EESE 7/98


Lingua Franca English
Characteristics of successful non-native-/ non-native-speaker discourse1


Christiane Meierkord (Erfurt)


  1. Introduction
  2. The Data
  3. Methods and Units of Analysis
    1. Analyzing Discourse Structure
    2. Analyzing Politeness Phenomena
  4. Results
    1. Lingua Franca Discourse Structure
      1. Conversational Phases
      2. Topics
      3. Turns
      4. Pauses
      5. Simultaneous Talk
    2. Politeness Phenomena in Lingua Franca English
      1. Ritual Illocutions in Openings and Closings
      2. Back-Channels
      3. Other Gambits
      4. Choice of Topics
  5. Summary
  6. Bibliography
  7. Key to Transkription Symbols
  8. Appendix
  9. Sound files available

1. Introduction

The growing globalization of the world's economic markets, increased travel opportunities and better communication facilities have created a situation in which people from different linguistic and cultural backgrounds need to communicate with each other. When speakers need to resort to a language which is the mother tongue to neither of them, they use a lingua franca. English is one of the languages that are used as a communicative tool in these situations. Unlike native / non-native communication, this discourse type, which shows both interlanguage as well as lingua franca traits, has up to now received only little attention.2 English used as a lingua franca is characterised by features which are typically found with learner languages and interferences with the speakers' mother tongues as well as by strategies to simplify the target language. At the same time, speakers bring their own culture's communicative norms and styles into lingua franca communication. Thus a new speech community arises as Hüllen (1982: 86) points out:

What does communication in this speech community look like? How does conversation differ from that of native speakers, when non-native interlocutors need to communicate with restricted linguistic resources, and when at the same time they operate in a situation where each one is uncertain about the others' norms and values. This article summarizes the findings of field-work carried out in 1991/1992, examining non-native speakers during small talk conversations.3 Considering both discourse structure and politeness phenomena, it attempts to characterise the pragmatics of a variety of English about which Kasper (1981: 80) wrote:

2. The data

The corpus underlying this article comprises 23 small talk conversations with a total of 13.5 hours, which were tape-recorded in a student hall of residence in Great Britain. However, this is not a homogeneous corpus, in that some of the factors influencing communication, i.e. the demographic characteristics of the speakers, have not been nor could be controlled. The participants of the single conversations were of both sexes, aged 20 to 30 with a range of 17 different mother tongues. They were grouped together into broad cultural groups: European, African, Arab, Indian/Pakistan and Asiatic speakers. A further factor was the speakers' communicative competence, which may be understood as their competence for using the English language appropriately,5 and speakers were divided intuitively into more and less competent speakers, motivated by a definition by Zimmermann (1984: 100):

Each conversation was influenced by the different speaker variables, which cannot be isolated for analysis, i.e. it is not possible to make statements on the linguistic performance of e.g. competent male speakers or female Asian speakers. However, this was not the aim of the study, which - as stated above - was rather to characterize lingua franca English in more general terms. Tables presenting the distribution of speakers regarding the above factors are in the appendix. The data resulting from my corpus were compared - wherever possible - to the results that research on native speaker as well as on learner discourse has yielded. The data used for comparisons were Oreström (1983) for turn-taking, Bublitz (1988) and Schneider (1987 and 1988) for topic development and choice, Kasper (1981) for gambits and Edmondson and House (1981) for the identification of back-channels, gambits and illocutions.

3. Methods and units of analysis

The method used to describe lingua franca English involves an adaption of quantitative conversation analysis. Even though it can clearly be argued that quantitative analyses alone cannot sufficiently explain communication and needs to be supplemented by qualitative analyses, I chose to work within the quantitative paradigm, as the native speaker data, which were available for comparison had been analyzed in the same vein.

3.1. Analyzing Discourse Structure

Discourse is considered - for the purpose of this analysis - to consist basically of an opening, a core and a closing phase. In the opening phase, speakers come together and establish the conversation circle. Talk is assumed. During the core phase interlocutors talk about at least one topic (see below). After the core phase has been finished, i.e. the last topic to be discussed has been finished or abandoned, talk is brought to an end in the closing phase during which speakers prepare to separate again.7 The core phase itself may at times be interrupted by what I call entry and exit phases, i.e. phases during which a further speaker joins the group or one of the interlocutors leaves it.8

Usually, the use of special illocutions9 is required to link the opening or closing phase to the core phase. These are (cf. Edmondson/House 1981: 210) REMARKs, identified to pave the way from the greeting into the core phase and EXTRACTORs and OKAYs which lead the interlocutors into the closing phase of a conversation. Whereas EXTRACTORs anticipate the withdrawal from a conversation which has not really reached its conclusion (e.g. Well, I must be off.), OKAYs (e.g. Okay. Right.) sum up the results of the core phase. The core phase of the conversation can be further subdivided into the different topics speakers are concerned with. However, the concept of a conversational topic has so far not been defined precisely, and this may well remain an impossible task.

Topics will vary with regard to their length as well as their depth, i.e. they may be treated superficially, but at the same time may be developed into several subtopics and in such a case show a high involvement of the speakers into what is being discussed as the current point of talk. Inside the individual topical units, the structure is composed of the individual turns, which speakers contribute to the topic. Discussing turns will also involve pauses and simultaneous speech:

3.2. Analyzing Politeness Phenomena

Modifying Koike's (1989) definition, the term 'politeness phenomena' - for the context of this article - is used to refer to all those linguistic phenomena that are employed by speakers to establish, deepen or maintain a positive social relationship with her interlocutors. Basically they comprise routine formulae or ritual illocutions, which occur during the opening and closing phase of a conversation, as well as back channels and other gambits that appear in the core phase. For the purpose of this article, choice of topic is also regarded as being a politeness phenomenon. By using specific illocutions during the opening phase, such as GREETs (e.g. Hello there.) and HOW-ARE-YOUs (e.g. How's it going?) (cf. Edmondson und House 1981), speakers establish a relationship with each other. Ritual illocutions or routine formulae differ from all other illocutions in that they "cannot be said to inform about, or express a speaker's attitude towards some external event or state of affairs". (Edmondson and House 1981: 59). They occur at highly predictable places in a conversation and their meaning is socially as well as pragmatically conditioned by the actual situation. At the same time, they provide speakers with a comfortable means of assuring pragmatic acceptability:

This relationship between the speakers is deepended and maintained during the core phase of the conversation by the use of gambits and back channels. Gambits can best be defined as instances of the speaker's speech, which have no interactional function. A gambit needs to be contrasted with an utterance, which "plays a part in building up the ongoing conversation, being significant both with respect to what has already been said in the conversation, and with respect to what may follow in the conversation" (Edmondson and House 1981: 36). Gambits instead "clearly 'support' the speaker, and show interest, understanding and so on, but do not in themselves contribute towards the structure of the conversation." (Edmondson and House 1981: 61). Gambits can be subdivided into cajolers (e.g. I mean, you know, you see), underscorers (e.g. look, listen, the point is), appealers (e.g. question-tags, right, okay) and starters (e.g. well). Gambits enable the speaker to involve her listener(s) in her utterances. She shows concern for her listeners' feelings and opinions and this makes it easier for the listener to react.

The listener on the other side signals her attention and interest by uttering back channels. Back channels are sometimes subsumed under the heading of gambits. However, I feel they should be treated separately, as they are - unlike the other gambits - associated with the listener's role.

Oreström distinguishes between supports (e.g. Mhm, yes, yeah, I know, I see), exclamations (e.g. Oh, God, Bloody hell), exclamatory questions (e.g. what, really, did he), sentence completions and restatements. The choice of non-offending topics avoids face-threatening situations. This is especially important in intercultural communication situations as mostly interlocutors are insecure about what constitutes an acceptable topic in their fellow speakers' culture. In the closing phase, just like in the opening phase, ritual illocutions such as LEAVE-TAKEs and WISH-WELLs serve the purpose of maintaining the relationship beyond the individual conversation.

4. Results

4.1. Discourse Structure
4.1.1. Conversational Phases

Speakers generally failed to link opening / entry and closing / exit phases to the core phase of a conversation with any of the illocutions mentioned above. There is a general failure to produce suitable REMARKs to move from the opening phase into the core phase, which means that the core phase is entered quite abruptly as is demonstrated in example 1. In lines 1 to 3 a new speaker enters the group and is greeted by those already engaged in talk. After this short entry phase, a long pause occurs. Without any move being made to link the opening to the core phase, Shiraz (male, Pakistan, competent) proceeds by introducing a new topic, the planned refurbishment of the student hall of residence.

EXAMPLE 1 12

Illocutions providing a smooth transition from one phase of the conversation into the other are usually also missing at the end of a talk. Especially extractors, which would have been expected when a speaker withdraws from an ongoing conversation, were hardly ever used. Rather, speakers made a long pause to indicate their desire to end a talk. When EXTRACTORs do occur, there is a high chance of their being misinterpreted as example 2 shows.

EXAMPLE 2

This example presents an elaborated exit phase initiated by Tsu (male, Malaysian, competent) with an EXTRACTOR in line 7. He obviously wishes to indicate that his meal is now finished. Apparently though, his interlocutors fail to react appropriately, or simply do not realize his intention. Thus, one possible explanation for the low use of EXTRACTORs - besides the speakers' inability to produce them - may be fear of misunderstanding. After further eight seconds, Tsu clearly marks the exit phase with a LEAVE-TAKE in line 9.

4.1.2. Topics

Speakers displayed a preference for short topics, dedicating an average of six to ten turns to the individual topics. Only in relatively few instances i.e. 20% of all cases, more than 15 turns were spent on a single topic. Of a total of 324 topics, 274 were completely changed after their completion. There were only 50 instances, in which topics were shifted resulting in the elaboration of more complex topics with single sub-topics.13 When topic shifts occurred, again half of these elaborated topics contained just one topic shift. Only very rarely (i.e. in eight cases) do sequences of related topics extend over more than three topics. It was not possible to show any correlation between speakers' sex, culture or competence and their willingness to introduce, shift or change a topic. Different interpretations may hold for the short length of topics: speakers may deliberately change to a new topic as they are uncertain about the latter topic's acceptability. But they may also need to give up a topic as they have to realise their own or their communication partner's inability to cope with a topic because of language problems.

4.1.3. Turns

For the analysis of turn-taking,14 simultaneous talk and pauses, a total of 2.412 turns were chosen from the corpus.15 Individual turns were on average 8.5 words long, but showed a lot of variation, with lengths ranging from one to 129 words. Turn length correlated with the speakers' sex, L2-competence and cultural background, displaying significant differences, which have been summarized as follows:

 mean
female speakers7.3
male speakers9.4
competent speakers10.0
less competent speakers6.8
European speakers9.3
African speakers10.2
Arab speakers9.1
Pakistan speakers10.2
Asiatic speakers6.8

Female speakers' turns are on average shorter than those of the male speakers. Also, less competent speakers' turns - as we would expect - are shorter than is the case with competent speakers. It is, however, not an easy task to interpret the calculated figures, as gender, culture and competence are factors, which influence one another. Thus the groups of speakers from Pakistan or African countries have a higher share of more competent speakers than the three other groups. Also, male speakers in this corpus tend to be more competent than females.

Even though the overall mean may obscure the existing heterogeneity, it serves the purpose of enabling a contrast with the native-speaker-data computed by Oreström (1983). Table 1 presents a comparison of the figures.

turn
length

native speakers

(Oreström 1983)

non-native speakers

(this study)

mean

mode

median

28,2

1 - 10

11 - 20

8,5

1 - 10

1 - 10

Table 1

A first look at the figures reveals a very highly significant difference between the mean turn lengths of native speakers and non-native speakers, native speakers' turns being more than three times the length of the non-natives' ones. It is however problematic to compare the present data to the ones found in Oreström (1983), as the conversations he analyzed were of a dyadic nature, whereas the greater part of my data comes from group conversations. This alone may account for a certain difference of figures. Unfortunately, data on group conversations among native speakers have not been available for comparison. As the mean may be a less suitable figure for conversational data as Buttler (1985: 33) explains, mode16 and median17 have also been calculated. These display a high similarity to the native speakers' figures. In any case, the non-native speakers' turns are shorter than those produced by the native speakers.

4.1.4. Pauses

Pauses occur - on average - after each 26 to 50 words. The mode of the pauses' length is two, the median four seconds. More competent speakers generally made less pauses, which indicates that a main explanation for the amount and length of pauses is the production problem of the speaker as is apparent in example 3.

EXAMPLE 3

Maria (female, Spanish, less competent) tells her interlocutors about the job she has just found as an au-pair. She apparently struggles with her vocabulary, which is signaled by the pauses she produces when seeking for words and phrases. The fact, that the pauses displayed in the above example probably stem from production problems of a less competent speaker and thus indicate that she takes some time to plan or rearrange her turn, is further corroborated by the hesitation phenomena, such as eh and uhm, the cut off sentence in line 1 and her generally incomplete sentences. Much different from the use of pauses as in example 3, are the pauses often employed by more competent non-native speakers with a specific communicative goal, i.e. to trigger a reaction from the interlocutors. This deliberate use of pauses can be seen in example 4.

EXAMPLE 4

In his explanation about an exam, doctors from overseas need to pass to be permitted to practice in Great Britain, Ibo (male, Nigerian, competent) pauses for short instances to give his interlocutors a chance to either indicate their understanding by a back-channel, or to utter problems in understanding and ask for clarification. It is important to note that his pauses occur immediately after a certain piece of information has been given. Thus they sort of structure the whole information process and divide its context into smaller, more-easy-to-process, units. He thus makes sure that even less competent participants of the conversation will understand what he says, contributing to a cooperative atmosphere.

Of all the pauses analysed, 484 occurred between speakers' turns. Of these, a majority of 58 % led to turn-taking, whereas 42% did not. There is, however, no significant difference in the length of pauses leading to turn-taking and those, which did not. Pauses in-between turns furthermore have an influence on the continuation or discontinuation of a topic, as can be seen in example 5. Pauses are often employed to indicate the wish to terminate or change a topic. In this they substitute verbal means normally found in the speech of native speakers.

EXAMPLE 5

In line 5, Anja (female, German, competent) tries to continue the topic "address forms in Urdu", which was the topic in lines one to three, after a pause of seven seconds. However, Shiraz (male, Pakistan, competent) had already planned to introduce a new topic. Now, two different topics are available for continuation. Shiraz clearly demonstrates that he is not willing to accept the previous topic, by taking the turn after a short period of simultaneous talk.

4.1.5. Simultaneous Talk

In 11% (i.e. in 191 cases) of the turns analyzed, interlocutors talked simultaneously for a while. Simultaneous talk mainly occurred at grammatical boundaries (in 54 cases), but also at pauses within grammatical units (in 32 cases). Especially the last figure is lower than expected and may imply that pauses occurring within a single turn are less often misinterpreted as turn-taking signals as prior research for native / non-native discourse has suggested.18

The average length of simultaneous talk was 2.1 words. Thus, overlaps were significantly shorter than those calculated by Oreström for native speakers of British English. In his corpus, the mean length of overlaps was three words. This difference in length may be due to the non-native speakers' insecurity about their interlocutors' cultural norms regarding simultaneous talk - which always implies interrupting the other party -, so that they rather refrain from longer overlaps.

Simultaneous speech may either occur by chance or be deliberately employed by a speaker as a turn-taking device. In both cases, the listener signals her wish to take the turn, but in the former, she reluctantly gives up her attempt until she can safely assume that the speaker has finished her turn, or she completely abandons her attempt to take the turn. Simultaneous talk in these instances is unintended, as can be illustrated by example 6. The sequence taken from an informative talk on dinners, Tsu (male, Malaysian, competent) has to attend to be admitted to the barrister's exam, shows the typical kind of a rather short sequence of simultaneous talk followed by turn-taking.

EXAMPLE 6

In lines 1 and 2, Hashif (male, Pakistan, competent) requests information on the consequences of failure to attend these dinners. His turn spans two moves. With the first one Hashif requests information about what the consequences of not attending the dinners are, and with the second one he asks whether attendance is compulsory. Tsu obviously did not expect the second question and in line 3, he reacts with an answer after Hashif's first move. When he realises that Hashif has not yet ended his turn, he cuts off his utterance, proceeds the second move of Hashif and answers this second question first. Whereas this period of simultaneous talk was not planned by either of the interlocutors, the overlap in example 7 is a specimen of its strategic use.

EXAMPLE 7

Ibo (male, Nigerian, competent) and Mina (female, Kenyan/Pakistan, competent) both are sort of experts in their conversational topic, malaria. Ibo is a doctor, and Mina is from a country with a high occurrence of this illness. As Ibo points to its problematic treatment, Mina wishes to inform him about an endemic malaria virus, which does not require any treatment. To do so, she interrupts Ibo's turn (lines 1-12) three times (lines 5, 11 and 13). She always does so at a pause (lines 5 and 13) or at a grammatical boundary (line 11). When Mina realises that Ibo is not willing to relinquish his turn, she cuts off her utterance, until in line 13 she finally takes the turn after Ibo has completed his one in line 12. The example shows that the non-native speakers use simultaneous talk for strategic pusposes, but seem to obey constraints inherent in the intercultural situation, which cause them to refrain from longer stretches of simultaneous talk. Furthermore, this example is also of interest regarding the use of gambits in two of Mina's attempts to take the turn, i.e. in lines 5 and 11.

4.2. Politeness Phenomena

4.2.1. Ritual Illocutions in Opening and Closing Phases

During the opening phase of a conversation, native speakers of British English (cf. E/H 1981) establish their relationship by using typical illocutions such as GREETs, HOW-ARE-YOUs, WELCOMEs and DISCLOSEs. In the conversations observed, openings were generally short and mainly consisted of GREETs and HOW-ARE-YOUs only. The vast majority of these again were mostly stereotypically realised by one or two different expressions, e.g. GREETs usually came as a form of either Hello. or Good Morning. Illocutions typically employed to maintain the social relationship beyond a conversation, are LEAVE-TAKEs, WISH-WELLs and - when applicable - INTERRUPTORs. Here, speakers basically restricted themselves to WISHWELLs and LEAVE-TAKEs. These again hardly showed any variation. Speakers in general used the phrases Have a nice day. for WISH-WELLs and See you. for LEAVE-TAKEs.

These results may reflect an economic language acquisition behaviour. Thus, learners do mainly learn the one or the few forms, which ensure successful communication. On the other hand, the same restricted variation did occur with rather competent speakers, who had lived in Great Britain for several years. These may restrict themselves to using a well-known phrase to spare themselves and their interlocutors any misunderstanding. Another possible explanation is a phenomenon which has been referred to as fossilization, i.e. speakers may have ceased learning at a point when they considered their knowledge of English to be sufficient to meet their communicative needs and therefore fail to reach target language competence.

4.2.2. Back-Channels

Forty percent of all turns were either accomplished or followed by a back-channel from the listener. When back-channels occurred within a turn, up to three were used by the listener. Just as native speakers do, non-natives in the corpus used supports, exclamations, exclamatory questions, sentence completions and restatements as reactions to what the speaker had said. However, the distribution of these back channels in lingua franca English, which is summarized in table 2, is somewhat different form the figures found by Oreström (1983) for native speakers and by Kasper (1981) in native / non-native conversations.

back channels

native speakers

(Oreström 1983)

non-natives

(Kasper 1981)

this study

supports

laughter

exclamations

exclamatory questions

sentence completions

restatements

92,6

 

2,6

1,6

3,2

80,9

 

12,6

 

 

6,5

48,4

33,0

7,5

1,6

3,5

6,0

Table 2

The vast majority of all back-channels were supports (81.4%), which is almost the figure Kasper calculated for the non-native speakers in her study. At the same time the figure is significantly lower than the one of Oreström's study. It is important to note the extremely high share of non-verbal supportive back-channels, i.e. the use of laughter in the present data.19

The typical verbal back-channel occurring after a peace of information has been given by the speaker is displayed in example 8.

EXAMPLE 8

Each time Steffan (male, German, less competent) has done so, Meong Hee (female, Korean, less competent) signals her understanding of his utterance (lines 2,8, 10 and 12) and thus makes him continue. Her excessive use of back-channels may be triggered by her problem to understand Steffan in line 4. This over-use of back-channels has also been observed by Kasper (1981: 247) with German learners of English:

EXAMPLE 9

Apparently though, non-native speakers use less supports than native speakers do. On the other hand, they employ a high amount of sentence completions and restatements, as table 2 shows. By using sentence completions, speakers help each other to collaboratively finish sentences. This is especially helpful, when one speaker struggles to express her ideas. Restatements probably serve the purpose of checking / making sure that the other party has understood what a speaker wanted to express. Hence their high occurrence.

4.2.3. Other Gambits

The speakers in my data use cajolers, underscorers, appealers and starters, just as native speakers do, but to a very different extent. Whereas the underscorers' and cajolers' share in the total gambits is similar to that found with native speakers by Kasper (1981), starters and appealers are obviously used in a different way, as can be seen from figure 1:

Figure 1

Non-native speakers in the present study use starters to a much lesser extent than the native speakers in Kasper's study. Appealers, however, occurred three times as much. The high frequency of appealers and cajolers is a further indicator of the non-native speakers' attempt to support each other during the conversation. Example 10 illustrates the use of appealers, which Tsu (male, Malaysian, competent) uses in lines 1 (right) and 6 (you see) to involve Anja (female, German, competent).

EXAMPLE 9

The same cooperative behaviour can be seen in example 11. Here, Ibo (male, Nigerian, competent) and Hashif (male, Pakistan, competent) both behave in a similar way as Tsu does in example 10. Moreover, they also use cajolers in lines 3 and 9 (I mean) to soften their utterances.

4.2.4. Choice of Topic

Speakers chose situational and safe topics.21 The topics either referred to the immediate situation, such as comments on food and life in the student hall of residence; studies or careers of speakers; third persons; daily routine and information about the home countries of the speakers. Table 3 provides a breakdown of the corresponding figures.

 

conversational topics

 
   

actual situation (meals, hostel)

50

school, job

38

third persons

22

past and future events

19

information on home countries

19

speaker's / listener's person

15

Table 3

The non-native speakers' choice of topics is thus much similar to the choice native speakers make in informal conversations. Schneider (1987: 254) has identified the following areas suitable for choosing topics:

  1. the immediate situation, such as, PARTY, as the medium level or, with regard to its topic potential, the most neutral and noncommitting level (frame elements)
  2. the external situation, the outer circle, which represents the larger context of the immediate situation and could be called the 'supersituation'; its topic potential is the least limited of the three.
  3. the communication situation, that is, the interactants, as a subset of the immediate situation; the inner circle of the model. Topics range from comparatively safe ones like hobbies to taboo topics such as income, sex or even age.
  4. The starting point for small talk is always the medium level. The continuation which involves the least commitment is talking exclusively about the supersituation. However, in some cultures it is not acceptable not to show any interest in the identity of the partner. In this case only uncontroversial issues are selected.

Topics implying possible sources of conflict are only chosen seldom (e.g. politics and religion). Interpreted from an intercultural perspective, this may be because speakers try to avoid conflict as much as possible, as they are insecure about the acceptability of arguing in other cultures. Taking into account aspects of second language acquisition, the avoidance of these topics may also be due to the deficits (esp. vocabulary) in the interlocutors' second language. More competent speakers may prefer less complicated topics to help their interlocutors save their faces, i.e. not having to admit deficits in English.

5. Summary

The lingua franca variety employed by the non-native speakers on the one hand showed much similarity with the standard varieties British and American English for length of turns, simultaneous speech and back channel behaviour. On the other hand, characteristics generally attributed to learner language were also found. These include low variation in ritual speech acts, failure to produce certain speech acts - e.g. extractors - and preference of 'safe topics', such as talking about meals or university classes.

Specific characteristics of lingua franca English in the situations analyzed comprised the following: speakers make use of laughter to substitute verbal back channels and to create a friendly and cooperative atmosphere. Pauses serve to indicate topic changes and to mark the transition between different phases of a conversation. Speakers encourage each other by an extensive use of gambits, especially cajolers and underscorers, and back-channels with supports being their favourite choice. Conversations are built up collaboratively and speakers used a comparatively high amount of sentence completions and restatements. These characteristics do not result from interferences with the speakers' mother tongues. Rather, they are specific phenomena of this interlanguage lingua franca.

The results show that non-native speakers establish a special variety of English, which is effective in informal conversations like the ones analyzed. Due to their cooperative behaviour, speakers manage to communicate successfully despite their restricted linguistic means. However, breakdowns also did occur, though only four instances are in the corpus.22 Still, we are left with the question, why communication breakdowns are not solved. Apparently, this question cannot be answered without assistance from other disciplines - e.g. the application of the psychological concept of empathy.

So far, learners should be equipped with strategies to overcome their own productive problems and to react to their partners' difficulties. Learners should be trained and provided with the respective strategies they can use when facing own communicative problems, such as the use of synonyms or circumscription or direct and indirect requests for help. To be able to react adequately when their communication partners get stuck, they furthermore need to be trained how to react in these phases of a conversation. Such a training must include sociolinguistic information on different cultural norms, e.g. how directly may I address my partner's problem.

6. Bibliography

Blum-Kulka, Shoshanna (1989). "Playing it safe: The role of conventionality in indirectness". In: Blum-Kulka et al. (Eds.), 37-70.

Bublitz, Wolfram (1986). "Gesprächsthema und thematische Handlungen im Englischen". In: Burkhardt, Armin und Körner, Karl-Hermann (Eds.). Pragmantax. Akten des 20. Linguistischen Kolloquiums Branschweig 1985, 225-234.

------- (1988). Supportive Fellow-Speakers and Cooperative Conversations. Amsterdam und Philadelphia: John Benjamins.

Buttjes, Dieter (1991). "Interkulturelles Lernen im Englischunterricht". Der Fremdsprachliche Unterricht 25(1), 2-9.

Buttler, Christopher (1985). Statistics in Linguistics. Oxford: Blackwell.

Coulmas, Florian (Ed.) (1981). Conversational Routine: Explorations in Standardized Communication Situations and Pre-patterned Speech. The Hague: Mouton.

Edmondson, Willis and House, Juliane (1981). Let's talk and talk about it. München, Wien und Baltimore: Urban und Schwarzenberg.

Enninger, Werner. (1987). "What interactants do with non-talk across cultures". In: Knapp, Karlfried et al. (Eds.). Analyzing Intercultural Communication. Berlin und New York: Mouton de Gruyter, 269-301.

Firth, Alan (1990). "'Lingua Franca' Negotiations: Toward an Interactional Approach". World Englishes 9/3, 269-280.

Firth, Alan (1996). "The Discoursive Accomplishment of Normality: On 'Lingua Franca' English and Conversation Analysis". Journal of Pragmatics 26:2, 237-260.

Götz, Dieter (1977). "Analyse einer in der Fremdsprache (Englisch) durchgeführten Konversation". In: Hunfeld, Hans (Ed.). Neue Perspektiven der Fremdsprachendidaktik. Kronberg: Scriptor, 71-81.

------- (1980). Englische Gespräche Deutscher Anglistikstudenten: Ausgabe und Kommentar. Augsburg: Universität Augsburg.

Goodwin, Charles (1989). "Turn construction and conversational organization". In: Dervin, Brenda et al. (Eds.). Rethinking Communication. Vol. 2: Paradigm Exemplars. Newbury Park, California, London and New Delhi: SAGE Publications, 88-102.

Gramkow-Andersen, Karsten (1993). Lingua Franca Discourse: An Investigation of the Use of English in an International Business Context. Unpublished M.A. thesis, Aalborg Universitestcenter.

House, Juliane (1982). "Opening and closing phases in German and English dialogues". Grazer Linguistische Studien 16, 52-82.

Hübler, Axel (1985). Einander Verstehen. Englisch im Kontext internationaler Kommunikation. Tübingen: Narr.

Hüllen, Werner (1982). "Teaching a foreign language as `lingua franca´". Grazer Linguistische Studien 16, 83-88.

Hymes, Dell H.(1972). "On communicative competence". In: Pride, J.B. and Homes, Janes (eds.). Sociolinguistics. Harmondsworth: Penguin, 269-293.

Kasper, Gabriele (1981). Pragmatische Aspekte in der Interimsprache. Tübingen: Narr.

Keller, Eric (1979). "Gambits: Conversational strategy signals". Journal of Pragmatics 3(3/4), 219-238.

Knapp, Karlfried (1991). Linguistische Aspekte interkultureller Kommunikationsfähigkeit. Unveröffentlichte Habilitationsschrift, eingereicht bei der Philosophischen Fakultät der Heinrich-Heine-Universität Düsseldorf im Mai 1991.

Koike, Dale April (1989). "Requests and the role of deixis in politeness". Journal of Pragmatics 13, 187-202.

Kultusministerium des Landes Nordrhein-Westfalen (Ed.) (1993). Richtlinien und Lehrpläne für das Gymnasium - Sekundarstufe I - in Nordrhein-Westfalen. Englisch. Frechen: Ritterbach.

Meeuwis, Michael (1994). "Nonnative-Nonnative Intercultural Communication: An Analysis of Instruction Sessions for Foreign Engineers in a Belgian Company". Multilingua 13:1-2, 59-82.

Meierkord, Christiane (1996). Englisch als Medium der interkulturellen Kommunikation. Untersuchungen zum non-native-/ non-native-speaker-Diskurs. Frankfurt et al.: Lang.

Nelson, Cecil Linwood (1984). Intelligibility: The Case of Non-Native Varieties of English. Ann Arbor, Mich.: University Microfilms International.

Oreström, Bengt (1983). Turn-Taking in English Conversation. Lund: Liber (CWK Gleerup).

Poel, Kris van de (1991). Modification in Phatic Endphases: A Study of Cross-Linguistic and Interlanguage Aspects. Unveröffentl. Dissertation. Universität Edinburgh 1991.

Sacks, Harvey et al. (1974). "A simplest systematics for the organization of turn-taking for conversation". Language 50/4, 696-735.

Schegloff, Emanuel A. und Sacks, Harvey (1973). "Opening up closings". Semiotica 8, 289-327.

Schneider, Klaus Peter (1987). "Topic selection in phatic communication". Multilingua 6/3, 247-256.

------- (1988). Small Talk: Analysing Phatic Discourse. Marburg / Lahn: Hitzeroth. (Univ. Diss. Marburg / Lahn).

Schwartz, Joan (1980). "The Negotiation for Meaning: Repair in Conversations between Second Language Learners of English". In: Larsen-Freeman, Diane (ed.). Discourse Analysis in Second Language Research, 138-153.

Smith, Larry and Nelson, Cecil (1985). "International Intelligibility of English: Directions and Resources". World Englishes 4.2, 333-342.

Tannen, Deborah and Öztek, Piyale Cömert (1981). "Health to our mouths: Formulaic expressions in Turkish and Greek". In: Coulmas, Florian (Ed.), 37-54.

Varonis, Evangeline M. and Gass, Susan (1985). "Non-Native / Non-Native Conversations: A Model for Negotiation of Meaning. Applied Linguistics 6, 71-90.

Westheide, Henning (1991). "Dialogstrukturierende Routineformeln". In: Stati, Sorin et al. (Eds.). Dialoganalyse III. Referate der 3. Arbeitstagung Bologna 1990. Teil 2. Tübingen: Niemeyer, 325-337.

Yamada, Haru (1990). "Topic management and turn distribution in business meetings: American versus Japanese strategies". Text 10(3), 271-295.

Zimmermann, Rainer (1984). Pragmalinguistik und kommunikativer Fremdsprachenunterricht. Heidelberg: Groos.

Notes

1 An earlier poster version of this article was presented at the 11th AILA World Congress of Applied Linguistics, 1996 in Jyvaskylä, Finland.

2 The studies which deal with lingua franca English so far mainly addressed the question of mutual intelligibility of different varieties - including learner language varieties - of English (for example Nelson (1984) and Smith and Nelson (1985)). Analyses of interaction between non-native speakers of English are still scarce. Two of these, Schwartz (1980) and Varonis and Gass (1985) investigated the negotiation of meaning from a learner language perspective. Studies focussing on discourse structure include Firth (1990 and 1996), Gramkow (1993), Meeuwis (1994) and Meierkord (1996).

3 Theoretical as well as methodological problems associated with the analysis and interpretation of non-native-/ non-native discourse are dealt with in Meierkord (1996) and (1998). These topics are not dealt with in depth here.

4 At the present time, however, we know too little about the processes and conditions involved in international communication to set a norm for the kind of English that is adequate for this type of communication and that could be used as a basis for the identification of learner-language utterances. [my translation, CM]

5 Hymes (1972) assumes that speakers do not only acquire competence for grammar, but also a competence for use.

6 A speaker is considered to be communicatively competent, if s/he can form turns, which are (1) grammatically structured, (2) adapted to the linguistic resources available to her, (3) suitable for the circumstances relevant to the conversation and (4) commonly used in the respective situation.

7 Schegloff and Sacks (1973) provide a theory of the structure of closing sequences and Poel (1991) investigates phatic endphases in interlanguage communication.

8 These differ slightly form opening and closing phases, both regarding their conversational functions and the illocutions found with them (cf. Meierkord 1996: 52f.).

9 The individual illocutions have been identified according to Edmondson and House (1981), who define illocutions as the speaker's communicative intent (1981: 48).

10 For the subsequent analyses, a turn will be defined as any utterance of a speaker, which furthers the topic of the conversation, and which ends either when another speaker takes the turn or by a long pause. A turn may be interrupted by short pauses up to two seconds as well as by simultaneous speech by another speaker, if this does not result in turn-taking. Turn-taking itself may occur with a pause between turns, with a non-comprehensible pause or with overlapping turns. In case turns overlap, the overlapping speech is considered to belong to all individual turns respectively.

11 Coulmas (1981), Blum-Kulka (1989) and Westheide (1991) provide further evidence on this topic.

12 topic change differs from a topic shift with regard to the content of the following topic. Whereas in the case of a shift its content is related to that of the precious topic, there is no relation between both topics after a change has occurred.

13 For a detailed account of the turn-taking system see Sacks et al. (1974), Oreström (1983) and Goodwin (1989).

14 The motivation and exact procedure is stipulated in Meierkord (1996: 108).

15 The mode is that value which has the highest frequency. (Cf. Buttler 1985: 32).

16 "The median value is that value of an arranged set of figures in order from highest to lowest, that is, in 'rank' order, which has equal numbers of observations above it and below it." (Buttler 1985: 29-30)

17 Cf. e.g. Götz (1977 and 1980) and Enninger (1987).

18 Laughter also plays an important role in repair sequences, i.e. stretches of talk which occur after the conversation had been interrupted due to misunderstandings. In these cases, speakers lacked a certain vocabulary item and had to jointly negotiate its meaning. During such sequences, the speakers' use of laughter helped them cope with these potentially face-threatening situations. A thorough description and interpretation of negotiation sequences can be found in Meierkord (1996).

19 L [the learner] assumes that N [a native speaker] assumes, that due to a lack of (receptive) competence in the English language, L does not (fully) understand what N says; L therefore lets N know by frequently sending back-channels that she does understand.

20 For a discussion of the notion of safe topic see Schneider (1988: 26).

21 For a description and interpretation of these breakdowns see Meierkord (1996: 205 ff.).