Keywords in Garside and Raven

Working on a bibliographic entry for Devil turn'd hermit makes one realize what an incredible effort Garside and Raven is. Even though they situate themselves within prior work and cataloguing of first-edition texts, the amount of sheer compilation, correction, and discovery involved in it makes me wonder how much time, exactly, was spent on this work. But that's not the point.

We are looking at a collection that is in some ways, then, fantastically broad, but also carefully defined in order to present a cohesive collection. Their years are 1770 - 1829. All novels within are published in the British Isles, situating themselves in a place. They limit themselves to what they term "prose fiction", allying themselves with certain generic conventions. These includetranslational works, but not including children's novels, "related works of non-fiction", "didactic fiction", and reprints and reissues. They also use popular reviews as sources of information for these texts.

A brief look at the Tables and Figures section of the 1770-1829 book gives a brief look at what questions Garside and Raven have already addressed through focusing on these particular times, places, and genres. "Publication of New Novels in Britain and Ireland 1770 - 1799", "Publication of Epistolary Fiction 1770 - 1799", "Republication rates before 1801 of Novels First Published 1770 - 1799", "Authorship of New Novels 1770 - 1799" (which includes male/female/anonymous/other author categorizations) indicates that we are spiritual siblings to the questions that Garside and Raven have already asked. Specific patterns are elucidated upon in both volumes of the text in their introductions. We get an interesting depiction of relationships between women and men, authors and their publishers, England to other regions (in terms of publication), and specific subgenres to the overall genre of "prose fiction".

This brief and token resummary already invites particular questions. For example, in their prominent use of keywords in titles to broadly grasp at some generic convention, they give us some pretty broad, but interesting fractions. In epistolary fiction, for example, they give us a sense of rising popularity and then decreasing popularity (1750 - 1760, 10%. 1770 - 1790, 20%, afterwards falling), but we don't know what they judge as "epistolary fiction". One sort of assumes that an epistolary fiction would be 100% letter, but that is probably rare. At any rate, the level of epistolarity in an epistolary novel is not addressed, and the precision to which they judge a novel as epistolary or not seems to be absolute (one column in Table 2. Publication of Epistolary Fiction, 1770 - 1790 is "Epistolary Titles as % of Total Year).

I have a moderate pet peeve in that they also don't offer tables for all of their conclusions, and some of their tabling suggests bias in evidence selection towards their overall argument. I looked on the internet database (or what I assumed to be their internet database), and the search abilities don't really seem to be sophisticated enough to generate really really subtle statistical comparisons that are based in the form of the novel, and the results are unspecific enough so that I wish that this didn't result in just an introduction, but more carefully parsed analysis. For an example, Table 2 in Vol. 2, Keywords in Titles, 1800 - 1829 traces how many times the word "Novel", "Romance", and "Tale/s" are used throughout those years.

"While sometimes these terms can seem virtually interchangeable, generally the term 'Novel' is found in titles which contain contemporary domestic matter, whereas 'Romance' more usually describes fiction with an historical, fantastical, or exotic orientation. 'Tale' is naturally used for shorter works, and in the plural for compilations, though its relative freedom from negative connotations also gave it an attraction as an alternative for Novel" (51, v.2)

This is moderately objectionable because in the previous volume, there is a different method of defining subgenres within the novel through title keywords based on "epistolary" (with "letters" as its keyword), "sentimental" (with "sentimental" as its keyword), and Gothic (which is presumably more internally recognizable). (31 - 34) This generic parsing based on the keywords of titles is blunt, but perhaps necessary. As they state in the second volume, "In a period where large numbers of new novels were published anonymously, titles naturally played a crucial part in signalling the specific generic character of indivdual works" (49). They don't, however, trace the appearance and frequency of all of the keywords used over all of the era surveyed. I refuse to believe that "novel", "romance", and "tale/s" failed to show up before 1790, for example, and the lack of stretching these keyword-based generic conventions across the entire time surveyed strikes me as bizarre. There is a problem that arises not only between the largely arbitrary distinctions of individual decades surveyed, but also between introductions between the two volumes of the bibliography.

Similarly, in their addressing of the prominence of the "Adventures of" and "Sentimental" types of tales in the first volume, we could ask (but they don't) what other types of keywords exist, as well as the frequency of these other keywords. Part of the problem, as they clearly already note, is that the differences between the "Adventures of", "Sentimental", "Romance", "Novel", "Tale/s", "letters of" works of prose fiction aren't necessarily going to be distinguishable.
They do note, "Even the basic labels of Novel etc. were expendable to give a more specific generic indication, an additional epithet sometimes overriding in importance the root term (as in Fashionable Novel/Fable/Tale). Seen in unison, various elements of a title might point more closely to distinctions within a broad mode, such as Gothic". Still, they are highly explicit with some of their data regarding this, and not all of it - is "sentimental" popular across the entire timespan, though the epistolary novel loses popularity before the 19th century?

Some basic questions also arise from their use of tables. When I read Table 2 in vol 2., "Keywords in Titles, 1800-1829", they give, separated by decade:

Use of the word Novel by frequency
Use of the word Novel by percentage
Use of the word Romance by frequency
Use of the word Romance by percentage
Use of the word Tale/s by frequency
Use of the word Tale/s by percentage

They do not include (but one could figure out easily) total novels published over 1800-1829, and it's probably someplace else in the text, but it would still be useful here. Furthermore, given the length of the titles and the commonality of the words, I suspect that "Novel", "Romance", and "Tale/s" can possibly appear more than once in one title, and they don't address that. If they don't appear more than once in the same title ever, that might actually be interesting. One might assume that a "novel" is not a "romance" is not a "tale", but that's not certain, and if it's true, then one can also logically assume that these are at least, in terms of titles, completely distinct works, which would be cool.

The internet, of course, helps with all of this, and one could run searches without too much difficulty to answer some of these questions. I feel like somewhere, someplace, someone has already figured out all of the questions I've asked, if at least because Garside and Raven make a gesture towards concluding something from the titles and seem to back off rather quickly because of the paratextual nature of titles.

"Any full study of generic movements within fiction at this time would soon encounter problems such as those arising from the generic instability of a number of the works under view. The following observations concerning trends are therefore necessarily tentative and provisional. While title-page information offers a useful initial indicator of mode, first-hand knowledge of texts is utilized whenever possible, and guidance is also sought from a number of previous studies of specific genres"

Still, Garside and Raven offer a wealth of statistical information about titles, but what they finally use and conclude from is a pretty narrow range among a fairly limited set of words in titles. In retrospect, this got overly specific, ranty, and didn't lead me anywhere, really, but it was fun and educational.