Green and Hecht (1992) devised a test in which 300 German learners of English and 50 native speakers of English were shown twelve ungrammatical sentences. They were asked to correct the errors and provide explanations or rules for these ungrammatical sentences. These errors in the sentences were chosen because the grammatical concepts had been included in the syllabus and they were frequently encountered in learners’ communicative tasks. The results showed that only fewer than half of the participants could provide a correct rule while they could achieve 78 percent of the error corrections. Besides, for the connection of correct rules with
20
correct error corrections, it seemed that correct rules could lead to correct error corrections. That is, if students had a correct rule available, they could make an accurate correction. However, learners were still able to make accurate corrections even if they learned incorrect rules (corrections 70 % correct) or when they knew no rule at all (corrections 55% correct). Green and Hecht concluded that a complex relationship seemed to exist between explicit rules and error corrections. Students were likely to correct the errors with implicit knowledge, which then prompted the use of explicit knowledge. In other words, the participants’ ability to use implicit knowledge (i.e. error correction) appeared to surpass their ability to use explicit knowledge (rule provision) and their explicit knowledge of the rules was only included as a subsection of implicit knowledge.
Macrory and Stone (2000) conducted a project in a comprehensive school in the north-west of England. Ten students participated in the study and none of them showed particular talent or difficulty in learning French. The researchers investigated the learners’ own perceptions of their knowledge about the perfect tense in French by asking the students to recall what they knew about the rule of the perfect tense and in what situation they would use it. Moreover, the researchers assessed the participants’
actual knowledge of forms within the perfect tense by devising a gap-filling test in an oral and a written form to elicit their auxiliary use of participle. Finally, a semi-structured interview was conducted to see if the participants were able to use the structure in speaking and writing. The findings revealed that most participants were able to state clearly their knowledge about the perfect tense and all reported that they would use the structure when referring to actions in the past. On the gap-filling tests, they generally gave an auxiliary verb and a past participle as answers but when required to use the perfect tense in speaking and writing, they did not supply any auxiliary. Macrory and Stone thus indicated that there was a weak relationship
21
between the participants’ use and knowledge of the perfect tense in French. The learners could provide a clear understanding of the structure but they were less able to make use of the structure in spontaneous language production especially when referring to a third person. The study displayed a discrepancy between knowledge, which was considered to be related to explicit knowledge, and spontaneous language use, which was a representation of implicit knowledge.
In the study of Hu (2002), he involved 64 Chinese learners of English at a university in Singapore in order to examine the extent to which metalinguistic knowledge could be employed in spontaneous writing tasks. The participants were required to complete a spontaneous writing task by writing a narrative and an argumentative in the first round. After the writing task, they were given a rule-verbalization task and an untimed error-correction task to raise their consciousness of the focused six structures in the study. Then the second round spontaneous writing task was undertaken, followed by another error-correction task with time pressure. Finally, the participants were provided with a prototypicality judgment task in which six pairs of structures were listed for them to judge which one was more prototypical. Hu discovered that when the learners were equipped with correct metalinguistic knowledge, their accuracy with prototypical structures increased and their prototypical uses of the structures were less susceptible to differences in their attention to form. Besides, the participants’ accuracy with the six structures increased in the second spontaneous writing task after the participants were drawn attention to form through the post-consciousness-raising activities, indicating their more use of and access to metalinguistic knowledge. More explicit knowledge appeared to result in better performance in spontaneous writing tasks.
The studies reviewed above were to some extent connected to each other in their design. Green & Hecht (1992) and Macrory & Stone (2000) investigated the
22
relationship between implicit and explicit knowledge while Hu (2002) explored the extent to which metalinguistic knowledge could contribute to writing tasks. Among the studies, explicit knowledge was mostly operationalized as the learners’ ability to explain rules or correct errors in sentences, and implicit knowledge was considered to emerge in spontaneous oral language production. Many other studies further discussed the issue that whether or not implicit knowledge could be converted into explicit knowledge or vice versa.
Holding the non-interface position, Paradis (1994) claimed that “explicit and implicit memory rely on different cerebral systems and are differentially involved during the acquisition/learning of a foreign language” (p.393). In his opinion, explicit knowledge could only be used consciously and accessed with deliberate attention. It could be used to check the accuracy of the utterances produced automatically by means of implicit knowledge but it was not likely to evolve into the automatic production. Explicit knowledge could only be employed when learners tried to make sentences slowly or judge the grammaticality of a sentence. Whereas some people argued that explicit knowledge could gradually become automatic through practice, Paradis (1994) negated the transformation between the two and compared the non-interface relationship to the phenomenon that once a caterpillar became a butterfly there was no caterpillar. He suggested that implicit and explicit knowledge
“have different contents, and hence one cannot become the other, or be converted to the other or be transferred to the other” (p.405). According to Paradis (2009), four reasons could account for the non-transformation between implicit and explicit knowledge: (1) They are two distinct sources of knowledge and metalinguistic knowledge remains the same even if implicit knowledge is acquired. (2) They have qualitatively different nature. (3) They rely on separate and anatomically different cerebral systems. (4) One cannot notice what becomes internalized. Instead of
23
practicing the rule repeatedly, learners could possibly practice the utterances where the rule was used to automaticize the process by focusing attention on the form to be practiced. It was the “frequency of occurrence” and then the “frequency of practice”
(Ellis, 2002) that was crucial in facilitating its “incorporation” (Paradis, 2009) into implicit knowledge.
Another theory of second language acquisition that has been influential and associated with the strong non-interface perspective was the one proposed by Stephen Krashen (1982). According to him, “acquisition” and “learning” are two ways for adults to learn a second language. Acquisition will take place when learners are exposed to samples of the second language and it resembles how learners acquired their first language without paying conscious attention to form. Therefore, acquisition was primarily connected with implicit knowledge while explicit knowledge was seen relevant to “learning,” which developed via conscious attention to form and rules.
For Krashen, acquisition was more prevalent and important since it was available in fluent communication and he asserted that learning could not transform into acquisition. Learning, instead, served as a monitor or an editor to correct the utterances produced by the acquisition system. Learners would use the monitor model only when they have sufficient time, when they pay attention to form, or when they know the rules.
Likewise, Bialystok (1994) maintained that explicit knowledge differed from implicit knowledge and it was not likely for explicit knowledge to become implicit.
Instead, it was the access to knowledge that changed and it was the easier access that resulted in more automatic processing and more fluent performance. Despite the change of access to knowledge, “the representation of the knowledge, in terms of clarity of structure, can never become less explicit” (p. 567). Bialystok’s position, therefore, was regarded as a weaker form of the noninterface position (Ellis, 2005).