Allard Jongman, Ph.D. - US grants

The broad objective of the proposed project is to gain a more thorough understanding of the production and perception of English fricatives. To date, no single metric has been able to classify place of articulation for all English fricatives with a high degree of accuracy. The proposed research will investigate two recent metrics for classifying place of articulation in fricatives - spectral moments and locus equations. While these metrics have successfully been applied to stop consonants, virtually no work has extended these approaches to fricatives. With appropriate modifications, these metrics seem particularly promising for fricative classification. Performance of these global metrics will be compared to a local metric - spectral peak location. Other properties such as noise duration, noise amplitude, and relative amplitude will also be evaluated in terms of their utility in classifying fricative place of articulation.The general research strategy includes detailed acoustic analyses of natural speech tokens produced by a large sample of speakers. In addition, perception experiments using computer-edited natural speech and synthetic speech will be conducted to help evaluate the accuracy of classification metrics and to investigate the "psychological reality" of these metrics. Finally, experiments are planned to specifically explore the separate contribution of auditory, visual, audio-visual, and contextual linguistic information to the perception of sounds which, it is often claimed, rely primarily on non-acoustic properties. By relating acoustic and perceptual data, and by comparing the role of auditory, contextual, and visual information, the proposed research thus aims at a comprehensive account of the acoustic and perceptual properties of English fricatives. In addition to increasing our general understanding of the normal processes of speech production and perception, this research will have practical benefits, including direct applications to hearing-impaired populations.

This project funds a new networked multimedia computer and audiovisual equipment for use by undergraduates in order to link and reform the undergraduate component in two independent language research labs. This new equipment involves high-end satellite work stations with CD-ROM and multimedia capacities in each lab, to be centralized in one new interdisciplinary Cognitive Studies undergraduate research laboratory/computer room. The project assists labs in developing innovative methods of involving undergraduates in hands-on scientific (hypothesis-testing) lab experiences that they can integrate with the theoretical study of linguistics. A larger number of undergraduates can be integrated in a sustained research experience, providing a model formal mechanism through the university's recently established Cognitive Studies program for undergraduate interaction and sharing of resources across labs. In addition, the project introduces and integrates a third research group, Adult Second Language Acquisition. Several new courses are being planned on the basis of the new lab improvement. Demonstration materials regarding language knowledge and organization, including multimedia learning modules, available on-line, are being developed and integrated in new and existing courses.

Despite its importance as one of the world's major languages, Arabic has not been the topic of much phonetic research. This project aims to contribute to the description and analysis of Arabic by focusing on one of its unique features - emphasis. Emphasis is a distinctive feature of Semitic languages such as Arabic and Hebrew that refers to a group of consonants produced with a secondary constriction in the posterior vocal tract and a primary constriction typically in the dental/alveolar region. The characteristics of emphasis are not confined to emphatic consonants themselves but can spread to adjacent segments as well. With support from the Linguistics Program and the Office of International Science & Engineering, Dr. Allard Jongman will conduct a detailed acoustic and perceptual investigation of emphasis. Acoustic analysis will include a variety of measurements of both consonants and vowels. Perception experiments will then evaluate the relative contribution of the emphatic consonant and adjacent vowels to the perception of emphasis. This research will document and enhance an understanding of the sound structure of Arabic. This project will examine a sizeable number of stimuli produced by a number of different speakers, thereby allowing for an assessment of differences in gender and dialect. In addition, the symmetrical or asymmetrical nature of the spread, the extent to which it diminishes as the distance from the emphatic increases, and the characteristics of potential blocking segments will have significant implications for phonological theory.

This research will enable linguists to test hypotheses on many aspects of the sound structure of Arabic in addition to emphasis, with the complete set of audio recordings from the project available through a web-based database. These data can also be used by Arabic language teachers to develop teaching materials that provide detailed information about unique aspects of the Arabic language. The broader impact of this project includes opportunities for undergraduate and graduate students to participate in linguistic research. This research will also be the start of a collaboration with several universities in Jordan. As such, this project will foster international cooperation, with the hope that this collaboration will eventually develop into a partnership, an exchange between faculty and students, and ultimately contribute to better cultural understanding.

When native speakers of American English begin learning Spanish, their acquisition of native-like pronunciation can be hampered by the tap versus trill distinction in words like CARO 'expensive' and CARRO 'car'. The trill proves difficult because it does not exist in English. The tap exists as an allophone of /t/ and /d/ in English words like 'writer' and 'rider', but English speakers must learn to process it as a phoneme rather than an allophone. Similarly, learners have difficulty acquiring the spirantization of voiced stops, where the /d/ in CODO 'elbow' is produced as a voiced dental fricative, which is more like the 'th' sound in English.

This study investigates whether American English-speaking learners of Spanish can be trained to perceive and produce the intervocalic tap, trill, /t/, and /d/ contrasts in Spanish. Participants will be trained using both perceptual and production training methods. Past research has reported that perceptual training alone improves both perception and production and that production training alone improves both as well, but the production training studies have not been limited to production as trainees have been able to listen to the training stimuli.

Both training modalities will be systematically controlled in this study so that they can be directly compared. A third training methodology will be introduced that includes both perception and production to discover whether perceptual training, production training, or a combination of the two is most effective. This study will use cross-modal priming and ERP data in addition to traditional tasks (identification and production tasks) to evaluate the effect of training, an innovative use of both tasks to determine if trainees not only perceive and produce the trained L2 contrasts but also if they unconsciously process these contrasts and if they have built new phonemic categories for these sounds.