Reviewing the language annotations behind clause counts
Editable parts of speech now appear alongside the transcript, making the annotations used for clause counts available for review and correction.
Work done September 9, 2026 · Last edited September 25, 2026
On this page
Arc now shows the part-of-speech labels used to calculate clause counts directly beneath the words in the transcript. These labels make the language annotations behind the analysis available for review and correction.
Parts of speech alongside the words
An initial set of part-of-speech labels is generated when a sample is imported. The labels are editable, so an SLP can review how words have been classified and correct an inappropriate label in context.
The report’s clause counts use these annotations. Corrections therefore contribute to the analysis rather than remaining separate notes on the transcript.
This work also replaces the earlier verb-list approach with rules that use parts of speech and the surrounding words to identify clauses.
Utterance boundaries in the initial draft
The initial draft also uses pauses in the recording and a rule for dividing utterances joined by repeated “and” clauses. The resulting divisions provide a starting point for transcript review; utterance boundaries remain editable.
