. E

 

_ A

:_

.

—

,2
w;

__

‘ _

f __

.

.____.__:___*__:

C?F 1" -~

AN WWESTEG ’Ct'iufl

Ei‘fl'ED

:: AC:

A

NE 7’5‘3'

2
s

LEPREADERS TO DEYERM

.YS‘I‘Li

i

J“
‘09

YLLABLE 0? P

t‘
.1

jaws

 

ABSTRACT

AN INVESTIGATION OF THE ABILITY OF UNSKILLED
LIPREADERS TO DETERMINE THE ACCENTED
SYLLABLE OF POLYSYLLABIC WORDS
by Janet D. Greene

The purpose of this study was to determine whether
or not syllable accent of polysyllabic words could be lip-
read. Thirty unskilled lipreaders and two talkers were
used in the study. Eighteen of the lipreaders were females
and twelve were males, all of whom were students enrolled
in speech courses at Michigan State University. The two
talkers were both graduate students in the Department of
‘ Speech and Hearing Science.

One talker was filmed while enunciating the sixty—
three syllable nonsense words, which were especially designed
for this study. These sixty, three—syllable, nonsense
words comprise the Test of Visual Accent Placement. The
other talker repeated aloud the number of each word before
it appeared on the screen. The lipreaders viewed the silent,
color film and responded by checking the syllable position
of the syllable which they thought was accented. This
was done individually for each of the sixty test words.

The score received on the test was determined by counting
the number of words in which the accented syllable had been

correctly identified.

Janet D. Greene

The data were statistically analyzed by t tests and by
analysis of variance. The raw scores of Test I (the first
half of the words as they appear on the film) and Test II
(the second half of the words as they appear on the film)
of the TVAP were individually subjected to t test. The
results of the analysis showed that lipreaders did equally
as well when accents were placed on a particular syllable
in accordance with the number of times that syllable
position is accented in three syllable words in the language
as when accents were assigned to a syllable arbitrarily.

On the basis of the findings obtained from the
statistical analysis the following conclusion was made:

It is possible for unskilled lipreaders to lipread
accent placement in polysyllabic words and to obtain a

score greater than would be expected by chance.

AN INVESTIGATION OF THE ABILITY OF UNSKILLED
LIPREADERS TO DETERMINE THE ACCENTED
SYLLABLE OF POLYSYLLABIC WORDS

By I

f.
11'.

Janet DIfGreene

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS
Department of Speech

196A

ACKNOWLEDGMENTS

To Mr. Richard Franks, who filmed the test, and

to Miss Bari Schoenkerman a special note of thanks is

extended.

ii

TABLE OF CONTENTS

ACKNOWLEDGMENTS

LIST OF APPENDICES

Chapter

I.

II.

III.

IV.

STATEMENT OF THE PROBLEM.

Introduction. .
Statement of Problem .
Hypotheses

Importance of the Study
Definition of Terms
Organization of Thesis

REVIEW OF THE LITERATURE.

Non—Visual Cues.

Visual Cues

Phonema Pressure
Discussion of Literature.

SUBJECTS, EQUIPMENT, MATERIALS, AND
PROCEDURE. . . .

Viewer Subjects.

Talkers

Equipment.

Materials. .

Test Situation . . . .
Procedure. . . . . . . .

RESULTS AND DISCUSSION

Results
Discussion

iii

Page
ii

|_J

CD NONU'IJZ‘NH

l2
l3

15

15
15
15
16
17
18

2O

2O
22

Chapter Page

V. SUMMARY AND CONCLUSIONS . . . . . . . . 24

Summary . . . . . . . . . . . . . 24
Conclusion . . . . . . . . . . . . 25
BIBLIOGRAPHY. . . . . . . . . . . . . . . 27
APPENDICES . . . . . . . . . . . . . . . 30

iv

Appendix
A.

B.

LIST OF APPENDICES

Three—Syllable Nonsense Test Words
Answer Sheet
Raw Scores

Floor Plan and Seating Arrangement

Page
31
3A
37
38

CHAPTER I

STATEMENT OF THE PROBLEM

Introduction

 

A need which remains paramount throughout life is that
of communicating with others. Many factors are involved
and contribute to the success or failure of this complex
act. One of the most disabling handicaps to communication
is a severe aural impairment. An individual with an aural
impairment must compensate for this disability by demanding
that his other sensory channels not only perform well the
intricate and highly specialized tasks for which they were
intended, but must also take on the added responsibility
of "hearing." For this individual, the eyes become one of
the tremendously important receptors by which the visible
component of oral language are received and transmitted for
decoding. Consequently, proficiency in lipreading ability
is an indispensable requisite as well as a valuable asset.

Lipreading involves an insurmountable number of
factors, one of which is determining which syllable of a
polysyllable word is accented. Our voices like everything

on earth, have to do with motion.1 We learn to speak with

 

lM. Faircloth, Lip Reading Study and Practice (Toronto:
The Ryerson Press, 19A6), p. 56.

l

rhythmic motion and stress our words and syllables according
to the rules of inflection.l The accenting of syllables
within a word is one of the determining factors of the
rhythm of the English Language. Moreover, the accenting of
the correct syllable is not a matter which can be arbitrarily
decided by each individual speaker because the rhythm of the
language demands that certain syllables of certain words be
accented. Therefore, it can be logically assumed that in
order for a lipreader to reproduce words with the customary
rhythm, especially words with which he is unfamiliar, he
must be able, through the use of lipreading, to determine

which syllable of a word is being accented by the speaker.

Statement of Problem

 

The purpose of this study is to determine whether
or not individuals with no previous lipreading training
; can, by lipreading alone, determine which syllable of a
trisyllabic word is accented; and if this is possible, to
what degree or extent are they proficient. In order to
investigate this problem a special test was devised. The
Test of Visual Accent Placement (TVAP) which was constructed
by the investigator especially for this study, consisted of
sixty nonsense words, each of which contained three syllables.
These words were filmed while being enunciated by one of the

talkers and this film was used as visual stimulus material.

lIbid.

Nonsense rather than actual words were used so that the
subjects would not gain cues because of their familiarity
with the word. If the subjects had viewed and recognized

a word with which they were familiar, their answer probably
would have been biased. In this case, they would have
given the correct response because they were familiar with
the word and knew which syllable should have been accented
rather than having determined the accented syllable by
lipreading.

Three—syllable words were used in order to minimize
the probable number of correct answers occuring by chance.
The nonsense words used were constructed by randomly
combining syllables which actually appear in authentic
three syllable words of the English Language.

The first half (1-30) of the nonsense words, as they
appear on the film, comprise Test I of the TVAP. The
second half (31-60) of the nonsense words, as they appear
on the film, comprise Test II of the TVAP (see Appendix A).
In Test I, the number of words having a particular syllable
position (first, second or third) accented is proportional
to the number of three syllable words which occur in the
English language that have that same syllable position
accented. These proportions were obtained by taking a
sample of five hundred, three—syllable words from the works
of five contemporary authors and computing the percentage

of accents received by each of the three syllable positions.

It was found that the first syllable is the syllable most
often accented, with the second syllable being the next in
prominence, and the third syllable receiving the accent
least often. In Test II all syllable positions were
accented an equal number of times. In all other respects,

for both tests, the position of accent was randomly determined;L

Hypotheses

 

The questions to be investigated and their corresponding
null hypotheses are as follows:

1. Do lipreaders score higher than would be expected
by chance on Test I of the TVAP?

Null Hypothesis: The mean score on Test I of the

 

TVAP is less than or equal to the chance score of 33.3 per
cent.

2. Do lipreaders score higher than would be expected
by chance on Test II of the TVAP?

Null Hypothesis: The mean score on Test II of the

 

TVAP is less than or equal to the chance score of 33.3
per cent.

3. Do lipreaders obtain different mean scores
between Test I and Test II of the TVAP?

Null Hypothesis: The mean score for TVAP, Test I

 

is equal to the mean score for TVAP, Test II.

 

1This refers to the test used which was especially
designed for this study by the author-—Test of Visual Accent
Placement.

A. Are scores of lipreaders on the TVAP distributed
with the same variance in Test I as in Test II?

Null Hypothesis: The variance of the scores on TVAP,

 

Test I is equal to the variance of the scores on TVAP, Test

II.

Importance of the Study

 

Rhythm is an important aspect of any language; the
English Language being no exception. The speech reader
must observe the rhythm, stress, and accent of speech.1
Close observation of these elements of the language in-
creases the aurally handicapped individual's proficiency
in speech reading. The purpose of speech reading training
is to restore the ability of an aurally handicapped
person to understand speech.2 Since syllable accent is a
determiner of the rhythm of language, it is hoped that

the results and_conclusions brought out by this study

 

will aid speech therapists and teachers of the deaf and hard
of hearing in assisting the aurally handicapped individual

to incorporate the customary rhythm of the language into

 

his speech. It is also hoped that this information can be

 

 

 

}? lM. Bethel Clifford and Robert G. Chaplin, Seeing and
99 Hearing, The Art of Lipreading (Published by the Michigan
Association for Better Hearing, Inc., Lansing, Michigan, n.d.).

2Boris V. Morkovin, "Rehabilitation of the Aurally
Handicapped Through the Study of Speech Reading in Life
Situations," Journal of Speech and Hearing Disorders, V. 12
(19A7), p- 363—368.

 

 

be used to aid the aurally handicapped individual in under-

standing the speech of others.

Definition of Terms
For the purpose of this study, the terms used are
defined in the following manner:

Accent or Stress.-—The means by which a syllable or

 

syllables are given their promenence over the unaccented
syllables in a word by a relative increase in the muscular
effort exerted during their production.1

Lip Reading or Speech Reading.—-The process which

 

enables an individual to determine what is being said
and which syllable of a polysyllabic word is accented by
utilizing only visual cues.

Talker.-—The person who is seen saying the words on
the filmed test.

Subjects.-—Those persons to whom the filmed test is
administered.

Syllable?—-A motor unit; each syllable is basically
a movement complex in which the larger, underlying movement
is the breath pulse, or thrust of the chest musculature,

which creates a compression of air in the lungs.

 

1L. F. Brosnahan, The Sounds of Language (Cambridge,
England: W. Heffer & Sons LTD., 1961), p. 175.

2R. H. Stetson quoted by R.M. S. Heffner, General

Phonetics (Madison, Wisconsin: The University of WiSconsin
Press, 1949).

 

Organization of the Thesis

 

Chapter One includes the statement of the problem,
the hypotheses and the corresponding null hypotheses to be
tested, the importance of the study, and definitions of
terms which will be used. Chapter Two contains a review
of the literature which pertains to this topic. In
Chapter Three a discussion of the subjects, equipment,
materials and testing procedures employed in conducting
this study is presented. The analysis of the data and a
discussion of the results of the study are brought forth
in Chapter Four. Chapter Five consists of the summary
and conclusions of the study with recommendations for future

research.

CHAPTER II
REVIEW OF THE LITERATURE

Non—Visual Cues

D. B. Fryl investigated the effects of duration and
intensity of vowels as they relate to linguistic stress.
Using six pairs of words in which a change of meaning,
from noun to verb, is accompanied by a change of strong
stress from the first to the second syllable, and using
twelve American speakers, spectrograms were made of the
words as they were uttered. Some of the test words were:
contract, subject, and digest. The measurements showed
that when the stress was shifted from the first to the
second syllable the most marked variation took place in
the relative duration and intensity of the "vowel" portions
of the speech wave, while other parts of the wave remained
remarkably constant in these respects.2 The test words
were recorded having the intensity and duration of the two
vowels vary over a set range. A test was made from this

material and a group of listeners was asked to judge, for

 

1D. B. Fry, "Duration and Intensity as Physical
Correlates of Linguistic Stress," Journal of the Acoustical
Society of America, V. 26, no. 1 (195A), p. 138.

2Ibid.

 

each test item, whether the noun or the verb had been uttered.
The results revealed that there is a good correlation between
the perception of stress and the vowel duration and vowel
intensity ratios.l

Phillip Lieberman2 conducted a study which was very
much like the study done by Fry. Lieberman used sixteen
speakers and twenty—five pairs of words which could either
be nouns or verbs, depending on syllable stress. He found
that fundamental frequency seemed to be one of the most
relevant unidimensional accoustic correlates of syllable
stress.

Gleason speaks of suprasegmental phonemes which are
stressed and intonations of pitch. One of the useful
aspects of suprasegmentals is that they can be used as
syntactic devices. The use of suprasegmentals as syntactic
devices is of primary importance to the linguist when he
is endeavoring to decide if two, or more, grammatical
constructions are comparable. If two grammatical construc—
tions have similar patterns of stress and intonation, they

may be compared.3

 

Ibid.

2Phillip Lieberman, "Some Acoustic Correlates of Word
Stress in American English," Journal of the Acoustical
Society of America, V. 32, no. A (1960), p. HBl.

 

 

3H. A. Gleason, Jr., An Introduction to Descriptive
Linguistics (New York: Holt, Rinehart and Winston, 1955),

p- 135.

10

Both phrase and syllable stress require careful
listening to be identified. Both have an extensive effect
on pronunciation and also play a very important part in
determining the over-all rhythm of American speech.l
Suprasegmental phonemes also aid an individual who is
engaging in conversation by denoting the emotional tone
of the conversation. An individual with a severe hearing
impairment cannot, by acoustical means, determine the
rhythm of the language, nor can he determine the emotional
tone of the conversation. Therefore, in order to gain this

information he must rely heavily on visual cues.

Visual Cues
The aurally handicapped individual readily learns to
make use of visual cues in understanding speech. O'Neill
and Oyer state that in this situation the eye is the primary
receptor, with the ear affording some slight assistance.
For the individual with a moderate to severe hearing loss,
the visual shape and movement of a speaker's articulators

become the important communicative elements.2

 

lJames Carrell and William Tiffany, Phonetics: Theory
and Application to Speech Improvement (New York: McGraw—Hill
Book Co., Inc., 1960), p. 25“.

2John J. O'Neill and Herbert J. Oyer, Visual Communi—
cation for the Hard of Hearing (New Jersey: Prentice Hall,
1961), p. l.

 

 

11

Stone1 investigated the effect of facial cues in
lipreading. The three variables which he took into consider—
ation were: lip mobility, facial expression, and facial
exposure. The results of his study revealed that lip
mobility had the most pronounced and consistent effect.

It was concluded that mobile lips were easier to lipread than
tight ones. The next variable, in order of importance, was
facial expression. Here, it was concluded that a grim

facial expression is easier to lipread than is a smiling

one. Facial exposure was the least important of the three
variables. However, it is thought that the greater the

area of face exposed, the more successful lipreading is
likely to be.

Hartman2 conducted a study of lipreading using eighty-
seven deaf students. The students watched two speakers
who were filmed. The mouths of the speakers were blocked
out. One speaker used greatly exaggerated facial expression
and the other speaker used only limited expressions around
the mouth and lower jaw. The students were able to correctly

identify 66 per cent of the sentences spoken by the speaker

 

lLewis Stone, "Facial Cues of Context in Lipreading,"
John Tracy Clinic Research Papers, V (December, 1957).

2N. Hartman, Quoted by W. J. Bechinger, "Convention to
Further Informing and Educating Teachers of Schools for the
Deaf, the Hard of Hearing, and Speech Defectives in Baden-
Wurttenberg, Heidelberg, January 9 and 10, 1961," Neue Blatter
fur Taubstummenbildung, 15 (1961), 168-184. Cited in dsh
Abstracts 2, No. 3, (1962), pp. 209—210.

 

 

12

with over—emphasized facial expression, while they were

only able to correctly identify AA per cent of the sentences
Spoken by the speaker who used limited expressions around
themouth and jaw.

Brannon and Kodman did a study in which they focused
their attention on the visibility of vowel and consonants
and vertical mouth Opening as factors influencing word
intelligibility. These factors were found not to be
significant, although they do state that the visibility
of the total—movement form of a word is the best cue that
the lipreader has in identifying it, and that the visual
identification of phoneme elements of words is directly

related to the place of articulation of the element.1

Phoneme Pressure

 

Black2 conducted a study which investigated the
amount of air pressure in the mouth when producing certain
consonants. Eight consonants and three vowels were chosen
to be used. The vowels were combined with consonants to
form syllables. Two instruments, a Killsman Altimeter
and a Hill stall-proximity indicator were used for measuring

the in—the—mouth air pressure. The results of this study

 

1John B. Brannon, Jr. and Frank Kidman, Jr., "The
perceptual Process in Speech Reading," AMA Archives of
Otolaryngology, V., 70 (1959), pp. 11A—119.

2John W. Black, "The Pressure Component In the
Production of Consonants," Journal of Speech and Hearing
Disorders, V., 15, (1950), pp. 207—210.

 

 

13

indicated that the four voiceless consonants that were
studied had greater amounts of air pressure than did the
four voiced consonants. The four voiceless consonants were
identified more often by the subjects than were the voiced
consonants.l

O'Neill also conducted a study of vowel and consonant
pressure and found that the sound pressure level was not
important in the visual recognition of vowels and
consonants, but it did appear that vision had the greatest
apparent effect on the identification of consonants, and
lesser effect, in order on the recognition of vowels,

words and phrases.2

Discussion of Literature

 

Although the literature was thoroughly searched,
this investigator was unable to find studies dealing with
the visual aspect of syllabic accent. The literature
contains studies on numerous facets of lipreading, some of
thich are related to the present study. Visual and facial
cues are always important factors of lipreading. The lip—
reader can discern many things from these cues, such as
what word is being spoken, the rate at which the word is

being spoken and even what syllable is being accented. The

 

lIbid.

2John J. O'Neill, "The Contribution of the Visual
Components of Oral Symbols to the Speech Comprehension of
Listerner," Journal of Speech and Hearing Disorders, V., 19

(195“): pp' E29-u390

 

1A

hearing participant in conversation determines syllable
stress by listening closely to the duration and intensity
of the vowels. Since this is the method used for detecting
syllable accent, the hearing impaired individual is at a
disadvantage, and must find some way to cope with this
problem.

Although great advancement has been made through the
years, more research is needed in order to determine what
factors or what cues gained from the speaker's face make
it possible for the lipreader to determine which syllable
of a polysyllabic word is being accented. This information
would be extremely useful to those who are concerned with
the education and rehabilitation of hard of hearing and

deaf individuals.

CHAPTER III

SUBJECTS, EQUIPMENT, MATERIALS,
AND PROCEDURE

Viewer Subjects

 

The experimental test population was composed of
thirty students enrolled in undergraduate speech courses
at Michigan State University. This group consisted of
twelve males and eighteen females. The ages of the subjects
ranged from eighteen to thirty-seven, with a mean age of
21.5 years. None of the subjects had had previous lip—
reading training; moreover, all subjects reported normal

or corrected 20/20 vision, normal hearing.

Talkers
Two adult females were used as talkers. One female
adult was filmed while enunciating the sixty nonsense
words which composed the test. The other adult female
said aloud the number of each word before it appeared on

the screen.

Equipment

 

Projector.——A Bell and Howell portable projector

 

(model 265A) with autoload was used to project the film onto

the screen. This projector was also equipped with a zoom

15

l6

lense which could be used to vary the size of the picture
on the screen.
Camera.—-The filming of the Speaker enunciating the
sixty monsense words was accomplished by using a Nikkorex 8,
portable, home movie camera equipped with an F 1.8 lense.
Screen.--A DaLite Versatol Screen was used. This
was a portable home movie screen which measured A0 x A0
inches.

StOp Watch.--A manually operated stop watch was used

 

to measure response intervals of six seconds.

Movie Film.—-Kodak, Kodachrome II color film was

 

used to make the test film.

Materials

 

Answer Sheet.——A three—page answer sheet was devised.

 

The format included labeled blanks for routine data such

as name, age, date, sex, and seat number; written instructions;
word numbers; syllable numbers; and response blanks (See
Appendix B).

Te§£.——The Test of Visual Accent Placement (TVAP) which
was constructed by the investigator especially for this study
consisted of sixty, three-syllable, nonsense words. These
words were constructed by taking a pronouncing dictionary
and randomly turning to a page; then upon finding the first
three-syllable word, the first syllable of that word was

recorded as the first syllable of a nonsense word. This had

17

to be repeated three times in order to construct one three-
syllable, nonsense word (See Appendix A).

Test Film.-—The filmed test was used as visual

 

stimulus material. The film was silent and in color. The
female talker was filmed while enunciating the polysyllabic,
nonsense, test words. Throughout the filming the talker
attempted to maintain a natural facial expression. The
filmed View of the speaker included only the area from the

head to the upper shoulders.

Test Situation

 

The room in which the experiment was conducted was
equipped with built—in, black—out shades, at the windows,
for movie viewing. The screen was situated in the front
of the room. The distance from the top of the screen to
the floor measured six feet and four inches. There were
three rows of chairs. The front row consisted of five
chairs which were placed eight feet away from the screen.
The back row consisted of four chairs whihh were placed
fourteen feet away from the screen. Each chair was equipped
with a writing arm. The projector was placed fifteen feet
away from the screen and was focused so that the picture
covered the viewing area of the screen in its entirety,

(See Appendix D).

18

Procedure

 

Before the subjects entered the room all of the
necessary equipment had been properly arranged and answer
sheets and pencils had been placed on the writing arm of
each chair. Once seated, the subjects were asked to fill
in the routine data at the top of the sheet. Next, they
were asked to read the instructions silently while the
investigator read them aloud. The instructions were
as follows:

You will view sixty (60) nonsense words all
of which have three syllables. Check the number
of the syllable which you think is accented. Do
this for each word. If you can not decide which
syllable is accented, make a guess. You must

have an answer for each item even if it is only a
guess. Are there any questions?

 

 

After the above instructions had been read, the questions
which the subjects had concerning what they were expected
to do were answered. Additional oral instructions were
given. They were as follows:

The number of each word will be spoken aloud
before the word appears on the screen. The answer
blanks are numbered. Make sure the blank in which
you respond has the same number as the number that
is spoken aloud. You will have six seconds to
answer.

So that you can get an idea of what the test
is like, we will have a short practice session.

The first and second words I will say aloud two
times, then I will go through only the lip movements,
and you will be asked to tell which syllable of that
word was accented. Now, for the first word (the
procedure described above was used for each word
individually).

19

The three practice words used were: amethyst, projector,
and happiness. The actual test words were, of course, not
real words.

The film was stopped after each word was projected
to allow six seconds for the subjects to respond. The
number of each word was spoken aloud by the female assistant,
before it appeared on the screen.

The film was shown three times in order to accommodate
all of the subjects. The test film was viewed only once
by each subject. Fifteen subjects viewed the first showing,
seven viewed the second showing and eight viewed the third

showing.

CHAPTER IV

RESULTS AND DISCUSSION

Results
The data consisted of a raw score for each subject.
The raw scores were tabulated by counting the number of
words in which the accented syllable had been correctly
identified by the subject (See Appendix C).

For the purpose of review, the null hypotheses that

were tested, as stated in Chapter One are:

l. The mean score on Test I of the TVAP is less
than or equal to the chance score of 33.3 per
cent.

2. The mean score on Test II of the TVAP is less
than or equal to the chance score of 33.3 per
cent.

3. The mean score for TVAP, Test I is equal to the
mean:$ore for TVAP, Test II.

A. The variance of the scores on TVAP, Test I is
equal to the variance of the scores on TVAP,
Test II.

For the purpose of this study the data were subjected

l

to t tests. The data were divided into two distinct sets.

 

lHubert M. Blalock, Social Statistics (New York; McGraw—
Hill Book Company, Inc., 1960), p. 1A8.

20

 

21

The raw scores from the TVAP, Test I composed the first
set, and the raw scores from the TVAP, Test II composed
the second set. A t test analysis was performed on both
sets of data, individually. The data were also subjected
to a statistical analysis of variance.

Mean Scores on Test I.--With twenty-nine degrees

 

of freedom, a t of 1.699 or greater is required in order

to yield significance in a one—tailed test at the .05

level of confidence. The mean for Test I of the TVAP was 70
per cent. The 3 statistic for Test I of the TVAP was 28,
therefore the following null hypothesis was rejected: The
mean score on Test I of the TVAP is less than or equal to
the chance score of 33.3 per cent. Therefore, lipreaders

do score higher than would be expected by chance on Test I
of the TVAP.

Mean Scores on Test II.-—With twenty—nine degrees of

 

freedom, a t of 1.699 or greater is required in order to
yield significance in a one—tailed test at the .05 level

of confidence. The mean for Test II of the TVAP was 70 per
cent. The t statistic for Test II of the TVAP was 18,
therefore the following null hypothesis was rejected: The
mean score on Test II of the TVAP is less than or equal to
the chance score of 33.3 per cent. Therefore, lipreaders
do score higher than would be expected by chance on Test II

of the TVAP.

22

Difference Between Means of Test I and Test II.——

 

The means for both, Test I and Test 11 were computed and
compared. The mean for the scores of Test I was 70 per cent
and the mean for the scores of Test II was 70 per cent;
therefore the following hypothesis was not rejected: The
mean score for TVAP, Test I is equal to the mean score for
TVAP, Test II.

Difference Between Variances of Test I and Test II.——
The variances for both, Test I and Test II were computed
and compared. The variance for the scores of Test I was
A9 and the variance for the scores of Test II was 121; with
twenty-nine degrees of freedom an F ratio of 1.90 or
_greater is required in order to yield significance at the
.05 level of confidence. An F ratio of 2.10 was obtained
after comparing the two variances, therefore the following
hypothesis was rejected: The variance of the scores for
TVAP, Test I is equal to the variance of the scores for
TVAP, Test II. Therefore, the scores of lipreaders on the
TVAP were not distributed with the same variance in Test I

as in Test II.

Discussion

 

The results of the statistical analysis of the data
indicates that persons who have had no previous lipreading
training can, by lipreading determine which syllable of a
three syllable word is accented. Since the mean score

for Test I was exactly the same as the mean score for Test II,

23

this indicates that the subjects as a group did not respond
differently between conditions when accents were assigned
to syllables arbitrarily and when accents were assigned to
a particular syllable in accordance with the number of
times that syllable position is accented in three syllable
words in the language. In both cases, however, the scores
made by the subjects were much higher than the score one
would expect by chance (33.3 per cent).

On Test II the variance was greater than it was for
Test I. At this time, only a hypothetical answer can be
given as to the cause of this. The hypothetical reason,
as concluded by the investigator is that accents on the
first syllable of a three syllable word are easier to
detect than are accents on the second and third syllable.
Since there were more words with the first syllable accented
on Test I of the TVAP than there were on Test II of the
TVAP, it is suspected that this, in part, accounts for the
smaller variance of the scores of Test I. It is also
suspected that it is possible to detect accented syllables
of words due to the size of the mouth opening and the

longer length of time taken to pronounce the vowel of the

accented syllable.

CHAPTER V

SUMMARY AND CONCLUSIONS

Summary
When auditory stimuli cannot be received by the

sensory organ most suited for that purpose, other sensory
organs must compensate for the disability. The eyes

become one of the very important channels by which auditory
stimuli are received. The aurally handicapped individual
must now depend on lipreading as a substitute for hearing.
He must be able to discern many details of speech and of
the language by closely observing the face of the speaker.

One such detail is that of determining which syllable
of a polysyllabic word is accented. The purpose of this
study was to determine whether or not persons with no
previous lipreading training can accomplish this, and if
so, how proficient are they.

Thirty subjects were used in this study, twelve of
which were males and eighteen of which were females. All
of the subjects were students enrolled in undergraduate
speech courses at Michigan State University. Two female
talkers were used, both of whom were graduate students at
the above mentioned university. One of the female speakers
was filmed while enunciating the sixty, three-syllable,

2 A '

25

nonsense words. The second female talker said aloud the
item number for each word before it appeared on the screen.

The silent, color, motion-picture film was viewed
by the thirty subjects. An answer sheet (See Appendis B)
was provided for each of them, on which they checked the
position of the syllable, for each word, which they thought
was accented.

One point was given for each correct answer. The
scores were analyzed by t tests and analysis of variance.
The results showed that individuals having no previous
lipreading training can lipread the placement of accent

in nonsense words of three syllables.

Conclusions

 

The statistical analysis of the data revealed that
lipreaders score higher than would be expected by chance
on Test I and on Test II of the TVAP. It also revealed
that the scores of the lipreaders on TVAP were not
distributed with the same variance in Test I as in Test II.

The findings obtained in this study were significant
at the .05 level of significance. Therefore, it seems
reasonable to conclude that unskilled lipreaders can lip-
read accent placement of trisyllabic words when tested

under the conditions used in this study.

26

Implications for Future Research

Information received from this study indicates that
research pertaining to the factors that enable skilled
and/or unskilled lipreaders to determine the accented
syllable of polysyllabic words would very likely yield
valuable information. For the purpose of comparison,
skilled lipreaders could be used in an investigation using
the same design as was used in the present study, since
this study was conducted using unskilled lipreaders in the
eXperimental population. An investigation of the ability
of skilled and/or unskilled lipreaders to determine what
word is being said when actual words are presented as
many times as they have syllables, with a different syllable
being accented each time the word is presented. seems
worthwhile.

In the present study the talker happened to be the
experimenter and it is possible that the experimenter
might have inadvertently and unconsciously overly stressed
the accented syllables during the filming of the test since
she was aware of the purpose of the study. Therefore, it
would seem beneficial to conduct a study using the same
design as was used in the present study, but without allowing

the talker to know the purpose of the study.

BIBLIOGRAPHY

27

BIBLIOGRAPHY

Books

Brosnahan, L. F. The Sounds of Language. Cambridge,
England: W. Heffer & Sons LTD., 1961.

 

Carrell, J. and Tiffany, W. Phonetics: Theory and Appli-
cation to Speech Improvement. New York: McGraw—
Hill Book Co., Inc., 1960.

 

Faircloth, M. Lip Reading Study and Practice. Toronto:
The Ryerson Press, 19A6.

 

Gleason, H. A. An Introduction to Descriptive Linguistics.
New York: Holt, Rinehart and Winston, 1955.

 

O'Neill, J. J. and Oyer, H. J. Visual Communication for
the Hard of Hearing. New Jersey: Prentice Hall,
1961.

 

Periodicals

 

Black, J. W. "The Pressure Component in the Production of
Comsonants," Journal of Speech and Hearing
Disorders, 15 (1950), 207-210.

 

Brannon, J. B. and Kodman F. ”The Perceptual Process in
Speech Reading,” AMA Archives of Otolaryngology,
70 (1959), 11A-119.

 

Fry, D. B. "Duration and Intensity as Physical Correlates
of Linguistic Stress,” Journal of the Acoustical
Society of America, 32 (1960), A51.

 

 

Hartman, N. cited by Bechinger, W. J. "Convention to
Further Informing and Educating Teachers of SChools
for the Deaf, the Hard of Hearing, and Speech
Defectives," dsh Abstracts, 2 (1962), 209—210.

 

Lieberman, P. "Some Acoustic Correlates of Word Stress in
American English," Journal of the Acoustical Society
of America, 32 (1960), A51.

28

 

29

Morkovin, B. V. "Rehabilitation of the Aurally Handicapped
Through the Study of Speech Reading in Life Situations,"

Journal of Speech and Hearing Disorders, 12 (19A7),
363-368.

O'Neill, J. J. "The Contribution of the Visual Components
of Oral Symbols to the Speech Comprehension of

Listeners," Journal of Speech and Hearing Disorders,
19 (195“), “29-439-

Reports

Stone, L. ”Facial Cues of Context in Lipreading," John
Tracy Clinic Research Papers, V (December, 1957).

Pamphlets

 

Clifford, B. M. and Chaplin, R. G. Seeing and Hearing: The
Art of Lipreading (Published by the Michigan Association
for Better Hearing, Inc., Lansing, Michigan, n.d.).

 

APPENDICES

30

APPENDIX A

Three-Syllable Nonsense Test Words

Kr ni-dr—jan
brILté—VIl
mI-wibn
dBnlrI-blo
glA—bCLtam
SimllI-Kast
ino—vavt
SKCLtIS-pQHL
pjuLmDr-tIv

I n—mdb :zu
miLtry—lalz
prIn—te-bId
hetvas—lv
aLpe—vant
KwEs—graf—poz
pleldferry
hiegII—Oia
bam—p 1&3~ -gard
ge-Kse—mas

I

Swa—lo—nI'S

31

32

Anlho-ru
flvlrai-lrs
Cdja—reKLta
magvprc—Sand
qchJILbl
ne—zalrin
prILrIl—tant
b 4731K nz—fIk
hju-bJLVIl
saILttn-E’K
Kwad—mg3-t8r’
aI—panljul
gar—mré—bi
mc—ta—faI’
ha-fjr-fal
KoLdJai—zu
ne’eslwnat
praI—tL—nas
sno—fju: 3
tcf ks-ta-Sralrz
mQL1I-pat
[yigz-tum

/
KrL-Ka-San

lI—ju-lan/

33

sfemIn—narz
dg—wIn—tr}
bjt-naz—avt
tI—er—rol
O—lILSIV

/
a3f-KI7—lam
Kavn-wa—navns
par-ka-Sant
sch—spo—van
OiLKAm—go
dJu—fgk-tran
mAs—fa—bra'
KfLmo-djrn
KwIn-ni-ras
aénrtist-viy

I
hol—bI-sIl

Name

blilllfli I

ANSWER SHEET

 

Sex

 

INSTRUCTIONS:

10.
11°
12.
13.
1A.

Age

 

Seat No.

23

.
b9

. “w“:nmv; I («‘5' van ‘-

Ibu will view sixty (60) nonsense words all of which

have three syllables.
which you think is accented.

@1983.

IIIIHHHHH 4

SYLLABLES

w!” Illfil~

HIH!

3A

Check the number of the syllabl:

Do this for each word,
If you can not decide which syllable is accented, mak: a

‘MY‘ .‘ W “‘1'?

a
i

i

Yen must have an answer for each item even if it
is only a guess. Are there any questions?

..,:_::::::::_::
:2::::::_:::_::?

,::.:::E::::: ,

15.
16.
17.
18.
19.
20.
21.
22.
25.
a.
25.
25.
27.
28.
29.
50.
31.
52.
55.
54.
55
56.
57.
58.

59.

A9.
50.

6

bk.)

HIHHHIIHIHHHI! '°

HHIH

m
m
m
M
m
m
w
M
m
m
cm
w
m
m
-0: “a

Scores

42
‘15
AT
55

A2
A}

47
35
A7
53
17

A1

A:
#1

M
’41

A)
. ‘7

APPENDIX(3

RAW SCORES

“01.4mm
d‘OfiJmO-fimug‘umk

N

19

19
55
18
2o
20
21
21
22
22
21

25

26
19
19

20
21

17

I9
20
21
20
57
21
20
20

‘a'fl
111112131111:fiwzxxquxzxxxwwwx-y'nmww

APPENDIX D

Floor Plan and Seating Arrangement

 

 

1::::Jscreen

 

X ‘x X Chairs
)< )1 )(

X X X

[:1 Projector

 

 

38