Stanford University| Department of Linguistics

Autumn quarter regular meetings are Mondays at 7pm.
Please check back regularly for updates and special announcements.

2009-2010 Schedule

Fall 2009

• Monday, November 16, 7:00 pm.
         Greenberg Room, Margaret Jacks Hall, 460-r126
         Yao Yao, University of California, Berkeley
         A corpus-based study on the effect of phonological neighborhood density on speech production
         ▽ Abstract

  This study investigates the effect of phonological neighborhood density on speech production in spontaneous speech. Phonological neighborhood density (PND) refers to the number of words that are phonologically similar to a given word. Previous literature has shown that unlike other frequency- or probability-based measures (e.g. lexical frequency, contextual predictability), PND has opposite effects on perception (i.e. inhibitory) and production (i.e. facilitatory). Therefore it is an interesting question whether in actual speech, the speaker will hyperarticulate high-PND words for the sake of the listener or hypoarticulate them as a result of easy production.

  The current study uses data from a spontaneous speech corpus. The dataset consists of around 13,000 word tokens, all monomorphemic CVC content words. Separate linear mixed-effect models are built to predict word duration and vowel dispersion based on neighborhood density, average frequency of neighbors, as well as a wide range of control factors from the speaker, the word and the context. The results show that everything else being equal, high-PND words are shorter than low-PND words, and words with high-frequency neighbors are realized with more dispersed vowels than words with less frequent neighbors. In other words, the results with word duration provide evidence for speaker-internal forces in speech production, but the results with vowel dispersion indicate the pressure to distinguish confusing words in the lexicon. Implications of these results for the speech production model are also discussed.

  
  
         --> Dinner starting at 6:50pm <--
  
  
• Monday, November 30, 7:00 pm.
         Greenberg Room, Margaret Jacks Hall, 460-r126
         Abby Kaplan, University of California, Santa Cruz
         TBA
  
  ----Previous meetings from Fall 2009----
  
  
• Saturday, October 17, 10:30am.
         Windy Hill Open Space Preserve
         Special event! Phonetics/Phonology Hike
  
  
• Monday, November 2, 7:00 pm.
         Greenberg Room, Margaret Jacks Hall, 460-r126
         Hideki Zamma, Stanford University
         "Factorial Typology of Prosodic and Morphological Constraints and Classhood in English"
         ▽ Abstract

  Since Chomsky and Halle (1968), it has been widely assumed that English suffixes can be divided into two major categories (Siegel (1974), Allen (1978), Kiparsky (1982), Halle and Mohanan (1986), Halle and Vergnaud (1987), Benua (1997), etc.). The defining characteristics of classhood include, for example, (i) whether or not a suffix is stress-neutral and (ii) whether or not a suffix is capable of attaching to the root base. From time to time in the literature, however, it is pointed out that some suffixes have "dual membership" in both of the classes. In other words, the same suffix sometimes exhibits both stress-neutral and root-attaching behaviors (Aronoff (1976), Selkirk (1982), Fudge (1984), Szpyra (1989), Giegerich (1999), etc.). This fact has posed a serious problem to any theory of lexicon with dichotomy, because it suggests that some of the characteristics in suffixation might not be attributed to the difference between the two classes.

  This paper proposes that a proper analysis of the problem at hand can be given within the framework of Partial Ordering Theory (Anttila and Cho (1998), Anttila (2002)). Assuming that various groupings in the lexicon can have different constraint rankings, the dual membership suffixes can be analyzed as having the same rankings as Class 1 and 2 at the same time. Moreover, a POT analysis predicts a different type of suffixes not observed to date. This kind of suffixes is actually borne out in an overwhelming investigation utilizing the SOED.

  Based on a research on the same dictionary, it is also investigated how many suffixes (out of 120) actually belong to each class. Considering the distributional facts revealed from the investigation, it is further discussed if they can be theoretically predicted from POT.

  
  
  
• Monday, November 9, 7:00 pm.
         Greenberg Room, Margaret Jacks Hall, 460-r126
         Matthew Adams, Stanford University
         "Welsh meter: An Optimality-Theoretic account of the cynghanedd lusg and cynghanedd groes"

Winter 2010

• TBA
  

Spring 2010

• TBA

When and where

Regular meetings of Phonology Workshop Fall 2009 are Mondays at 7pm. Unless otherwise specified, we meet the Greenberg Room (460-126), Margaret Jacks Hall, Department of Linguistics, Stanford University (Building 460). Please stay tuned for more information.

Participating and attending is free and open to everyone. Here is some information about getting to Stanford.

If you are not on the Phonology Workshop mailing list, you can join by filling out your information HERE. Likewise, you may also unsubscribe at the same address.

2009-10 organizers: Arto Anttila and Stephanie Shih