Evidence against the mismatched interlanguage speech intelligibility benefit hypothesis

Area of Science:

• Linguistics
• Phonetics
• Second Language Acquisition

Background:

• Investigating speech intelligibility is crucial for effective cross-linguistic communication.
• Previous research by Bent and Bradlow (2003) established a foundation for understanding non-native speaker intelligibility.
• Understanding how first language (L1) background influences intelligibility is key in globalized contexts.

Purpose of the Study:

• To examine the impact of talker proficiency and first language (L1) background on keyword intelligibility.
• To determine if a mismatched L1 between listener and low-proficiency non-native speaker creates an intelligibility detriment.
• To compare the intelligibility of native English speakers and high-proficiency non-native English speakers.

Main Methods:

• Five talker groups (Korean high/low proficiency, Saudi Arabian high/low proficiency, native English) read carrier sentences with keywords.
• Fifty listeners from four L1 groups (Korean, Saudi Arabian, native English, mixed) participated in a word recognition test.
• Statistical analysis was used to compare intelligibility scores across listener and talker groups.

Main Results:

• Low-proficiency non-native speakers were significantly less intelligible to listeners with a different L1.
• Listeners did not find low-proficiency speakers with a shared L1 to be similarly unintelligible.
• No significant intelligibility benefit was found for high-proficiency non-native speakers over native speakers for listeners with different L1 backgrounds.

Conclusions:

• A significant "mismatched interlanguage speech intelligibility detriment" exists for low-proficiency non-native speakers.
• Accents of low-proficiency non-native speakers unfamiliar to the listener pose an intelligibility challenge.
• High-proficiency non-native speakers' intelligibility is comparable to native speakers for listeners with different L1 backgrounds.