Bowler’s talk at Tübingen

Posted on March 10, 2021 by



LEL’s Margit Bowler will be the guest at this week’s colloquium organised by the C1 project at the University of Tübingen. She will be presenting on

“Perspectives on cross-linguistic variation in degree constructions”.

It’s an open research colloquium, so anyone interested in hearing about Margit’s research (and Emily’s and Ryan’s and Andrew’s – they all have a finger in this pie!) is welcome to attend.

The talk will take place on Thursday, 11th March 021, 3-5 GMT. A Zoom link can be requested by email from either Margit or Vera. The abstract is below.

Beck, et al. 2009 propose the Degree Semantics Parameter (DSP), by which languages vary in the semantics of their property concept lexemes. Languages with a positive setting of the DSP (English, Thai, Hungarian) have grammatically explicit comparatives (John is taller than Bill), measure phrases (Mary is 6ft tall), and differential comparatives (Mary is 6 inches taller than Bill), among other degree constructions. Languages with a negative setting of the DSP (perhaps only Washo; Bochnak 2015) lack these constructions. However, recent work has uncovered a number of languages that display “mixed” behavior with respect to the DSP. For example, Nez Perce is argued by Deal & Hohaus 2019 to have a negative setting of the DSP, despite the fact that it has an explicit comparative morpheme like English -er/more. We argue that “mixed” languages like Nez Perce argue against the binary DSP analysis, and support an analysis of degreefulness as arising through functional morphology introducing degrees into the composition. Property concept words (tall, big) denote relations between individuals and states; degree semantics arise via a measure on these states, which is introduced by functional degree elements like -er/more, measure phrases (6ft), and so on.

The question then arises whether all languages with “mixed” degreeful behavior should be analysed along these lines. Are degreeless, Kleinian analyses (as in Bochnak 2015 for Washo) appropriate for any such languages? I argue that this is so for Warlpiri (Pama-Nyungan, Australia), based on original fieldwork data. Conjoined comparatives in Warlpiri (of the form John is tall, Bill is short) are felicitous in crisp judgment contexts, in which the objects being compared differ only a small amount with respect to the relevant property. This is unexpected, given a model of conjoined comparison as a conjunction of positive, vague predications. I show that Warlpiri property concept words are nouns, and as nouns, can be interpreted definitely when combined with a phonologically null iota operator. I argue that Warlpiri conjoined comparatives can be interpreted as conjunctions of definite property concept predicates (John is the tall one, Bill is the short one), which restrict the context to the two uniquely tall and short individuals. This predicts their acceptability in crisp judgment contexts (Kennedy 2007). The absence of any comparative morphology in Warlpiri conjoined comparatives argues against a degreeful analysis in which degrees are introduced by functional morphology.

Featured photo represents a Zoom talk experience.