Language uses a distributed network. Grammar leans on procedural memory and vocabulary on declarative. Listening and speaking share circuits, and chunks are processed as units. Useful framing, but brain findings don’t license classroom prescriptions on their own.
Claims summarised
- No single “language centre”: a distributed network.
- Declarative/procedural model: vocabulary is declarative (temporal lobe); grammar is more procedural (basal ganglia, frontal). Ullman’s model, cited in a table but missing from the reference list.
- Complex syntax recruits more frontal/parietal regions.
- Listening and speaking share circuits (Hickok & Poeppel 2007).
- Formulaic chunks are processed holistically (Wray 2002; Van Lancker Sidtis 2004).
- Early L2 processing is effortful and becomes automatic; L1 stays active (Kroll et al. 2008; Abutalebi 2008).
- Prediction during comprehension (Federmeier & Kutas 1999; DeLong et al. 2005); emotional salience aids memory (Kensinger & Corkin 2003).
Classroom implications (as given)
Multisensory input (gesture, TPR, captioned video); stems and chunks recycled across topics; complex grammar introduced gradually; lots of active listening before speaking; spaced recall for vocabulary, structured repetition for grammar; routines that let learners predict.
Limits & critique
- Neuro-to-pedagogy leap: “shared circuits” doesn’t show that more listening improves speaking. That needs intervention studies. (Bruer’s “bridge too far” argument, 1997.) (Claude)
- Citation inconsistencies: Ullman 2001 and Friederici 2011 are cited but not listed. Check before citing in DELTA work.
- “Ballistic processing” is a non-standard label for prediction.
Apply in my context (Claude)
- Good material for a YouTube explainer if framed honestly: “what brain research suggests, and what it can’t tell us”. That’s an angle few ELT channels take.
- For DELTA, cite the behavioural SLA evidence (e.g. Ullman’s declarative/procedural model as a theory), not fMRI headlines.