#
QP
306
16W'
*
• A
Mose's
:
·
.
Interbr
}
彎
​..
A
ations in
F
'
new method
tata:
1
1
pala lography
24
29
ÄsÄurut ketma

*
BRARY
GENE
#8
MICHIGAN
C Y
2131
ARTE
OF
UNIVERSIT
#
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
.
GP
دیا پیسے تو
رورد
ELBERT pomand
By
MOSES, Jr., M.Sc., Ph.D.
Ohio State University
Columbus, Ohio
EDWARDS BROTHERS, INC.
ANN ARBOR,
ARBOR, MICHIGAN
1940
Copyright 1940
ELBERT R. MOSES, Jr.
PRINTED IN U.S.A.
Litho printed by Edwards Brothers, Inc., Litho printers and Publishers
Ann Arbor, Michigan, 1940
этому
General Lib
Stock
1-23-40
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY1
by
Elbert R. Moses, Jr., M.Sc., Ph.D.
Department of Speech
Ohio State University
GP
306
Palatography, the study of tongue contacts upon the palate during
articulation, has been used for interpreting dynamic phonetics since its incep-
tion in the latter half of the nineteenth century. The early phoneticians,
J. Oakley Coles, Grützner, Techmer, Kingsley, Meyer, Lenz, Haeglin, Vietor,
Rousselot, Josselyn, Panconcelli-Calzia, contributed to the study or to the
techniques involved in the field of palatography; but no one contributed definite
findings in phonetics. Dr. Muyskens of the University of Michigan, however, in
1925 integrated his findings in palatography, and evolved from them definite
conclusions in dynamic phonetics. He also employed in his project a greater num-
ber of palatograms (diagrams of the tongue contacts on the hard palate) than those
used in the experiments of his predecessors, thus giving to the field of pala-
tography its most extensive study to that date.
I experimented also with new techniques in making positive casts and
pseudo-palates, and tried for a more accurate means of recording the contact
patterns made upon the pseudo-palate in its contact with the tongue.
C
After some preliminary experimentation, this particular project was
undertaken in the Speech Laboratory of the University of Michigan in 1933. It
involved the palatograms of nine students, each using the same set of twenty-four
sounds. The casting method then current in the University of Michigan Speech
Laboratory was used in these projects, that of making a positive cast from a
negative impression of Kerr's Compound, and into this swedging a sheet of ash
metal to form a pseudo-palate. (Cf. fig. 1) The reproduction of the image left
on the pseudo-palate was made on parchment paper and later blue-printed.
P
This experiment in the technique for preparing pseudo-palates was ac-
companied by the development of new methods of recording the contacts. I con-
ceived the idea of cutting the casts in half along a line from the line from
the division between the medial incisors to the soft palate, marking the con-
tact on this cut cast, and recording the variations of the same sound articulated
three times. (Cf. figs. 2, 3, and ff.)
1. The following article is an abstract of the doctorate thesis, Palatography, A Critical Study
and Analysis of Contacts, which was completed under the direction in partial fulfillment of
the requirements for a Ph.D. in the Horace H. Rackham School of Graduate Studies of the
University of Michigan.
G G
W
The author would like to set forth the advantages of dividing the cast
for measuring and for photographing for record. First, the lines would appear
separate in a photograph of the cast entire. This method does show more clearly
the actual contours which would be fore-shortened and distorted in a ground-plan
photograph, instead of the side-view which dividing permits. Second, the cross-
section indicated makes it clear how the pressure will spread the mass of the
tongue.
-
2
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
Two more separate projects resulted from this decision to work with split
casts. In Project II thirty pairs of sounds articulated by three individuals
were recorded on plaster casts, each sound being repeated three times. These
sounds were sonants and surds preceded by and preceding vowels. Project III, at
the suggestion of Dr. Muyskens, is a direct check upon his study. The sounds
which he employed were investigated with the new technique of recording the con-
tact.
The Statement of the Problem
From these three projects, the blue-print one of 1933-1934, and the two
more recent ones employing plaster casts, evolved this problem, which may be
stated thus:2
1. With the use of a new technique in recording the contact, would those
statements made by Dr. Muyskens at the conclusion of his thesis hold true?
These I prepared to check. Dr. Muyskens made the following statements:
a. Contacts made by a subject are specific and are the result
of a movement touch in a fixed habit.
b. They lie within small and fixed boundaries. Variations are
due to tonicity variations.
c. Contacts for voiced consonants have greater energy coefficients
than do the corresponding contacts of voiceless consonants.
d. Contacts for consonants preceding the vowel are more regular
and larger than the same consonants after the vowel.
e. The vowel (i) preceding a consonant has a narrower, closer
contact and is more regular than when following the consonant.
f. If the vowel (a) precedes a consonant, the contacts show less
area then if following the consonant. The flabbier tongue,
however, may make a broader contact. The mean percentage of
the relative contacts before and after the vowel is 78.22 with
contacts preceding the vowel greater.3
2. In addition to the above, I wished to get data regarding several
other features of palatography:
a. In making a series of sounds, each sound being repeated three
times, would the sonants or surds in some of these sounds show
less variation in the contact; that is, would the sonant have
greater contact that the surd in one trial, and less contact
that the surd in another trial?
b. Would sonants or surds have greater variation in movement pre-
ceding or following the vowel?
A problem of this nature, involving a new technique in recording the
contact area, required, too, new techniques in the pouring of plaster casts, and
in making permanent and easily handled records of the project through the medium
of photography.
The following method was used in attacking the problem:
Palatography Technique
The old method of taking a negative impression with Kerr's Compound and
from this pouring a positive one was found impracticable for a project requiring
2. I have used 888 blue-printed palatograms and 375 photographic cuts of casts showing contact
variations.
3. John H. Muyskens. Thesis. The Hypha, pp. 45-46. University of Michigan, 1925.
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
useȧ:
3
a great number of casts. It became necessary, therefore, to devise some method
by which numerous plaster casts might be made from one negative impression. Two
new methods were tried and used; one, however, the glue-mold process, proved very
satisfactory and inexpensive - the method used by sculptors.
the method used by sculptors. A negative impres-
sion of Kerr's Compound was made, a positive cast run up, and from this a nega-
tive mold made of glue. These glue-molds can be used indefinitely for pouring
positive casts. The method was found to be quick and accurate.
For the pseudo-palates, ash metal #7 was used.
In Projects I and II the following vowel and consonant combinations were
Project I:
(The symbol "a" is used for the phonetic symbol "ae" in all
examples throughout this study.) ad, at, da, ta, ɛd, ɛt, dɛ, tɛ, id,
It, dr, ti, ag, ak, ga, ka, ɛg, ɛk, gɛ, kɛ, ig, ık, gı, kı, az, as, za, sa,
ɛZ, ɛs, zɛ, sɛ, IZ, IS, ZI, SI, a 3, aS, 3a, Sa, ɛz, ɛs, zɛ, Sɛ, 13, IS, 31,
SI, la, al, lɛ, ɛl, lı, ıl, ra, ar, rɛ, εr, ri, Ir, ab, ap, ba, pa, ɛb, ɛp, bɛ,
pɛ, ib, ip, bi, pi, av, af, va, fa, ɛv, ɛf, vɛ, fɛ, iv, if, vi, fi, að, að, da,
θα, εδ, εθ, δε, θε, 1δ, 1θ, δί, θi, ab, ba, ap, pa, eb, be.
Project II:ɑd-at,da,ta, ɛd, et, dɛ, tɛ, Id, it, dï‚tî, ag,ak, ga, ka, ɛg,ɛk.
gɛ, kɛ, Ig, Ik, gı, kı, az, as, za, sa, ez, ES, Zɛ, Sɛ, IZ, IS, ZI, SI, âz, ds,za,sa,
Ez, ES, Zɛ, Sɛ, IZ, IS, az, as, za, Sa, e3, eS, 3€, Sɛ, 13, 1 S, 31, Sĩ, la, al, le, el, lï,
Il, ra, ar, rε, er, ri, ir.
rɛ,
In Project III the following vowel and consonant combinations were used:
az, as,
ak,
Ja, Da, da, ta, ga, ka, za, sa, að, a Q, ad, at, ag,
Ji, Qi, di, ti,
gi, ki, zi, si, ið, i☺, id, it, ig, ik, iz, is, (Note: The symbol "a is used for the phonetic
symbol "œ" in all examples throughout this study.)
Each sound was articulated three times by each person, each particular
image being recorded directly on the plaster cast in a different color (black,
first trial; red, second trial; blue, third trial). In transferring the contact
pattern from the pseudo-palate to the positive cast, the pseudo-palate was placed
in the palate of the cast; certain directional marks were made on the cast; and,
with the aforesaid marks as guides, and calipers for measuring the width of
contact, an outline of the contact pattern was drawn upon the positive cast.
After all contact patterns had been recorded on all the casts, each cast
was cut in half along a line from the division between the medial incisors fol-
lowing the median line of the hard palate.
mo
In order that a useable and permanent record of these contact patterns
with depth might be made, the casts were photographed. The pictures were made as
near life size as possible. In Project II only one half of the cast was photo-
graphed; while in Project III both the right and left sides, or halves, were
photographed.
4
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
な
​Calculations were made from the plaster casts in the following manner:
1. For each phase of palatography which was to be studied, a special
check sheet was prepared. On this sheet were written the sounds
employed, and the letters and marks pertaining to the individual
subject.
2. All estimates were recorded on these sheets and totaled.
3. The sheets then were arranged according to projects, there being, in
some cases, duplicate sheets included in the various projects.
4. When some problem included material in all three projects, a special
tabulation sheet was prepared, and on this was made a summarizing
check.
5. From these tabulation sheets and their results the necessary deduc-
tions were made to prove or disprove those questions in palatography
which were made the objects of research at the beginning of the
problem.
4
In each project estimates were made to determine (1) whether sonant
contacts are greater than surd contacts (2) whether consonants before the vowel
have greater contact than consonants following the vowel (3) whether there is
any difference in the amount of variation of contact pattern of consonants
articulated before and after the vowel.
The results of the three projects may be summarized as follows:
1. The results are not significant enough to warrant a definite
statement that surds have greater contact areas than sonants.
In Project I, 148 sonant contacts are greater than the
corresponding surd contacts; 179 surd contacts are greater
than the corresponding sonant contacts; and there are 18
indeterminate contacts.
In Project II, 94 sonant contacts are greater than the
corresponding surd contacts; 113 surd contacts are greater
than the corresponding sonant contacts; and there are 5 in-
determinate contacts.
P
In Project III, 76 sonant contacts are greater than the
corresponding surd contacts; 68 surd contacts are greater than
the corresponding sonant contacts; and there are 8 indeterminate
contacts. (For surd contact patterns Cf. figs. 2, 3, 4, 8, 10,
13, 15, 16, 17; for sonant contact patterns Cf. figs. 5, 6, 7, 9,
11, 18, 19, 20.)
2. Consonants after the vowel have greater contact areas than con-
sonants before the vowel.
-
In Project I, 206 consonants after the vowel have greater contact
areas; 74 consonants before the vowel have greater contact areas; and there
are 36 indeterminate contacts.
4. (a) This particular problem is based on a study of English sounds only.
(b) The palatograms of Project I were chosen at random by a class of speech students and one
subject among these (no. 3) did not have a normal speech pattern (had a cleft palate
correct ).
(c) One subject of Project III (S) did not have a normal speech pattern (stutterer)
1
In Projects II and III, 135 consonants after the vowel have greater
contact areas; 74 consonants before the vowels have greater contact areas; and
there are 12 determinate contacts.5 (Cf. figs. 2, 3, etc.)
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
3. The vowel (i) preceding labial consonants has a narrower, closer space
between side contacts than when following the same consonants.
4. Contacts made by a subject lie within a small and fixed zone.
From a list of sounds (28 out of a total 46, containing 19 surds and 9
sonants) from Projects II and III, it has been found that surds showed greater
coincidence in their contact patterns over sonants in this list. (Cf. figs.
2, 3, 4, 5, 6, (da), 8 10, 17 (Ik), 18 (Ig), 20 (Ig).
In addition to the above checks, a measurement of the variations in
contact was made to the nearest millimeter. It was found that ga, gI have more
variations variation more than 1 mm, in two out of three repetitions.
(cf. figs. 7, 18, 19, 20).
G
-
SUMMARY
270 sounds out of 372, however, have contacts within 1 mm. in two out of
three repetitions.
5. In certain sound combinations the statement that consonants after the
vowel show greater variation in contact than consonants before the vowel is con-
tradicted.
There is greater variation in ga than in ag (Cf. fig. 7); in kI than IK
(Cf. figs. 15, 16 17).
A new method which involves dividing the marked cast made possible a
careful check of all palatograms and casts. The following statements can be
made:
5
1. There is not enough evidence to warrant the definite conclusion
that surds have greater contacts than sonants.
2. Evidence shows that there is a tendency for consonants after the
vowel to have greater contact than the same consonants before
the vowel.
3. The vowel (i) before a labial consonant has a narrower, closer
space between side contacts than when following the same consonant.
This fact agrees with Muyskens' conclusion. p. 2.
4. The contour of the contact pattern lies within a very small zone.
This fact agrees with Dr. Muyskens' conclusion. p.2.
The following contact patterns are exceptions to the above state-
ment: ga, gl, kl.
5. The same sounds were compared. It is evident that in Project I there is a definite
correlation between the larger contact after the vowel sound, and a greater tonicity
variation when the consonant follows the vowel.
P
6
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
FIGURE 1
Full casts from left to right E-E', B, M, with pseudo-palates set into the casts.
FIGURE 2
Cast E', including both left and right halves of the cast, shows that the contour
of the contact pattern lies within a very small zone in three trials. The total
contact lies between the lingual surface of the teeth and the contour line.
FIGURE 3
Cast E', including both left and right half of the cast, shows that the contour
of the contact pattern lies within a very small zone in three trials.
total contact lies between the lingual surface of the teeth and the contour
line.
The
FIGURE 4
Cast M shows that the contour of the contact pattern lies within a very small
zone in three trials. The total contact lies between the lingual surface of
the teeth and the contour line.
!
FIGURE 5
Cast E', showing both left and right half of the cast, indicates that the
contour of the contact pattern lies within a very small zone in three trials.
The total contact lies between the lingual surface of the teeth and the con-
tour line.

3
Figurel
ta
at
Fig.
da
Fig. 7
Figla
at
Fig 5
E'
Figure 1
Figure 1
E'
E'
Figure 2
Figure 3
Figure 4
Figure 5
E
P
É
Fig 2.
Fig.3
ta
Fig. 5
at
M Figit
ta
de
8
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
FIGURE 6
Cast E shows that the da contour patterns made at different times conform
to the E' contour patterns. The ad in comparison with the da shows that there
has been more pressure used in making ad. Consonants after the vowel show
greater contact than those before the vowel. The da contact patterns lie
The total contact lies between the
within a very small zone in three trials.
lingual surface of the teeth and the contour line.
FIGURE 7
Cast M shows that ga has more variation in contact pattern than ag. There is
a cross of contour lines in ga which is characteristic of contour patterns of
g.
Cf. Figure 19. The total contact lies between the lingual surface of the
second bicuspid, the first and second molars, and the contour line.
FIGURE 8
Cast M shows that ak and ka do not show any variation in the three trials, but
do show that the contour of the contact pattern lies within a very small zone
in three trials. The total contact lies between the lingual surface of the
first and second molars and the contour line.
FIGURE 9
The total contact lies
Cast E' shows both left and right side of the cast.
between the lingual surface of the teeth and the contour line.
FIGURE 10
Cast E' shows the vowel (i) following the consonant t. The total contact
lies between the lingual surface of the teeth and the contour line.

کیا
M
Σ
di
ti
Fig. 6
Fig. 1
Fig:8
ag
ак
Fig. 9.
ad
Figo
E*
Figure 6
Figure 7
Figure 8
E
Figure 9
E
Figure 10
E
M
M
Fig. 6
Ů
Fig. 7
Fig. 8
Fig.
Fig 10
da
ga
Ka
di
ti
10
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
FIGURE 11
Cast E' shows the vowel (1) following the consonant z. The total contact lies
between the lingual surface of the teeth and the contour line.
FIGURE 12
Cast E' shows the vowel (1) preceding the consonant d. It also shows that
there is a tendency for consonants after the vowel to have a greater contact
than consonants before the vowel. Cf. Figure 9. The total contact lies be-
tween the lingual surface of the teeth and the contour line.
FIGURE 13
Cast E' shows the vowel (1) preceding the consonant t. It also indicates that
there is a tendency for consonants after the vowel to have a greater contact
than consonants before the vowel. The total contact lies between the lingual
surface of the teeth and the contour line.
FIGURE 14
Cast E' shows the vowel (1) before the consonant z. The total contact lies
between the lingual surface of the teeth and the contour line.
FIGURE 15
Cast M shows the vowel (I) before and after the consonant k. The total
contact lies between the lingual surface of the second bicuspid, the first
and second molars, and the contour line.

Zi
id
it
iz
Ik
Fig. "
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Σ
Ε
E'
لیا
E
E'
Figure 11
E'
Figure 12
E'
Figure 13
F
E'
Figure 14
Figure 15
Fig. 11
Fig. 12
Fig. 13
Fi5-14
KI
Zi
id
it
iz
Fig 15
M
12
INTERPRETATIONS OF A NEW METHOD IN PALATOGRAPHY
FIGURE 16
Cast E shows less variation in the contour of the contact pattern of I with k
in three trials than with g in Figure 19. The total contact lies between the
lingual surface of the second bicuspid, first, second, and third molars, and
the contour line.
.
FIGURE 17
Cast B shows a narrow, deep palate. It also shows that the contour of the
contact pattern with Ik lies within a very small zone, and with kI in two
out of three trials. kI shows also greater variation in the contour of
the contact pattern. The total contact lies between the first and second
molars and the contour line.
FIGURE 18
Cast B shows that the contour of the contact pattern with Ig lies within a
very small zone. There is greater variation of the contact pattern in gI.
The total contact lies between the second bicuspid, the first and second
molars, and the contour line.
FIGURE 19
Cast E. The g causes more variation in the contour of the contact pattern
than does k. Cf. Figure 16. The g before the vowel (I) causes more varia-
tion than when it comes after (I). The total contact lies between the first
and second bicuspids, the first, second, and third molars, and the contour
line.
FIGURE 20
Cast M. With Ig the contour of the contact pattern lies within a very
small zone in three trials. With gI there is more variation in the
contour of the contact pattern. Cf. Figure 19. The total contact lies
between the second bicuspid, the first and second molars, and the contour
line.

Ik
IK
Ig
19
Ig
Fig. 16
Fig 17
Fig 18
Fig. 19 E
Fig. 20
B
M
E
Figure 16
Figure 17
Figure 18
Figure 19
Figure 20
KI
KI
JI
SI
91
Fig: 16
Fig. 7
Fig. 18
Fig. 20
E
M
B
Fig: 19 E
B
BIBLIOGRAPHY
Muyskens, John Henry. The Hypha. Thesis, University of Michigan Library,
1925.
Rousselot, L'Abbé, P. J. Principes de Phonétique Expérimentale.
Edition. Paris, 1924.
Nouvelle
Scripture, Edward Wheeler. The Elements of Experimental Phonetics.
Charles Scribner's Sons. New York, 1902.
Vietor, Wilhelm. Kleine Phonetik.
0. R. Reisland.
Leipzig, 1903.

A
A
100

BOUND
FEB 6 1942
UNIV. OF MICH
LIBRAR
·
D
..
}
...
**
•
...
•
?
1
!
F
2
•
S
V
7

1.





$
fot
S
xxx
S
P
That

....
***
SPAX
AA
.....