Back to listing

Seminar: Could a language evolve triple articulation – the case of Nen number, Nick Evans, 23 Feb

Australian National University, Nicholas Evans, Shape

Date: 16 February 2018

Seminar: Could a language evolve triple articulation – the case of Nen number

Speaker: Nick Evans

Venue: Engma Room (5019), HC Coombs Building, ANU

When: 3.30pm-5.00pm, 23 Feb 2018


A key design feature of language is ‘double articulation’ – the arbitrary combining of meaningless segments to produce meaningful assemblages. For example, the phonemes /p/ and /b/, /k/ and /g/ in the English quadruplet pick / bic / pig /big mean nothing in themselves, and it is only when they are put together with other phonemes that they mean something. Double articulation is a powerful tool because of the extensive semiotic recombinations it then allows in producing linguistic signs.

Could a language then go a step further and evolve ‘triple articulation’ – in which it is now sequences of phonemes (something like what were traditionally called ‘morphemes’, but bled of their meaning) can be recombined in similar ways, only meaning something once they are assembled? While in the canonical case, words will follow a principle of compositionality, there has been an increasing realisation in many theories (e.g. Anderson 2005, Aronoff 1994 , Stump 2016) that complex paradigms force us to take the whole word as the unit of analysis without being able to assign clear meanings to each recurring string. However, while new models such as those supplied by these theories allow us to handle the informational mapping more adequately, they leave open the problem of how such non-compositional morphologies arise in the first place. Viewing them through the lense of emergent triple articulation helps us understand how morphological systems of this type emerge.

In this talk I examine the case of Nen, a Papuan language (Yam family) of Southern New Guinea which has developed a fourth value of grammatical number for verbal agreement with absolutive arguments, namely a ‘large plural’ on top of a more consistently marked three-way contrast (singular, dual, plural). Intriguingly, ‘large plurals’ are formed by half a dozen quite different methods, according to the type of verb involved, and it is not possible to motivate them by clear semantic principles. Rather, a number of different kinds of morphological deponency seem to be involved, collectively conspiring to create a new semantic target. All methods involve ‘distributed deponency’ in the sense that they combine morphomes from at least two sites and give at least one of them a ‘misleading’ semantic value.

While these developments amplify the system’s semantic expressivity by adding a fourth inflectional number value, in doing so they move away from a system in which morphological elements are always directly interpretable semantically. As such, Nen is an intriguing example of rapid paradigmaticisation leading to the creation of distributed deponency. In charting the limits of what can be a possible word,  ‘distributed deponency’ needs to be considered as a type of extended deponency in which strings at multiple sites are loosened from their base semantics, thereby allowing new combinations of material which were previously disallowed but logically possible if one ignores semantic constraints.  In this sense distributed deponency creates a sort of ‘triple articulation’ of the linguistic sign in which not only phonemes, but also strings of phonemes, can be combined without direct semantic links to the individual elements. This in turn helps understand one pathway by which extreme fusional systems may arise.

  • Australian Government
  • The University of Queensland
  • Australian National University
  • The University of Melbourne
  • Western Sydney University