John Benjamins Publishing Company
This is a contribution from Corpora, Grammar and Discourse. In honour of Susan Hunston. Edited by Nicholas Groom, Maggie Charles and Suganthi John. © 2015. John Benjamins Publishing Company This electronic file may not be altered in any way. The author(s) of this article is/are permitted to use this PDF file to generate printed copies to be used by way of offprints, for their personal use only. Permission is granted by the publishers to post this file on a closed server which is accessible to members (students and staff) only of the author’s/s’ institute, it is not permitted to post this PDF on the open internet. For any other use of this material prior written permission should be obtained from the publishers or through the Copyright Clearance Center (for USA: www.copyright.com). Please contact
[email protected] or consult our website: www.benjamins.com Tables of Contents, abstracts and guidelines are available at www.benjamins.com
chapter 9
General extenders in learner language Karin Aijmer
University of Gothenburg The aim of the paper is to study how general extenders are used by Swedish learners of English in comparison with native speakers. The study is based on a corpus of Swedish learners’ spoken English compiled within the international LINDSEI project. The analysis is both quantitative and qualitative. It is shown that the Swedish learners do not use general extenders in the same way as native speakers; in particular, they ’underuse’ and ’overuse’ certain forms and use fewer variants. The qualitative analysis shows that the function of general extenders is also linked to aspects of speaking fluently. Comparisons are also made with the use of general extenders by French, Dutch and German learners on the basis of other spoken learner corpora. The analysis shows that learners’ use of general extenders is affected both by the resources available in the native language and by the social norms and values regarding how the extenders should be used.
1. Introduction I don’t know I’ve always liked university and things or studying and so on and . been thinking about what to do and I can’t really think of anything it’s like teacher maybe but eh which would be interesting but I . like I’ve done . bit of substitute teaching for: grades seven to nine and it’s basically keeping order on the kids and not so much about . the actual subject or . whatever so and that’s what interests me I don’t really . like . sort of being an extra father for some teenagers so so then yeah obviously you have to go . a bit further and . I like the environment here and everything so . I wouldn’t mind staying on (LINDSEI- SW: Swedish component of the Louvain International Database of Spoken English Interlanguage)
DOI 10.1075/scl.73.10aij © 2015 John Benjamins Publishing Company
212 Karin Aijmer
‘Final tags’ such as and things, and everything, or whatever have interested scholars concerned with variation since Dines (1980) in a pioneering study drew attention to their formal and social variation. Dines (1980) used Labov’s model of variation analysis to show that and stuff like that and variants (or something like that, and things, etc.) could be regarded as a discourse variable whose realisations were ‘differentially distributed’ in the discourse community. Not surprisingly, utterance-final tags beginning with and and or have been studied from many different perspectives. There is, however, no consistent terminology. The term ‘general extender’ (which will be used here) has been more common in recent work to describe “phrase- or clause-final constructions which exhibit extensive variability” (Pichler & Levey 2011: 442).1 However, general extenders are ‘slippery’ since they can be realised in many different ways. They are multifunctional and are used differently depending on who the speakers are, whether they know each other or not, the discourse type, formality, etc. More recently, learner corpora have made it possible to study how non-native speakers use general extenders and to compare native and non-native speakers’ use of general extenders. The extract above, for example, illustrates how a Swedish learner uses general extenders. The data from learner corpora can show if learners overuse or underuse the extenders and if they use them like L1 users do. The aim of this paper is to study how general extenders are used by (Swedish) learners of English in comparison with native speakers. The analysis will be both quantitative and qualitative and will address the following questions: do native and non-native speakers use the same general extenders and with the same frequencies? Are there any differences in how they are used? We can expect that learners’ use of general extenders is affected both by the existence of similar resources in their native language and by cultural and social norms and values specific to that language. The extenders used by Swedish learners, for instance, are not the same as those used by learners with other mother tongues. We also need to distinguish between ‘the corpus as a whole’ and how individual learners use general extenders.
1. The term ‘general extender’ has for instance been used by Overstreet (1999) and by Cheshire (2007). In some recent work (Evison et al. 2007; O’Keeffe 2004, 2006) the general extenders are referred to as Vague Category Markers (VCMs). Other terms which have been used are ‘generalized list completer’ (Jefferson 1991), ‘extension particle’ (Dubois 1993), ‘setmarking tag’ (Dines 1980; Ward & Birner 1993). However these terms seem to identify the general extenders with a single function (cf. Overstreet 1999).
© 2015. John Benjamins Publishing Company All rights reserved
Chapter 9. General extenders in learner language 213
2. Previous work Tagliamonte and Denis (2010) describe general extenders as a unique opportunity to study social and linguistic influence on discourse pragmatic variation. They demonstrate how variability can be correlated with locality (Toronto) and with the age of the speaker. General extenders have also been studied in dialects (Pichler & Levey 2011), in several major varieties of English (Aijmer, 2013), and in adolescent language (Cheshire 2007; Stenström et al. 2002). The frequency and use of general extenders can also be related to the norms of the discourse type, for example whether it is accepted to be implicit only. Some interesting work on general extenders has been carried out by O’Keeffe (2004, 2006; cf. also Evison et al. 2007). The focus in these studies is on how general extenders (referred to as ‘vagueness markers’) are indexically associated with speakers and their identities, professional roles, and their relationship to the hearer. A key to their function is ‘underspecification’. General extenders should not be regarded as vague, uninformative and sloppy but rather represent “the creative forefront of language use and the collaborative use of meaning” (Evison et al. 2007: 142). Speakers construct an identity and a social relationship with the hearer by signalling that the hearer will be able to fill in the missing information. For example, in an informal conversation general extenders “draw on what is given and shared within the participation framework” (O’Keeffe 2004: 5). However, there are differences between discourse types. In an informal conversation a speaker who ‘says everything’ may be regarded as pedantic or a bore. On the other hand, explicitness is valued in a business transaction or in the courtroom. What is ‘given’ information also depends on who the speakers are and the type of discourse. In a corpus study of an Irish radio phone-in show, for example, O’Keeffe (2004) found that the general extenders projected a high degree of knowledge specific to the Irish community. The speakers used general extenders to refer to certain ‘common denominators’, for example, social practices and responsibilities, thereby identifying themselves as a socially-aware middle class group. The speakers in my study were university students who were interviewed by a native speaker. The interviewees were encouraged to talk on certain topics (see Section 3). As a result, what O’Keeffe (2006) refers to as the (socio-cultural) reference domain is to some extent pre-determined. However, while O’Keeffe (2004) emphasises ‘speaker-addressee interdependence’ and the ‘co-construction of meaning’, learners may use general extenders in new and different ways. They may, for instance, use general extenders because they cannot think of a word, because they are uncertain or because they think that a certain extender sounds English. © 2015. John Benjamins Publishing Company All rights reserved
214 Karin Aijmer
3. Data This study is made possible by the availability of learner corpora. In order to study Swedish learners, I have used the Swedish Component of the Louvain International Database of Spoken English Interlanguage Corpus (LINDSEI-SW) (Gilquin et al. 2010). The Swedish corpus consists of 50 interviews with a native speaker interviewer totalling 95,164 words. For the present study, I have only included the learners’ turns (= B turns), amounting to 67,822 words. The non-native speakers are advanced level, i.e. they have studied English for three or four years. The interview proceeds as follows. The interviewee is asked to talk on a set topic. This could be a film or play which the interviewee thought was good or bad or (most often) a country they visited which impressed them. A large part of the interview consists of free discussion. At the end of the interview, the interviewee is asked to describe a cartoon representing a young girl who has had her portrait painted by an artist but is not happy with the result. The Swedish learner data will be compared with data from a corpus of comparable interviews with adolescent native speakers compiled according to the same principles as the learner subcorpus: the Louvain Corpus of Native English Conversation (LOCNEC). The LOCNEC Corpus contains 125,666 words, or 71,853 words if only B-turns are counted. Differences in frequency between the groups can therefore be explained as being due to the type of speaker (native or nonnative). In addition, comparisons will be made with other spoken learner corpora compiled within the LINDSEI project. I have, for instance, made a comparison with the French, Dutch and German components of the LINDSEI Corpus. 4. General extenders and frameworks General extenders “are recognizable chunks of language” (O’Keeffe 2006: 130) or “recurrent sequences of words” (De Cock 2004: 226). They can be described as ‘collocational frameworks’ (Renouf & Sinclair 1991)2 which differ depending on whether they are introduced by and or by or. The head noun can be thing or stuff. Some main patterns of general extenders are presented in Figure 1; in patterns 3 and 4 we have a quantifier followed by a comparative (like that).
2. Renouf and Sinclair’s (1991) examples of collocational frameworks consist of pairings of grammatical words such as ‘a * of ’ where both the choice of word-class and the collocate need to be specified.
© 2015. John Benjamins Publishing Company All rights reserved
Chapter 9. General extenders in learner language 215 that this
sort kind
and
things
like
that
Pattern 3
and
all everything
like
that this
Pattern 4
or
something anything
like
that
Pattern 1
and
Pattern 2
of
thing stuff
Figure 1. Four patterns of general extenders
To some extent, general extenders are variable and flexible. Speakers can, for example, choose between and that (sort of thing) and and this (sort of thing). The connective can be deleted (Ø things like that). There is also variation between short forms (and things) and the longer forms to which they are related (and things like that). In addition to extenders which are flexible, there are extenders with a fixed form such as or so, or whatever, and so on, et cetera.
5. General frequencies Previous research suggests that general extenders are used less frequently overall by learners than they are by native speakers. This is in line with earlier observations that learners underuse pragmatic markers generally (Gilquin 2008). To take just one example, Hasselgren (2002: 118) found that “and things/everything/stuff that and like were virtually unused” by Norwegian 14–15 year-old learners of English. Advanced Swedish learners, however, do use general extenders, although not as frequently as native speakers. There are also differences depending on whether the extender is introduced by and or by or (see Tables 1 and 2). Table 1 shows the different patterns with and found in the native and non-native speaker corpora, while Table 2 does the same thing for or patterns.3 The method used to collect the examples of general extenders from the two corpora was to search for combinations with and and or and words they are known to frequently co-occur with. The paradigms include examples where there is no connective (and things like that>things like that).The Log Likelihood Calculator 〈http://ucrel.lancs.ac.uk/llwizard.html〉 was used to compare frequencies in the native and non-native speaker corpus.
3. Where normalised frequencies are referred to in Tables (‘norm.’) in this Chapter, these are per 100,000 words.
© 2015. John Benjamins Publishing Company All rights reserved
216 Karin Aijmer
Table 1. The distribution of and-extenders in the non-native and native speaker corpora LINDSEI-SW
LOCNEC
tokens
norm.
tokens
norm.
34
47
1
and stuff
38
56
2
and so on
17
25***
2
3***
3
and things
12
18***
52
72***
4
and things like that
12
18**
29
40**
5
and everything
12
18***
45
63***
6
and stuff like that
18
27
17
24
7
and all that
5
7
3
4
8
things like that
2
3***
9
and all
2
3
–
–
10
and (yeah) everything like that
1
1
–
–
11
and all this
1
1
2
3
12
and all these
1
1
–
–
13
Ø everything like that
1
1
–
–
14
and all the rest (of it)
1
1
3
4
15
and everything else
1
1
–
–
16
and eh whatever
1
1
–
–
17
and all that stuff
–
–
1
1
18
and all that kind of stuff
–
1
1
19
and all that kind of thing
–
3
4
20
and all things like that
–
1
1
21
and that sort of thing
–
2
3
23
32***
22
and all these things
–
1
1
23
Ø that sort of thing
–
8
11
24
et cetera
–
2
3
25
and places like that
–
4
6
26
and that kind of thing
–
9
13
Total
124
183***
242
337***
**indicates that the difference is significant at p