L2 stress perception: the reliance on different acoustic cues

Qian Wang2008 · DOI

TL;DR

While native English speakers use a balanced "cocktail" of pitch, loudness, and timing to identify word stress, Chinese learners of English rely almost exclusively on pitch (F0). This study reveals that "prosodic transfer" from a tonal language isn't just about phonology—it’s a deep-seated phonetic bias where learners remain "deaf" to duration and intensity cues that native speakers take for granted.

What Problem Does This Paper Address?

In English, lexical stress (e.g., RE-cord vs. re-CORD) is signaled by a combination of three acoustic cues: fundamental frequency (F0/pitch), intensity (loudness), and duration (length). For native speakers, these cues work in tandem. However, second-language (L2) learners often struggle with this.

Previous research suggested that learners from "fixed-stress" languages like French are "stress deaf," while learners from tonal languages like Chinese might suffer from "tonal transfer." But what does that transfer actually look like at the acoustic level? The author, Qian Wang, investigates whether Chinese learners are truly "deaf" to English stress or if they are simply tuned into the wrong frequencies—specifically, whether they over-rely on pitch because their native language uses pitch to distinguish word meanings (tones).

Methods

The researcher designed a perception experiment using systematically manipulated nonsense words: latmab, nizdit, and tetsep. Using nonsense words ensured that participants’ performance wasn't biased by their existing English vocabulary.

The study employed a 5x5x5 factorial design:

  • Three Acoustic Cues: F0, Duration, and Intensity.
  • Five Levels of Manipulation: For each cue, the researcher created five steps ranging from "strongly favoring the first syllable" to "strongly favoring the second syllable."
  • Stimuli: This resulted in 375 unique tokens (3 words × 125 combinations of cues).

Two groups—34 Chinese learners (CE) and 25 native English speakers (NE)—listened to these tokens and identified which syllable sounded stressed. The researcher then used ANOVA (Analysis of Variance) to calculate how much each acoustic cue contributed to the listeners' decisions.

Datasets

The study utilized a custom-built stimulus set rather than a public database:

  • Primary Stimuli: 375 manipulated nonsense tokens based on recordings by a trained phonetician.
  • Foils: 100 real English words used to screen participants (only those with >80% accuracy on real words were included in the final analysis).
  • Participants: Data was collected from university students in Nanjing, China (CE group) and the University of Victoria, Canada (NE group).

Key Insights and Ideas

  • Selective Deafness: Unlike French speakers who struggle to hear stress at all, Chinese learners can perceive English stress quite well—but only if the pitch cue is present. They are effectively "deaf" to changes in duration and intensity when making stress judgments.
  • The "Pitch is King" Bias: In Chinese (a tonal language), pitch is the primary discriminator for meaning. The study suggests that learners don't just transfer the concept of tone; they transfer the perceptual weight they give to pitch.
  • Phonetic vs. Phonological Transfer: The paper argues that L2 acquisition hurdles aren't just about abstract grammar rules (phonology). Instead, they are rooted in "low-level" phonetic processing—the way our brains are wired to prioritize specific acoustic signals based on our first language.

Key Results

The quantitative divide between the two groups was stark:

  • Native English Speakers (NE): All three cues were significant. F0 was the strongest predictor (explaining 58.6% of the variance), followed by intensity (40.9%) and duration (10.7%).
  • Chinese Learners (CE): F0 was the only decisive factor, explaining a massive 91.2% of the variance.
  • Sensitivity Shifts: When pitch was manipulated from the lowest to the highest level, the percentage of "initial stress" judgments for Chinese learners jumped from 7% to 86%. In contrast, changing duration or intensity for the same group resulted in almost no change in their stress judgments (a flat line on the graph).

Limitations and Future Work

The author acknowledges a few areas for further exploration:

  • Scope of Cues: The study focused on F0, intensity, and duration, but did not investigate vowel quality (the reduction of unstressed vowels to a "schwa"), which is another critical cue for English stress.
  • Language Diversity: The findings are specific to Chinese-to-English transfer. The author suggests that future studies should look at Japanese learners, whose native language uses "pitch accent" (a different prosodic system), to see if they exhibit similar or different perceptual biases.
  • Production vs. Perception: While this study covers how learners hear stress, it doesn't address whether their production of English stress follows the same lopsided reliance on pitch.

Summary updated

Abstract

Second language learners of English often experience difficulties in English lexical stress perception because of prosodic transfer of their first language (L1).It has thus been hypothesized that the problem of Chinese learners with English stress arises from tonal transfer.However, little research has been devoted to the investigation of the phonetic details of second language (L2) stress perception.The present research focused on Chinese L2 learners of English.Their reliance on the three acoustic cues, F0, intensity and duration, is compared against a baseline of native English speakers, in terms of English stress perception of manipulated nonsense tokens.The results show that while all three cues have significant effects on stress perception for native English speakers, only F0 has a decisive effect on Chinese learners' stress judgments.This study suggests that there is transfer of reliance on F0 in the acquisition of L2 English stress, rather than transfer of tone at the phonological level.
BibTeX
@misc{wang2008l,
  author = {Qian Wang},
  title = {{L2 stress perception: the reliance on different acoustic cues}},
  year = {2008},
  doi = {10.21437/speechprosody.2008-141},
}
Figure