how grammar and vocabulary interact with oral … papers from the twenty-fourth international...

16
Selected Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency Giles, Witton-Davies (朱杰佑) National Taiwan University [email protected] Oral proficiency in foreign languages is often assessed using distinct criteria, including fluency, lexis, grammar and pronunciation. These criteria are to some extent independent of each other: some speakers are fluent, but have poor grammar, while those with good grammar may not be fluent. However, as proficiency improves, these four skills generally develop together. This study investigates whether hesitations and repairs are more related to problems with grammar or vocabulary, and also examines to what extent grammatical and lexical abilities affect oral fluency. The data was provided by seventeen students of English at a Taiwanese university. They performed picture narrative tasks, which were subsequently recorded, transcribed, and analyzed using Praat software. Fluency analysis involved measures of speech rate, hesitation and repair, while the range and accuracy of grammar and lexis were also measured. Hesitations were then categorized according to whether they occurred before form or function words, this being taken to indicate whether the disfluency related to lexical or grammatical difficulty. Repairs were also separated into two groups, depending on whether changes were made to lexis or to grammar. The figures for the two types of hesitations and repairs could thus be compared. Finally, correlations were calculated between the fluency and the grammatical and lexical measures. Overall, the results suggest that lexical knowledge affects fluency more than grammatical range and accuracy. Reasons for this finding are discussed in the context of theories of speech production, and some implications for teaching oral English are considered. INTRODUCTION This research arises out of 18 years of experience teaching oral English at university level in Taiwan. Taiwanese university undergraduates often have a great amount of explicit knowledge of English, and may be able to manipulate the language in order to pass required English tests, based on what they have learned in at least ten years of English classes at schools and language schools. However, they are often unable to use the language to communicate effectively in speech or in writing. Speech is often slow, hesitant and full of self corrections. University oral training or conversation classes aim to redress the balance, but improvement, if any, is very gradual in my experience. Although it is easy to conclude what students do not need (more traditional grammar rules and exercises), it is more difficult to establish what they do need, and what will contribute to improved oral skills and fluency. This research provides evidence for the key role played in fluency by vocabulary. What is fluency and how can it be measured? It will help to explain briefly how the word fluency is used in this paper. In everyday life, linguistic fluency is often equated to overall competence in a second language. However, as Lennon (1990, p. 389) points out, “in its narrower sense, fluency in EFL refers to one, presumably isolatable, component of oral proficiency.” This component involves temporal factors such as speed, hesitations and repairs, and is distinct from other elements of proficiency such as accuracy and range of language. Within this notion of fluency, Segalowitz draws a three-way distinction between perceived, utterance, and cognitive fluency (Segalowitz, 2010). Perceived fluency is the perception of another person's speech as being produced smoothly and efficiently in a given context, while

Upload: hatuong

Post on 09-Mar-2018

222 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Selected Papers from the Twenty-fourth International Symposium on English Teaching

173

How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

National Taiwan University [email protected]

Oral proficiency in foreign languages is often assessed using distinct criteria, including fluency, lexis, grammar and pronunciation. These criteria are to some extent independent of each other: some speakers are fluent, but have poor grammar, while those with good grammar may not be fluent. However, as proficiency improves, these four skills generally develop together. This study investigates whether hesitations and repairs are more related to problems with grammar or vocabulary, and also examines to what extent grammatical and lexical abilities affect oral fluency. The data was provided by seventeen students of English at a Taiwanese university. They performed picture narrative tasks, which were subsequently recorded, transcribed, and analyzed using Praat software. Fluency analysis involved measures of speech rate, hesitation and repair, while the range and accuracy of grammar and lexis were also measured. Hesitations were then categorized according to whether they occurred before form or function words, this being taken to indicate whether the disfluency related to lexical or grammatical difficulty. Repairs were also separated into two groups, depending on whether changes were made to lexis or to grammar. The figures for the two types of hesitations and repairs could thus be compared. Finally, correlations were calculated between the fluency and the grammatical and lexical measures. Overall, the results suggest that lexical knowledge affects fluency more than grammatical range and accuracy. Reasons for this finding are discussed in the context of theories of speech production, and some implications for teaching oral English are considered.

INTRODUCTION This research arises out of 18 years of experience teaching oral English at university level

in Taiwan. Taiwanese university undergraduates often have a great amount of explicit knowledge of English, and may be able to manipulate the language in order to pass required English tests, based on what they have learned in at least ten years of English classes at schools and language schools. However, they are often unable to use the language to communicate effectively in speech or in writing. Speech is often slow, hesitant and full of self corrections. University oral training or conversation classes aim to redress the balance, but improvement, if any, is very gradual in my experience. Although it is easy to conclude what students do not need (more traditional grammar rules and exercises), it is more difficult to establish what they do need, and what will contribute to improved oral skills and fluency. This research provides evidence for the key role played in fluency by vocabulary.

What is fluency and how can it be measured? It will help to explain briefly how the word fluency is used in this paper. In everyday life,

linguistic fluency is often equated to overall competence in a second language. However, as Lennon (1990, p. 389) points out, “in its narrower sense, fluency in EFL refers to one, presumably isolatable, component of oral proficiency.” This component involves temporal factors such as speed, hesitations and repairs, and is distinct from other elements of proficiency such as accuracy and range of language.

Within this notion of fluency, Segalowitz draws a three-way distinction between perceived, utterance, and cognitive fluency (Segalowitz, 2010). Perceived fluency is the perception of another person's speech as being produced smoothly and efficiently in a given context, while

Page 2: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

174

utterance fluency refers to the features of speech that make it sound fluent. Cognitive fluency, on the other hand, is the label given to the underlying processing knowledge and skills that allow for speech to be produced fluently. Perceived fluency can be measured by asking for raters' opinions of speech samples, but unfortunately people tend to use different criteria to judge fluency. Cognitive fluency, meanwhile, is not easy to measure directly, although it has been attempted with some success (De Jong, Steinel, Florijn, Schoonen, & Hulstijn, 2013; Segalowitz, 2010). This leaves utterance fluency, embodied by a set of temporal variables, as the commonest target for research and discussion.

In its turn, utterance fluency has often been divided into three aspects – speed, breakdown and repair fluency (Skehan, 2003; Tavakoli & Skehan, 2005). Speed fluency is measured as articulation or speech rate; breakdown fluency relates to different kinds of pausing; while repair fluency is the result of the amount (or relative absence) of repair. Given that some pausing is a necessary feature of all speech, and that it is only when pauses appear in the wrong places and last for too long that they detract from fluency, I prefer the term “hesitation fluency” to the rather misleading label “breakdown fluency, and use hesitations to refer not only to silent pauses, but also filled pauses and repetitions.

Measuring fluency In this section I will consider how fluency can best be measured. Single construct measures (e.g. AR – articulation rate) are best for research into the nature of fluency (Bosker, Pinget, Quené, Sanders, & De Jong, 2013), while combined measures (e.g. SR and PSR – speech rate and pruned speech rate) are more useful for global fluency assessment where detailed measures of different aspects of fluency are not required, and a single, simple measure of fluency is more useful (Mehnert, 1998).

Speed fluency is best measured by AR, rather than SR, which combines speed with pausing. An alternative is mean length of syllable, used by de Jong and others in their research (Bosker et al., 2013; De Jong, Groenhout, Schoonen, & Hulstijn, 2013).

Hesitation fluency should include not only silent pauses, but also filled pauses, and the two are best combined, although this has been done only occasionally in fluency research (Duez, 1985; De Jong, Steinel, et al., 2013). The choice of one kind of pause rather than the other seems to be a matter of a speaker's personal style rather than of speaking ability. Repetitions should also be included as hesitations rather than as repairs, as they do not involve any self-correction, but rather serve the same function as pauses, in that they allow time for the speaker to plan what to say next (Duez, 1985).

Pauses are a natural part of most kinds of unscripted speech, and they only become hesitations when they are too frequent, too long and/ or come in the middle of clauses or constituents. Pause location certainly needs to be taken into account, as this is what most clearly distinguishes native speaker from learner pausing patterns (Foster & Tavakoli, 2009; Tavakoli, 2011). It is therefore the overall time and frequency of hesitations, together with their location, that should be considered when evaluating fluency.

While speed and hesitations may be thought to reflect linguistic processing ability (or “cognitive fluency”), repair seems to represent a different facet of fluency – namely, the effects of monitoring and self-correction. When a speaker self-corrects, or stops and restarts, speech will tend to sound disfluent. False starts and reformulations of grammar and lexis are the two basic kinds of repair. The amount of repair in speech depends not only on the accuracy of speech (which determines the need for repair), but also on speakers' confidence in their own accuracy (often correct language is corrected unnecessarily or erroneously), their ability to monitor their own output, and the extent to which they wish to interrupt the flow of speech in order to self-correct. Using factor analysis, Tavakoli and Skehan (2005) found that repair fluency loaded onto a different factor to speed and breakdown (hesitation) fluency, thus

Page 3: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

How Grammar and Vocabulary Interact with Oral Fluency

175

supporting the idea that repair fluency is rather different from speed and hesitation fluency.

Fluency and lexis It is interesting to speculate whether grammar or lexis have a greater repercussion on

fluency. One way to address this question is by looking for correlations between different types of linguistic knowledge and fluency. In one study, knowledge of vocabulary and the ability to produce correct sentence intonation were found to be the best predictors of speaking proficiency (de Jong, Steinel, Florijn, Schoonen, & Hulstijn, 2012). Lexical knowledge on its own has therefore sometimes been used as an overall indicator of oral proficiency (De Jong, Groenhout, et al., 2013). The relevance of grammar to fluency is, however, less certain. In a study related the the two just cited, knowledge of lexis found to correlate more strongly than grammatical knowledge with fluency variables such as AR (the equivalent used here was mean duration of syllable) silent pauses, filled pauses, repetitions and reformulations (De Jong, Steinel, et al., 2013).

Another approach is to analyse fluency, and in particular hesitations, to see to what extent these are influenced by grammatical and lexical factors. Goldman Eisler (1958) measured the transitional probabilities of words, with some words in spontaneous speech showing higher statistical probabilities of occurrence than others. Pauses were found most likely to occur after the most predictable words and immediately before the least predictable ones. Words after hesitations were found to be ten times less predictable than words before hesitations. Similarly, Beattie and Butterworth (1979) provided evidence for hesitations being associated with intervals before less frequent vocabulary, while they found that hesitations also came before words of low contextual probability (even when frequency was constant).

Building on Goldman Eisler's work, Maclay and Osgood (1959) for L1 speech and Fathman (1980) for L2 speech found evidence from the positioning of hesitations before content words that they were often the result of a search for lexis. Maclay and Osgood looked at the words following hesitations, arguing that where these were lexical words, this indicated that he speaker was looking for vocabulary, whereas grammatical or function words were a sign of grammatical problems. They found that most pauses and repeats came before lexical rather than function words. Fathman also found that among young learners of second language English most repairs were focused on meaning, as they involved lexical, rather than grammatical, changes.

Unfortunately, little recent fluency research has followed up on the findings described in the two previous paragraphs on the nature of particular hesitations and repairs in second language speech. However, analyses used in work on lexical density were used in the present study for the purpose of distinguishing lexical from function words. Lexical density can be defined as the proportion of lexical or content words (such as most nouns, verbs and adjectives), as compared to form or grammar words (e.g. determiners, auxiliary and modal verbs, and linking words). Halliday (1989) and Ure (1971) used lexical density as a way of distinguishing the qualities of different qualities of text, with higher figures in both written language as well as in more formal, scripted kinds of speech, and lower figures in spontaneous speech, as exemplified by everyday conversation (O’Loughlin, 1995). Lexical density is not used directly as a measure in this study, but the distinction between lexical and grammatical items is used to assess the comparative influence of lexis and grammar on hesitations.

Levelt's Blueprint of the Speaker One of the most widely accepted models of speaking used in second language

acquisition research is that of Levelt (1989, 1999). The model, amongst many other insights, provides an account of the parts played by implicit and explicit knowledge. The basic model distinguishes three processes of language production: conceptualisation, formulation and

Page 4: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

176

articulation. Conceptualisation is the least problematic in a second language, as it involves the preparation of a preverbal message, and is therefore independent of any particular language. It is the move to formulation that is more difficult in an L2, as it involves, on the one hand, a search for required lexis and, on the other, the adoption of appropriate morphosyntactic forms and patterns. Articulation is the third stage, where a phonological plan is converted into actual sounds. Although this too can often be very problematic in a second language, as it may be difficult to produce the necessary sounds and intonation, these difficulties will not necessarily affect fluency.

Of the three stages in speech production, therefore, formulation processes are most likely to be slow in a second language, and will also tend to produce errors of form, syntax and lexical choice. Formulation of grammar in L1 is viewed by Levelt as an automatic process, without any need for attention, while lexical knowledge is seen as conscious, meaning that lexical choices can also be made consciously, especially in cases of difficulty.

This leaves two possibilities for the fluency of L2 speech, depending on whether formulation is automatic or not in a second language. If L2 speaking follows the same pattern as that of L1, then it is likely that delays in making lexical choices will affect fluency more than problems relating to morphosyntax. This is because while delays accessing lexis will inevitably lead to hesitations, as speech cannot proceed until a lexical choice is made, formulation of grammar will be processed automatically (while errors will inevitably be made). If, on the other hand, as suggested by Paradis (2009), L2 speakers use explicit or declarative knowledge in formulation, rather than implicit knowledge, then grammatical processing will be much slower than in L1, meaning that grammatical problems will cause at least as much disfluency as do lexical ones.

Aims of this study This study explores the relationships between fluency, on the one hand, and lexis and

grammar on the other. It does so by considering the following research questions:

1. Do hesitations (silent pauses, filled pauses and repetitions) tend to come more often before lexical or grammatical words? 2. Are reformulations and false starts more often related to grammatical or to lexical problems? 3. Which correlates more strongly with fluency measures – grammatical range and accuracy or lexical range and accuracy?

METHODOLOGY Data

The data used here comes from a study of the fluency development over four years at university of a group of 17 English learners in Taiwan. They completed two tasks, a monologue picture-story and a discussion-dialogue, in the first and final (eighth) semesters of their degree courses, but only the picture narrative monologues are used here. Each student had 2-4 mins for telling their story, although some spoke for shorter or longer times than this. Ten minutes of preparation time, together with materials for each of the two tasks, were provided before the tasks were completed. Students spoke in pairs, in order to create a more communicative context, with the listener having a set of pictures to put in order as they listened. A total of 34 narratives, from 17 speakers in year 1 and year 4 of their courses, are used here. Full details of the previous study can be found in Witton-Davies (2014). However, the present study includes new analysis of the data, particularly concerning the contexts of hesitations, as described below.

Page 5: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

How Grammar and Vocabulary Interact with Oral Fluency

177

Analysis Transcriptions were made of the spoken data using AS units, following Foster et al. (2000), while Praat analysis was used for making temporal measurements (de Jong & Wempe, 2007). In this way detailed measures of speed, hesitation and repair could be achieved. These cover aspects such as rate (e.g. words per minute), frequency (e.g. pauses per 100 words), time (e.g. filled pause time per 100 words), and duration (e.g. mean length of pause).

Two principal measures of speed were used – articulation rate (AR) and pruned speech rate (PSR). AR is a measure of pure speed of speech, without taking into account pausing or what is actually said. PSR, on the other hand, is the number of words spoken per minute, including pause time, and with repairs (or more properly, reparanda) removed. It therefore reflects speech rate while taking into account pauses and repair, and serves as the best global fluency measure.

In this study silent and filled pauses were analyzed according to their location, with pauses between clauses (external) being distinguished from those within clauses (internal). Internal pauses were then divided into three further locations: a) between constituents (constituent pauses); b) between words but within constituents (word pauses); and c) after words that are later repeated, reformulated or abandoned (repair pauses). External pauses are not considered here, as they are a necessary part of speech, marking the boundaries between clauses, and should not be considered as hesitations. They occur approximately as much in native-speaker as in learner speech (Tavakoli, 2011) and generally do not negatively affect perceived fluency (Boomer & Dittmann, 1962). It is also thought that external pauses are used to plan the next clause as a whole, so the nature of the next word is unimportant. In contrast, this study focuses on the words immediately after internal pauses, as these are more likely to reflect the reason for the hesitation.

Silent and filled pauses were counted separately, but later combined, as it was found that they play very similar roles, and the choice of one or the other simply reflects personal speaking style. Repetitions were also considered together with pauses, reflecting their use to “buy time” for the speaker, rather than to make corrections. Together, pauses and repetitions are here described as “hesitations”.

Lexical density Measurement of lexical density requires words to be separated into form or grammatical words, on the one hand, and content or lexical words on the other. The same distinction is used for the classification of pauses described below. O'Loughlin (1995, p. 228) describes in some detail how this can be done, and his framework was used for this study.

Lexical and form related hesitations Following Maclay and Osgood (1959) and Fathman (1980), the main types of

hesitation (silent pauses, filled pauses and repetitions) were classified according to whether they came before content (lexical) or form (grammar) words. This is based on the assumption that the word(s) after a hesitation are likely to be the cause of that hesitation. Thus a pause or repetition before a lexical word will indicate difficulty in accessing that word, while a pause before form words such as articles, auxiliary verbs or prepositions indicates uncertainty with grammar.

The following are examples of hesitations before content or lexical words (a noun, a verb, an adjective and a verb):

Page 6: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

178

|a-n-d they watched the beautiful ….. {uh} scenery on the mountain|

|so he wants ::to {hm} .....find something {um} ….. exciting and adventurous|

|so he had to …... take a plane there|

Form related hesitations: the examples below show, respectively, a repetition before an auxiliary verb; a pause before a preposition; and two pauses before articles:

|but however {they d-} they didn't|

|because they were too tired ….. on ….. the day the trip|

he ‘s in charge of ….. {uh} the human resource department|

[Note: ….. = silent pause, uh, hm = filled pause, words in {} = words that are “repaired” or repeated.]

In cases like the final two examples above, pauses before some grammatical words, such as articles, might equally indicate a search for lexis. On the other hand, pauses before lexis do in most cases appear to be clearly caused by lexical difficulties. It may therefore be that the figures for lexical and grammatical pauses underestimate the former and exaggerate the latter.

Lexical and grammatical range and accuracy Lexical range was assessed using D_Tools, which is a measure of the variety of lexis used. It is thus a form of type-token ratio, adjusted to make it work with different text lengths in a similar way to similar to “D” (Malvern & Richards, 1997). D-Tools is claimed to be particularly suitable for short texts of the kind used here (Meara, n.d.). Grammatical range was measured by the number of pruned words per clause (with repeated and corrected words removed), as longer clauses usually involve more complexity. An alternative measure, a subordination index where the number of clauses is divided by the number of speech units (AS units), was found neither to correlate with any other measure (whether linguistic or temporal) used in this study, nor to serve as an indicator of progress made over time.

Accuracy of lexis and grammar were measured by errors per 100 words. Lexical errors covered cases where word-choice, word formation, or collocation were clearly wrong. Grammatical errors included clear mistakes in word order, verb endings, tenses, articles, or where missing or extra words made clauses or phrases clearly incorrect. Total errors were divided by the complete number of words, and multiplied by 100.

Lexical and grammatical repairs As explained above, repetitions were included under the umbrella of hesitations. This left two categories of repair – reformulations and false starts. Reformulations are where a speaker self-corrects, changing either the lexis or grammar of the original words uttered. Where word order, morphology or function words are changed, they are considered grammatical reformulations. Where lexical words are changed, the repair is a lexical repair. An attempt was made to treat false starts in a similar way, according to whether the word likely to come after the interruption (and so the likely cause of the false start) was a grammatical or lexical word. However, the low number of false starts and the uncertainty as to what word would have come next (were it not for the interruption and new start) made this part of the analysis unsatisfactory. For this reason, results for false starts are not given below.

Page 7: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

How Grammar and Vocabulary Interact with Oral Fluency

179

RESULTS I will deal with the three research questions (RQ 1-3) separately.

RQ1: Do hesitations (silent pauses, filled pauses and repetitions) tend to come more often before lexical or grammatical words?

Table 1. Hesitations before lexical and grammatical words

Table 1 Grammatical pauses, repeats Lexical pauses,

repeats

Frequency Time Frequency Time

Total no. 990 536.77 1165 600.72

mean 29.12 15.79 34.26 17.67

% of pauses 45.94 47.19 54.06 52.81

St. dev. 18.8 10.83 17.58 11.06

Av length pause (ALP) 0.54 0.52

Table 1 shows the figures for hesitations (internal pauses and repetitions) before

lexical words and grammar words. As explained above, external (between-clause) pauses are excluded from this analysis. Taking all hesitations together, there is a slight majority of lexical, as opposed to grammatical, hesitations (54.1% by frequency, and 52.8% by time). The average length of each pause type is similar.

Table 2. Hesitations according to location before grammar and lexical words

Table 2 Constituent pauses Gram

Constituent pauses Lexis

Word pauses Gram

Word pauses Lexis

Repair pauses Gram

Repair pauses Lexis

Freq Time Freq Time Freq Time Freq Time Freq Time Freq Time

mean 11.03 5.39 6.15 3.06 2.26 1.03 7.76 3.6 9.88 6.29 11.85 6.72

% of pauses 64.21 63.78 35.79 36.22 22.58 22.17 77.42 77.83 45.47 48.35 54.53 51.65

St.dev 7.07 3.15 3.03 1.6 3.16 1.21 3.88 2.2 8.87 6.07 8.81 5.76

ALP 0.49 0.5 0.45 0.46 0.64 0.57

Table 2 shows figures for silent and filled pauses by location. Constituent pauses are

the greatest in number, and they tend to come before form words, as constituents often begin with determiners (in the case of noun phrases); prepositions (as with prepositional phrases but also many adverbials, e.g. on the other hand, at first); auxiliaries and modals (verb phrases); or linking adverbs. It is generally only in the case of single word constituents (e.g. verbs in the simple present/ past, adverbs of manner, or names) that constituent pauses come before lexical words. Not surprisingly, therefore, constituent pauses tend to be form-related (64.2% of occurrences) rather than lexical (35.6%). Word pauses, on the other hand, come mostly before lexical items (77.4%), while repair/ repetition pauses are more mixed, with slightly under half coming before form word, and slightly over half before lexical words.

Page 8: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

180

The figures in Tables 1 and 2 do not appear to provide support for either lexis or grammar being a greater influence on fluency, as the numbers of lexical and grammatical hesitations are similar. However, in interpreting these figures, it is important to bear in mind lexical density – the overall proportion of form and lexical words. The first five columns in Table 3 shows that just over a third of words in this sample are lexical, while almost two thirds are grammatical. If hesitations were equally likely to be caused by grammatical and lexical difficulties, we should expect most hesitations to come before grammar words, yet this is not the case. To clarify the situation, a calculation was made of the probability of there being a pause before each lexical and grammar word, by dividing the number of lexical /grammatical pauses by the number of lexical /grammatical words. These figures appear in the last two columns in Table 3. Column 6 shows the probability of a pause before each lexical word, and column 7 shows the same information for each grammatical word.

Table 3. Lexical density and hesitation probabilities per word

Total words

Lexical words

% Lexical words

Form words

% Form words

Prob Lex hesitation

Prob Gra hesitation

total 8680 2966 1168.45 5802 2274.01 1413.38 588.15

average 255.29 87.24 34.37 170.65 66.88 41.57% 17.3%

stdev 93.01 31.31 3.03 70.67 13.92 20.68 8.81

The results, once the higher proportion of grammar words is taken into account,

suggest a much stronger link between hesitations and lexical choices than with grammatical choices. It can be seen that the chance of a hesitation before a lexical word was 41.57%, while for grammar words it was a mere 17.3%. The probability of pauses before lexical words is thus two-and-a-half times greater than before grammatical words. With 34 narratives but only 17 different speakers at two different times, a non-parametric equivalent of the T-test, the Wilcoxon Signed Ranks Test, was used to test the significance of this difference. The result confirmed the difference between the pause probabilities of lexical and grammatical words as significant, with a test statistic (Z-value) of 5.0, p< 0.001.

RQ2: Are reformulations and false starts more often related to grammatical or to lexical problems?

There were slightly more repairs related to lexis than to grammar, but the differences were not great. For example, there were on average 1.09 reformulations of grammar per 100 words, and 0.90 reformulations of lexis per 100 words. Compared to Fathman's (1980) young learners, the participants in this study were clearly more interested in correcting their grammar.

However, it was also decided to compare correlations between repetitions or reformulations and the remaining fluency measures – i.e. speed and hesitation fluency measures. Because of the sample size, a non-parametric test – the Spearman Rank Order Correlation – was used. The results are shown in Table 4, where stronger correlations can be seen for lexical reformulations and repetitions than for grammatical ones. For example, for AR: lexical reformulation /100 words, rho=.270; for AR: lexical repetition /100 words, rho=.285. In contrast, the frequency of grammatical reformulations and repairs did not correlate significantly: AR: grammatical reformulations, rho=.110; AR: grammatical repetitions, rho=.061. Similarly, other fluency measures consistently correlate more strongly with lexical reformulations and repetitions than grammatical ones. Repetitions show the same story, except

Page 9: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

How Grammar and Vocabulary Interact with Oral Fluency

181

in the case of external repetition frequency, which as has already been explained is not considered a fluency marker.

Table 4. Repair fluency correlations – grammar and lexis

RQ3: Which correlate more strongly with fluency measures–grammatical range and accuracy or lexical range and accuracy? Correlations were calculated between a selection of fluency scores and the measures of grammatical and lexical range and accuracy. Again a non-parametric test was used because of the small data sample – the Spearman Rank Order correlation. As there is not space here to include all the correlations between fluency measures and those measuring linguistic range and accuracy, I will focus on the linguistic correlations of the most global indicator of fluency – pruned speech rate (PSR), which reflects speed, hesitation and repair fluency. The results are shown in Table 5. PSR correlates strongly with lexical range (rho=.571), lexical accuracy (rho=.514), and more weakly with grammatical accuracy (rho=.384), all of these correlations being significant (p=0.01). Correlations with grammatical complexity (pruned words per clause/unit) are, in contrast, very weak or non-existent.

Table 5. Correlations between overall fluency and measures of language range and accuracy

GRA error VOC error dLEX pruned words per clause

pruned words per unit

rho sig. rho sig. rho sig. rho sig. rho sig.

Pruned Speech Rate –.384** 0 –.571** 0 .514** 0 –.007 0.96 0.13 0.31

Page 10: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

182

DISCUSSION The results presented above can be summarized as follows:

a) Hesitations of all kinds (silent pauses, filled pauses and repetitions) are, on average, two-and-a-half times as likely to occur before lexical or content words as before grammar or function words.

b) Grammatical reformulations and repetitions are slightly more frequent than those affecting only lexis. However, lexical reformulation and repetition frequencies correlate moderately with fluency measures, while their grammatical equivalents do not.

c) Lexical range and accuracy correlate more strongly with fluency measures, as represented by the combined measure of pruned speech rate, than do the equivalent grammatical measures.

The first finding strongly suggests that hesitations are preponderantly caused by searches for lexis, rather than for grammatical processing. It is quite likely that even more hesitations are lexis related than shown here, as some hesitations before grammatical words (e.g. articles) may be used to search for lexis coming up immediately after these words. However, it could also be argued that some hesitations before lexical words could be grammar-related, as for example when the speaker hesitates before a verb while deciding which tense to use.

The second finding indicates that reformulations and repetitions of lexis, which indicate problems in accessing vocabulary, are a sign of less fluent speech, and that more fluent speakers rely less on such devices to keep going. Here the claim is not so much that lexical repairs and repetitions detract from fluency (although they may do so) as that they are signs of lexical difficulties, which are also symptomatic of disfluency.

The third finding implies that oral fluency depends more on lexical knowledge (or more precisely, on the ability to choose words correctly from a broad range of lexis) than it does on grammatical knowledge (or the ability to speak accurately, with complex clauses and units of speech). This is not to deny the importance of grammar, as grammatical accuracy is a central element in proficiency, and essential for efficient communication, but simply to suggest that grammatical knowledge does not contribute as much to oral fluency as lexical knowledge does.

This study therefore provides support, from three different sources, for the existence of a connection between lexical knowledge and fluency. Lexical range and accuracy seem to contribute to fluency much more than do the abilities to achieve grammatical complexity and accuracy. This overall finding is consistent with the idea that, following Levelt's (1989, 1999) blueprint of the L1 speaker, L2 grammatical formulation may also be unconscious and automatic just as they are in L1. That would account for why grammatical difficulties do not seem to slow speech down in the same way lexical choices do.

The findings are also compatible with the MOGUL model proposed by Michael Sharwood Smith and John Truscott (2014). MOGUL adopts a modular view of brain processing, which includes a language module that is highly specialized in carrying out morphosyntactic and phonological processing, but the workings of which are automatic and not open to conscious introspection. Lexical meanings, on the other hand, are conceptual, situated outside the language module, and therefore open to consciousness. This model would therefore predict that lexical knowledge is more relevant to fluency than grammatical ability. Speech production may well be halted or slowed down while lexis is searched for or accessed. On the other hand, the language module will simply use any available connections and representations to produce syntactic patterns (whether correct or erroneous), either based on L1 structures or the current state of the learner's interlanguage, without any need for delay or even for the speaker to be aware of what is happening.

This study does not, on the other hand, provide any support for Paradis's (2009) belief that

Page 11: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

How Grammar and Vocabulary Interact with Oral Fluency

183

L2 processing is based on declarative or explicit knowledge. A logical consequence of the Paradis view would be that grammatical knowledge should have a marked effect on fluency, and the results described above to not support that hypothesis.

With regard to pedagogical implications of these findings, it is always dangerous to draw conclusions from studies like this about how languages should be learned or taught. However, one clear implication of these findings would be the perhaps surprising one that building a good vocabulary, and practicing its use in communicative contexts, is a good way to promote oral fluency. Another, more controversial implication would be that, in view of the discussion in the previous paragraph, greater effort should be put into building implicit knowledge of grammar, rather than the mainly explicit knowledge generally taught in schools in Taiwan and elsewhere. This would probably not contribute greatly to fluency, but might lead to more accurate and sophisticated grammar in speech.

As already mentioned, there are methodological questions about the validity of taking the nature of the word after a hesitation as providing the key to understanding the reason for that hesitation. Other ways to identify the reasons for pauses, such as think aloud protocols, might be used in future studies to provide another kind of evidence on this question. The roles of lexis and grammar in language processing lie at the heart of understanding second language acquisition, and these questions are not going to be answered in the near future, and certainly not as a result of small-scale studies like this. The most that could be hoped for is that this study makes a contribution to that debate. REFERENCES Beattie, G. W., & Butterworth, B. L. (1979). Contextual probability and word frequency as

determinants of pauses and errors in spontaneous speech. Language and Speech, 22(3), 201–211. Retrieved from http://doi.org/10.1177/002383097902200301

Boomer, D. S., & Dittmann, A. T. (1962). Hesitation pauses and juncture pauses in speech. Language and Speech, 5(4), 215–220. Retrieved from http://doi.org/10.1177/002383096200500404

Bosker, H. R., Pinget, A. F., Quené, H., Sanders, T., & de Jong, N. H. (2013). What makes speech sound fluent? The contributions of pauses, speed and repairs. Language Testing, 30(2), 159–175. Retrieved from http://doi.org/10.1177/0265532212455394

De Jong, N. H., Groenhout, R., Schoonen, R., & Hulstijn, J. H. (2013). Second language fluency: Speaking style or proficiency? Correcting measures of second language fluency for first language behavior. Applied Psycholinguistics, 36(2), 1–21. Retrieved from http://doi.org/10.1017/S0142716413000210

De Jong, N. H., Steinel, M. P., Florijn, A. F., Schoonen, R., & Hulstijn, J. H. (2012). Facets of speaking proficiency. Studies in Second Language Acquisition, 34(01), 5–34. Retrieved from http://doi.org/10.1017/S0272263111000489

De Jong, N. H., Steinel, M. P., Florijn, A., Schoonen, R., & Hulstijn, J. H. (2013). Linguistic skills and speaking fluency in a second language. Applied Psycholinguistics, 34(5), 1–24. Retrieved from http://doi.org/10.1017/S0142716412000069

De Jong, N. H., & Wempe, T. (2007). Automatic measurement of speech rate in spoken Dutch. ACLC Working Papers, 2(2), 49–58.

Duez, D. (1985). Perception of silent pauses in continuous speech. Language and Speech, 28(4), 377–389. Retrieved from http://doi.org/10.1177/002383098502800403

Fathman, A. (1980). Repetition and correction as an indication of speech planning and execution processes among second language learners. In H. Dechert & M. Raupach (Eds.), Repetition and correction as an indication of speech planning and execution processes

Page 12: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Giles, Witton-Davies (朱杰佑)

184

among second language learners (pp. 77–85). Frankfurt: Peter Lang. Foster, P., & Tavakoli, P. (2009). Native speakers and task performance: Comparing effects on

complexity, fluency, and lexical diversity. Language Learning, 59(4), 866–896. Foster, P., Tonkyn, A., & Wigglesworth, G. (2000). Measuring spoken language: A unit for all

reasons. Applied Linguistics, 21(3), 354–375. Retrieved from http://doi.org/10.1093/applin/21.3.354

Goldman-Eisler, F. (1958). The predictability of words in context and the length of pauses in speech. Language and Speech, 1(3), 226–231. Retrieved from http://doi.org/10.1177/002383095800100308

Halliday, M. A. K. (1989). Spoken and written language. Oxford: Oxford University Press. Lennon, P. (1990). Investigating fluency in EFL: A quantitative approach. Language Learning,

40(3), 387–417. Retrieved from http://doi.org/10.1111/j.1467-1770.1990.tb00669.x Levelt, W. (1989). Speaking: from intention to articulation. Cambridge, MA: MIT Press. Levelt, W. (1999). Producing spoken language: a blueprint of the speaker. In C. Brown & P.

Hagoort (Eds.), The neurocognition of language (pp. 83–122). Oxford: Oxford University Press.

Maclay, H., & Osgood, C. (1959). Hesitation phenomena in spontaneous English speech. Word, 15, 19–44.

Malvern, D. D., & Richards, B. (1997). A new measure of lexical diversity. In A. Ryan & A. Wray (Eds.), Evolving models of language. Clevedon, England: Multilingual Matters.

Meara, P. (n.d.). Lognostics~tools. Retrieved November 4, 2013, from http://www.lognostics.co.uk/tools/index.htm

Mehnert, U. (1998). The effects of different lengths of time for planning on second language performance. Studies in Second Language Acquisition, 20(1), 83–108.

O’Loughlin, K. (1995). Lexical density in candidate output on direct and semi-direct versions of an oral proficiency test. Language Testing, 12(2), 217–237. Retrieved from http://doi.org/10.1177/026553229501200205

Paradis, M. (2009). Declarative and procedural determinants of second languages. Amsterdam; Phildadelphia: John Benjamins.

Segalowitz, N. (2010). Cognitive bases of second language fluency. London: Routledge. Sharwood Smith, M. (2014). The multilingual mind: A modular processing perspective. New

York: Cambridge University Press. Skehan, P. (2003). Task-based instruction. Language Teaching, 36(01), 1–14. Retrieved from

http://doi.org/10.1017/S026144480200188X avakoli, P. (2011). Pausing patterns: differences between L2 learners and native speakers. ELT

Journal, 65(1), 71–79. Retrieved from http://doi.org/10.1093/elt/ccq020 Tavakoli, P., & Skehan, P. (2005). Strategic planning, task structure, and performance testing.

In R. Ellis (Ed.), Planning and task performance in a second language (pp. 239–273). B.V.: John Benjamins.

Ure, J. (1971). Lexical density and register differentiation. In G. E. Perren & J. L. M. Trim (Eds.), Applications of linguistics: Selected papers of the second International Congress of Applied Linguistics, Cambridge 1969 (pp. 443–452). Cambridge: Cambridge University Press.

Witton-Davies, G. (2014). The study of fluency and its development in monologue and dialogue (Unpublished PhD thesis). Lancaster University, Lancaster, UK.

Page 13: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

FROM THE EDITOR We are delighted to present this volume of Selected Papers from the

Twenty-fourth International Symposium on English Teaching in CD-ROM version to our members and colleagues. This year’s conference theme is Teaching and Learning English for a Multilingual World. Featured topic areas include the global and local issues in language teaching and learning, literature, technology, and research.

We received 130 local and 18 international paper proposals, 10 workshops, 2 panels, and 1 poster this year. Among 25 full papers submitted to us, we chose 18 of them after the recommendation from the reviewers. The blind review committee members are scholars from various areas of the TESOL profession.

The topics of the papers received cover a wide spectrum of issues such as basic language skills, pedagogy, global education, World Englishes, storytelling, grammar, and others.

Also included in this volume are 8 papers contributed by our invited speakers. These papers address issues in pedagogy, professional development, assessment, writing, and children literature.

Due to time constraint, we may not be able to revise or edit all the papers. Mistakes are therefore inevitable. Although we have endeavored our utmost to keep each writer’s style, we did make some changes and adjust headings, sub-headings, and tables to conform to the APA citation style.

It is hoped that this volume will not only provide insights for graduate students, teaching professionals, and researchers, but also serve as valuable resources for them.

The production of this volume of selected papers would not have been possible without the assistance from Wei-teh Lee, Su-yi Hwang, and Avis Teng. We also would like to take this opportunity to express our deepest gratitude to the major sponsors for their generous support in the production of this volume: the Ministry of Science and Technology, the Ministry of Education, Taipei City Government, the Bureau of Education, Taipei City, Takming University of Science and Technology, and Minghsin University of Science and Technology. We would also like to thank the Crane Publishing Company for its continuous support to the publication of the Selected Papers and the Language Training and Testing Center for their sponsorship of the best paper awards. Last but not least, special thanks should go to Professors Wei-yang Dai, Johanna Katchen, Yu-ching Chan, Yiu-nam Leung, Wen-li Tsou, Chung-shun Hsia, Tzu-chia Chao, Sy-ying Lee, Yu-li Chen, Tai-yuan Tseng, John Truscott, and Jessica Wu (LTTC) for their meticulous efforts spent in reviewing the papers submitted for inclusion in this volume. Yiu-nam Leung (梁耀南), President/Executive Editor October 25, 2015

Page 14: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

TABLE OF CONTENTS INVITED PAPERS

Chan, Leo Tak-hung (陳德鴻), Post-Communicative Pedagogies: Revisiting the Translation Method of Teaching English in East Asia · · · · · · · · · · · · · · · · · · · · · · ·

Christison, Mary Ann, Engaging Students in Learning in a Multicultural World · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Goh, Christine C. M., English Language Teachers’ Professional Learning: Impacts, Challenges and a Multi-Themed Approach · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Green, Anthony, What Do Teachers Need to Know About Language Assessment · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Kubota, Ryuko, Questioning Language Myths in English Language Teaching: Toward Border-Crossing Communication · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Lado, Ana, Applying Tellability to Teaching English Learners with Picture Books, Radio Scripts, and Technology · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Liu, Dilin (劉迪麟), Contrastive Analysis-Revisited: The case of Chinese and English and its Implications for English Learning/Teaching · · · · · · · · · · · · · · · · · · · ·

Matsuda, Paul Kei, Let’s Face It: Language Issues and the Writing Program Administrator · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

1-15

16-22

23-35

36-43

44-57

58-66

67-81

82-97

WINNERS OF BEST PAPER AWARD

Lan, Shu-wen (藍淑雯), Intertextual Analysis of EFL College Classroom Discourse Using Authentic Materials · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Liao, Yen-fen (廖彥棻), A Closer Look at Lexico-grammatical Knowledge as a Predictor of L2 Listening Ability · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

98-110

111-124

PAPERS

Asakawa, Kazuya, Holistic and Global Education in English Teaching · · · · · · · · · · · ·

Balaz, Allen, Falling on Bard Times: Shakespeare in a World of Englishes · · · · · ·

Chang, Ching-shyang (錢清香), The Effect of Doing Extended Listening-focused Activity on Developing L2 Listening Fluency through Extensive Reading While Listening to Audio Graded Readers · · · · · · · · · · · · · · · · · ·

125-129

130-141

142-154

Page 15: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency

Chen,Ya-chen (陳雅貞) & Teng, Huei-chun (鄧慧君), Effects of Summary Writing Instruction on EFL Reading and Writing · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Chu, Po-ying (朱珀瑩), What Puzzles Teachers? · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Giles, Witton-Davies (朱杰佑), How Grammar and Vocabulary Interact with

Oral Fluency · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Gong, Byron (龔 營), Taiwan Needs a Unified Scale for College English

Language Proficiency · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Huang, Da-han (黃達翰) & Chao, Yu-gi (趙玉芝), Content Analysis Study of

English Textbooks in Taiwanese General and Bilingual Senior High Schools · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Huang, Tzyh-lai (黃自來), Understanding English Grammar and Its Place in

TEFL ·· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Lin, Yu-yuan (林毓淵) & Chuang, Li-ling (莊琍玲), Pacing Storytelling for the

Differentiated Classroom ·· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Liou, Hsien-chin (劉顯親 ) & Lin, Kuan-hui (林冠慧 ), Citations and

Knowledge Claims in MA-TESOL Students’ Theses · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Shiue, Connie (薛紹楣), Bragg, Nigel & Roebl K. M., A Meta-Analysis Study

of the Development of Reading Comprehension Skills in EFL University Level Students · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Ueda, Mami (植田麻実), Abe, Emika (阿部恵美佳), Sugino, Toshiko

(杉野俊子) & Shimizu, Sunao (清水順), What it Means to be Members of a Global Society · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Wang, Hui-chih, Shih, Harry & Dai, Jei-ru, A Study on Extensive Reading and

Student Engagement in College English Class · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Wang, Wei-ti (王維蒂) & Lo, Gloria Yi-hsuan (駱藝瑄), The Effects of Audio

Story Books on Teenage EFL Learners · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · White, G. Benjamin (白班哲明), Taiwanese Students’ Transition Markers

Use with Comparisons to SSCI Articles · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

155-165 166-172

173-184 185-197

198-207

208-224

225-239

240-250

251-259

260-267

268-281

282-294

295-308

Page 16: How Grammar and Vocabulary Interact with Oral … Papers from the Twenty-fourth International Symposium on English Teaching 173 How Grammar and Vocabulary Interact with Oral Fluency