The exact relationship between implicit sequence learning and developmental dyslexia continues to be a topic of debate, limiting potential advancements in educational interventions. This chapter delves into exploring this relationship by analysing recent studies that have investigated implicit learning using the artificial grammar learning paradigm (AGL) in children with dyslexia. The implications, drawbacks, and future possibilities identified in this context are discussed. Ultimately, valuable insights are extracted from the body of literature on implicit artificial grammar learning, suggesting that incorporating implicit learning methods in educational settings might hold significant and relevant promise for enhancing reading abilities. This questions the conventional notion that dyslexia interventions should solely rely on explicit instructional methods, and instead reinforces the concept of a cooperative strategy towards instructional approaches.
Top1. Implicit Artificial Grammar Learning Paradigm: An Overview
Implicit learning terminology was coined by Reber (1967). It pertains to the cognitive process of acquiring knowledge unintentional through systematic exposure to patterns in the environment, without a conscious realization of the acquired knowledge. Implicit learning encompasses various aspects of information acquisition, including implicit sequence learning, that involves acquiring unconscious knowledge of structural patterns or regularities in a sequence of stimuli. Indeed, in his seminal research work Reber observed that participants could successfully classify string sequences derived from an implicit artificial grammar learning (AGL) paradigm above chance levels. He suggested that implicit sequence learning is fundamental to natural language acquisition because language, like AGL, involves the exploitation of regularities and sequences.
Usually, the AGL paradigm involves two main phases: the acquisition and the grammaticality classification phase. In the acquisition phase, participants undertake a short-term memory task, in which they are presented with symbol sequences that are produces using an artificial grammar. Crucially, participants are unaware that the sequences they are exposed to adhere to the rules of a grammar. Following this phase, participants are made aware that the sequences they saw adhere to a complex set of rules. Subsequently, they are tasked with categorizing novel items as either grammatical or non-grammatical based on their instinctive response (gut feeling). In this phase there are two approaches to classification: explicit classification, known as grammaticality classification, where participants are explicitly informed about the grammar’s presence, and implicit classification, known as preference classification, where participants are not provided any information and must categorize new sequences based on personal preference. Implicitly acquired knowledge could be more effectively assessed through preference classification since participants without being aware of the fundamental process generating the outcomes (Folia et al., 2008). In both grammaticality and preference classification tasks, participants perform above chance level, displaying comparable behavioural and neuroimaging outcomes (Peterson et al., 2012). These findings indicate that participants possess knowledge of the underlying generative structure, and the classification performance likely relies on implicit acquisition mechanisms, as previously demonstrated in studies by Forkstam et al. (2006) and Petersson et al. (2004).
Furthermore, AGL can be manipulated in terms of associative chunk strength design (ACS), which is a measure of familiarity of the items presented during the acquisition phase in terms of bigrams and trigrams, compared to the items encountered in the classification phase (Forkstam et al., 2006). Sensitivity to ACS suggests a learning process that relies on statistical fragments. On the other hand, sensitivity to grammaticality status, regardless of ACS, is associated with an implicit structure-based acquisition mechanism. Consequently, reduced reliance on ACS indicates a higher level of implicit structure-based knowledge concerning the actual rules. Furthermore, AGL sequences can be created using various elements, including linguistic and non-linguistic stimuli. Moreover, AGL can be explored through diverse sensory modalities like visual, auditory, or tactile methods. This makes AGL an adaptable and versatile approach for examining various facets of the language learning process, including incidental learning from exemplars without feedback, implicit extraction of linguistic rules, and syntax learning.