Effective speech communication relies on a speaker's ability to convey a
message. The most basic requirement for speech is a sound source, normally
the larynx (voice box). The normal speaker depends on the fine-tuning
capabilities of the larynx to vary F0, intensity and duration, which are
essential to convey the prosodic structure of an utterance. Pitch changes,
for example, accentuate the important words in a sentence. Words are also
lengthened or pauses inserted after words to signal the end of a phrase or
sentence. The research presented in this dissertation focused on speakers
who have had a laryngectomy (surgical removal of the larynx). These
alaryngeal speakers rely on an alternative sound source, namely mucosa and
muscle situated at the entrance to the esophagus. Alaryngeal speakers'
control over this alternative voice is limited.
A series of perception experiments revealed that alaryngeal speakers who
were able to vary the relevant prosodic cues consistently, conveyed
prosodic intent more accurately than speakers who could not. However,
speakers who had no, or no consistent control over the relevant prosodic
cues, often managed to signal, for example, the intended accented word, or
managed to convey the correct phrasing. This was achieved by manipulating,
albeit inconsistently, other – sometimes unexpected - prosodic cues that
are not normally associated with the prosodic function in question. It is
therefore important to investigate which prosodic cues are still present in
an alaryngeal speaker's speech. Through subsequent training of those cues
that are still available, it might be possible to improve the speaker's
overall communicative effectiveness.