Spring 2015 Colloquium

The University of Arizona

Department of Linguistics

Spring 2015

Linguistics Colloquium

Ling 495/595A

Coordinator and convener: Professor Adam Ussishkin (Associate Professor, Department of Linguistics)


Fridays 3:00-4:30 PM in Communication 311

Please see dates for Colloquia below.


Friday, January 23, 2015: Robert Henderson, Wayne State University

Note special time and location: this talk takes place on Friday, January 23 from 4-5 pm in Communication 311.

Title: Mayan positional roots at the syntax-semantics interface


This talk investigates an enigmatic root class in Mayan languages, called positional in the descriptive literature, and argues that these roots should receive a scalar semantics. Example (1) presents some instances of positional roots in Kaqchikel, while (2) shows a few of their canonical derivations. Note that they can be derived into stems of a variety of syntactic categories.

        (1) Positionals                       (2) Derived Positional

               a. ✓ch’eq ‘wet’                    a. ri ch’eq-ech’ïk che’ ‘the very wet tree’

               b. ✓sët ‘circular’                  b. Set-ël. ‘It’s circular.’

               c. ✓köt ‘twisted’                   c. X-kot-e’. ‘It twisted.’

               d. ✓jot ‘elevated’                 d. Xu-jot-ob’a. ‘He elevated it.’

Core Proposal: Positional roots denote measure functions of type〈e, d〉 (from individuals to degrees on a scale).

After mustering distributional arguments for a degree-based account of positional roots, I then expand the analysis along three routes. First, I show how a series of positional-specific morphological puzzles can be solved when positional derivations (like those above) are reanalyzed as degree morphology. Second, given the cross-categorial distribution of scalar items, I show how the analysis lets us understand why positionals are so category neutral: They lexicalize the scalar core underlying gradable predicates across categories. Finally, I consider how to integrate derived positionals into clause-level degree constructions like the comparative. All along the way there will be tension between giving positionals a scalar semantics and preventing them from collapsing on bona fide root adjectives, which pattern differently in a variety of ways. This will open up a way to think about different sources of gradability in natural language.


Friday, January 30, 2015: Jason Haugen, Oberlin College

Title: Uto-Aztecan Lexicostatistics 2.0


Much work in contemporary historical linguistics revives a long-standing interest in applying quantitative methods to questions of language comparison and classification. The Uto-Aztecan language family has figured into such discussions for decades, including work in lexicostatistics (Miller 1984; Cortina-Borja and Valiñas 1989), phonostatistics (Brown et al. 2008), and mixed methods (Cortina-Borja et al. 2002). Thanks to a half-century of research by scholars such as Wick Miller and Ken Hill, we now have in hand a massive corpus of Uto-Aztecan cognate sets (K. Hill 2014, n > 1200), which is ripe for computational investigation. In this talk I report on the results of our current Oberlin College-based lexicostatistical research utilizing this corpus. Questions to be asked and answered include: (i) is there an optimal clustering algorithm for generating phylogenetic trees for languages?; (ii) is there an ideal wordlist for lexicostatistical comparisons?; and (iii.) is there an ideal wordlist size for lexicostatistical comparisons?


Friday, February 6, 2015: Susan Penfield, University of Arizona

Title: Interdisciplinary Approach to Language Documentation


Classic language documentation was often limited to the study of specific areas of linguistics or genres of language which could be recorded and archived for further study in an effort to preserve the salient features of a given language or language family. It has taken sometime for researchers to fully embrace the feasibility and utility of broadening basic language documentation to be more inclusive of other disciplines.

In this talk, the presenter shares four papers presented at the International Conference of Language Documentation and Conservation (Honolulu, 2013) which brought to light how researchers can think about and incorporate other disciplines into their language documentation efforts. These four studies show how specific field research in language documentation has incorporated methods from studies of child language acquisition, botany and ethnobotany, areal sociolinguistics, and geographic research into the research design. This talk raises issues of both possibilities and problems related to crossing disciplines in research contexts. As well, the presenter will also highlight areas of her own research involving Native American communities which particularly invite the union of language documentation with input from other disciplines to better document not only language but also cultural practices for posterity.


Friday, February 27, 2015: Joe Salmons, University of Wisconsin, Madison

Title: Everything you thought you know about i-umlaut is wrong: New data and analysis


Germanic languages show a regressive vowel-to-vowel assimilation pattern known as i-umlaut, one of the most discussed changes in historical phonology, especially in Old High German (OHG). This paper shows, first, that previous analyses, including recent ones, are based on faulty evidence about umlaut’s unfolding in OHG. The traditional, still-dominant view posits early and uniform allophonic umlaut — fronting and/or raising of all back vowels before *i/*j — which is phonemicized as one phonological rule at one point in time with the loss of those triggers. This analysis leads to a set of famous paradoxes, including why the fronted allophones did not revert on the loss of triggers. We have created a corpus from manuscript facsimiles and draw on work in language variation and change and acoustic and perceptual phonetics to interpret this data. Our data shows a gradual and systematic unfolding of umlaut, patterns consistent with attested complex sound changes. The present analysis eliminates the paradoxes and accounts for a broader range of data than earlier work.


Friday, March 6, 2015: Jeff Punske, Southern Illinois University, Carbondale

Title: Form and Interpretation in Distributed Morphology


In this talk, I examine the nature of non-compositionality within the framework of Distributed Morphology--with special focus on idioms and compounds. Tying together strands from Harley (2011), Punske and Stone (2014, in prep) and Jackson and Punske (2013, in prep), I argue that traditional approaches to non-compositionality runs counter to the basic tenets of DM, but that outlined approaches ameliorate this problem. After developing the theoretic approach to non-compositionality, I revisit some of the long outstanding issues in non-compositionality (many of which formed the basis of the so-called "Linguistics Wars") and discuss how DM has solved these once intractable issues.


Friday, April 10, 2015: Carrie Gillon, Arizona State University

Title: Reanalyzing Michif “Determiners”


Michif, a mixed language derived from Plains Cree and Laurentian French, offers a unique insight into the effects of language contact. These effects are potentially even more interesting given the typological differences between French (Romance) and Cree (Algonquian). Michif has been characterized as having French DPs and Cree VPs, each with their own grammatical system stemming from their source language (Bakker 1997). However, the situation is more complicated than that: while the “determiners,” possessives, numerals and most quantifiers are derived from French, the demonstratives and some quantifiers are derived from Cree. We argue that though what looks like a determiner system in Michif is historically French, the system does not display properties typical of French determiners, and that the determiners appear to have been reanalyzed as Class prefixes, contra Rosen (2003). If these French-source “determiners” are in fact affixes in Michif, it means that a French syntactic item has been reanalyzed into the Algonquian nominal morphology. This lends support to arguments that Michif is typologically an Algonquian language, with heavy borrowing from French, rather than a mixed language with two intertwined grammars, as argued in Bakker (1997).


Friday, May 1, 2015: Department of Linguistics

Title: The 3rd Annual Undergraduate Research Forum

More Information about the Undergraduate Research Forum can be found here.