Part II Forms and functions
4 Beyond verb second – a matter of novel information-structural effects? Evidence from Norwegian, Swedish, German and Dutch
Introduction
It has been relatively clear over the last few years that multilingual settings support new linguistic variation and the emergence of new linguistic patterns. Contemporary urban vernaculars have emerged among adolescents in multilingual settings in large cities throughout Europe. In this chapter, we use the label contemporary urban vernaculars (see Chapter 2), as it is problematic to label youth vernaculars using a technical term such as ‘ethnolect’ or a lay term such as straattaal (‘street language’).1 These labels essentialize groups and their language use, obscure speakers' styling practices and suggest systematic linguistic differences that reflect a pre-existing social category such as youth or ethnicity (cf. Cornips and de Rooij Reference Cornips, de Rooij, Siemund, Gogolin, Davydova and Schulz2013; Jaspers Reference Jaspers2008: 85, 87; Chapter 3, this volume).
This chapter addresses one of the structural characteristics that has been pointed to in the descriptions of contemporary urban vernaculars across Europe, namely deviations from the syntactic verb second constraint (V2) in Germanic languages. Several studies have revealed that this feature – although widespread among learners of Germanic languages as a second language – cannot be seen solely as a second-language feature in the language of youths in contemporary multilingual urban settings. In several studies, both simultaneous and early successive bilingual child acquirers and speakers without second-language speaking backgrounds have been found to use this feature in their everyday language (Dirim and Auer Reference Dirim and Auer2004; Ganuza Reference Ganuza2008a, Reference Ganuza2008b, Reference Ganuza, Quist and Svendsen2010; Keim Reference Keim2007a, Reference Keim and Auer2007b; Opsahl Reference Opsahl2009a; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010; Wiese Reference Wiese2009).
We address the question of whether the multilingual settings that serve as a backdrop for (emergent) contemporary urban vernaculars provide contexts not only with loosened syntactic restrictions (that is, new linguistic variation) per se, but also provide a context where – as a consequence of softer grammatical restrictions (Sorace Reference Sorace, Cornips and Corrigan2005) – information-structural preferences may be realized in novel ways. We contend that the deviations from V2 found in contemporary urban vernaculars across Europe is not a matter of free syntactic variation or incomplete language acquisition, but rather forms an integral part of the grammar of the mentioned vernaculars.2 To this purpose we tie together the evidence of several empirical sources from different but typologically similar languages to one comprehensive study – a procedure which enables us to draw an overall picture of the syntactical variational limits and possibilities which urban vernaculars of different origins share.
We begin with a short presentation of the speech communities from which our data is drawn. Afterwards, for potential readers with limited knowledge of the V2 feature in Germanic languages, a short overview of the phenomenon is presented. We then present data from the respective vernaculars that illustrate new word order patterns within declarative sentences. In the analysis section we reveal our findings and analyses with respect to V2 deviations and their relation to novel realizations of information-structural preferences in Norwegian, Swedish, German and Dutch.
Contemporary urban vernaculars in Norway, Sweden, Germany and the Netherlands: data collection procedures
Norway
Before the UPUS-project (Utviklingsprosesser i urbane språkmiljø [Developmental Processes in Urban Linguistic Settings]) began in 2006, little research on contemporary urban vernaculars had been done in Norway apart from descriptions of loan words from immigrant languages (see Aasheim Reference Aasheim1995; Drange Reference Drange, Drange, Kotsinas and Stenstrøm2002). The sociodemographics of the time pointed, however, to the fact that one would be likely to find what was often termed ‘multiethnolects’ – such as those already described in neighbouring Scandinavian countries – in Oslo.3 The fact that new ways of speaking Norwegian had emerged in multiethnic neighbourhoods (at least as far as new lexical items were concerned) was already well established in the media discourse when the UPUS-project began, and the label ‘Kebab Norwegian’ was frequently used. This label has far less resonance in the youth groups visited, and the young speakers themselves tend not to label their speech styles (‘just the way we speak’) or to connect it to the name of their local community (Aarsæther Reference Aarsæther, Quist and Svendsen2010). The UPUS research group chose to refer to the new urban linguistic practices as ‘multiethnolectal speech styles’.
Data were collected in two city districts where the immigrant population is higher than the average: in the inner-city district of Gamle Oslo (Old Oslo), where immigrants comprised 36 per cent of the population, and in the suburban city district of Søndre Nordstrand (Southern Nordstrand), where immigrants comprised 48 per cent of the population. A majority of people with immigrant background in these two areas have their origins in Asia, Africa, South and Latin America and Turkey.4 Most of the young people in the study report that they have friends whose family backgrounds originate in more than one of these parts of the world.
The findings of the UPUS-project revealed that the Norwegian spoken among adolescents in multiethnic areas in Oslo has several characteristics regarding lexical, phonological, morpho-syntactic and pragmatic features, and forms part of complex stylistic repertoires in the adolescents' everyday identity work and linguistic practices (Aarsæther Reference Aarsæther, Quist and Svendsen2010; Brunstad, Røyneland and Opsahl Reference Brunstad, Røyneland, Opsahl and Terkourafi2010; Nistov and Opsahl Reference Nistov, Opsahl, Mæhlum and Åfarli2014; Opsahl Reference Opsahl2009a, Reference Opsahl2009b; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010; Svendsen and Røyneland Reference Svendsen and Røyneland2008; Chapters 10 and 12, this volume). Most importantly, the studies revealed that multiethnolectal speech styles are not restricted to speakers with migrant background. As is the case with Germany (see below), the linguistic repertoire of speakers in multiethnic areas in Oslo covers a variety of registers including in-group and out-group informal speech, bilingual registers containing code-mixing as well as stylized variants of (the notion of) the standard language, traditional urban vernaculars or new urban vernacular speech styles themselves.
The Norwegian examples drawn upon in this chapter are based on a sample of data from the UPUS-Oslo corpus. The corpus consists of video-recorded peer conversations and interviews, which took place at youth clubs in the adolescents' neighbourhoods. No adult researchers were present during the recording of the peer conversation. The data are made available through an Internet-based interface – where transcripts and audio and video files are linked together – developed by Tekstlaboratoriet (‘The Text Laboratory’) at the University of Oslo. For the following analysis, data from 22 of the 56 adolescents who participated in the study are included. The adolescents have various parental backgrounds, but they were all born and raised in Norway. A total of 194 cases of deviations from the so-called V2 constraint are obtained from the data. As is also the case in the Swedish data (see below), most of the examples of XSV (‘X’ = (non-subject) sentence-initial element, ‘S’ = Subject, ‘V’ = finite Verb) were found in the peer conversations.
Sweden
In Sweden, studies of linguistic practices among youths in multilingual settings were first carried out in the mid-1980s by the language researcher Ulla-Britt Kotsinas (e.g. Reference Kotsinas1994, Reference Kotsinas, Androutsopolous and Scholz1998). Kotsinas focused primarily on the language use of a group of youths in the multilingual suburb of Rinkeby in Stockholm. She began to refer to their way of speaking Swedish as rinkebysvenska (‘Rinkeby Swedish’), a term that has since been widely dispersed and is often used to refer to ways of speaking Swedish among youths in any multilingual setting in Sweden, and/or sometimes even to refer to Swedish spoken by anyone with an immigrant background (e.g. Fraurud and Bijvoet Reference Fraurud, Bijvoet, Hyltenstam and Lindberg2004). In more recent research, the label Rinkeby Swedish is often avoided, as it carries negative connotations. Rather, more general labels are used, such as ‘multiethnic youth language’ or ‘suburban slang’ (e.g. Bijvoet and Fraurud Reference Bijvoet and Fraurud2006), although these labels may be equally problematic for the reasons mentioned in the introduction (see above).
Kotsinas (Reference Kotsinas1994, Reference Kotsinas, Androutsopolous and Scholz1998) described certain linguistic features that she found typical of the Swedish spoken by some of the youths in Rinkeby. Among other things she mentioned the youths' frequent ‘replacement of the inverted word order by SV order in sentences with a short temporal or locative adverbial’ (Kotsinas Reference Kotsinas, Androutsopolous and Scholz1998: 137). The intention of the extensive research project Language and Language Use Among Adolescents in Multilingual Urban Settings5 (the SUF project), which began in 2002, was to describe, analyse and compare ways of speaking Swedish among adolescents in several multilingual areas in Stockholm, Gothenburg and Malmö, and to provide multidimensional perspectives on the language practices of youths in contemporary multilingual urban settings in Sweden (e.g. Boyd Reference Boyd, Quist and Svendsen2010; Chapters 5 and 14, this volume). In total, 222 adolescents from eight upper secondary schools participated in the SUF project, and data was gathered from a number of settings (interviews, self-recordings, group conversations, movie retellings, oral presentations and written essays, as well as elicited oral and written data). The participants represented a wide range of linguistic backgrounds, and included youths who grew up in monolingual Swedish family settings as well as youths from bilingual families. Some of the bilingual participants began to learn Swedish from birth or an early age, whereas others began as late as eight to ten years of age.
The SUF project data has been analysed in relation to several grammatical aspects (e.g. Ekberg Reference Ekberg, Quist and Svendsen2010; Ganuza Reference Ganuza2008a, Reference Ganuza2008b, Reference Ganuza, Quist and Svendsen2010; Tingsell Reference Tingsell2007), phonology (e.g. Bodén Reference Bodén, Quist and Svendsen2010), lexicon (Prentice Reference Prentice2010) and discourse (Svensson Reference Svensson2009). There are also studies within the SUF project that have primarily focused on youths' identity work (Almér Reference Almér2011; Werndin, forthcoming). Similar to the German and Norwegian findings reported in this section, the SUF-project data includes a wide variety of ways of speaking Swedish. The studies based on the SUF corpus display large variation; there are no simple correlations to be found between the use of certain linguistic features and participants' linguistic backgrounds. Nevertheless, the use of XSV has been shown to be more frequent in the language of adolescents who have grown up in a multilingual setting and/or who report that they have many friends with a multilingual background. The use of XSV is, however, not necessarily related to the speakers' status as second-language speakers of Swedish (e.g. Ganuza Reference Ganuza2008a, Reference Ganuza2008b, Reference Ganuza, Quist and Svendsen2010). Examples of XSV were found in the everyday language of both monolingual Swedish-speaking youths and of bilingual youths who began to learn Swedish at an early age.
The syntactic analyses and examples drawn upon in this chapter are based on a sub-sample of data from the SUF corpus. It includes data from 34 participants, from peer conversations, a retelling task and written essays. The participants produced a total of 218 sentences that deviate from V2.6 Most of the examples of XSV were found in recordings of peer conversations, either in self-recordings or in recorded group conversations in the schools between peers from the same class. Very few examples of XSV were found in the written essays (Ganuza Reference Ganuza2008a, Reference Ganuza2008b, Reference Ganuza, Quist and Svendsen2010). Only six of the participants in the sub-sample produced no examples of XSV in the contexts discussed here.
Germany
In Germany, new linguistic practices have emerged among adolescents in multiethnic urban areas of larger cities (such as Berlin, Mannheim, Frankfurt/Main, Hamburg) during the last few decades. These practices can be subsumed under the category Kiezdeutsch (literally ‘hood German’, Wiese Reference Wiese2009, Reference Wiese2012). This label makes use of the northern German word Kiez (‘neighbourhood’), which is informally used to refer to individual quarters of towns and cities, particularly in Berlin. The locals in the respective neighbourhoods are usually very proud of their Kiez; the term carries a very positive connotation. Speakers of Kiezdeutsch are predominantly youths whose parents or grandparents have immigrated to Germany. These speakers have usually grown up in a bilingual environment. Apart from the majority language, German, the languages involved range from Turkish, Kurdish, Persian and Arabic to Bosnian, Croatian and Polish, among others. A fundamental aspect of Kiezdeutsch is that monolingual speakers of German also refer to themselves as speakers of Kiezdeutsch. Most importantly, specific linguistic characteristics that can be found in this urban vernacular are not tied to specific linguistic backgrounds (Freywald, Mayr, Özçelik and Wiese Reference Freywald, Mayr, Özçelik, Wiese, Kern and Selting2011; Wiese Reference Wiese2009). The linguistic repertoire of Kiezdeutsch speakers includes not only the majority language (plus a particular family language if bilingual), but also a variety of registers within these languages, including in-group and out-group informal speech, bilingual registers containing code-switching or code-mixing, formal, standard-like variants of German, and even stylized variants of, for example, the local dialect or of Kiezdeutsch itself (cf. e.g. Freywald, Mayr, Schalowski and Wiese Reference Freywald, Mayr, Schalowski and Wiese2010; Keim Reference Keim2007a, Reference Keim and Auer2007b).
Various studies on Kiezdeutsch – or Kanaksprache (‘wog language’) (Deppermann Reference Deppermann and Auer2007; Füglein Reference Füglein2000; Zaimoğlu Reference Zaimoğlu1995), Ghettodeutsch (‘ghetto German’) (Keim Reference Keim2004) or Türkendeutsch (‘Turks' German’) (Androutsopoulos Reference Androutsopoulos2001; Kern and Selting Reference Kern and Selting2006), as it is also called in the literature (terms that however indicate a certain amount of bias and are often used in a derogatory manner in everyday language) – describe convergent linguistic features of this way of speaking with regard to the lexical level as well as to the grammatical domains phonology, inflectional morphology, syntax and semantics (see e.g. Auer Reference Auer and Buhofer2003; Dirim and Auer Reference Dirim and Auer2004; Füglein Reference Füglein2000; Jannedy, Weirich and Brunner Reference Jannedy, Weirich and Brunner2011; Selting and Kern Reference Selting and Kern2009; Wiese Reference Wiese2006, Reference Wiese2009, Reference Wiese2011a, Reference Wiese, Gregersen, Parrott and Quist2011b, Reference Wiese2012).
The following analyses rest upon and result from investigations conducted within the Kiezdeutsch project7 at the University of Potsdam, where the Kiezdeutsch-Korpus (KiDKo) is currently being built up (see Wiese, Freywald, Schalowski and Mayr Reference Wiese, Freywald, Schalowski and Mayr2012). The corpus compiles data of spontaneous speech in multiethnic neighbourhoods, based on self-recordings of adolescents from Berlin-Kreuzberg (17 anchor speakers, aged 14 to 17, approximately 228,000 tokens). It is complemented by a corresponding sample of self-recordings made by adolescents from a monoethnic neighbourhood (Berlin-Hellersdorf) with comparable socioeconomic indicators (six anchor speakers, aged 15 to 16; approximately 105,000 tokens).8 All anchor speakers recorded informal conversations with their peers during leisure time with no researchers or other adults present. The recordings were conducted in 2008. Transcription conventions follow for the most part the transcription system GAT 2; capitalizations in the examples below mark accents (see Selting, Auer and Barth-Weingarten Reference Selting, Auer and Barth-Weingarten2009 for details, including a list of transcription symbols). A manual search in KiDKo revealed 55 instances of XSV order.9
The Netherlands
Urban vernacular speech has been examined in the larger cities in the culturally and politically dominant western part of the Netherlands. The studies undertaken are small compared to those done in Germany, Norway and Sweden. For Dutch, the data from three case studies were examined.
(1) Eight mutual friends in the Utrecht neighbourhood Lombok and Transvaal, all in their early twenties, were interviewed in a bilateral setting between the interviewer (the researcher) and the adolescent. Only in one setting did two friends interview each other. The friends have different language backgrounds: four speak Moroccan Arabic/Berber/Dutch, three speak Turkish/Dutch and one Surinamese (Hindi)/Dutch. The interviews were all recorded between 1998 and 2001 (approximately 8 hours) in the local youth centre (Cornips Reference Cornips, Bennis, Extra, Muysken and Nortier2002).
(2) Four male adolescents of Surinamese, Creole descent between the ages of 15 and 18 were recorded in 2002 in the Rotterdam neighbourhood Feijenoord. The fieldworker made contact with these youngsters through friends. The young people allowed the fieldworker to interview them and to hang out with them and their group of friends for four months. From the hours spent with these youngsters, the fieldworker recorded approximately two hours out ‘in the street’ (Cornips and de Rooij Reference Cornips, de Rooij, Siemund, Gogolin, Davydova and Schulz2013).
(3) 12 youngsters between the ages of 18 and 20 were selected by means of the classical Labovian methodology, and divided into three groups according to their language backgrounds (Hinskens Reference Hinskens, Kern and Selting2011). This variable is based on three values: Dutch (n = 4), Moroccan Arabic/Berber/Dutch (n = 4) and Turkish/Dutch (n = 4). Pairs of speakers who did not necessarily know each other but who belonged to the same cell according to ‘ethnicity’ (in-group and out-group conversations) were recorded. The recordings took place at their schools in Amsterdam and Nijmegen. From this third case study, ten minutes of each conversation have been transcribed.
New word order patterns in contemporary urban vernaculars
The V2 feature in Modern Germanic
The respective standard languages considered here all display the so-called V2, which means that exactly one constituent precedes the finite verb in declarative clauses.10 In this chapter, V2 constraint, which should be taken as a neutral description with regard to the various theoretical instantiations of this phenomenon.11 Most commonly, the first position of the clause is occupied by the subject, followed by the finite verb (see (1a–4a)). However, the first position is not restricted to subjects in any of the four languages described here. It can host a variety of constituents, such as objects, adverbials and even verbal phrases. The only requirement is that only one constituent precedes the finite verb. Thus, whenever a declarative clause begins with something other than the subject, the subject must follow the finite verb, as in (1b–4b).
b. I går var jeg på kino. (*I går jeg var på kino)
b. Igår var jag på bio. (*Igår jag var på bio.)
b. Gestern war ich im Kino. (*Gestern ich war im Kino.)
b. Gisteren was ik in de bioscoop. (*Gisteren ik was in de bioscoop.)
The following sections give an overview of newly developed word order patterns within the left sentence periphery in Germanic-based urban vernaculars throughout Europe. Some urban vernaculars show deviations from this usually very strict constraint whereas other languages, such as Dutch, do so rarely (for a discussion, see below). We examine the ways in which these patterns might be influenced by information-structural factors, or, in other words, to what extent different urban vernaculars make use of newly available syntactic means in order to mark the information-structural status of phrases.
Deviations from V2 in contemporary urban vernaculars: Norwegian, Swedish and German
With the exception of Wiese (Reference Wiese2006, Reference Wiese2009, Reference Wiese2012), Kern and Selting (Reference Kern and Selting2006) and Selting and Kern (Reference Selting and Kern2009), syntactic phenomena have played only a minor role in studies on urban vernaculars in Germany. In Sweden and Norway, however, the syntactic characteristics of contemporary urban vernaculars have played a central role in descriptions made thus far (Ganuza Reference Ganuza2008a, Reference Ganuza2008b, Reference Ganuza, Quist and Svendsen2010; Nistov and Opsahl Reference Nistov, Opsahl, Mæhlum and Åfarli2014; Opsahl Reference Opsahl2009a; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010). The same holds for urban vernaculars in Dutch, in particular with respect to grammatical gender-related phenomena in the adnominal domain, word order in embedded clauses and the use of the auxiliary gaan (Cornips Reference Cornips, Bennis, Extra, Muysken and Nortier2002, Reference Cornips and van der Sijs2005, Reference Cornips2008; El Aissati, Boumans, Cornips, Dorleijn and Nortier Reference El Aissati, Boumans, Cornips, Dorleijn, Nortier and van der Sijs2005).12 A syntactic pattern that is addressed not only by these papers but that is repeatedly mentioned in the literature on Norwegian, Swedish and German urban vernaculars is the pre-positioning of material in front of an otherwise ‘normal’ verb second declarative clause, i.e. an XSV order, see (5a–5c):
As the languages in question are V2 languages, the pattern in (5) is highly unexpected. As shown above, main clauses normally do not allow for more than one constituent preceding the finite verb. In contemporary urban vernaculars, however, this restriction seems to be less robust, and the questions arise (i) as to whether we are dealing with a genuine and general violability of V2 in these vernaculars or whether we see here a clearly delimitable construction (that can be filled with varying lexical material), and (ii) the functional motivation behind it. To answer these questions, we will first take a closer look at the structure of this word order pattern, and will then examine the functional contribution of this syntactic construction from an information-structural point of view.
The typical appearance of what appears to be a violation of V2 in the Norwegian, Swedish and German data is the order ‘adverbial – subject – finite verb’ (AdvSV), as in (5) above. Further examples are given in (6) and (7) (the finite verb is in italics, capitals mark accent):
The overall picture points to a systematic pattern: the elements that precede the finite verb show a rather coherent behaviour with respect to their syntactic functions, their semantics and their discourse pragmatics across the languages considered here.
A pattern arises in relation to syntactic categories and functions: the constituent that directly precedes the finite verb is, as far as we can see, almost without exception the subject (this is the case in all cases of XSV attested in Norwegian, in 217 of the 218 examples in Swedish, and in 51 of the 55 examples in German). These subject constituents are in most cases pronominal. Full subject DPs are comparatively rare; in the Norwegian data, for instance, 170 of the 194 subjects examined are pronominal. The same pattern emerges in the Swedish data, where 191 of the 218 examples contain a pronominal subject, and also in the German examples where 41 of 51 subjects have the form of a pronoun. Hence, we are dealing with subjects that consist of little phonetic material and are virtually always unaccented.
The very first position of AdvSV clauses is invariably filled by adverbials.15 They occur as PPs (see (8)), as DPs (9) and even as CPs (10), but most often they have the form of simple adverbs ((5)–(7) above and (11)):
With regard to semantics, the vast majority of adverbials in first position have temporal meaning; for example, gestern (Ge.), i går (No., Sw.) (‘yesterday’); jetzt (Ge.), nå (No.), nu (Sw.) (‘now’); danach (Ge.), etterpå (No.) (‘afterwards’); and jedes Jahr (Ge.) (‘every year’). Further temporal adverbials attested are i dag (No.) (‘today’); dann (Ge.) (‘then’); nachher (Ge.) (‘afterwards’), among others. In addition, a comparatively large number of conditional adverbials can be found, whereas local, modal and causal adverbials are less frequent. In the Swedish data, examples of XSV most commonly begin with the connective adverb (å) sen(‘(and) then’) (95 of 218, i.e. 43.6%) (Ganuza Reference Ganuza2008a: 97–8; Ganuza Reference Ganuza, Quist and Svendsen2010: 38), a pattern also very common in the German corpus.
The case of Dutch
In Dutch, too, the violation of V2 is the order ‘adverbial – subject – finite verb’, as illustrated in (12) below. This order is, however, very rare in all three case studies. In this respect it differs crucially from the spontaneous speech of adult second-language learners of Dutch, who are claimed in the literature to produce this order (Appel and Muysken Reference Appel and Muysken1987: 91), and recently migrated Turkish- and Moroccan-speaking children in the beginning of the 1980s (Appel Reference Appel1984; see also Schwartz and Sprouse Reference Schwartz, Sprouse and Archibald2000):
The corpora of the three examined case studies of Dutch show only three tokens of V2 violations. One is the adverbial toen (‘then’), which requires V2 in Standard Dutch.16 The other two tokens contain the adverb daarom (‘that's why’). All examples are presented below.
Table 4.1 shows the (potential) occurrences of V2 and deviations from it in the three corpora for daarom (‘that's why’) (13b, 13c), misschien (‘maybe’), nu (‘now’), dan (‘then’) and eigenlijk (‘actually’). Only in the case of the adverb dan (‘then’) and toen (‘then’) are there enough tokens (n = 537 and 121, respectively) to claim that the Dutch urban vernaculars do not show a structural pattern of V2 variations.
Table 4.1. Occurrences of V2 and XSV patterns with selected adverbials in Dutch.
Notes.
a This corpus is not accessible for researchers outside the ‘Roots of Ethnolects project’ yet. Therefore, the countings are taken from Lukassen (Reference Lukassen2011).
b In this interview setting the youngsters were asked about their future plans which elicited many ‘(and) then’ sentences.
It is important to point out that the AdvSV order is not the one that shows up when youths in Dutch multilingual urban settings are stereotyped linguistically (Nortier Reference Nortier2001). This would provide more evidence that this order is not available in Dutch. In Swedish, however, this order is strongly associated with ways of speaking among youths in multilingual urban settings, and is not least apparent in media discussions and in literary representations of contemporary urban Swedish vernaculars (e.g. Källström Reference Källström, Ledin, Palicki, Melin, Nilsson, Wirdenäs and Åbrink2006, Reference Källström, Quist and Svendsen2010). In Germany, the AdvSV pattern is associated with youth speech in multilingual urban settings as well, both in stylized variants/imitations used by various comedians (Kotthoff Reference Kotthoff, Lewandowska-Tomaszczyk and Pułaczewska2010) and to some extent in the media. The same can be said of the situation in Norway (although it has not been systematically analysed).
How then to account for the fact that there does not seem to be a structural pattern of V2 violation in Dutch? There are three factors to consider. First, the almost complete absence of V2 violations may indicate that Dutch is a stricter V2 language than are German, Norwegian and Swedish. A second factor may simply be the size of the corpora, since the Dutch corpus is considerably smaller than the German, Swedish and Norwegian corpora. Finally, one must consider methodology. It might be the case that interview settings in Utrecht and classical Labovian in- and out-group settings in Amsterdam/Nijmegen are not the optimal settings in which to elicit V2 deviations, as the peer-group conversations in Germany, Sweden and Norway do. The Rotterdam corpus was however collected through participant observation and recordings of peer-group conversations with no adults present. Thus, from a methodological perspective, it is difficult to account for why the four Surinamese-Dutch-speaking youngsters do not show V2 violations. Seen together, it seems to be the case that Dutch does not allow loosened grammatical restrictions with respect to the XSV order.
A functional interpretation in terms of discourse pragmatics
In order to find motivation for the emergence of AdvSV constructions, it is worthwhile to take into account the information structure of the utterance. It is well known that speakers give their utterances an informational relief by separating focal from background information, new information from familiar facts and so on. An utterance can thus usually be divided into topic (= what the sentence is about; typically referring to already introduced or inferable entities) and comment (= what is said about the topic; this domain consists of or contains the focused element(s), i.e. new information).17 As Wiese (Reference Wiese2007, Reference Wiese2009) first argued for Kiezdeutsch, AdvSV structures follow a rather consistent information-structural pattern, which can be specified as sequence of frame setter plus topic to the left of the finite verb. The examples attested in the Norwegian, Swedish and – by now extended – German data confirm this analysis: the information-structural status of the preverbal subject phrase is always that of a topic. The kind of topic can be even more specifically described as that of a familiarity topic, which means that it refers to discourse referents that have been mentioned in the previous context, or that are generally known (this corresponds perfectly to the fact that the topic is mostly pronominal in form).18
The left-most adverbial in an AdvSV constructions fulfils the function of providing an interpretational frame or anchor for the following statement, first, in terms of time, place, condition (in the case of adverbials meaning ‘from now on’, ‘yesterday’, ‘every year’, ‘if you are in school’ and so on), or second, more abstractly, in terms of discourse linking (as is the case in certain uses of the equivalents of ‘then’ and ‘afterwards’). The usage of adverbials in the first function correlates with the notion of frame setters, whose basic function is ‘to limit the applicability of the main predication to a certain restricted domain’ (Chafe Reference Chafe and Li1976: 50). Used in this way, adverbials are no longer directly event-related (Maienborn Reference Maienborn2001: 191); rather, they carry discourse pragmatic functions. In cases like (14) (as well as (6), (8b, 8c) and (9)), the adverbial at the beginning of the sentence limits the validity of the statement to a specified temporal domain (implying a contrast to other time periods for which the uttered statement does not hold):
The second function of fronted adverbials covers usages as contextualizers. Serializing adverbials with the meaning ‘then’/‘afterwards’, as in (7) and (11c), appear as discourse-structuring elements, which subdivide a larger stretch of discourse into smaller units (which can still be larger than sentences) and arrange them in an easily processible linear order. Naturally, the ordering of events within a narrative resembles the ordering of the narrated events in reality. Adverbials that express the chronological order of events might therefore function simultaneously as serializers with respect to the pure textual structure of a narration (Schalowski, in press). Examples from German and Swedish are given in (15) and (16), respectively:
(16) så dom kutar upp för dom här spa+19 vagnarna # å sen så du vet dom kutar dom kutar å allting å sen dom tittar ner å sen dom ser att dom e på väg upp dom hära # zombies eller va dom är # å sen eh # ah å sen de e du vet hon kollar upp hon börjar skrika å du vet han kommer efter han ba VAD # sen står de en kille där me såhär skydd å allting han står såhära # han å du vet hon blev rädd att han va en av dom han ba ah men kom kom # hon ba du vet såhära han ba knacka på första dörren # så han släppte för väg dom å så stod han kvar så slog han på dom här som kom
[so they run up on these spa+ shopping carts # and then you know they run they run and everything and then they look down and then they see that they're on their way up these # zombies or whatever # and then eh # yeah and then there's you know she looks up she starts screaming and you know he comes after he's just like WHAT # then there's a guy there with protection and everything he stands there # he and you know she got scared that he was one of those he just but come come # she just you know he just knocks on the first door # so he lets them through and then he remains there and then he hits these that came]
Both frame setters and (familiarity) topics generally tend to occupy a position as far left as possible (Erteschik-Shir Reference Erteschik-Shir2007: 105; Molnár Reference Molnár and Reis1993: 178). Likewise, discourse linkers must be in front of the sentence or discourse unit they link to a previous item. In Germanic V2 languages, however, the possibilities to mark these discourse pragmatic functions by syntactic means, namely by placing the respective elements at the left edge of a sentence, are strictly limited by the V2 rule. Here, the newly developed word order options in the German, Norwegian and Swedish urban vernaculars allow a more liberal realization of information-structural strategies, for it is feasible to have more than one constituent in front of the verb in order to mark their pragmatic function(s) syntactically (see also Wiese Reference Wiese2009; Wiese, Freywald and Mayr Reference Wiese, Freywald and Mayr2009).20
Conclusion
In this chapter, we addressed a syntactic pattern that has been pointed out in several descriptions of Germanic contemporary urban vernaculars across Europe. In these vernaculars, a word order pattern has developed that does not obey V2 in declarative sentences, a constraint that is strict in the respective standard languages. What can instead be found is a serialization type with two constituents preceding the finite verb (XSV order).
Using data from natural speech in Norwegian, Swedish, German and Dutch, we have provided comparative descriptions and analyses of this pattern, thus going beyond the studies on individual languages available thus far. The comparative approach revealed striking parallels between Norwegian, Swedish and German urban vernaculars with regard to both structural shape and usage of the XSV pattern.21 It became apparent that XSV typically occurs in peer conversations; instances of XSV are remarkably rare or even entirely absent in interviews and written texts (Freywald, Mayr, Schalowski and Wiese Reference Freywald, Mayr, Schalowski and Wiese2010; Ganuza Reference Ganuza2008a, Reference Ganuza, Quist and Svendsen2010; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010). As to the categorical status of the two constituents in front of the finite verb, there is a confinement to ‘adverbial plus subject’ observable across all languages considered here; the pattern has therefore been labelled AdvSV. The motivation for the occurrence of AdvSV lies in discourse pragmatics. The prefinite subject predominantly represents a familiarity topic (a fact that is reflected in that it mainly assumes pronominal form). The adverbial that can be placed in front of the subject and the finite verb delimits a statement temporally (or, less often, locally or modally), or expresses a condition under which the uttered proposition holds – that is, it functions as a frame setter. At the same time, it helps structure the (narrative) discourse, the most common adverbials being the respective equivalents of the cohesive elements ‘then’, ‘afterwards’, ‘after this’. Apart from these clearly defined functional domains, the standard-like V2 pattern does not dwindle in general in the contemporary urban vernaculars studied here, but remains a regularly occurring pattern in declarative sentences. Considering this scenario on a large scale, one is perhaps reminded of the process of the loss of V2 during the Middle English period. As Los (Reference Los2009), for example, points out, during the course of losing V2, i.e. of establishing strict SVO order, the preverbal position is less and less often filled by deictic, inversion-causing adverbials but rather becomes increasingly confined to subjects. After the loss of V2, pragmatically less prominent, preverbal elements must have the form of the subject, their function being that of ‘unmarked discourse linking’. Preposed non-subject constituents, however, are interpreted as ‘marked themes’ and serve – in the case of adverbials – as a means of text cohesion (Los Reference Los2009; for another detailed study of this transitional process, see e.g. Kemenade and Westergaard Reference Kemenade, Westergaard, Meurman-Solin, Los and López-Couso2012). However, whether we witness the beginnings of a similar development in the urban vernaculars considered in this chapter remains a matter of pure speculation at this stage of research. To study syntactic developments of this kind thoroughly, detailed long-term investigations are needed. So, we must leave this open question to future research at this point.
Only very few examples of AdvSV could be found in Dutch, which must be considered an exception in this respect. As long as no further data collections are available, Dutch provides some evidence that the observed novel syntactic ways of realizing information-structural preferences might be applicable only when grammatical restrictions are loosened. If this condition is not fulfilled, as is the case in Dutch, AdvSV may not occur.
Taken together, the emergence of AdvSV is not a matter of random syntactic variation or incomplete language acquisition, but rather follows a systematic, delimitable pattern. What we observe is the emergence of a structural option that is not a reflex of dropping V2 in general, but is confined to the order ‘adverbial – (pronominal) subject – finite verb’. As we have shown, this restriction is clearly motivated by information-structural grounds: by using the AdvSV structure, both frame-setting and topical constituents are enabled to appear at the left edge of the sentence, that is, on the left of the finite verb. Leaving Dutch aside, the languages investigated – Norwegian, Swedish and German – behave strikingly similarly with regard to AdvSV order, in both structural and (discourse-)functional respects.
1 Considering the multilingual/multiethnic situation in these speech communities the term ‘multiethnolect’ has been coined (see e.g. Clyne Reference Cornips and van der Sijs2000; Freywald, Mayr, Özçelik and Wiese Reference Freywald, Mayr, Özçelik, Wiese, Kern and Selting2011b; Quist Reference Quist2000, Reference Quist2008; Svendsen and Røyneland Reference Svendsen and Røyneland2008). While this term avoids the establishing of correlations between language use and individual ethnicities, it still might involve problems in that it correlates the use of particular linguistic practices with the social category of ethnicity.
2 The development of this grammatical feature might be part of larger processes of ongoing language change. For some brief speculations on developmental processes, see below.
3 By 1 January 2011 the immigrant population in Oslo was 28.4 per cent; that is, 170,000 of nearly 600,000 inhabitants.
5 Financed by the Bank of Sweden Tercentenary Foundation.
6 This figure does not include examples of kanske_S_V ‘maybe_S_V’, which are well-known exceptions to the V2 constraint in standard and regionalized varieties of Swedish (e.g. SAG 1999).
7 The project with the current title ‘The KiezDeutsch Corpus. Analyses at the Periphery’ (PI: Heike Wiese) is part of the Collaborative Research Centre ‘Information Structure’ (SFB 632) at the University of Potsdam, Humboldt-University Berlin and Free University Berlin, funded by the German Research Foundation (DFG).
8 For further details, see www.sfb632.uni-potsdam.de/en/cprojects/b6.html.
9 This search was conducted in two steps. (1) A random sample of 30 declarative sentences was manually searched for XSV structures for each anchor speaker. For those who produced any instances of XSV the entire recordings were searched. (2) Based on the results of step (1), a search of the entire corpus was carried out covering only the adverbs danach, nachher (both ‘afterwards’) and gestern (‘yesterday’), which were repeatedly produced in XSV structures by several speakers (see also Schalowski, Freywald and Wiese Reference Schalowski, Freywald and Wiese2010). For technical reasons, exhaustive quantitative syntactic analyses are not yet feasible.
10 This is also the case for the other Germanic languages, except English (see e.g. Haider and Prinzhorn Reference Haider and Prinzhorn1986; König and van der Auwera Reference König and van der Auwera1994; Wechsler Reference Wechsler, Leffel and Bouchard1991; and in particular Vikner Reference Vikner1995). In all four languages under consideration there is, however, a limited number of previously known exceptions to the V2 constraint, such as left dislocation constructions (which – unlike the structures discussed in this chapter – require a resumptive element, e.g. a pronoun or the respective cognates of the No. så ‘so’) and sentences with fronted kanske/kanskje ‘maybe’ in Swedish and Norwegian (cf. Sw. då kanske han ser det, lit. ‘then maybe he sees it’; see also footnote 6 above). These constructions are, however, not dealt with in this paper.
11 In Standard German, it is almost a consensus that a root clause is a CP and that the finite verb lands in C after having undergone movement through TNS and AGR, if one implements the SplitInfl account in which I is split into AGR and TNS (see e.g. Grewendorf Reference Grewendorf, Jacobs, von Stechow, Sternefeld and Vennemann1995; Schwartz and Vikner Reference Schwartz, Vikner, Belletti and Rizzi1996). However, Zwart (Reference Zwart1997) has proposed for Dutch that movement to C is variable in root clauses and that the finite verb moves up to spec-C, via AGR and TNS, in root clauses with inversion only, but stays below C, namely in spec-AGR, in root clauses with a straight order.
12 As is the case with the thorough code-switching studies of Dutch/Turkish in Tilburg and Dutch/Moroccan-Arabic in Utrecht by Backus (Reference Backus1996) and Boumans (Reference Boumans1998), respectively.
13 The abbreviations used to identify speakers consist of four parts: neighbourhood of speaker, speaker number, sex of speaker and family language of speaker. The abbreviation in (6b) indicates a speaker who lives in a multiethnic neighbourhood (→ Mu, as opposed to Mo = monoethnic neighbourhood), who has the number 9, who is female (→ W, as opposed to M = male) and whose family language is Turkish (→ T, as opposed to D = German, A = Arabic, K = Kurdish).
14 The code used to identify speakers in the SUF-corpus contains information about the school attended by the speaker (Stockholm schools are B, K and L, Gothenburg schools are P and S, and Malmö schools are C, D and E) and speaker number (which also identifies the sex of the speaker: 1–24 are male speakers, and 25–46 are female speakers).
15 It should be noted that the results for German partly rest upon a search that takes into consideration only preselected adverbs, and thus might be slightly biased (see also footnote 9).
16 Utterances that show a long silence or hesitation, such as ‘uhhh’, as illustrated in (i), or restarts, as in (ii), are not analysed as violations:
17 We follow the definition of ‘topic’ given in Reinhart (Reference Reinhart1981) – adopted and further developed by Lambrecht (Reference Lambrecht1994), Jacobs (Reference Jacobs2001) and Krifka (Reference Krifka2008), among others – who metaphorically describes the topic–comment structure as a kind of file card system. The topic indicates the entry under which the information given in the comment should be stored.
18 Given information – as conveyed by familiarity topics – is prosodically not prominent, and is typically represented by anaphoric expressions such as pronouns, clitics or definite phrases (see Krifka Reference Krifka2008: 262–4).
19 Key to transcript: + = self-interruption, # = short pause, capital letters = emphasis.
20 For a broader discussion of how linguistic modules and subsystems interact with extralinguistic domains (such as the general hierarchization of information) and how the expression of extralinguistic concepts must be formally adjusted to the grammatical possibilities available in a particular language, see for instance Culicover and Jackendoff (Reference Culicover and Jackendoff2005), Jackendoff (Reference Jackendoff1997, Reference Jackendoff2002), Wiese (Reference Wiese2003, Reference Wiese, Härtl and Tappe2004) and Wiese, Freywald and Mayr (Reference Wiese, Freywald and Mayr2009).
21 For a similar cross-linguistic approach to parallel developments of a grammatical phenomenon, namely the functionalization of comparative/deictic particles, see Chapter 5.
5 Functional gains: a cross-linguistic case study of three particles in Swedish, Norwegian and German
Introduction
Contemporary urban multilingual settings in Europe lead to rich sources of language contact involving the majority languages and a range of typologically diverse minority languages, such as Arabic, Urdu, Turkish, Kurdish, Croatian and Sranan. The linguistic outcomes of these settings may result in the emergence of what we shall refer to as contemporary urban vernaculars (see Chapter 2). By choosing this label, we make clear that this is a way of speaking related not to a particular ethnicity, but rather to contemporary urban areas in Europe (see also Chapter 3 for a discussion of terms such as ‘ethnolect’, which assume pre-existing social categories such as ‘ethnicity’).
Multilingual urban settings constitute a linguistic environment that is particularly open to linguistic variation and innovation, and might also support a faster pace of language change – in comparison not only to national standard varieties that are more restricted by normative processes, but also compared to informal varieties and styles that are set in more monoethnic/monolingual speech communities and that cannot draw on this kind of language diversity. This linguistic dynamics makes contemporary urban vernaculars particularly interesting for investigations that target the linguistic system and the interactions of different grammatical and extragrammatical subsystems in the emergence of new patterns. This chapter is devoted to a case study from this domain that brings together similar lexical items from three Germanic languages, namely Swedish sån, Norwegian sånn and German so (‘such (a)’). We will show that these items can be used in a similar way across the three languages as grammatical and pragmatic markers, building on a pattern of semantic bleaching and functional gain.
The rationale underlying our choice of study is the fact that the accumulation of studies on linguistic developments in urban multilingual settings over the last decades (see Chapter 1, this volume; Quist and Svendsen Reference Quist and Svendsen2010; Wiese Reference Wiese2009) has cleared the ground for a stage where we are now able to investigate parallel grammatical developments in different urban settings across Europe: the time is ripe to make the step from studies across Europe to cross-European studies.
This has the potential to reveal general tendencies of language variation and change reflected in these new linguistic developments, and by doing so, will also emphasize another important aspect for the understanding of contemporary multilingual urban settings. The studies performed so far have provided us with valuable insight into the complexities associated with a multicultural and multilingual urban reality. However, one aspect we wish to highlight is that in addition to the advantageous environment that multilingual speech communities provide for language change, there might also be a perceptual advantage in the salience of these phenomena in such communities. The (relatively) recent attention given to contemporary spoken urban vernaculars – among linguists, the media and the general public alike – may have put a spotlight on phenomena that might be first observed here, but perhaps constitute instances of more general cases of language variation and change. Some of the linguistic traits associated with contemporary urban vernaculars are not necessarily restricted to multiethnic urban areas, and in some cases might not even be significantly more widespread here than in more monolingual settings.
Ekberg (Reference Ekberg, Quist and Svendsen2010: 29) poses an important question: could it be that the foreign-sounding way of speaking Swedish that is associated with the multiethnic speech style described in Malmö known as ‘Rosengård Swedish’ triggers the listener to search for (other) non-standard traits? If we include the researcher in the group of listeners pointed to by Ekberg in her question, we approach another explanation for our choice of object of study. The following presentation of some of the grammatical aspects associated with contemporary urban vernaculars in Germany, Sweden and Norway reveals interesting parallels, but we also find examples of characteristic linguistic traits that may not be restricted to multilingual settings per se. This is especially true when we turn to phenomena associated with the grammar–pragmatics interface. These facts do not undermine, but rather reinforce, the importance of examining the special linguistic dynamics at play in urban multiethnic areas, as we will show in the next section.
In the introduction to this volume we saw how a linguistic–anthropological viewpoint reveals how language form, practices and ideologies may be understood as interwoven phenomena, and how various aspects must be taken into consideration if one wants to grasp the ‘total linguistic fact’. A cross-linguistic viewpoint as employed here, that covers related grammatical phenomena arising in parallel urban environments, may contribute to a further understanding of these issues, especially so in combination with studies carried out from an interactional point of view (e.g. in Part V of this volume). At a meta-level of identity and us–them categories, our study contributes to ‘normalis[ing] the kind of urban speech we are examining, moving it out of the “marked” margins’ (Rampton Reference Rampton and Deppermann2013:78). As we show here, the developments we find in new urban vernaculars are not idiosyncratic or alien to the linguistic domain of the respective majority languages, but part and parcel of the variation we find there. They are embedded in the range of options we find, in our case, for Swedish, Norwegian and German, respectively, in a way that defies a distinction of ‘allochthonous’ versus ‘autochthonous’ varieties.
Linguistic characteristics associated with contemporary urban vernaculars
Our focus on three of the Germanic languages, namely Swedish, Norwegian and German, allows us to target three closely related lexical elements with our analysis, namely sån, sånn and so ‘such (a)’ (more on these elements in the following section), and to investigate the strikingly parallel development they undergo cross-linguistically. Since, as we will show below, the findings from this study point to general patterns of functionalization, they will be relevant to other European linguistic settings as well, in particular given the cross-linguistic similarities of contemporary urban vernaculars at a general level that have so far emerged from different national studies (see also Wiese Reference Wiese2009).
In general, we can identify three patterns of development in contemporary urban vernaculars: (1)contact-induced changes, such as the emergence of new lexical loans from some of the background languages available in the communities, (2) new patterns that arise from an extension of lexical material or grammatical patterns offered by the respective majority languages and (3) developments that reflect phenomena of general language change, such as the interaction of weaker grammatical constraints and a more direct realization of information-structural preferences.
It is only in the first case that elements specific to particular background languages come into play. In contrast to this, the other patterns are supported language-internally (either from the point of view of particular majority languages or from general linguistic tendencies). This means that we can expect similar developments cross-linguistically, and that contemporary urban vernaculars are not exotic, but rather form an integrated part of the linguistic spectrum found within the majority languages.
An example of the first phenomenon is given in the German data in (1), where the Arabic abu (‘father’, here as an abbreviation for more general insults involving the addressee’s father) is used as an exclamation signalling displeasure, best translated into English as ‘my!’
(1) (German)
Innovative uses of lexical items are not only based on loan words from migrant languages, but also on new or extended uses of existing linguistic material already present in other varieties of the respective languages, realizing the second pattern mentioned above. In Sweden, for instance, the tag du vet (‘you know’) – as can be seen twice in the Swedish example (2) – has been pointed to as characteristic for speakers in multiethnic areas in Malmö (Svensson Reference Svensson and Ekberg2007), and is also observable in our Norwegian data (Opsahl Reference Opsahl2009a: 140). An elaboration and/or generalization of grammatical options offered by the linguistic system of the majority languages can also be observed at the morpho-syntactic level. Examples of this are some of the phenomena described by Kotsinas (Reference Kotsinas, Allén, Andersson, Löfström, Nordenstam and Ralph1988a, 1988b, Reference Kotsinas, Daun and Klein1996) for Swedish, such as the elaboration of the domain and function of certain prepositions, or productive light verb constructions in German (Wiese Reference Wiese2006). Yet another example of morpho-syntactic characteristics is the emergence of new determiners, as described for instance by Ekberg (Reference Ekberg, Quist and Svendsen2010), and visible in (2) (‘så var de(t) sån lite(n) bebis’), elaborated on in the next section of this chapter. (2), from Ekberg (Reference Ekberg, Quist and Svendsen2010: 18), is repeated with more context in (18) below.
(2) (Swedish)
A phenomenon that has been observed across Germanic V2 languages is new word order options for the left periphery (Ganuza Reference Ganuza, Quist and Svendsen2010; Kern and Selting Reference Kern and Selting2006; Opsahl Reference Opsahl2009a; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010; Wiese Reference Wiese2009), which might be related to a more direct realization of information-structural preferences (see Wiese Reference Wiese, Gregersen, Parrott and Quist2011b for German and Chapter 4, this volume, for further studies including Swedish and Norwegian data). The Norwegian data in (3) give an example of such a new word order option (‘Etterpå jeg skal ta’, AdvSV), together with an example of the absence of definiteness marking on the NP (‘alle de brus’, where one would expect ‘all den brusen’ (-en def) or ‘alle de brusene’ (-ene def pl)), in line with the second pattern identified above.
Related to this pattern of extending existing grammatical options, multiethnic areas might also support a faster pace of language change in general. In Oslo, for instance, this seems to be the case regarding the apical pronunciation of the /kj-/ sound, resulting in a well-known merger – at least for Norwegians – of the two sounds /kj-/ and /skj-/ (Opsahl and Røyneland Reference Opsahl and Røyneland2009). In Germany, a phenomenon known from youth language in general, namely the use of voll (‘full’) as an intensifier, seems to be more widespread in multiethnic speech communities (Wiese Reference Wiese and Deppermann2013).
The characteristics we observe at lexical, morphological and syntactic levels are supported by the particular linguistic dynamics of contemporary urban vernaculars. A large number of speakers of such vernaculars have grown up with two or more languages, leading to multilingual repertoires that encourage a higher linguistic flexibility than in monolingual speakers. This does not, however, exclude monolingual speakers from using these features, or from taking part in linguistic practices associated with a modern urban multicultural reality. All the communities from whom our data are drawn include monolingual speakers of the majority languages, and speakers use these features regardless of their language backgrounds.
The examples presented above and those elaborated on below are types of linguistic variation that are not random deviations from (standard) majority languages, but rather constitute coherent (sub)systems of their own (see also Quist and Svendsen Reference Quist and Svendsen2010). This point is especially clear if we turn to the main focus of this chapter, the case of sån, sånn and so. What is particularly interesting is that we find two similar functional developments for elements with a common primary semantic meaning (‘comparison’) in a similar way across the three languages under investigation. This suggests that what we observe here are not idiosyncratic developments, but rather instances of general processes that support similar patterns of functionalization in Germanic. These patterns are highlighted in multilingual environments, but they are – as pointed out above – not restricted to them. Rather, what we find here are options that are in principle available in a linguistic system and will also be realized in other variants of informal language, if possibly less extensively in terms of frequency or entrenchment.
Sån, sånn and so
In their original meaning, sån, sånn and so can be translated into English as ‘such a’/‘such’, with an indexical modal meaning answering to ‘how’. Norwegian sånn and Swedish sån are short forms of sådan, which originates in Plattdeutsch. In the standard varieties of Swedish and Norwegian, sån/sånn are typically used as an attribute in an indefinite noun phrase (sån säng ‘such [a] bed’), in Norwegian also as an adverb (as in du kan ikke gjøre sånn ‘you can’t do like this!’). German so is a polyfunctional element that can be used as an adverb and also, based on this, as a conjunction, and a complementizer, similar to Swedish and Norwegian så. The usages as conjunction or complementizer build on the indexical lexical content (Wiese Reference Wiese2011a).
In addition to their usage as content words, sån, sånn and so can also be used as purely functional markers. In this usage, the original meaning of these elements is lost; they undergo semantic bleaching, trading content for function. It is this functional usage that we investigate in this chapter. We have chosen these particular items for our case study on functionalization because they bring together developments at two levels: (1) from a content word to a grammatical marker, namely to an element with a determiner function, and (2) from a content word to a pragmatic marker, namely to a focus marker. Let us briefly make clear what we understand to be the outcome in the two cases.
(1) In speaking of an element ‘with a determiner function’ rather than ‘a determiner’, we account for the fact that these elements fulfil some, but not necessarily all, functions that are usually associated with determiners. (For a detailed discussion and overview, see Himmelmann Reference Himmelmann, Haspelmath, König, Oesterreicher and Raible2001.) In particular, they mark nominal elements, and, if one adopts the syntactic perspective of a determiner phrase (DP), they can thus be understood to provide the outer functional layer of noun phrases. Unlike articles, they do not mark the entire range of morphological categories that a full determiner might indicate in the respective languages (i.e., number, case, gender).
(2) By a ‘focus marker’ we understand an element whose co-constituent is a focus expression; that is, an expression whose referent is focused at the level of information structure. This focusing characteristically involves alternatives or contrasts, and hence can be understood as an answer to an implicit or explicit question that requests a choice from a set of such alternatives (see, for instance, Jackendoff Reference Jackendoff1972; Jacobs Reference Jacobs1983; König Reference König1991; Krifka Reference Krifka, Féry, Fanselow and Krifka2007; Rooth Reference Rooth1985). When speaking of ‘focus markers’, we refer to dedicated functional elements that support this organization at the level of information structure, but that do not contribute any additional semantic content. That is, we refer to what König (Reference König1991: 29) calls ‘pure’ focus markers, in contrast to focus particles such as only, even or also that contribute their own restrictive, scalar or additive semantics (see Horn Reference Horn1996 and König Reference König1991 for a detailed analysis of only).
It should be noted that in Swedish, in addition to sån which is common among young speakers in Malmö, another item that originates in a comparative expression, namely såhär, can be used as a focus marker, in particular among adolescents in Stockholm. Såhär is a fusion of a deictic expression, så här, idiomatically ‘like this’, where the first element, så (‘so’), is a comparative-demonstrative adverb and the second, här (‘here’), a proximal demonstrative. Såhär is also written sär, representing the assimilated pronunciation. A Norwegian parallel to the Swedish såhär is the fact that Norwegian sånn is sometimes used in combination with herre (a variant of her ‘here’) and quite often with derre (a variant of der ‘there’) (Lie Reference Lie, Johannessen and Hagen2008: 87).1
In this chapter, we demonstrate that for this field of cognates – sån/såhär, sånn and so – we can observe a common pattern of functionalization that holds cross-linguistically. This functionalization leads from lexical words with full semantics to semantically bleached elements; that is, to elements that have lost their lexical semantics in favour of functional gains. Specifically, lexical words are transformed here into functional elements with focus marker and determiner functions. As we show below, the cognitive motivation of this functionalization is a core meaning of comparison and deixis we find associated with the lexical semantics on which the development builds.
Data
The findings we present below draw on three main empirical sources: the Malmö corpus, the UPUS/Oslo corpus and the Kiezdeutsch-Korpus (KiDKo). The Swedish data, here referred to as the Malmö corpus, is part of a larger corpus collected within the realm of the project Language and Language Use Among Young People in Multilingual Urban Settings during the years 2002 and 2003 (see Chapter 14), an elaborate collection of speech data that has become a well-established reference corpus for contemporary urban vernaculars in Sweden.
The Malmö corpus consists of data collected in two upper secondary schools in Malmö. The proportion of multilingual students differs between the schools, but in the classes from which the participants were selected, the number of students with foreign background was similar, approximately 65 per cent.2 The data consists of spontaneous speech collected with no researcher or other adult present. The participants either took part in semi-directed group conversations at school, or conducted self-recordings in situations they chose themselves, interacting with peers. In both cases, the recordings were made on mini-discs. The present study is based on a part of the Malmö corpus that comprises nearly 6 hours (≈ 46,000 tokens) of recorded speech of participants from two peer groups. The first, referred to as the C group, consisted of four bilingual students, whereas the other, the E group, consisted of three monolingual students. A perception experiment showed that two of the participants in the C group were regarded as speaking Rosengård Swedish, the local multiethnic variety, whereas two of the participants in the E group were regarded as not speaking Rosengård Swedish (Bodén Reference Bodén and Ekberg2007, Reference Bodén, Quist and Svendsen2010; Hansson and Svensson Reference Hansson [Bodén], Svensson, Branderud and Traunmüller2004). (The perception experiment included specimens of speech from all four informants in the C group, and two of the three informants in the E group.) Thus, the Swedish data referred to in this study very likely includes language use regarded as an instance of a multiethnic variety, as well as language use not regarded as an instance of a multiethnic variety.
The Norwegian data is primarily drawn from the UPUS-Oslo corpus, a spoken language corpus developed by the Oslo group of the UPUS-project (Utviklingsprosesser i urbane språkmiljø – ‘Developmental Processes in Urban Linguistic Settings’). The aim of the Oslo UPUS-project (2006–10) was to explore linguistic practices among adolescents in multilingual settings in Oslo and the possible emergence of so-called multiethnolectal speech styles, from both structural and functional perspectives (see Aarsæther Reference Aarsæther, Quist and Svendsen2010; Opsahl Reference Opsahl2009a, Reference Opsahl2009b; Opsahl and Nistov Reference Opsahl, Nistov, Quist and Svendsen2010; Svendsen and Røyneland Reference Svendsen and Røyneland2008).
Data was collected in the inner city district of Gamle Oslo, where the migrant population – according to Statistics Norway – is 36 per cent, and in the suburban city district of Søndre Nordstrand, where the migrant population is 48 per cent. The 65 adolescents who participated in the study were between the ages of 13 and 19, and they were all born and raised in Oslo. The data consists of video-recorded peer conversations and interviews, which typically took place at youth clubs in the adolescents’ neighbourhoods. A few of them also provided the project with self-recordings. The transcripts were automatically annotated with morphological information. The data is made available for researchers through an Internet-based interface – where the transcripts, audio- and video-files are linked together – developed by Tekstlaboratoriet (‘The Text Laboratory’) at the Faculty of Humanities, University of Oslo. At the time of writing, the database had not yet been completed (i.e. there are still some transcripts and video files that need to be linked to the interface). It is therefore not possible to provide the exact number of tokens in the corpus. An estimate for the total duration of transcribed recordings is 30 hours.
For the German data, we draw mainly on KiDKo (Wiese, Freywald, Schalowski and Mayr Reference Wiese, Freywald, Schalowski and Mayr2012, www.kiezdeutschkorpus.de), in addition to some data from listening-in situations and from interviews with young people in multiethnic neighbourhoods of Berlin. This corpus has been compiled within Project B6 (Kiezdeutsch, 2008–15, principal investigator H. Wiese) of the Special Research Area (SFB) 632 Information Structure: The Linguistic Means for Structuring Utterances, Sentences and Texts of the University of Potsdam, Humboldt-University Berlin and Free University Berlin. KiDKo is a corpus of spontaneous, informal conversations between adolescents in peer-group situations, conducted mostly in German. The corpus is based on self-recordings of young people (14 to 17 years of age) in Berlin. The recordings (audio-files) are accompanied by aligned transcriptions in XML format that will, in the future, also have syntactic annotations (syntactic categories, phrases, topological fields). The corpus uses EXMARaLDA (Extensible Markups Language for Discourse Annotation, cf. Schmidt and Wörner Reference Schmidt and Wörner2005), a corpus system that allows automatic searches at all levels (transcription, annotations, metadata on speakers and speech situations). The corpus consists of two parts. The main corpus contains speech data from young people in Kreuzberg, a multiethnic and multilingual neighbourhood of Berlin, with 17 anchor speakers and approximately 205,000 tokens of transcribed recordings. Speakers are multilingual and monolingual, coming from migrant (mostly Turkish, Arabic, Kurdish) as well as non-migrant (German) backgrounds. The supplementary corpus contains data from young people in a community that is mostly monoethnic but has similar socioeconomic indicators, Berlin-Hellersdorf, with six anchor speakers and approximately 105,000 tokens of transcribed recordings. All speakers in the supplementary corpus are of non-migrant, monolingual German background.
In the following section, we first present an analysis of the functional usages observable here for sån/såhär, sånn and so, respectively. Against this background, the final section will summarize our findings, highlighting commonalities and differences across the three languages.
Functional usages of sån/såhär, sånn and so
The present section investigates the similar uses and parallel functionalization of the lexical items Swedish sån/såhär, Norwegian sånn and German so. Semantically these items have a common basis. In their primary lexical uses, sån, sånn and so have a comparative meaning, along with a more or less salient deictic/demonstrative meaning. According to our data (see our introduction), sån, sånn and so may function both as focus markers – pointing to the rhematic element in the utterance, i.e. the element containing the new information – and as determiners, as a substitution for the indefinite article. A telling example from the Norwegian data of this dual function is given in (4), where the first instance of sånn is interpreted as determiner-sånn and the second instance is interpreted as a pragmatic marker. In the first instance, sånn is combined with the name of a well-known (for Norwegians) television channel which is introduced here, suggesting that sånn substitutes an indefinite article. In the second instance, sånn is used to mark the rhematic element (kornåkerprogram ‘field.of.barley.programme’), while simultaneously signalling reservation regarding the chosen linguistic form.
(4) Norwegian (Opsahl Reference Opsahl2009a: 109)
In addition, in Swedish, the element såhär (see our introduction) can also be used in a focus-marking function. In the following, we will account for the usage of sån/såhär, sånn and so as focus markers, and the usage of sån, sånn and so in a determiner function. As mentioned in the introduction, we argue that the core meaning of comparison and deixis is the cognitive motivation of the semantic change of the three items. Against this background, we will suggest a possible path of development from lexical to functional usages at pragmatic and morphosyntactic levels.
Sån/såhär, sånn and so as focus markers
The particles sån/såhär, sånn and so have all been attested in a focus-marking function in informal, spoken varieties of Swedish, Norwegian and German, respectively; cf. (5)–(7). In this function they are unstressed, and precede the expression carrying the main stress. While the German so is fully semantically bleached (it has lost all of its lexical semantic content) and non-referential in this usage, Swedish sån/såhär and Norwegian sånn retain some of their primary comparative/deictic meaning (Ekberg Reference Ekberg, Quist and Svendsen2010; Johannessen Reference Johannessen, Enger, Faarlund and Vannebo2012).
(6) Norwegian (Opsahl Reference Opsahl2009a: 109)
(7) German (Wiese Reference Wiese2011a: 992)
While sån/såhär, sånn and so alike can take on the function of a focus marker, the range in which they can do so seems to be different in the three languages. The focus-marking function appears to be most elaborated in German, where we find evidence of so used both to mark information focus, as in (7) above, and to mark contrastive focus, cf. (8) (from Wiese Reference Wiese2011a: 993).
In comparison to the German data, there is less collected data in Swedish and Norwegian of sån/sånn as a focus marker. The existing data do, however, indicate a function similar to German so: sån and sånn are used to point to the rhematic element in an utterance, cf. (5) and (6) above. However, some of the examples found in the Swedish and the Norwegian corpora are difficult to analyse in regard to the pragmatic function, due to the overlapping function of sån/sånn as a determiner. Sån/sånn are also unstressed in a determiner function, and may precede a noun phrase that receives the main stress. For instance, in the Norwegian example in (6), sånn is interpreted as a determiner with a certain pragmatic content. Sånn often conveys an element of reservation towards the chosen linguistic form or to the proposition uttered, and invites the hearer to cooperate by mobilizing socially shared knowledge in order to arrive at an interpretation of the following constituent. This ability has inspired Johannessen (Reference Johannessen, Enger, Faarlund and Vannebo2012) to call sånn a modal determiner. The pragmatic function of reservation is rather clear in the phrase sånn kornåkerprogram (‘such field.of.barley.programme’) in (4), above. It may also to some extent be said to be present in (6), but in this case the pragmatic content is bleached, and we are approaching a situation where sånn is seemingly a substitute for the indefinite article.
Likely cases of a pure pragmatic function occur when sån/sånn do not agree with the following noun, as in the Norwegian examples in (9) (from Opsahl Reference Opsahl2009a: 109). In (9), the second noun område ‘area’ is neuter, whereas sånn is used in its common gender (the same holds for the Swedish example in (5), above, where the noun busfrö is neuter while sån is used in its common gender). In (9b), sånn, which is singular, is followed by a noun in plural. The non-congruent form can be taken as an indication that sån/sånn are not part of the noun phrase (Lie Reference Lie, Johannessen and Hagen2008), thus functioning as a pragmatic particle rather than as a determiner (this is also the position taken in Ekberg Reference Ekberg, Kern and Selting2011).
Wiese (Reference Wiese2009, Reference Wiese2011a) proposes a pragmaticalization path of the German so, leading from the propositional core meaning of comparison, via a semantically bleached meaning of vagueness – so denoting a generalized kind – to a function as a genuine focus marker, devoid of semantic content. As a vagueness marker, so may be used for hedging. Hypothetically, the crucial transitional phase from a lexical element to a focus marker is when so is used pragmatically as a hedging device. As hedging is most relevant in relation to information-heavy constituents, so will frequently co-occur with focus expressions, which, as Wiese argues, ‘make[s] it likely for it to be reinterpreted as a focus marker’ (Reference Wiese2011a: 1021).
The examples in (10) illustrate the dual function of so as a focus marker and as an element used for hedging (Wiese Reference Wiese2011a: 1022). In (10a), which is taken from a monolingual German context,3 this is further indicated by the preceding item irgendwie (‘somehow’); in (10b), which is taken from the multiethnic part of the KiDKo corpus (a discussion of an ongoing football match), the epistemic ‘I have a feeling’ sets up a suitable context for hedging. By bringing together both options for so, (10) exemplifies an ambiguous context necessary for a reinterpretation of so as a focus marker; that is, a context that is ambiguous between focus and hedging, and can thus support transitions from one interpretation to the other.
Swedish sån and Norwegian sånn may also be used as hedging devices in combination with marking the new information, as illustrated in (4), above, where the second instance of sånn signals reservation with regard to the linguistic form of the focused expression. See also the example in (13), below, where sån is used both as a determiner and as a hedging device. For Swedish and Norwegian sån/sånn, we can thus assume a developmental path as for focus-marking German so.
Evidence for the possible reinterpretation of a semantically full usage of comparatives as focus markers comes from a novel usage of the German term Ausdruck in multiethnic speech communities in Berlin (and quite possibly elsewhere). In its conventional usage, Ausdruck means ‘expression’ / ‘word’ and as such is a neutral term.4 In (11) and (12), however, it has taken on a negative semantic feature, meaning something like ‘bad expression’/‘swear word’ (from informal interviews in Berlin-Kreuzberg).
The new meaning for Ausdruck might derive from such contexts as in (12): teachers admonishing children and adolescents not to use swear words (at kindergarten or school) frequently do so by telling them not to use ‘such expressions’, as illustrated in (12).
In such contexts, the comparative (‘such’) stands adjacent to the focus-carrying expression, Ausdruck, which receives main sentence stress. In this position, solch (‘such’) might be reinterpreted as a focus marker. As such, solch would lose its semantic content in favour of functionalization, and as a result, Ausdruck alone would then be the term for ‘swear word’.
A parallel case to the use and development of sån/sånn/so is English like, which has a similar comparative lexical meaning and may function as a focus marker (e.g. Underhill Reference Underhill1988: 234). There is, however, a further parallel. Romaine and Lange (Reference Romaine and Lange1991) distinguish an additional use of like, namely the function as a quotative, which is hypothesized to be developed from its use as focus marker (see also Ferrara and Bell Reference Ferrara and Bell1995; Tagliamonte and Hudson Reference Tagliamonte and Hudson1999). In addition to their use as focus markers, Norwegian sånn, German so and sporadically Swedish såhär are also used as quotative markers (Ekberg Reference Ekberg, Kern and Selting2011; Golato Reference Golato2000; Hasund, Opsahl and Svennevig Reference Hasund, Opsahl, Svennevig, Buchstaller and van Alphen2012; Opsahl and Svennevig Reference Opsahl, Svennevig, Enger, Faarlund and Vannebo2012). As quotative markers they are not restricted to reported speech, but are used for dramatized events as well (Golato Reference Golato2000; Hasund, Opsahl and Svennevig Reference Hasund, Opsahl, Svennevig, Buchstaller and van Alphen2012). A further parallel example of grammaticalization from a focus marker to a quotative is Swedish ba (Eriksson Reference Eriksson1997). Ba is a reduced form of bara with the lexical meaning ‘only; just’. The same development has been described in Norwegian, but the Norwegian quotative is used in a non-reduced form, bare (Opsahl and Svennevig Reference Opsahl and Svennevig2007, Reference Opsahl, Svennevig, Enger, Faarlund and Vannebo2012).
Sån, sånn and so as determiner
In colloquial Swedish, Norwegian and German, in particular in multilingual urban contexts, we also find evidence of sån, sånn and so used in a determiner function; that is, in positions where they seemingly act as substitutes for the indefinite article (Ekberg Reference Ekberg and Ekberg2007, Reference Ekberg, Quist and Svendsen2010, Reference Ekberg, Kern and Selting2011; Opsahl Reference Opsahl2009a; Wiese Reference Wiese2006). The use of Swedish sån as a determiner seems, however, to be restricted to the south of Sweden; it has for example not been attested in Stockholm (approximately 600 km north of Malmö).
In the standard varieties of Swedish, Norwegian and German, an indefinite determiner, or a quantifier, must precede a singular count noun phrase. When no quantifier is present, the indefinite article is obligatory when the noun phrase has argumental status. In the Swedish corpus, there is a strong tendency to leave out the obligatory indefinite article en/ett (‘a/an’) in a noun phrase with sån (for details, see Ekberg Reference Ekberg, Kern and Selting2011). Consider the example in (13) where the singular indefinite NP lacks the article en. In Standard Swedish, sån NP would be preceded by en; that is, ensån NP (‘a such NP’) or merely construed as enNP (‘an NP’). Parallel examples for Norwegian and German are given in (14) and in (15).
(13) Swedish
(14) Norwegian (Opsahl Reference Opsahl2009a: 109)
(15) German (Wiese Reference Wiese2006: 256)
The primary meaning of Swedish and Norwegian sån/sånn is to compare two referents concerning type; in the Swedish example in (16) the comparand (B) serving to identify the comparee (A) is expressed in a relative clause introduced by som (‘as’).
The meaning of comparison is central also of German so, where the object of comparison (B) can be either explicit (introduced by wie (‘how’/‘as’)) or implicit (Wiese Reference Wiese2011a: 996):
The meaning of sån/sånn/so is not only comparative, but also deictic/demonstrative; the speaker is pointing to B, linguistically and/or physically, in order to denote A. As the comparand is needed to identify the designated entity, the comparand is part of the (immediate) semantic scope of sån (Langacker Reference Langacker1990).
Formally, sån/sånn/so replace the indefinite article in (13)–(15). The question is whether they also fulfil the function of introducing a new discourse referent. We will demonstrate this reasoning by looking closely at Swedish sån with a determiner function.
As pointed out by Heine (Reference Heine1997: 72), an early developmental stage of indefinite articles is the use as a presentative marker to introduce a new discourse participant presumed to be unknown to the listener. One of the contexts in the Swedish corpus where sån occurs without an indefinite article is precisely as part of a noun phrase; that is, the logical subject of a presentational construction, as is shown in (18). Here, the informant is telling her friends what happened when she was at the bank the other day, namely that she saw a cute little baby. She introduces the new discourse referent (the baby) by construing it as the logical subject in a presentational construction (var de(t) sån lite(n) bebis, ‘there was sån little baby’).
However, a closer look at the examples in (13) and (18) indicates that sån is semantically and functionally more complex than the indefinite article. Consider again the example in (13), where sån is used for the first mention of the referent säng ‘bed’; that is, for introducing a new discourse referent. In Ekberg (Reference Ekberg and Ekberg2007, Reference Ekberg, Quist and Svendsen2010, Reference Ekberg, Kern and Selting2011) the use of sån in (13) is analysed as having a recognitional function (Diessel Reference Diessel1999; Himmelmann Reference Himmelmann and Fox1996, Reference Himmelmann1997; Lindström Reference Lindström, Botley and McEnery2000; Sacks and Schegloff Reference Sacks, Schegloff and Psathas1979). The speaker introduces a new referent in the current discourse that is construed as known to the listener.6 The Norwegian data (Opsahl Reference Opsahl2009a) indicates that the same analysis can be applied to sånn in a determiner function, as in (19) ((6) from above is repeated here as (19c)):
More specifically, when using sån/sånn the speaker is referring to a referent that she assumes the listener is able to identify through shared knowledge and experience. Sån points to an imagined comparand needed to identify the comparee (the referent). The characteristics of the referent are then successively elaborated in the continuation of the discourse (cf. the second utterance from Gordana in (13)). This cataphoric use of determiner sån is equal to the one observed in the Norwegian example in (14), where the first instance of sånn points forward to the more precise mention of the introduced NP. The construal of the new referent as something known is supported by the co-occurrence of the pragmatic particle du vet (‘you know’), which emphasizes the closeness between speaker and listener while construing a common frame of reference (Svensson Reference Svensson and Ekberg2007). In the recognitional use of sån, the deictic/demonstrative meaning inherent in the primary comparative use has thus been extended to the mental domain.
At the same time, both Swedish sån and Norwegian sånn are used for hedging in their recognitional function. In (13) and (14), the speaker ‘asks’ for cooperation in denoting the specific referent when using sån/sånn to point at an imagined type of referent (Ekberg Reference Ekberg, Kern and Selting2011). Also when used as a quotative (see the former section), Norwegian sånn functions as a means to encourage the listener to imagine a type of event in order to understand or visualize the specific event (Opsahl and Svennevig Reference Opsahl, Svennevig, Enger, Faarlund and Vannebo2012; see also Johannessen Reference Johannessen, Enger, Faarlund and Vannebo2012). In contrast to the primary use, the immediate scope of sån/sånn NP includes only one entity, the comparee (the referent). However, we propose that the comparand is part of what Langacker (Reference Langacker1990) refers to as the maximal scope, since when using recognitional sån, the speaker is referring to an imagined comparand, necessary to identify the designated entity.
To conclude, the comparative meaning of sån/sånn is also present, although backgrounded, in its function as a determiner. In (18), for example, sån lite(n) bebis (‘sån little baby’) is construed as a type of baby known to the listener, although the referent is not previously mentioned and is presumably unknown. In other words, sån/sånn refers to an indefinite entity, belonging to a type of entity that is construed as known to the listener. The overall indefinite meaning thus comprises a definite element arising from the conception of (a known) type inherent in sån/sånn.
Interestingly, the use of German so with a determiner function again seems to parallel the use of sån/sånn in regard to this complex meaning. This holds both for the basic form so and for an extended form son- that derives from so combined with a cliticized indefinite article (and accordingly can undergo inflection). Hole and Klumpp (Reference Hole and Klumpp2000) described son as a new determiner in spoken, informal German that, parallel to the findings we report here for sån/sånn, simultaneously denotes definite type reference and indefinite token reference.
The comparative/deictic meaning of sån/sånn/so is hypothesized to be retained in the function as a determiner. In their lexical use, the items in question point to an (explicit or implicit) object of comparison; in their functional use, the act of pointing is extended to the level of discourse: sån/sånn/so pointing to an imagined type of referent needed to identify the indefinite instance of the type. The pointing is simultaneously a means to create joint attention of the referent that is being construed and denoted. When using sån/sånn/so, the speaker invites the listener to co-construct the referent (Ekberg Reference Ekberg, Kern and Selting2011). Consider again the example in (14), where the speaker introduces a new discourse referent (‘field.of.barley.programme’) via the mention of a type referent (‘programme’). The use of sånn as a determiner to introduce the adequate token referent – the second instance of sånn in (14) – is thus also an instance of hedging.
We hypothesize that the contextual coexistence of recognitional meaning and pragmatic hedging on the one side, and the introduction of a new discourse referent on the other (as in the examples in (13), (14) and (18)) triggers the reanalysis of a lexical (comparative/deictic) item into a grammatical item (see Traugott and Dasher Reference Traugott and Dasher2002). The grammatical determiner function is an invited inference arising from the regular association of the recognitional use of sån/sånn/so to introduce a new discourse referent.
In the previous section, we proposed that German so has developed from a comparative lexical item to a focus marker via its function as a hedging device. Thus, hedging may be the common trigger, or transitional phase, from a lexical comparative use to either a pragmatic or a grammatical (morphosyntactic) use. The cognitive motivation for the semantic change however lies in the inherent comparison and (type-)deixis of sån/sånn/so, since pointing to an imagined comparand is crucial for the type meaning that gives rise to the function of focus marker as well as determiner.
Conclusions
Taken together, the phenomena we discuss in this chapter point to striking parallels in the development of sån, sånn and so. From a similar lexical basis with a deictic comparative meaning, they undergo functionalizations leading to similar pragmatic and grammatical usages, namely that of focus markers and of elements with determiner functions, respectively, and these functional gains involve two similar patterns of loss in the semantic domain.
As pragmatic focus markers, sån, sånn and so undergo semantic bleaching that leads to a complete (in the case of so) or at least partial (for sån and sånn) loss of semantic content in favour of a pragmatic function that targets the level of information structure. As the linguistic evidence in the previous section showed, this bleaching reduces the deictic power present in the initial, lexical usage, whose semantic contribution can be paraphrased as ‘of this kind’, leading to a weaker, more general meaning ‘of a/some kind’, or even to a complete disposal of semantic content.
In the development of a grammatical item, sån, sånn and so undergo a different kind of semantic loss which, however, is again similar across languages: in this case, what we saw was not so much a reduction, but rather a backgrounding of deictic power, based on its relocation to the discourse level. This relocation allows the speaker to introduce a new referent, identifiable by the listener through shared knowledge, through an implicit or imagined comparand for the referent. This kind of ‘recognitional’ usage supports a grammatical usage of sån, sånn and so in the place of indefinite articles; that is, with a determiner function.
Figure 5.1 brings together these two different, and shared, paths for sån, sånn and so, from comparative deictics to functional elements at the pragmatic or grammatical level.
Figure 5.1. Two common paths of functionalization for Swedish sån, Norweigan sånn and German so (‘such (a)’).
These findings underline a view of developments in contemporary urban vernaculars not as mere simplifications, but as systematic patterns that result in elaborations, in this case in the rise of pragmatic and grammatical function words. While some salient linguistic phenomena from multiethnic speech communities might at first glance appear as unsystematic reductions such as ‘lack of articles’ or ‘overuse of filler elements’ – and are often perceived as such in the public discussion in particular – a closer look reveals systematic developments that build on an interaction of semantic loss and functional gain. As such, these developments allow us insights into possible routes from content words to function words and the semantic cuts that pave the way. This receives further support from the fact that, as the examples in this chapter have demonstrated, these developments are not restricted to multilingual speech communities. They might be more salient or, in some cases, more widespread in these contexts. However, they are not idiosyncratic to them, but can also be observed in other informal linguistic contexts outside multiethnic speech communities.
Participating in and building on general developmental tendencies of the respective majority languages, new urban vernaculars are embedded in the general linguistic domain they establish, by a network of ties in different grammatical areas. In our example, we found such ties, for instance, in the domain of quotative and hedging uses for sån, sånn and so, and also in the use of so as a focus marker, which has been described for informal spoken variants of German in general (Wiese Reference Wiese2009). This focus-marker usage seems to be more widespread and possibly more systematic in multilingual speech communities, but it is not restricted to them. As argued at the beginning of this chapter, we see this as an indication for the integration of such new urban vernaculars into the linguistic domain of the respective majority languages. From this perspective, we should expect a substantial amount of overlap with other, standard and non-standard, variants, rather than restricting our approach to these new vernaculars to features we consider outside the range of ‘autochthonous’ varieties.7
Urban vernaculars, with their multilingual speech communities and the special linguistic dynamics that they espouse, put the spotlight on new linguistic developments, thereby providing us with a unique opportunity to observe and investigate such phenomena. They are, however, no strangers to the linguistic landscape in which they emerge. Rather, we find in these vernaculars possible variations on a common theme of linguistic innovations that might follow the same patterns across languages, and might be found in other standard and non-standard varieties as well.
Transcription conventions
- capitalization
marks main sentence stress
- bold
marks elements that are emphasized in the analysis
- (.)
pause
- +
incomplete word
- (t)
standard orthographic addition
- ee
marks hesitation
- vpart
verb particle
- def
definiteness marker
- pl
plural
Background information on the speakers
Speakers in the Swedish examples (in alphabetical order, by pseudonym)
Bodil: monolingual Swedish; born in Sweden to parents of Swedish background
Gordana: bilingual Swedish/Bosnian; came to Sweden at the age of eight, parents of Bosnian background
Märta: monolingual Swedish; born in Sweden to parents of Swedish background
Sabaah: bilingual Swedish/Arabic; born in Sweden to parents of Syrian background
Speakers in the Norwegian examples (in alphabetical order, by pseudonym)
Anders: born in Norway; one parent of Norwegian, one of North African background
(claims to be a speaker of ‘the minority’s dialect’)
Aud-Jeanette: multilingual Norwegian/Tagalog/English; born in Norway, one parent of Norwegian background, one of Philippine background
Lukas: multilingual, born to parents of Ethiopian background
Olav: monolingual Norwegian; born in Norway to parents of Norwegian background
Roger: monolingual Norwegian; born in Norway to parents of Norwegian background
Waqar: multilingual Norwegian/Urdu/Punjabi/English; born in Norway to parents of Pakistani background
Speakers in the German examples (in alphabetical order, by pseudonym)
Angela: monolingual German; born in Germany to parents of German and Italian background
Ben: bilingual Thai/German; born in Germany to parents of Thai background
Cennet: bilingual Turkish/German; born in Germany to parents of Turkish background
Jessica: monolingual German; born in Germany to parents of German background
Merdan: multilingual Kurdish/Turkish/Arabic/German; born in Germany to parents of Kurdish/Turkish background
Nico: monolingual German; born in Germany to parents of German background
Stefanie: monolingual German; born in Germany to parents of German background
1 It should be noted that sånn is quite frequent in spoken Norwegian and probably not restricted to particular geographical areas. As the amount of spoken language corpora has increased, sånn has been the subject of several studies from both functional and grammatical perspectives over the last few years. These are not restricted to – and sometimes do not even include – multilingual speech communities (Hasund, Opsahl and Svennevig Reference Hasund, Opsahl, Svennevig, Buchstaller and van Alphen2012; Johannessen Reference Johannessen, Enger, Faarlund and Vannebo2012; Lie Reference Lie, Johannessen and Hagen2008).
2 That is, either the students themselves or at least one of their parents was born abroad.
3 A TV discussion of literature, from the DWDS corpus of spoken and written German. DWDS (Digitales Wörterbuch der Deutschen Sprache des 20. Jahrhunderts), ‘Digital Lexicon of German in the Twentieth Century’, Berlin-Brandenburg Academy of Science: Corpus Gesprochene Sprache ‘Spoken Language’, encompassing transcripts from the 20th century with approximately 2.5 million tokens. The speaker is Iris Radisch, a German literature journalist.
4 Accordingly, this neutral usage, without any negative semantic components, is what one finds for example in the DWDS (‘Digital Lexicon of German in the Twentieth Century’, Berlin-Brandenburg Academy of Sciences).
5 Solch (‘such’) is the counterpart of so that is used with plural nouns in standard German. (In informal varieties, in particular in northern Germany, one also finds a plural form sone- of so, where a plural suffix -e is attached to the form son- that initially derives from so with a cliticized indefinite article; see next section.)
6 A parallel case is when a demonstrative is used in the initial mention of a noun phrase, introducing a referent that is new in the current discourse (cf. Himmelmann Reference Himmelmann and Fox1996: 230).
7 The latter seems to be a perspective implied, for instance, by Auer (Reference Auer and Deppermann2013), who provides further data in support of Wiese’s (Reference Wiese2009, Reference Wiese2012) findings of focus marking so in German outside multiethnic Kiezdeutsch, and claims on this basis that ‘this usage of so is simply not kiezdeutsch’ (Auer Reference Auer and Deppermann2013: 29).
