Speaker Series

The Linguistics Speaker Series is a series of invited talks, organized each semester by the grad students in the Penn Department of Linguistics. We invite students and professors from various subfields and various universities to speak about their current research. All talks are open to the greater Penn linguistics community.

For Fall 2009, the speaker series is organized by Josef Fruehwald. You can see a list of past speakers here.

Schedule

Talks are held on Thursday afternoons in the IRCS Large Conference Room unless otherwise indicated. IRCS is at 3401 Walnut Street, 4th floor, suite 400A. Make two lefts out of the elevators, and the Large Conference Room is the very first door on the left within IRCS, room 470 (Directions|Parking|Map).

Announcements of talks are posted below, in the department, and to the Penguists mailing list.

Speakers for Fall 2009:

• September 24 - Rena Torres Cacoullos, Pennsylvania State University
  Grammaticalization Through Variation
  Using multivariate analysis, this study tracks the configuration of factors conditioning variation between the Spanish Progressive (estar ‘to be (located)’ + Verb-ndo) and the simple Present, in 15th, 17th and 19th century data. The Progressive begins as a locative construction, as shown by strong early effects of co-occurring locatives. Over time locative meaning weakens and aspectual meaning, which is inherently part of the locative construction, strengthens. The changing relative magnitude of effect of the factor groups in the multivariate analyses shows the gradual development of a progressive-nonprogressive opposition. Thus, in the course of speakers’ recurrent choices in discourse between functionally overlapping variant forms, the variants themselves evolve, developing different functions.
• October 1 - William Idsardi, University of Maryland
  Learning Vowel Phonemes
  In this talk I will review and critique the standard understanding of vowel phoneme acquisition. The usual conception is that there is an initial stage which estimates phonetically significant categories (allophones) followed by a procedure to group the discovered phones into phonemes. This view of phoneme acquisition has had recent statistical polishing with model-based clustering (Vallabha et al 2007) and Kullback-Leibler divergence (Peperkamp et al 2005). Following a review of some of the shortcomings of these methods, I will sketch an alternative that renders the allophonic level epiphenomenal. The alternative view uses hierarchical statistical models of vowel phonemes and phonological processes to derive the allophones, resulting in more parsimonious coverage of the data.
• October 29 - Raffaella Zanuttini, Yale University
  Subject-Verb Agreement in Appalachian English

  Appalachian English displays a different pattern of subject-verb agreement than standard American English. In particular, in the present tense the verb displays verbal -s in co-occurrence with plural lexical subjects, as shown in (1); in contrast, it does not display verbal -s with plural pronominal subjects, as shown in (2):

  1. (1)
     1. a. All preachers likes fried chicken. (GC, DOH) (Appalachian English)
     2. b. The Freewills believes in feet washing. (EC, DOH)
  2. (2)
     1. a. We even grind the cornmeal.
     2. b. Cause you go and buy, over there you buy, uh, macaroni, in Mountain City.

  The literature on other varieties of English that exhibit a similar pattern has proposed that examples of the type in (1) show lack of subject-verb agreement. We take a different approach and suggest instead that verbal -s expresses different features in different varieties of English: while in standard English it expresses agreement in number, in Appalachian English verbal -s expresses agreement in person. Hence, the examples in (1) constitute instances of subject-verb agreement in person, and not of lack of agreement.

  This view of what the patterns mean leads us to the following hypothesis about the difference between standard English and Appalachian English: while standard English does not have a feature person in the syntactic representation of a clause, Appalachian English does. This has two interesting consequences, which I will explore in my talk. The first is that the feature person must be matched by an overt element that has a person morpheme in Appalachian English, but not in standard English. This morpheme can be either on the subject or on the verb. The second is that the syntax of Appalachian English must be closer to that of Icelandic, which is said to have person marking, than to that of the mainland Scandinavian languages, which are said to lack it.

• November 12 - Abigail Cohn, Cornell University
  What are the Elements of Phonology? And Where Do They Come From?

  A basic issue in phonology that is often taken for granted is what are the elements of phonology and where do they come from. What is the role of features, segments, prosodic structure, and what is the source of such structure? I argue that our underlying and often implicit assumptions about the nature of phonology frame our answers to these questions and warrant reconsideration. I pursue two points: 1) the need to reconsider the close linkage of the nature of cross-linguistic phonological patterns and the patterns of language acquisition and 2) the sources of phonological universals, that is, cross-linguistic commonalities in sound systems. The discussion is philosophical in nature, drawing on the empirical and experimental literature on sounds patterns and their acquisition.

• November 19 - Paul Portner, Georgetown University
  Permission and Choice

  This talk investigates the interface between discourse semantics and modality by focusing on the two topics mentioned in the title. I begin by developing an analysis of permission imperatives:

  1. 1) Have a piece of fruit!

  This analysis is based on the dynamic analysis of imperatives discussed in Portner (2004, 2007), and draws crucially on the idea that imperatives serve to impose an order on the worlds compatible with the common ground.

  I then show that this analysis naturally provides explanations for free choice phenomena with imperatives and for Ross's paradox:

  2. 2) Have an apple or a pear! --> The addressee is permitted to have an apple, and the addressee is permitted to have a pear.
  3. 3) Have an apple! -/-> Have an apple or a pear!

  Of course permission readings and free choice phenomena don't only occur with imperatives; they are typically discussed in the context of modal sentences:

  4. 4) You may have a piece of fruit.
  5. 5)
     1. a)You may have an apple or a pear.
     2. b)That might be John or Harry.

  I show that the theory developed for imperatives can be extended directly to a certain modals, namely the performative ones (which includes the permission-giving ones like (4)). I then argue that choice phenomena with descriptive (non-performative) modals are less robust than those with performative modals, and should therefore receive a slightly different account.

  In sum, this talk aims to make three contributions:

  • An analysis of permission.
  • An analysis of choice phenomena with imperatives and performative modals.
  • A better understanding of the empirical status of choice phenomena across a range of modal subtypes.
• December 3 - Thomas Roeper, University of Massachusetts Amherst