University of Surrey

English Language Institute

 

MA in Linguistics (TESOL)

 

MA Dissertation

 

 

Effects of Technologically Enhanced L2 Learning on Fluency and Accuracy:

The Case for Video and Audiotapes

 

Tutor: Ms Smiljka Gee

 

Charles Hohmann                                     17 October 2000

 

Abstract

 

The present study compares possible effects of two types of technologically enhanced language learning practices on the development of speech proficiency in an untutored L2 learning environment. After exposure to practices that are either visually/aurally (video, film) or only aurally based (audiotape), the two speech proficiency variables fluency and accuracy are analysed to determine whether any differences in L2 acquisitional patterns have ensued. The outcome of the experiment implies that, in an untutored acquisitional context, the use of a single medium (audiotape) can better lead to an improvement of the speech proficiency variable accuracy than a combination with media like video. Such a conclusion contradicts the common assumption that "more media is better" when teaching an L2 and suggests that the choice of one medium or a combination of media should take into account processing requirements and also the specific language skill which is to be improved.

 

In addition, the study also includes experiments with various operationalised fluency and accuracy sub-variables to test the limits of a quantitative approach to speech proficiency and makes suggestions for further research. It concludes by evoking implications for L2 learning and teaching practice.

 

 

Acknowledgements

 

Special thanks are due to Smiljka Gee for her pertinent and detailed comments on the manuscript, as well as to the staff at the English Language Institute of the University of Surrey and their director Dr. Glenn Fulcher who through their unflagging assistance made my first distance learning experience such an exciting and rewarding enterprise.

 

Contents

 

Abstract........................................................................................................ 3

 

Introduction................................................................................................ 6

 

Chapter 1: Technologically Enhanced Language Teaching

1.1. The role of narratives in language learning......................................... 7

1.2. Technologically enhanced language teaching.................................... 8

1.3. The limitation of attentional resources................................................. 8

1.4. Hypotheses............................................................................................ 9

 

Chapter 2: Review of the Literature

2.1. The role of listening in SLA................................................................ 11

2.2. Acquisition by means of listening materials ................................... 11

2.3. Acquisition by means of authored reading materials..................... 12

2.4. Acquisition by means of videos or multimedia................................ 12

2.5. The speech proficiency variables accuracy and fluency................ 14

2.5.1. Fluency...................................................................................... 15

2.5.1.1. Primary variables: .................................................... 16

2.5.1.2. Secondary Variables................................................ 16

2.5.2 Accuracy..................................................................................... 18

2.6. Mark-up conventions................................................................... 19

 

Chapter 3: The Experiment

3.1. Participants......................................................................................... 20

3.2. Materials.............................................................................................. 20

3.3. Design 21

3.4. Limitations of the design.................................................................... 22

3.5. Findings............................................................................................... 23

 

Chapter 4: Discussion

4.1. The pre-test......................................................................................... 24

4.2. The post-test: fluency.......................................................................... 24

4.3. The post-test: accuracy...................................................................... 26

4.4. Difficulties encountered in the approach with the variables........... 27

4.5. The scores for accuracy..................................................................... 27

 

Chapter 5: Conclusion

5.1. Implications for untutored L2 learning............................................... 29

5.2. Implications for tutored L2 learning................................................... 29

5.3. Implications for multimedia L2 learning............................................ 30

5.4. Summary............................................................................................. 30

 

Bibliographic References......................................................................... 32

Appendix 1: Sample of an analysis.......................................................... 39

Appendix 2: Statistical overview.............................................................. 42

 

Introduction

 

An assumption as a result of increasing multimedia acceptancy in current L2 teaching is that the simultaneous mediation of language through different perceptive channels improves L2 acquisition. To date, however, very little is known about how such a combination of media affects SLA. Two key factors to consider when combining different media are the impact of these media on specific language skills such as speech and writing and their respective subskills and the different processing requirements of the learner. Furthermore, it is necessary for research to distinguish between two contexts, one where acquisition occurs naturally as in L1 learning and a second context in which materials have been specifically processed to facilitate SLA. Research findings in the former context are a prerequisite for the development of L2 learning material if designers are to implement effective technological systems and teachers to use them purposefully. The present study attempts to provide some insights into the effects of single and combined media on acquisition in an untutored, natural context.

 

In view of the intricacy of the issue, the study limits itself to the impact of two types of technologically enhanced language learning practices on the development of speech proficiency in an L2 learning environment. After exposure to practices that are either visually/aurally (video, film) or only aurally based (audiotape), the two speech variables fluency and accuracy and their sub-variables are analysed to determine whether any differences in L2 acquisitional patterns have ensued. The context chosen is that of narrative sequences since the latter provide the mental schemata necessary to facilitate thematic planning and to enable coherent production in L2 recall operations. Speech proficiency being a highly complex and little understood linguistic phenomenon, the study also tests the usefulness of the different speech proficiency variables within the scope of quantitative analysis and points to some limitations of such an approach.

 

Chapter 1:

Technologically Enhanced Language Teaching

 

1.1. The role of narratives in language learning

The focus of the present study on narrative structures[1] needs to be justified since it may be challenged  that narration is not a typical example of truly communicative "face-to-face" language as some supporters of the Communicative Language Teaching (CLT) may claim. For in narration there are no turns, i.e. there is no form of alternation between a speaker and a listener. However, it can be argued that such a concept of "authenticity" or of "real life" English is too narrowly conceived.[2] Narratives already play an important role in the transmission of cultural knowledge from an early age onward and are part and parcel of our daily exposure to language in the media and, therefore, must be included in a communicative view of language.[3] When people listen to (or read) narratives they activate and test hypotheses about their possible contents (Rumelhart 1977, 1978, 1987, Dechert 1987: 229). These hypotheses, also called schemata, are evaluated against the sentences as they are read or listened to.[4] But besides being models of understanding, schemata are also models of production. Very often they help speakers to overcome or to avoid planning problems on a thematic level and therefore contribute considerably to making speech consistent and fluent (Dechert 1987: 234).

 

Discourse analysis applied to language teaching has also highlighted the importance of context in discourse processing and stressed the value of larger coherent and cohesive text structures (Lewis 1993: 15; 103-104; Segalowitz & Gatbonton 1995: 132). Oller (1983:12, as quoted in Joiner 1997: 94), to add a classroom study, claims that a text will be easier to produce, understand, and recall to the extent that it is motivated and structured episodically.

 

The importance of narrative structures in mental processing is also born out by studies in neighbouring fields. Research in artificial intelligence and psychology, for example, views “understanding” as a script-based phenomenon, scripts being instantiated by “stories” (Schank & Abelson 1977: 41). In the area of neuroscience, to quote another example, biological evidence has been found for the use of narrative patterns in mental processing.[5] In light of the above considerations, a study of the possible influence of narrative structures as input data for language acquisition merits investigating.

 

1.2. Technologically enhanced language teaching

While videotaped narratives with their motley of spoken and pictorial materials enjoy a great popularity among young L2 learners, audio-books are finding increased acceptance as an alternative to conventional reading, as their growing number in catalogues of publishing houses demonstrates. A more recent incursion of technology into an area traditionally dominated by written texts is the advent of computerised narratives. What most of these new technologies have in common is that they mediate narrative contents through different perceptive channels simultaneously. To date very little is known about how these multimedia technologies affect L2 acquisition (Brett 2000). However, such knowledge is essential if L2 learning material designers are to implement effective technological systems. Some of the questions future empirical research has to continue addressing can be summarised as follows:

 

a)     Are there differences in the impact various technologies may have on specific linguistic abilities such as listening, reading, speech and writing and their respective sub-skills?

b)     How are learners' limited attentional resources affected by being distributed simultaneously onto different perceptive channels?

c)      When input is presented to different channels simultaneously, does processing occur in an additive fashion or are there blending effects?

d)     Should certain materials be preferred to others when made available to a determinate channel?

e)     Are there materials that are better presented through a single channel (e.g. audiotapes) than through more than one channel (e.g. videotapes or multimedia, as in the case of modern CALL)?

 

The present study will address the first question in particular (a) and compare the possible effects of videotaped input (aural and visual) as opposed to audiotaped stimuli (only aural) on speech proficiency after extended exposure to the same fictional texts in both media. The focus will be on the way the two speech variables fluency and accuracy may be affected by attentional limitations (b) and possible blending effects (c). Based on the findings of the present and similar studies, the usefulness of the two technologies for presenting narratives could be assessed and a decision could be taken as to their respective roles in current practices of L2 teaching (d,e).

 

1.3. The limitation of attentional resources

It can be assumed that the amount of attentional resources an individual has at his disposal is limited, and spreading these resources simultaneously on several channels could mean reducing their perceptual effect on a single channel. This limited attentional capacity of the speaker-hearer, according to Sajavaara (1987: 55) is not constant and increases with task difficulty to a peak, from which it suddenly begins to decline again.

 

Attentional limitations can be shown to exist within a single perceptual channel in the apparent dissociation of language use from its related analytic system. When a speaker/hearer has to attend to meaning, he is distracted from attention to form (Skehan 1998: 27, 73). This prioritisation has been explained by the operation of two relatively autonomous language production and comprehension systems which compete for the same attentional resources: a rule-based analytic system and a formulaic, exemplar-based system which operates on language chunks (Skehan 1998: 54). According to this model, under the pressure of a communicative situation, the exemplar-based system is often prioritised to ease the processing load.

 

In view of such evidence, it does not seem unreasonable to look for pro-cessing limitations in other areas of language production as, for example, when several perceptive channels are involved. We assume, that within an L2 multimedia learning environment, which is characterised by the simultaneous or shifting stimulation of sight, hearing and speech, language processing faculties will vie for the limited attentional resources at hand. Conversely, as in the case of purely audio materials, where the learner can turn all his attention to the receiving (aural) channel, this “undivided” use of language processing resources could reduce the strain on processing capabilities and permit focusing on partial aspects of the language. It may be assumed, therefore, that the impact of either multimedia or merely aural training on acquisition will differ.

 

1.4. Hypotheses

A study of such effects, however, is fraught with difficulties because of the many psychological and linguistic variables involved in a multimedia learning environment. It is for this reason that a first approach to this field of inquiry should limit the number of linguistic and technological variables. Based on such considerations we may formulate the following hypotheses:

 

H0 : Assuming that the speech proficiency variables fluency and accuracy can be enhanced through exposure to L2 materials, it can be demonstrated that the nature of this impact on any of these variables is not dependent on the choice of either a combined aural and visual (videotape) or a simple aural language presentation technology (audio-book).

 

H1: In consequence the alternative hypothesis (H1) states that aural/visual and aural presentation technologies affect language acquisition differently, and that, therefore, depending on the type of technology, the development of the two speech proficiency variables fluency and accuracy will occur along different lines.

 

It will be argued that due to the small-scale of this study, its results cannot claim strong external validity which can only be obtained through repeated tests with larger samples. One of the values of the approach, however, is that it experiments with various speech variables to test the range of their applicability.

 

The questions raised are of importance for various areas in SLA practice. In an untutored context, on the one hand, as in the case of the autonomous learner, it may give him some guidance as to the choice of his materials, e.g. if his L2 accuracy is an area to be improved, then he may, if the hypothesis is confirmed, opt for an audio-book over a written text. In an instructed context, on the other hand, as in multimedia, a designer might, for example, choose certain purely acoustic sequences in his media mix to allow for the specific development of SLA accuracy.

 

Chapter 2: Review of the literature

 

2.1. The role of listening in SLA

Since this study analyses effects of technologically enhanced language teaching on a receptive skill like listening, this chapter gives a brief review of how some researchers see its possible role within a competence model.

 

A great deal of research into L2 acquisition has been devoted to the question of whether language competence is unitary or divisible. The Unitary Linguistic Competence Hypothesis was most extensively presented by Oller (1979, as quoted in Brown 1994: 264-265) who claimed the existence of a single general factor underlying all language skills. The divisibility hypothesis, on the other hand, is based on the claim that language proficiency consists of separate components or aspects of them (Bachman & Palmer, 1983, as quoted in Reves & Levine 1988: 327). Several researchers have suggested a compromise between the two opposing views, namely that language proficiency most probably consists of a general factor as well as of multiple contributing factors (Canale, 1983; Palmer, 1983; Vollmer & Sang, 1983; all quoted in Reves et al. 1988: 327). Reves and Levine (1988: 327-336), in a comparative study of reading and listening skills, found evidence for this latter hypothesis. They came to the conclusion that the skills of reading and listening consist of similar receptive sub-skills, and that listening is a more integrated skill, in which sub-skills tend to cumulatively contribute to the holistic comprehension of the message. According to Sajavaara (1987: 52) the mental apparatus involved in this process can be compared to a "time-sharing system"[6] which permits the hearer to make simultaneous use of phonological, syntactic, pragmatic, and other knowledge. If we consider, furthermore, that listening accounts for roughly half the time we spend using our L1 (Joiner 1997: 77) and that it is the main resource of L1 acquisition, its traditional pivotal role in many L2 teaching methods seems justified.

 

2.2. Acquisition by means of listening materials

Technology has been used to manipulate audio material in several ways. Modern equipment, for example, enables pausing and re-listening, thereby making the "immediacy skill of listening" resemble the "recursive skill of reading" (Joiner 1997: 82). It also permits the alteration of the rate of presentation which in turn may affect understanding in at least two ways: While listening to "time-expanded (slowed) speech", learner confidence is enhanced (Ko 1992, as quoted in Joiner 1997: 82-83), and "time-compressed (accelerated) speech" helps increase concentration (Duker 1974, Ostermeier 1991, as quoted in Joiner 1997: 83).

 

2.3. Acquisition by means of authored reading materials

Practically no L2 lesson takes place without the use of printed texts. Reading is similar to listening in that it uses comparable cognitive strategies, albeit probably in a less holistic way (cf. 2.1. above). However, it also differs in that by the very nature of its medium, it permits recapturing of information. Furthermore, information in writing tends to be more densely packed and its grammatical and discourse structure is often more complex (McDonough 1993: 129-130).

 

Hypermedia-annotated texts are a recent technological enhancement of reading materials. Working with the target text and the textual annotations, students "learn both about the language and with the language in a contextualized way, and they become actively involved in the reading process" (Martínez-Lage 1997: 121). The enhancement incorporates acoustic and visual support elements. Several studies (Omaggio 1993; Hudson 1982; both quoted in Martínez-Lage 1997: 143-144) support the use of such visual assistance during the pre-reading stage to facilitate the contextualisation of what is being read.

 

2.4. Acquisition by means of videos / films or multimedia

Foreign language teachers tend to think of television as being the more complete medium because of the extralinguistic and contextual clues to meaning provided by the visual image. As Joiner (1997: 84) remarks, "video engages the senses of both sight and hearing, thereby approaching the face-to-face situation". Balatova (1994, as quoted in Brett 2000) compared the effects of video, video with sound, and sound only on comprehension and found the visual clues to be more effective. Other studies (Herron, et al , 1995; Secules et al. 1992; both quoted in Brett 2000) also found that video improved listening skills.

 

Subtitled videos combine audio soundtrack, moving visual image, and written text and provide learners, therefore, with three different sources of input. Such a mix of media appears to ideally facilitate comprehension and to lead to improved language recall as a great number of studies evidence (Brett 2000).

 

However, research on the effects of video materials on L2 learners has also led to less favourable conclusions. These critical studies focus on the two topics emphasised in the above hypothesis: processing limitations and processing differences.

 

The limitation of attentional resources seems to explain why Long (1991, as quoted in Joiner 1997: 85) found that some students she observed had difficulty processing the auditory and visual material of an authentic Spanish video simultaneously and needed repeated viewing to take in the various stimuli present in the text. In another study, Thompson and Rubin (forthcoming, as quoted by Joiner 1997: 104) reported that after viewing a video, several L2 students commented that they sometimes elected to make use of one channel at a time. While some stated that they preferred viewing without sound in order to get a general idea first, others found the visuals distracting and reported that turning the picture off was helpful when listening for the second time. Finally, in a recent study that compared the way (1) video, (2) video and subtitles, (3) video subtitles and tasks, and (4) video and tasks affected the retention of ideas and language, Brett (2000) discovered that the last two groups (3 & 4) performed worst, a fact which he attributed to the much heavier processing loads the multimedia environment demanded.

 

A possible strategy to palliate the demands on perception of both visual and auditory stimuli is the McGurk-effect. McGurk and MacDonald observed that if a person was shown a film of someone saying 'ga' but on the soundtrack the speaker says 'ba', then what the hearer understands is 'da'. The visual and the auditory cues are integrated into unitary perception and this combination of stimuli produces a different message than the aural input alone. The "Gestalt" properties of the mind seem to make available to consciousness something that appears to correspond neither to the sound wave itself nor to the articulatory movements that produced it, but to more abstract categories of speech (McGurk & MacDonald 1976, Summerfield 1987, Langenmayr 1997: 568-569).[7] The McGurk-phenome-non could suggest that since combined visual and auditory input is pro-cessed at a more abstract level, attention to form is diminished, an aspect that is addressed in the next paragraph.

 

Evidence for differences in processing modes was found in attention and comprehension studies of young TV viewers. Meringoff et al. (1983: 174), for example, found that folktales for children (6 to 10 years-old) were processed differently depending upon whether they were presented as animated televised film or were read aloud from a picture book with illustrations. While children exposed to the televised story "included more of the characters' actions in their verbal retellings and did more physical gesturing," children who were read the story recalled more "figurative language and based their inferences more on textual content." In another study, Duquette & Painchaud (1996), who analysed vocabulary acquisition in both audio and video contexts, found some evidence for the fact that strategies developed by students were different depending on whether linguistic input was accompanied by visuals or not:

 

... when learners can both see and hear, it seems that less attention is focused on purely linguistic cues ... when only audio is available, some internal linguistic cues, particularly if the word is a cognate, facilitate the inference of words that are not specific to the dialogue's theme and that have a more general meaning (158).

 

It is the result of these findings which has led to the formulation of our main hypothesis, namely, that depending on the type of input, i.e. whether it is audio-visual or merely visual, an L2 learner's speech proficiency, i.e. his ability to speak easily, smoothly, expressively and accurately, will reflect a prioritisation either of communicative ability (fluency) or of focus on form (accuracy). In the following section we shall attempt to demonstrate how these two concepts can be defined and accommodated within the framework of Skehan's dual code processing hypothesis. The methodological focus will be on psycholinguistic aspects of production, a perspective omitting the cultural, social, communicative and intentional conditions of speech. Such a broader view would go beyond the scope of this study.

 

2.5. The speech proficiency variables accuracy and fluency

The dual code processing hypothesis claims that grammatical rule pro-cessing and memorisation are two psychologically, linguistically and neurolinguistically distinct processes, one of which can be impaired relative to the other (Gopnick 1994 as quoted in Skehan 1998: 220). Native speakers, therefore, have the possibility of producing the same language items from two independent sources – generatively using their acquired competence, or by recalling sentences as learned wholes (Cowie 1988: 126-139; Skehan 1998: 88-89; Lewis 1993: 95-96; Nattinger & DeCarrico 1992; Pawley & Syder 1983).

 

This interpretation derives from Jakobson's (1956) view of language abilities. According to Jakobson, who refers to de Saussure's distinction between syntagmatic and associative relations in language, language has a bipolar structure which manifests itself either in paradigmatic or metaphoric and syntagmatic or metonymic relationships:

 

The development of a discourse may take place along different semantic lines: one topic may lead to another either through their similarity or their contiguity. The metaphorical way would be the more appropriate term for the first case and the metonymic for the second, since they find their most condensed expression in metaphor and metonymy respectively. In aphasia one or the other of these two processes is blocked. ... In normal verbal behaviour both processes are continually operative, but careful observation will reveal that under the influence of a cultural pattern, personality, and verbal style, preference is given to one of the two pro-cesses over the other.[8]

 

Research findings suggest that to improve in one area often seems to be at the expense of improvement elsewhere, i.e. that trade-off effects are operating and that selective rather than across-the-board improvement apply (Skehan 1998: 112). These two aspects of performance that seem somewhat independent of one another Skehan calls fluency and accuracy[9]:

Earlier studies addressed the issue of fluency and accuracy under somewhat different terms, focusing on the way utterance planning, sometimes also referred to as “monitoring”, and repair mechanisms (e.g. “ehm”/”eh” tentative versions of phrases followed by improved versions, new starts) affect L2 production (Rehbein 1987: 100-101; Krashen 1978; Seliger 1980: 88; Ellis 1994: 393). The generic term for the above phenomena used in these studies is that of utterance planning and correction behaviour (UPC) (Seliger 1980: 88). These UPC phenomena have been subdivided into two basic types, temporal variables and hesitation phenomena. According to these studies progress in fluency skills is reflected by the gradual acceleration of UPC behaviour until it can be performed nearly simultaneously with the execution stage.

 

The insights these earlier studies provide are invaluable and have been included in the present approach. The only difficulty encountered was in the use of the variable hesitation phenomena since it does not distinguish between two types of hesitation that were encountered in the present corpus, namely those that seem to reflect searches in the mental lexicon, or to use Jakobson's terminology, that are of "paradigmatic nature", and those that precede false starts and certain self-corrections and that, again to quote Jakobson, are of "syntagmatic nature". In the present study the former will be referred to as fluency deficiencies and the latter as deficiencies of linguistic knowledge (accuracy).[10]

 

2.5.1. Fluency

As noted above, a central aspect of fluency is the ability to produce lexical phrases as unanalysed wholes or “chunks” (Lewis 1993: 95).[11] In some studies this process is also referred to as automatization, a complex decision-making process, which enables speakers to short-cut and simplify complex activities previously learned in order to reduce the number of mental operations required for the activity. This process makes the costly construction of some structures superfluous (since the outcome is already available), or speeds up the search in the mental lexicon (Wiese 1984: 15).

 

Fluency has been analysed in terms of temporal variables which again can be subdivided into primary variables and secondary variables (Grosjean 1980:40-42).[12]

 

2.5.1.1. Primary variables:

Primary Variables, following Grosjean (1980), are always present in language output and can be subdivided into Complex and Simple Variables. Complex Variables can again be separated into Speech Rate and Phonation-Time Ratio (P.T.R.).

 

Speech Rate is measured in syllables per second and is obtained by dividing the total number of syllables in the utterance by the total speech time and dividing the product by 60.

 

Phonation-Time Ratio (P.T.R.), usually expressed as a percentage, indicates the amount of time spent articulating during the utterance. It is obtained by dividing the time spent articulating by the total speech time and multiplying the product by 100.

 

The second type of Primary Variables, the Simple Variables, consist of the three variables Articulation Rate, Length of Silent Pauses and Length of Runs.

 

Articulation Rate is the actual phonation rate of the speaker. It is usually expressed in syllables/sec and obtained by dividing the number of syllables in the utterance by the speaker’s articulation time (total speech time minus the pause time).

 

Length of Pauses, refers to all filled and unfilled pauses >= 1 sec as measured on a pen recording. It is obtained by calculating the mean length of pauses above the threshold value of 1 sec.

 

Length of Run is defined as the speech that occurs between two pauses. It can reflect a word, a phrase, a sentence or a series of sentences depending on the task and the rate of the output. It is often characterised by "prosodic features like specific intonation phenomena, intensity and stress” (Raupach 1984: 117). The formulaic units or “chunks” as described above appear to coincide with the concept “length of run”. It is obtained by calculating the mean numbers of syllables between pauses.

 

2.5.1.2. Secondary Variables

Secondary Variables belong to the hesitation phenomena category. Unlike the primary variables, their presence is not required in speech and they rarely occur in oral reading and in very fluent speech. Grosjean (1980) lists four types of Secondary Variables: filled pauses (e.g. “ehm”), drawls (unnatural lengthening of syllables) repeats (any repetition that does not add to the meaning of the utterance) and / or self-corrections (any unfinished phrase or sentence).[13]

 

In accordance with the dual code approach and in the light of Jakobson's theory, we shall distinguish between hesitations predominantly due to either paradigmatic selection or to morphosyntactic or syntagmatic combinations. For example, repeats, on the one hand, usually serve the purpose of providing additional time for lexical search (Lennon 1984: 66) and are, therefore, hesitation phenomena that can be assigned to the fluency category. Self-corrections, on the other hand, can be hesitation phenomena of either lexical or syntactic origin. While in the former case, i.e. when they reflect lexical replacement or substitution (selection), they will be assigned to fluency, in the latter, where they indicate a deficiency of linguistic knowledge (syntagmatic or combinatory), they will be classified as relating to accuracy. The following four examples from our corpus serve to illustrate this difference:

 

(1)     "... it drove down out ..."

(2)     "... but but he he wasn't up to ..."

(3)     "... the next brother cames in comes in who is the ..."

(4)     "... the policeman is a has to take care of Carl ..."

 

In the above examples the lexical replacement ("out" for "down") (1) and the two repetitions ("but", "he") (2) are instances of hesitation phenomena of paradigmatical nature or selection and pertain to fluency, whereas the next examples are hesitation phenomena of syntagmatic origin or of combination, since they repair syntax ("cames" vs. "comes") (3) and sentence structure ("is a" vs. "has to take care of") (4), respectively. This adaptation of UPC variables forces us to abandon the distinction suggested by Grosjean between primary and secondary as well as between simple and complex categories of variables since they overlap.

 

The dichotomous approach to hesitation phenomena proposed above is useful from an analytical point of view but must not make us blind to the fact that both criteria only describe predominant aspects of what in language use forms a unity. In fact, sometimes the main source of a hesitation phenomenon cannot be established clearly, or, to quote Wiese (1984: 21), "sometimes one cannot decide whether the deficiency is one of linguistic knowledge or automatization (or both)."

 

In accordance with the criteria outlined above we can, therefore, distinguish the following hesitation phenomena as belonging to fluency:

·        repeats

·        corrections in the meaning of an utterance

·        topic changes

·        corrections of lexicon where one or two words are changed in an utterance

·        collocational repair

 

2.5.2 Accuracy

Skehan (1998: 5) defines accuracy as a learner's attempt to use an interlanguage system of a particular level to produce correct, but possibly limited language. The limitations are perceived by a NS as deviations from his adult grammar and reflect the interlanguage competence of the speaker. Usually the speaker is not aware that part of his IL-utterance is merely an intermediate structure not present in the target language. However, when his IL is in a phase of stabilisation, he acquires the ability to correct himself. In the present analysis we have opted for treating both hesitation phenomena separately as either self-corrections or "errors". As already outlined above, deficiencies in accuracy imply faulty combinations of morphemes and syntactic order.

 

Accuracy is measured by dividing the speaker’s articulation time (total speech time minus the pause time) through either self-corrections or IL forms ("errors"). The numeral obtained represents the mean time during speech measured in seconds in which no interruption occurs. [14]

 

Marking L2 speech is different from marking L2 writing since incomplete sentences, the lack of explicit logical connectors (e.g. however, moreover), the replacing/refining of expressions, the use of generalised vocabulary (e.g. thing, stuff, etc.), repetitions of the same syntactic structures and the use of pauses and 'fillers', etc. cannot be considered "errors" in the common sense, implying that the marker has to be rather tolerant. It is also important to recall in this context that speaking with fluency in an L1 does not always imply an uninterrupted flow of error-free speech. The "good" NS speaker "knows" how to hesitate, how to be silent, how to self-correct, how to interrupt, and how to complete expressions or leave them unfinished following an intricate pattern of social and cultural rules (Sajavaara 1987: 62). In the case of the L2 learner who still lacks these skills, however, these phenomena are determined by the semantic and syntactic requirements of the target language and follow to a lesser degree pragmatic rules. As a consequence, the variables listed above would not be applicable in the same way to an L1 learning context.

 

 

Below is a summary of the different speech variables that have been discussed in this section:

 


 


Table 1: Overview of speech proficiency variables

 

 

2.6. Mark-up conventions

The mark-up symbols used in the transcriptions (cf. appendix I) refer to the above variables and are as follows:

 

Fluency:

the passengers           (highlighted text) length of run

(ehm), (yeah)              filled pauses

/....../                         (text between forward slashes) hesitations, paradigmatic self-corrections and repeats

[2,5"]                         Pauses above threshold of 1 sec

...                              (three dots) minor hesitation < 1 sec.

 

Occasional code-switches were included in the count and are indicated by a parenthesis and a reference to the language in question (G = German; SG = Swiss German)

 

Accuracy:

/their/ his                    (text between forward slashes and crossed out once) syntactical self-corrections, i.e. self-perceived errors

he comed                   (text crossed out twice) IL-forms, unperceived syntactical errors

 

 

Chapter 3: The Experiment

 

3.1. Participants

The sample consisted of 21 (7 females and 16 males) intermediate learners of English aged between 16 and 18, having German as their L1 and stemming from two parallel classes in a German Swiss school (Gymnasium). Most of them had learned English at school for a period ranging from three to four years involving three 45' periods per week. However, two sample members had benefited from exposure to English in English speaking countries and were partly bilingual.

 

3.2. Materials

Pre-test:

To test students’ level IL of fluency and accuracy in English, they were asked to give a summary of and comment freely on a narrative that had been previously presented to them in the form of an audio-book (Skyjack by Tom Vicary[15]) for five minutes. Such a use of spoken language has been described as "transactional", or as focusing on factual and propositional information as opposed to "interactional communication", where the learner attempts to establish and maintain social roles (Anderson & Lynch 1988: 8; McDonough & Shaw 1993, 155). To encourage language production the students were given a set of pictures related to the narrative.

 

Intermediate phase:

In a second phase the L2 students were split up into two groups of approximately the same size and each group was exposed for 10 periods of 45' either to a video or an audio-taped version of a fictional narrative characterised by a high rate of direct discourse, in the present case, to Doyle's The Hound of the Baskervilles. This text was chosen since it exists in versions that have been specially rewritten for both media, namely as a dramatised audio-version for a BBC radio play and in two movie versions.[16] These adaptations of one and the same story line to the particularities of both media have the advantage of permitting a better assessment of the medial impact on acquisition as well as facilitating experimental comparison across media.

 

Post-test:

The students of both groups were asked to give a partial summary of and to comment freely on either the dramatised or the video versions of C. Doyle's The Hound of the Baskervilles for five minutes. To encourage language production the students in both groups were given the same set of pictures that illustrated an episode of the narrative.[17]

 

3.3. Design

Sample-members all knew the researcher in his capacity as ESL teacher well and no one else was present. Furthermore, sample-members were aware that they were being recorded and were not permitted to use notes. Since they were prompted to re-tell their stories or comment on them, their output can be classified as "experimentally elicited data"[18]. In addition, no planning time was allotted, since this could have affected the variables in uncontrollable ways.[19]

 

The students were allowed to go through the materials at least twice so that acquisition could be reinforced through repetition, a key factor for retention (Segalowitz & Gatbonton 1995: 140). After 10 hours of intensive practice, fluency and accuracy levels were tested again using the same approach as in the pre-test.

 

The audio-book group:

Listening strategies can be classified according to the degree of verbal interaction they entail and range from attentive listening with minimal verbal interaction to intensive listening (focusing the students' attention on language form), selective listening (directing the student's attention to key words, discourse sequence cues, etc.) and interactive listening (maximum of interaction) (Rost 1991: 10-11). For the present experiment attentive listening with a minimum of interaction, little teacher intervention and input manipulation seemed to be the most adequate listening strategy. Care was taken not to attract explicit attention to formal features of the language (intensive listening) and to provide both groups with the same visual aids (an illustrated structural mapping of the story and its main characters on a poster) and just enough vocabulary explanation so as to assure comprehension. When students found difficulties in understanding passages, these were played again rather than paraphrased or elucidated by the teacher.

 

The video group:

Individual scenes and short sequences from the video were presented at first and then repeated until the students felt they were following the narrative. Whenever necessary the teacher pointed to the same illustrated structural map that was presented to the audio-group. Again, no attention was drawn to formal language features and teacher interaction was restricted to the bare essentials.

 

3.4. Limitations of the design

The experiment assumes that input frequency will have affected output within the span of the two tests (pre-test & post-test). While, as Ellis (1994: 270) points out, it makes little sense to investigate input-output relationships in data collected at the same time, there is no consensus among researchers as to the length of a gestation phase. Furthermore, as Ellis (1994: 288) also points out, input combines with other factors, such as the learner's L1, the learner's communicative need to express certain meanings and the learner's internal processing mechanisms.

 

A dimension of learning style which is very prominent in a formal classroom setting is the preference that learners show toward either visual or auditory input. Visual learners tend to prefer reading and studying charts and drawings, etc., while learners with an auditory bias prefer listening to lectures and audiotapes (Brown 1994: 113). Therefore, students with visual preferences could have benefited more from the video sessions, while students with an auditory bias would have been at an advantage in the aural group. An analogous distinction can be drawn on the basis of the dual code processing theory, which distinguishes between learners who tend to be excessively rule-oriented and are predisposed to emphasise restructuring and accuracy, and learners who rapidly acquire communicative fluency by preferring exemplar-based learning (Skehan 1998: 269-270). In the present study these differences in learning styles and linguistic aptitude types were not addressed since it was assumed that both types of cognitive preferences were equally distributed among the two groups.

 

Another set of variables that may influence data eliciting procedures is the difference between high-knowledge and low-knowledge subjects (within the same domain) reading the same text (Seliger 1980: 90; Appel 1984: 186-187). High-knowledge individuals not only recall more propositions of a story, but they also provide a more accurate account of the sequence of important events in the text. Under the assumption that both high- and low-knowledge subjects were uniformly represented in both groups, these variables were not controlled.

 

A further consideration to keep in mind is that L2 speech proficiency (fluency, accuracy, complexity and control of content) also depends to some extent on the type of task set. Depending on whether the student is confronted with a personal task, a narrative task or a decision task, for example, and given the limited attentional capacities of L2 users and the trade-off effects operating between the variables, these will be affected differently. In Skehan's & Foster's (1997: 185) words:

 

Tasks which contain clearer inherent structure, when planned, seem to favour accuracy, whereas tasks which require more on-line processing or which have complex outcomes, when planned, produce greater complexity.

It could be, therefore, that the retelling task employed in this study favours one variable in particular and that its results might not be relevant for other task types.

 

Length of gestation, cognitive preferences, cognitive aptitudes and task dependency are aspects which, therefore, could have had an impact on our results and the validity of the results, therefore, must be seen within the scope of these limitations.

 

3.5. Findings

The tape-recorded materials were transcribed and analysed in terms of the various aspects of fluency and accuracy as outlined in chapter 2. Finally, whenever necessary a Student's t-test was applied to the variables to test their significance at the 0,05 level.[20]

 

In the tables below, the scores for the different fluency variables in both the pre-test and the post-test are summarised for the two groups. The last column indicates whether the 0-hypothesis must be accepted (Y), i.e. the result is not significant (ns) or must be rejected (N) because the results are significant (s).

 

a)     Pre-test scores

 

Fluency

Pre-test A

(aural)

Means

Pre-test B

(aural)

Means

P value

(two tailed)

p>0,05

(Y/N)

Speech rate (syl/sec)

2,1

2,0

 

Y/ns

Phonation-Time Ratio

85,6

77,6

0,1504

Y/ns

Articulation rate (syl/sec)

2,5

2,6

 

Y/ns

Pause length (sec)

3,4

3,5

 

Y/ns

Length of run (syl/run)

22,4

16,6

0,0766

N/s

Repetitions, self-corr. (semant.) (no)

15,9

18,2

NA

NA

Accuracy

 

 

 

 

Self-corrections (syntax) (secs)

96,7

97,4

 

Y/ns

Errors (syntax) (secs)

16,3

10,6

0,3726

Y/ns

 

b)     Post-tests: scores in experimental groups (video and aural)

 

Fluency

Post-test (aural)

Means

Post-test

(video)

Means

P value

(two tailed)

p>0,05

(Y/N)

Speech rate (syl/sec)

2,0

2,2

 

Y/ns

Phonation-Time Ratio

83,8

83,1

 

Y/ns

Articulation rate (syl/sec)

2,4

2,7

0,2366

Y/ns

Pause length (sec)

3,0

3,6

0,1348

Y/ns

Length of run (syl/run)

22,9

17,6

0,1838

Y/ns

Repetitions, self-corr. (semant.) (no)

19,7

23,0

NA

NA

Accuracy

 

 

 

 

Self-corrections (syntax) (secs)

65,0

79,5

0,3132

Y/ns

Errors (syntax) (secs)

19,7

9,9

0,0514

N/s

 

The results of the tests are discussed in the following chapter.

 

Chapter 4: Discussion

 

The outcome of the experiment is only partly conclusive, a fact which may also be due to the small size of the samples, the brevity of the treatment, to the various additional variables that were not taken into account (cf. 3.4. above) and to limitations inherent in a purely quantitative approach to speech proficiency. The results are discussed in detail below.

 

4.1. The pre-test:

Except for the variable length of run that was weakly significant, all variables confirmed the 0-hypothesis as expected from the pre-test.

 

4.2. The post-test: fluency

4.2.1. Speech rate

(Aural group: 2,0 syl/s; video group: 2,2 syl/s; ns)

The speech rate which represents the amount of syllables per second including pauses did not vary in any of the two experimental groups and, therefore, these results support the 0-hypothesis.

 

4.2.2. Phonation-Time Ratio

(Aural group 83,8 %; video group V: 83,1%; ns).

While Phonation-Time Ratio, which indicates the amount of time spent articulating during the whole utterance, differed slightly but not significantly in the pre-test, in the post-test it was identical in both groups, thereby confirming the 0-hypothesis.

 

4.2.3. Articulation rate

(Aural group 2,4 syl/s; video group: 2,7syl/s, ns).

The articulation rate which measures the number of syllables per second during effective speech time showed no significant change in the aural group when compared to the pre-test; however, there was a slight improvement in the video group, albeit not significant, suggesting that video stimuli could have affected this aspect of fluency. Nevertheless, the 0-hypothesis must be accepted.

 

4.2.4. Pause length

(Aural group: 3,0s; video group: 3,6s; ns)

The variable pause length which measures the average duration of a pause in seconds and includes in this test filled (e.g. "ehm", "yeah", etc.) and unfilled pauses yields results that appear to confirm the 0-hypothesis.[21] However students in the aural group made slightly shorter pauses and could have, therefore, improved in this aspect of fluency. Such a tendency is unexpected and is probably due to the fact that the computation of this variable remained imprecise. The cut-off points for a pause are often difficult to determine since this variable seems to enter into conflict with the variable "length of run", as will be shown below. This difficulty in defining what can be considered a pause finds its reflection in the different threshold levels researchers propose (e.g. Lennon 1984: >0,2s; Grosjean 1980: >0,25s; Skehan & Foster 1997: >1s) and for which there does not seem to be a rationale.

 

4.2.5. Length of run

(Aural group: 22,9 s; video group: 17,6 s)

The values of this variable which indicate the average duration of speech between pauses, like the previous one, confirms the 0-hypothesis. Again, the small difference that suggests that the aural group made longer runs and, therefore, became more fluent is unexpected. However, as observed above, measuring the length of a run includes counting drawls and repeats, time gaining devices that ought to be given the same status as pauses. The preference for either repeats, drawls, or filled and unfilled pauses appears to depend more on learner speech strategy and psychological factors like the degree of self-confidence of a student. An additional difficulty is the fact that in an L2 learning context the natural flow of language is more often interrupted than in L1 speech, where interruptions coincide more with "chunks" or "episodic units", and often breathing pauses occur within parts of speech in an erratic way. Therefore, measuring the length of a run like the length of a pause in L2 speech can only yield approximate values.

 

A further aspect which merits consideration is the role of the variables "control of content" and "complexity", variables which were not considered in this experiment. Some hesitations that at first appear as difficulties of syntactic alignment are, in fact, difficulties with reordering content. Decisions as to topic selection and general meaning, therefore, may also induce hesitations and affect the length of pauses and, therefore, shorten runs.

 

4.2.6. Repetitions, self-corrections (semantic)

(Aural group: 19,7; video group: 23,0)

The use of this variable was unsatisfactory. Repeats can be time-gaining devices like pauses but also ploys to keep the listener's attention, and result, therefore, from the application of a speech strategy. The occurrence of repeats with different functions entails that the researcher must make assumptions about whether a repeat is a time gaining device which should not enter into the syllable count or merely a rhetorical device. This type of dilemma points to the limitations of a quantitative approach to speech proficiency since the researcher's decision will also affect other variables like Phonation-Time Ratio and the articulation rate. Furthermore, some repeats contain synonyms, thereby implying lexical searches and they can, therefore, by definition, not be considered time-gaining devices, as the example below illustrates:

 

(5) "... all the people who are all (yeah) the people who live in Baskerville Hall..."

In the above examples a phrase is repeated, albeit with the replacement of one of the key expressions through a word with a similar meaning, i.e. without changing any content of the utterance.

 

As outlined above, self-corrections that implied a different paradigmatic choice were also considered under the present category. Consider the following example:

 

(6) "... Ah that's why because (ehm) Sir Henry is the last of the Baskervilles ..."

 

In the sentence above, the student seems to realise that he has made a wrong word choice ("why") and he substitutes it with what he considers a more appropriate selection ("because").

 

4.3. The post-test: accuracy

Accuracy is analysed using two variables, self-corrections of syntagmatic nature, i.e. word combinations perceived as wrong and unperceived IL-forms, i.e. inadvertent "errors" of syntax. The former occurred only every 97s while the latter happened after every 10s (video group) or 20s (aural group) respectively.

 

4.3.1. Self-corrections (syntax)

(Aural group: 65,0; video group: 79,5; ns)

Compared to the pre-test both groups corrected their utterance more often in the post-test. In the pretest the number of syntagmatic self-corrections was practically the same for both groups; in the post-test, however, while both groups corrected themselves more, the aural group outdid the video group. It must be recalled that a greater ability to correct oneself in an L2 is an indication of greater IL-stability (cf. 2.4.2). In other words, while the students' accuracy in both groups benefited from the treatment generally, the aural group improved their accuracy levels more. However, statistically, these results were not significant enough to warrant a claim. An analysis of the types of self-corrections yields the following classes (the self-corrections are highlighted):

 

a) False starts:

(7) "... and there is (ehm) the the owner died ..."

(8) "... he didn't want to cause ... most probably he felt a bit annoyed ..."

 

b) Omissions, deletions:

(9) "... telling him to give up and ... but Stapelton didn't want to ..."

(10) "... Then they went through the moor and ...  but they didn't find him ..."

 

c) Additions of words or phrases:

(11) "... he was waiting her for her at the door ..."

(12) "... is that the all the inhabitants of this place ..."

(13) "... and then but he couldn't find him and then he was waiting in a hut but he couldn't find him..."

 

 

 

d) Replacements / substitutions:

(13) "... and that was used because (ehm) ... it was used to catch the scent..."

(14) "...so he gave justice he was brought to justice ..."

 

e) Re-orderings:

(15) "... because he didn't want to let Watson know ... if Watson knew that he was here ..."

 

4.3.2. Errors (syntax)

(Aural group: 19,7; video group: 9,9; s)

This variable, like the previous one, indicates the amount of time that elapses before an unperceived "error" occurs. It is obtained by dividing total speech time minus pauses by the total number of "errors." Interestingly the video group, which had performed slightly worse than the aural group in the pre-test evidenced no progress whatsoever in the post-test. The aural group, however, improved its performance, so that the difference between the two groups became significant in the post-test and, therefore, the 0-hypothesis must be rejected in the latter case. In other words, the exposition to a mere aural stimulus appears to permit more attention to formal aspects of the language and, thereby, to improve accuracy scores better. This result of the test is explained in more detail in the following section.

 

4.4. Difficulties encountered in the approach with the variables

Several difficulties in the use of the variables were described above, such as decisions as to whether a phenomenon like a repeat was because of a mental search operation or due a to rhetorical ploy as in L1 speech, or whether a hesitation phenomenon was the result either of an operation of syntagmatic realignment or of content ordering. In such cases the analyst is in a classical linguist's dilemma and is forced to make mentalistic assumptions about the intentions of the L2 learner, a fact which weakens the internal validity of the quantitative approach and its variables. This issue is important and needs to be resolved if L2 input parsers are to developed for CALL.

 

4.5. The scores for accuracy

The findings for accuracy need to be interpreted in more detail. If listening is a more integrated skill in which sub-skills tend to cumulatively contribute to the holistic comprehension of the message (cf. 2.1. above) and if, in turn, the learner may acquire certain language skills better by directing his "undivided" attentional resources exclusively to the receiving (aural) channel (cf. 1.3. above), then we must conclude that listening plays a primordial role in language acquisition. In the present case, in particular, listening in an untutored context appears to foster formal "apperception", i.e. the noticing of formal aspects of the input (Chapelle 1998: 23) better than a combination of acoustic and visual stimuli. In other words, the “undivided” use of language processing resources while listening could have the effect of raising the learner's attention to linguistic code features, a process also called "focus on form" or "noticing" (Schmidt 1990). Such noticing, which actually takes place within short term memory, constrains what the learner can attend to effectively so that its result becomes available for practice and storage into long-term memory (Skehan 1998: 49-52).

 

The fact that the accuracy variable "syntagmatic self-correction" produced better scores in the aural group - however, it must be admitted, not significant in a statistical sense – also needs to be interpreted. If, as was observed above, what has been noticed is integrated into long term memory, or has become "intake" and has been integrated into the linguistic system of the speaker/hearer (Chapelle 1998: 23), then this, in turn, will affect his L2 output. This output is not only an observable result of the process but is also an important contributor to linguistic development since it forces the learner to use the newly acquired syntactic features. Consequently, in the case of untutored listening the heightened awareness for syntactic input appears to lead to a heightened awareness for syntactic form in the output and, therefore, to more self-corrective behaviour.

 

These insights – albeit in a different research setting - tend to confirm the so-called "zero option" hypothesis, i.e. the assumption that classroom learners may acquire grammatical competence without any formal instruction, as proposed by Prabhu (1987) and others (cf. Ellis 1997: 47-48)[22]. However, the assumption that the formal intake described above would apply to all syntactic phenomena could be too strong. Maybe, as some researchers have already pointed out in cross-sectional studies, not all linguistic forms can be acquired implicitly to the same extent. [23] These investigations into IL forms which, in an untutored environment, either allow or resist assimilation are important since they are a prerequisite for studies of L2 developmental sequences.[24] Finally, analogous investigations could be made into reading patterns to determine how attention to form manifests itself within the context of a print medium and comparisons drawn as to the type of syntactic structures acquired in reading and their possible order.

 

The fact that the major results of the present research - notwithstanding the caveats expressed above - dovetail with previous studies heightens their value, and it is our belief that they contribute an important piece to a jigsaw puzzle that applied linguistics is gradually putting together, namely the processes involved in learning a foreign language. Furthermore, by demonstrating that by being exposed to a mere aural stimulus a foreign language learner's linguistic accuracy can be improved and self-corrective behaviour encouraged, a new prioritisation of L2 teaching practices must ensue. It is to this new focus on certain established L2 methods that we shall now turn.

 

Chapter 5: Conclusion

 

It was remarked in the previous chapter that the findings of the present paper appear to give some support to the "zero option" in SLA, without, however, endorsing the rejection of teacher intervention. In fact, quite a large body of research suggests that a planned manipulation of input that enforces "noticing" can effectively foster the acquisition of grammatical forms (Doughty & Williams 1998). Actually, as the "Natural Approach" suggests (Krashen & Terell 1983), these two approaches need not be incompatible, and it can be maintained that a knowledge of assimilative and resilient target language forms is also a prerequisite for a "focus on form" approach which could particularly address those structures that prove resistant to implicit learning. The following section will attempt to elucidate some implications of these insights to both contexts of L2 learning, namely tutored and untutored acquisition through listening practices.

 

5.1 Implications for untutored L2 learning

It is a well-established and also embarrassing fact that foreign language classroom tuition seldom leads to advanced proficiency levels (Lewis 1993: 122). Learners, e.g. those preparing language proficiency tests or those having gone through some kind of immersion learning[25], therefore, often need to complement their language skills. In such cases, depending on the type of linguistic skill they wish to develop, they can be given guidance as to the natural source materials that will best suit their requirements. If accuracy needs to be improved, for example, a teacher-adviser could recommend a phase of purely acoustic input in the guise of radio news or audio-books in lieu of written materials.

 

5.2. Implications for tutored L2 learning

As observed above, proponents of the "Natural Approach" (Krashen & Terell 1983) claim that, in order to be successful, language learning procedures should be based on natural language acquisition or assimilation. The main statement this approach makes is that comprehensible input is a necessary and sufficient condition for successful L2 acquisition (Richards & Rodgers 1986: 140). Teaching procedures in the wake of this approach attempt to single out those features of the acquisitional process that are amenable to guided practice to develop their methods (Richards & Rogers 1986, 128 ff., Klein 1986: 54 ff.). Naturally, as the learner progresses the number of methodical interventions can be reduced and, in accordance with the premises of the approach, L2 learning will eventually proceed to an untutored stage. Again, considering the advantages of acoustic input outlined above, the use of aural material will be recommended.

 

A similar case can be made for the "Lexical Approach" (Lewis 1993) which champions the use of "real" materials. According to this approach, long term acquisition is facilitated by material which is only partially understood or as Lewis (1993:186) states: "Providing the task is chosen so that students can understand something, it is absolutely not essential that they understand everything" (highlighting is the author's). The approach requires the selection of "real" material, and the preparation of supplementary material around it (Lewis 1993: 186). Such a use of learning materials is in accordance with the findings outlined above.

 

5.3. Implications for multimedia L2 learning

New speech technologies are developing at a breathtaking pace.[26] In their wake, the number of available multimedia-based learning materials is also increasing rapidly, sometimes outpacing the corresponding body of research into their implications for L2 pedagogy (e.g. Jamieson & Chapelle 1988; Harben 1999). At the present stage, L2 pedagogy endorses the view that multimedia applications are particularly effective in an L2 environment since they are adaptable to the individual learner's age, his interests, his learning styles, aptitudes, motivations and the learning situation. Of further importance is also the fact that multimedia learning materials can be task-driven and can draw on the whole gamut of language production and reception faculties, including non-linguistic or paralinguistic information, to facilitate comprehension.

 

Nevertheless, as was observed previously, research results on the utility of visual material in combination with listening activities were not always conclusive, in particular, when sub-skills like fluency and accuracy were considered (cf. 2.4. above). It was suggested that there is a threshold above which the combination of visual information with linguistic input fosters partial skills at the expense of others, a factor attributable to an overload of the learner's processing capabilities. This insight paves the way for a more effective exploitation of the possibilities multimedia technology offers to control the media mix. Not only does this control facility enable the teacher to tailor the materials to the requirements of his students, but the student himself, aware of his learner's profile and linguistic necessities, can adapt the technology to suit his needs. Multimedia materials specially designed to permit self-accessing listening practice and on-line choices allowing learners the flexibility to decide their own way of working through the materials are fortunately available today (Harben 1999).

 

5.4. Summary

The present study attempted to find evidence for the hypothesis that aural/visual or aural presentation technologies affected language acquisition differently, and that depending on the choice of a technology, the development of the two speech proficiency variables fluency and accuracy would occur along different lines. While the study could find some evidence for the fact that aural input alone led to better accuracy levels, it could not detect any major impact on the variables at stake when aural and visual stimuli were combined. Moreover, because of the small-scale of the study, the exclusion of certain situational variables, and constraints on quantification inherent to some of the variables used, the results cannot claim strong external validity.

 

The difference in impact of the two technologies on speech proficiency was explained by the operation of attentional constraints. It was maintained that to ease the cognitive processing demands of input through two channels, the learner would attend less to formal aspects of the language. Conversely, the processing of input through one channel appeared to permit an “undivided” use of language processing resources, thereby allowing a heightened awareness to form and, therefore, to more formal intake.

 

In the course of the analysis, variables other researchers have proposed for the study of speech proficiency were revised and extended to include a distinction between hesitation phenomena due to operations of a formulaic, exemplar-based speech system and others due to a rule-based analytic system, a distinction that was shown to parallel Jakobson's bipolar view of language production as being either determined by paradigmatic choices or syntagmatic combination operations. Furthermore, it was concluded that the sub-variables length of pause, length of run and hesitations could not be applied to the study L1 speech proficiency since, in the case of the L1 speaker, these sub-variables are not determined by the semantic and syntactic requirements of the language but follow pragmatic rules.

 

Finally, in analysing implications of the research results for L2 learning, it was shown that, whether in tutored or untutored contexts, the choice of a single medium or of a combination of media must be made according to processing requirements and pedagogic decisions concerning the specific speech skill to be improved. This is an important insight which essentially contradicts the widespread assumption that "more media is better" in L2 learning.

 

 

 

Bibliographic References

 

Anderson, A. & Lynch, T. (1988). Listening.Oxford: Oxford University Press.

 

Appel, G. (1984). "Improving Second Language Production”. In Dechert et al. (1984): 186-210.

 

Baddock, B. (1996). Using Films in the English Class. Hemel Hempstead: Phoenix ELT.

 

Balatova, I. (1994). "The impact of Video on the Comprehension Skills of Core French Students." The Canadian Modern Language Review 50: 507-523.

 

Barnbrook, G. (1996). Language and Computers. Edinburgh: Edinburgh University Press.

 

Brenzel, E. (1984). “Summarizing in a Second Language.” In Dechert, W. et al. (eds.), (1984): 69-113.

 

Brett, P. (Forthcoming). "Too Many Media in the Multimedia? A Study of the Effects of Combinations of Media on a Recall Task".

 

Brown, H. D. (1994). Principles of Language Learning and Teaching. Englewood Cliffs, NJ: Prentice Hall Regents.

 

Brumfit, C. J. (1980). Communicative Methodology in Language Teaching: the Roles of Fluency and Accuracy. Cambridge: Cambridge University Press.

 

Brumfit, C. J., Phillips, M. & Skehan, P. (eds.). (1985). Computers in English Language Teaching. A View from the Classroom. ELT Documents 122. Oxford: Pergamon Press.

 

Bryant, J. & Anderson, D. R. (1983). Children's Understanding of Television. New York: Academic Press.

 

Bull, S. (1994). "Student Modelling for Second Language Acquisition." Computers & Education 23, 1/2: 13-20.

 

Bush, D. & Terry, R. M. (eds.) (1997). Technology Enhanced Language Learning. Lincoln, Illinois USA: National Textbook Company.

 

Canale, M. (1983). "On some Dimensions of Language Proficiency." In Oller, J. W., Jr. (ed.). Issues in Language Testing Research, 333-342. Rowley, MA: Newbury House.

 

Cater, R. & McCarthy, M. (eds.) (1988). Vocabulary and Language Teaching. London: Longman.

 

Chapelle, C. A. (1998). "Multimedia CALL: Lessons to be Learned from Research on Instructed SLA." Language & Learning Technology, 2, 1: 22-34.

 

Cotton, J. W. & Klatzky, R. (eds.) (1978). Semantic factors in cognition. Hillsdale, NJ: Erlbaum.

 

Cowie, A. P. (1988). "Stable and Creative Aspects of Language Use." In Carter, R. & McCarthy, M. (eds.) (1988).

 

Dechert, H. W. & Raupach, W. (eds.) (1980a). Towards a Cross-Linguistic Assessment of Speech Production. Frankfurt a. M.: Peter Lang.

 

Dechert, H. W. & Raupach, W. (eds.) (1980b).Temporal Variables in Speech. The Hague: Mouton.

 

Dechert, H. W. (1984). "Second Language Production: Six Hypotheses". In Dechert, H., Möhle, D & Raupach, M. (eds.) (1984): 211-230.

 

Dechert, H. W. (1987). "Understanding Producing." In Dechert, H. W. & Raupach, M. (eds.) (1987): 229-237.

 

Dechert, H. W. & Raupach, M. (eds.) (1987). Psycholinguistic Models of Production. Norwood, NJ: Alex Publishing Corporation.

 

Dechert, H., Möhle, D. & Raupach, M. (eds.) (1984). Second Language Production. Tübingen: Gunter Narr.

 

Dodd, B. & Campbell, R. (eds.) (1987). Hearing by Eye: The Psychology of Lip-Reading. London: Lawrence Erlbaum.

 

Doughty, C. & Williams, J. (eds.) (1998). Focus on Form in Classroom Second Language Acquisition. Cambridge: Cambridge University Press.

 

Duker, S. (1974). Time-compressed Speech: An Anthology and Bibliography in Three Volumes. Metuchen, NJ: The Scarecrow Press Inc.

 

Duquette, L. & Painchaud, G. (1996). “A comparison of vocabulary acquisition in audio and video contexts.” The Canadian Modern Language Review  53,1:143-72.

 

Ellis, R. (ed.) (1987). Second Language Acquisition in Context. London: Prentice-Hall International.

 

Ellis, R. (1994).The Study of Second Language Acquisition. Oxford: Oxford University Press.

 

Ellis, R. (1997). SLA Research and Language Teaching. Oxford: Oxford University Press.

 

Fathman, A. K. (1980). “Repetition and Correction as an Indication of Speech Planning and Execution Processes Among Second Language Learners.” In Dechert, H. W. & Raupach, M. (eds.), 1980a: 77-85.

 

Fillmore, C. J., Kempler, D. & Wand, W. S-Y. (eds.) (1979). Individual Differences in Language Ability and Language Behavior. New York: Academic Press.

 

Fillmore, J.C. (1979). “On Fluency”. In Fillmore, C. J., Kempler, D. & Wang, W. S-Y. (eds.) (1979): 85-101.

 

Foster, P. & Skehan, P. (1996). "The Influence of Planning on Performance in Task-based learning." Studies in Second Language Acquisition 18: 299-324.

 

Foster, P. & Skehan, P. (1997). "Modifying the Task: the Effects of Surprise, Time and Planning Type on Task-based Foreign Language Instruction." Thames Valley Working Papers in English Language Teaching. Vol. 4.

 

Fox Tree, J. E. & Meijer, P. J. A. (1999). "Building Syntactic Structure in Speech." Journal of Psycholinguistic Research, 28, 1: 71-92.

 

Godwin-Jones, B. (2000). “Emerging Technologies: Speech Technologies for Language Learning.” Language Learning & Technology 3: 26-29. [On-Line] 26.08.(2000). Available URL: http://llt.msu.edu/vol3num2/emerging/index.html.

 

Goldman-Eisler, F. (1968). Psycholinguistics: Experiments in Spontaneous Speech. New York: Academic Press.

 

Gopnick, M. (1994). "Impairment of Tense in a Familial Language Disorder." Journal of Neurolinguistics 8: 109-133.

 

Grosjean, F. (1980). “Temporal Variables Within and Between Languages.” In Dechert, H. W. & Raupach, M. (eds.) (1980): 39-53.

 

Hall, S. S. (1999). “Journey to the Center of the Mind." [On-Line] 14.5.(2000). Available URL: http://www.fmri.org/journey.html.

 

Harben, P. (1999) "An Exercise in Applying Pedagogical Principles to Multimedia CALL Materials Design." ReCALL, 11, 3: [On-Line] 24.7.(2000). Available URL: http://www.rdg.ac.uk/AcaDepts/cl/ CALS/ph/multimed.htm.

 

Herron, C., Morris, M., Secules, T., and L. Curtis (1995). "A comparison of the effects of video-based versus text-based instruction in the foreign language classroom." The French Review 68: 775-796.

 

Higgins, J. (1995). Computers and English Language Learning. Oxford: Intellect Books.

 

Hoey, M. (1993). Data, Description, Discourse. Papers on the English Language in Honour John Mc H Sinclair. London: Harper Collins.

 

Hoven, D. (1999). "A Model for Listening and Viewing Comprehension in Multimedia Environments." Language Learning & Technology 3, 1: 88-103. [On-Line] 14.5.(2000). Available URL: http://llt.msu.edu/vol3num1/hoven/.

 

Hudson, T. (1982). "The Effects of Induced Schemata on the Short Circuit in L2 Reading: Non-decoding Factors in L2 Reading Performance." Language Learning, 32: 1-31.

 

Jakobson, R. & Halle, M. (1956). Fundamentals of Language. The Hague: Mouton.

 

Jamieson, J. & Chapelle, C. (1988). “Using CALL Effectively: What Do We Need to Know about Students?” System.16, 2: 151-162.

 

Johnson, P.N. (1988). The Computer and the Mind. Cambridge, Mass.: Harvard University Press.

 

Joiner, E. G. (1997). "Teaching Listening: How Technology Can Help." In Bush, M. D. & Terry, M. E. (eds.) (1997): 77-120.

 

 

Kecskés, I. (1988). “Computer Programs to Develop Both Accuracy and Fluency.” System 16, 1: 29-35.

 

Klein, W. (1986). Second Language Acquisition. Cambridge: Cambridge University Press.

 

Ko, P. (1992). "The Effects of Rate-controled Speech on Advanced Chinese EFL Learners' Short Term Listening Comprehension and Confidence." Unpublished Ph.D. dissertation, Florida State University, Tallahassee.

 

Krashen, S. (1978). "Individual Variation in the Use of the Monitor." In Ritchie, W. (ed.) (1978).

 

Krashen, S. D. & Terell, T. D. (1983). The Natural Approach: Language Acquisition in the Classroom. Oxford: Pergamon Press.

LaBerge, D. & Samuels, S. J. (eds.), (1977). Basic Processes in Reading: Perception and Comprehension. Hillsdale, NJ: Lawrence Erlbaum.

 

Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.

 

Langenmayr, A. (1997). Sprachpsychologie: Ein Lehrbuch. Göttingen: Hogrefe.

 

Leech, G., Myers, G., Thomas, J. (eds.) (1995). Spoken English on Computer. Harlow: Longman.

 

Leeson, R. (1975). Fluency and Language Teaching. London: Longman.

 

Lennon, P. (1984). “Retelling a Story in English as a Second Language.” In Dechert, W. et al. (eds.) (1984): 50-68.

 

Lennon, P. (1987). “Introspection and Intentionality in Advanced Second-language Acquisition”. Language Learning 39: 375-395.

 

Levelt, W. (1989). Speech. Cambridge, Mass: MIT Press.

 

Levy M. (1997). Computer-Assisted Learning. Context and Conceptualisation. Oxford: Oxford University Press.

 

Lewis, M. (1993). The Lexical Approach. Hove: Language Teaching Publications.

 

Lodge, D. (1979). The Modes of Modern Writing. London: Edward Arnold.

 

Long, D. R. (1991). "What Foreign Language Learners Say They Think About When Listening to Authentic Texts." Unpublished manuscript.

 

McDonough, J. & Shaw, C. (1993). Materials and Methods in ELT. Oxford: Blackwell.

 

McGurk, H. & MacDonald, J. (1976). "Hearing Lips and Seeing Voices." Nature 26, 4: 746-748.

 

Martínez-Lage, A. (1997). "Hypermedia Technology for Teaching Reading". In Bush, D. & Terry, R. M. (eds.) (1997): 121-163.

 

Mehnert, U. (Forthcoming). "The Effects of Different Lengths of Time for Planning on Second Language Acquisition." Studies in Second Language Acquisition.

 

Meringoff, L. K., Vibbert, M. M., Car, C. A., Fernie, D. E., Banker, G. S. & Gardner, H. (1983). "How Is Children's Learning from Television Distinctive? Exploiting the Medium Methodologically." In Bryant & Anderson (eds.) (1983): 151-179.

 

Nation, P. (1989). “Improving Speech Fluency.” System 17, 3: 377-384.

 

Nattinger, J. R. & DeCarrico, J. S. (1992). Lexical Phrases and Language Teaching. Oxford: Oxford University Press.

 

Oller, J. W. (1979). Language Tests at School: A Pragmatic Approach. London: Longman.

 

Oller, J. W. (1983). "Some Working Ideas for Language Teaching." In Oller, J. W. & Richard-Amato, P. A. (eds) Methods That Work. Rowley, MA: Newbury House.

 

Oller, J. W., Jr. (1983). Issues in Language Teaching and Research. Rowley, MA: Newbury House.

 

Omaggio, H. A. (1993). Teaching Language in Context: Proficiency-Oriented Instruction. 2nd ed. Boston: Heinle & Heinle.

 

Ostermeier, T. H. (1991). "Fast Talkers and Speeding Listeners: Television/Radio Commercials." Journal of the International Listening Association 5: 22-35.

 

Palmer, A. S. (1983). "Compartmentability and Integrated Control. In Oller, J. W. Jr. (ed.) (1983): 323-332.

 

Pawley, A. & Syder, F. (1983). "Two Puzzles for Linguistic Theory: Nativelike Selection and Nativelike Fluency." In Richards, J. C. & Schmidt, R. (eds.) (1983).

 

Pennington, M. C. (ed.) (1996). The Power of CALL. Houston: Athelstan.

 

Prabhu, N. S. (1987). Second Language Planning Pedagogy. Oxford: Oxford University Press.

 

Raupach, M. (1980). “Cross-Linguistic Descriptions of Speech Performance as a Contribution to ‘Contrastive Psycholinguistics’”. In Dechert, H. W. & Raupach, M. (eds.) (1980): 9-22.

 

Raupach, M. (1984). “Formulae in Second Language Speech Production.” In Dechert, et al. (eds.) (1984):114-137.

 

Rehbein, J. (1987). “On Fluency in Second Language Speech.” In: Dechert, W. & Raupach, M. (eds.) (1987): 97-105.

 

Reves, T. & Levine, A. (1988). “The FL Receptive Skills: Same or Different?” System. 16, 3: 327-336.

 

Richards, J.C. & Rodgers, T. S. (1986). Approaches and Methods in Language Teaching. Cambridge: Cambridge University Press.

 

Richards, J. C. & Schmidt, R. (eds.) (1983). Language and Communication. London: Longman.

 

Ritchie, W. (ed.) (1978). Second Language Acquisition Research. New York: Academic Press.

 

Rost, M. (1991). Listening in Action. New York: Prentice Hall.

 

Rubin, D. L. & Rafoth, B. A. In press. Oral language: A Criterion for Selecting Listenable Material.

 

Rumelhart, D. E. (1977). “Understanding and Summarizing Brief Stories.” In LaBerge, D. & Samuels, S. J. (eds.) (1977): 50-68.

 

Rumelhart, D. E., & Norman, D. A. (1978). "Accretion, Tuning and Restructuring: Three Modes of Learning." In Cotton, J. W. & Klatzky, R. (eds.) (1978): 37–53.

 

Rumelhart, D. E. (1987). "Understanding Understanding." In Dechert, H. W. & Raupach, M. (1987): 209-228.

 

Sajavaara, K. (1987). "Second Language Speech Production: Factors Affecting Fluency." In Dechert, H. W. & Raupach, M. (1987): 45-65.

 

Schmidt, R. (1990). "The Role of Consciousness in Second Language Learning." Applied Linguistics 11: 17-46.

 

Schmidt, R. (1992). “Psychological Mechanisms Underlying Second Language Fluency”. Studies in Second Language Acquisition 1, 4: 357-385.

 

Secules, T., Herron, C., and Tomessello, M. (1992). "The effect of video context on foreign language learning." The Modern Language Journal 76: 480-490.

 

Segalowitz, N. & Gatbonton, E. (1995). “Automaticity and Lexical Skills in Second Language Fluency: Implications for Computer Assisted Language Learning.” Computer Assisted Language Learning 82, 3: 129-149.

 

Seliger, H. W. (1980). "Utterance Planning and Correction Behavior: Its Function in the Grammar Construction Process for Second Language Learners." In Dechert, H. W. & Raupach, M. (eds.) (1980): 87-99.

 

Shank, R. C. & Abelson, R. P. (1977). Scripts, Plans and Goals and Understanding. Hillsdale, NJ: Lawrence Erlbaum Ass.

 

Singer J. L. & Singer, D. G. (1983). "Implications of Childhood Television Viewing for Cognition, Imagination, and Emotion." In Bryant & Anderson (eds.) (1983): 265-295.

 

Skehan, P. & Foster, P. (1997). "Task Type and Task Processing Conditions as Influences on Foreign Language Performance." Language Teaching Research 1, 3: 185-211.

 

Skehan, P. (1998). A Cognitive Approach to Language Learning. Oxford: Oxford University Press.

 

Summerfield, Q. (1987). “Some Preliminaries to a Comprehensive Account of Audio-visual Speech Perception.” In Dodd, B. & Campbell, R. (eds.) (1987): 3-51.

 

Sweller, J. (1994). "Cognitive load theory learning difficulty and instructional design." Learning and Instruction, 4: 295-312.

 

Tarone, E. E., Gass, S. M. & Cohen, A. D. (1994). Research Methodology in Second-Language Acquisition. Hillsdale, NJ: Lawrence Erlbaum Ass.

 

Thompson, I. & Rubin, J. (Forthcoming). "Can Strategy Training Improve Listening Comprehension?" Foreign Language Annals.

 

Towell, R. (1987). “Variability and Progress in the Language Development of Advanced Learners of a Foreign Language.” In Ellis (ed.) (1987).

 

Underwood, M. (1990). "Task-related Variation in Past-tense Morphology." Unpublished M.A. thesis, Institute of Education, London University.

 

Vollmer, H. J. & Sang, F. (1983). "Competing Hypotheses about Second Language Ability: a Plea for Caution." In Oller, J. W. Jr. (ed.) (1983): 29-79.

 

White, G. (1998). Listening. Oxford: Oxford University Press.

 

Wiese, R. (1984). “Language Production in Foreign and Native Languages: Same or Different?” In Dechert et al. (eds.) 1984:11-25.

 

Appendix 1

 

N.N.    (SR= Speech Rate / PTR=Phonation-time ratio / AR: Artic. Rate / LP: Length Pause / LR: Length Run / Rep: Repeat)

 

Posttest Group T: Fluency 26.6.00  Disk 4/5

SR

LP

LR

Rep

04"

OK, the first picture there is Sir Henry, (ehm) Dr Watson and Sherlock Holmes (ehm) they come back to (ehm) isn't it Stapelton House? (teacher prompt) Merripit House OK

35

 

11

8

4

 

18"

This is the house where the Stapelton's live (ehm) (ehm) ... [4,0"] ... /Sir Stapelton/ (ehm) Mr Stapelton

19

4,0

22

4

1

28"

he just died in the moor. /He/ he got into the quicksands somehow and there /he sunk/ he

21

 

 

2

38"

sunk down (ehm) ... [3,0"] ...Sherlock Holmes and Dr Watson Sir Henry /they/ they couldn't help him

18

3,0

28

1

48"

because they were quite far away and it was too dangerous to enter the moor because Dr Watson also sank in this

31

 

 

 

58"

quicksands /but then he/ (ehm) but Sherlock Holmes was able to help him. So he could survive. But

20

 

52

1

1'08"

Mr Stapelton just died so they returned to Merripit House to tell (ehm) ... [1,7"] ... Mrs Stapelton what happened with her

28

1,7

33

 

1'18"

husband. So (ehm). They arrive at the house and /they/ they recognize that

15

 

13

1

1'28"

the door is locked so they need to crack the door. And there /they/ they

15

 

 

1

1'38"

find Mrs Stapelton (ehm) .... [2,4] ... She was kind of hang up. /She was/ she isn't killed but (ehm)

18

2,4

34

12

 

1'48"

/she can't/ she can't move She was helpless ... [2,8"] ... so /they/ they save her

14

2,8

9

2

1'58"

then they take her down from this (yeah) ...................[5,6"] ... (G: Seils) (teacher prompt)

8

5,6

12

 

2'08"

......... ropes (yes) (ehm) .... [3,8"] ... Then (ehm) Miss Stapelton (ehm) wakes up. And she tells

11

3,8

2/1

4

 

2'18"

them /that he escaped/ that he got away. She tells everything. She

16

 

 

1

2'28"

tells /that/ that he is a monster and that he is the murderer and she's scared of him. She's very confused.

26

 

47

1

2'38"

....................... [7,4"] .../She tells/ She tells Sherlock Holmes about the the hiding

13

7,4

 

1

2'48"

place that is an old mine /where she/ where /he/ he hides himself. Well (ehm)

15

 

28

2

2'58"

(yeah) ................ [8,4" ]... But Sir didn't you tell me that he already sank

12

8,4

 

 

3'08"

/into the/ into the .... [2,2"] ... quicksands? But Stapelton just tells them in the second last

20

2,2

18

1

3'18"

picture that /where he used to/ where he used to hide himself. ... [4,1"] ... Well OK Sherlock Holmes

20

4,1

28

1

3'28"

knows where he could find him and next morning they arrive at exactly this place where Miss Stapelton told them

25

 

31

 

3'38"

(ehm) ...[3,1"] ...  there /they can only/ they can find just find (ehm) /an/ an shoe of Mr

16

3,1

10

2

3'48"

Stapelton. .... [2,7"] ... So they then really know that ... [1,8"] .............(teacher prompt -5')

10

2,7

9

7

 

3'58"

..... No this is Sir Henry's shoe I think. Because Mr Stapelton (ehm) needed this one to let /the

23

1,8

16

1

4'08"

dog/ the killer dog know (ehm) what Sir Henry smells like yes exactly

16

 

13

 

4'18"

/So they had/ /they/ they knew exactly how /he/ he made everything with the dog. /They found also/, they found out that the dog

30

 

 

3

4'28"

the bliming and that came from the dog was /in the worst/ in the worst sense paint and something like that

23

 

 

1

4'38"

So everything was clear and now he's dead /so he gave justice/ ...[2,4"] ...

16

2,4

77

1

4'48"

he was brought to justice in a way. .......... [4,4"]...So that's it. ...(4'57")

9

4,4

9

 

4'58"

..................... (teacher prompt –20')

 

 

9

3

 

5'08"

....................

 

2,0

 

 

5'18"

This was because he found out that Stapelton was (ehm) ... [2,0"] ... like a cousin to Sir Henry

20

2,0

12

1

5'28"

Yes a cousin and (ehm) .... [2,0"] ... /he lived/ he lived sometime in South America ... [3,3"] ...

17

3,3

13

13

 

5'38

He tried /to/ to kill all these follows /of/ of (ehm) /Sir Charles/ Sir Hugo So he killed Sir Charles and now

23

 

11

3

5'48"

Sir Henry and he would be the only follow and he could own everything. All would be /his/ his

25

 

 

2

5'58"

own /in his own/ ...(6'08")

4

 

41

30

 

 

632

67

635

 

 

 

 

19

34

 

 

Pauses:  67"  ; total speaking time: 6,04"-25" (teacher prompts)= 5,39", speaking time minus pauses:  4'32"

 

 

 

 

 

SR: 634/5,6"=113/60=  1,9 syll/sec

 

 

 

 

 

PTR: (4,5"*100)/5,6"= 80  %

 

 

 

 

 

AR: 634/4,5"= 101/60= 2,3 syll/sec

 

 

 

 

 

LP: 67"/19=  3,5 sec

 

 

 

 

 

LR:634/34=18,6  syll

 

 

 

 

 

Rep: 30

 

 

 

 

 

 

 

 

 

 

 

 

Posttest Group T: Accuracy 26.6.00  Disk 4/5

SC

ER

04"

OK, the first picture there is Sir Henry, (ehm) Dr Watson and Sherlock Holmes (ehm) they come back to (ehm) isn't it Stapelton House? (teacher prompt) Merripit House OK

 

 

18"

This is the house where the Stapelton's live (ehm) (ehm) ... [4,0"] ... /Sir Stapelton/ (ehm) Mr Stapelton

 

 

28"

he just died in the moor. /He/ he got into the quicksands somehow and there /he sunk/ he

 

1

38"

sunk down (ehm) ... [3,0"] ...Sherlock Holmes and Dr Watson Sir Henry /they/ they couldn't help him

 

 

48"

because they were quite far away and it was too dangerous to enter the moor because Dr Watson also sank in this

 

1

58"

quicksands /but then he/ (ehm) but Sherlock Holmes was able to help him. So he could survive. But

 

 

1'08"

Mr Stapelton just died so they returned to Merripit House to tell (ehm) ... [1,7"] ... Mrs Stapelton what happened with her

 

1

1'18"

husband. So (ehm). They arrive at the house and /they/ they recognize that

 

 

1'28"

the door is locked so they need to crack the door. And there /they/ they

 

 

1'38"

find Mrs Stapelton (ehm) .... [2,4] ... She was kind of hang up. /She was/ she isn't killed but (ehm)

1

2

1'48"

/she can't/ she can't move She was helpless ... [2,8"] ... so /they/ they save her

 

1

1'58"

then they take her down from this (yeah) ...................[5,6"] ... (G: Seils) (teacher prompt)

 

 

2'08"

......... ropes (yes) (ehm) .... [3,8"] ... Then (ehm) Miss Stapelton (ehm) wakes up. And she tells

 

 

2'18"

them /that he escaped/ that he got away. She tells /.../ everything. She

 

1

2'28"

tells /.../ /that/ that he is a monster and that he is the murderer and she's scared of him. She's very confused.

 

 

2'38"

....................... [7,4"] .../She tells/ She tells Sherlock Holmes about the the hiding

 

 

2'48"

place that is an old mine /where she/ where /he/ he hides himself. Well (ehm)

 

 

2'58"

(yeah) ................ [8,4" ]... But Sir didn't you tell me that he already sank

 

 

3'08"

/into the/ into the .... [2,2"] ... quicksands? But Stapelton just tells them in the second last

 

 

3'18"

picture that /where he used to/ where he used to hide himself. ... [4,1"] ... Well OK Sherlock Holmes

1

1

3'28"

knows where he could find him and next morning they arrive at exactly this place where Miss Stapelton told them

 

3

3'38"

(ehm) ...[3,1"] ...  there /they can only/ they can find just/ find (ehm) /an/ an shoe of Mr

 

1

3'48"

Stapelton. .... [2,7"] ... So they then really know that ... [1,8"] .............(teacher prompt -5')

 

1

3'58"

..... No this is Sir Henry's shoe I think. Because Mr Stapelton (ehm) needed this one to let /the

 

 

4'08"

dog/ the killer dog know (ehm) what Sir Henry smells like yes exactly

 

 

4'18"

/So they had/ /they/ they knew exactly how /he/ he made everything with the dog. /They found also/, they found out that the dog

1

1

4'28"

the bliming and that came from the dog was /in the worst/ in the worst sense paint and something like that

 

1

4'38"

So everything was clear and now he's dead /so he gave justice/ ...[2,4"] ...

 

 

4'48"

he was brought to justice in a way. .......... [4,4"]...So that's it. ...(4'57")

 

 

4'58"

..................... (teacher prompt –20')

 

 

5'08"

 

 

 

5'18"

This was because he found out that Stapelton was (ehm) ... [2,0"] ... like a cousin to Sir Henry

 

 

5'28"

Yes a cousin and (ehm) .... [2,0"] ... /he lived/ he lived /.../ sometime in South America ... [3,3"] ...

 

1

5'38

He tried /to/ to kill all these follows /of/ of (ehm) /Sir Charles/ Sir Hugo So he killed Sir Charles and now

 

1

5'48"

Sir Henry and he would be the only follow and he could own everything. All would be /his/ his

 

 

5'58"

own /in his own/ ...(6'08")

 

1

 

 

3

18

 

Self-corrections: 272"/3=90,6"          / IL-forms, unperceived errors: 272"/18=15,1"

 

 

 

 

 

 

 

Name

SR

syll/sec

PTR

%

AR

syll/sec

LP

secs

LR

syll

Runs

Rep

SC

ER

 

N.N.

1,9

80

2,3

3,5

18,6

34

30

91

15

 

Appendix 2

 

Pretest

 

Audio group

 

SR

PTR

AR

LP

LR

Rep

SCS

ER

 

SR

PTR

AR

LP

LR

Rep

SCS

ER

 

Test subjects

syll/

sec

%

syll/

sec

secs

syll/

run

no

secs

secs

 

syll/

sec

%

syll/

sec

secs

syll/

run

no

secs

secs

 

T1 (fem)

2,4

95

2,5

2,6

32

11

152

9

 

2,0

91

2,2

2,3

24

14

51

11

 

T2 (male)

2,4

90

2,6

3,0

19

39

54

18

 

2,6

85

3,0

2,7

24

25

56

26

 

T3 (male)

2,1

80

2,7

2,8

20

17

58

7

 

2,0

75

2,7

3,5

22

22

108

5

 

T4 (fem)

2,0

88

2,2

4,7

14

11

70

6

 

1,8

77

2,3

3,5

13

16

39

7

 

T5 (fem)

1,9

78

2,4

2,9

18

11

74

9

 

2,0

81

2,5

3,6

33

15

90

12

 

T6 (male)

1,8

81

2,2

3,7

11

35

52

10

 

1,9

80

2,3

3,5

19

30

91

15

 

T7 (male)

2,7

90

3,0

2,3

25

15

136

19

 

2,4

83

2,9

3,1

21

27

38

30

 

T8 (male)

2,0

90

2,2

2,2

29

10

53

20

 

1,9

93

2,1

2,8

34

19

61

26

 

T9 (male)

2,0

92

2,2

2,5

15

12

69

6

 

1,6

83

1,9

2,7

14

13

41

15

 

T10 (fem)

2,5

83

3,0

6,0

40

8

272

68

 

2,1

90

2,4

2,7

25

16

75

50

 

T11 (fem)

1,8

75

2,5

5,0

25

6

74

7

 

***

***

***

***

***

***

***

***

 

Means

2,1

85,6

2,5

3,4

22,4

15,9

96,7

16,3

 

2,0

83,8

2,4

3,0

22,9

19,7

65,0

19,7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pretest

 

Video group

 

V1 (male)

2,3

81

2,8

3,6

20

22

47

8

 

2,2

84

2,7

4,1

19

23

46

9

 

V2 (male)

1,1

40

2,7

3,5

10

9

94

6

 

1,3

65

2,1

3,3

9

21

50

9

 

V3 (male)

1,7

70

2,4

3,8

16

23

48

26

 

2,0

83

2,4

2,6

11

34

78

20

 

V4 (fem)

2,8

90

3,1

2,0

29

2

205

26

 

2,4

92

2,6

2,4

43

5

146

22

 

V5 (fem)

2,2

90

2,5

2,4

14

10

78

8

 

2,1

88

2,4

4,6

11

17

54

7

 

V6 (male)

2,2

82

2,7

2,9

13

43

62

5

 

3,1

90

3,4

2,6

21

36

70

6

 

V7 (male)

1,7

81

2,1

4,2

12

19

270

5

 

1,8

84

2,2

3,1

9

17

57

5

 

V8 (male)

2,2

82

2,7

4,2

17

20

41

9

 

2,7

78

3,4

4,3

18

23

78

8

 

V9 (male)

2,1

85

2,5

3,7

18

14

68

8

 

2,2

77

2,8

5,4

18

25

73

7

 

V10 (male)

2,1

75

2,8

4,3

17

20

61

5

 

2,3

90

2,5

3,4

17

29

143

6

 

Means

2,0

77,6

2,6

3,5

16,6

18,2

97,4

10,6

 

2,2

83,1

2,7

3,6

17,6

23,0

79,5

9,9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


[1] In the present context, narration is broadly understood as a type of discourse characterised by description of a sequence of fictional events in time.

[2] Such a restricted view of L2 teaching was prevalent during the early phase of the "communicative revolution." (McDonough & Shaw 1993: 32)

[3] The overall assumption on which the present study rests is that both interactive and non-interactive input will lead to some form of acquisition. To date such an assumption is still controversial since we only dispose of indirect evidence (Ellis 1994: 286).

[4] The notion of ‘schema’ can be traced back to Immanuel Kant who describes it as a kind of rule for the synthesis of all possible representations of a concept (Brenzel 1984: 69).

[5] Using functional MRI (magnetic resonance imaging), neuroscientists have shown that the brain parcels out two related capabilities – one to create stories and the other to articulate them in speech – at essentially the same place in the two hemispheres. The storytelling area is located in a region called the “inferior frontal gyrus” and is the mirror image of the site on the left, dominant side of the brain that controls speech (Hall 1999: 4-5).

[6] The metaphor stems from computer technology and refers to "an operating system feature allowing several users to run several tasks concurrently on one processor, or in parallel on many processors ... time-sharing is multitasking for multiple users." Free Online Dictionary of Computing sv. "time-sharing". [On-Line] 9.08.(2000). Available URL: http://foldoc.doc.ic.ac.uk/foldoc/index.html.

[7] A different, asynthetic view, was advocated by researchers such as Singer and Singer (1983: 271) who came to the conclusion that auditory stimuli appeared to be processed by a more action-oriented system linked especially to the left hemisphere of the brain, while visual processing involved "a more receptive, global, spatially oriented or imagery focused" activity linked to the right half of the brain.

[8] As quoted in Lodge (1979: 79). Jakobson extends this distinction to describe different literary periods and styles. D. Lodge, among others, applied this dichotomy successfully to literary criticism in his The Modes of Modern Writing (1979).

[9] Skehan (1998: 5, 108; 1997: 191) adds as a third variable, complexity, "a willingness to take risks, to try out new forms even though they may not be completely correct" and which he measures by analysing the degree of subordination per communication unit. Nation (1989: 381-382), in his comparison of several renderings of the same text, introduces the variable, "control of content", i.e. other ways besides "speaking faster" of "fitting the talk into the reduced time available" like "omitting pieces of information and changing the grammatical constructions so that the same information is presented more efficiently." An analysis of these additional variables would go beyond the scope of the present study.

[10] Wiese (1984: 18) suggests associating hesitation phenomena like pauses and repetitions as means of facilitating monitoring in the planning stage, while corrections would be evidence for post-articulatory monitoring. Monitoring is not understood here in Krashen's special sense, but in the sense that "all speakers monitor their speech to maintain well-formedness and, even more, to communicate their intentions" (Wiese 1984: 22).

[11] Several researchers distinguish between chunks or formulas that function as "fillers" and others that act as "organisers." While "fillers" give time for further planning activities, "organisers" contribute to the development of ongoing speech by helping learners to establish a structure for phrases and sentences (Ellis 1994: 394-395; Raupach 1984: 123-128; Towell 1987; Lennon 1984). In line with the present approach the former may be classified as paradigmatic and pertaining to fluency and the latter as syntagmatic and attributable to accuracy.

[12] The following categorisation of temporal speech variables expands on Grosjean (1980: 39-42).

[13] Grosjean (1980) makes a distinction between filled pauses (e.g. "ehm", "yeah", "OK") and unfilled pauses. However, in our experiment we found no rationale for this distinction. Whether a pause was filled or unfilled appeared to depend on learner characteristics and individual speech strategies, as will be shown below.

[14] Skehan (1997: 185) opts for a different method to analyse the above variables by measuring fluency through the number of pauses and accuracy through the percentage of error-free clauses. This line of approach, however, was considered too restrictive for our purposes.

[15] Oxford Bookworms, stage 3, Oxford University Press, 1989

[16] The audio-version of the original novel was not used for two reasons: 1. As a rendering of the original text, it would have taken much more time to listen to than the rendering of the film and, therefore, the variable "time" would have become uncontrollable, 2. As an audio-book it follows the written text which was designed to be read in the first place as opposed to the dramatised play which was specifically created to be listened to.

[17] The pictures used as a support for the oral activity were culled from a comic strip version of Doyle's book (Foreman S, & Larkham, R. (1998). The Hound of the Baskervilles, Detective English. London: Chancerel International Publishers).

[18] Ellis (1994: 672) lists among means of eliciting experimental data: reading aloud, structured exercises, completion tasks, imitation tasks, translation, guided composition, and story-retelling.

[19] As a matter of fact, several studies have shown that planning time modifies fluency and complexity processing, leaving accuracy unaffected. (Foster & Skehan 1996, 1997; see also studies by Ellis 1987; Underwood, 1990; Mehnert (forthcoming).

[20] For the t-test I used a java applet freely available on the net. [On-Line] 24.7.(2000). Available URL: http://pomacea.cox.miami.edu/~rob/Students_t.html

[21] The distinction between filled and unfilled pauses which Grosjean (1980) suggests was not taken into account since a perusal of the data did not support it. While some learners would fill most pauses, others made sparse use of this device. Whether a pause was filled of unfilled seemed rather to depend on learner characteristics and speech strategies.

[22] The "zero position" also entails a rejection of planned and unplanned intervention in L2 learning, a sequel to the hypothesis which, however, is not being endorsed in the present study.

[23] For a research overview see Ellis (1997: 50-51).

[24] For a list of research findings in developmental sequences cf. Ellis (1994: 20, 73-117).

[25] A typical case of immersion learners are semi-bilinguals who are fluent but whose accuracy performance levels demand repair.

[26] For a comprehensive overview of new developments, cf. Godwin Jones 2000.