Introduction
See the short text on 'he' versus 'she' for background.
Experiment 1: the use of kinship terms such as 'father', 'mother', 'sister' and the use of male and female first names in the corpus per million words
A straightforward comparison to make across the span covered by the corpus, 1900-2019, is to compare the uses of common kinship terms and also, in this case, the first names of characters in the books mapped to their most likely gender.
Below are several plots that show the development in the frequency of use of 'mother', 'father', 'sister' and male and female first names.
As to the use of 'mother' and 'sister', female writers (green lines) use these words more frequently than than male writers (orange lines), while the use of 'father' seems more dependent on the individual books published and incorporated into the corpus in the various decades (Figures 1-3).
Figure 1. The use of 'mother' in texts written by female and male writers per million words
Figure 2. The use of 'father' in texts written by female and male writers per million words
Figure 3. The use of 'sister' in texts written by female and male writers per million words
The use of male and female first names, with the caveat that the library I use to guess gender by first names gets it right in most cases, are more evenly distributed until the 1990s when female writes start using both male and female names more often than male writers. (Figures 4 and 5). Again, there must be something about the books published and incorporated into the corpus in that decade that contribute to this sharp rise in the use of first names.
Figure 4. The use of male first names in texts written by female and male writers per million words
Figure 5. The use of female first names in texts written by male and female writers per million words
Experiment 2: the use of Mr, Mrs, Miss and Ms per million words
To see if the texts of my corpus are in line with what has been observed from studies of other corpora, I have also plotted the use of 'Mrs', 'Mr', 'Miss' and 'Ms' per million words. Paul Baker writes in an article in Quartz that "My research has showed that Sir, Mr, Miss, and Mrs have all plummeted in usage in written English, with “Mr” showing the most dramatic falls, as the graph below shows [see the graph in the article]. If the trends continue, in a few decades these terms will be practically obsolete."
So what about the use of these honorifics in the CBF.
Male and female writers use 'Mrs', 'Mr' and 'Miss' to more or less the same extent by 2010-2019 (Figures 6, 7 and 8). And, my data show the same sharp decline in the use of these titles as witnessed by Baker.
Figure 6. The use or Mrs in texts written by female and male writers per million words
Figure 7. The use or Mr in texts written by female and male writers per million words
Figure 8. The use of Miss in texts written by female and male writers per million words
What about the use of Ms? Paul Baker writes in the Quarts article "One term of address which has shown an increase is Ms but it was starting from nowhere, is still a rarity, and encounters resistance. I’m doubtful it will fully catch on any time soon."
This seems to be contradicted by my data for modern fiction. In my material the use of Ms per million words rises sharply from around 1970, especially in novels by female writers.
Figure 9. The use or Ms in texts written by female and male writers per million words
Although, Baker is right that its use starts from 'nowhere' and that its use by male writers may show the beginning of its decline.
Finally, here are two plots of the use of 'uncle' and 'aunt' respectively, including trendlines. It seems that writers of modern, British fiction have diminishing interests in portraying uncles and aunts.
Experiment 3: the use of kinship words according to genre per million words
Another way of plotting the frequency of use of these kinship terms is by genre. After the discussion of the aptness of the original genre labels used for the texts in the CBF (see Investigating genre labels of individual texts), the CBF texts are now re-labelled and divided into four (broad) genres: crime, general, historical and romance.
In what follows, we will plot the frequency of 'mother', ' father', 'sister', 'Mrs' and 'Ms' according genre, to see if we get a more nuanced picture of the use of these kinship terms than in the plots above where male and female gender of writers where the variable used.
Figure 10. The freq. of use of 'mother' by genre per one million words
In Figure 1, we saw that female writers used 'mother' much more than male writers. What this plot in Figure 10 shows, is that 'mother' is used quite a lot more in books labelled general than crime and also more than in historical novels, irrespective of the gender of the author.
Similar trends can be seen for 'father' and 'sister', i.e., that crime uses these kinship terms less that the other genres. Cf. Fig. 11 and 12.
Figure 11. The freq. of use of 'father' by genre per one million words
Figure 12. The freq. of use of 'sister' by genre per one million words
The recent rise in the use of 'sister' we saw in Figure 3 is even more pronounced in Figure 12, except for a small dip in the latest decade for historical.
What then about the use of e.g. 'Mrs' and 'Ms'.
As to the use of 'Mrs', the frequency of use by genre shows the same as when plotted according to gender. Perhaps a bit surprising is the infrequent use of 'Mrs' in historical novels, which is close to zero by 2010 (Fig. 13).
Figure 13. The freq. of use of 'Mrs' by genre per one million words
The steep rise in the use of 'Ms' we noticed in Figure 9 seems due, mostly, to the genres crime, general and romance and not historical.
Figure 14. The freq. of use of 'Ms' by genre per one million words
As a final warning and caveat when reading these graphs: the number of texts per genre per decade is very uneven. The genres historical and romance are thinly distributed across the corpus. As an example, the genre romance is represented by only 14 out 176 books in the decade that spans 2000 - 2009. This means that very frequent use of 'Ms' in only one or two books may be responsible for the spike we see in Figure 14 in that decade.