Answers to exercises: Chapter Four practical activities
A4-1) Compare and contrast two recent grammars of English which claim to be based on corpus data.
This question has no single answer – your response to this activity will be very much dependent on which grammars you have compared.
Modern grammars contain many areas where they agree, but also areas where they at least generate different data if they do not disagree outright:
- Some of the difference is due to theoretical approach (for instance, compare Huddleston and Pullum 2002 on tense and aspect to Biber et al. 1999 on tense and aspect.)
- Some is due to the differences in data that they are using (for example, Biber et al. 1999 made more comprehensive use of a much greater range of corpus data than Quirk et al. 1984).
- Some of them are due to different explanations being forwarded for the same phenomenon (see for example the phenomena that are described as lexical bundles by Biber et al. 1999, but not considered separately from other outcomes of the language system by most earlier grammars).
This means that the grammars, while agreeing on a broadly common core, exhibit significant differences.
A4-2) How much of a problem does the grammatical structure of sentences with filled pauses cause for traditional descriptive grammar based on written English?
Typically grammars based on writing do not account for such features at all, or if they do so they dismiss them merely as hesitations or disfluencies. Grammars which embrace the differences between speech and writing are much more likely to identify a range of functions for such filled-pause vocalisations, claiming that they express functions that are, to some extent, part of the grammar of the language themselves.
Even if we do not wish to introduce the analysis of filled-pause vocalisations into the “grammar” proper, however, we will still have to take account of how they problematise other analyses. For instance, filled pauses that occur mid-clause often indicate an on-line revision of the structure of the clause – that is a shift from a clause with one kind of structure to a clause with another kind of structure. The overall utterance that results is likely to resist analysis in terms of traditional grammatical description, for example constituency parsing or dependency parsing.
A4-3) To what extent do modern English language teaching textbooks use corpus data?
As with A4-1, your answer here will be extrmeley dependent on which book you have looked at. But here are some general comments on what you might observe.
It is very likely that a recently-published textbook will be based, at least in part, on corpus data – especially any lexicographical elements, as corpus data is the sine qua non of modern learner dictionaries.
However, it is also very likely that the reader will be “insulated” from the actual corpus data itself. The account presented in the textbook will have been created by the authors with (at least some degree of) reference to a corpus; but in most cases only the end-result will be present in the actual book, not the corpus data itself.
This is based on the (probably correct) assumption that while being based on real language usage is a selling point, actual corpus data would be offputting to the market for these books, i.e. learners of English as a foreign language.
So, a textbook is quite likely to reference the use of corpus data as part of its introduction. It may go as far as to mention the corpus used on the front or back cover of the book. However, within the text, it is unlikely that you will find corpus examples cited directly. You are likely to find modified corpus examples at best in the text itself.