Search

Building a corpus entails the principled collection of a dataset, and corpora designed for general purposes commonly require the submission of that data to an annotation process whereby each item is ‘tagged’ according to its part of speech (POS). In some cases, a ready-made tag-set is applied to the data, and in other cases a bespoke tag-set is required. Corpora require a host infrastructure; building or sourcing this is the other essential element of corpus design. The creation of these components, along with a semantic tagger (to mark up the data for meaning as opposed to part of speech) and its own tag-set, plus the bespoke pedagogic toolkit (Y Tiwtiadur) constituted the CorCenCC construction plan. Decisions relating to the user-driven infrastructure and the collection and processing of data, in particular, present some specific challenges in the context of minoritised languages. In this chapter we outline how these challenges were addressed in the CorCenCC project.

As noted in Chap. 1.1, CorCenCC’s vision rested on two guiding principles: first, the language data captured within the corpus should be as representative as possible of the ways in which Welsh is currently used and encountered. Second, the corpus infrastructure should address the needs of the multiple community user groups operating in and/or engaging with the Welsh language. In this chapter we evaluate the extent to which these principles are represented in the project outcomes. First, we reflect on CorCenCC’s vision of representativeness, by considering the size and scope of the corpus, and the provenance of data in terms of language topic, context, genre and mode on the one hand, and contributor profiles on the other. We then document some of the specific challenges, including unanticipated ones, encountered during the project, and the steps taken to mitigate their impact. Next, we consider the extent to which the project has identified, and addressed, the specific needs of user group communities. This necessitates a reflection on our engagement and communication with potential corpus user groups throughout the project. Finally, we summarise the technical tools and resources developed to support the realisation of the CorCenCC vision. Throughout the chapter we report on the dynamic decision-making regarding method and approach that shaped the project as it progressed.

This paper offers a solution to a significant problem for teachers and researchers of\ud language learning that confounds their interpretations and expectations of test data:\ud the apparent simplicity of tests of vocabulary knowledge masks the complexity of the\ud constructs they claim to measure. We first scrutinise task elements in two widely cited\ud productive vocabulary measures, Lex30 (Meara and Fitzpatrick, 2000) and the\ud Lexical Frequency Profile (LFP, Laufer and Nation, 1995), in order to gain a more\ud precise understanding of the relationship between test performance and learner\ud knowledge. Next, in three empirical studies (N = 80, 80, 100) we compare L2\ud learners’ performance on Lex30, as the static point of reference, with LFP and with\ud two new tests designed to investigate specific elements of the vocabulary test tasks.\ud Correlation analyses indicate systematic differences in the tests’ capacity to capture\ud information about the quality of learners’ word knowledge and the size of their\ud vocabulary resource. Using the findings from this empirical work, we formulate a\ud model of vocabulary ‘capture’ onto which test tasks can be mapped. We demonstrate\ud how capturing key elements of the relationship between test scores and lexical\ud competence can guide teachers and researchers in applying and interpreting\ud vocabulary tests.

What measurements should linguists use when comparing texts written by different writers? We report aspects of a systematic evaluation of 381 different language measures derived from 200 analytic tools, carried out during the pilot for a study exploring genetic contributions to language variation. The measures covered lexis, structure, meaning, and discourse features, and were evaluated with a focus on capturing numerically the qualitative features that linguists consider central to differentiating one text from another. We review principles for selecting analytic tools, and the choices faced by the researcher in processing and analysing data. We then identify and demonstrate five of the measures, which between them provide a useful profile of different linguistic features, and note correlations with psychometric measures taken for each writer. We conclude with some caveats regarding general issues of validity and some indications about potential links between our work and research into authorship attribution for forensic purposes

Adult second language learners often come across as proficient but not native-like, with a command of the words and grammar of the language, but not of its idiomaticity. Idiomaticity resides in selecting the most native-like turn of phrase from a larger set of ways in which a particular message might be grammatically expressed. This article investigates what happens when learners are pushed to use native-like phrases in conversation. In an experiment inspired by TALK (a computer-assisted speech aid for the disabled), L2 learners of English were provided with native-like ways of expressing specific messages for targeted conversations. Both TALK and the L2 study represent artificial approaches to conversational interaction, retaining characteristics of ‘normal’ conversation along with a range of additional features arising from the limitations of being forced to rely on prefabricated material. Comparative evaluations of the two types of interaction indicate that even entirely fixed formulaic language can be highly effective in conversation. However, the extreme conditions of language use explored here also expose persistent weaknesses inherent in relying too heavily on formulaic material. The study exemplifies the value of research at the interface of lexical and grammatical processing and identifies related implications for the language learner.

The volume contains articles focussing on the role of prefabricated items in language. It provides both a discussion of Sinclair's seminal work including unpublished pieces of his manuscripts as well as articles from linguists of various fields demonstrating the relevance of this topic in theoretical as well as more applied contexts.

This paper describes an easy-to-administer test of productive vocabulary. The test requires Ss to produce a set of word association responses to a small set of stimulus words. The stimulus words are chosen so that with native speaker Ss they typically generate a wide range of responses, and a high proportion of low frequency responses. The paper argues that a test of this sort might tap the extent of non-native speakers' productive vocabulary more effectively than some other tests in current use.