Elena Tarasheva, PhD New Bulgarian University
Jan 06, 2016
Elena Tarasheva, PhDNew Bulgarian University
Conclusions at last year’s BETA conference
Method and effect of the previous study2 corpora: students’ writing – film analysis; 1st &
2nd draft
Analysis of the message of Michael Palin’s Travel series about Eastern Europe
The first draft corpus revealed features of informal style, lack of cohesive devices through Key words; Concordancing revealed racial language.
The second draft showed improvement
Corpora and method for this studyStudents’ summaries of a text; 300 words
longThe text for the summary
Key words for the summaries are compared to the key words for the text
Concordancing reveals how each key word is used for the summaries and in the original
Word lists suggest common mistakes in the writing
The corpora13 student summaries1st year students at Sofia UniversityLanguage proficiency – C1Taught about cohesion, coherence, linking
devices, text flow4 hours of teaching per weekThe summary was set as a homework. The
work was submitted via the electronic platform Moodle
statisticstokens (running words) in text 2,667 types (distinct words) 729type/token ratio (TTR) 27.33standardised TTR basis 1,000mean word length (in characters) 4.99sentences 2,792 mean (in words) 18.99paragraphs 377 mean (in words) 100.93
Key wordsCalculated via a statistical procedureAn automatic function of The Word SmithCompares the frequency list of a task corpus
to the frequency list of a balanced corpus – in this case: the British National Corpus
The words with a frequency higher than in the balanced corpus tend to show the ‘about-ness’ of a text, specific language idiosyncrasies, particulars of the style etc.
The words with a frequency lower than in the balanced corpus show what the author avoided
Key words in the original
Key words in the summaries
RhetoricWeWritingRadicalWriterTowardOur ConservativeStudentCompositionHonestCoherenceLearns
RhetoricWritingConservativeWriterRadicalLiberalCoherenceTextCompositionExamplesAuthorDoesns
ConcordancingAll the cases when a word is used in the
corpus The word is highlightedThe context to the left and right is shownThe concordances can be arranged to criteria
pre-set by the researcher – L or R, -2, +3 etc.
The degree of formality
The possessive ‘s
The Use of linking devices
Linking devices
Linking devices
Clusters Automatically calculated by the Word SmithWords which occur togetherThe number of words can be pre-set – 3, for
this study, but it can be 4, 5, 6
Major key words in contextOriginal collocates Summary collocates
A missing key word
concordancingIn this case – arranged according to -L1Because the word is a noun and we are
interested in the attributes used with itIf it was a verb and we wanted to see what
objects follow, then we would arrange for +R1
Relative clauses
Relative clauses
Language Functions
Language functions
Negative keyness
original summary
Conclusions
Comparing the key word lists of a text and its summary reveal whether the gist of the text has been grasped
Concordances show differences in focus, suggests possible misunderstandings
Word lists allow assessing the acquisition of grammatical features, such as relative clauses, linking words etc.
High key-ness can reveal language functionsWord lists also reveal adherence to a style