US-China Foreign Language, January 2019, Vol. 17, No. 1, 43-47
doi:10.17265/1539-8080/2019.01.006 A Critical Review: Evaluating the Effectiveness of Explicit
Instruction on Implicit and Explicit L2 Knowledge
Sami Sulaiman Alsalmi
School of education, University of Bristol, Bristol, United Kingdom
This review discuss Akakura’s rearch study, entitled “valuating the Effectiveness of Explicit Instruction on
Implicit and Explicit L2 Knowledge”. The argument prented here will be developed by means of a critique of
Akakura’s study, in turn addressing a summary of the study with the focus placed on the method and st
atistical
techniques, as well as the evaluation of the study. Some explanations will be added to the summary of study to
make some points more obvious, specifically tho that do not receive enough description in the study. In addition,
given that the study is rather broad and includes many situations that are worthy of discussion but that they cannot
be covered in a paper, the evaluation will be narrowed down to concentrate on two aspects of the study: measures
of implicit knowledge and explicit instructions (treatment stage).
Keywords: explicit instruction, implicit knowledge, explicit knowledge
Summary of the Study
Akakura’s (2012) study sought to explore to what extent explicit instruction can develop cond-language
learners’ implicit and explicit knowledge of English articles. Explicit instruction is concerned with “developing a metalinguistic awareness of the target rule” (Ellis, 2009, p. 54). That is, learners are provided with the instruction of the target grammatical rules. Implicit knowledge refers to the procedures comprising “knowledge which can be easily and rapidly accesd in unplanned language u. In contrast, explicit knowledge exists as a declarative fact that can only be accesd through the application of attentional process” (Ellis, 2009, p. 12). The study claims that rearch has not enough discusd which measures can best test the spontaneous status of the implicit grammatical knowledge.
The study employs a quasi-experimental design with a pretest/posttest and delayed test model entailing two groups: experimental (N = 49) and the control group (N = 45). In each testing stage, participants were expod to four measures: elicited imitation task, oral production task (for implicit knowledge), grammaticality judgement task, and metalinguistic knowledge task (for explicit knowledge). A pretest was run first for the two groups, and then the experimental group was expod to explicit instruction, using computer-assisted language learning, for one week following th
e pretest. The form/function mappings of articles were explained to participants, and then the participants were provided with form-focud exercis and quizzes. The posttest was administered after the participants completed article lessons achieved by explicit instruction. The delayed posttest was then completed six weeks after the treatment.
Sami Sulaiman Alsalmi, Ph.D. candidate at Bristol University in the UK.
All Rights Rerved.
EV ALUATING THE EFFECTIVENESS OF EXPLICIT INSTRUCTION
44 As indicated above, four instruments were employed to measure the two types of knowledge, and a succinct description of each measure is provided below.
Elicited Imitation Task (EIT)
This task required a participant to listen to a storey while looking at a quence of pictures depicting it. Half the recorded storey contained ntences that are incompatible with the pictures and simultaneously included grammatical (N = 10) and ungrammatical (N = 10) articles. Participants were then asked to decide whether or not the ntences match the picture, and to repeat the state
ment when they heard a bell sound. The inclusion of picture plausibility of the storey is to ensure that a participant’s attention is on meaning and not form. The study does not describe the overall goal of this task. According to literature in the field of implicit knowledge acquisition, the underlying assumption of this task is that if a participant could repeat the statement under time constrains and orally correct the ungrammatical articles spontaneously, it would imply that the participant had internalized the target articles.
Oral Production Task (OPT)
In this task, participants were required to narrate the identical storey that they had been expod to in ELT but in their own words. It was hypothesized by the authors that that using the identical pictures could minimize cognitive load during performance and rai the possibility of language complexity. Participants, furthermore, were required to think that their audiences were children. The authors hypothesized that this technique can reduce the likelihood of reliance on hearer knowledge and hence the definite article was excessively ud. Grammaticality Judgement Task (GJT)
Participants in this task were asked to grammatically judge the underlined portions of ntences. The
judgement scale was created as a confidence measure requiring coding such as: 1) correct, 2) probably correct,
3) probably incorrect, and 4) incorrect. Although such a task allows a participant to process the ntence for its form, time-constraint is hypothesized to stimulate a participant to access implicit knowledge. That is, the likelihood of the participant re-examining and monitoring the respon is heavily reduced and that of intuitive linguistic judgement is raid, indicating a high degree of automaticity of the implicit knowledge.
Metalinguistic Knowledge Task (MKT)
Participants were required to correct 10 ntences and each ntence included an article error that was underlined (N = 10). Next, participants were required to give written explanations for the ungrammatical articles (N = 5). Participants were provided unlimited time to complete the task, and conducted two practice items prior to commencing. Respons were scored as either correct (1 point) or incorrect (0 points).
As has already been described, the participants of experimental and control groups were expod to each measure three times (pretest, posttest, and delayed test). To find out if the explicit instruction e
xerted an influence on implicit and explicit knowledge, the mean of the obrved data in each measure was calculated to explore the extent to which the mean of (ex. EIT) in the pretest was statistically different from the posttest and delayed test and different between the two groups. A statistical test in this ca should be employed to test the strength of mean differences between the three testing stages within one group and between the two groups. One-way ANOVA, thus, was the appropriate technique to employ in this ca.
The study provided a detailed and complex description of statistical data; for instance, the obrved data of each measure in each group (experimental and control) was divided into four ctions: non-generic articles,
All Rights Rerved.
EV ALUATING THE EFFECTIVENESS OF EXPLICIT INSTRUCTION 45
generic articles, grammatical articles, and ungrammatical articles. Descriptive statistics were also calculated for each ction. The author has sought to reorganize part of the obrved data of one measure in a way that gives a patent reprentative sample of the results.
Table 1
Elicited Imitation Task (Sample)
20 articles
Experimental group Control group
M SD N D M SD N
Pretest 10.29 3.202 49 -0.20 10.96 3.567 45
Posttest 13.80 2.993 49 0.79 11.14 3.690 45
Delayed test 14.98 3.058 49 1.37 10.64 3.276 45
The output prented above shows that the scores generally incread over both posttest and delayed test in the Experimental Group, which outperformed the control group. ANOVA was computed to test the mean
differences of the three testing stages of the elicited imitation task between the two groups. A statistical
difference was found in the posttest (F(1,92) = 14.866, p = 0.000) with an increa in the delayed test (F(1,92)
= 44.023, p = 0.000). The results elucidated, bad on the measure “EIT”, that implicit knowledge can be
promoted by explicit instruction.
ANOVA was also computed for the other tasks and revealed that there is no significant difference between groups in the posttest of OPT (F(1,92) = 1.609, p = 0.208), but there was a significant difference in the delayed
task (F(1,92) = 5.161, p = 0.025). It also offered no significant difference between the groups in the posttest of
GJT (F(1,92) = 3.496, p = 0.065). However, it revealed a significant difference in the delayed test (F(1,92) =
4.457, p = 0.037). Finally, ANOVA showed a significant difference between the groups in the posttest of MKT
(F(1,92) = 28.787, p = 0.000). This was sustained in the delayed test (F(1,92) = 27.344, p = 0.000). (In
statistics, a p-value can never be exactly zero, but the zero here was reported bad on SPSS output.) The overall findings suggest that implicit and explicit knowledge can be developed as a result of explicit instruction. In addition, the study demonstrated that the measures of implicit and explicit grammatical
knowledge can be reasonably parate. The two measures of implicit knowledge required time constraints and a
focus placed on the meaning. The other measures of explicit knowledge did not entail time pressure and a focus
placed on form (greater discussion will be prented under “critique of the study”).
Critique of the Study
As suggested in the introduction, the author is going to prent a conci critical discussion of measures of implicit knowledge and explicit instruction. The former was lected becau one of the
biggest challenges in
psycholinguistics-bad rearch is how best to asss the spontaneous level of acquired language (implicit
knowledge). The latter was chon becau it is the independent variable upon which the change in the
dependent variables (explicit and implicit knowledge) occurs.
Measuring implicit knowledge entails more cautious treatment to ensure that it accurately asss the unconscious status of specific acquired structures, unlike explicit knowledge. Historically, oral production task
has been employed to measure implicit knowledge, and although it supplies an amount of natural speech, it
lacks the accurate elicitation of the spontaneous u of a specific language structure (Ellis, 2009). Put
differently, it might give the learner a chance to u his/her explicit knowledge (to plan and monitor their All Rights Rerved.
EV ALUATING THE EFFECTIVENESS OF EXPLICIT INSTRUCTION
46 respons) rather than to test the spontaneous, contextualized u of a specific structure. This appears in Akakura’s study, where the oral production task reflected the difficulty of generating the specific articles. Thus, we find that the impact of explicit instruction was evident only in the delayed test, although the effect was suppod to appear in the posttest.
The elicited imitation test task has been later developed (more description of this task is provided under The Summary of the Study), but some threats have appeared that might potentially affect the validity level of the EIT. Studies have revealed that an L2 learner’s attention could be turned to the form of the ntence rather than the meaning. Other studies have also elucidated that L2 learners imitate the stimulus statement by rote (Erlam, Loewen, & Philp, 2009); they repeat the statement verbatim without understanding the stimulus ntence. Akakura’s (2012) rearch study could achieve a pioneering success in enhancing the control of the two limitations to provide a higher level of validity. The picture plausibility is employed in the task to make a participant’s attention focus
on meaning rather than on form. In addition, it provides a chance for delaying repetition so that participants do not repeat the ntence verbatim. In Rebuschat and William’s (2012) study, mi-artificial grammar is ud and participants are required to listen to statements on an item-by-item basis, to judge the plausibility of the mantics of the stimulus statement, and then to repeat the statement. In Erlam, Loewen, and Philp’s (2009) study, the statements are designed to enable the subjects to decide whether they agree with, disagree with, or do not comprehend a statement. However, a storey-bad elicited imitation test had not been previously employed to measure implicit knowledge, and it is considered, to my knowledge, that its first u was in the Akakura (2012) study.
When a deeper scrutiny is applied to measures of implicit knowledge, we explore that the EIT includes a choice that can increa the level of internal validity, such as the choice not sure (if the ntence fits with the
picture). This is becau, if some participants guesd the option correctly more by luck than by judgement, it is
expected that they would not correctly guess during the posttest or delayed-test pha, and then that they would obtain a low score. This threat, statistically, implies that the scores in the distribution regr
ess to the mean as a result of guesswork, not of the explicit instruction itlf (Gravetter & Forzano, 2011).
The oral production test, as indicated above, might fail to bolster the rigour of the measure of implicit knowledge, and, additionally, it could be influenced by the tutors’ personal prejudices, resulting in a low level of reliability. For instance, some tutors might lo control or their confidence when they are asssing enormous amounts of free natural speech, and, accordingly, they might be inclined to give participants scores in the middle range to ward off vere errors (Morgan, Dunn, Parry, & O’Reilly, 2003). The study further failed to provide a patent description of how the oral production test is achieved and how the data are gathered in a numerical pattern.
However, at the treatment stage of the study procedure, the study does not exactly elucidate the role of the rearcher—for example, regarding who has taught the participants, the rearcher himlf or another hired tutor. In addition, the treatment stage is confined to only one learning condition. Other learning conditions of explicit instruction are not addresd in the study. Snobul and Schmitt (2013), for instance, employed three learning conditions—enriched input, enhanced input, and decontextualized input—to evaluate under which conditions both adult native speakers and advanced non-native speakers of English acquire collocations. Tagarelli, Mota and Rebuschat (2015)
ud two conditions: implicit and explicit input. In the implicit learning condition, subjects were aware of neither the underlying goal of the experiment nor the target knowledge that would be learned or tested . The explicit learning condition is similar to the condition employed in Akakura’s
All Rights Rerved.
EV ALUATING THE EFFECTIVENESS OF EXPLICIT INSTRUCTION 47
(2012) study, where participants were aware of what knowledge they would acquire. The author considers that
the Akakura (2012) study could be enhanced if more learning conditions were included to determine which
implicit knowledge and explicit knowledge of English articles might best be developed.
In summary, the study succeeded in bolstering control of the limitations of EIT, in clearly describing the instruction process and in employing the appropriate statistical test (ANOVA). Nonetheless, some threats in the
study need to be reduced by a more careful treatment related to validity and reliability, such as guesswork and
an oral production task. Some further directions have been suggested to improve the treatment stage, such as
employing more than one learning condition in the treatment stage.
Finally, the method ud has many crucial implications, and the most promising one appears to be pedagogy. For instance, when an L2 learner obtains a high score in grammar, this does not imply that the learner
has externalized the target rule and thus can u it spontaneously and in unplanned language u. Rather, it shows
how well a learner might apply the rule in a context in which clo analysis of text is involved. Therefore, policy
makers in education should be aware of the learning conditions that enhance not only explicit knowledge but also
implicit knowledge, which is considered the chief aim of language learning.
References
Akakura, M. (2012). Evaluating the effectiveness of explicit instruction on implicit and explicit L2 knowledge. Language Teaching Rearch, 16(1), 9-37.
Ellis, R. (2009). Implicit and explicit knowledge in cond language learning, testing and teaching. London: Multilingual Matter.
Erlam, R., Loewen, S., & Philp, J. (2009). Form-focud instruction and the acquisition of implicit and explicit knowledge. In R.
Ellis (Ed.), Implicit and explicit knowledge in cond language learning, testing and teaching (pp. 237-261). Bristol:
Multilingual Matters.
Gravetter, F., & Forzano, L. A. (2011). Rearch methods for the behavioral sciences. New York, NY: Cengage Learning.
Morgan, C., Dunn, L., Parry, S., & O’Reilly, M. (2003). The student asssment handbook: New directions in traditional and online asssment. New York, NY: Routledge.
Rebuschat, P., & Williams, J. N. (2012). Implicit and explicit knowledge in cond language acquisition. Applied Psycholinguistics, 33(4), 829-885.
Sonbul, S., & Schmitt, N. (2013). Explicit and implicit lexical knowledge: Acquisition of collocations under different input conditions. Language Learning, 63(1), 121-159.
Tagarelli, K. M., Mota, M. B., & Rebuschat, P. (2015). Working memory, learning conditions and the acquisition of L2 syntax. In Z. E. Wen, M. B. Mota, & A. McNeill (Eds.), Working Memory in Second Language Acquisition and Processing (pp.
224-247). Bristol, UK: Multilingual Matters.
All Rights Rerved.
