Implicit learning is the process of acquiring knowledge about regularities in the environment without conscious awareness of what has been learned or even that learning has occurred. Arthur Reber's (1967) artificial grammar learning experiments launched the field: participants exposed to strings generated by a finite-state grammar (e.g., MTVRX, MTTVRX) subsequently classified new strings as grammatical or ungrammatical above chance, despite being unable to articulate the rules of the grammar.
Key Paradigms
Three paradigms dominate implicit learning research. In artificial grammar learning, participants classify strings after exposure to grammatical exemplars. In the serial reaction time (SRT) task, participants respond to sequentially appearing targets, gradually speeding up when the sequence has a hidden pattern and slowing down when the pattern is disrupted. In dynamic systems control tasks, participants learn to control a complex system (like managing a simulated sugar factory) by interacting with it, acquiring knowledge of the input-output relationships without being able to state the governing rules.
Characteristics
Implicit learning has several distinctive properties: it is incidental (occurs without intention to learn), produces knowledge that is difficult to verbalize, is relatively robust across age and IQ (unlike explicit learning, which varies substantially with these factors), and is relatively preserved in amnesia and various clinical conditions. These properties suggest that implicit learning depends on a different cognitive and neural system than explicit, declarative learning.
Whether implicit and explicit learning are truly distinct systems, or whether "implicit learning" is really just explicit learning of small fragments, remains debated. Proponents of the dual-system view point to the different properties and neural substrates of the two forms. Critics argue that careful measurement often reveals some conscious awareness of the learned regularities. The debate hinges partly on how to measure awareness — a notoriously difficult methodological challenge.