The purpose of the present paper is to review L1 word recognition research and discuss its implications for L2 vocabulary research. Word recognition research thus far has focused on the mechanism involved when we recognize visually presented words, and various word recognition models have been proposed. Classical studies suggested that, in word recognition processes, not only is the target word (e.g., clam) activated in the mental lexicon, but other orthographically similar words, called neighbours (e.g., calm, cram), are also activated.
Recent studies have investigated how this so-called parallel activation occurs. One key issue is how to reveal the degree of similarity of different types of neighbours, since the similarity between words affects the degree of parallel activation. For example, it is important to show whether calm is more similar to the word clam than cram. Three types of framework to explain the similarity of neighbours have been proposed: position specific coding, local context coding, and spatial coding. This paper reviews the theoretical predictions and empirical evidence of these coding schemes and observes that position specific coding has fundamental problems.
Word recognition research has thus far mainly focused on the internal mechanisms of word recognition processes, and few studies have investigated how the mechanisms are acquired. However, recent research has shed some light on the developmental issue. The present paper also reviews some hypotheses in terms of the development of word recognition system and discusses implications for L2 vocabulary research.