LINGUIST List 2.816

Sat 23 Nov 1991

Disc: Segments and Speech Interfaces

Editor for this issue: <>

Directory

1. Jacques Guy, Chinese: a segmentless language?
2. , HCI/Speech Interfaces

Time to air a perfectly outrageous proposition which has been on my mind
for a few years. So outrageous that I am not sure I believe in it myself.
And again... Here: the order of the elements of a Chinese syllable (Peking
dialect, at least) is not emic. Enough said.

Re. Peter Benson's query about HCI research on speech
interfaces, here are some which touch on human errors
& effects of extended periods of ASR use:
Frankish C & Noyes J "Sources of Human Error in
Data Entry Tasks Using Speech Input". Human Factors
l990 32(6) 697-716.
Frankish C Jones D & Hapeshi K. "Maintaining Recognition
Accuracy during Data Entry Tasks Using Speech Input".
Contemporary Ergonomics (Ed EJ Lovesay) Taylor & Francis
l991 (pp 445-449).
Shortly to appear in Int. J. of Man Machine Studies:
Frankish C Jones D & Hapeshi K: "Decline in Accuracy
of ASR as a Function of Time on Task".
HR Kirby & PJ Roach of Leeds University also did some
work in l987 (?) on 'voice degradation' after long periods
of using ASR interfaces, & its effect on recogniser
performance - but I don't know if/where it was published.
Alison Murray.