Corpus Linguistics Case study 2

1
Corpus Linguistics Case study 2

• Grammatical studies based on morphemes or words.
  G Kennedy (1998) An introduction to corpus
  linguistics, London Longman, pp. 121-137

2
Introduction

• Survey of corpus-based studies of uses of verbs
  and verb forms, taken from various sources
• Besides considering the results, we should also
  consider what the parameters are for each search,
  and how we could express them
• Analysis of Brown and LOB corpora reveal that
  nearly 20 of words are verbs
• Apparently 224 different verb forms are possible

3
Ota (1963)

• Manual (pre-computers) analysis of small corpus
  (150k words), mainly transcribed speech
• 17,166 finite verb forms (977 of them passives)
• Table (next slide) shows distribution of tenses
• Looked at relative frequency of adverbs with
  different tenses
• today and this year more commonly found with past
  tense
• be most frequent (30), together with 6 other
  verbs account for 50 of all verbs

4

• notice predominance of simple present tense and
  simple past over other tenses
• simple past equal or more than present with said,
  went, did, came, made, took
• present progressive hardly used, except going,
  doing, coming

5
Ota (1963)

• Remember that Otas corpus is mostly spoken
  language
• Predominance of present over past also found in
  other studies
• Stative verbs (know, want) rarely found in
  progressive, and 10x more frequently with 1st
  person subject

6
Joos (1964)

• Also manual
• Study of 9100 finite and non-finite verb forms in
  a book, an account of a courtroom trial
• 23 most frequent forms listed on next slide
• These cover 95 of data
• Another 56 forms found in remaining 5 of data
• 145 of the 224 possible forms were not found at
  all

7
Again, 62 are simple present or simple past
8
Three studies

• Ota (1963) transcribed speech
• Joos (1964) written account of a verbal process
• George (1963) analysed 108,783 verb forms from
  a 0.5m word corpus of expository texts mostly
  from newspapers, nonfiction and references books
• Despite genre differences, findings are very
  similar

9

• Perfect and progressive less frequent than simple
  tenses in all cases,
• Present more frequent than past in these compound
  tenses in speech
• Vice versa in written English
• Results slightly biased by presence of be, which
  is rarely used with perfect or progressive aspect

10
Manual counts confirmed

• Early manual studies benefited from replication
  with computerized corpora, which confirmed the
  findings (eg with Brown corpus)
• Might make a good project for some of you
• Note how much distribution of tenses varies with
  genre

11
(No Transcript)
12
informative imaginative
13
Interlude

• How would you define the finite tense forms in a
  corpus search?
• Combination of verb form
• bare form (?infinitive), -s, -ed
• note that simple past and past participle differ
  with strong verbs, but not others
• Auxiliaries (have, be) participles (-ing, -ed)
• Various modals (if will and would indicate
  tenses, so do can and must)
• Active vs passive
• Infinitive with to non-finite tense forms
  difficult to distinguish
• Some verb groups are discontinous, eg should
  normally have been
• And many verbs form homograph pairs with nouns

14
George (1963)

• Finite vs non-finite verb forms

15
Francis Kucera (1982)

• Analysed syntactic and semantic functions of
  different verb forms in Brown corpus, eg

16
Modals

• Modals make up 7.6 of verb forms in Brown corpus
• Coates (1983) studied both distribution and use
• Major differences between spoken and written
  English

17
(No Transcript)
18
Modals

• epistemic use eg You must be exhausted, more
  frequent than root meaning in spoken corpus
• Major genre differences between root and
  epistemic use of some modals

19

• Main use with bare form or with passive
• shall, would rarely seen with passive
• can more usually with passive

20
Mindt (1995)

• Classified uses of modals, eg should
• Advisability/desirability (eg You should plant
  potatoes) 55
• Hypothetical event/result (eg I should have left)
  36
• Politeness (eg I should like to thank you) 9
• Other non-modal meanings include tense
• Past time (I told him I should have gone)
  38
• Future time (This should be done soon)
  25
• Present time (I dont think we should wait)
  19
• Timeless (You should brush your teeth regularly)
  18

21
Voice active vs passive
22
Use of passive and genre

• Study also showed that agentless passive (80)
  much more usual than passive with by-agent

23
Other topics

• Verbs particles
• Use of subjunctive (if it were, important that
  he join us) more prevalent in AmE than BrE
  apparently
• Prepositions
• Frequency
• Immediate right collocates
• Use as prepobj markers (look at, wait for etc)
• Semantic function (locative/time vs other)
• Conjunctions (eg since, when, because)
• Frequency
• Function, esp. as indicative of genre
• More vs less (and fewer)