Introduction

P_Lex was designed as a way of assessing the vocabulary used by L2 learners when they produce texts. The program was designed to address two problems with the standard ways of evaluating productive vocabulary use in L2. Malvern and Richards' D measure, which we described in Chapter 2, is good at describing lexical diversity, but it does not tackle lexical sophistication, and we felt that this was an important aspect of development in second language speakers. The second standard measure is Laufer and Nation's Lexical Frequency Profile (LFP; Laufer & Nation, 1995). LFP does address the issue of lexical sophistication, but it does so by describing the vocabulary in a text in terms of the frequency of the words the text contains. The Profile is a set of values, rather than a single value, however, and this makes it rather difficult to interpret, or to make comparisons between sets of texts whose profiles vary in detail. Additionally, both LFP and D need relatively long texts as input – 200–300 words is a safe minimum, but texts of this length are not easily generated by low-level learners, and this restricts the type of research that you can carry out with these measures. P_Lex was an attempt to overcome these problems. What we were aiming at was an assessment tool that could work on relatively short texts, and provide a single overall score that was easy to understand and transparent in its computation.

Briefly, P_Lex divides a text up into 10-word segments, and counts the number of 'difficult' or 'unusual' words that appear in each segment. The score it generates is the mean number of difficult words in each segment. P_Lex has the advantage that it can be used with relatively short texts, although the scores it produces become more reliable with longer texts.