Researchers have long sought to determine which of writing or reading tasks best facilitate acquisition of new words in a second language (L2). Studies suggest a higher efficiency of writing tasks (Hulstijn & Laufer, 2001; Laufer, 2003) while others (Barcroft, 2004) point to reading tasks as being more efficient teaching activities. A more recent study (Webb, 2005) helped clear the picture, but provided further data that do not seem to be best explained by the TOPRA model of vocabulary acquisition (Barcroft, 2002). I present here a new alternate theoretical construct of L2 vocabulary acquisition based on the type of processing. The Concreteness and Learning Intent (CLI) model is assumed to better reconcile mitigated data in the reading vs. writing debate. It also makes suggestions for the use and design of teaching activities involving writing. The purpose of this study was to confront both models to verify which one best explains intentional acquisition of L2 concrete words, through a simple writing task for which the two hypotheses predict different outcomes.

Participants (N ≈ 80) were French-speaking Canadian university students learning Spanish as a L2 and were tested in spring 2006. Participants viewed a sequence of 12 new concrete nouns presented in random order, each with a picture provided as meaning-support. They were asked to write a sentence containing each word, and to try to learn the words. Following the presentation and writing task, the participants did either a free recall or a cued recall test. The TOPRA model predicts higher scores on the cued recall test, while the CLI model predicts similar scores. Analyses of recall scores are presented. Implications for curriculum design are discussed in light of the debate over the relative efficiency of writing vs. reading tasks.