Recent decades of research in linguistics have seen a shift towards empirical methods and an increased use of data from corpora as a basis for making claims about language (Sampson 2005). This trend has made its mark on research on the Nordic languages also, and the current special issue aims to show some of the breadth of research in this field. The issue is in its entirety devoted to contributions that use the methodology of corpus linguistics on Nordic language data. This includes research that investigates both historical and contemporary aspects of the languages of the Nordic region.

The development of grammatical markers has been described from several theoretical perspectives over the last decade: Grammaticalization Theory (Hopper & Traugott 2003, Heine, Claudi & Hünnemeyer (1991), the Minimalist Program (Roberts & Roussou 2003, van Gelderen 2004), and Lexical-Functional Grammar (Vincent 2001), see also the overview in (Börjars & Vincent 2010). It has recently been addressed in Construction Grammar, where it is argued that a shift towards a constructional perspective on change may yield new insights into the workings of grammaticalization (Bergs & Diewald 2008, Hoffmann & Trousdale 2013, Traugott & Trousdale 2013). This paper should be taken as a contribution to a constructional view on grammaticalization. It is about the rise of the concessive subordinator fast(än) in the history of Swedish occurring in a construction or clause type called universal concessive conditional (Haspelmath & Köning 1998), in Swedish generaliserande bisats (SAG 1999). The Swedish fast, etymologically (and still productively) as an adjective in the meaning ‘steady’, ‘robust’ is used as an intensifier, ‘very’, ‘much’, in early Modern Swedish, eventually established as a concessive marker ‘even if’, ‘although’ in the 18th century. The conventionalization of a concessive inference is highly interesting and may be traced back to specific constructions in the 16th and 17th centuries. On the basis of an extensive corpus study, I analyze the critical contexts and discuss the development as constructional change rather than lexical change, arguing that a remapping between form and function takes place in concessive conditional constructions due to processes of inferencing and mismatch.

Keyword analysis has been used to investigate properties of style and genre, as a tool in discourse analysis, and as a method of identifying differences between the speech of distinct social groups. It has often been criticised as a blunt tool which can exaggerate what differences are present and fails to distinguish between quite distinct phenomena. However, it remains a very powerful tool for wide analysis of systematic differences between corpora when used with sufficient scepticism. This paper uses keyword analysis to examine differences between the speech of male and female characters in the Íslendingasögur, narrative prose texts composed in Iceland in the 13th and 14th centuries. This dataset is of particular interest because such representations of speech are the only window on the language of social groups who were not involved in text production in medieval societies. It aims to demonstrate a rigorous application of keyword analysis, exemplifying what it can and, crucially, what it cannot show.

This article presents a corpus-based investigation of the motivations behind the use of the s-passive and the bli-passive in contemporary written Swedish. Following a probabilistic approach to language variation and building on observations in the literature, we examine the simultaneous effects of semantic and syntactic factors by means of a multivariate statistical analysis. Our corpus sample includes 1,197 passive sentences of three Swedish verbs, which alternate in their use of the passive (acceptera ‘accept’, behandla ‘treat’ and välja ‘choose’). The results suggest that the choice of passive form is significantly associated with five predictor variables: Subject Animacy, Subject Number, Modal Verb, Av-phrase (‘by’-phrase) and Aktionsart. Among these, Subject Animacy and Modal Verb appear to yield the strongest impact effect on the choice of passive form. The study adds to earlier research in that it allows for a more accurate analysis of the simultaneous effect and relative strength of each factor on the speaker's choice of one of the passive forms.

Previous studies have shown the productive nature of eye and how it enters into patterns of a more or less non-compositional nature (e.g. Sinclair 1991a, Więcławska 2012). This paper adds a contrastive dimension and explores the cross-linguistic phraseology of the English–Norwegian cognates eye and øye on the basis of monolingual, bilingual and multilingual corpora. Starting with a survey of uses in the bidirectional English–Norwegian Parallel Corpus+ (ENPC+), the contrastive analysis reveals that while the two languages overlap in many of their uses of eye/øye-expressions, differences also emerge, particularly with regard to the number of recurrent patterns recorded and their conditions of use. English has more recurrent patterns with eye, but Norwegian has by far the most frequent pattern, få øye på ‘catch sight of’ (lit.: get eye on). Following this general cross-linguistic survey, a focused contrastive case study of få øye på and its English correspondences shows how a combination of bilingual and monolingual corpora may complement each other in contrastive research. The study uncovers that English has three main correspondences – catch sight of, see and spot – of which the first is the one favoured by bilingual dictionaries. An in-depth analysis of få øye på and catch sight of and their extended context, i.e. when they are part of extended units of meaning (e.g. Sinclair 1996), suggests that although the two patterns are perfectly matched, there are substantial differences when it comes to their frequency of use. This contributes to the relatively low mutual correspondences in the bidirectional translation material at hand.

This study analyses the relationship between four near-synonymous Swedish prepositions, namely angående, beträffande, gällande and rörande, which are used to establish what we call a topic-marking relation, as in statens avtal angående finansieringen ‘the agreement of the state regarding the financing’. By focusing on a single, loosely defined genre consisting of the written texts included in the Swedish PAROLE corpus, we address the question of what differences there are among these four prepositions, which intuitively seem highly similar and mutually interchangeable. In order to find out which contextual and grammatical factors might influence the choice of one preposition over the others, two complementary analyses were performed. First, a so-called collostructional analysis (see Stefanowitsch & Gries 2003, Gries & Stefanowitsch 2004) was performed on 791 cases of these prepositions found in the PAROLE corpus. Secondly, the corpus examples were annotated according to ten syntactic and four semantic criteria and a multinomial logistic regression analysis was performed on the annotated data set. The results show some tendencies pointing to differing usage patterns of the four prepositions. Beträffande stands out as the most frequent of them all and is also preferably used when no explicit head element is present, typically in sentence-initial position. Angående prefers words of communication while rörande is used when another topic-marking preposition is also present. On the other hand, neither of the two analyses leads to a clear distinction among the four prepositions, thus pointing to the fact that these topic-marking prepositions indeed constitute a fairly good case of adpositional synonymy, with few distinguishing factors separating one from the other.

This study investigates the linguistic characteristics of Swedish clinical text in radiology reports and doctor's daily notes from electronic health records (EHRs) in comparison to general Swedish and biomedical journal text. We quantify linguistic features through a comparative register analysis to determine how the free text of EHRs differ from general and biomedical Swedish text in terms of lexical complexity, word and sentence composition, and common sentence structures. The linguistic features are extracted using state-of-the-art computational tools: a tokenizer, a part-of-speech tagger, and scripts for statistical analysis. Results show that technical terms and abbreviations are more frequent in clinical text, and lexical variance is low. Moreover, clinical text frequently omit subjects, verbs, and function words resulting in shorter sentences. Clinical text not only differs from general Swedish, but also internally, across its sub-domains, e.g. sentences lacking verbs are significantly more frequent in radiology reports. These results provide a foundation for future development of automatic methods for EHR simplification or clarification.