03/05/05

 

Erik W. Willis, Ph.D.  (University of Illinois at Urbana-Champaign)

Assistant Professor of Linguistics and Spanish

Department of Languages and Linguistics

New Mexico State University

ewillis@nmsu.edu

 

Research interests: Spanish, Phonetics and Phonology, Intonation Theory, Prosody, Acoustics, Variation, Caribbean Spanish, Border Spanish, Bilingualism.

Selected presentations and papers (all abstracts are listed below)

•  Dominican Spanish Intonation: Findings and Analysis.  Ph.D. Dissertation, University of Illinois at Urbana-Champaign. (abstract and pdf files of the dissertation, Willis-diss-2003).

•  An Acoustic Study of the Trill in Dominican Spanish.  The 35th Linguistic Symposium on Romance Languages. The University of Texas at Austin, February 24-27, 2005.  (PowerPoint presentation with sound files, Willis-LSRL-2005)

•  Tonal Levels and Boundary Tones of Broad Focus Declaratives, Absolute Interrogatives, and Pronominal Interrogatives in Puebla Mexico Spanish.  Paper presented at the Sixth High Desert Linguistics Conference, University of New Mexico, November 4-6, 2004.  (PowerPoint presentation with sound file, Willis-HDLS-2004).

•  (forthcoming).  Tonal Levels in Puebla Mexico Spanish Declaratives and Absolute Interrogatives.  In (eds. Randall Gess and Ed Rubin) Proceedings of the Linguistic Symposium on Romance Languages, University of Utah, March 2004.

•  2003.  Prenuclear Low Tone Alignment in Dominican Spanish.   Proceedings of the XV International Congress of Phonetic Sciences, Barcelona, Aug. 3-9, 2003. pp 2941-2944. (Figures and sound files, Willis-ICPhS-2003).

•  2002. Is There a Spanish Imperative Intonation Revisited: Local Considerations.  Linguistics 40:2 pp 347-374.

********************************************************************************************************************************************************************************************************************************************

•  An Acoustic Study of the Trill in Dominican Spanish.  The 35th Linguistic Symposium on Romance Languages. The University of Texas at Austin, February 24-27, 2005.

This paper reports the findings of an acoustic study on the production of the Spanish trill in intervocalic position and word initial position.  There are variable reports on the acoustic realization of the /r/ in Dominican Spanish.  Jiménez Sabater (1975) reports findings that conflict with previous accounts by Navarro Tomás (1957) and Amado Alonso (1967).  Several phonological analyses are based on these previous descriptions (Bradley in press, Núñez Cedeño 1989, 1994), yet the phonetic facts are limited and contrast among the accounts.

The current study examines the production of the trill in the Cibao dialect spoken in Santiago, Dominican Republic.  The informants for the study were ten university students native to the Cibao region, six males and six females, studying at various universities in the same city.  The data for the study is based on a narrative telling of a children’s picture book by Mercer Mayer, Frog Where Are You?.  The story is about a perro ‘dog’,  a rana ‘frog’  and a little boy,  and induced multiple productions.  The trill was often produced both within the tonic syllable as well as atonically due to the speaker’s spontaneous addition of a diminutive morpheme.  Each speaker was engaged in an informal interview about himself or herself to gather additional tokens.  Each speaker produced an average of 15-25 tokens of the trill.  The informants were recorded using USBpre external sound card and a Shure 512 microphone.  Each production of the trill was extracted from the interview recording and analyzed using PRAAT.  The analysis of the data takes into consideration the sex of the speaker and the phonological context (vowel quality and stress location).  The acoustic measurements include the duration of the ‘trill’, the duration of the preaspiration, the number of occlusions and the presence of voicing.

The data indicate a strong tendency toward a “preaspirated tap” resulting in two distinct articulatory gestures produced as a single segment.  The period of preaspiration was typically more than 50 ms, and longer than the following occlusion(s).  The data provide a much clearer understanding of the rhotic segment than earlier auditory descriptions.  The findings are discussed with respect to the previous studies, including the generational gap between the current informants and earlier generations, and the implications for dialectal variation and a phonological characterization of the Dominican Spanish trill.

 

•  Tonal Levels and Boundary Tones of  Broad Declaratives, Absolute Interrogatives, and Pronominal Interrogatives in Puebla Mexico Spanish.  Paper presented at the Sixth High Desert Linguistics Conference, University of New Mexico, November 4-6, 2004. 

This paper examines tonal levels and boundary tones of read speech of Puebla Mexico Spanish declaratives and interrogatives produced in response to a broad focus contextual prompt. The characteristics examined include tonal levels measured at the boundaries and pitch accent tonal targets, and the type of boundary tones. Differences in intonational tonal levels are reported to serve as a determining factor in distinguishing between declaratives and interrogatives in Spanish (Navarro Tomás 1944; Prieto 2003, forthcoming; Sosa 1999). Willis (ms) reports that while tonal levels and variances in prominence are determining factors in distinguishing between Dominican Spanish declaratives and interrogatives, prominence relations in Dominican Spanish function differently than previously reported dialects. Currently there are few empirical studies that compare dialectal variation of intonation across utterance types. This paper provides an additional dialectal analysis of Puebla Mexico Spanish utterance types for comparison with the quantitative claims reported for Peninsular Spanish and Dominican Spanish. The read speech data was produced by three young female speakers studying at the Universidad de las Américas, Puebla from Puebla Mexico.

The present corpus includes over 30 repetitions of each utterance type of phonologically similar sentences (see [1]). The data was analyzed using Xwaves and Praat. The results of the comparisons of tonal levels for the three utterance types indicate that each utterance type can be distinguished by intonational means. The primary difference between the declarative productions and the interrogatives is the directionality of the final boundary tone, L% versus H%. The declaratives were produced with a final tonal fall, while both interrogatives were typically produced with a final tonal rise. A significantly higher initial tonal value and prenuclear pitch accent distinguished the pronominal interrogatives from the absolute interrogatives.

These findings on Puebla Mexico Spanish differ significantly from previous dialectal characterizations of Spanish (Prieto 2003, 2004; Sosa 1999; Willis ms) and suggest considerable variability among the dialectal systems. This paper contributes to our general knowledge of intonational patterns in Spanish and utterance signaling with new set of empirical data from a previously unexamined dialect of Spanish. These findings identify specific characteristics of intonational variation within Spanish dialects, and suggest the need for additional dialectal studies before assuming a comprehensive characterization of the intonational characteristics responsible for distinguishing between Spanish utterance types.

 

•   Tonal Levels in Puebla Mexico Spanish Declaratives and Absolute Interrogatives.  In (eds. Randall Gess and Ed Rubin) Proceedings of the Linguistic Symposium on Romance Languages, University of Utah, March 2004. (forthcoming).

Since Navarro Tomás’ (1944) classic characterization of Spanish intonation, it has been generally accepted, that compared to declaratives, interrogatives are produced at a higher tonal level from the onset of the utterance.  Prieto (forthcoming) verifies that interrogatives are distinguished from declaratives by an earlier tonal prominence associated with a prenuclear pitch accent tone based on laboratory data from Peninsular Spanish.  However, Willis (2003) claims that Dominican Spanish absolute interrogatives are not distinguished from declaratives at the onset of the utterance.  Rather, Willis argues that the Dominican Spanish data indicate that a target late in the utterance, the nuclear pitch accent, serves as the most reliable target for distinguishing tonal prominences and suggests that intonational disambiguation strategies may be dialect specific.  This paper continues the search for dialectal variation and examines the intonational differences in Puebla Mexico Spanish declaratives and absolute interrogatives through a comparison of tonal levels of five potential targets.  Similar to Dominican Spanish, Puebla Mexico Spanish does not disambiguate between utterance types at utterance initial tonal targets.

The corpus is based on a controlled experimental data set consisting of read speech of utterances in broad focus and includes an empirical examination of tonal levels of particular tonal targets.  The data were collected from three educated females at the Universidad de las Américas Puebla in Puebla, Mexico.  There were a total of 36 broad focus declaratives and 32 broad focus absolute interrogatives per speaker.  Each sentence was preceded by a context designed to elicit the particular reading.

The results of the tonal comparisons indicate that initial tonal values (IT) do not serve as a target for categorical distinctions between declaratives and interrogatives in this Spanish dialect.  The box plots also illustrate that the significant distinction between declaratives and interrogatives is produced on the final tonal boundary. The current paper adds to our knowledge of Spanish intonational patterns by empirically characterizing tonal differences between Puebla Mexico Spanish declaratives and interrogatives at key tonal targets.  The current findings also challenge long-held assumptions concerning the intonational signaling of utterance types in Spanish.  Finally, this paper highlights the dialectal variability present in the Spanish intonational system.

 

  2003.  Prenuclear Low Tone Alignment in Dominican Spanish.   Proceedings of the XV International Congress of Phonetic Sciences, Barcelona, Aug. 3-9, 2003. pp 2941-2944.

Spanish prenuclear rising pitch accents have been described as having a Low tone consistently aligned near the onset of the tonic syllable and a High tone, whose alignment may vary depending on focus marking.  An examination of Dominican Spanish (DS) reveals a previously unreported alignment pattern for Spanish.  The DS data reveals a variable Low tone alignment; an early alignment near the tonic onset, and a late Low alignment, past the tonic midpoint, typically 100-180 ms into the tonic syllable.  The High tone does not demonstrate two categories of alignment seen with the Low tone.  The variation observed in the DS prenuclear pitch accent alignments provide clear evidence of intonational dialectal variation and contributes to the growing body of research on intonational variation in general.

 

•  2002. Is There a Spanish Imperative Intonation Revisited: Local Considerations.  Linguistics 40:2 pp 347-374.

This paper reports the results of an experiment designed to reexamine the existence of an imperative intonation in Spanish by investigating acoustic cues at the local pitch accent level.  Navarro Tomás (1944) claims that intonation alone is sufficient to distinguish between declarative and imperative sentences.  Kvavik (1987, 1988) finds inconclusive evidence for the claim that intonation can distinguish between the two types of sentences for a Cuban dialect of Spanish based on global or utterance level correlates.  In the present study, intonational differences are examined at the local pitch accent level with a larger corpus of Mexican Spanish.

The current study finds that speakers differ in their use of local intonational strategies between imperative and declarative utterances, but that the results are not categorical.  The local differences between imperative and declarative utterances include: an increased tonal range at the local pitch accent level, reduced intonational deaccenting, an increased use of an early H tone pitch accent associated with contrastive focus, and modifications of duration. 

 

 

This site was last updated 03/05/05