The use of case forms in Modern Danish – an empirical study

ABSTRACT

This paper investigates the distribution of the nominative (nom) and the oblique (obl) form in personal pronouns in Modern Danish. As per a primary rule, pronoun NPs functioning as subjects are typically in the nom form, whilst the obl form is used in all other syntactic functions. Nonetheless, this rule is subject to certain exceptions in a number of constructions. In this paper, we present four studies concerning the distribution of nom and obl in selected constructions, where instances of variation are observed. These investigations are based on data from two separate corpora, one representing written language and the other, spoken language. Our findings suggest that both stylistic and structural characteristics, alongside considerations concerning processing, significantly influence the distribution of nom and obl. Furthermore, we note a considerable disparity in distribution when comparing spoken and written data.

KEYWORDS:

• Modern Standard Danish
• nominative and oblique forms
• personal pronouns
• case forms
• written and spoken language

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Jørgensen (2000) consulted the corpus Korpus DK87–90, but did not conduct frequency counts.

² Or case-impoverished (Sigurðsson and Heltoft 2021).

³ This set of pronouns with only 7 lemmas and only two case forms dates back to Early Middle Danish (Jensen 2011, 2018a).

⁴ Unless otherwise stated, examples are constructed by the authors.

⁵ In addition to the proposals presented in Section 3, Erik Hansen ([1972] 2001) observes that the distribution of the two case forms correlates with stress patterns: when a pronoun, owing to its syntactic function, is obligatorily stressed, it takes obl; when it can be unstressed, it takes nom. This observation, the ‘stress proposal’, was very influential in Danish grammar studies, especially in the 1970s. In all the constructions chosen for empirical examination in this paper, the pronoun is obligatorily stressed.

⁶ As also noted by Heltoft, the idea of a correlation between nom and an anaphorical use stems from Aage Hansen (1965, 100).

⁷ A point also made by Erik Hansen ([1972] 2001).

⁸ As shown in (14) and (15), variation is also found in certain other constructions, such as vi/os andre (‘we/us other’), I/jer begge (‘you/you both’), hun/hende selv (‘she/herself’) and more (see also Diderichsen 1946, 113; Hansen and Heltoft 2011, 443). Due to time and space limitations, these are left out of this paper.

⁹ For the spoken language corpus, this resulted in up to 240,000 examples. Across all datasets, irrelevant examples were deleted, e.g., examples with comparative som ‘as, like’ in studies 1 and 2 and, examples of conjunctions with full clauses in study 3 (e.g., end at de gjorde det ‘than they did it’). From the spoken language datasets, we removed also all examples containing self-corrections and false starts.

¹⁰ The studies presented in 5.1 were first published in Schack et al. (2019) and Petersen et al. (2020). The study in section 5.2, however, is an original contribution and has not been previously published. As for the study presented in 5.3, while the results concerning the spoken data were published in Jensen (2018a, 2019a, 2019b), the findings concerning written data are derived entirely from fresh material.

¹¹ When subjects adopt the form according to the function in the subordinate clause, (21), this is designated as ‘subordinate clause attraction’ (“bisætningsattraktion”, Diderichsen 1941, 140; 1946, 278); when noun phrases within a relative clause adopt the form according to the function in the matrix clause, this is referred to as ‘main clause attraction’ (“hovedsætningsattraktion”).

¹² χ² = 166.94 (df 1, N = 878), p < 0.001. All statistic tests are performed in Excel, on the website Social Science Statistics, or by use of VassarStats.

¹³ I.e., a girlfriend.

¹⁴ χ² = 31.65 (df 1, N = 5,442), p < 0.001. The choice of preposition correlates statistically significantly with the distribution but in a yet unexplained manner, see Petersen et al. (2020).

¹⁵ This traditional rule is a remnant of structures where words and constituents with affinity to each other would agree. Before Early Modern Danish (1500–1700), also subject complements would occur in nom, agreeing with the subject. In Modern Danish, the subject complement follows the main rule, i.e., because it is not a subject it is always in obl, cf. (6) (Jensen 2017, 2018a).

¹⁶ There is evidence to suggest that the second comparative element is not, in fact, a subject within a reduced clause. In clauses with reflexive constructions, in Modern Standard Danish, the NP is constructed with the possessive pronoun sin (declined for person and gender) if the entity under discussion refers to the subject of the clause. Conversely, the genitive form of the personal pronoun is used if the entity in question does not refer to the subject of the sentence. In han er lige så klog som sin bror ‘he is just as clever as his brother’, the pronoun employed to establish the reflexive construction is the one referring to the subject han ‘he’. If it were the case that the comparative element was a reduced clause, the pronoun used to establish the reflexive construction would have been the genitive form of the personal pronoun hans ‘his’ (han er lige så klog som hans/*sin bror er ‘he is just as clever as his.gen/*his.poss brother is’) (Hansen and Heltoft 2011, 442–443; Schack 2021).

¹⁷ Although sympathetic to the analysis of these words as prepositions, in this paper, we follow the tradition and call them conjunctions.

¹⁸ Som ‘as, like’ vs end ’than’: χ² = 4.9 (df 1, N = 996), p < 0.027. The difference between end ‘than’ and ligesom ‘just as, just like’ is not statistically significant: χ² = 0.446 (df 1, N = 892), p = 0.495.

¹⁹ The difference between the three conjunctions is statistically significant: ligesom ‘just as, just like’ vs som ‘as, like’ χ² = 8.9 (df 1, N = 604), p < 0.003; som ‘as, like’ vs end ‘than’, χ² = 9.9 (df 1, N = 917), p < 0.002.

²⁰ The difference is not statistically significant: χ² = 0.27 (df 1, N = 247), p > 0.6.

²¹ 1p.sg vs 2p.sg: χ² (df 1, N = 97), = 2.9, p > 0.087; 2sg vs 3sg: statistic value is 0.66 (Fisher exact test, df 1, N = 40); 1p.sg vs 3p.sg: χ² = 2.3 (df 1, N = 115), p < 0.001.

²² 1p.sg vs 3p.sg: χ² = 2.4 (df 1, N = 263), p > 0.119; 3p.sg vs 2p.sg: χ² = 1.0 (df 1, N = 146), p > 0.308; 1sg vs 2p.sg: χ² = 4.4 (df 1, N = 197), p < 0.036.

²³ For end ‘than’: χ² = 14.007 (df 1, N = 473), p < 0.001. For som ‘as, like’: χ² = 8.687 (df 1, N = 444), p = 0.003. The conjunction ligesom ‘just as, just like’ eludes statistical testing due to its low frequency.

²⁴ Examples (37) and (38) are from Brink and Lund (1975, 666).

²⁵ Jensen (2018b, 2019a) used a different dataset for the count on written language. Therefore, the numbers are not identical to the count in the present study. The overall tendencies, however, are the same.

²⁶ I.e., anaphoric or cataphoric, section 3.3.

²⁷ One such study is now available. Johannsen (2023) has examined data from Facebook (representing written language) and from the reality shows Ex on the Beach and Big Brother (representing spoken data).

²⁸ Sigurðsson and van de Weijer (2021), however, does indeed give examples of obl in subject complements in some Swedish varieties.

Additional information

Funding

This work was supported by the Kulturministeriets Forskningspulje, Denmark [FPK.2020-0008].

Notes on contributors

Eva Skafte Jensen

Eva Skafte Jensen is a senior researcher at the Danish Language Council. She earned her PhD from the University of Copenhagen 2001 and her higher doctorate from Roskilde University 2011. Both theses are on grammatical change in Danish. Her research interests include the morphology and syntax of Middle and Modern Danish, diachronically and synchronically. She is currently working on morpho-syntactic change in Late Modern Danish (from the nineteenth century to the present day).

Jørgen Schack

Jørgen Schack is a senior researcher at The Danish Language Council. In his research, he primarily focuses on grammar, orthographic studies, and vocabulary studies. Currently, he is working with Eva Skafte Jensen on a project on 19th century morphology as a key to modern norm-related issues.

Jan Heegård Petersen

Jan Heegård Petersen is an associate professor of Danish studies at the University of Copenhagen. His main research interests include Danish as a heritage language, grammar of spoken Danish, and field linguistics. He has published on the grammar of spoken Danish since 2012.