Workshop

METHODOLOGICAL PROBLEMS IN THE STUDY OF FIRST LANGUAGE ATTRITION IN A SECOND LANGUAGE ENVIRONMENT

M.C.J. (Merel) Keijzer

This research project examines the non-pathological language attrition of Dutch as a first language among emigrants in the English-speaking part of Canada. More specifically, it examines whether the phenomena witnessed in this situation can be accounted for on the basis of the regression hypothesis (first postulated in a linguistic context by Jakobson in 1941). This hypothesis suggests that the order in which a language is lost is the reverse of that in which it was first acquired in childhood. There appears to be general consensus that this hypothesis cannot hold for pathological language loss, for example in aphasia patients (cf. Berko-Gleason, 1982; Seliger, 1991), but it has never been tested in a data-driven research on non-pathological first language attrition.

The very fact that data will be collected as part of an experimental research design, while working within the very much theoretical framework of the regression hypothesis is not without problems. First and foremost: how do you keep your method as close as possible to the theory, while at the same time developing a test battery that is a practical device for data collection? Not only is the development of the test battery problematic, difficulties were also encountered in the data collection process itself, as well as in the subsequent process of data analysis, all of which will be discussed on the basis of a completed pilot study.

To start with the development of the test battery, the main tests within this battery had to focus on morphology and syntax, since those are the linguistic features showing a gradual process of acquisition in Dutch children, which can be compared to the potentially gradual process of attrition in Dutch emigrants in Canada. For this purpose the main tests in the test battery are a modified version of Berko’s wug test (1958) as well as a grammaticality judgment task. The other tasks in the test battery serve to establish a general pattern as well as background information and overall language proficiency. They include, for example, a sociolinguistic questionnaire, C-test and film retelling task. In selecting tasks for the test battery and modifying them, consideration had to be given to the fact that hardly any of these tasks were uncontroversial. The major methodological problem in developing the test battery therefore was how to take into account what is known about language testing in general and particular language tasks in particular, while at the same time focusing on your own individual research aims and objectives. A prime example in this respect is the grammaticality judgment task. This meta-linguistic task, as it is often thought to be, has been scrutinized for being used in second language acquisition research, but does that automatically mean that it cannot be used for the affiliated research field of first language attrition in a second language environment?

Similarly, and more generally, most tests that have traditionally been used to attest language attrition originally come from second language acquisition research. However, no conscious attempt has so far been made to adapt them to first language attrition research. How should language tests from one discipline be adapted to fit another, related discipline?

In addition, the process of data collection is not entirely unproblematic either. It is intuitively obvious that the researcher’s conduct as well as instructions may impact very heavily on the data actually collected. The term ‘observer’s paradox’ fits only too well here. How can this be controlled for? One obvious solution is to include a second, more neutral observer in the data collection process, but what in the case of a Ph.D. project with limited financial means when this is simply not feasible?

The issue of external judges also plays a role in the final stage of any experimental research, the analysis of the data. Certainly the researcher alone cannot judge whether a certain deviation found in the data constitutes an ‘error’ that may indicate attrition. Equally problematic is the baseline for comparison, since ‘unattrited’ speech is by no means error-free either. The ‘error’ issue is discussed elaborately in Schmid (forthc.), who advocates that studies on language attrition should “combine the analysis of ‘what went wrong’ with data on ‘what went right’”. While intuitively obvious, what is the best way to use such a combinatory approach?

The methodological issues discussed above will need to be addressed directly in order to develop research designs as well as methods of data collection and analysis that go some way towards explaining what really happens when first language attrition occurs in a second language environment.

Finally, despite the fact that these methodological issues are here discussed in relation to first language attrition research, they are just as relevant to the affiliated discipline of second language acquisition research. A discussion of the problems involved in the development of language tests, the process of data collection as well as that of data analysis is therefore hoped to be useful to all participants of the doctoral student workshop at EUROSLA 14.

 

THE AGE OF ONSET OF L2 ACQUISITION AS AN EXTERNAL VARIABLE IN AN L1 ATTRITION SITUATION

Emanuel Bylund

In current research there seems to be converging evidence indicating that L1 attrition in children is a fast and deep erosion, whereas in the case of adults attrition is far less significant (for an overview, see Köpke & Schmid, 2004). This has generated speculations about a certain age limit conditioning the outcome of the reduced L1 contact. However, because of the diverging sets of test variables and subjects examined in the studies general conclusions remain problematic.

This study aims at investigating the role of age of onset of L2 acquisition / bilingualism as an external variable in L1 (Spanish) attrition in an L2 (Swedish) environment. The subjects’ age of arrival ranges from 2 years to adult age and the selection was made consistent with the frequency of L1 use. Furthermore, the L2 proficiency of these subjects is studied in a project on maturational constraints and ultimate attainment in second language acquisition (Hyltenstam & Abrahamsson, 2003).

In order to assess age-related variation across different linguistic domains the test battery comprises production and perception of VOT, oral production and grammaticality judgements. With respect to grammar there will be a focus on [± interpretable] features within generative theory, so as to investigate whether [+ interpretable] features (categorical features and nominal phi-features) are more prone to attrition than [– interpretable] features (pertinent to morpho-syntax, such as agreement and case), as proposed by Sorace (2000). At a second stage of the study, the L1 proficiency data will be compared with those obtained in the L2 Swedish test, with the intention of reaching a global assessment of the subjects’ bilingual proficiency.

Firstly, it will be suggested that age plays a crucial role in attrition, i.e. that environmental and attitudinal factors cannot solely account for the considerable variation of L1 maintenance between child and adult attriters. Secondly, with respect to the proposal of a critical period in L1 attrition, this study suggests an interpretation of the age variable within the framework of L2 acquisition proposed by Hyltenstam & Abrahamsson (2003), i.e. stressing the gradual – rather than time span-constrained – effects of maturational factors. Thus, from an attrition perspective, the maturation period would imply L1 maintenance to be increasingly effortless and less affected by the amount of contact.

For discussion:

I: To what degree can the gradient of maturational factors be discerned and what claims can be made concerning age in view of the fact that the amount of L1 contact is a crucial variable in child attrition (Pallier et al., 2003)?

II: What are the possibilities of adopting an SLA-based framework of the maturation period, given that L2 contact in L1 attrition can be regarded as a relatively constant variable, whereas L1 contact is of a more dynamic character and thus somewhat difficult to measure?


 

THE ACQUISITION AND USE OF DISCOURSE MARKERS IN L2 SPANISH

Helena Lindqvist

Different notions of communicative competence, such as those proposed by Hymes (1971) from the point of view of linguistic anthropology or by Habermas (1984) from the perspective of social philosophy, have served as guidelines for establishing communicative competence as the overall goal of language learning and teaching. Specific models of communicative competence, including concepts of discourse and socio-pragmatic competence have been suggested by among others Canale & Swain (1983) and Bachman (1990). Interlanguage pragmatics (ILP) has since then been a prospering research field. However, most studies in ILP have focused on second language use rather than learning (Kasper & Schmidt 1996). They have tended to adopt a comparative, cross-cultural rather than an acquisitional approach. Because of this, several researchers have called for more longitudinal studies with a developmental orientation (Kasper & Rose 2001, 2003).

The current research project aims at investigating the acquisition and use of interactionally relevant discourse markers, adopting a cross-sectional as well as a longitudinal exploratory model. The first aim is to explore developmental patterns in the acquisition and use of discourse markers in instructed, semi-instructed and naturalistic learners of L2 Spanish. At a second stage, learners' use of interactionally relevant discourse markers will be compared to their performance regarding structural phenomena which traditionally have been given a higher priority in the classroom, such as verb inflection and verb and noun phrase agreement. A third aim is to examine the role of psychological factors such as attention, awareness and the importance of motivation in the acquisition of discourse markers. The study will thus be conducted within the conceptual framework of intra-personally oriented models, such as acculturation and cognitive processing theory, as well as from an inter-personal perspective, including socio-cultural theory and models for language socialization and interactional competence.

Our data consists of conversations, interviews and retellings, belonging to the AKSAM-and CorALEs corpus established at Stockholm University.

Methodological issues of importance are the following:

• How do we define instructed, semi-instructed and naturalistic learners? Any need for more precision?

• Which are the appropriate data collection methods for measuring motivation and acculturation?

• Does a native speaker control group have to be studied and compared with non-natives for reliable results to be obtained? Is the quantitative score of both groups relevant and comparable data?

• Which are the strengths and weaknesses of cross-sectional and longitudinal models respectively? How important are individual developmental differences for this type of study? In the cross-sectional model, it is not possible to take individual developmental differences into account.


 

ASSESSING L2 EMBEDDED VOCABULARY THROUGH OBJECTIVE COMPUTERISED TEXT ANALYSIS

Soraya Moreno

This doctoral research project focuses on the assessment of L2 embedded vocabulary through different electronic analysers.

We aim at proving that a single assessor can measure L2 vocabulary in written compositions through objective computerised text analysis in a valid and reliable way. We believe that these results will show a positive correlation with the manual assessment of L2 embedded vocabulary undertaken by two different raters.

In the last decades, we have seen how information technology has been introduced within the field of L2 assessment. Thus, we come across different measures of lexical richness in L2 written production which make use of computational tools, such as the Lexical Frequency Profile (LFP) (Laufer & Nation 1995), P_Lex (Meara & Bell 2001) and WordSmith Tools (Scott 1997), amongst others.

The same as vocabulary cannot be measured by a single test, since vocabulary knowledge is many faceted (Laufer 2001), we believe that we should approach objective assessment of L2 embedded vocabulary by taking into account different technological instruments, since they may provide different insights into lexis. Thus, amongst other things, we can: (a) analyse collocational patterns; (b) measure lexical richness based on infrequency of occurrence; and (c) match L2 vocabulary against frequency lists. By means of all that information, we may get a rounded picture of L2 embedded vocabulary assessment and analyse the advantages and disadvantages of using one or another computational tool; and moreover investigate whether they actually report any benefit as opposed to time-consuming manual scoring.

Thus, the following research questions are specifically addressed in this study:

1. Can we obtain a feasible assessment of L2 vocabulary in written compositions by using a single rater helped with objective computerised text analysis?
2. Will there be a positive correlation between the technological assessment of L2 embedded vocabulary and its manual scoring?
3. Do we obtain any significant advantage in using computerised vs. manual assessment in order to measure L2 embedded vocabulary?

As far as we know, our study differs from previous ones in the sense that: (a) we aim at exploring the insight that different computational tools provide into the assessment of L2 vocabulary; (b) we have not come across any investigation which has made use of the LFP, P_Lex and WordSmith Tools in order to assess the same sample of written compositions; and (c) most of the researches that have employed any of our assessing instruments have used them to measure the lexical richness of heterogeneous groups of university students, whereas our sample of informants comprise a rather homogeneous group of Spanish secondary learners of English as a foreign language.


 

METHODS FOR STUDYING LOW-LEVEL SECOND LANGUAGE (L2) LEARNERS’ CLASSROOM VOCABULARY ACQUISITION

S. Catherine Howell

Word learning is a crucial part of the second language learner’s task, and nowhere is vocabulary teaching and learning as important as in the low-level language classroom. Low-level students do not yet have the language resources to be able to easily understand or acquire new words they encounter in written or oral texts. Second language classrooms provide critical support for low-level learners acquiring the vocabulary they need to understand enough language outside the classroom to pick up new words from context. My dissertation addresses the beginner’s vocabulary paradox from the point of view of instructional practices in the low-level ESL classroom. This paper focuses on two methodological difficulties I have encountered in my research: describing intensive vocabulary instruction in a low-level classroom and assessing those classroom practices via measures of vocabulary learning.

Since low-level learners need exposure to multiple repetitions of new words in different contexts, they can benefit from maximally explicit, intensive vocabulary instruction. Two types of data are necessary to test this claim: low-level vocabulary instruction needs to be described in detail and student learning of the vocabulary taught in the classroom needs to be measured. However, research has generally focused narrowly on part of one or the other of these needs: either describing in detail particular classroom practices or pre- and post-testing the efficacy of isolated, individual tasks.

This research needs to be expanded in two ways. First, there is no adequate conceptual framework for describing and analyzing all the types of classroom tasks and interactions that provide contexts for multiple exposures to the same word. In a pilot study, I addressed this need by developing methods for tracking exposure to specific vocabulary items over time in a classroom. In particular, I focused on intensive vocabulary instruction, tracking words that were used more than ten times in a thematic unit. In my current work, I am refining this analytical tool to identify and better understand intensive vocabulary instruction in low-level classrooms.

Second, little research on classroom discourse has adopted a pre-test, post-test methodology, as intensive vocabulary emerges through classroom practice and cannot be predicted beforehand. In my current work, I am developing methods to measure students’ learning of vocabulary used intensively in class. In particular, I am pre-testing using the Vocabulary Levels Test (Nation 1983; Nation 1990; Schmitt 2000; Schmitt, Schmitt et al. 2001), to determine learners’ level of vocabulary knowledge based on the frequency of words in general English. After the pre-test, I videotape and transcribe classroom discourse data to identify words used intensively. I then develop and administer a post-test. The post-test includes the words used in intensively in class as well as a comparable number of words from the same general frequency level in English that introduced in class but used less than five times.

Although this method does not rule out learners’ prior knowledge of the target words, it will show if learners know more of the words used intensively than the words not used intensively. This information can then be used to analyze the effectiveness of the intensive instruction.

This talk will focus on my methodological goals. One is the development of a fine-grained understanding of the classroom practices that underpin intensive vocabulary instruction. The second is the development of methods for measuring changes in students’ knowledge of vocabulary in a classroom-based setting. My ultimate goal is linking measurements of student knowledge to classroom practices and actions. Open discussion with other researchers studying classroom-based vocabulary acquisition would help me to refine my methods for describing intensive vocabulary instruction, measuring student vocabulary learning, and assessing the effectiveness of different types of intensive instruction.

 

FORMULAIC SEQUENCES AND THEIR FUNCTION IN THE ACQUISITION OF SPANISH AS A FOREIGN LANGUAGE

Inmaculada Moreno Teva

One basic assumptions of this study is that “a language user has available to him a large number of semi-preconstructed phrases that constitutes singles choices, even though they might appear to be analysable into segments” (Sinclair 1991: 110). The formulaic sequence is defined as “a sequence, continuous or discontinuous, of words or other elements, which is, or appears to be, prefabricated: that is, stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar” (Wray 2002: 9). Fixed and, above all, semi-fixed formulaic sequences are presented by Wray as the frame in which language is constructed. Semi-fixed constructions are defined as having one or several open slots which differ in form and size according to speakers’ needs that vary from moment to moment and from speaker to speaker. The syntactic variability of stored multi-word items is accounted for by cognitive linguistic models such as construction grammar (e.g. Fillmore et al. 1988, Goldberg 1995) and usage-based models (e.g. Bybee 1998).

The main research questions are:

- Do different kinds of formulaic sequences prevail at different levels of acquisition of Spanish L2? and, in that case, what is their function in the acquisition of Spanish L2?
- Is there any kind of instruction that favours the use of formulaic sequences by Spanish L2 learners?
- Is there any difference in the use of formulaic language in conversation or in writing?
- What are general and what are individual patterns in the acquisition and use of formulaic sequences in Spanish L2 learners?
- Does the use of formulaic language by the learners reflect a higher degree of identification or involvement with the culture of the target language?
 

Our empirical material consists of spoken and written production of Swedish students of Spanish L2 taken from the AKSAM corpus (discussions and negotiations) and from the CorALEs corpus (conversations, interviews, re-tellings and written production) recollected by the Spanish department of the Stockholm University.

 

PROBLEMS IN ANALYSING DATA OF LANGUAGE LEARNING BEGINNERS

Susan Sayehli

The aim of my research is to study the order in which certain syntactical and morphological structures emerge in L2 acquisition. Special consideration is given to the influence of the first or other previously learned languages on this process. The theoretical frame of my work is Processability Theory (PT; Pienemann 1998) from which hypotheses will be derived and tested. To this purpose spontaneous oral data is collected in a pseudo-longitudinal study from Swedish native speakers with English as their first foreign language who are learning German in a classroom context. All informants are language learning beginners, even though they have been studying German to different degrees coming from four different grades. They represent thus a range of beginner proficiencies.

Language learning beginners naturally find it very hard to express themselves in the target language. Their talk typically abounds of long pauses, self-corrections, gap-fillers, new starts and code-switches. A first but crucial step in the analysis of especially spontaneous data is to decide what to include into a further syntactical or morphological analysis. The decision is at best guided by carefully considered criteria, which are consistently applied. However, selecting criteria there is the danger to unconsciously favour those that are fitting the respective theoretical frame. The overall question then runs, which empirical/statistical/methodological and theoretical arguments there are to support the use of the respective criterion. Is there a methodology to reduce the ambiguities? The question of criteria selection turns up in different shapes in a database. The following areas are particularly interesting analysing oral, natural occurring data of very beginners of language learning in a trilingual context:

1. In order to study word order, it is crucial to differentiate new starts on the one hand from pauses and on the other hand from self-corrections, if one wants to delineate an utterance. How can they however be differentiated? Furthermore, especially in morphological analysis, what part of these newly started or self-corrected utterances can reasonably be included in the analysis?

2. Oral production is characterised by a tight time frame that is given for its production. The planning time is only a fraction of seconds as well as the possibilities for monitoring the output are limited. Applying a processing theory, which refers to the automation underlying language production and as such only applies to oral production, it is important to distinguish automation from more monitored production. Utterances of language learning beginners are typically full of pauses and are characterised by a slow speech rate. How is this to judge in respect of automation? Can utterances with a low speech rate, self-corrections and long pauses still be assumed to involve underlying automatic processing procedures? Are there cognitive/psychological methods to further examine this question?

3. There are language samples of, and code-switches between, at least three different languages to be expected. How can it be decided which language (L1, L2 or LN) was the target language? And hence which utterances shall be included in the analysis of German Interlanguage?


 

THE BILINGUAL FOREIGN LANGUAGE LEARNER -IMMIGRANT STUDENTS LEARNING GERMAN AS A FOREIGN LANGUAGE

Johanna Klawitter Beuch

The goal of my doctoral thesis is to investigate what role bilingualism and individual differences play when learning and understanding a new foreign language (i.e. an L3, L4 etc.). During one “school year” (from September 2001 to June 2002) I followed two groups of students, one monolingual and one bilingual, learning German as a foreign language at three different high schools in Sweden. The students completed German tests and questionnaires created for my research project. Approximately one third of the students was also interview and asked how they solved the German tests. The method used was a mixture between an interview and a think aloud protocol.

In the questionnaires the students gave information about family / social background, grades in other foreign languages, maths, science and Swedish, estimated knowledge in the mother tongue for the bilingual students, estimated knowledge in German, motivation, attitudes and language anxiety and learning strategies and techniques.

The language tests all dealt with German vocabulary problems . They ranged from quite basic, e.g. recognizing German words among other foreign words, to rather complicated, e.g. trying to put a “scrambled” German text in the right order.

I also carried out a similar research project in Sydney, Australia, where I spent last semester as a visiting academic at University of New South Wales. The results from this study are not yet evaluated. I am not sure if it is possible to integrate this project as a part of my thesis or if I will try to publish the results separately in an article.

I have encountered a number of theoretical and methodological problems when working with my research and thesis writing. I am now somewhere in the middle of the doctoral programme in the department of German and Dutch at Göteborg University (I have completed two out of four years), which means that I am in the process of writing the theoretical part and evaluating my results. I am also working on a theoretical framework that consists of the theories from a number of researchers, among others Gardner (The socio-educational model of second language learning) and Bialystok’s theory about metalinguistic knowledge.

What I would like to do at the conference is to present my project and some of the difficulties that I struggle with. These concern some areas of the theoretical part of my thesis e.g. bilingualism, strategies and metalinguistic knowledge, but also the empirical part, e.g. the evaluation of the think aloud protocols. It would be an honour to talk to Ellen Bialystok about my project since she has great knowledge in all these areas.



 

VERBAL MORFOLOGY AND LEARNING STRATEGIES IN THE INTERLANGUAGE OF SOME SWEDISH LEARNERS OF L2 SPANISH

Rakel Östeberg

Why do learners of Spanish L2 frequently use stereo-typed verbal morphology ? And why do often learners create certain functional words or structures, for instance a double form of the verb (soy hablo español) in order to find an equivalent to the swedish personal pronoun?

I will try to show how these phenomena, and others, are related to factors in the formal learning context (FLC), such as:

- conditions and conventions which govern the work in the classroom in an explicit and implicit way.
- the kind of teaching, the learner’s attitude toward SLA, the general attitude toward SLA in society, reflected in the political plans (curricula) which establish and control the national goals of the language teaching.
- what is awarded in tests.
- personal motivation.
- the degree of identification with the target language.

.
The main hipothesis of this work is that the formal acquisition of L2 Spanish in high degree is governed by the FLC.

The purpose is to contrast the use of stereo-typed verbal morphology and certain cognitive strategies to the FLC. Firstly some different learners profiles will be described, secondly I will show the relationship between different kinds of FLC, for instance the degree of identification with the target language, and the use of different kinds of learner strategies which in this example could be the overuse of the infinitive, by applying both qualitative methods and co relational techniques, also called multivariate statistics.

I will investigate the influence of the FLC through a semi-longitudinal study in which I will study both written and spoken interlanguage produced by ca 30 young Swedish learners in two highschools. The two highschools are chosen since the applied pedagogical methods differ and thus the FCA. Deep interviews will be held with some of the students in order to grasp both their personal goals and experiences of SLA as well as their attitudes toward language learning. Quantitative inquieries on the FCA will also be used.

I intend to use a group of control with learners who have acquired Spanish mainly in an informal context in order to evaluate the results of my investigation.

I would like to discuss the design of the corpus and the methods since the matter is socio-linguistic and a combination of different kinds of methods is needed. How could I define and measure the different components ( such as the not target like use of infinitives and the degree of identification with the target language) of the qualitative case study? How could these elements be integrated in a multivariate statistic?

It would also be interesting to focus on the main features of SLA in formal contexts. It is clear that cognitive strategies and universal features govern the interlanguage and the order of acquisition. My intention is to take a further step and investigate the influence of the learner’s context. In other words, a phenomenon might have two explanations, firstly it could be understood as a cognitive feature, secondly it can be highly related to an external factor in the learner’s context.



 

THE MULTI-FEATURE HYPOTHESIS: EFFECTS OF TASK CHARACTERISTICS ON THE PROCESSING AND ACQUISITION OF L2 SPANISH VOCABULARY

Moonen, M.L.I.

Theoretical background

There is rather broad agreement that SLA to a large extend takes place by performing language activities (Wagner-Gough, 1975). In this view SLA is being seen as a function of the way input and output are processed (Long & Crookes, 1987). In this context the presence of feedback, modified output and attention to form is supposed to be essential. And that is why interaction (‘negotiation of meaning’) is thought to be crucial. These ideas have resulted in a so called ‘task-based approach’ (TBA) (Bygate, Skehan, & Swain, 2001).

Although there is some theoretical underpinning e.g. in Long’s interaction hypothesis, there is no explanation for the assumed role of these criteria in terms of a more generic learning theory. Some insights from cognitive psychology might be helpful in this respect. According to this insights, the product of a cognitive learning process should not be perceived as a template, but as a more or less open mental structure of neural units (Bereiter, 1991; Anderson, 1995). The network structure of a concept is distributed over many units, each of which can also participate in the representation of other concepts (Gasser, 1990). The term linguistic and non-linguistic ‘features’ is often used for these units. A concept consists of a distinctive combination of features. The essence of a concept is not in the units, but in the combination in which they are activated. Activation of a network can start from any connected feature, dependent on the type of stimulus that is received. The stronger the connection, the sooner and faster the activation.

Taking the Task Based Approach as a starting point, the multi-feature hypothesis (Westhoff, 2003) predicts that the SLA process will be more effective when learners perform tasks that focus on (a combination of) different features of a concept in the foreign language. It is expected that tasks that elicit mental actions involving:

-more features,
-more different categories of features,
-with more frequency,
-in more frequent combinations and
-simultaneously,

will have a more positive effect on the acquisition of the concept in the second language. A learner that performs a task that is adapted according to this hypothesis, is expected to perform more mental actions and to operate upon more and different features etc. of the concept. Therefore, the learner is expected to demonstrate a more effective second language acquisition process.
During the PhD project, the hypothesis will be tested in authentic classroom settings (Spanish and French), with original textbooks and different task types/skills (reading, conversation, vocabulary, etc).

Machteld Moonen, IVLOS, University of Utrecht (m.l.i.moonen@ivlos.uu.nl)

Description Pilot Study, Spring – Autumn 2004

The multi-feature hypothesis is tested in a study on the effectiveness of ‘original’ and modified textbook tasks on the acquisition of Spanish vocabulary by students in the lower levels of secondary education. It is expected that tasks that are modified according to the multi-feature hypothesis will have a more positive effect on the acquisition of vocabulary than the original textbook tasks.

The textbooks and materials used in the course will be analysed using the checklist that is being developed at this moment. Important aspects of the checklist are the nature of the tasks (e.g. producing formulaic speech, content oriented processing of input), the mental actions that are elicited (e.g. categorizing, elaborating) and the features that are addressed (e.g. morphology, syntax). Tasks that do not ‘force’ students to operate on more, different etc (linguistic or non-linguistic) features will be modified. This will be done using the checklist. One group will carry out the original tasks, and the other group will do the modified tasks. The students belong to the same year and level and have the same teacher.

In the pilot study that will be carried out this spring, think-aloud protocols will be used to study the quantity and quality of mental actions that the learners perform while working on the original and modified tasks. This part will focus on the learning process.

The second part of the pilot study (autumn 2004) will also focus on acquisition as the result of the learning process in a pretest-posttest design. These tests will be used to measure the performance on vocabulary tasks. In order to gain insight into the effects of recall and retention, tests will take place directly after the tasks and after a month.

During the rest of the PhD project, studies will be carried out in different classrooms on a larger scale (Spanish and French), with different textbooks and different task types/skills (reading, conversation, vocabulary, etc).

I would like to discuss the following methodological issues:

• The use of think- aloud protocols
• The design of the tests used as pretests and posttests
• Differences in ‘time-on-task’
• Treatment design: disentangling specific effects of independent variables and task types on the performance (on both think-aloud procedures and tests)
 

INTERLANGUAGE PRAGMATICS AND THE LANGUAGE OF CONFLICT. HOW POLITENESS IS ACQUIRED IN L1 AND L2

Doris Dippold

Interlanguage Pragmatics has only recently begun to adress the question of how pragmatics develops in a second language. In this paper, I want to investigate several questions in the area of second language pragmatic development which I intend to answer in my PhD dissertation.

Many researchers tend to explain NNS pragmatic failure through two factors: Learners’ transfer of L1 pragmatic rules into the L2, and teaching-induction. However, though weaknesses in students’ pragmatic socialisation through classroom discourse have been demonstrated, I want to argue that we must go beyond adherence to L1 pragmatic norms and teaching-induction to explain where learners’ inadequate pragmatic behaviour comes from. My hypothesis is that the cognitive development of L2 learners plays an important role for the development of pragmatics and that learners pragmatic performance shows characteristics that are similar to those of L1 learners at different developmental stages.

Disagreement is a speech event which is particularly adept to show developmental patterns in learners’ pragmatic performance. When people disagree, they usually pursue two goals: the transactional goal of persuading the interlocutor of their opinion, and the social goal of not jeopardizing the relationship with the interlocutor. New research on argumentative discourse shows clearly that, if learners are cognitively not ready to do so, they may not acquire new strategies whatever training or exposure is provided for them.
Both L1 and L2 learners, when involved in argumentative discourse, tend to be more focused on the content of their message than on message format or politeness (Kotthoff 1989; Felton, 2001). Whereas learners’ performance in the L1 improves with age, familiarity with the task proves more successful than direct instruction in the L2 (De Capua, 1998; Németh & Kormos, 2001). One single explanation works for both L1 and L2: Learners are able to pay more attention to the form of their utterance, because they have more cognitive space available to perform the task.

Though politeness norms differ cross-culturally, we can assume that both L1 learning children and L2 learning adults draw evidence from a universal pragmatic knowledge base, as proposed by many theories of politeness and communication, such as Grice (1976), Brown & Levinson (1983) and quite new proposals such as Spencer-Oatey and Jang’s (2003) Sociopragmatic Interactional Principles. For disagreement, this knowledge base would specify that disagreement must be linguistically marked in some way. However, how it is marked in specific languages remains subject to cultural variation.

It has been proposed that learners use both their L1 and universal pragmatic knowledge to acquire an L2, i.e. Bialystok’s (1993) language processing hypothesis. According to this hypothesis, learners’ primary task is one of achieving control over already available pragmatic knowledge in order to select contextually appropriate language for a specific context. From these hypotheses and observations follows my research questions:
When do L2 learners become aware of pragmatic rules and when do they apply them themselves? Do these rules reflect universal norms, L1 norms or L2 norms? What developmental stages do learners go through? Are these stages similar or different in L1 and L2 development? What is the role of teaching in pragmatic development?

Data shall be collected through the same discussion task carried out by groups of adult native speakers of German, adult native speakers of English and adult English learners of German at different proficiency levels. In addition, the L2 groups will also be asked to rate video clips for politeness in order to trace the development of pragmatic perception. All groups will subsequently receive instructional treatment in pragmatics, with the results of treatment tested through a further discussion task.

It is hoped that the proposed study with its developmental focus will enrich the field of Interlanguage Pragmatics with cognitive processing receiving stronger attention in the future than has been the case to date.
At the doctoral workshop, I would discuss problems relating to universal pragmatic principles, pragmatic development and data analysis, in particular:

• I suspect that there is a considerable overlap between universal norms of pragmatics and L1 / L2 norms. How can the pragmatic norms learners are resorting to at different developmental stages be identified?
• Is quantitative or a qualitative analysis more suited to analyzing the problem on hand? What is the most adequate method of transcription and analysis?
• It has been detected that the experimental design of language studies can lead to anomalous language production. How can the study be designed to enable learners to show their actual pragmatic competence?

Resource persons who could help me solve those problems are Ellen Bialystok, Juliane House and Gabriele Kasper (if available) or any other experienced researchers in the field of Interlanguage Pragmatics.


 

L2 PHONOLOGICAL ACQUISITION: THE CASE OF ITALIAN VOWELS IN THE INTERLANGUAGE OF MOROCCANS

Laura Mori

The aim of the present paper is to discuss some experimental phonetic data from cross-language production in Italian by Moroccan Arabic-speakers immigrated in Italy.

This is a first partial description which refers to a larger research project (author’s PhD thesis in progress) on the segment-based phonetic variation in the interlanguage of thirty Moroccans.

The final objective is to focus on the production of new phones, those which characterize the target language and are not in the native language, and to see how L2 phonological contrasts are produced as far as the vowel system and some consonants are concerned.

The starting hypotheses are based on a descriptive contrastive evaluation of the two phonological inventories involved in the acquisition process: the Moroccan Arabic phonological system (source language) and the Italian one (target language), taking in due consideration that the input is the regional Italian variety spoken in Latium where the speakers live.

As for consonant analysis we focus our attention on some case-studies, taking into consideration only those phones which are not included in L1 phonemics such as the stop /p/, the fricative /v/ and the affricate sounds (/t/, /d/; /ts/, dz/). We also analyse the production of the stop /k/ since in Moroccan Arabic (unlike Italian) the opposition between the velar stop (/k/) and the uvular one (/q/) is functional. Do these phonemic differences interfere in the interlanguage of Moroccans?

We suppose that these features could cause a number of phonological processes such as transfer-based substitutions, overlapping and lenition.

As far as the vowels are concerned, in the L1 vowel system both quality and quantity contrasts exist (short and long vowels): according to experimental studies (Hamdi, 1991 ; Barkat, 2002, Al-Tamimi/Barkat-Defradas, 2004) Moroccan Arabic vowel inventory comprises the three Classic Arabic long vowels (/a:/, /i:/, /u:/) the back /u/ and the central vocoid //.

Does this vocalic trapezium affect the realization of L2 vowels where only quality contrast is pertinent? Furthermore, from a quality point of view, Italian has a seven-based vowel system compared to the five-based Moroccan Arabic one. Does phonological interference occur in the production of the Italian medium vowels (e/;o/) that aren’t included in the L1 system?

Our hypothesis is that speakers tend to use their native vowel system: this means that they could categorize vowels according to quantity opposition and avoid producing medium vowels by making extensive use only of their native vowels.

The objective of the present paper is to describe experimentally the interlanguage vowel system of a sub-group of ten male speakers presenting both temporal measurements and formant values compared to those reported in literature for Italian and Moroccan Arabic spoken by natives.

The reference corpus (ArabIt) has been collected by the author for her PhD thesis and it comprises a collection of thirty semi-spontaneous interviews led with sixteen males and fourteen females (average age: 35) living in Italy (Latium region) from a different number of years.

ArabIt comprises a representative collection of Italian spoken by a sociolinguistic diverse community: various social backgrounds, degree of literacy, heterogeneous age of arrival and length of residence in Italy, asymmetric language use of L1 and different competence of Italian as L2 and sometimes as L3 (in the case of cultivated people who have a bilingual competence of French since their childhood).

The only common feature concerns their linguistic competence: all the speakers have Moroccan Arabic as first language and have learnt Italian as second (or third) language in a spontaneous context, not as a foreign language taught at school.

The corpus comprises speech elicitated in many ways in order to comply with different needs: the semi-guided conversation on biographical themes (migratory experience, language-use in their daily life, personal opinion on cultural habits and differences, etc..), story-telling through sequence of images and word list through visual stimulus.

Though we took care of speech quality and used a DAT recorder with a head-microphone, recordings took place in different environments (speaker’s working place or house, at the Mosque, at the University or at cultural association sites) according to speakers’ availability in order to minimize the artificiality of the recording situation and obtain pieces of spontaneous speech.

OLD ISSUES, NEW EVIDENCE: PARAMETER RESETTING IN NON-NATIVE LANGUAGE ACQUISITION

Junkal Gutierrez

In this paper we present evidence showing that parameter resetting is possible in non-native language acquisition (L2A). We report on the results from an experimental investigation eliciting English Long Distance (LD) wh-questions (see sample (1)) from L1 Basque/Spanish bilingual subjects. Crucially, both our subjects’ L1s (Basque and Spanish) and the Target Language (TL) form LD wh-questions by moving the wh-phrase from the embedded clause to the matrix. That is, they share the same parametric option for LD wh-questions. Under most current proposals it is basically assumed that L2 parameter resetting takes place when the L2 input cannot be accommodated by the subjects’ L1 grammar. It follows that production of questions featuring a parameter setting other than that present in the L1s and TL of our subjects is not expected. Interestingly, we found that some of our subjects produce non-adult questions featuring a ‘new’ parameter setting possible in other natural languages such as Bavarian or Hindi among others. These non-adult questions involve two related wh-phrases, one in the matrix CP and the other in the embedded CP (sample (2)). Furthermore, the same type of non-adult questions were reported for the acquisition of L1 English in Thornton (1990). From these non-adult questions we conclude that at least those subjects who produce these questions in L2 English start out with a parameter setting distinct from that instantiated in their L1s, and TL. The evidence presented argues against a view of L2 initial states involving L1 transfer, direct access to L2 parametric values or no access to UG hypotheses. We also present an alternative hypothesis which tries to accommodate these data, allowing access to UG and parameter resetting in non-native acquisition.

(1) Who do you think he saw?

(2) What do you think who he saw? / Who do you think who he saw?

 

VOWEL REDUCTION FOR THE NON-NATIVE LEARNER

Esther Gómez Lacabex


An experimental study is described which intends to contribute to the knowledge of the relationship between perception and production and the effect of pronunciation instruction in the foreign language classroom. These two issues are investigated in the acquisition/learning process of a phonological element, vowel reduction in a second language formal acquisition setting.

Vowel reduction is a characteristic phonological process in English. Along with intonation and rhythm, vowel reduction is an unintelligibility renderer as authors like Gimson, Flege, Cenoz and García Lecumberri or Kohler have suggested. English vowel reduction is not always phonetic, optional or variable (depending on style, speed and due to ease of articulation in connected speech). It is often phonological: the reduced vowel is part o f the phonological structure in many English words (mother, national etc.) and, thus, it may become contrastive as in “violet”  vs. “violate”.

We wish to contribute to the investigation of the relationship between production and perception within the area of second/foreign language formal acquisition.While some authors refer to these elements as independent from one another, there are views which see this relationship as a dependent one. Our work intends to look into the relationship between these two elements within the context of formal language learning, seeing this relationship as mutually facillitative.

Additionally, this work aims at analyzing the effect of including pronunciation instruction within the foreign language classroom. This aim arises from the frequent finding that in English as a foreign language formal instruction, oral skills (speaking and listening) lag behind written skills (reading and writing), as has been noted consistently in the work of the Research in English Applied Linguistics (REAL) group (University of the Basque Country) (Cenoz & García Lecumberri, 1999) and has also been given attention by other authors such as Elliot (1995) and Ferguson (2000).

The main research questions that we propose to address are:

1. What is the relationship between preception and production in the context of formal acquisition? Is this relatioship a dependent one or is it rather mutually facilitative?
2. Is vowel reduction a phonetic or phonological characteristic to be learnt/acquired by the non-native learner?
3. Is the training in vowel reduction effective? Are there differences between the instruction based on perception and the one based on production?
 

Our methodology will consist of 3 different steps:

Pre-test
A corpus of minimal syllable/word pairs ( reduced vowel- full vowel in closed and unstressed syllable) of the type: robot /"rObOt/ vs. abbot /"{b@t/ or violate /"vaI@leIt/ vs violate /"vaI@l@t/ will draw the students perception and production in syllable level, word level and connected speech level.
 

Instruction period
Three groups will undergo three different training periods:
Experimental group 1: pronunciation instruction based on production
Experimental group 2: pronunciation instruction based on perception
Control group : no pronunciation instruction
 

Post-test
The same tasks given to the students in the pre-test will enable us to draw conclusions about the effectiveness of the instruction period.
 

Possible areas for discussion:
 

1. How to design an instruction period based on perception of schwa in the foreign language classroom and how to do likewise in the case of instruction of the production of schwa.
2. Is vowel reduction phonetic of phonological for the native leaner? That is, is it an internalised full vowel with realization constraints due to frequency, ease of articulation, etc.. or is it a phonological category internalised as such?
3. Is it always the case that perception must lead production in the language acquisition process? Is this the case in foreign language formal leaning?