Enhancing Listening and Spoken Skills in Spanish Connected Speech for Anglophones

Elena Paz Vizcaya, Dublin Institute of Technology

Document Type Theses, Ph.D

Submitted for PhD thesis to School of languages/Digital media Centre,Dublin Institute of Technology June 2012.



Native speech is directed towards native listeners, not designed for comprehension and analysis by language learners. Speed of delivery, or economy of effort, produces a speech signal to which the native listener can assign the correct words. There are no discrete words in the speech signal itself therefore there is often a linguistic barrier in dealing with the local spoken language.

The creation, development and application of the Dynamic Spanish Speech Corpus (DSSC) facilitated an empirically-based appreciation of speaking speed and prosody as obstacles to intelligibility for learners of Spanish. “Duologues”, natural, relaxed dialogues recorded in such a manner that each interlocutor’s performance can be studied in isolation, thus avoiding problems normally caused by cross-talk and back-channelling, made possible the identification of the key phonetic features of informal native-native dialogue, and ultimately, the creation of high quality assets/ research data based on natural (unscripted) dialogues recorded at industry audio standards.

These assets were used in this study, which involved documenting productive and receptive intelligibility problems when L2 users are exposed to the Spanish speech of native speakers. The aim was to observe where intelligibility problems occur and to determine the reasons for this, based on effects of the first language of the subjects, and other criteria, such as number of years learning/using Spanish, previous exposure to spoken Spanish and gender. This was achieved by playing recorded extracts/ snippets from the DSSC to which a time-scaling tool was applied.