributions from both interviewers and interviewees are included.

⁷ The list includes all verbs which take the auxiliary be, and have or/and be according to the Danish Dictionary, DDO (I want to thank the editor of the Danish Dictionary Lars Trap-Jensen for supplying the list), as well as verbs mentioned in earlier research (Götzsche 1997; Hansen and Heltoft 2011; Jensen and Maegaard 2012; Johannisson 1945; Jul Nielsen 1998; Sandford Pedersen 1997; Thestrup 2000), and the prefixed forms of those verbs.

⁸ The irrelevant constructions, which were excluded, were typically passive or predicative constructions or transitive uses of some verbs.

⁹ The main reason for choosing the two most frequent verbs was to get as many relevant examples as possible. As spoken language corpora tend to be relatively small, it is often difficult to get sufficient numbers of grammatical constructions in certain linguistic environments (cf. Cheshire, Kerswill, and Williams 2005).

¹⁰ The chi-square test was used: G1 vs. G2: χ² = 1.7958, df = 1, p-value = 0.1802.

¹¹ The chi-square test was used: G1 vs. G3: χ² = 19.2148, df = 1, p-value = 1.168e-05; G2 vs. G3: χ² = 8.9442, df = 1, p-value = 0.002783.

¹² The chi-square test was used: G1-S1 vs. G1-S2: χ² = 2.6479, df = 1, p-value = 0.1037; G2-S1 vs. G2-S2: χ² = 0.0317, df = 1, p-value = 0.8586.

¹³ All examples in the following sections are from the corpus, if not otherwise stated. Boldface added by the present author.

¹⁴ Sandford Pedersen (1997, 81) also points to this significant difference in usage between what she calls the ‘prototypical meaning’ and the overall frequency of the motion verbs gå ‘go’ and løbe ‘run’.

¹⁵ sic!

¹⁶ Note, though, that the sentences may demonstrate a possible case of variation, as native speakers of Danish also tend to accept sentences (7a–b) with the auxiliary have.

¹⁷ The high frequency is due to the fact that the LANCHART corpus consists of sociolinguistic interviews, where the participants were explicitly asked to talk about their life stories, including the childhood and school years.

¹⁸ Examples in Pedersen (1997) are from Ømålsordbogen (the Dictionary of Danish Insular Dialects).

¹⁹ The difference between these two meanings is expressed prosodically. The social activity meaning has unit accentuation (signalling noun incorporation), while the atelic motional process has primary stress on the verb gå (Pedersen 1997, 233).

²⁰ The phrasal verb komme ind (with the meaning ‘be admitted (to school etc.)’) is as a rule used with be, cf. (i).