# Object of the gerund

The rise of the zero object of the gerund (the contentment of meeting you as opposed to meeting of you) shows another kind of regularity: the most important factor throughout its diffusion proves to be female gender. The first VARBRUL analysis in Table 9.4. looks at the process in its new-and- vigorous stage (at about 30 per cent) and indicates that the incoming variant is favoured slightly more by the Royal Court than by the other areas. Later on, regional differences continue to provide the same average range of variation. The register variable shows even less variation throughout the process. These findings were already in evidence in Chapter 8, where the object of the gerund was found to be more evenly distributed in regional terms than either the third-person -s or the short possessive determiners (the momentary dip shown by Court in mid-range being partly due to a number of older informants; for further discussion, see 8.4.1.).

The which ^ which

Our data catch only the tail end of the spread of the relative pronoun which. It exceeds the frequency of 60 per cent in the fifteenth century and is as good as completed in the first half of the sixteenth. All three factor groups turn out

Table 9.4. VARBRUL analysis of the object of the gerund variable (all data), period mean showing the relative frequency of zero forms. Min./Max. = minimum/maximum weight of a factor in a factor group, respectively

 Period/mean (%) Factor group Min. Max. Range Total N 1520-59 28% Region Gender Register .419 (L) .401 (m) .473 (nf) .575 (C) .599 (w) .527 (ff) .156 .198 .054 492 492 492 1560-99 42% Region Gender Register .405 (C) .353 (m) .489 (ff) .561 (N) .647 (w) .511 (nf) .156 .294 .022 652 652 652 1600-39 70% Region Gender Register .425 (N) .401 (m) .481 (ff) .562 (C) .599 (w) .519 (nf) .137 .198 .038 623 623 623

(Abbreviations: Region: L = London, C = Court, N = East Anglia + North; Gender: w = female, m = male; Register: ff = family and friends, nf = non-family)

Table 9.5. VARBRUL analysis of the the which/which variable (all data), period mean showing the relative frequency of which. Min./Max. = minimum/ maximum weight of a factor in a factor group, respectively

 Period/mean (%) Factor group Min. Max. Range Total N 1440-79 70% Region Gender Register .133 (L) .227 (w) .412 (nf) .856 (C) .773 (m) .588 (ff) .723 .546 .176 908 908 908 1520-59 92% Region Gender Register .444 (N) .332 (w) .426 (nf) .597 (C) .668 (m) .574 (ff) .153 .336 .148 1823 1823 1823

(Abbreviations: Region: L = London, C = Court, E = East Anglia, N = North; Gender: w = female, m = male; Register: ff = family and friends, nf = non-family)

to be relevant: the form is promoted by the Royal Court and East Anglia (0.856 and 0.761 in the first period, respectively), by men writing to their immediate family.

In the second period regional differences become less pronounced and gender now provides the widest range of variation. A systematic male advantage in both periods is interesting as women rarely wrote their own letters in the fifteenth century. Since a gender difference nevertheless emerges, it suggests that the Paston women's letters, for example, were probably taken down from dictation by their male scribes more or less verbatim (see also Figure 6.14). For both men and women, register differences play a less significant role at this advanced stage of the process.