Boundary tone

The term boundary tone refers to a rise or fall in pitch that occurs in speech at the end of a sentence or other utterance, or, if a sentence is divided into two or more intonational phrases, at the end of each intonational phrase.

The term was first introduced in a PhD thesis on English intonation by Mark Liberman in 1975 but without being developed further.

[2] In Pierrehumbert's model, which later developed into the ToBI system of intonational transcription, every intonational phrase is marked as ending in a boundary tone, written either H% when the speaker's voice rises up or remains high, or L% when it falls or remains low.

If it is pronounced the second way, the words sister and Mary both have a falling intonation, and each one is transcribed by Pierrehumbert as H* L− L%.

Because of its simplicity compared with previous attempts at transcribing English intonation, Pierrehumbert's model has been influential[8] and has been successfully adapted to several other languages, for example Persian,[9] German,[10] and Dutch.

A corpus-based study of yes–no questions in American English found that the great majority of them (approximately 90%) ended in a high boundary tone (H%), most frequently (80%) using a "low-rise" final contour transcribed L*H-H%.

A typical low-rise question transcribed in the study is And do you still work for a veterinarian?, with the syllable ve- marked as L* followed by a smooth rise to a high pitch at the end.