Mikey Does

Procedural Knowledge

Procedural knowledge is the implicit, automatic, skill-based knowing of how to do something — as opposed to declarative knowledge (explicit, conscious, fact-based knowing of that something is the case). In language use, procedural knowledge is what native speakers and fluent L2 users draw on to produce grammatical, contextually appropriate speech in real time: the execution is automatic, fast, and largely unconscious. Procedural knowledge cannot be fully verbalized — you can do it, but you may not be able to explain how you do it. Developing procedural knowledge from initially declarative language knowledge is one of the most important — and most practically difficult — challenges in SLA.

In-Depth Explanation

Declarative vs. procedural: the practical gap:

A learner who has studied Japanese for one year may be able to state: “The て-form of a verb is formed by replacing the final syllable of the dictionary form according to the て-form rules, and the te-iru construction indicates ongoing or resultant state.” This is declarative knowledge.

The same learner, asked in real-time conversation “何をしていましたか？” (nani o shite imashita ka? — “What were you doing?”), may stutter, lose their train of thought, or produce an ungrammatical response — because the proceduralized ability to rapidly produce て-form constructions in conversational conditions is not yet present.

The declarative-to-procedural gap is the reason that “knowing the grammar” and “speaking fluently” describe very different competences.

ACT-R and proceduralization:

John Anderson’s ACT-R cognitive architecture (1982, 1983) provides the cognitive model for proceduralization:

  1. Declarative encoding: A skill is initially represented as explicit, conscious declarative knowledge (“first do X, then Y”).
  2. Knowledge compilation: With guided practice, declarative representations are compiled into production rules — condition-action pairs that fire automatically given the appropriate context.
  3. Procedural tuning: Compiled procedural rules are refined through practice — strengthened when they lead to correct performance, weakened otherwise.

The result of this process is automatized procedural performance: fast, accurate, low-effort execution that no longer requires conscious deliberation.

Procedural memory in the brain:

Michael Ullman’s Declarative/Procedural (DP) model of language (2001) proposes that:

  • Lexical knowledge (words and their forms) is stored in declarative memory (mediated by the hippocampus and temporal lobes)
  • Grammatical computation is implemented by procedural memory (the basal ganglia and frontal-parietal circuits)

For native speakers, both systems function smoothly in parallel. For late L2 learners, who cannot rely on the same procedural systems for the L2, there is greater dependence on declarative memory — which is why L2 performance may feel more effortful and may use different neural resources than L1 performance.

How procedural knowledge develops:

SLA research suggests several conditions facilitate proceduralization:

  1. Meaningful practice: Using language in communicative contexts (not just drilling forms in isolation)
  2. Input volume: Extensive exposure to the language in context (immersion) provides the pattern frequency that drives proceduralization
  3. Output practice: Producing language under time pressure forces procedural execution rather than deliberate rule application
  4. Feedback: Error feedback in meaningful contexts helps correct and refine procedural rules
  5. Time: Proceduralization is slow; automatization of complex grammatical patterns takes hundreds of hours of meaningful use

Procedural knowledge and fluency:

What we recognize as fluency — smooth, quick, accurate production without pauses, reformulations, or conscious rule consulting — is the product of highly developed procedural knowledge. Fluency is not simply speaking fast; it is speaking with low processing cost. This is why fluency is an outcome of extended meaningful practice, not a target that can be directly instructed.

History

The declarative/procedural distinction in cognitive science originates with Gilbert Ryle’s (1949) distinction between “knowing that” and “knowing how.” Computational formalization came through Anderson’s ACT-R framework (1982, 1983). DeKeyser’s Skill Acquisition Theory applied this directly to SLA (1997 onward), arguing that L2 proceduralization follows the same cognitive architecture as other skill learning. Ullman’s neurobiological DP model (2001) provided neuroimaging evidence for the brain basis of the distinction.

Common Misconceptions

  • “Procedural knowledge = unconscious knowledge.” Partially true — procedural knowledge operates without conscious attention during execution. But it was initially built from consciously practiced declarative knowledge in many cases.
  • “Procedural knowledge in L2 is the same as in L1.” Late L2 learners show evidence of greater declarative memory involvement in L2 grammatical processing than native speakers show in L1 — proceduralization in L2 is incomplete and more effortful.
  • “Doing grammar drills develops procedural knowledge.” Drills may help if they involve meaningful use. Mechanical drills with isolated sentences develop narrow procedural patterns that may not transfer to conversational use.

Practical Application

  • To build procedural knowledge, move from studying rules to using them in meaningful contexts as quickly as possible. A conversation partner session where you must respond in real time forces procedural execution.
  • Sentence mining and speaking from memory (shadowing, output practice) contribute to proceduralization more than passive review.
  • Expect “performance anxiety” errors (procedure breaking down under stress/pressure) — this is normal and improves as proceduralization deepens.
  • High-volume immersion (listening and reading) is training procedural parsing of the L2 — every hour of comprehensible input is proceduralization practice for reception skills.

Related Terms

Sources

mikeydoes.com