Simulating Speech Error Patterns Across Languages and Different Datasets.
Sofia Strömbergsson, Jana Götze, Jens Edlund and 1 others
PMID 33637011WHAT IT FOUND
Backing caused the largest simulated sound-pattern effects in Swedish, Norwegian, and English child speech.
Stopping caused the next largest effects. The simulated measures did not match clinicians' intelligibility rankings.
Key findings
01Backing was consistently ranked as having the most severe phonological effects across all measures and all three languages.
02Stopping was ranked as the second most severe error pattern, with effects often larger for English than for the two Scandinavian languages.
0333 speech-language pathologists ranked six error patterns by likely intelligibility effect, and none of the computed severity measures correlated with their rankings.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study simulated errors in transcripts, so it does not measure real children's speech or intelligibility. The computed severity measures did not correlate with clinicians' intuitive intelligibility rankings. The clinician survey came from a Swedish speech-language clinic and may not reflect wider practice. The corpora were not equally representative across languages, and child speech data varied in age and source. English results were based on American pronunciation, which can affect r-related patterns. Error patterns were applied across whole word sets, not in natural contexts.
Declared interests
The paper reports funding by the Swedish Research Council. It states that the authors disclosed receipt of financial support.
The easy way to misread this
Do not treat backing as proven to worsen intelligibility or as a tested treatment target. The study simulated errors in transcripts and measured phonological structure, not real speech intelligibility, and the computed measures did not match clinicians' rankings.