Corpus

Corpus

1. Collection of texts

   • Large
   • Computer readable
   • Designed for linguistic analysis

2. Applications

   • Depend on
     • Desing of corpora
     • Observational methods of analysis
     • Interpretation of analysis
   • Translation studies
   • Stylistics
   • Forensic linguistics
   • Cultural representation & key words
   • Psycholinguistics
   • Theoretical linguistics

3. Modern corpora & software

   • Principles
     • Observer must not influence what is observed
     • Repeated events are significant
   • Available corpora
     • 1960s --> 1990 (First generation)
       • Small but carefully designed
     • Carefully designed REFERENCE corpora
   • Corpus design
     • Must be balanced
     • Must include consideranble data
       • Running words
       • Size of the audience for the text in corpus
     • Must combine
       • Large general corpora
       • Small corpora for specific knowledge
       • Opportunistic text collections
   • Some types of corpora (according to process)
     • Raw
     • Lemmatized
     • Annotated

4. Empirical linguistics

   • Computer technology is essential
   • Requires observation
     • No single method
     • Uses concordances
       • Concordance lines
       • Concordance data

5. New findings & descriptions

   • Word frequency
     • Varies according text-types
     • May have differences senses
     • Requires interpretation by material designers & teachers
     • AWL can be used as a guide
   • Phrase frequency
     • Determines word frequency
     • Phrase-like units
       • Basic units of meaning
   • Phrases
     • Collocations
       • 1, 2, or 3 words co-occurring frequently
     • Recurrent phrases
       • Frequent multi-word strings
         • Identified using computer programs
           • Identify patterns
   • Semantic prefence, discourse prosody, and extended lexical units
     • Collocation
     • Colligation
     • Semantic preference
     • Discourse prosody
     • Strength & attraction between nodes & collocates
     • Position of nodes & collocates
     • Distribution
   • Grammar, co-text, and text-types
     • Corpus can reveal characteristics
       • Type-token ratio
       • Lexical density
       • % of everyday & academic vocabulary