CONVERGING CORPUS ANALYSES Sample Clauses

CONVERGING CORPUS ANALYSES. Although the rates of collective attraction in experiments 3 and 4 are consistent with the hypothesis that the difference between British and American speakers is primarily one of grammatical number categorization, they were low enough to make converging evidence desirable. To strengthen the conclu- sion, we examined more closely the corpus distributions of plural verbs used with the collectives from our sample. Figure 4 shows the results for the collectives that occurred at least once as the subjects of verbs in the British and American corpora. The graph indicates the percentage range of plural verbs that accompanied each collective when the collective headed the verb’s subject. The incidence of plural verb use is given in 10% intervals between 0 and 100%. As before, percentages were calculated from the number of unambiguous plural verbs relative to the total of unambiguous singular and unambiguous plural verbs. For British collective usage, the figure reveals a bimodal distribution. More than two-thirds of the collectives (nineteen) are either singular-dominant or plural-dominant, with the remainder in between. American usage, in contrast, is unsurprisingly singular- dominant for twenty of the collectives. There is little difference between the dialects in the numbers of collectives occupying the intermediate ranges where lexically inde- pendent, notionally controlled usage might be reflected. However, the specific words that occur are not the same. Only gang and majority appear in the midranges for both dialects. It might be arguedthat the distributions reflect some bias within our relatively small sample of collectives. Perhaps few of them are amenable to notional variation, either because of their semantics or their typical contexts of occurrence. To examine this, we also compared the distributions in the corpora for plural pronouns used with collective antecedents. Just as we did for the verb counts, we calculated the percentages of plural pronouns (out of all singulars and plurals) whose antecedents were unambiguously identifiable as one of the collectives in our sample. Figure 5 displays the results. Unlike the verbs, the uses of plural pronouns with collective antecedents are distributed rela- tively evenly across the full frequency range, for both British and American English. In terms of their pronominal anaphors, some collective antecedents are predominantly singular, some are predominantly plural, but the majority fall somewhere in between.