Acquiring productive collocation knowledge via retrieval activities: the effects of teaching method and incongruence type

The study investigated how the acquisition of productive knowledge of incongruent collocations achieved through output retrieval tasks is affected by two factors: (1) teaching method [non-contrastive (NCON): L2-only vs. contrastive (CON): L1–L2 contrastive analysis] and (2) collocation incongruence type (lexical vs. syntactic). Lexically incongruent collocations (LEXcs) are defined as collocations with a similar structure in L1 but a different collocate, e.g. “follow instructions” (“obey instructions” in Hebrew). Syntactically incongruent collocations (SYNcs) have different phrase structures in L1 and L2 and are translated by a syntactic paraphrase in L1, e.g. “ridiculously expensive” (expensive in a ridiculous way in Hebrew). Seventy-five Israeli undergraduate EFL learners at a B1–B2 CEFR proficiency level learned sixty collocations under four conditions: LEXcs/NCON, LEXcs/CON, SYNcs/NCON, and SYNcs/CON. Productive knowledge was measured through immediate and delayed form recall posttests. Results showed significant effects for both variables, with LEXcs and contrastive teaching yielding better outcomes. The findings are discussed in relation to L1–L2 awareness raising, the competition model, selective attention, pushed output hypothesis, involvement load hypothesis, and technique feature analysis.