LINGUIST List 4.1

Wed 6 Jan 1993

Sum: Measuring Pause

Editor for this issue: <>

Directory

1. , Measuring pause: summary

This summary was written by Peter Carroll, a Ph.D. student in
Anthropology, Northern Territory University, Darwin,
Australia, on whose behalf the original query was posted. The
request on 2 November 1992 by Patrick McConvell was for advice
on the measurement of pause length.

15 responses were received; many of the responses were based
on study of pause in conversation, which is a little different
from our concern, which is the use of pause as a significant
feature in the analysis of narrative texts in an Australian
Aboriginal Language.

Various methods were suggested
 .the use of a stop watch - this was regarded as 'crude'
by some but seen as adequate for some situations by others.
 .the use of a sonogram - limited value for texts if print
outs are limited to 2-3 seconds at a time.

Computer software is available which assists in measuring
pause length among other things - the advice relates to
Macintosh programs.

 Signalyze - analyses up to 49 secs
 InfoSignal Inc -Lausanne Switzerland fax +41 21 881 6202
 Email:76357.1213  COMPUSERVE.COM

 Macrecorder Sound Edit - analyses up to 168 secs with 4
 meg of RAM
 Farallon Computing Inc 2000 Powell Street, Suite 600,
 Emeryville CA 94608 (415) 596-9000

 MacSpeech Lab II- The Darwin Applecenter advises that
 this program has been superseded by Sound Scope 16 and
 Sound Scope Classic

General Comments Received

It is important to determine the purpose in measuring pause.
Will it be comparison of speakers or comparison of speech
events?
Meaningful pauses will vary from speaker to speaker and within
the speech of one speaker. Variation will occur in relation
to overall speech speed.
Pauses will be relative to the overall speed and context of
the utterance.
Accurate measurement could be made in milliseconds which is
probably beyond the range of human perception. Pauses could
act as a contextualising cue.
A more emic description than an etic one would be more
appropriate.
Perceived pauses do not always correspond to silence
Consider the activities that frame or trigger the change in
speed.

Several people responded to the part of the original query
bringing up the question of of 'mean speed of speaking'.-

Various suggestions were made as to how to measure the rate of
speech:
 words/minute, syllables/minute, stressed syllable/minute,
 segments/minute
As word length in Kunwinjku (the Aboriginal Language of the
texts) varies from one syllable words (CV) to multisyllable
complex verbs the speech rate is being measured in
syllables/sec. Initial measurements have shown variation in
rate between speakers, between different texts of 1 speaker
and within a single text. The Speech rate (syllables per
total text time) can be compared with the Articulation rate -
syllables per total articulation time (ie total time less
pause time) (Terminology from Griffiths 1991).

Other measurements suggested were:
 the ratio of sound to silence
 the ratio of the duration of pause to the duration of the
utterance.

Andy Butcher has been in Australia studying the phonetics of
Australian Aboriginal languages. He recommended the use of
the Signalyze program for the Macintosh and made these
observations in relation to the analysis of Kunwinjku
Aboriginal language texts
 .don't ignore intonation patterns
 .consider the segments preceding each pause for common
 patterns
 .look at rhythm and speed
 .consider language specific features and personal style
 .consider categorising stories

References I have seen, with comments

Dankovicova, Jana (1992)
 "Minimum Pause Duration in Spontaneous Speech" in
Progress Reports from Oxford Phonetics Vol 5, June 1992.

experiment to gather data on pauses ranging from 0.10 to 0.20
secs in spontaneous speech; a KAY DSP Spectograph 5500 was
used for the experiment; pauses less than 0.13secs are
regarded as articulatory pauses; pauses longer than 0.13secs
serve the encoding process.

Griffiths, Roger (1991)
 "Pausological Research in an L2 Context:A Rationale, and
Review of Selected Studies' Applied Linguistics 12.4

defines 'pausology' as "the study of temporal variables in
speech"; accurate measurement of very short pauses is
problematical; has used a cut off point of 0.1 sec

References not yet reviewed

Atkinson M & J Heritage eds. (1984) "Structures of Social
Action" - introduction

Fiksdal, Susan "The Right Time and Pace: A Microanalysis of
Cross-Cultural Gatekeeping Interviews" - book

Fischer, Susan - a study on measuring pauses in ASL in
Cognition 1972 issue 1

Goldman-Eisler, F. (1958) "The predictability of words in
context & the length of pauses in speech" Language & Speech I
Part 3:226-231

Goldman-Eisler, F. (1968) Psycholinguistics: Experiments in
Spontaneous Speech" New York: Academic Press.

Longacre, Robert "Anatomy of Speech Notion"

Sacks, Schlegoff & Jefferson (1974) "A Simplest Systematics
for the Organisation of Turn Taking in Conversation" Language
50(4):696-735.

A number of further contacts for additional information, and
leads for additional references were received; I will post a
further report if these yield anything new. A frequently cited
source is Fr. Daniel O'Connell.

I have begun to use Mac Sound Edit and it seems to be working
fairly well so far to measure pause, but assigning analytical
significance to different pause lengths is more difficult.