Seeking and sought share no collocates with seek at all.
Even if we confine ourselves to the nominative singular forms of naine 'woman' and ema 'mother' as node words, we get the desired results for naine, but in the case of ema there is no automatic way of telling apart the collocates of the grammatically homonymous forms ema NomSg, ema GenSg, and ema PartSg.
In addition, the collocates from the 3:3 range have been weighted, so that the collocate that is closer to the node gets a higher value, relying on the principle that the farther a word is positioned from the node word the less regular is its relation to the latter.
As the collocates analysed occurred at the right of the node-word, Court takes the syntactic role of a subject in such instances, while verbs following after are used in the active voice.
After scrutinizing the concordance it turned out that the collocates extracted are embedded in identifiable colligational patterns.
According to surrounding clause elements, the extracted collocates could be grouped into two major grammatical categories: verbs taking that-complement clauses or transitive verbs.
A maximum of 20 collocates for each head-word is given: down to a cut-off point of 15 joint head-word/collocate occurrences.
I have studied a sample of 1,000 of the headwords with their associated collocates and concordance lines: I simply took every tenth word alphabetically.
I will talk of collocates co-occurring with a node within a span of word-forms to left and right.
The notional sets derived from collocates of the node word 'blood' (See Appendix A) yield lexical patterns that clearly connect with the theme of damaged identity and enhance the themes of the novel in general.
The notional sets of 'green' collocates (See Appendix C) similarly suggest qualities that are not readily apparent with syntagmatic reading.
The evidence is overwhelming in favour of this interpretation: forty-one collocates state or imply an inherent destructive force, whereas only three indicate positive development.
But in order to remedy the inequities that Olsen admits exist in the Tresor, we have sought the frequencies and collocates of those target words in a data base modified to include women's writing.
We next produced the men's corpus and submitted the target words to a statistical inquiry designed to reveal their collocates -- the words and phrases that characteristically tended to appear in their proximity within a sentence -- in the writing of both the women and the men.
Both the frequencies and the collocates in each set of samples bear out this tendency of women's writing, rather than men's, to reflect a preoccupation with the experience, quality, and definition of self (and of other women) within the domestic role.