The Impact of Accent, Noise, and Linguistic Predictability on the Intelligibility of Non-Native Speakers of English.

Abstract

In many situations today non-native speakers of English must speak English as an international language or as a common language between two non-native speakers. Such communication is often complicated by adverse listening conditions such as noise and high stress levels. This study examined the effects of linguistic predictability and noise factors on the intelligibility of non-native speakers of English with varying degrees of accent when their listeners were native English speakers. Speech recordings were elicited from four adult male native speakers of Brazilian Portuguese and one native speaker of English. Sentences from the Speech Perception in Noise lists were read by each speaker, representing native, mild, mild-moderate, moderate-strong, and strong foreign accents. Sentences were mixed with multi-talker babble with a signal-to-noise ratio of 6 dB, 10 dB, and 15 dB. Target words in half of the sentences were highly predictable, and the remaining half were of low predictability. All 30 listeners were native speakers of English. They wrote down the last word of each SPIN sentence from recordings of random selections of speakers and noise levels and rated spontaneous speech samples for degree of perceived accent and intelligibility pre- and post- SPIN listening task. Analyses of the data suggest that all three factors--accent, noise, and predictability-had a combined effect on the intelligibility of non-native speakers of English. Even the intelligibility of the native speaker was compromised when the signal-to-noise ratio was low and when the linguistic predictability was also low. When the native listener was challenged further by the addition of a foreign accent, intelligibility was even more compromised. This effect was greater as the degree of accent became progressively stronger.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 03, 1999
Accession Number
ADA370565

Entities

People

  • Kimberly R. Scott

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Ground and Sea Platforms

DTIC Thesaurus Topics

  • Accuracy
  • Air Force
  • Automated Speech Recognition
  • Communication Disorders
  • Computational Science
  • Diseases And Disorders
  • Grammars
  • Health Services
  • Hearing Loss
  • Intelligibility
  • Language
  • Linguistics
  • Measurement
  • Measuring Instruments
  • Medical Personnel
  • Neurobehavioral Manifestations
  • Statistical Analysis

Fields of Study

  • Linguistics

Readers

  • Speech Processing/Speech Recognition.