Phrasal verbs
Consider the following questions:
- How often is the verb set used on its own (e.g. set an agenda for a meeting) and how often as a
phrasal verb (e.g. set up an appointment)?
- What are the most common phrasal verbs with set, and does it vary according to text type?
If you're interested in the behaviour of Phrasal verbs, you will need to look for a lexical
verb and an adverb particle, i.e. a word like up, in, on, about. In the CLAWS "C7" tagset, lexical verbs
are tagged VV... and adverbial particles are tagged RP.
But first, let's just look for set on its own.
- Find the [spoken - demog] section of the BNC Sampler corpus. CLEAR PREVIOUS if you used the written files
before. Now select all the demog files.
- Do a search for the different forms of set as a verb
set_VV*/sets_VVZ/setting_VVG
and make a note of the frequency.
Note the use of * to mean any "any letters which fit the space". So set_VV* includes all of
set_VV0, set_VVI, set_VVD, set_VVN.
(We could even put sets_VVZ and setting_VV* to save more typing)
- Now we are going to change the search to specify the particle, tagged RP. Because we do not know which words are
going to appear, it's best to specify this as *_RP (this means "any word that has the tag RP".)
You will need to type this in the Context window, then specify that it appears between "zero words to the
left" (0L) and "eight words to the right" (8R). We suggest eight because a word attached to its
POS-tag counts as 2 words in WordSmith - so really you are only looking for a particle 4 words to the right of
set.
- How many hits?
- Which are the main phrasal verbs forming with set?
Tip: click the Resort button , and set it to 2R.
- What proportion of the total number of occurrences of set as a verb is accounted for by set as a
phrasal verb?
- If you have time, look at whether set up is used in the same way in this section of the corpus as in
the written informative texts.
Optional further practice
Read through the tagset and use Concord to look up other tags and tag combinations. Check
out if the type of text makes a difference to the frequency. If you're not so interested in verbs, why not look for
adjectives, or comparative and superlative adjectives, e.g. the adjectives used to describe the words man and
woman, or the nouns which follow lovely or wonderful or miserable.
If you're comparing differently sized corpora, don't forget to normalize the scores of
each corpus.
|