Simon Fraser University Speech Error Databasesfused320x240_2015-11-13

Objectives

  • To create two large databases, one for English speech errors, another for Cantonese speech errors.
  • To extend and refine existing methodologies for speech error collection and classification.
  • To develop a classification system that supports rigorous testing of psychological effects and probing linguistically sophisticated language structures.
  • To support examination of phonetic errors, distinct from phonological errors, through classification and audio recordings.
  • To partner with language production researchers on research projects that enrich the database, and ultimately make the database a public resource (projected release date is 2018).

Progress

Results so far have assembled approximately 9,000 English speech errors and 1,000 Cantonese speech errors from audio recordings. Speech errors are documented in a database with 62 fields organized into field classes: record fields (facts about how record was made), example fields (superficial facts of the error), major and special class fields (analyze the error type and related facts), and word and sound fields. The intended and error words or phrases are phonetically transcribed, and analyzed in a host of linguistic dimensions (e.g., part of speech tag, open/closed class, syllable position and role, semantic and morphological relatedness). Preliminary results show that there are important differences between our method of collecting errors offline (listening to audio recordings) and most prior research collecting errors online (on-the-spot observation). Also, while we find that different data collectors hear different errors, after completing our training regime, collectors seem to detect the same types of speech errors, which justifies the use of multiple data collectors. The SFUSED database makes the distinction between phonological errors that substitute, add, or delete phonological segments, and phonetic errors that involve ambiguous, transitional, or intrusive phonetic structures.

Research team

Principal Investigator:

John Alderete (Simon Fraser University)

Collaborators:

Stefan Frisch (University of South Florida)

Alexei Kochetov (University of Toronto)

Research assistants:

Rebecca Cho

Monica Davies

Gloria Fan

Holly Wilbee

Jennifer Williams

Olivia Nickel

Crystal Ng

Queenie Chan