Abstract
This study makes use of a large amount of 21st-century corpus material in order to examine a number of differences between the be-passive and the get-passive in contemporary American English. Previous studies on 20th-century material have indicated a decline in the use of the more frequent and more formal be-passive and an increase in the frequency of the rarer get-passive, especially in written American English. The present study is based on the Corpus of American Soap Operas; this corpus may be seen as an attempt at representing spoken language, which makes it a new departure point for the study of passive voice in English. Frequencies for the two constructions are tracked across the first decade of the century, and the number of true passives is estimated based on manually examined subsets. The differences between central be- and get-passives are discussed in terms of semantic preferences, colloquialization, and stylistic prescriptivism. Since prescriptive influence is assumed to be a factor in the use of passive voice, other indicators of prescriptivism (relativizer which, pied-piping of wh-relativizers) are also examined.
Appendix A
Appendix B
Tables with raw frequencies which correspond to figures
Notes
1 ‘AMC’ stands for All My Children. A complete list of abbreviations can be found in Appendix A.
2 The fact that there is no expressed by-agent does not disqualify the construction from being categorized as a central passive; on the contrary, it fulfills the criteria that there is an active version of the sentence (I heard a rabbi was marrying you tonight, actually), and that agent addition is possible (I heard you were getting married by a rabbi tonight, actually), and it would not be possible to replace getting with feeling or seeming. Nor can the past participle be co-ordinated with an adjective or modified by quite or rather. Huddleston and Pullum (Citation2002: 1441) and Leech et al. (Citation2009: 156) only include get married as a central passive when a by-agent is present; however, it was felt that the same criteria should be applied to both auxiliaries for a fair comparison to be possible. Instances such as that in (15) were thus included as central passives.
3 See Tables 3a and 3b in Appendix B for figures. There were 449 be-passives in 2002 and 426 in 2010; there were 550 get-passives in 2002 and 587 in 2010.
4 Change in centrality of be-passives: Chi-square = 6.43; d.f. = 2; p = 0.040.Change in centrality of GET-passives: Chi-square = 3.24; d.f. = 2; p = 0.198.
5 Chi-square = 19.5; d.f. = 2; p < 0.001.
6 Chi-square = 4.09; d.f. = 2; p = 0.129.
7 See Appendix C, .
8 Note that who and whom are not the only wh-words with which pied-piping and stranding of prepositions are possible. Which and what are other examples (see Quirk et al Citation1985: 664, Huddleston and Pullum Citation2002: 58). However, it was felt that who and whom formed a neatly comparable pair, while which’s partner, that, cannot be used in a pied-piping construction. More exhaustive studies, such as Fukaya (Citation1993) or Hoffmann (Citation2014), are excellent sources of information on the placement of prepositions.
9 Unfortunately, the search terms inevitably return a great many instances where the preposition has a different complement than who or whom, such as She’s the one who’s going to go to prison. (BB, 2002). The figures are presented in Appendix C, .
10 See Appendix C, for the exact frequencies of interrogative and relative clauses with prepositions.
11 See Appendix C, for the frequencies for stranding vs. pied-piping.
12 An example is he got fired vs he was fired, where the reading of the version with got can only be verbal, but the reading of the version with was could be either verbal or adjectival.
13 The reason that the totals do not add up to 2000 in this table is that a small number of the be-passives in the data resisted classification for centrality and were left undetermined rather than being forced into a category.