Then, the texts obtained, which contained 196,245 tokens, were stored in raw text format and processed with Wordsmith 5.0 (Scott 2008) resulting into a type list (8) of 14,686 items that could be analysed and compared with the ones based on BLaRC, our legal corpus (the corpus obtained by scanning the three textbooks will be referred to as LeG-TeXT henceforth).
After extracting and validating the STWs in LeG-TeXT applying Drouin's (2003) ATR method (following the same steps as in the processing of BLaRC), it was attested that 67% of the SWTs identified were already present in the term lists obtained from our legal corpus using the same ATR method, a considerably high percentage taking into account the fact that the textbooks employed as reference deal with many different types of both private and public legal documents and topics apart from law reports.
LeG-TeXT was also analysed with Heatley and Nation's (1996) software Range adapting our term list (the SWTs identified in BLaRC) to become a base-word list used as reference by the software (instead of employing the ones provided by default with the software programme from GSL, AWL or BNC (9)) with the purpose of establishing the percentage of running words in LeG-TeXT covered by our list.
Even so, the specialised terms in BLaRC, which would fit into Nation's category of technical words (5% predicted text range), covered almost three times as many words as might be expected according to Nation (2001).
The overall number of coinciding terms and the percentage of text coverage provided by the term inventory extracted from BLaRC explains why we decided to employ it as a source to design a set of activities for the teaching of legal terminology.
Prior to the implementation of these corpus-based tasks, the students should also be instructed on the use of concordancers (11) so as to be able to access easily the information requested from the corpus used as support for the legal English class (BLaRC in this case).
Words like appeal, claim or law form other terms by derivation whose usage learners would have to attest through the search of concordance lines in BLaRC. These concordances would serve not only to confirm their guesses, but also to study the context of usage and meaning.
After doing so, they would be offered different examples extracted from BLaRC to fill in the gaps with the appropriate prepositions to guarantee the validity of their observations.
Table 1 illustrates the most frequent senses of the sub-technical terms party and offence selected from amongst the ones defined in the OED (2002) and the concordances obtained from BLaRC, our legal corpus, and LACELL, the general one.
Further below, table 3 presents some of the concordances for these compound terms found in BLaRC.