Explicit pronunciation instruction improves controlled production of targeted sounds and prosody, with smaller and less certain gains in spontaneous speech¶

Claim · All claims
Evidence · 3 studies · 2 quant-synthesis r?, 1 review r2 · q2–q3 · i2–i3
Currency · newest study 2019 · 1 of 3 entries checked against their sources (automated) · not reviewed by a person

Subclaims¶

q2 i? The article states that the general consensus of pronunciation research is that clear and demonstrable results are obtained when teachers have students pay explicit attention to pronunciation features and dedicate class time to meaningful and communicative practice. → Darcy 2018 q3 i3 A meta-analysis of 86 reports finds a generally large effect of pronunciation instruction (d = 0.80 between groups), larger with longer interventions and with feedback; the abstract does not isolate communicative practice. → Lee et al. 2015 q3 i2 A re-analysis of 77 studies finds instruction reliably improves controlled production of specific segmental and suprasegmental features, but gains on global measures and in spontaneous speech are not distinguishable from zero between groups. → Saito & Plonsky 2019

Evidence¶

Darcy 2018¶

Darcy, I. (2018). Powerful and effective pronunciation instruction: How can we achieve it? The CATESOL Journal, 30(1). doi:10.5070/b5.35963

q2 · i? · review · r2

Narrative synthesis of prior studies and meta-analyses (e.g., Lee, Jang, & Plonsky, 2015; Saito, 2012), which the article says confirmed global improvement from instruction sessions lasting only a few weeks, including improved intelligibility and comprehensibility.

"the general consensus of pronunciation research today is that clear and demonstrable results are obtained when teachers have students pay explicit attention to pronunciation features and dedicate class time to meaningful and communicative pronunciation practice."

Lee et al. 2015¶

Lee, J., Jang, J., & Plonsky, L. (2015). The Effectiveness of Second Language Pronunciation Instruction: A Meta-Analysis. Applied Linguistics, 36(3), 345–366. doi:10.1093/applin/amu040

q3 · meta-analysis (abstract only) · i3 · large effect, d=0.80 between-group · n=86 reports · quant-synthesis · r?

A meta-analysis of 86 unique reports testing second-language pronunciation instruction, coded for study features and outcomes as Cohen's d. Instruction had a generally large effect: d = 0.89 for N-weighted within-group (pre-post) contrasts and d = 0.80 for between-group contrasts. Effects were larger for longer interventions, for treatments that provided feedback, and on more controlled outcome measures. Read from the abstract only: the abstract does not say whether the instruction studied included communicative practice, does not give confidence intervals, and does not report how large the effects were on spontaneous or communicative speech, so it supports "pronunciation instruction improves pronunciation" rather than the communicative half of this claim.

"Aggregated results showed a generally large effect for PI (d = 0.89 and 0.80 for N-weighted within- and between-group contrasts, respectively)."

Saito & Plonsky 2019¶

Saito, K., & Plonsky, L. (2019). Effects of Second Language Pronunciation Teaching Revisited: A Proposed Measurement Framework and Meta‐Analysis. Language Learning, 69(3), 652–708. doi:10.1111/lang.12345

q3 · meta-analysis · i2 · medium effect, d=0.68 between-group · n=77 studies · quant-synthesis · r?

A re-analysis of 77 L2 pronunciation teaching studies (1982–2017), coding each outcome by what it measured (global impressions such as comprehensibility vs specific segmental and suprasegmental accuracy), how it was scored (human raters vs acoustic analysis) and what speech it elicited (controlled vs spontaneous). Overall, instruction helped: d = 0.68 (95% CI 0.49–0.86) for between-group contrasts and d = 0.73 (95% CI 0.69–0.78) for within-group contrasts, against d = 0.31 for control groups simply retested. But the gains were reliable only for controlled production of specific features (between-group, expert-rated d = 0.75, CI 0.42–1.07; acoustically measured d = 0.84, CI 0.45–1.24). For global measures and for spontaneous speech, every between-group confidence interval crossed zero (e.g. global, controlled d = 0.33, CI −0.18 to 0.85). Read from the authors' accepted manuscript (UCL Discovery), whose numbers the published abstract's conclusions match.

"According to the results, instruction is most effective when it targets learners' monitored production of specific segmental or suprasegmental features. The efficacy of instruction remains relatively unclear when gains are measured globally via subjective, human judgments, especially at a spontaneous level."

Studies bearing on this claim¶

Does explicit second-language pronunciation instruction improve pronunciation, and does the gain carry into spontaneous speech? This table gathers the studies behind this claim and the 1 narrower claims that bear on it: 3 distinct studies (2 quant-synthesis, 1 review), q1–q3, 2 of 3 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

The two meta-analyses support the first half of this claim, that explicit attention to pronunciation produces demonstrable improvement, and qualify the rest. Saito & Plonsky (2019) find the improvement is clear for learners' controlled production of targeted sounds and prosodic features, and unclear when measured as global comprehensibility or in spontaneous speech, which is the speech communicative practice is meant to develop. Neither synthesis, as read here, compares instruction with and without communicative practice, so the earlier title's "with communicative practice" was the source article's framing, not something these studies establish, and its "clear and demonstrable" overstated the evidence for spontaneous speech; the title now states what the syntheses show.

Related Claims¶

• CAPT software shows promising results for segmental pronunciation, while prosodic features and fluency still require further research and development — related
• Guided ASR practice did not produce statistically significant improvement in any specific pronunciation error type, with some errors unimproved or regressed — reports the opposite
• Profy practice shows larger observed pronunciation-intelligibility improvement than elicited imitation, with non-overlapping pre/post confidence intervals — a narrower finding that bears on this claim
• Learners receiving both form- and meaning-focused instruction improved more than a form-focused group (attributed to Park, 2000) — related
• The Fluspeak ASR-based pronunciation software gave good results with beginners focusing on phoneme production but poor results overall for advanced learners seeking fluency — related