The effects of two types of enhanced input on intake and the acquisition of implicit and explicit knowledge
as published here:
Reinders, H. & Ellis, Rod 2009 ‘The Effects of Two Types of Positive Enhanced Input on Intake and L2 acquisition’. In: Ellis, R., Loewen, S., Erlam, R., Philp, J., Elder, C., Reinders, H. Implicit and Explicit Knowledge in a Second Language. Clevedon: Multilingual Matters.
The tables are not showing proper formatting but email me and I’ll send them to you.
HAYO REINDERS and ROD ELLIS
The importance of ample input for second language acquisition is uncontroversial. At the same time, evidence exists (for example from studies in immersion settings) to show that even with massive exposure certain aspects of the language develop slowly or not at all (Swain, 1988). This appears to apply especially to formal features that are semantically redundant and/or that are difficult to notice. The study of the incidental acquisition of 3rd person –s reported in the preceding chapter of this book provided clear evidence of this, as the learners failed to improve their accuracy of this feature despite intensive exposure to it. It appears that such aspects require some form of instructional intervention, although it remains unclear what type of intervention is most effective. One instructional possibility is ‘input enhancement’.
The term input enhancement was used by Sharwood-Smith (1991, 1993) to refer to attempts to direct the learner’s attention to a specific linguistic form in the input. Sharwood-Smith argued that this term is to be preferred to the earlier term he used to refer to same idea (‘consciousness-raising’) because it makes no assumption as to whether the input alters the learner’s mental state. ‘Input enhancement implies only that we can manipulate aspects of the input but makes no further assumptions about the consequences of that input for the learner’ (1993, p. 176). Sharwood-Smith includes a number of techniques under the umbrella term of ‘input enhancement’ and makes a distinction between positive and negative input enhancement. The former refers to the manipulation of the input learners are exposed to. The latter refers to input that is enhanced by means of explicit instruction and/or corrective feedback. In this article we are concerned only with positive input enhancement.
This chapter reports a study that investigated the effect of two different types of input enhancement (input enrichment and input enrichment + noticing instruction) on both the intake and acquisition of a difficult grammatical structure (negative adverbs). As in the previous studies in this part of the book, the effect of the instruction will be measured in terms of both implicit and explicit L2 knowledge. First, the key constructs that inform the study will be defined. Then a number of studies that have examined the types of input enhancement we are interested in will be examined.
Definition of the key constructs
The specific types of input enhancement we are interested in are (1) ‘enriched input’ (i.e. input that has been seeded with the target structure so that learners are exposed to a high frequency over a period of time) and (2) enriched input combined with an explicit instruction to the learners to pay attention to the target structure – i.e. ‘noticing instruction’. Both constitute focus on form techniques, as this construct was defined by Doughty and Williams (1998a). That is, focus-on-form instruction is an attempt to focus learners’ attention on form in the context of an activity where their primary attention is on meaning. The particular feature of focus on form instruction that the two types of input enhancement address is what Doughty and Williams (1998b) refer to as ‘learner attention’, which they differentiate in terms of whether the technique involves ‘attracted’ attention or ‘directed’ attention. Enriched input, we would argue, constitutes an example of attracted attention, as the high density of sentences containing the target structure is predicted to cause the learners to notice it. In contrast, enriched input combined with an explicit instruction to pay attention to the target structure constitutes ‘directed attention’. Both types can be considered examples of unobtrusive focus-on-form in Doughty and Williams’ taxonomy. In this respect, they contrast with obtrusive techniques such as input-processing instruction (VanPatten, 1996) and consciousness-raising tasks (Fotos & R. Ellis, 1991).
Enriched input is input that has been manipulated in some way. There are various ways of doing this – glossing, bolding, underlining, or increasing the frequency of the target feature (sometimes referred to as ‘input flooding’). Studies that have investigated enriched input draw on Schmidt’s Noticing Hypothesis (1990, 1994), which states that, in order for learners to acquire from input, they must first pay conscious attention to exemplars of particular forms. By artificially increasing the saliency of the target structure, it is thought that learners will notice and thus acquire the structure more easily. The technique used in the study reported in this chapter was that of artificially increasing the frequency of the target structure. The reason for choosing this type of enriched input was because it is one of the least obtrusive of the available options. Whereas typographical enhancement and glossing direct participants’ attention to the target structure, increasing the frequency of the target feature simply makes it more likely that the learners will notice it.
Enriched input in the context of a meaning-focused activity caters to incidental learning. This is defined operationally by Hulstijn (2003) as the learning that results when learners are provided with L2 input without telling them that they will be tested afterwards. One way in which this can be achieved is by engaging learners in a communicative activity where their attention is focused on extracting meaning from input and then testing whether they have learned a specific linguistic feature in the input. As Hulstijn points out (and as noted in the preceding chapter) most of the studies of incidental acquisition have examined vocabulary and there are very few studies that have investigated grammar learning. In the study reported in this chapter participants in the enriched input condition were asked to complete meaning-focused tasks but were not told they would be tested on the target structure (or tested at all). This condition, then, involved incidental learning. It should be noted, however, that this condition for incidental learning differs from that in the preceding chapter in that here no attempt was made to distract the learners’ attention by focusing on a different grammatical structure. Learners were free to attend to the target structure (negative adverbials) as they processed the input for meaning
Noticing refers to the cognitive activity that learners engage in when they consciously attend to some linguistic feature in the input. Once learners have noticed a feature they are able to rehearse it in short-term memory and thus increase the likelihood of acquiring it (i.e. integrating it into their interlanguage). Input can be enhanced by means of an instruction to the learners to pay attention to a specific feature. The instruction might simply ask the learners to look out for exemplars of the target feature or it might ask them to try to work out the rule to explain how the target feature works. Both types of instruction are likely to encourage intentional learning but the former probably less than the latter. In a context where the learners’ attention is primarily focused on the meaning of the input and where they are simply asked to look out for the target structure and are not forewarned they will be tested on the structure it is less clear that they will engage in intentional learning. It is for this reason that we labeled this condition in our study the Noticing Condition rather than the Intentional Learning Condition.
The study also draws on two other constructs; intake and acquisition. As McLaughlin (1987) pointed out, the term intake ‘has taken on a number of different meanings, and it is not always clear what a particular investigator means in using it’ (p. 13). Some theorists view intake as an initial stage of learning, intermediate between input and acquisition. Gass (1997), for example, distinguishes a number of stages starting from raw input. Several factors (including time pressure, frequency, affect, salience, associations and prior knowledge) influence whether input gets noticed, or apperceived. Apperception is conceptualized as a priming device that prepares the learner for the possibility of subsequent analysis and intake, which Gass defines as the ‘process of assimilating linguistic material’ (p. 5). Intake can thus be conceptualized as apperceived input that has been further processed. Other theorists, however, use the term to refer to the entire process of acquisition. Chaudron (1985), for example, defines it as ‘the mediating process between the target language available to learners as input and the learners’ internalized set of L2 rules and strategies for second language development’ (p. 1). Kumaravadivelu (1994) likewise defines intake as a complex process starting with detection and ending with acquisition. It is difficult to see how intake can be distinguished from learning in such definitions. In this study we adopt Gass’ position and seek to distinguish intake from acquisition. We define intake as a subset of the detected input (comprehended or not) that is held in short-term memory and from which connections with long-term memory may be created or strengthened.
Not surprisingly given the differences in the definition of intake, a range of operationalizations of this construct exist. Rosa & O’Neill (1999) recommend using performance measures such as recall protocols, cloze tests, grammaticality judgements, and rule formation, all to be administered soon after the treatment or exposure to the target input. Leow (1993, 1995) also used multiple-choice recognition tasks and gave participants very limited time to complete their tasks, which were administered immediately after exposure. Shook (1994) made use of both production tests (cloze test, sentence completion) and a recognition test (multiple-choice sentence completion) all of which were administered immediately following the exposure. Shook claims that ‘it is most improbable that the data collection procedures used could reflect anything except the immediacy of Process I [the input-to-intake stage], and thus this study does not reflect any acquisition of the grammatical input’ (p. 85). What is common to all these methods is the attempt to probe what is held beyond short-term memory and to avoid measuring existing knowledge. The key lies in assessing what learners have noticed immediately after (but not during) exposure to input. In this study we used a production measure; we took correct use of the target structures in written output produced shortly after exposure to the enhanced input as evidence of intake.
The final construct we will consider is acquisition. As in the rest of this book, two types of knowledge are distinguished – implicit knowledge and explicit knowledge. These two types of knowledge were defined in Chapter 1 of the book. The acquisition of these two types of knowledge can be measured using grammaticality judgment tests. As shown in Chapters 2 and 4, where it was shown that a grammaticality judgement test with limited response times predispose learners to draw more on implicit knowledge while a test with unlimited response times can allow learners to access more explicit knowledge, especially in the case of the ungrammatical sentences in the test. This is the approach to measuring acquisition that will be followed in this study.
Previous studies of enhanced input
The following review will only consider studies where the target feature was grammatical. The review will consider studies where learners were simply exposed to enriched input and studies where learners’ attention was directed towards the target structure (i.e. an attempt was made to induce noticing of the target structure)..
A key question regarding the efficacy of enriched input is whether learners actually notice the target structure. This was investigated in a study by Jourdenais, Ota, Stauffer, Boyson and Doughty (1995). They found that English speaking learners of L2 Spanish were more likely to make explicit reference to preterit and imperfect verb forms when thinking aloud during a narrative writing task if they had previously read texts where the forms were graphologically highlighted. They also found that the learners exposed to the enhanced text were more likely to use past tense forms than the learners who read the non-enhanced text even though both texts had been enriched. It should be noted, however, that in this study the target structure was highlighted. In this respect the enriched input of Jourdenais et al’s study differed from the enriched input of the study reported in this article.
A number of studies have investigated whether enriched input results in acquisition. Trahey and White (1993) examined whether an ‘input flood’ (viewed as ‘positive input’) was sufficient to enable francophone learners of L2 English to learn that English permits adverb placement between the subject and the verb (French does not) but does not permit placement between the verb and object (French does). Exposure occurred 1 hour a day for 10 days. The target structure was not highlighted in any way. The learners succeeded in learning the SAV position but failed to ‘unlearn’ the ungrammatical SVAO position. In a follow-up test administered one year after the treatment, however, Trahey (1996) found that the beneficial effects of the input flood on the acquisition of SAV had disappeared.
J. White (1998) compared the effects of three types of enriched input; (1) typographically enhanced input flood plus extensive listening and reading, (2) typographically enhanced input by itself, and (3) typically unenhanced input flood. This study found that the three types of enriched input worked equally effectively in assisting Francophone learners to acquire the possessive pronouns his and her, leading White to conclude that the target structure was equally salient in all three.
While several studies have investigated the effects of enriched input, very few have investigated the effect of enhanced input involving noticing instructions. Leeman, Artergoitia, Fridman and Doughty (1995) examined the effects of input enhancement on the acquisition of preterit and imperfect Spanish verbs forms that were highlighted in written input. The learners were told to pay special attention to how temporal relations were expressed in Spanish and received corrective feedback from the teacher. Posttests showed that the learners outperformed a comparison group that did not receive the enhanced input. However, because they received instruction involving several options, it is not possible to claim that the benefits were solely due to the enriched input.
Leow’s (1998) study also investigated the effects of a noticing instruction. Following Tomlin and Villa (1994), Leow distinguished three levels of noticing (alertness, orientation and detection) and set out to investigate these by asking learners of L2 Spanish to perform a crossword that required attention to the irregular third-person singular and plural preterit forms of stem changing –ir verbs. Orientation was operationalized through a noticing instruction: ‘Please note that some of the forms of the verbs are irregular’. The opportunity for detection was provided by ensuring that the irregular forms needed to complete some of the clues were available in a number of the other clues. While all four groups were designated as + alertness, they differed in terms of whether they were – orientation/ – detection (Group 1 – the control group), + orientation/ – detection (Group 2), + orientation/ + detection (Group 3) and – orientation/ + detection (Group 4). The results showed that Groups 3 and 4 outperformed both the control group and Group 2 on all the posttests but did not themselves differ significantly. In other words, the groups that had the opportunity to detect the target forms in the input outperformed those that did not, and simply orientating the learners to the existence of the form without the opportunity for detection had no effect.
A number of studies have investigated the effects of instruction that involved simple exposure to the target structure through enriched input and instruction that included explicit reference to the target structure (often in the form of rule presentation).
Alanen (1995) conducted a study with four groups; (1) a control group, (2) an ‘enhancement group’ which received just enriched input in two fifteen minute instructional periods, (3) a ‘rule group’ that received just explicit instruction and (4) a ‘rule + enhanced group’ that received both enriched input and explicit instruction. The enriched input took the form of two short texts in which the target features had been italicized. Learning was measured by means of a sentence completion task, a grammaticality judgement task, and a rule statement task. The learners were also asked to think aloud during the treatment. The main finding was that groups (3) and (4) outperformed groups (1) and (2). Also, there was no difference between groups (1) and (2) or between groups (3) and (4). One reason why the enriched input had no clear effect on acquisition in this study might have been that the period of instruction was too short.
Rosa and O’Neill (1999) compared the effects of instruction directed at learning the Spanish contrary to fact conditional (a complex structure) by university-level learners of L2 Spanish. Four types of instruction were included in this study; (1) rule explanation + rule search, (2) rule explanation + no rule search; (3) no rule explanation + rule search; (4) no rule explanation + no rule search. Acquisition was measured by means of a time-pressured multiple-choice recognition task while think-aloud protocols were used to measure awareness of the rule. Awareness was operationalized as a verbal reference to the target feature during task execution and thus might be considered a measure of intake. Two types of awareness were distinguished – “noticing” if no reference was made to the underlying rules and “understanding” if there was. All the groups improved from pre- to posttest. The instructed condition (i.e. (1)) proved superior to the enriched input only condition (i.e. (4)). Also, more aware participants, both those showing greater “noticing” and those showing greater “understanding”, performed better on the multiple-choice recognition task.
Radwan (2005) also investigated the effects of instruction involving a focus on meaning only compared with input enhancement and rule provision on learning, and awareness of English dative alternation. He also investigated if differences in awareness affected learning. Forty-two lower-intermediate participants were pretested for prior knowledge of the target structure, and one day later given a short story to read which contained a high number of datives. Reading of the short story was followed by comprehension questions. The next day, a similar treatment was administered but in addition participants were given a narration task which involved describing a set of pictures. Participants were asked to think aloud while completing the task in order for the researcher to gauge their awareness. The treatments were followed by a posttest (one day later) and a delayed posttest (one month later). A control group only completed the tests. Radwan found a significant advantage for the rule-group over the other groups, which failed to make significant progress. This advantage was maintained on the delayed posttest. He also found that participants showing a greater degree of awareness during the narration task did better on the tests. However, awareness at the level of noticing was not as good a predictor of learning as awareness at the level of understanding.
It is not easy to draw clear conclusions from these studies. Although they have investigated what appear to be similar constructs (e.g. enhanced input, noticing, intake, directed learning) they have operationalized these in very different ways, drawing on very different disciplines in doing so (i.e. different schools of psychology and language pedagogy). As a result, it is difficult to compare results. The following conclusions, therefore, must be viewed as tentative:
1. There is some evidence that enriched input can help L2 learners acquire some new grammatical features and use partially learned features more consistently although it may not enable learners to eradicate erroneous rules from their interlanguage. Enriched input appears to work best if the instructional treatment provided learners with extensive exposure to the target features and was relatively prolonged (i.e. ‘input flooding’). Enriched input both with and without highlighting of the target structure has been shown to assist acquisition.
2. Enhanced input consisting of a noticing instruction may assist noticing and acquisition.
3. Noticing appears to be related to learning especially if it involves ‘detection’ and ‘awareness’.
4. Simple exposure to enriched input typically results in low levels of awareness of the target structure.
5. Exposure to enriched input has been consistently shown to be less effective than instruction that is more explicit (e.g. a rule-search or an explicit instruction condition).
Finally, it is worth noting that none of the studies reviewed above attempted to distinguish the effects of enhanced input on the acquisition of implicit and explicit L2 knowledge. It should also be noted that many of the studies used grammaticality judgement tests to measure acquisition but invariably these were of the untimed type.
The study reported below builds on the previous research by examining the effects of enhanced input in two conditions: (1) input that has been enriched by seeding with the target structure and (2) enriched input combined with a noticing instruction. A unique feature of the study is that it will examine the effects of these two types of enhanced input on both intake and acquisition. A further feature is that we will attempt to distinguish acquisition in terms of implicit and explicit knowledge.
The research questions this study addressed were:
1) What are the effects of enhanced input on a) intake and b) acquisition of English negative adverbs?
2) What difference is there in the effects of two types of enhanced input (i.e. enriched input and enriched input + noticing) on a) intake and b) acquisition of English negative adverbs?
Enhanced input was operationalized by means of three reproduction tasks. In the case of the enriched input condition learners completed meaning-focused tasks that had been seeded with several examples of negative adverbs. In the case of the enriched input + noticing condition learners completed the same tasks but were also instructed to pay attention to the position of the auxiliary verb in the sentences in the tasks’ input. Measures of intake were obtained from the participants’ performance of the treatment tasks. Acquisition was measured by means of a timed and an untimed grammaticality judgment test.
Participants completed the pretests and were then randomly assigned to one of the two treatment conditions (enriched input vs. enriched input + noticing). Each participant completed a treatment task on three separate occasions in their selected condition. Immediately following the third occasion the posttests (a timed and an untimed grammaticality judgment test) were administered. The delayed posttests were administered one week later. There was no separate control group in the study. However, as an alternative to a control group, the participants’ performance on the items measuring knowledge of negative adverbs (the target structure) was compared with their performance on distractor items in the tests. Table 1 summarizes the design of the study.
Table 1: Design of the study
Week 1 – Pretest (all participants)
Week 2-3-4 – Treatments
Enriched input N=17
Enriched input + noticing N=11
Week 4 Immediate posttest
Week5 Delayed posttest
Participants were 28 students from an upper-intermediate proficiency level in a New Zealand private English language school. They volunteered to join the study in exchange for financial compensation of approximately NZD$10 per hour. Sixteen of the participants were female and 12 male. Fifteen of them came from East Asia (Japan, Korea, China) and four participants came from Switzerland. Altogether, participants came from a total of eleven different countries and had ten different first languages. Most of the participants had lived in an English speaking country for less than six months.
The participants had been given an in-house placement test earlier in the year to determine their class level. After one week, consultation between the student and the classroom teacher, and where necessary the Director of Studies, took place. The school considered upper-intermediate level students to be the equivalent of level B2 of the European Framework. That is, it was expected that students:
Can understand the main ideas of complex text on both concrete and abstract topics, including technical discussions in his/her own field of specialization. Can interact with a degree of fluency and spontaneity that makes regular interaction with native speakers quite possible without strain for either party. Can produce clear detailed text on a wide range of subjects and explain a viewpoint on a topical issue giving the advantages and disadvantages of various options (Council of Europe, 1996)
Learners at the upper-intermediate level were used in this study in the attempt to ensure that they were developmentally ready to acquire the target structure (negative adverbs) but without having yet done so.
The target structure was negative adverbs with inversion of subject and auxiliary, as in this example:
Seldom had he seen such a beautiful woman.
Other negative adverbs requiring subject-verb inversion are “never”, “rarely”, “seldom” and “hardly.
R. Ellis (2006) proposed a number of criteria for determining the level of difficulty of grammatical structures as implicit and explicit knowledge. The difficulty of negative adverbs is now considered in the light of these criteria.
1. Difficulty of negative adverbs as implicit knowledge
Input frequency; negative adverbs are relatively rare as confirmed by an analysis of the British National Corpus (frequency ranged from 276 occurrences for ‘not only was’ to 3 for ‘seldom do’ and fewer for a range of other adverb/auxiliary combinations).
Saliency; negative adverbs can be considered salient in that they are sentence-initial but the inversion of subject and verb, which involves the use of an auxiliary, is probably less salient as the auxiliary is typically unstressed.
Functional value; the ‘function’ of this structure (the negative meaning associated with the adverb) is conveyed lexically and thus the subject-verb inversion is redundant.
Regularity; only negative adverbs require subject-verb inversion – other adverbs of time, place and manner (e.g. “yesterday”, “there” and “rapidly”) take normal subject-verb word order.
Processability; in terms of Pienemann’s (1998, 2005) hierarchical processing operations, negative adverbs with subject-verb inversion will be late acquired (i.e. they involve what Pienemann refers to as S-procedure).
2. Difficulty of negative adverbs as explicit knowledge
Conceptual clarity; negative adverbs with subject-verb inversion are functionally relatively simple but formally complex as they involve a variety of auxiliary forms; the declarative rule required to explain them is also not easily extractible from data.
Metalanguage: it will be difficult to avoid the use of metalanguage in articulating the declarative rule for negative adverbs (e.g. ‘’adverb”, “negative” “auxiliary”, “subject”, “main verb”).
In short, Ellis’ criteria indicate that negative adverbs with subject verb inversion constitute a difficult structure as both implicit and explicit knowledge. It should be noted, however, that Robinson (1996) used negative adverbs as his ‘easy rule’. Perhaps it can be considered ‘easy’ in comparison with the other structure Robinson investigated (pseudo-clefts of location) but it is likely that many learners will fail to acquire negative adverb structures without some form of instructional intervention. The purpose of the study was to examine whether intervention in the form of enhanced input can assist them to acquire it.
This study used three types of treatment tasks (described below). Each treatment task consisted of oral or written input in the form of a text about a range of general interest topics that had been seeded with several instances of the target structure (i.e. the texts were enhanced in terms of the frequency of the target structure). Altogether each student was exposed to 36 sentences with negative adverbs. The nature of the exposure differed. In the case of the enriched input condition the learners were simply instructed to complete the tasks. That is, they were given no indication of what to look out for. In the case of the enriched input + noticing condition they learners were given the following instructions:
‘Listen carefully and pay attention to where the auxiliary verb comes in each
sentence. For example in the sentence ‘Rarely has so much rain fallen in such a
short time’ the auxiliary is ‘has’ and it comes before the subject of the sentence ‘so
That is, in the noticing condition their attention was specifically and explicitly drawn to the target structure.
Each student completed one the three types of tasks described below. Each student took part in three treatment sessions involving the same type of task. It should be noted that although different students completed different tasks, all three types of tasks figured equally in both the enriched input and the enriched input + noticing conditions. In the case of the former, five students completed the dictation task, 5 the individual reconstruction task and 7 the collaborative reconstruction task. In the case of the latter, 4 students completed the dictation task, 3 the individual reconstruction task and 4 the collaborative reconstruction task.
The treatment tasks were:
1. Dictation. Participants were asked to listen to a passage of about 60-70 words on a computer and write it out section by section as in a standard dictation. Before the actual treatment, they completed three practice passages. Participants first listened to the entire passage and then again section by section while writing out each section as they heard it. Each section contained no more than 10 words and mostly around seven or eight. The treatment thus involved immediate recall. There were four passages containing three target sentences in each treatment session.
2. Individual reconstruction. The individual reconstruction treatment involved delayed rather than immediate recall of the texts. Participants were asked to listen twice to a passage of about 60-70 words. Participants were allowed to take notes. They then attempted to reconstruct it by writing it out. While they were doing this, they were asked to talk-aloud (Ericsson & Simon, 1993). Instructions for this treatment were in the form of a video demonstrating talk-aloud.
3. Collaborative reconstruction. The collaborative reconstruction treatment was similar to the individual reconstruction treatment except that two participants were paired and were asked to reconstruct the text together. It therefore also involved delayed recall.
It should be noted that differences in performance of these three tasks was not the focus of the study reported in this chapter as both the enriched input and enriched input + noticing conditions involved all three tasks.
The same tests were administered on three occasions – as a pretest, an immediate posttest and a delayed posttest (see Table 1). There were two tests; a timed grammaticality judgment test (GJT) and an untimed GJT. A description of these follows.
Timed GJT. This test consisted of 50 sentences, 20 of which contained negative adverbs. Of these 10 were grammatical and 10 ungrammatical sentences. The other 30 items consisted of sentences with other adverbial structures relating to adverb position and to the difference in form between adverbs and adjectives. In this test sentences were shown on screen and participants had to press the “enter” key if they thought the sentence on the screen was correct, and the left-hand “shift” key if they thought it was not. The keys were labelled with stickers indicating “correct” and “incorrect”. There were eight practice sentences during which the researcher was present to give clarification where needed. The tests were first trialled on native speakers and similar learners in order to establish a time limit for each sentence. The time limit for each sentence was longer than the mean time taken by the native speakers on that sentence but shorter than that of the non-native speakers. The learners were given relatively more time on the earlier than the later items in the test. They were told that they might not be able to respond to all the items in time but that they should try to answer as many as they could.
Untimed GJT: The untimed test was the same as the timed test (i.e. consisted of the same sentences, although in a different order). However, there was no time limit for judging each sentence. Students entered their responses on the computer as for the timed GJT. Previous research (see Chapter 2) has shown that timed and untimed GJTs measure separate constructs. In line with the findings of these studies, we propose that the timed GJT (especially the grammatical sentences) provides a measure of the learners’ implicit knowledge and the untimed GLT (ungrammatical sentences) provides a measure of their explicit knowledge.
The reliabilities of negative adverb items and the control items were assessed by means of Cronbach Alpha and are shown in Table 2. The alphas ranged from a high of .928 to a low of .605.
Table 2 Reliability figures for the grammaticality judgement tests
Pre Tim Pre Unt Post Tim Post Unt Dpt Tim Dpt Unt
Control 0.841 0.725 0.928 0.789 0.84 0.852
Negative adverb 0.605 0.770 0.779 0.882 0.732 0.857
Pre = pretest, Post = posttest, Dpt = delayed posttest, Tim = timed, Unt = untimed.
The tasks and tests completed by the learners were used to obtain the following measures: (1) intake of the target structure, (2) acquisition of L2 implicit knowledge of the target structure and (3) acquisition of L2 explicit knowledge of the target structure. The measures are described below.
Intake. As discussed in the introduction, intake can be operationalized as information held in short-term memory after exposure to the target language. Intake then, needs to be determined immediately after exposure to the target feature. In the present study correct suppliance of the target items during the treatments was taken as a measure of intake. The time between hearing the input and reproducing it was sufficiently long to prohibit mimicking but sufficiently short for it to remain in short-term memory. This is self-evidently the case for the individual and collaborative reconstruction tasks as the learners could not have memorized the whole texts they had heard. It was also likely in the dictation task as the chunks the learners were asked to reproduce were too long for easy memorization. The reproductions of the learners were inspected and occasions where they attempted to reproduce a sentence with a negative adverb identified. Responses were judged as correct as long as the participants inverted subject and auxiliary. Spelling and other errors not relating to the target structures were discounted (e.g. one learner spelt ‘do’ as ‘to’). Also, there was no expectancy that learners would reproduce the exact words of an input sentence. For example, for the sentence “No sooner does there seem to be a solution then another problem arises”, one learner responded:
No sooner is it solution … and the other problem is the ice
Here the wrong auxiliary was chosen but the word order was correct so the sentence was scored as correct. However, any sentence starting with an adverb and followed by a subject were scored as “incorrect”. Sentences with no auxiliary (e.g. No sooner that I arrived…) or without a subject (e.g. No sooner had arrive) were also scored as incorrect.
Implicit L2 knowledge. The implicit knowledge scores were arrived at by totalling the number of correct judgments that the learners made in the timed GJT. Total scores and also separate scores for the 10 grammatical and the 10 ungrammatical sentences were calculated as previous research has indicated that these measure separate constructs (R. Ellis, 2005; Hedgcock, 1993). To measure acquisition of implicit L2 knowledge, gain scores from pre- to immediate posttest, from pre- to delayed posttest and from immediate to delayed posttest were calculated.
Explicit L2 knowledge. A similar scoring procedure was followed for measuring explicit L2 knowledge but this time the responses to the untimed GJT were used. Again, total scores and separate scores for the grammatical and ungrammatical sentences were calculated. To measure acquisition of explicit L2 knowledge gain scores from pre- to immediate posttest, from pre- to delayed posttest and from immediate to delayed posttest were calculated.
Control items. Learners’ responses to the 30 items in the GJTs that did not contain the target structure were used as the control items in this study. Total scores on these items together with scores for the grammatical and ungrammatical items separately were calculated. Gain scores were then computed.
As participants in the study completed multiple treatments and tests, repeated measures analysis of variance models (ANOVAs) were used to investigate group differences. For post-hoc analyses the Least Significant Differences (LSD) method was used. This method is considered liberal in that it compares means for all possible data sources separately, rather than combined. Considering the fairly small number of data sources, and considering that the present study was exploratory, the use of LSD was deemed acceptable. For all statistical analyses the alpha level was set at .05.
First the results for intake will be presented followed by those for acquisition.
Table 3 below shows the enriched input and enriched input + noticing groups’ scores for intake of negative adverbs. The results show a clear improvement for scores obtained by both groups from time one to time two and from time two to time three. The time difference was statistically significant (F(1,81)=.28.82, p<.001). Scores for the enriched input + noticing group are higher at treatment time two and three than those for the enriched input group. However, a one-way ANOVA did not show a significant effect for treatment condition (F(1,81)=2.41, p=.124). A t-test for two independent groups also failed to show a significant difference for the Time 3 scores (t = 1.436; df 26; p = .348). Table 3: Intake scores for negative adverbs. Time 1 Mean SD Time 2 Mean SD Time 3 Mean SD Enriched input(N=17) .155 .152 .328 .236 .441 .276 nnriched input + Noticing (N=11) .126 .113 .454 .356 .606 .327 Acquisition Gain scores for the timed and untimed GJTs were calculated separately. Table 4 shows the mean gain scores 1) from pretest to posttest, 2) from pretest to delayed posttest, and 3) from posttest to delayed posttest for the timed GJTs. Gain scores from pre- to posttests were higher for the enriched input than the enriched input + noticing Group. They were also higher for the grammatical than the ungrammatical items. Table 4: Gain scores negative adverbs and controls on the timed GJTs NA Timed tests Negative Adverbs Grammatical Ungramm. Gain SD Gain SD Control Grammatical Ungramm. Gain SD Gain SD Pretest to posttest Enriched input (n=17) .241 .245 .058 .2 -.041 .197 .15 .213 Enriched input + noticing (n=11) .118 .357 .072 .241 -.009 .347 .127 .2 Pretest to delayed posttest Enriched input (n=17) .311 .228 .076 .301 .052 .18 .12 .141 Enriched in put + noticing (=11) .081 .354 -.027 .296 -.081 .389 .213 .23 Posttest to delayed posttest Enriched input (n=17) .07 .323 .017 .283 .094 .265 -.029 .261 Enriched input + noticing (n=11) -.036 .456 -.1 .322 -.072 .337 .086 .282 First, the differences between total gain scores on target and control items were compared by means of a 2 (negative adverbs/control) x 3 (gain scores) repeated measures ANOVA. This showed no statistically significant difference (F (1,333) = 1.16, p=.283). In other words, the instruction had no effect on acquisition of negative adverbs as measured by total scores of the timed tests. However, for the grammatical items in the timed tests the gain scores for the negative adverbs were significantly greater than for the control items (F (1,165 )= 9.71, p= .002) with a medium effect size (d=.48). There was also a significant difference on the ungrammatical items (F (1,165) = 4.49, p= .035), but this was to the advantage of the control items. Next, an ANOVA was performed to establish if there was an effect for instructional condition. This was not the case for gain scores on the grammatical items from pretest to posttest (F (1,54) = .31, p=.581). However, from pretest to delayed posttest there was a difference (F (1,54) = 4.95, p=.03), to the advantage of the enriched input condition. The effect size was (d=.62). Descriptive statistics for the Untimed GJTs are shown in Table 5. The gain scores for the negative adverb grammatical (but not ungrammatical) items are generally larger than those for the control items. Both the enriched input and the enriched input and noticing groups manifested gains on the grammatical but not the ungrammatical items. Table 5: Gain scores for negative adverbs and controls on the untimed GJTs NA Untimed tests Negative Adverbs Grammatical Ungramm. Gain SD Gain SD Control Grammatical Ungramm. Gain SD Gain SD Pretest to posttest Enriched input (n=17) .229 .271 -.058 .19 -.035 .176 0 .269 Enriched input + noticing (n=11) .218 .292 -.1 .282 .054 .211 .04 .204 Pretest to delayed posttest Enriched input (n=17) .252 .316 -.041 .308 -.011 .226 .05 .22 Enriched input + noticing (n=11) .254 .5 -.236 .372 .09 .344 -.004 .328 Posttest to delayed posttest Enriched input (n=17) .023 .185 .017 .255 .023 .204 .05 .235 Enriched input + noticing (n=11) .136 .441 -.136 .297 .036 .254 -.045 .342 A 2 (target/control) x 3 (gain scores) repeated measures ANOVA using total scores showed no significant difference between negative adverb and control items (F (1,333) = .147, p=.225) indicating that instruction had no overall effect on participants’ acquisition of negative adverbs as measured by the untimed tests. However, in the case of the grammatical items a significant difference was found (F (1,165 )= 15.75, p<.001) with a medium effect size (d=.611). Gains were greater for the negative adverb items. For ungrammatical items there was a significant difference between gain scores on negative adverb and control items (F (1,165) = 5.37, p=.021), however this was to the advantage of the control items. Next, ANOVAs were performed to establish if there was an effect for instructional condition. No statistically significant differences were found on the gain scores for the grammatical items (pretest to posttest (F(1,54)=.3, p=.586); pretest to delayed posttest (F(1,54)=1.06, p=.307; posttest to delayed posttest (F(1,54)=.74, p=.394)). Summary The following is a summary of the main results: 1. Intake scores as a whole rose over the period of instruction but there was no difference between the enriched input and enriched input + noticing groups. 2. Overall the instruction had no effect on acquisition as measured by the total scores of the timed GJTs. However, effects were evident when the grammatical and ungrammatical sentences were examined separately. The instruction resulted in higher scores for negative adverb items in the case of the grammatical sentences but in lower scores than the control items for the ungrammatical sentences. The enriched input group outperformed the enriched input + noticing Group in the long term (i.e. in gain scores between pre- and delayed posttest) on the grammatical sentences in the timed GJTs. 3. Overall the instruction had no effect on acquisition as measured by total scores on the untimed GJTs. However, again, effects were evident when the grammatical and ungrammatical sentences were examined separately with the same pattern of results as for the timed GJTs. There was no difference in any of the untimed GJT gain scores between the enriched input and enriched input + noticing Groups. Discussion The first research question asked what effects the enhanced input had on intake. Intake was measured in terms of the learners’ use of the target structure (negative adverbs with subject-verb inversion) in three different reproduction tasks (dictation, individual reconstruction and collaborative reconstruction) which were completed on three different occasions. An inspection of the means scores in Table 3 shows that on the first occasion intake was negligible (only 16% for the enriched input Group and 13% for the enriched input + noticing group). Over time, however, intake increases steadily so that by the third occasion intake scores have reached 44% for the enriched input condition and 61% for the enriched + noticing Group. The time difference was statistically significant. Thus, it would seem that intake increases along with exposure to the target form. However, there was no statistically significant difference in the intake scores of the enriched input group (which only received exposure to sentences containing negative adverbs) and the enriched input + noticing Group (which received the same exposure but was also directed to pay attention to the sentences with negative adverbs). In other words, the noticing instruction did not lead to significantly greater intake. This indicates that it was repeated exposure to the target structure that enabled the learners to notice the target structure and rehearse it in short-term memory sufficiently to reproduce it. Leow (1998) also found that an orienting instruction had no effect on learners’ acquisition of irregular Spanish verb forms. This study reports that a very similar orienting instruction had no effect on learners’ intake of a difficult syntactic feature. Clearly, it would be premature to conclude that noticing instruction is ineffective in assisting acquisition but, to date, there is no evidence that it is. The second research question concerned whether the enhanced input treatments resulted in acquisition. Acquisition was measured by means of timed and untimed GJTs with a view to providing relatively separate measures of implicit and explicit knowledge. The intake scores suggest that the learners obtained sufficient information from the tasks to make acquisition of the target structure possible. However, intake does not guarantee acquisition. It is possible that learners are able to notice and rehearse a grammatical form in their short-term memories and therefore reproduce it and yet be unable to integrate it into their interlanguage systems. This is the very point of the theoretical distinction between ‘intake’ and ‘acquisition, as in Gass’ (1997) model. The results in Table 4 show that overall the learners did not perform better on the target items than on the control items in the timed GJTs. In other words, there was no evidence that the instruction led to acquisition. However, when the grammatical and ungrammatical items were examined separately, it emerged that the learners performed better on the target items in the case of the former and worse in the case of the latter. Similar results were obtained for the untimed GJTs. This asymmetry in performance on grammatical and ungrammatical items has been observed in other studies (e.g. R. Ellis, 2005; Hedgcock, 1993). One interpretation of these results is that the enhanced input assisted the acquisition of implicit knowledge but not explicit knowledge. Such an interpretation makes sense given that the enhanced input treatment of both groups in this study favoured the development of implicit rather than explicit knowledge. There was no deductive or inductive explicit instruction that might have assisted the development of explicit knowledge. In fact, the noticing instruction seems to have had a deleterious effect on the learners’ explicit knowledge of negative adverbs as the gain scores on the ungrammatical items in both the timed and untimed GJTs were very small and frequently negative. Indeed, in the case of the ungrammatical items of the untimed GJT (arguably the best measure of explicit knowledge), the gain scores of the enriched input + noticing group were all negative. The likely explanation for this is the cognitive difficulty of understanding how this particular structure works without the assistance of detailed explicit instruction. Given that there was no difference in the intake levels of the two treatment groups it might be predicted that there should be no difference in their levels of acquisition. This proved to be the case for the gain scores from pre- to immediate posttest on the grammatical items of the timed GJT but not for the gain scores from pre- to delayed posttest. That is, in the long term, the incidental exposure to enriched input worked better than the enriched input combined with the noticing instruction. No group differences were evident on the ungrammatical items. What these results suggest is that asking students to consciously attend to the target structure can actually impede the acquisition of implicit knowledge. This result accords with the findings of other studies of incidental instruction. N. Ellis (1993), for example, found that incidental instruction consisting of enriched input worked better than a more explicit form of instruction when the structure was a difficult one. No group differences were evident on the grammatical or ungrammatical items in the untimed test. It might have been expected that the enriched input + noticing condition would have helped learners improve their ability to judge the ungrammatical items of the untimed GJT if, as we have argued, this constitutes a measure of explicit knowledge. However, as we noted above, the conceptual difficulty of negative adverbs may have prevented the learners from benefiting from deliberate attention to this structure. They were simply unable to work out the rule. To sum up, enhanced input in the form of oral texts seeded with exemplars of a difficult target structure resulted in intake and also in the acquisition of implicit knowledge (as measured by the grammatical sentences of a timed GJT). However, it did not benefit explicit knowledge (as measured by an untimed GJT). Providing learners with a noticing instruction in addition to the enriched input conferred no advantage for either intake or acquisition, possibly because of the conceptual difficulty of the particular target structure of this study. Conclusion This study has shown that enhanced input in the form of enriched input resulted in intake and assisted the acquisition of implicit knowledge. It has also shown that asking students to pay attention to the target structure conferred no additional advantage for either intake or acquisition. The study is supportive of the claims that have been advanced on behalf of focus-on-form instruction (Doughty & Williams, 1998) and show shows that even a very unobtrusive focus-on-form strategy can be effective. However, the results of this study do not support Norris and Ortega’s (2000) general finding, namely that explicit instruction is more effective than implicit instruction. This might have been because the noticing instruction provided in this study was insufficiently explicit to assist the learners. We have attempted to look at very specific instructional options. In this respect, our study differs from many others which have tended to investigate form-focused instruction through treatments that combine a number of options. While such studies may have ecological validity in that they reflect common pedagogical practice, they are problematic where SLA theory testing is concerned. Norris and Ortega (2000) complained that the essential features that distinguish one type of instruction from another have been inconsistently operationalized. This problem can only be overcome if researchers investigate very clearly defined instructional options. If we want to know what effect different forms of input have on L2 acquisition we need to isolate specific instructional strategies and test for their effect on acquisition. It is also important to attempt to distinguish the effects of instruction on implicit and explicit knowledge. As both Norris and Ortega (2000) and Doughty (2003) have argued, the tests that have been typically used have been biased in favour of explicit knowledge. In this study we have tried to obtain separate measures of implicit and explicit knowledge using a timed and untimed GJT. We recognize that GJTs in general are controversial (R. Ellis, 1991) and we acknowledge that some readers may remain sceptical of the construct validity of the tests. However, we note that we did obtain different results for the two tests (and also for the grammatical and ungrammatical sentences in the tests) and that these results are interpretable in terms of the general finding reported in Chapters 2 and 3, namely that GJTs can be used to provide relatively separate measures of implicit and explicit knowledge. Finally, we wish to acknowledge a number of weaknesses of our study. The sample size was relatively small (the enriched in put + noticing group had only 11 learners). There was no control group, although we were able to use the non-target items in the GJTs as a point of comparison. The total exposure time provided by the instruction was relatively limited; arguably exposure to 36 exemplars of the target structure does not amount to an input-flood. But then it is perhaps all the more impressive that it produced a measurable effect. References Alanen, R. (1995). Input enhancement and rule presentation in second language acquisition. In R. Schmidt (Ed.), Attention and awareness in foreign language learning (pp. 259-302).Honolulu: University of Hawai'i. Chaudron, C. (1985). Intake: on models and methods for discovering learners' processing of input. Studies in Second Language Acquisition, 7, 1-14. Council of Europe (1996). The Common European Framework. [Online document]. Last retrieved January 16, 2007 from: http://www.coe.int/t/dg4/linguistic/Source/Framework_EN.pdf Doughty, C. (2003). Instructed SLA: Constraints, compensation and enhancement. In C. Doughty & M. Long (Eds.), The handbook of second language acquisition (pp. 256-310). Malden, MA.: Blackwell. Doughty, C. & Williams, J. (1998a). Issues and terminology. In C. Doughty and J. Williams (Eds.), Focus on form in classroom second language acquisition (pp. 1-11). New York: Cambridge University Press. Doughty, C. & Williams, J. (1998b). Pedagogical choices in focus on form. In C. Doughty and J. Williams (Eds.), Focus on form in classroom second language acquisition (pp. 197-261). New York: Cambridge University Press. Ellis, N. (1993). Rules and instances in foreign language learning: interactions of explicit and implicit knowledge. European Journal of Cognitive Psychology, 5, 289-318. Ellis, N. C. (Ed.) (1994). Implicit and explicit learning of languages. London: Academic Press. Ellis, R. (1991). Grammaticality judgments and second language acquisition. Studies in Second Language Acquisition, 13, 161-186. Ellis, R. (2004). The definition and measurement of L2 explicit knowledge. Language Learning, 54, 227-275. Ellis, R. (2005). Measuring implicit and explicit knowledge of a second language: a psychometric study. Studies in Second Language Acquisition, 27, 141-172. Ellis, R. (2006). Modelling learning difficulty and second language proficiency: The differential contributions of implicit and explicit knowledge. Applied Linguistics, 27, 431-463. Ericsson, K., & Simon, H. (1993). Protocol analysis: Verbal reports as data (2nd ed.). Boston: MIT Press. Fotos, S. & Ellis, R. (1991). Communicating about grammar: a task-based approach.TESOL Quarterly, 25, 605-628. Gass, S. M. (1997). Input, interaction, and the second language learner. Mahwah: Lawrence Erlbaum. Han, Y., & Ellis, R. (1998). Implicit knowledge, explicit knowledge and general language proficiency. Language Teaching Research, 2, 1-23. Hedgcock, J. (1993). Well-formed vs. ill-formed strings in L2 metalingual tasks: Specifying features of grammaticality judgements. Second Language Research, 9, 1-21. Hulstijn, J. 2003. Incidental and intentional learning. In C. Doughty & M. Long (Eds.), The handbook of second language acquisition (pp. 349-381). Krashen, S. (1981). Second language acquisition and second language learning. Oxford: Pergamon Press. Kumaravadivelu, B. (1994). Intake factors and intake processes in adult language learning. Applied Language Learning, 5, 33-71. Leeman, J., Arteagoitia, I.,Fridman, D & Doughty, C. (1995). Integrating attention to form incontent-based Spanish instruction. In R. Schmidt (Ed.), Attention and awareness in foreign language learning (pp. 217-258). Honolulu: University of Hawai'i. Leow, R. P. (1993). To simplify or not to simplify. Studies in Second Language Acquisition, 15, 333-355. Leow, R. P. (1995). Modality and intake in second language acquisition. Studies in Second Language Acquisition, 17, 79-89. Leow, R. P. (1998b). Toward operationalizing the process of attention in second language acquisition: evidence for Tomlin and Villa's (1994) fine-grained analysis of attention. Applied Psycholinguistics, 19, 133-159. McLaughlin, B. (1978). The monitor model: some methodological considerations. Language and Society, 7, 183-213. Norris, J. M., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning, 50, 417-528. Jourdenais, R., Ota, M., Stauffer, S., Boyson, B. & Doughty, C. (1995). Does textual enhancement promote noticing? A think-aloud protocol analysis. In R. Schmidt (Ed.), Attention and awareness in foreign language learning (pp. 183-216). Honolulu: University of Hawai'i Press. Pienemann, M. (1998). Language processing and second-language development: processability theory. Amsterdam: John Benjamins. Pienemann, M. (2005). An introduction to processability theory. In M. Pienemann (Ed.), Cross-linguistic asspects of processability theory (pp. 1-60). Amsterdam: John Benjamins. Radwan, A. A. (2005). The effectiveness of explicit attention to form in language learning. System, 33, 69-87. Robinson, P. (1996). Learning simple and complex second language rules under implicit,incidental, rule-search, and instructed conditions. Studies in Second Language Acquisition, 18, 27-67. Rosa, E., & O'Neill, M. D. (1999). Explicitness, intake, and the issue of awareness. Another piece to the puzzle. Studies in Second Language Acquisition, 21, 511-556. Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics, 11, 129-158. Schmidt, R. (1994). Deconstructing consciousness in search of useful definitions for applied linguistics. AILA Review, 11, 11-26. Sharwood Smith, M. (1991). Speaking to many minds: on the relevance of different types of language information for the L2 learner. Second Language Research, 7, 118-132. Sharwood-Smith, M. (1993). Input enhancement and instructed second language acquisition. Studies in Second Language Acquisition, 15, 65-85. Shook, D. J. (1994). FL/L2 reading, grammatical information, and the input-to-intake phenomenon. Applied Language Learning, 5, 57-93. Swain, M. (1988). Manipulating and complementing content teaching to maximize second language learning. TESL Canada Journal, 6, 68-83. Tomlin, R. & Villa, V. (1994). Attention in cognitive science and second language acquisition. Studies in Second Language Acquisition, 16, 183-203. Trahey, M. 1996. Positive evidence in second language acquisition: Some long-term effects. Second language Research, 12, 111-139. Trahey, M. & White, L. (1993). Positive evidence and preemption in the second language classroom. Studies in Second Language Acquisition, 15, 181-204. White, J. (1998). Getting learners’ attention: A typographical input enhancement study. In C. Doughty & J. Williams (Eds.), Focus on form in classroom second language acquisition (pp. 85-113). New York: Cambridge University Press.