it’s a comprehensive reference grammar, innit

Last Wednesday, Elizabeth Zeitoun’s recently published Grammar of Mantauran (Rukai) arrived in my mailbox at SOAS from Academic Sinica in Taipei. This is a beautifully produced description of a dialect of Rukai, one of the Endangered Languages of Taiwan and at 551 pages is a sizeable account of the language.
So I got to thinking: this is a pretty impressive comprehensive reference grammar of an endangered language. And then, well what counts as a ‘comprehensive (reference) grammar’? The term gets used quite a bit in relation to endangered and minority languages. For example the February 2007 newsletter [pdf] of La Trobe University’s Research Centre for Linguistic Typology, the most recent one available, contains over 25 uses of the term, and all 10 PhD students associated with the Centre are said to be writing a ‘comprehensive grammar’ of a small language. A Google search for “comprehensive reference grammar” returns 1,130 hits, and for “comprehensive grammar” 128,000 hits, though that includes things like A Comprehensive Grammar of the English Language by Randolph Quirk,Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik, which doesn’t really count for our purposes, nor does Matthews and Yip’s Cantonese: A Comprehensive Reference Grammar.
So I then adopted a tried and true sampling method of language typologists, namely have a look at the grammars of smaller languages that are on my book shelves at SOAS and pick the fattest ones (ok, ok, I know real typologists don’t do sampling like this any longer, but bear with me for the purposes of this exercise). What I came up with is summarised in the following table (astute readers will notice that I am not controlling for factors like margin width, page size, font type size and line spacing, but I’m only human):

Now if we count the number of pages in each
book that actually deal with grammatical description we get a slightly
different picture (some of the fatter ones are padded out with Sample
Texts; and Vocabulary at the back):

And the winner is, on both counts, Keren Rice’s Grammar of Slave (and what a deserving language it is too, mind-bendingly complex in its morphology and having plenty of phonology and syntax too). Jeffrey Heath’s grammar of Nunggubuyu comes in second (it’s another mind-bendingly complex language, but unfortunately the grammar is now out of print) – complemented by his 556 page separate volume of Ethnographic and Narrative Texts.
Of course simple quantitative measures like page counts don’t give us a real picture of how we might determine what comprehensive means in qualitative terms in relation to grammars, but I think the exercise is interesting and informative none-the-less. Perhaps it is time to get to work on parameters for a reliable qualitative approach to determining what we mean when we refer to a work as a “comprehensive (reference) grammar” (cf. my previous blog post about possible qualitative measures for text corpora).
Now, back to reading about Rukai.