Stratified Analysis of Multisyllables (SAP)

Check the syllabification and phone alignment of each record. SAP uses the stored alignment as it is: a production whose alignment is missing, has an empty column, or does not cover every phone is reported as unscored rather than realigned. Assign a speaker to each record so that its results appear in the correct participant section.

The analysis opens at Query: Data Tiers, which returns every aligned word of the Orthography tier. A word is analyzed when its IPA Target or its IPA Actual has more than one syllable, so a monosyllabic target produced as two syllables is included. Words with one syllable on both sides are counted as ineligible and otherwise ignored.

The report shows five summary tables for all participants, then for each participant and each session. Within each session, a section named #record orthography /IPA Target/ → [IPA Actual] holds the per-production tables.

Each per-production table, Word Level, Syllable Level and Phone Level, has the columns Process, Target, Actual and Note. One row is one process. An empty span is shown as ∅. An unscored production has empty tables and a Status line giving the reason.

Spans at this level keep their stress marks.

• Syllable addition ∅ → σ and Syllable deletion σ → ∅ for every syllable aligned with a gap. A stressed syllable also gets Primary-stress addition, Secondary-stress addition, Primary-stress deletion or Secondary-stress deletion.
• Primary-stress preservation or Secondary-stress preservation for an aligned pair with the same stress. Unstressed pairs produce no row.
• Stress change for an aligned pair whose stress differs; the note gives the direction, for example primary → none.
• Stress-pattern change once per production when the whole-word patterns differ, written with one σ per syllable, for example ˈσ → ˌσ.ˈσ.

Only aligned syllable pairs are examined. Spans hold audible phones only.

• Syllable-shape preservation or Syllable-shape change for every aligned pair, with the shapes in the note, for example CCVCC → CVC. Shapes use C and V.
• Onset reduction, Nucleus reduction, Coda reduction when a constituent loses phones; Onset-cluster reduction or Coda-cluster reduction when the target constituent had two or more.
• Epenthesis when phones are added to a constituent the target did not have; the note names the constituent. Onset expansion, Nucleus expansion or Coda expansion when a constituent the target had grows.
• Complex-onset resolution or Complex-coda resolution when a target cluster ends up in different syllables of the production. The suffix by epenthesis marks the case where an inserted vowel separates them, for example bɹ → bi.w.
• Resyllabification once per production when any phone moved to a different syllable, with the word shapes, for example CCVCC → CV.CVC.

An inserted vowel that creates a syllable is counted once as a syllable addition at the word level; the cluster resolution line describes its structural effect, not a second addition.

Each aligned pair of different phones, diacritics included, produces one row per matching phonological process, for example Devoicing d → t or Gliding ɹ → w. When no process in the library matches, the row is Substitution. The note lists the feature difference: features only in the target with −, features only in the actual with +. Identical pairs produce no row but count toward Phone Accuracy.

Process names are the English names from the phonological process library. The session's language decides which processes of the library apply. When the library cannot be used for a production, its rows are labelled Substitution and the Status line says why.

For /ˌbæn.ˈdæ.nə/ → [ə.ˈtæ] with the alignment ∅↔ə, b↔∅, æ↔∅, n↔∅, d↔t, æ↔æ, n↔∅, ə↔∅:

For /ˈbɹænʧ/ → [ˌbi.ˈwæʧ] with the alignment b↔b, ∅↔i, ɹ↔w, æ↔æ, n↔∅, ʧ↔ʧ:

Neither word lists ə, i, bæn, nə or n at the phone level.