NEGOTIATION, NOTICING, AND THE ROLE OF SELECTIVE CROSSLINGUAL STRATEGIES IN FOREIGN LANGUAGE CLASSROOMS MA in TEFL thesis, William R. Pellowe, 1998
PREVIOUS	TITLE PAGE TABLE OF CONTENTS REFERENCES	NEXT
http://www2.gol.com/users/billp/thesis/index.html et seq billp@gol.com

In chapter 3, we saw that restricting L1 in certain types of tasks may provide learners with the confidence to employ L2 to a greater extent in other types of classroom discourse. One is prone to assume that students would be less inclined to use L1 with non-native L1 teachers; however, if students know that recourse to L1 is available, even "sanctioned", will they opt out of using L2? In this chapter, I examine written test texts produced during an extensive reading program; on these tests, learners were allowed to use written Japanese freely. In this chapter, the amounts and functional distribution of L1 use, as well as the proportion of L1 to L2, are described, and possible motivations for language choice are discussed.

4.1 The extensive reading program tests

Given the benefits of extensive reading (for overviews, see Kim & Krashen, 1997; Nation, 1997; Bamford & Day, 1997), even within the constraints of graded, simplified texts (cf. Hill, 1997), I implemented an extensive reading program. The students (n=11) were required to read four self-selected titles from our graded reader library during the second half of the academic year (September 1997 - January 1998). The graded readers are basic level, with headword counts of 400 to 700 words. One half of each Friday lesson was devoted to the program. On these days, students who had finished books handed them in, and were given tests to complete, while others used this time to read. Most reading, though, was to be done outside of class; furthermore, students could take tests during their free time whenever they so chose. Students who read more than the required 4 would receive extra points on their final term grades, while those who read fewer would have points deducted.

4.1.1

The single, simple object of the tests was to determine whether or not the student had read the book. The test consisted of a photocopy of two consecutive pages from a random section of the book, and a response paper with two questions:

1. Tell me something about what happened before this scene.
2. Tell me something about what happened after this scene.
   
   

The written instructions clearly stated that students were allowed to write in Japanese if they so chose, but that English was encouraged: "Please try to write in English. If you have to write in Japanese, that's OK!" This decision to allow Japanese was motivated by research from immediate recall tasks (both reading and listening) indicating that allowing L1 more accurately taps true comprehension, as responses are not limited by L2 productive ability (Schmit-Rinehart, 1994), while disallowing L1 in student responses can have the adverse effect of "trivializing" the students' commentary (Modica, 1994, p. 303). There was no specification as to how long the answers had to be; in fact, students were encouraged to be brief.

4.1.2 How far did the teacher influence language choice?

In introducing the program and the test to the students, I told them (lightheartedly) that their written Japanese would be good for my language study, but that I would seek help (from another teacher) if needed. I pointed out that their written English would not only be easier on me, but would benefit their own writing practice while highlighting for me some areas of English which might be usefully targeted in subsequent lessons. In short, the encouragement to use English was given as an appeal (to make things easier for me) and as a pedagogic opportunity. Dictionary use was allowed (yet some did not always remember to bring them). The texts created in response to these test questions are analyzed below.

4.2 Results

The 11 students took a total of 48 tests over the five-month period. Given that they had the freedom to opt to use Japanese, it is striking that the majority of the texts (62.5%) were composed entirely in English (Table 1). Four students did not use Japanese at all, and one student used Japanese exclusively. (However, S5, with 1 text in L2, revealed

Table 1: Language of Test Texts Per Student
STUDENT*	LANGUAGE OF TEST TEXTS				TOTAL
	English	Mixed		Japanese
		(English with Japanese insertions)	(Japanese with English insertions)
S1	2	2	0	0	4
S2	6	0	0	0	6
S3	5	0	0	0	5
S4	2	2	0	1	5
S5	1	0	0	4	5
S6	5	0	0	0	5
S7	1	3	0	0	4
S8	3	0	0	0	3
S9	4	1	0	0	5
S10	1	0	1	0	2
S11	0	0	0	4	4
TOTAL	30	8	1	9	48
PERCENT	62.5%	16.7%	2%	18.7%	99.9%
* Students are numbered by rank on oral proficiency tests. [top]

that she was initially unaware that L1 was allowed; subsequent texts were exclusively L1.) Of the 9 texts exhibiting language switching, one is classified as Japanese with English insertions as 13 of its 16 sentences were in Japanese. The remaining 8 contained 42 instances of Japanese use. One learner, though, used the same L1 word 6 times in her text; counted as one usage, the total is reduced to 37 (Table 2).

Table 2: Japanese Usages In Mixed Texts (English with Japanese insertions)
STUDENT	NUMBER OF MIXED TEXTS	NUMBER OF USAGES OF JAPANESE
S1	2	5
S4	2	5
S7	3	26
S9	1	1
TOTAL	8	37

4.2.1

Simply counting the number of occurrences of Japanese is meaningful only within the context of the amount of English used. The ratio of English words to Japanese words [3] (Table 3) reveals that the quantity of Japanese insertions varied widely, from one single word for S9, up to over 16% of one text for S7.

Table 3:Distribution of English Words vs. Japanese Words
STUDENT (ONE TEXT PER ROW)	NUMBER OF ENGLISH WORDS	NUMBER OF JAPANESE WORDS	TOTAL WORDS	PERCENT JAPANESE WORDS
S1	114	8	122	6.56
S1	52	1	53	1.89
S4	83	4	87	4.60
S4	90	5	95	5.26
S7	62	8	70	11.43
S7	130	25	155	16.13
S7	143	15	158	9.49
S9	159	1	160	0.63
TOTAL	674	66	740	8.91
MEAN	104.13	8.38	112.50	7.00

4.3 Functions of Switches

The functions of L1 varied considerably. Switches fell into three main categories: fill ins, clarifying translations of English, and single Japanese particles. Fill-ins and clarifying translations each contained instances of single words, clauses or full sentences (Table 4), discussed below.

4.3.1 Clarifying Translations

Clarifying translations serve as a side-note explanation of English text, clarifying the intended meaning. In Example 1, the learner marks this relationship with an arrow drawn from the Japanese word to its English equivalent.

The student is unsure that "Black people" is the right term. The spoken counterpart of clarifying translations would perhaps be L2 marked both intonationally and with gestures (rising intonation, direct eye contact, perhaps raised eyebrows) as a clarifying

Table 4: Functional Distribution of Japanese Code Switches by Type
	SINGLE WORDS	PHRASES, CLAUSES	FULL SENTENCES	TOTALS
Clarifying Translations	3	1	2	6
Fill-Ins	13	11	5	29
Japanese Particles	3	not applicable	not applicable	3
TOTALS	19	12	7	38

check to solicit confirmation of the accuracy of the term. Given that the written mode does not allow for this kind of interactive negotiation or solicitation of feedback, the clarifying check anticipates a possible error while serving to eliminate ambiguity resulting from inaccuracies. Thus, clarifying translations mark the L2 as tentatively offered. The five other instances of clarifying translations were paired with single words (2), a phrase, and whole sentences (2). In each case, the clarifying translation was set apart from the context by the inclusion of arrows, as in example 1, or by enclosure between parentheses.

4.3.2 Fill-ins

Clarifying translations are different from the more common fill-ins (Examples 2 through 8, below) in that clarifying translations are additional L1 items supplementing what is already on the page in L2, whereas fill-ins are instances of L1 used in place of L2 to bridge a (usually lexical) gap. Fill-ins accounted for 29 of the 38 instances of Japanese text insertions. The most common use of fill-ins were single words (13). A further 11 instances were phrases and clauses, and 5 were full sentences.

4.3.2.1 Fill-ins: Single Words: nouns

Examples 2 is from a test on a 600-headword version of The Wonderful Wizard of Oz (Baum, 1981). The book did not contain the word "tornado", but described the tornado as "a big wind". However, the student tries to explain what happened in terms of an analogy to a typhoon:

Whether or not the student knew from other sources that the storm was a tornado is uncertain, as the Oz story is not as well known in Japan. This circumlocution as a communication strategy is rendered less effective, though, when she cannot come up with the word "typhoon", and thus opts for a single-word fill-in. "Typhoon" is not an English loanword; it was borrowed into English from Japanese, where the pronunciation is similar to "tie who", not as similar to the English pronunciation as it appears to be in the romanization of the word.

Note that taifu is preceded by an indefinite article, which is interesting because in contrast to English, Japanese has no articles. This indicates that the Japanese nouns in the data were treated as English items. Although taifu is somewhat close to its English equivalent, this similarity cannot be cited as a probable reason for its "English lexicalization", as other Japanese nouns sharing no similarities with their English equivalents were likewise modified:

And later:

The correct use of the definite article in Example 3, and the possessive "s" in Example 4 show typical ways in which the learners localized their Japanese insertions to conform as closely as possible with English conventions.

4.3.2.2 Fill-ins: Single Words: verbs

The three verbs in the data showed less English lexicalization than was evident with nouns. While verbs occurred in SVO patterns, rather than Japanese (S)OV patterns, they were not conjugated with English endings. Two of the verbs were given Japanese grammatical modifications for tense, but one was given in simple dictionary form:

4.3.2.3 Fill-ins: Phrases and Clauses

The second most common type of fill-in was for phrases and clauses. As with single word insertions, the majority of these were inserted to retain English word order. In some cases, it was difficult to classify phrases across the language gap, as occasionally a "phrase" may be considered a single word in Japanese, yet the equivalent in English would be a strong collocation or a fixed phrase:

However, most cases were clearer:

4.3.2.4 Fill-ins: Full Sentences

A further five of the fill-ins were full sentences. Their function in the text was to move the text forward from one L2-based situation to another. (As these are meaningless outside of their full contexts, no examples are provided, but see Endnote 3.)

4.3.3 Japanese Grammatical Particles

The third type of Japanese insertion involved putting single grammatical particles into English sentences. In Example 8, the topic marker "wa" is an example of a Japanese particle insertion:

These grammatical particle insertions were not common; one student used this three times, yet never without additional Japanese elsewhere in the sentence.

4.4 Discussion

What factors prompted or influenced these learners in their language encoding choices? When I tried asking one student directly why she used Japanese, her reply was "Sorry". Even after assuring her that my question was not censorious, her reply was apologetic, with an explanation that she did not know the English words. This experience dissuaded me from asking others directly. While it would be convenient to draw generalizable conclusions based on connections between the students' language choices and some other contributing factors, I cannot advance any generalizable, all encompassing statements other than that the choice was an individual one, yet within subsets of students, some weak patterns emerged. Possible contributing factors to these individual decisions are considered and discussed in the points below.

• Timing of tests

While some individuals used less English in earlier tests than in later ones, it is not true that all students progressively decreased the amount of Japanese they used; in short, there is no significant correlation between when the tests were taken and the amount, if any, of the Japanese within the texts.

• Proficiency

Comparing proficiency data from two sources (the TOEIC standardized test and oral interview scores with an English NS unknown to the students) with the language choices on each test, it is apparent that proficiency does not correlate for the high or middle sections of the class. S1, who ranks highest on both proficiency measures, used Japanese, yet there were many who chose not to use Japanese at all. While it is true that one of the least proficient students opted for Japanese on every test, a student in the middle ranking of the class (S5) used Japanese on every test save one, and that one occasion was by mistake (as noted in 4.2.0).

• Time Available

One contributing factor may be time; when S4 chose not to use Japanese, the length of her texts were much greater (word counts of 245 and 138) than when she used a mixture (87 and 95) or used only Japanese. S7 and S1's English texts were similar in length to their mixed ones. Perhaps given the constraints of time available, these students chose the most economical use of time and effort. Obviously, though, the most economical use of time would have been to compose entirely in Japanese. Their use of English rather than Japanese can only be seen in terms of wanting to make the most of the learning experience (cf. the next point).

• English Use as a Pedagogic Opportunity

Students made no concessions to the fact that I was a non-native Japanese speaker when they wrote in Japanese. Concessions to non-native speakers of Japanese include using romanized Japanese, avoiding all but the most basic kanji characters, writing unnaturally full sentences, and using excessive "furigana" (phonetic superscripts for kanji characters). Instead, the Japanese texts were written as if to a Japanese person. This gains in significance when we recall that the initial encouragement to use English (4.1.2) was given as both an appeal to make things easier on me and as a pedagogic opportunity for them. As the former apparently held little to no sway for these students, the later must be considered a significant factor in their use of English.

4.4.1

And what of the corollary to the two preceding points? Do texts written entirely in Japanese therefore represent a lack of desire to take advantage of pedagogical opportunities? Sadly, the answer seems to be "yes". The two students who produced the largest number of Japanese texts (S11, 4 out of 4; S5, 4 out of 5, with 1 L2 text done unaware that L1 was allowed, see 4.2.0) commonly engaged in behavior incongruent with the assumed classroom goals of learning English, seldom taking constructive advantage of pedagogical opportunities. They were often off-task, commonly holding non-pedagogically-motivated L1 conversations at loud volume as they were not in adjacent rows. Despite repeated warnings and complaints from other students, it took a general class meeting with their homeroom teacher to impress on them the deviance of their behavior, and the level at which the others in the class were being disturbed. What we can generalize from their tests is that students who opt for the bare minimum or opt out of learning opportunities were the only ones who consistently opted for only Japanese on the tests.

4.5 Conclusion

We saw that when given the choice to use as much Japanese as they want to on written tests, most of these learners used Japanese minimally, if they used it at all. When Japanese is used within English texts, it serves two main functions; one is to move the discourse forward, and the other is to preempt possible misunderstandings or mistakes by providing clarifying translations. In both functions, it serves to bridge a deficiency, whether real or imagined, within the overarching constraints of time and linguistic resources available. This limited L1 use, which helps move the discourse forward, thereby provides opportunities for further output.

4.5.1

A few students relied on a great deal of Japanese to produce their texts. Based on their texts, it was clear that these students had read and understood the books, thus meeting the goals of the program, a fact which some of them may not have been able to demonstrate through an English-only test.

4.5.2

The instances of L1 use within English texts provide the teacher with useful information regarding students' errors. It is axiomatic that learners produce errors. These errors are not always evidence of a deficient interlanguage system; some errors may instead be temporary slips or mistakes (cf. Sharwood Smith, 1994, p. 19; Corder, 1967/1975, p. 25). Errors are only apparent when the learner attempts to use a form, and are not apparent if the learner elects to avoid using problematic forms. When effective circumlocution is not available, recourse to L1 removes the necessity of using (other) avoidance strategies. The existence of L1 provides the teacher with an indication of a probable gap in the student's knowledge. It may be that the learner is unaware of the range a familiar lexical or structural item covers, or it may be that the L1 use indicates a larger gap. Thus, allowing limited L1 use within L2 texts in this type of written task highlights gaps in the students' L2 knowledge more effectively than if L1 were proscribed.

4.5.3

Last, the medium of written communication differs from oral communication in that it allows students (operating within the same constraints of L2 accessibility) to think first in L1, and subsequently compose a translation. This process is not open to inspection in the data here, as absence of clarifying translation provides no indications on the presence or absence of this L1 to L2 process. However, the few instances of clarifying translation are significant from an interlanguage point of view in that they highlight uncertainty. From a pedagogical point of view, these clarifying translations provide the instructor with sufficient data to interpret L2 which may have otherwise been ambiguous.

4.5.4

The initial concern stated in 4.0 was that students would use too much L1 if its use were sanctioned. Within the context of written texts, this concern is unfounded for the majority of the students. However, in order to ensure the benefits of limited L1 recourse, it would be desirable to discourage L1-only texts by reframing the allowance of L1 within the positive strategies described above. In short, learners who initially produce L1-only texts should be made aware that while L1 can usefully clarify L2 and make up for L2 gaps, an exclusive use of L1 is neither expected nor ultimately beneficial.