SEARCH

SEARCH BY CITATION

Keywords:

  • second language acquisition;
  • usage-based linguistics;
  • second language interaction;
  • longitudinal;
  • English negation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

This article explores the usage- and exemplar-based roots of second language (L2) negation construction learning. Based on two longitudinal case studies involving two adult L2 English learners and a corpus of 63 three-hour sessions of recorded classroom interactions, the study shows that L2 learning follows the predictions of usage-based models of language knowledge and acquisition, as the two participants’ learning of English negation constructions is found to go from recurring expressions toward an increasingly schematic, dynamic inventory of linguistic resources. Furthermore, exploring the evolution of two negation patterns in ongoing discourse, I suggest that local usage and long-term learning are inseparable and call for further detailed investigations of how locally contextualized interactions influence L2 development.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

Usage-based linguistics (UBL) views language learning as the emergence of constructions from “the collaboration of the memories of all of the utterances in a learner's entire history of language use and the frequency-biased abstraction of regularities within them” (N. Ellis & Larsen-Freeman, 2009, p. 92). This outlook on language learning builds on and empirically supports the view of language knowledge as a structured inventory of symbolic units, that is, form–meaning pairings (Langacker, 1987), which vary in size from morphemes to full expressions and range from the totally specific (e.g., words and fixed expressions) to the maximally general (schematic templates sanctioning the single instantiations). This in essence spells out the adherence to a noncompartmentalized view of language knowledge, implying that all linguistic units are psycholinguistically identical (Croft & Cruse, 2004) with fixed formulas and productive schematic templates simply occupying “opposite ends of the continuum of linguistic structures” (Achard, 2007, p. 1). Thus, among other things, UBL does away with the strict division between lexis and syntax and assumes that linguistic structure in the form of constructional form–meaning patterns emerges from language use as people engage in interaction in the real world (e.g., Kemmer & Barlow, 2000; Langacker, 1987, 2000; Tummers, Heylen, & Geeraerts, 2005).

A usage-based approach has been gaining rapid attention within second language acquisition (SLA) research (e.g., N. Ellis & Cadierno, 2009; N. Ellis & Larsen-Freeman, 2009; Eskildsen, 2009). Other current SLA views on second language (L2) development also share the frequency-biased and usage-driven view on language and language learning, stressing the dynamic and experientially adaptive nature of language knowledge (e.g., Beckner et al., 2009; Herdina & Jessner, 2002; Verspoor, Lowie, & de Bot, 2011). As shown in Verspoor et al. (2011), a dynamic and adaptive approach in fact offers an explanatory tool to the variability and nonlinearity in development attested in empirical findings and theoretical discussions within the SLA field dating back to the late 1970s (e.g., Berdan, 1996; Cancino, Rosansky, & Schumann, 1978; Huebner, 1985; Larsen-Freeman, 1991; Meisel, Clahsen, & Pienemann, 1981; Young, 1988).

But if, in accordance with UBL, usage events drive the developmental trajectories of language learners, then equally crucial to a full understanding of language learning must be to trace the interactional environments in which the learners’ constructions evolve. This insight is well illuminated by the idea that language is not only emergent, as UBL holds, but also locally contingent and situated (Hopper, 1998; Schegloff, Ochs, & Thompson, 1996). If so, to understand the situated nature of the emergent linguistic inventory, it is necessary to investigate the environmental affordances surrounding a learner during language use to see if the nature of the interactions may prompt specific usage events and hence learning of specific patterns at certain points in time. The position thus extends UBL and calls for the complementary use of conversation analytic (CA) tools (Eskildsen, 2011).

In the present study, I draw on UBL and CA to investigate L2 negation learning as a usage- and exemplar-based process of constructing a repertoire of interrelated linguistic patterns. The article is organized as follows. I first discuss the theoretical motivation for the study. I then present the results of a first set of analyses which demonstrate the emergence of the two learners’ repertoires of interrelated negation patterns and constructions along the UBL trajectory of learning. Next, a second set of analyses using CA allows me to explore the relationship between two demonstrated developmental trajectories in the longitudinal corpus and the interactional environments (usage events) in which they evolved for one of the focal participants. This second part of the analyses constitutes a move toward a locally contextualized view of L2 learning whose foundation was developed in earlier work (Eskildsen, 2009, 2011). I conclude that UBL holds great promise to illuminate L2 development in ways that previous approaches are not equipped to do. Further, I contend that the fabric of the contexts in which linguistic development takes place is profoundly interwoven with that development and deserves minute attention. I therefore call for future investigations of the interactional contingencies of emergent L2 constructions.

Usage, Constructions, and Language Learning

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

The UBL notion of an item-based grammar that evolves over time in terms of increasing generalizability stems from MacWhinney (1975). Studying two Hungarian children learning their first language (L1), MacWhinney found that their early competence was based on a finite set of item-based patterns (translated into English, they include more+ X, X +too, and see+ X). These patterns are characterized by a recurring lexical item and an open slot for the insertion of semantically and structurally sanctioned items. In total, 42 such patterns accounted for 85–100% of a total of 11,077 utterances, and one central finding was that children's developing grammars are much more concrete than previously thought. In recent longitudinal research the focus has been on investigating how the continuum of schematicity at the heart of the linguistic inventory comes into being on the basis of concrete recurring linguistic material in use. As such, UBL's learning trajectory has been empirically validated in a number of child language studies (e.g., Dąbrowska & Lieven, 2005; Lieven, Salomo, & Tomasello, 2009; Tomasello 1992, 2003).

In terms of language learning, the continuum of schematicity proposed by UBL approaches is operationalized as an exemplar-based developmental trajectory from concrete constructional instantiations (such as fixed recurring multi-word expressions, e.g., I dunno) via partially schematic, partially concrete patterns known, among other terms, as utterance schemas (Tomasello, 2000) or item-based patterns (MacWhinney, 1975) (e.g., I don't Verb)1 to increasingly generalizable schematic constructions (e.g., NP do NEG VERB) based on systematic commonalities among patterns (e.g., N. Ellis, 2002; Tomasello, 2000). These commonalities are derived by language users in and through social interaction, as they encounter and learn a sufficient number of similar but different instantiations of a given pattern to be able to analogically form abstractions of what is generic to the pattern. This in turn enables language users to use, understand, and learn novel expressions of the same kind. UBL's exemplar-based path of language learning, then, denotes an incremental pattern-based process in which concrete patterns link to previously experienced concrete patterns. In time, with enough exemplars of a given construction in place, this results in the emergence of schematic language knowledge.

The ontological status of constructions in L2 learning has been empirically supported within SLA research (e.g., Bartning & Hammarberg, 2007; Collins & N. Ellis, 2009; Goldberg & Casenhiser, 2008; Gries & Wulff, 2005; Waara, 2004), and recent investigations (e.g., N. Ellis & Ferreira-Junior, 2009; Eskildsen, 2011; Mellow, 2006) have documented and discussed L2 learning in terms of an exemplar-based process where the gradual abstraction of common regularities link concrete expressions as schematic constructions.

However, the learning trajectory is not necessarily linear, because linguistic knowledge is constantly under construction and in flux as usage environments change (Hopper, 1998). In these changing environments, type and token frequencies are thought to determine matters of psycholinguistic entrenchment and schematicity of recurring expressions and constructions. Token frequency is the frequency of a concrete, specific expression; the higher the token frequency the more likely the specific expression is to become entrenched in the language learner's experience. Type frequency, referring to the number of different instantiations of the same construction, is key in determining the schematicity of that construction (e.g., Tomasello, 2003). However, entrenched high-frequency expressions are not replaced over time by emergent schematized constructional knowledge; rather, abstract constructions and their specific instantiations may cohabitate, in other words, coexist psycholinguistically, in individual linguistic inventories at multiple levels of schematicity (e.g., Achard, 2007; Langacker, 2000).

Emergence in Local Usage

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

In earlier work (Eskildsen, 2009, 2011), I have outlined a move toward a locally contextualized view of L2 learning, based on the tenet that language is not only emergent but emergence itself is locally contingent and situated (Hopper, 1998; Schegloff et al., 1996). The position advocates the combined use of UBL and CA (e.g., Kasper & Wagner, 2011; Schegloff, 2007) in a two-step analytical process. In the first step, I analyze the learning trajectory as a process of expanding a repertoire of interrelated linguistic constructions. This is documented through token and type frequency counts at several points in time. As will become clear, however, a deeper appreciation of the data often requires qualitative explanations which the quantitative counts cannot provide, as the constructions counted are, to varying degrees, contingent on interactional environments. To get to the core of the relation between such environments and the emergent constructions, the second analytical step makes use of CA and concepts such as action sequences, adjacency pairs, repair-initiation, and co-participants’ orientations, the focus being on the interactional job accomplished by the learner through employing the constructions in question (Eskildsen, 2011). The two-pronged approach then allows me to provide insights into the situated nature of the emergent linguistic inventory, as I demonstrate how the nature of the interactions prompts the use of specific linguistic resources and hence the learning of specific patterns over time.

Formulaic Language and Developmental Stages of English Negation

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

Two other branches of research are relevant to the present study and must be acknowledged. As a natural consequence of the exemplar-based starting point of UBL's trajectory of learning, the present study is related to SLA research concerned with formulaic language. I have argued elsewhere (Eskildsen, 2009; Eskildsen & Cadierno, 2007) that much research into the role of formulaic language in SLA differs on a fundamental level from the tenets of UBL because the former has tended to maintain the syntax-lexis distinction. Nonetheless, my approach finds kinship in empirical studies which have shown that formulaic language is developmentally significant, linked with and feeding into the rest of the linguistic system in learning (e.g., Nattinger & DeCarrico, 1992; Weinert, 1995). I argue that the UBL model, with its assumption that formulas and more creative patterns occupy opposite ends of a continuum of psycholinguistically identical structures, is more apt at capturing the dynamic interplay between such formulas and more creative patterns than an approach which makes a fundamental distinction between them a priori.

Another branch of research to which the present work is related deals with the L2 learning of English negation, a prolific empirical domain during the early years of the field which spawned seminal work on the oft-cited developmental sequences (e.g., Cancino et al., 1978; Hyltenstam, 1977; Schumann, 1979; Stauble, 1978; Stauble & Schumann, 1983). This body of research is too large to be discussed in detail here; suffice it to say that it examined stages of syntactic development in terms of a developmental sequence consisting of four stages defined by dominant structures used by L2 learners over time. In the order in which they were found to emerge in learning, these structures are: external negation, internal negation, auxiliary negation, and analyzed do-negation. Even though these early studies did not always agree on the details, the sequence itself became widely accepted in general terms as outlined in Schumann (1979), and since then it has been repeatedly reviewed and confirmed in SLA textbooks (e.g., R. Ellis, 1994; Tarone & Swierzbin, 2009) and handbooks (e.g., Doughty & Long, 2003; Pica, 2005). While the current study draws different developmental conclusions, it is influenced by the terminology of the sequence and will elaborate it when relevant in later sections.

The Present Longitudinal Learner Corpus

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

The data source for the present study is the Multimedia Adult English Learner Corpus (MAELC),2 which consists of audio-visual recordings of classroom interaction in an English as a Second Language (ESL) context in the United States. The classrooms were equipped with video cameras and students were given wireless microphones on a rotational basis; the teacher wore a microphone at all times (Reder, Harris, & Setzler, 2003). The present research is based on data from two adult Mexican Spanish-speaking students, Carlos and Valerio (pseudonyms). Carlos had been in the United States for 21 months, and Valerio for 9 months, before joining this ESL program, and both students progressed successfully through the four levels assigned to their classes by Portland Community College (PCC). According to PCC, the levels span from beginner (SPL 0–2) to high intermediate (SPL 4–6) (Reder, 2005). I have previously examined these and similar data in order to report on our ongoing investigations into the exemplar-based nature of L2 learning with a special focus on formulaic language (Eskildsen, 2009; Eskildsen & Cadierno, 2007). In Eskildsen (2011) a thorough analysis of the situated nature of formulaic language use and the locally contextualized aspects of L2 learning revealed a fundamental coupling between interactional environments and L2 construction learning. The present analyses seek to advance this understanding of L2 learning as rooted in specific constructional exemplars as well as interactional experience.

The database for this inquiry consists of transcripts from the sessions in which Carlos and Valerio are either wearing a microphone or sitting next to someone wearing a microphone. The database differs slightly from that of the research reported on in Eskildsen and Cadierno (2007) and Eskildsen (2009) due to the addition of new transcripts and a methodological improvement; in the present paper, uses that are afforded (van Lier, 2000), for example through language learning tasks in the classroom that provide interactionally relevant items which learners pick up and recycle in their own utterances, have been included because of the Vygotskian insight, to be further explored later, that the learning of a new resource involves several socially mediated encounters with it (Lantolf & Thorne, 2006). For language learning, this implies that affordances may play an important role in the process of learning new linguistic material in a slow and piecemeal fashion.

Following the same terms as the classes, the database has been divided into five recording periods for each student, as outlined in Table 1. Note that neither Carlos nor Valerio attended school consistently throughout their recording periods, hence the gaps; the corpus spans 2 years for Valerio and 31/2 years for Carlos. Throughout the presentation of results, the abbreviation RP is used to refer to recording period.

Table 1.  Timeline for Recording Periods (RP) for the Two Adult ESL Students
 ValerioCarlos
RP1July 1–August15, 2003September 27–November 29, 2001
RP2Sept. 23–November 21, 2003January 18–June 7, 2002
RP3January 27–April 9, 2004Sept. 23, 2003–March 2, 2004
RP4May 4–July 27 2004September 30, 2004–December 2, 2004
RP5June 30–August 4, 2005January 11–February 22, 2005

Results of UBL Analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

Carlos

Carlos's inventory of negated constructions consists of the following assortment of negation patterns:

  • (a) 
    Subject don't verb (e.g., I don't think so, I don't know)
    • i. 
      Short answer form, no I don’t
  • (b) 
    subject no verb (e.g., I no remember)
  • (c) 
    copula negation (e.g., it's no correct)
  • (d) 
    no-phrases (e.g., no problem)
    • i. 
      One negative imperative, don't worry, evolved from no worry
  • (e) 
    not-phrases (e.g., not too much, why not?)
  • (f) 
    can’t patterns (e.g., I can't come)

The categories are data-driven and not posited a priori, and their learning trajectories will be dealt with in turn. Starting with the constructions listed under a) and b), subject don't / no verb, this section discusses Carlos's learning trajectory as characterized by (1) an initial high frequency of a recurring exemplar of the target construction (I don't know); (2) a gradual increase in the use of other exemplars of the targetlike pattern, subject don't verb; and (3) a concomitant gradual waning of the nontargetlike pattern, subject no verb. Figure 1, adopted and revised from Eskildsen and Cadierno (2007), shows this development in terms of relative use of tokens of the three patterns over time. As to the nature of the processes involved in the gradual shift from nontargetlike to targetlike, my position is that it is experiential; the more exemplars of the targetlike construction Carlos encounters and puts to use himself, the more likely he is to internalize these and abstract their structural regularities as common constructional denominator.

image

Figure 1. Distribution of subject don't /no verb patterns over time, Carlos.

Download figure to PowerPoint

As the evidence is presented in this section, it will be useful to keep in mind that in UBL models the gradual entrenchment of the targetlike pattern is linked with the emergence of increasing schematicity of linguistic representation, which can be illustrated by way of type-token relationships. Tokens are the number of specific instantiations of the pattern under investigation, whereas types are the number of different instantiations of the same pattern. A high token frequency and a low type frequency results in a low type-token ratio and implies the possible entrenchment of one or a few specific exemplars. By contrast, the opposite scenario with a high type-token ratio implies a high degree of productivity of the pattern, indicating possible entrenchment of more schematic underlying knowledge.

Table 2 displays the type and token frequencies for Carlos's subject don't / no verb-patterns over time, showing the individual developmental trajectory of each of the two patterns.

Table 2.  Type and Token Frequencies for Carlos's subject don't / no verb Constructions
 Targetlike, subject don't verbNontargetlike, subject no verb
TokensTypesRatioMWEaTokensTypesRatio
  1. Note. RP = recording period; MWE = the fixed multi-word expression I don't know.

  2. ain tokens (and ratios).

RP11330.2310 (0.77)18130.72
RP22470.2913 (0.56)15110.73
RP340130.3321 (0.53)760.86
RP41890.50 6 (0.33)000.00
RP523110.4511 (0.48)000.00

The evidence in Table 2 shows an increase in productivity as indicated by the type-token ratio increasing from 0.23 to around 0.50 for the targetlike pattern. At the same time, the role played by the fixed multi-word expression (MWE) I don't know to sustain the targetlike pattern drops from a ratio of 0.77 to 0.33 in RP4 and goes back up again to 0.48 in RP5. The data also reveal a relatively high type-token ratio for the nontargetlike pattern, suggesting a generally higher degree of productivity for this pattern than for the targetlike variety. This might indicate that the targetlike pattern is generally more dependent on recurring expressions (not only the MWE I don't know; see Eskildsen & Cadierno, 2007, for more details) and that the nontargetlike variety, when in use, holds the position as the default structure for producing subject don't/no verb-patterns.

Carlos's data support the view of L2 negation development as exemplar-based, with initial development being crucially dependent on the MWE I don't know. The pattern expands on the basis of other recurring expressions and utterance schemas (I don't verb, you don't verb) and evolves in the direction of a more schematic representation (i.e., SUBJ AUX-dopresent/past tense NEG VERB) as Carlos uses an increasingly wider array of personal pronouns in subject position and ultimately conjugates the auxiliary for tense. Carlos's ability to conjugate for tense corresponds to what is traditionally described as using analyzed forms of don't (e.g., Schumann, 1979).

In addition to the uses displayed in Table 2, Carlos has two further expressions, no I don’t and don't worry, which were listed separately in the inventory overview as (a.i) and (d.i). The former is found at various points in specific activities targeted at practicing short answer forms, whereas the latter is found in one interaction only at the very end of Carlos's RP1. They constitute a taxonomical problem because they are not bona fide subject don't verb-patterns. Don't worry, in fact, seems more likely to have been derived from the no-phrases, to be discussed in due course, as it is preceded by a use of no worry. This development, however, from no worry to don't worry at the end of RP1 may be a reflection of an incipient broadening of the applicability of don’t as brought about by the gradual expansion of subject don't verb-patterns.

Whereas don't worry does not seem to undergo a process of generalization over time, that is, Carlos never uses other main verbs in a more productive negative imperative construction, no I don’t does exhibit some expansion potential, as RP4 sees one spontaneous use of no I didn’t.3 While don't worry, then, is likely to have developed from no worry, no I don’t, used 5 times in RP1, albeit only in practice situations where it is the targeted form, may have played a role in Carlos's general do-negation learning.

Post-negated copula negations, listed under c), are in use from the beginning. Thus, Carlos's development of copula negation does not immediately correspond with previous findings for English L2 users with Spanish L1, where the pre-negated copula pattern is often cited as a common structure for Spanish speaking learners of English L2 (e.g., Cancino et al., 1978; Schumann, 1979).While Carlos does use pre-negated copula four times in RP1 and once again, somewhat surprisingly perhaps, in RP4, the targetlike post-negated construction is much more frequent, and Carlos's initial use of the two patterns does not support the notion of a linear development from pre-negation to post-negation of copula. Furthermore, as indicated by the four-stage developmental sequence posited by the early research of the 1970s, as mentioned earlier, targetlike copula negation is usually reported to emerge at advanced stages around the same time as modal negation, but in Carlos's case targetlike copula negation is present at initial stages of development and emerges earlier than modal verbs.

In concord with the early research on developmental sequences, Carlos's first (in fact, his only)4 negated modal is can, listed under f) in his inventory, which he begins to use spontaneously in a targetlike fashion in RP3. However, contrary to previous findings, Carlos's targetlike use is not preceded by a period of non-target-like pre-negated can-patterns. He only has one such use, I no can, in RP1, which has been coded here as a subject no verb pattern, and thus included in Table 1, as Carlos seems to treat can as a main verb, making the construction resemble the other subject no verb-patterns. The two uses of can-negation in RP1, listed in Table 3 below, are I cannot, which Carlos repeats following the teacher's correction of his utterance I no can.

Table 3.  Copula negation and can-negation, Carlos
 Copula negation (targetlike)Copula negation (nontargetlike)Can-negation (targetlike)
TokensTypesRatioTokensTypesRatioTokensTypesRatio
RP1650.83441.0021
RP2331.00000.00000.00
RP3980.89000.00760.86
RP4540.8011320.67
RP511000.00000.00

Displaying the type and token frequencies for copula negation and can-negation, Table 3 shows that Carlos uses targetlike copula negation from the beginning. Table 3 even suggests that he does so at a very high level of productivity. However, a look beyond the numbers reveals that the learning trajectory of the copula negation seems to hinge on specific recurring patterns, namely she's no(t) here, is no(t) x and I am not (x). In RPs 1 and 2, Carlos's total use consists of she's no(t) here (twice), I’m not (an affordance), I am not finished, is no correct, and is no good. The short answer form, I'm not, and its related uses are often practiced in the classroom, as was the case with no I don't. They are considered full copula negations here, but instances where Carlos repeats the teacher's talk or reads out loud the teacher's writing on the whiteboard have not been included in the counts. Carlos's use of the targetlike copula negation from the beginning, however, may in part be due to the practices of the classroom. In RP3, the copula negation system is undergoing further change as evidenced by idiosyncratic uses of the I’m not (x)-pattern (I’m not change anything and no I aren’t), the development of is no(t) x to it's not x (e.g., it's not easy and it's not too hard) and the subsequent expansion of the pattern to include other subjects (e.g., the smoke is no good) as well as the addition of a new that's not x-pattern (e.g., that's not so and that's not true). Furthermore, the short answer form no I’m not and the it is not-pattern conflate to bring about a new form, I think it's not, and in RP5, a use of this note is not correct suggests that Carlos's copula negation pattern is becoming ever more productive.

The patterns listed as no/not-phrases under d) and e) in Carlos's inventory of negated constructions cover a variety of utterances whose basic common denominator is the use of no or not as negation particle. The negation particle is typically placed immediately before that which it negates, either utterance-initially (e.g., no Spanish) or utterance internally (e.g., I have no children). This category thus differs from what is traditionally known as external negations, that is, phrases where the negation particle is outside the utterance nucleus, often exemplified in the literature by no the sun shining, and originally identified as a particular feature of the first stage of the developmental sequence for negation learning in child language studies (Klima & Bellugi, 1966). Even though subsequent developmental SLA research did not consistently yield the same result (Schumann, 1979), the stage-defining status of the external negation has been upheld, albeit with the caveat that not all learners necessarily spend the same amount of time on each developmental stage (e.g., Doughty, 2003; R. Ellis, 1994). Yet, during his time in class, spanning an observation period of 31/2 years, Carlos never uses external negations.

In order to gain a comprehensive view of the no/not-phrases it is necessary to view them in terms of their constructional composition, that is, as either no-phrase or not-phrase, but also in terms of the extent to which they can be seen as targetlike, because the existence of, for example, no problem and no more shows that no-phrases cannot be a priori dismissed as faulty patterns. Table 4 shows that Carlos uses no twice as frequently as not from RP1 through RP4. It also indicates that he does so with a high degree of productivity, as evidenced by the consistently high type-token ratio. According to the table, the type-token ratio for not-phrases is also high, in fact maximal, in RP1, suggesting that Carlos also uses not freely as a negation particle from the beginning. However, a closer look at the data reveals that the numbers are misguiding, as Carlos's uses in RP1 (not write, not small, not here, and not this time) are all afforded by the immediate interactional environments; that is, he picks up these items from prior turns and recycles them in his own utterances. In fact, the first use of not in RP2 (but not very well) is also afforded, as the teacher has instructed the students to practice asking do you speak English? and answering yes I do but not very well. That is not to say that ultimately Carlos could not have produced them outside of these environments, but the empirical fact is that he does not.

Table 4. No/not-phrases, Carlos
 No-phrasesNot-phrasesNo-phrases, nontargetlikeNo(t)-phrases, targetlike
TokensTypesRatioTokensTypesRatioTokensTypesRatioTokensTypesRatio
RP112100.83441.001080.80661.00
RP213100.77630.50771.001260.50
RP310101.00520.40771.00850.63
RP4441.00221.0011551.00
RP5000.0011000.0011

Table 4 further indicates that Carlos's use of not is exemplar-based, as reflected in the lower type-token ratio in RPs 2 and 3. The recurring MWEs behind these numbers are why not? and not too much. Whenever he needs to negate more freely, however, he resorts to no, which most often results in nontargetlike expressions such as no married or no here but also in targetlike ones such as no children and no money. Nontargetlike uses seem to disappear toward the end of the data collection period with only one token in the two final RPs, which corroborates previous research (e.g., Stauble, 1978). Carlos's usage suggests that no+ phrase functioned initially as the default generic solution for negating utterances when other negation resources were not applicable because of their pattern-based nature (do-negation, copula negation and can-negation) or not available due to their interactionally contingent and exemplar-based nature, as mentioned in the cases of the afforded expressions and recycled MWEs above.

Valerio

Comparing Valerio's inventory of negated structures with Carlos's, similarities and differences appear. An outline of Valerio's negations takes the following shape:

  • (a) 
    subject don't verb (e.g., I don't think so, I don't know)
    • i. 
      short answer form, no I don’t
    • ii. 
      negative imperative, don't worry
  • (b) 
    subject no verb (e.g., I no remember)
  • (c) 
    copula negation, post-negated targetlike structure (e.g., it's no correct)
  • (d) 
    internal pre-negated copula, nontargetlike (e.g., this no is always)
  • (e) 
    external pre-negated copula, nontargetlike (e.g., no is melon)
  • (f) 
    idiosyncratic subject copula no verb-patterns (e.g., I am no say)
  • (g) 
    no-phrases (e.g., no problem, no always)
  • (h) 
    not-phrases (e.g., not yet)
  • (i) 
    can’t-patterns (e.g., I can’t)

Starting with patterns (a) and (b), subject don't / no verb, the most striking similarity between the two students is the co-presence of the MWE, I don't know, other instances of the targetlike do-negation pattern, and subject no verb. Figure 2 displays an overview of subject don't / no verb in Valerio's data, showing how their relative usage frequencies develop over time. Some developmental tendencies seem identical in Carlos and Valerio. Specifically, the MWE is frequent and retained throughout development (though decreasing from 75% to 22% of all do-negation uses in Valerio's case, whereas it was fluctuating in Carlos's case), and targetlike pattern usage increases over time.

image

Figure 2. Distribution of subject don't /no verb patterns over time, Valerio.

Download figure to PowerPoint

However, there are also individual differences. In Valerio's case the nontargetlike pattern plays a less significant role initially, constituting 8% (1 of 12) of all subject don't / no verb-patterns, but increases to 18% (3 of 17) in RP2 and 44% (24 of 54) in RP3. It lingers at roughly the same level in RP4 and finally drops to the brink of disappearance in RP5. Meanwhile, the targetlike pattern (excluding the MWE) is used relatively frequently in RPs 1 and 2, but drops to its lowest point in RP3, where the subject no verb-pattern is gaining a stronger foothold. In RPs 4 and 5, targetlike pattern usage increases again to reach a level of 69% relative use in RP5.

Displaying the type-token ratios for Valerio's subject don't / no verb, Table 5 further sustains the rather surprising finding that the targetlike don’t-pattern is initially much more frequent than the nontargetlike variety, and that the nontargetlike variety undergoes quite a development in RPs 3 and 4. The interactional environments in which this development takes place will be investigated in the second set of analyses reported later.

Table 5.  Type and Token Frequencies for Valerio's subject don't / no verb-Constructions
 Targetlike, subject don't verbNontargetlike, subject no verb
TokensTypesRatioMWEaTokensTypesRatio
  1. Note. RP = recording period; MWE = the fixed multi-word expression I don't know.

  2. ain tokens (and ratios).

RP11130.27 8 (0.73)11
RP21440.2910 (0.77)331.00
RP33080.2721 (0.70)24120.50
RP457180.3223 (0.40)34250.74
RP533120.36 8 (0.24)320.67

Table 5 shows the initial exemplar-dependence of subject don't verb, reflected in the frequent use of the MWE, especially in the first three recording periods where its ratio ranges from 70% to 77%, and in the relatively low type-token ratios which reaches a maximum of 0.36 in RP5. The nontargetlike variety, while much more productive, as displayed in the higher type-token ratios, also exhibits exemplar-based behaviour in learning, as will be shown in the second set of analyses later.

As was the case with Carlos, Valerio also uses don't worry as his only negative imperative,5 and is also exposed to the short answer form, no I don’t, including derived conjugations for person and tense, through language learning tasks in the classroom. For Valerio, don't worry emerges in RP4 and may thus be a result of a general expansion of the use of don’t. As to no I don’t, only the past tense variety seems to have any impact on Valerio's learning trajectory; his first uses of the past tense of do-negation, instantiated as I didn't see you and you didn't coming in RP4, occur only minutes after a situation in which the students have solved a task using no I didn’t. Valerio never uses this past tense form again, so whether these uses constitute the seed of the learning of this construction remains unanswered. I have listed both don't worry and no I don’t as subcategories under the subject don't verb-category in Valerio's inventory to indicate that although they are probably somehow related to this category they are not central members of it.

The next step in the investigation concerns copula negation, or items c) through f) in Valerio's inventory. Valerio employs a number of distinct copula negation patterns of varying frequency:

  • I. 
    targetlike copula negation (e.g., it's no correct)
  • II. 
    pre-negated copula constructions
    • i. 
      internal, with a subject (e.g., this no is always)
    • ii. 
      external, without a subject (e.g., no is melon)
  • III. 
    idiosyncratic subject copula no verb-patterns (e.g., I am no say)

In contrast to Carlos, Valerio thus uses two kinds of pre-negated copula constructions. The targetlike constructions and the pre-negated constructions without subject, although varying in frequency, are in use throughout, whereas pre-negated constructions containing a subject go in and out of use. The idiosyncratic pattern is used three times only in RP1 and RP3. Table 6 displays the distribution of the copula negation patterns.

Table 6.  Type and Token Frequencies for Valerio's Copula Negation
 Targetlike, post-negatedInternal, pre-negatedExternal, pre-negatedIdiosyncratic pattern
TokensTypesRatioTokensTypesRatioTokensTypesRatioTokensTypesRatio
  1. Note. RP = recording period.

RP11111760.8611
RP2420.5000221.0000
RP3420.50331.00661.00221.00
RP41080.80661.00771.0000
RP51100850.6300

The most noticeable aspect of the table concerns the variation in pattern dominance. The external pre-negation dominates in RP1 and RP5, but in RP2 through RP4 none of the patterns can be said to be dominant. This showcases the empirical fact of nonlinearity in L2 development; there is no straightforward development from no is to is no—just as the fairly frequent deployment of the no is-pattern in RP5 suggests that it is not developmentally related to the other no verb structures which were found to disappear by this time, as previously discussed. The use and hence learning of no as negation particle thus seems to be a pattern-specific matter.

Table 6 further suggests that the two nontargetlike patterns are less dependent on recurring expressions than the targetlike one as indicated by their higher type-token ratios. They are all, nonetheless, deeply rooted in recycled linguistic material. The external pre-negation is only instantiated as no is x, and the internal pre-negation comes out as subject no is x in all but the very last example caught on tape, which is I no am sleeping. In other words, both patterns exhibit item-based behaviour. The targetlike construction, on the other hand, displays a different developmental trajectory. The one instantiation in RP1 is it's no different, which Valerio picks up following the teacher's correction of his utterance no is different. Interestingly, and perhaps related to Valerio's idiosyncratic subject copula no verb-pattern, the next two uses are I am no, which Valerio seems to use in places where the short answer form I don’t would be more appropriate. The two final uses in RP2 are I’m not (quite) sure, which are among the linguistic resources introduced by the teacher to solve a specific task at hand; they are thus afforded by the environment. In addition to later repeated uses of I’m not sure, the pattern subsequently evolves into appropriate uses of I am no/I’m not, you’re not x, and this is no(t) x. Also, the data further suggest an incipient emergence of not verb-ing, instantiated by they’re not working and I’m not working in RPs 4 and 5.

Another interesting development that needs to be mentioned is the emergence of the subject no is-pattern in RP3 and RP4. At this point, Valerio is also using a targetlike copula negation of a similar kind, namely this is no(t), but the subject in the targetlike copula negation is only realized as this, whereas in subject no is it is instantiated by a wider array of subjects (food, she, this, this one, the batteries, and the lights). The nontargetlike variety, then, is again more productive than the more fixed targetlike variety.

Summing up the findings for Valerio's copula negation data, it seems safe to say that there is no developmental path from no is to is no. Instead, the two patterns are competing (in the semitechnical sense suggested, e.g., by R. Ellis, 1994; Schumann, 1979) throughout development; in particular, they seem to be competing for the domain of making references to people and objects in the world, as evidenced by the coexistence in Valerio's grammar of this is no(t) x and subject no is x. The targetlike variety is initially dependent on affordances but seems to be evolving into a productive schema, whereas the nontargetlike variety seems to become more entrenched as an item-based no is-pattern, perhaps as a result of L1 transfer.

Moving on to items (g), (h), and (i) in his inventory, Valerio further employs a host of target- and nontargetlike no/not-phrases as well as targetlike can-negation. The latter is found only in RP5, preceded by three uses of subject no can verb and one use of subject don't can verb in RP4. These were treated as subject no verb under the do-negation header above because at the time of their use they seemed to be more closely related to this pattern than to the yet-to-emerge targetlike can-negation. There are six such targetlike uses in RP5, namely I can’t (4), I can't live, and I can't work. The incipient emergence of targetlike can-negation thus also seems to exhibit item-based behaviour.

The distribution of no/not-phrases, however, is a more complex issue. As was the case with Carlos, it is necessary to view them in terms of both structural composition and targetlikeness. Table 7 presents this overview.

Table 7.  Type and Token Frequencies for Valerio's no/not-Phrases
 No-phrasesNot-phrasesNo-phrases, nontargetlikeNo(t)-phrases, targetlike
TokensTypesRatioTokensTypesRatioTokensTypesRatioTokensTypesRatio
  1. Note. RP = recording period.

RP1771.00000.00771.00000.00
RP21390.69000.00661.00730.43
RP314110.79210.501190.82540.80
RP415110.73551.00870.831280.67
RP5331.00331.00221.00441.00

It is very clear that Valerio has a propensity toward using no over not. Initially, he uses no not-phrases at all and, moreover, all his no-phrases are nontargetlike. The type-token ratios for no-phrases are also very high, indicating a high level of productivity throughout. For the targetlike no-phrases, on the other hand, the type-token ratios suggest a more exemplar-dependent learning trajectory. These exemplars include no problem and no more. When Valerio starts using not in RP3, the first uses are afforded, as he picks up not really (though mispronouncing it as not real) from the interactional environment. Another affordance is not yet in RP4, but all remaining uses seem to be Valerio's own creations, making for a quite productive not-inventory. Similar to Carlos's data as well as findings from the early research (e.g., Stauble, 1978), Valerio's use of nontargetlike no-phrases also seems to wane over time, giving leeway for targetlike uses of both no and not. An epitome of this is the development of why no to why not in Valerio's inventory.

One final empirical observation brought about by the UBL analyses that deserves to be mentioned is that initial negation learning seems to be pattern specific. The data for Carlos and Valerio suggest that the two learners do not operate on the basis of an early general no+ verb/phrase rule, contrary to what is traditionally postulated (e.g., Stauble & Schumann, 1983). Instead, this is a much more dynamic issue, which is both pattern- and learner-specific. Looking specifically at the two students’ first recording periods, it is clear that Carlos draws frequently on subject no verb rather than subject don't verb (18 of 31 tokens, 58.3%), but less often pre-negates copula (4 of 10 tokens, 40%). For the various negated no/not-phrases, Carlos also showed a tendency to initially prefer no, while uses of not were found to be dependent on affordances in RP1. Valerio shows opposite patterns; he uses subject don't verb predominantly (11 of 12 tokens, 91.7%) and has only one instance of subject no verb, but draws heavily on no copula (8 of 9 tokens, 88.9%) and no-phrases (7 of 7 tokens, 100%)—all within RP1.

Toward a Locally Contextualized View of the Emergence of Constructions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

The findings presented here thus far and those in previous usage-based research suggest that language knowledge is concrete, pattern-based, probabilistic, and emergent. Therefore, any a priori systemic and static conception of language will not do to describe and analyze the dynamics of L2 ontogenesis—a point already made by Cancino et al. (1978) and more recently by Larsen-Freeman (2006) and de Bot, Lowie, and Verspoor (2007), among others. Given the concrete nature of the pattern-based view of language knowledge as espoused by UBL, I would like to go further and explore the implications for L2 learning of the idea that language is not only emergent but also, to some extent, locally contingent and situated (e.g., Schegloff et al., 1996). It follows from this that any distributions of constructions uncovered in the two students’ emergent inventories can and should be investigated in more detail in terms of the interactional environments and affordances surrounding the students. This investigative step, drawing on CA (e.g., Kasper & Wagner, 2011; Schegloff, 2007), is necessary to understand the situated nature of the emergent linguistic inventory. The second set of analyses, whose results are presented in this section, thus address the locally contextualized and situated nature of L2 learning.

It was noted earlier for Valerio, in the discussion on Table 5, that his type-token ratios for the subject do/no verb-pattern suggested increasingly high productivity for both the nontarget- and targetlike variety. An especially intriguing finding was that the nontargetlike variety went from being relatively infrequent in RP1 and RP2 to becoming, in RP3, almost as frequent and twice as productive as the targetlike equivalent. Investigating the locally contextualized nature of L2 development, the present analyses focus on the causes of this development and why the pattern is retained by Valerio for so long as a seemingly productive pattern alongside the targetlike variety.

To explore local contextualization it is necessary to narrow the focus from general negation constructions to a particular way of instantiating it. This is done empirically. The data show that the subject no verb construction is dominated by the utterance schema you no verb, which in turn is dominated by a locally recurring MWE, you no write.Table 8 displays this in terms of token frequencies for subject no verb, you no verb and the MWE you no write in RPs 2, 3 and 4.

Table 8.  Type and Token Frequencies for you no verb in Relation to subject no verb Use (Valerio)
 Subject no verb TokensYou no verb
TokensRatioMWE you no writea
  1. Note. RP = recording period.

  2. a MWE tallies are included under the you no verb category count and disaggregated in this column for comparison.

RP 2310.330
RP 324150.63 9 (9/24 = 38% of subject no verb;
    9/15 = 60% of you no verb)
RP 434240.710

You no verb tokens constitute between 33 and 71% of all subject no verb tokens in RPs 2, 3, and 4. The primary thing to note, however, is that you no write is a dominant locally recurring MWE, constituting 9 of the 15 tokens (60%) of you no verb in RP3, the period which saw the sudden increase in subject no verb use. In the following analyses of Valerio's learning trajectory, I trace the local interactional contexts of the emergence of you no verb and its targetlike counterpart, you don't verb, respectively.

Emergence of You No Verb in Interaction

The most frequent purpose for which Valerio uses the you no and you don’t constructions is to make assertions (Pomerantz, 1984) or ask confirmatory questions (Turk, 1999). However, the you no construction is also used to achieve a specific purpose for which Valerio uses no other construction and for which he does not possess more relevant L2 resources, namely to instruct his fellow students on the task at hand. Valerio uses it for this purpose nine times in total, eight of which fall on the same day. Due due to space limitations, one extract will have to suffice as example: Extract 1, where the students are doing free movement interactions, a common praxis in this classroom in which the students walk around freely engaging relatively briefly in cued serial dyadic interactions (Hellermann & Cole, 2009). In this exercise, the students have been instructed by the teacher to write down the names of the students who answer their are you going to + verb questions affirmatively.

inline image

In lines 1–3, a classmate asks Valerio a task-question and Valerio answers. The student then starts writing, and Valerio instructs her not to write at this point, using you no write, the you no construction token under analytical focus here, as one of the resources to achieve this (lines 4–8). Valerio ends his turn in a verbally incomplete manner with you no: accompanied by a vertical open palm gesture (line 8–9). The gesture seemingly works as an embodied completion of the turn (Olsher, 2004), thus projecting a relevant point for speaker transition. In response, the co-participant, delivers a nothing (line 10) which displays understanding and thus helps maintain intersubjectivity, and Valerio immediately incorporates the item into his next turn, a repetition of the instruction, you no write nothing. These turns are exchanged without any delay or trouble, and the sequence is finally closed down as Valerio's partner, in line 12, thanks him following a change of state-token o:h (Heritage, 1984).

The sudden increase in the use of you no verb discussed earlier (cf. Table 8) happens on this day where Valerio uses the expression you no write eight times to instruct his classmates. There is a further use of you no write one month later, which also takes place in a task instruction environment. Apart from these nine occurrences, the expression is not found again in the data. It is indexical of a given interactional environment, as Valerio exclusively—and, importantly, successfully—deploys it for the purpose of instructing classmates in the task at hand.

The deployment of this particular pattern at this particular time, then, is a matter of using available resources allowing Valerio to achieve interactional goals that are otherwise beyond his L2 skills (Theodórsdóttir & Eskildsen, 2011). In February 2004, Valerio does not master any construction that would satisfactorily fulfill his purpose, his only negated imperative in use being don't worry, which functionally is more like an encouragement in the structural shape of an imperative. You don't verb is yet to emerge, and expressions like you’re not supposed to verb are not among his L2 resources. On the other hand, Valerio had used the you no verb-pattern before this occasion and thus it was already part of his linguistic experience, arguably making for the resource that gets him closest to his interactional target. The fact that the nontargetlike pattern, subject no verb, following the sudden increase in usage materialized in the you no write instantiations, becomes entrenched and evolves into a strong pattern in the competition against the target language pattern (as displayed in the type and token counts in Table 5) may in large part be due to the successful repeated use of you no write. From a UBL perspective, with the emergence of you no write in recurring interactions, the nontargetlike pattern becomes a statistical preference in Valerio's linguistic inventory, its use in later interactions a matter of the frequency-biased build of his linguistic inventory—he has not incorporated enough targetlike patterns into his linguistic inventory in order for it to outcompete the nontargetlike pattern. Given the highly contextualized nature of you no write and its numerical importance in relation to other subject no verb structures in use at the time, it is likely that different interactional environments (e.g., language learning tasks in the classroom) would have yielded a different trajectory of emergence for this particular feature of Valerio's negation construction inventory. While this speculative note shows the natural limits of the data, it also implies that the fabric of the contexts in which linguistic development is taking place is profoundly interwoven with that development and deserves minute attention.

Emergence of You Don't Verb in Interaction

This analysis traces the early emergence of the targetlike variety, you don't verb, in interaction. Valerio's first recorded use of you don’t falls on March 9, 2004, as the students are practicing short answers yes I do and no I don’t. As such, it does not contain a negated verb, but may have helped Valerio internalize the possibility of combining you and don’t. The afforded nature of this is displayed in Extract 2a.

inline image

Daniel asks Valerio if he has kids and receives a satisfactory answer to his question, following which he immediately moves on to answer the same question on his own behalf (lines 1–4). Valerio's next action (line 5) works as a confirmation request which Daniel confirms (line 6), but it may also be an embedded correction in that Valerio uses the form to be practiced, don’t, rather than simply recycling Daniel's didn’t. Daniel's yes might look like a disconfirmation but it is not, as evidenced by Valerio's next action, namely his writing which displays that Daniel's response was sufficient to continue solving the classroom task (line 7). Daniel's follow-up in the form of repeated displays of orientation to Valerio's writing and repetition of words from the task question further displays that their intersubjectivity is intact. The sequence is closed as Daniel accepts Valerio's writing (an emphatic yeas; line 10).

Forty-five minutes after this interaction, a situation arises in Extract 2b in which Valerio employs two forms that may also have helped shape his negation construction inventory. The teacher is instructing the students to do the next task in which they are supposed to express whether they agree or disagree with given statements. The teacher explains the notion of disagree to the whole class (line 1).

inline image

The teacher's explanation is not intended for Valerio only; and his response is an example of what is usually known as private speech (Lantolf & Thorne, 2006) and does not necessarily warrant a response. The interesting thing is that the first half of his utterance is recycled from the teacher's prior turn, whereas the second half is a new production on the basis of this recycled material. Seconds later, a fellow student asks Valerio a question in Spanish to which he responds, you like don't like. The students then embark on solving the task.

The four instantiations of don’t mentioned so far, then, may not all at face value be bona fide members of the you don't verb-construction. You don’t has no lexicalized verb, and don't believe it don't like it has no lexicalized subject. However, you like don't like, Valerio's first don’t-construction containing both a subject and a negated verb, can be shown to build on the affordance of you don't believe it that went immediately before it and to make use of the previously afforded you don’t combination (Extract 2a), which, it must be remembered, happened only 45 minutes before the interaction in Extract 2b. Together, these four instantiations mark the afforded, locally contingent, incipient emergence of you don't verb.

In the next extract, almost two months after the examples just discussed, Valerio produces his next you don't verb example, and again he is dependent on interactional affordances as he recycles what his classmate Sal has just said (Extract 2c). The students are reading some information in preparation of doing a task. Both students are reading loud in overlap (lines 1–2).

inline image

Smiling and gesturing “take it easy” with his palms down (Kendon, 2004), Sal mockingly complains that Valerio is reading out loud (lines 3–5). Valerio returns the complaint, mockingly repeating Sal's utterance, as Sal was just reading out loud, too (line 6–7). Smiling, Sal denies having read out loud; they both seem to be treating the situation as a joke. Note how Valerio's turn in line 6, the target utterance, seems to carry traces of the nontargetlike pattern, a distinctly audible “n” sound coming between you and don’t.

The next instantiation, three weeks later, is a self-repair of you no talk to you don't talk. The transcript is not shown, because the next part of the interaction is unintelligible. Seconds later, Valerio picks up the expression I don't mind from a fellow student across the classroom and repeats it out loud. The teacher hears this orientation to the expression and focuses for a while on getting Valerio to pronounce mind correctly. This results in two further repetitions of I don't mind. Then, the interaction in Extract 2d takes place. Here, in an off-task situation, Sal is yawning (line 1), which catches Valerio's attention.

inline image

Valerio asks Sal why he is tired in response to which Sal shrugs (lines 2–3). Then comes the target expression, you don't sleep. It is a confirmatory question used to pursue confirmation of what it negates (Turk, 1999); confirmation that is requested because the underlying assumption that Sal did sleep is questioned by something in the discourse (Sal's yawning). Sal orients to the question in precisely this way as he confirms that he did sleep (line 5), and the students leave the topic.

Within a very brief window of time, then, Valerio uses you don't sleep after having practiced pronouncing I don't mind and after his own self-repair of you no talk to you don't talk. Constituting an empirical observation of the output hypothesis at work (Swain, 1995), the self-repair is interesting as it may offer explicit evidence that his negation inventory is undergoing changes. The following use of you don't sleep, Valerio's first complete, nonafforded, nonrepaired use of you don't verb may be further evidence of this ongoing entrenchment process. The intervening repetitions of I don't mind are mentioned because they may also play a role in the expansion of don’t from I don't verb to you don't verb that is taking place.

The next you don’t-instantiation is shown in Extract 2e, which falls on the same day as Extract 2d. Sal and Valerio are doing pair work on vocabulary items. They have been instructed to imagine they are making a time capsule in the 1960s and then, from a vocabulary list, choose two items which they think people in the 1960s might have put in a time capsule. They are having problems agreeing on what items to choose, partly because they do not understand the meaning of all of them.

inline image

In line 1–2, Sal says he does not understand one of the items in the book. Recycling Sal's turn with a change in speaker perspective, Valerio then asks about this (line 3). After a pause during which Sal is looking in his papers he states that nail file is causing him trouble (lines 4–5).

In the teacher-fronted interaction in Extract 2f, recorded 15 minutes later, the class is still working on the time capsule theme. The focal lines are 2, 6, and 8, in which Valerio, respectively, recycles parts of Sal's prior turn and does a self-repair of a nontargetlike expression which he himself has used before, you no useh to you don't useh.7 Prior to the interaction in the extract, the teacher has introduced the students to another item from the 1960s, a plastic valve for medical use.

inline image

In line 1, Sal says that he does not know the item in response to which Valerio immediately orients physically to Sal (he turns around to face him), and, recycling his words with a change in speaker perspective (line 2), asks a confirmatory question. After a pause during which there is parallel talk in the class (denoted by the omitted lines in the transcript), Sal answers no, which closes down the sequence (line 3). Whether Valerio in fact knows the item remains uncertain, because at the moment where he could have explained it to Sal, the next item, contact lenses, is introduced by the teacher (lines 4–5). The teacher asks the class about contact lenses and Valerio embarks on a response using gesture followed by the nontargetlike pattern, you no useh glasses (line 6–7; note again the off-target pronunciation of ‘use’). In overlap (line 8), the teacher then acknowledges his response (a:h yeah). Next, the teacher and Valerio speak at the same time (lines 7–10); the teacher is rephrasing Valerio (they replaced glasses they were new) in overlap with which Valerio does a self-repair (you don't useh glasses) and elaborates on his response (again doing a self-repair) (it's lit- very small). At the end of the overlap, in line 11, Valerio displays understanding of the teacher's reformulation (yeah), and the teacher then gives another reformulation of Valerio's turn followed by an acknowledgment token, uhuh (line 13), which closes down the sequence.

Chronologically, then, Valerio's learning of you don't verb seems to originate in the afforded uses you don’t, don't believe it don't like it and you like don't like. Then, one month later, came another affordance, you ndon't read them loudly. Three weeks later, he did a self-repair of you no talk to you don't talk and within the next 60 minutes he further produced you don't sleep, the first nonafforded, nonrepaired use, as well as afforded uses what you don't understand and you don't know before another self-repair (you no use to you don't use). From then on, he begins using you don't know on a recurring basis and then gradually starts using a more varied you don't verb utterance schema. In short, a productive evolution of subject don't verb, initially dependent on affordances, self-repairs, and recycled main verbs, is dependent on the MWE you don't know, thus also showing exemplar-based characteristics in development.

The emergence of the two structures I have just traced indicates that for you no verb, a frequent local recurrence of a successful use of you no write may have spawned the entrenchment of the pattern, suggesting an essential link between interaction and learning trajectories. For you don't verb it seems that the interactional environments initially afforded Valerio's first uses of the pattern, which supports Vygotskian concepts of learning in the Zone of Proximal Development (Lantolf & Thorne, 2006), a psychological construct demarcating the gap between what a learner can do on his own and what he can do when scaffolded, that is, in collaboration with others. Transported into L2 learning contexts, the ZPD implies that emergent constructions originate in coachieved activities and become internalized over time in a movement from the social, interpersonal plane to the individual plane (Lantolf & Thorne, 2006). Moreover, stressing the experiential nature of language learning, when Valerio starts using the pattern on his own, he is dependent on recycled material from his own previous experience. Although not presented in this article for reasons of space, the same developmental trajectory was observed for Carlos's targetlike no/not-phrases; initially these were also picked up and recycled from the immediate environment, with subsequent uses consisting to a large extent of his own recurring phrases. In all, these observations point toward the necessity of investigating interactional contingencies of emergent L2 constructions.

General Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

In the present two-pronged investigation, the framework of UBL was used in a first set of analyses in order to inspect evidence for the development of English negation over 2 and 31/2 years, respectively, in interaction data for two adult ESL students. As predicted by UBL models of language knowledge and acquisition (e.g., N. Ellis, 2002; Tomasello, 2003), the data for both learners supported a view of learning of English negation from recurring expressions toward an increasingly schematic, dynamic inventory of linguistic resources. The targetlike varieties of the negated patterns were shown to be predominantly exemplar based, whereas the nontargetlike varieties seem to be more productive default solutions. An exception to this was Valerio's you no verb which was found to have an initial dependence on a local high recurrence of you no write. It was suggested that this particular expression and its successful deployment might have played an important role for the subject no verb pattern in the competition against the targetlike variety. The targetlike variety of this utterance schema, in turn, was dependent on two things: (1) an initial dependence on the locally constituted affordances and (2) a previously recurring pattern as you no useh evolved into you don't useh.

The current UBL findings lead to confirmation of some aspects of the findings about English negation from the early SLA studies. They also arrive at different developmental conclusions and they offer novel reconceptualizations of central issues. The evidence presented here (and in our previous study, see Eskildsen & Cadierno, 2007) supports the general consensus that no verb and don't verb are in competition (Schumann, 1979). The data also confirm the general finding that modal negation, usually instantiated by can as the first negated modal, emerges after an initial period of a piecemeal do-negation development. However, there is no support of the copula negation emerging at the same time as modal negation. For both Carlos and Valerio, targetlike copula negation is an earlier resource in the inventory. Combined with the fact that both students use targetlike do-negation from an early point, this suggests that they do not, at any stage, operate on the basis of an early general no+ verb/phrase rule, contrary to what is traditionally postulated (e.g., Stauble & Schumann, 1983). In sum, the targetlike patterns were generally increasing at the cost of the nontargetlike patterns, which echoes the distributional analyses in Cancino et al. (1978, pp. 211–217). The data, however, showed no acquisitional stage-defining pattern-dominance.

There are, moreover, some fundamental differences between the path of learning portrayed by UBL and uncovered in the present data and that generally conveyed by previous developmental sequence studies in traditional SLA studies. These differences show how the UBL framework can enable novel reconceptualizations of central L2 learning issues.

One difference resides particularly in the role played by the MWE, I don't know, in both Carlos and Valerio. In previous developmental studies, such MWEs are typically excluded from the analysis because they are thought of as unanalyzed chunks (Cancino et al., 1978; Schumann, 1979). This ties in well with the tradition of adopting a compartmentalized view of language knowledge which sustains the syntax-lexis division. In this view, MWEs such as I don't know are seen as somehow beyond the current combinatorial abilities sanctioned by the learners’ interlanguages. This is incongruent with UBL's trajectory of learning (for more detailed discussions, see Eskildsen, 2009; Eskildsen & Cadierno, 2007), which instead argues that the MWE is a crucial exemplar sparking the learning of the pattern. In this sense it is part and parcel of the emergent linguistic inventory and not to be ignored on the grounds that it is a memorized whole; it has already been mentioned that UBL views all linguistic items as psycholinguistically identical in nature (Croft & Cruse, 2004), but, following pioneering research such as Bolinger (1979), and Peters (1983), UBL also argues that all linguistic items are stored cognitive routines (e.g., Kemmer & Barlow, 2000; Langacker, 1987), implying that the feature of being memorized cannot be the sole distinguishing factor in the categorization of linguistic items. Furthermore, I don't know is not the only MWE in the data, leading to the issue of whether other MWE candidates (e.g., I don't think so, I don't remember, you no write) should be left out of the analyses too. In the present theoretical and methodological framework, then, no constructions should be discarded on grounds of being memorized patterns; instead, pattern behavior in terms of exemplar-dependence and productivity is a gradable issue measured, for example, in type-token distribution analyses carried out on the longitudinal production data.

Another fundamental difference is that the UBL approach focuses on the ontogenesis of singular patterns that share constructional features, whereas the traditional developmental sequence approach focused on the emergence of a syntactically rule-governed positioning of a broadly applicable negation particle. The present research builds on the empirical fact that the patterns in the data display individual acquisitional trajectories; neither targetlike nor nontargetlike features can be generalized to all negated patterns so what the linguist or the analyst calls negation does not seem to be learned as a rule-governed syntactic phenomenon to be deployed across diverse linguistic patterns in a broad-sweeping manner, but seems to emerge in different patterns in different ways at different points in time along, rather than across, constructional lines.

A third difference between UBL and other understandings of L2 development can be seen in the theoretical status of the entrenchment that was observed in Valerio's learning trajectory for copula negation (see Table 6). For many SLA researchers the evidence that no is remains in use at a fairly frequent level throughout may resemble fossilization (Selinker, 1972) or pidginization (Schumann, 1976), which both denote that the learning process has come to a halt. One of the informants in the work of Schumann and colleagues, Alberto, became the epitome of the pidginization hypothesis because of his seeming lack of L2 development. However, a later reanalysis of this data from a variationist, probabilistic perspective (Berdan, 1996) showed that his language was, in fact, developing toward an increasingly targetlike inventory of negation structures. Such research sits well with the present approach in its probabilistic principles. In the present case of Valerio, his seemingly frozen use of no is does not seem like a good candidate for a fossilized phenomenon, either. His linguistic inventory is generally dynamic and ever-changing (as evidenced by the constant development of the parallel targetlike copula negation). Therefore, because development is nonlinear, because no is constitutes a widely applicable pattern that has seemed to work in interaction, and because the apparent halt in development is restricted to this particular pattern, it is more apt to describe it as such, namely an entrenched item-based pattern which simply remains available for use for the time being. This description of the phenomenon captures the usage-based commitment and takes to heart the insight that grammar is emergent in the sense that it never reaches a final state of completion (Larsen-Freeman, 2005).

The present investigation of the emergence of L2 English negation did not stop with the UBL evidence. Instead, based on the argument that language is not only emergent but emergence itself is locally contingent and situated, a move toward a locally contextualized view of L2 learning was implemented analytically by exploring the evolution of two negation patterns in ongoing discourse with the aid of CA. Such analyses demonstrated how, at the point in time when Valerio finds himself in interactions requiring him to assist his fellow classmates in getting a task right, he uses the linguistic resources readily available to him. In this case, this resource was a nontargetlike, lexically specific pattern you no write. The usage events which prompted this locally heavy use of you no write may have laid the foundation for what was to become a seemingly statistical feature of Valerio's linguistic inventory, namely the preference for you no verb at the cost of a more targetlike do-negation pattern. This targetlike variety, you don't verb, was then shown to have derived from specific afforded instances of use and to be dependent on recurring lexical material. In sum, the data in this second set of analyses suggested a fundamental coupling of linguistic development and interactional requirements; interaction and learning, in other words, cannot be kept apart.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References

This study of L2 negation learning has shown how linguistic pattern emergence is exemplar- and item-based and very often based on affordances. It was also shown that most of the nontargetlike patterns were less item-based and hence more freely productive than the targetlike varieties. Furthermore, the data have suggested that learning L2 grammar is a matter of exemplar-deduced constructional tendencies that may or may not become schematized as abstract linguistic knowledge in ontogenesis. It is possible to track pattern development in great detail from the concrete item-based starting point of the patterns to the possible abstraction of regularities that link these patterns as schemas. Such possible abstraction, however, should not be the default starting point for longitudinal L2 learning studies, as research has shown that not all patterns in development lend themselves easily to abstraction (Eskildsen, 2009).

By implication, the data suggests that a compartmentalized view of language, that is, the idea that lexis and grammar are separate entities, should be abandoned. L2 learners do not seem to learn the two in a manner that justifies keeping them apart; they are inseparable, and learners acquire them together, not each in its own paradigmatic vacuum. Tomasello (2003) put it nicely when he said for first language acquisition that children must learn two faces of grammar: smaller elements and larger patterns. It seems that L2 learning research can benefit from this insight, but further studies are needed to investigate its validity.

The methodology proposed here has drawn on N. Ellis's (e.g., 2002) encouragement to use UBL's exemplar-based path of learning as the default way of investigating L2 learning. But it has also taken the usage-based commitment further and shown the situated and indexical nature of language and language learning in microanalyses of what happens in the usage events where constructions being learned are first put to use. These moments constitute the seed of construction learning, bearing in mind Tomasello's (2000) words that however abstract constructional knowledge may ultimately become, it derives invariably from specific occasions of use. Investigating longitudinal language learning through this lens on local usage events has proved fruitful, but more investigations are needed that study the socio-cognitive details of what happens when L2 constructions are put to work in interaction.

Notes
  • 1

    I will use “utterance schema” and “item-based pattern” interchangeably throughout.

  • 2

    MAELC was supported, in part, by grant R309B6002 from the Institute for Education Science, U.S. Dept. of Education, to the National Center for the Study of Adult Learning and Literacy (NCSALL) and was a partnership between Portland State University and Portland Community College. I thank Steve Reder for granting me access to MAELC.

  • 3

    In addition, through practice situations, Carlos is exposed to various uses of the past tense short answer construction no subject didn’t in RPs 1 and 2, but this no I didn’t in RP4 is his only spontaneous use of the past tense.

  • 4

    There are 3 instances of I won’t as practiced form in RP4.

  • 5

    He also has a use of don't laugh in a situation where a fellow classmate says don't laugh to him, and responds by saying why don't laugh?.

  • 6

    Transcript conventions: 01,02,03…: line numbers; Va, Sa…: speakers (Un = unknown student); -> = target utterance; (1)/(.) = pause, length indicated in seconds/pause shorter than one second; ((…)): transcriber's comments; : = stretched vowel; []= overlap; [DOWNWARDS ARROW], [UPWARDS ARROW]= falling / rising intonation; “-”= cut-off. “=”= continued lines.

  • 7

    This spelling denotes that Valerio pronounces use as /ju:sə/.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Usage, Constructions, and Language Learning
  5. Emergence in Local Usage
  6. Formulaic Language and Developmental Stages of English Negation
  7. The Present Longitudinal Learner Corpus
  8. Results of UBL Analyses
  9. Toward a Locally Contextualized View of the Emergence of Constructions
  10. General Discussion
  11. Conclusion
  12. References
  • Achard, M. (2007). Usage-based semantics: Meaning and distribution of three French “breaking” verbs. In M. Nenonen & S. Niemi (Eds.). Collocations and idioms 1: Papers from the first Nordic conference on syntactic freezes, Joensuu, May 19–20, 2007 (pp. 113). Joensuu , Finland : Joensuu University Press.
  • Bartning, I., & Hammarberg, B. (2007). The functions of a high-frequency collocation in native and learner discourse: The case of French c’est and Swedish det är. International Review of Applied Linguistics, 45, 143.
  • Beckner, C., Ellis, N. C., Blythe, R., Holland, J., Bybee, J., Ke, J., et al. (2009). Language is a complex adaptive system: Position paper. Language Learning, 59(Suppl. 1), 126.
  • Berdan, R. (1996). Disentangling language acquisition from language variation. In R. Bayley & D. R. Preston (Eds.), Second language acquisition and linguistic variation (pp. 203243). Amsterdam : John Benjamins.
  • Bolinger, D. (1979). Meaning and memory. In G. G. Haydu (Ed.), Experience forms. Their cultural and individual place and function (pp. 95111). The Hague: Mouton.
  • Cancino, H., Rosansky, E. J., & Schumann, J. H. (1978). The acquisition of English negatives and interrogatives by native Spanish speakers. In E. M. Hatch (Ed.), Second language acquisition: A book of readings (pp. 207230). Rowley , MA : Newbury House.
  • Collins, L., & Ellis, N. C. (Eds.). (2009). Input and second language acquisition: The roles of frequency, form, and function [special issue]. The Modern Language Journal 93(3).
  • Croft, W., & Cruse, A. (2004). Cognitive linguistics. Cambridge , UK : Cambridge University Press.
  • Dąbrowska, E., & Lieven, E. (2005). Towards a lexically specific grammar of children's question constructions. Cognitive Linguistics, 16, 437474.
  • de Bot, K., Lowie, W., & Verspoor, M. (2007). A dynamic systems theory approach to second language acquisition. Bilingualism: Language and Cognition, 10, 721.
  • Doughty, C. J. (2003). Instructed SLA: Constraints, compensation and enhancement. In C. J. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 256310). Malden , MA : Blackwell.
  • Doughty, C. J., & Long, M. H. (Eds.). (2003). The handbook of second language acquisition. Malden , MA : Blackwell.
  • Ellis, N. C. (2002). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24, 143188.
  • Ellis, N. C., & Cadierno, T. (2009). Constructing a second language: Introduction to the special section. Annual Review of Cognitive Linguistics, 7, 111139.
  • Ellis, N. C., & Ferreira-Junior, F. (2009). Construction learning as a function of frequency, frequency distribution, and function. The Modern Language Journal, 93, 370385.
  • Ellis, N. C., & Larsen-Freeman, D. (2009). Constructing a second language: Analyses and computational simulations of the emergence of linguistic constructions from usage. Language Learning, 59(Suppl. 1), 90125.
  • Ellis, R. (1994). The study of second language acquisition. Oxford , UK : Oxford University Press.
  • Eskildsen, S. W. (2009). Constructing another language: Usage-based linguistics in second language acquisition. Applied Linguistics, 30, 335357.
  • Eskildsen, S. W. (2011). The L2 inventory in action: Conversation analysis and usage-based linguistics in SLA. In G. Pallotti & J. Wagner (Eds.), L2 learning as social practice: Conversation-analytic perspectives (pp. 337373). Honolulu : University of Hawai‘i, National Foreign Language Resource Center.
  • Eskildsen, S. W., & Cadierno, T. (2007). Are recurring multi-word expressions really syntactic freezes? Second language acquisition from the perspective of usage-based linguistics. In M. Nenonen & S. Niemi (Eds.), Collocations and idioms 1: Papers from the First Nordic Conference on Syntactic Freezes, Joensuu, May 19–20, 2007 (pp. 86–99). Joensuu , Finland : Joensuu University Press.
  • Goldberg, A., & Casenhiser, D. (2008). Construction learning and second language acquisition. In P. Robinson & N. C. Ellis (Eds.), Handbook of cognitive linguistics and second language acquisition (pp. 197215). New York : Routledge.
  • Gries, S. T., & Wulff, S. (2005). Do foreign language learners also have constructions? Evidence from priming, sorting, and corpora. Annual Review of Cognitive Linguistics, 3, 182200.
  • Hellermann, J., & Cole, E. (2009). Practices for social interaction in the language learning classroom: Disengagements from dyadic task interaction. Applied Linguistics, 30, 186215.
  • Herdina, P., & Jessner, U. (2002). A dynamic model of multilingualism: Perspectives of change in psycholinguistics. Clevedon , UK : Multilingual Matters.
  • Heritage, J. (1984). A change-of-state token and aspects of its sequential placement. In J. Atkinson & J. Heritage (Eds.), Structures of social action: Studies in conversation analysis (pp. 299345). Cambridge , UK : Cambridge University Press.
  • Hopper, P. (1998). Emergent grammar. In M. Tomasello (Ed.), The new psychology of language: Cognitive and functional approaches to language structure (vol. 1, pp. 155175). Mahwah , NJ : Lawrence Erlbaum.
  • Huebner, T. (1985). System and variability in interlanguage syntax. Language Learning, 35, 141163.
  • Hyltenstam, K. (1977). Implicational patterns in interlanguage syntax variation. Language Learning, 27, 383411.
  • Kasper, G., & Wagner, J. (2011). A conversation-analytic approach to second language acquisition. In D. Atkinson (Ed.), Alternative approaches to second language acquisition (pp. 117142). New York : Taylor & Francis.
  • Kemmer, S., & Barlow, M. (2000). Introduction: A usage-based conception of language. In M. Barlow & S. Kemmer (Eds.), Usage-based models of language (pp. vii–xxviii). Stanford , CA : Center for the Study of Language and Information.
  • Kendon, A. (2004). Visible action as utterance. Cambridge , UK : Cambridge University Press.
  • Klima, E. S., & Bellugi, U. (1966). Syntactic regularities in the speech of children. In J. Lyons & R. J. Wales (Eds.), Psycholinguistic papers: The proceedings of the 1966 Edinburgh conference (pp. 183219). Edinburgh , UK : Edinburgh University Press.
  • Langacker, R. W. (1987). Foundations of cognitive grammar, vol. 1. Theoretical prerequisites. Stanford , CA : Stanford University Press.
  • Langacker, R. W. (2000). A dynamic usage-based model. In M. Barlow & S. Kemmer (Eds.). Usage-based models of language (pp. 164). Stanford , CA : Center for the Study of Language and Information.
  • Lantolf, J. P., & Thorne, S. L. (2006). Sociocultural theory and the genesis of second language development. Oxford , UK : Oxford University Press.
  • Larsen-Freeman, D. (1991). Staking out the territory. TESOL Quarterly, 25, 315150.
  • Larsen-Freeman, D. (2005). Second language acquisition and fossilization: There is no end, and there is no state. In Z.-H. Han & T. Odlin (Eds.), Studies of fossilization in second language acquisition (pp. 189200). Clevedon , UK : Multilingual Matters.
  • Larsen-Freeman, D. (2006). The emergence of complexity, fluency, and accuracy in the oral and written production of five Chinese learners of English. Applied Linguistics, 27, 590619.
  • Lieven, E., Salomo, D., & Tomasello, M. (2009). Two-year-old children's production of multiword utterances: A usage-based analysis. Cognitive Linguistics, 20, 481508.
  • MacWhinney, B. (1975). Pragmatic patterns in child syntax. Stanford Papers and Reports on Child Language Development, 10, 153165.
  • Meisel, J., Clahsen, H., & Pienemann, M. (1981). On determining developmental stages in natural second language acquisition. Studies in Second Language Acquisition, 3, 109135.
  • Mellow, J. D. (2006). The emergence of second language syntax: A case study of the acquisition of relative clauses. Applied Linguistics, 27, 645670.
  • Nattinger, J. P., & DeCarrico, J. S. (1992). Lexical phrases and language teaching. Oxford , UK : Oxford University Press.
  • Olsher, D. (2004). Talk and gesture: The embodied completion of sequential actions in spoken interaction. In R. Gardner & J. Wagner (Eds.), Second language conversations (pp. 221245). London : Continuum.
  • Pawley, A., & Syder, F. (1983).Two puzzles for linguistic theory: Nativelike selection and nativelike fluency. In J. C. Richards & R. W. Schmidt (Eds.), Language and communication (pp. 191226). London : Longman.
  • Peters, A. (1983).The units of language acquisition. Cambridge , UK : Cambridge University Press.
  • Pica, T. (2005). Second language acquisition research and applied linguistics. In E. Hinkel (Ed.), Handbook of research in second language teaching and learning (pp. 261283). Mahwah , NJ : Lawrence Erlbaum.
  • Pomerantz, A. (1984). Agreeing and disagreeing with assessments: Some features of preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action (pp. 57101). Cambridge , UK : Cambridge University Press.
  • Reder, S. (2005). The “Lab School.” Focus on Basics, 8a, 16.
  • Reder, S., Harris, K. A., & Setzler, K. (2003). A multimedia adult learner corpus. TESOL Quarterly, 37, 546557.
  • Schegloff, E. A. (2007). Sequential analysis: A primer in conversation analysis. Cambridge , UK : Cambridge University Press.
  • Schegloff, E. A., Ochs, E. & Thompson S. A.. (1996). Introduction. In E. Ochs, E. A. Schegloff, & S. A. Thompson (Eds.), Interaction and grammar (pp. 151). Cambridge , UK : Cambridge University Press.
  • Schumann, J. H. (1976). Second language acquisition: The Pidginization hypothesis. Language Learning, 26, 391408.
  • Schumann, J. H. (1979). The acquisition of English negation by speakers of Spanish: A review of the literature. In R. W. Andersen (Ed.), The acquisition and use of Spanish and English as first and second languages (pp. 332). Washington , DC : Teachers of English to Speakers of Other Languages.
  • Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics, 10, 209231.
  • Stauble, A. (1978). The process of decreolization: A model for second language development. Language Learning, 28, 2954.
  • Stauble, A., & Schumann, J. H. (1983). Towards a description of the Spanish-English Basilang. In K. M Bayley, M. H. Long, & S. Peck (Eds.), Second language acquisition studies (pp. 6882). Rowley , MA : Newbury House.
  • Swain, M. (1995). Three functions of output in second language learning. In G. Cook & B. Seidlhofer (Eds.), Principle and practice in applied linguistics (pp. 125144). Oxford , UK : Oxford University Press.
  • Tarone, E., & Swierzbin, B. (2009). Exploring learner language. Oxford , UK : Oxford University Press.
  • Theodórsdóttir, G., & Eskildsen, S. W. (2011). The use of English in everyday Icelandic as a second language: Establishing intersubjectivity and doing learning. NordAnd, 6(2), 5985.
  • Tomasello, M. (1992). First verbs: A case study of early grammatical development. Cambridge , UK : Cambridge University Press.
  • Tomasello, M. (2000). First steps toward a usage-based theory of language acquisition. Cognitive Linguistics, 11, 6182.
  • Tomasello, M. (2003). Constructing a language. Cambridge , UK : Cambridge University Press.
  • Tummers, J., Heylen, K., & Geeraerts, D. (2005). Usage-based approaches in cognitive linguistics: A technical state of the art. Corpus Linguistics and Linguistic Theory, 1–2, 225261.
  • Turk, M. J. (1999). Negatively formulated questions in interaction. In W. Klein & A. Barber (Eds.), Crossroads of language, interaction and culture (vol. 2, pp. 3948). Los Angeles : The Center for Language, Interaction, and Culture.
  • van Lier, L. (2000). From input to affordance: Social-interactive learning from an ecological perspective. In J. Lantolf (Ed.), Sociocultural Theory and Second Language Learning (pp. 246260). Oxford , UK : Oxford University Press.
  • Verspoor, M., Lowie, W., & de Bot, K. (2011). A dynamic approach to second language development: Methods and techniques. Amsterdam : John Benjamins.
  • Waara, R. (2004). Construal, convention, and constructions in L2 Speech. In M. Achard & S. Niemeier (Eds.), Cognitive linguistics, second language acquisition and foreign language pedagogy (pp. 51–75). Berlin : Mouton de Gruyter.
  • Weinert, R. (1995). The role of formulaic language in second language acquisition: A review. Applied Linguistics, 16, 180205.
  • Young, R. (1988). Variation and the interlanguage hypothesis. Studies in Second Language Acquisition, 10, 281302.