Explicit vs. implicit and declarative vs. procedural language learning

Bialystok (e.g. 1981b) develops a model of SLA based on two interacting types of mental representation of linguistic knowledge: implicit and explicit, corresponding to Krashen’s ‘acquired/learnt’ dichotomy. Implicit knowledge is “the unconscious knowledge of a much larger body of information that is the basis of automatic, spontaneous use of language” (Little 1994:103), which is intuitive, tacit, and proceduralised. It can be accessed without much conscious deliberation, but it is frequently unanalysed, and then—as in the case of most native speakers (NSs)—we can safely say that it is not available for conscious inspection and the speaker does not know what it is that s/he knows, or is unable to verbalise it. Linguistic forms and structures are internalised and used by the learner without him/her being able (nor, as a matter of fact, finding it necessary) to reflect upon their underlying rules. Implicit knowledge is the principal knowledge source drawn on in on-line communication where there is little planning time available. It comes in two types (Ellis 1994:355): rule-based knowledge (i.e. internalised, generalised grammatical principles) and formulaic knowledge (i.e. ready-made chunks).

Explicit knowledge is the conscious, cognitive knowledge of grammatical facts, which exists in an analysed form that we are able to articulate informally in our own words if the need arises, and may be the result of formal instruction. The learner discovers patterns and pigeonholes their underlying rules. This kind of knowledge is available to the learner as a conscious representation, hence the speaker knows what it is that s/he knows. The weakness of explicit knowledge is that it cannot be readily accessed in real-time interaction.

Bialystok suggests that implicit knowledge can be accumulated in duplo: (1) by ‘unconscious acquisition’, and (2) by automatising explicit knowledge through practice (to a point where its usage will “no longer diminish our pool of attentional resources”; Dakowska 2005:128). Thus, ‘acquired’ and ‘learnt’ knowledge can be related in terms of automaticity (McLaughlin et al. 1983:142). Bialystok’s revised (1994) theory, in contrast to her (1978) model, emphasises the explicit representation of linguistic knowledge: language-specific explicit knowledge is so from the very beginning, while implicit and universal linguistic and conceptual representations evolve from initial implicitness to growing explicitness (access to consciousness in the sense of ‘awareness’): this “increasing explicitness can almost serve as a definition of what we mean by ‘learning’” (1994:567). Han and Ellis (1998:6) recognise two more criteria that the explicit/implicit dichotomy entails: awareness and accessibility: implicit knowledge, held without awareness, is accessible without difficulty, while explicit knowledge, which is stored consciously, can only be accessed with controlled effort.

The explicit/implicit dichotomy receives some importance in this dissertation, I hence summarise the two notions in the form of a simple chart of two opposing categories on a single spectrum (see also e.g. de Bot (1992) for treating automatic and controlled processing in FL learners as a continuum, and Karmiloff-Smith’s pioneering (1986, 1992) model embellishing the dichotomy by positing several steps of increasing explicitness—or developing ‘representational redescription’—before consciousness is reached):

EXPLICIT KNOWLEDGE	IMPLICIT KNOWLEDGE
learnt conscious analysed metalingual overt controlled (processing) declarative	acquired subconscious, internalised unanalysed intuitive covert spontaneous, automatic typically procedural

Figure 1: Explicit and implicit knowledge

The terms in each column are not intended to be necessarily treated as 100% synonymous and interchangeable; they are only cover terms employed by different linguists and methodologists promiscue to denote the same or related concepts! This begs the valid question whether the terms ‘explicit’ and ‘implicit’ can be used interchangeably with ‘declarative’ (characterised by Anderson 1983 as “knowledge that,” consisting—in the case of language—of factual, content information about the L₂ that has not yet been integrated or automatised, but that is available to retrospection and theorising) and ‘procedural’, as they oftentimes are^[1]. Ellis prefers to draw a distinction between these knowledge types:

Table 2: Differences between explicit/implicit & declarative/procedural knowledge (after Ellis 1993a:94)^[2]

What is interesting for us in EFL settings is the question whether one can progress from Type A knowledge to Type D knowledge, and this is where most disagreement lies. According to Krashen’s non-interface position, Type D knowledge could only derive from Type C knowledge (yet, according to his (1989) skill-building hypothesis, rules are first learnt consciously to be gradually automatised through practice). In the weak interface position, Type A knowledge is allowed to turn into Type C knowledge as long as the learners are ready to accommodate it into their IL system. According to the proponents of the strong interface position, Type A knowledge can be converted into Type D knowledge through practice (Ellis 1993a). In this respect, Dakowska (2003:132) makes the following observation:

In typical human communication, automatic processing is not unconscious. It is subconscious in the sense of being the least attentional, and too fast for noticing – with decisions taking place in fractions of seconds.

McLaughlin et al. (1983) claim that “complex skills are learned and routinized (i.e. become automatic) only after earlier use of controlled processes.” Thus, the learner must go through a conscious stage of processing a grammar rule deeply and thoughtfully before being able to control grammatical structures automatically. Similarly, summarising recent psychological theory and research on the issue of consciousness in LL Schmidt (1990:149) concludes that “subliminal language learning is impossible, and intake is what learners consciously notice.” He bears out the view that a consciousness-raising process is necessary for adults to learn FL form, especially for redundant and communicatively less important features. This, coupled with a thorough review of relevant research findings, favours conscious grammar learning, to which we will pass in Chapter 2.

source: Paradowski, Michał B. (2007) Exploring the L₁/L₂ Interface. A Study of Polish Advanced EFL Learners. Institute of English Studies, University of Warsaw, pp. 11–13.