On 11 October 2014, linguists descended on Kyushu University’s Hakozaki Campus for an afternoon of linguistic sustenance and animated discussions. By dint of ingenious planning by Michinori Shimoji, two events were rolled into one: a workshop of the intransivitizing morphology in Japanonic and an invited talk by Toshihide Nakayama (Tokyo University of Foreign Studies) on the notion of “units” in linguistic description and theory.
The first paper of the workshop was by Fumikazu Niimura (Morioka University) and looked at Kesen, a variety (or in Niimura’s view a language) spoken by around 80,000 people in coastal towns in the southeast of Iwate Prefecture, such as Ofunato and Rikuzentakata. Niimura argued that Kesen ar-intransitives can express three functions: 1) resultant state, 2) ability, 3) spontaneous action (see examples 1–3):
‘I took the car’
I sinkansen=sa girgiri nosatta
I Shinkansen=DAT just in time ride-PAST
‘I could just get the Shinkansen in time’
arugu-tumori-datta-noni tui-kuruma=sa nosatta
walk-intention-had-but ultimately car=DAT rode
‘I was planning to walk, but in the end I went by car’
The talk built on earlier research by the same author which investigated two verbs – agaru ‘to open’ and nosaru ‘ride’. The dataset was extended to include about a dozen verbs.
The second paper was by Yusuke Hiratsuka (Seinan Gakuin University, now Shigakukan University), who is investigating the traditional dialect of the Koshiki Islands, just off the west coast of Kagoshima Prefecture in southern Kyushu. Following an introduction to the dialect, Hiratsuka analysed systematically: 1) verbs that can be intransivitized and which permit the addition of the intransitivizer –ar; 2) verbs that can be intransivitized which have a consonant stem in which –ar can be attached; 3) verbs that can be intransivitized that have a vowel stem in which –ar cannot easily be analysed (see examples 4–6 respectively; English translations can not always capture Japanese shades of meaning):
‘(He) tied the string’ (transitive)
‘The string was tied in a knot’ (intransitive)
‘The rain stopped’ (transitive)
‘The rain stopped’ (intransitive)
‘The ice melted’ (transitive)
‘The ice melted’ (intransitive)
In addition, Hiratsuka made some comparisons with the Fukuoka dialect and concluded that both show the existence of a unproductive –ar suffix, and that this can have both an intransitivizing and non-intransivitizing function.
The third paper in the session, by Mika Sakai (then Kyushu University, now NINJAL), switched to the Kumamoto dialect, which is spoken by about two million or so residents in Kumamoto Prefecture. The talk started with a detailed sketch of the phonology and morphological categories in the dialect before moving on to the bedevilling –ar construction, which Sakai dissected with admirable clarity. At issue then were constructions of the now familiar type, as below in (7–8):
hot water=SBJ boil-ar-PST
‘The water was boiled’
futon=SBJ hang out-ar-RSL
‘The quilt is hung out to dry’
As pointed out in the presentation, the two examples must be translated as passive sentences in English, but the usage differs from a typical passive or normal intransitive sentence and is not found in standard Japanese. Through a detailed analysis of the 13 verbal roots in Kumamoto dialect that can employ the –ar form, Sakai argued that the –ar construction expresses a spontaneous action and that essentially its function is anti-causative.
Following a short break for refreshments, the assembled cohort of linguists sat down for an engaging talk by Toshihide Nakayama. At issue was the notion of “units” in linguistic description and theory, a topic that Nakayama has mulled over for many years, not least through his documentation and analysis of Nuuchahnulth, an endangered Wakashan language spoken off the coast of Vancouver Island in Canada. In particular, the polysynthetic nature of this language calls forth numerous theoretical questions of analysis.
To question the various units, such as “the word” or “the clause”, is to shovel down to the very foundations – the fundamental axioms – of linguistic description. However, before we get the spade out, it is worth recalling how, at the beginning of the last century, mathematicians and philosophers struggled to provide descriptions of the most basic and fundamental axioms of mathematics in purely logical terms. And if it was an enormous challenge for scholars such as Bertrand Russell to define 1 + 1 = 2 in purely logical terms, then linguists will inevitably struggle to define “word”, “clause”, “morpheme” and such like. Although these are bound to remain slippery concepts, they are useful and necessary for an efficient and systematic description of language even though they seem to defy precise logical description.
At any rate, while the overall the structure and set up of grammars must rely on various concepts of units, Nakayama argued that this basic building-blocks approach can be combined with a longer term goal for language description that focuses on usage. Further, he noted that a usage-based grammar would be more in line with Bybee’s (2006) view that grammar is “the cognitive organization of one’s experience with language”, which considers functions of use and frequency of use, the patterns of language, integrating the Firthian observation that words (indeed units) are known by the company they keep.
However, as Nakayama made clear, such a detailed description of grammar cannot be managed in even the most detailed of our current descriptive grammars, be they 500, 1000 or 5000 page tomes. The only way this can be adequately achieved is through the use of technology, the compilation of corpora and databases that are sufficiently large enough to enable computers to statistically evaluate a language’s grammatical computation. In other words, computers would be able to recognize such things as productive patterns and formulaic expressions as well as ongoing change among speakers in, for instance, different age groups or of different genders, and even prosodic aspects could be integrated too (though tagging would presumably be needed and this too would have to be done at some interpretive level and involve units – unless this can be achieved otherwise with Artificial Intelligence). All this still seems a long way off, especially for the minority languages that we wish to document now, but Takayama provided a clear and utterly compelling vision of what grammars must strive towards.
After a busy afternoon, a bowl of chankonabe washed down with beer was very much in order and brought an interesting afternoon of linguistics to a close.