Neuropsychologia, 1971, Vol. 9, pp. 317 to 323. Pergamon Press. Printed in England
AUDITORY
PERCEPTION BY NORMAL JAPANESE "OF THE SOUNDS "L" AND "R"
ADULTS
HtROMU GOTO* Chikamori Hospital, Kochi, and Faculty of Medicine, Tokushima University Japan, (Received 11 iVlarch 1971)
AImtraet--1. Listening to words tape-recorded in the English language by both Americans and Japanese, the same eight (8) Americans and eleven (11) Japanese subjects were tested to see if they could distinguish which words had "L" and which words had "R". 2. The Japanese subjects were inaccurate in their auditory discrimination between "L" and "R" even when they heard the voices of the Americans. So they were considered to be unable to distinguish "L" and "R" by hearing. 3. The Japanese subjects whose English language ability was best had relatively good pronunciation of "L" and "R" compaired to their poor auditory discrimination. 4. The auditory discrimination of syllables is possible through practice in speech and hearing from early childhood. Without such practice, however, there is considerable difficulty in discriminating them. These findings indicate the importance of kinesthetic sensation in the subject's own speech when learning to articulate. IT MAY be said that iri sensation there are somewhat different yet correlative processes of sensation, perception and recognition. This may also be true in the case of auditotff sensation. Disregarding the semantic character of words, I would like to consider spoken sound as an element of a word, and especially the syllable or phoneme as a unit of spoken sound [1, 2]. F o r example, we can distinguish the sound " h a " f r o m " k a " . But are these two really so different when considered as sound? Can so called "congenital aphasiacs" who have scarcely spoken a word since birth, such as the cases investigated by WOrtSTERDROUGHT and ALLEN [3] and by L~NNE~RG [4], or primitive peoples whose language is undeveloped (if they still exist in the world today), or primates of higher intelligence trained by conditioned reflex, distinguish between the sound " h a " and " k a " ? When we hear the sound " h a " and then hear it again after a pause, we must m a k e considerable allowance to admit that the two sounds are the same syllable. This is because the repeated pronunciation of a sound by one person is not the same as the first time he pronounces it. Also, different people pronounce the sound " h a " differently. On the other hand, we must very clearly distinguish the sound of such syllables as " h a " and " k a " one f r o m the other. In fact it is easy and simple to distinguish the syllable " h a " from " k a " . In the Japanese language there are no particular syllables which are difficult to distinguish one f r o m the other. Anyone will grant that it is only natural when hearing two different syllables to recognize them as different. But judging f r o m the results of the following tests it seems that we can say that a certain training or practice f r o m the time of early childhood is necessary to be able to do this. * Address of author: 3-22, 2 chome, Shioyazaki, Kochi-shi, Kochi-Ken, Japan. 317
318
HmOMU GOTO
I n t h e J a p a n e s e l a n g u a g e t h e r e is a n i n t e r m e d i a r y s o u n d b e t w e e n t h e " L " a n d " R " o f E n g l i s h speech. It is said t h a t t h e J a p a n e s e are n o t g o o d at p r o n o u n c i n g " L " a n d " R " . N o w t h e q u e s t i o n is w h e t h e r o r n o t we J a p a n e s e c a n d i s t i n g u i s h " L " f r o m " R " w h e n it is e n u n c i a t e d b y n a t i v e s p e a k e r s o f E n g l i s h . M o s t p e o p l e h a v e t h o u g h t , t h a t we c o u l d c l e a r l y d i s t i n g u i s h t h e m since t h e n a t i v e t e a c h e r s w o u l d n a t u r a l l y e m p h a t i c a l l y differentiate them. METHOD
OF TESTING
"lest I The subjects were 8 Americans and 11 Japanese. Each subject tape-recorded the following 16 words. They were advised to pronounce each word as correctly as possible.
Japan
collect (correct)
Dick
play (pray)
save
lead (read)
thank
glass (grass)
when
flesh (fresh)
wash
bleed (breed)
visit
light (right)
zoo
(splint) (sprint)
Each even numbered word has " L " or " R " in it. For example, the words recorded by one subject would be: Japan, collect, Dick, pray, save, read, thank, g l a s s . . , splint. Then each subject listening to his own voice or to the recorded voices of others was asked to identify " L " or " R " in the even-numbered words, thus giving eight answers for each subject's voice. In the case of the above example, the following are the correct answers: 1, L; 2, R; 3, R; 4, L ; . . . 8, L. ~' The answers of a subject could range from 8/8 (full score) to 0/8 (zero score), though in fact no one scored 0/8. To make it easier to understand the value of the fractions, the answers are shown in the form of a circular graph on Table 1. The white portion of the circle shows the proportion of success. A perfect white circle indicates a full score. The letters along the top of Table 1 indicate the subjects who recorded their voices, while the letters along the left side of the Table indicate the subjects who identified the recorded voices. The double circles indicate the scores of the subjects identifying their own recording. The 8 subjects A to H are native Americans. All the others are Japanese. The scores of P1, P2, and P3 are all of the author, and are designed to examine how English lessons could effect the scores in these tests. P1 shows my score after one month of studying English conversation. P2 is after four months of study, and P3 is after ten months of study. Table 1 shows that most of the American subjects correctly identified the recordings of the other Americans. If we look carefully, however, we see that the pronunciation of some Americans presented a slight difficulty for other Americans. The pronunciation of subject F was mistaken once by four (4) other Americans. And subject F mistook the pronunciation of subject E. But all these mistakes were between "right" and "light". So it would seem that these mistakes are due to a peculiarity in speech in subject F. Subject E was mistaken once by both subjects G and H, the difference being in the distinction between "correct" and "collect". It is interesting to note that 4 of 5 Americans who answered got scores of 5/8 on subject 0 (Japanese subject). Their answers were the same for each word pronuonced by subject 0. For example, "correct" was correctly identified as " r " by all four. In the case of Japanese subjects, their answers were unstable and changing. Four (4) Japanese subjects were tested twice successively, but their answers in the second testing were different from their answers of the first testing. Among their eight (8) answers to each recorded voice, the degree of agreement between the the first and second testing was as follows: 7/8 was most frequent, 8/8 and 6/8 was relative and 5/8 was less frequent. This same tendency was found in each subject whether his answers were correct or not. It may be said then, that the mistakes between Americans described above, may be accepted as proof of the accuracy and sharpness of Americans in distinguishing the sounds " L " and " R " . Of the Japanese subjects, six (subjects I to N) were excellent in English conversation. Two of them had spent several years in America. The pronunciation of these six was correctly identified fairly well by the Americans as is shown in the middle upper part of Table 1. Bur their auditory perception of the Americans' voices was inferior. (The scores 8/8 and 7/8 may be admitted as the correct rate of discrimination.) Therefore, these six Japanese subjects have good articulation but poor hearing ability.
319
AUDITORY PERCEPTION BY NORMAL JAPANESE ADULTS OF THE SOUNDS " L " AND " R "
Table I A
B
C
O
Amencons ~ ~ E F G H I
Japenese J K L
M
N
0
P3 Pz P,¢ Q
R
S
"~.,O000, x~x x x O O x x x O x x x ~I x ~1(11 8000 x O ~ O O O O O ~ O ~ O ~ x x xx 00000'~00 O = O 0 0 O 0 (~ db ~ ( 1 1 0 d D ~ o O000!x~x x x O ! ( D x x x CD x x x ( ~ ( l l , x I~ ~ 0 0 O O O ( : D O O O O ~ ( : D O 0 ( ] I , O ~ I , ( ] I ~ I , ; ~ i~ ~ O 0 0 0 ~ O x x ~0 x x ~ ~ x ~~1(11 ~O 0 1 0 x l ~ O O O O ~ O ~ ~ ~ x x x x [ __. C O 0 x ( D O 0 x 0 O 0 0 ! x ~ x 0 x x x x x ' [ - T c ~ dl, t j ,~ ~ ! ~ x i ~ i i l i l l ~ O ~ i , : : : i l ~ ~ x MO~O(II'X ,', ¢ ¢ ~ @ ¢ ¢ X
', ¢ ¢
x x xx
X X X ~ X xI@O~ X X ~ ~ X ¢ ~X X I ¢ ~ , X ~ ~ ¢ ~
(9 ¢ X X X X (!I ~ X X X 1¢t X I X X I ¢ (111¢~¢
R CIl~x
x x x x x ~ x
s ~¢~1~1C~
x x ¢0
x CII xl:x x x l x x x ( i l
x x C~¢lIC~lx C~te~l,,C~l(ll
The other Japanese subjects from 0 to S were poorer in articulation and hearing. They are all university graduates and supposedly understand English books very well. However they are poor in English conversation. In the case of PI, P2, and P3, which are my own scores, I had studied English conversation very diligently for several months, because I had to make a very important speech in English in New York. It is possible that the Americans were able to quite correctly distinguish my pronunciation in P3 because I purposely pronounced the sound " R " like a " W " . But my own hearing ability in P3 did not show much progress when compared to my earlier scores. Most of the Japanese could not even clearly identify their own pronunciations. Tfst H
Because of the poor auditory discrimination between " L " and " R " of the Japanese subjects, the following test was given. Five (5) Americans recorded the same words as Test 1, but sometimes the same word was repeated and sometimes the " R " and " L " sounds were changed as in the following example: 1. lead and lead 2. right and light 3. glass and grass 4. breed and breed 5. pray and play 6. collect and collect 7. fresh and fresh 8. splint and sprint The five (5) Americans and the five (5) Japanese who answered were all subjects of Test I. The letters indicating the subjects on Table 2 are the same as those used on Table 1 and indicate the same subjects. If they judged the words to be the same, for example "lead and lead" (L and L), they were told to w~Ste " s " (the same). If they judged the words to be different, they were instructed to write " d " (different).
HIROMU GOTO
320
If a Japanese subject wrote as follows to an American's recording of the above words, he would get a full score: 1. L,L, s; 2. R,L, d; 3. L,R, d; 4. R,R, s; 5. R,L, d; 6. L,L, s;7. R,R,s; 8. L,R, d. The words with "L" and " R " shown above were arranged differently for each of the five (5) Americans who recorded, so the Japanese subjects had to judge by hearing alone whether the words had " L " or " R ' . We called this the "Lead and Lead Test" (named from one of the four possible combinations). The scores of this test are shown on the second line (L a LT) of each subject on Table 2. The scores of the " L or R Test" (shown as LRT on the first line) were borrowed from Table 1 for purposes of comparison. It was explained to each subject that the answers of "s" and " d " were more important, and that he might be allowed to write as follows: L.L. d, or R.L. s, etc. Table 2
- ~ Recorders Answers ~ . LRT
LoL7 SOT
Arnericons B I c l
E
G
!
~
©i0@(~ 0 ~(~(~0 ~ (~ (~ 0 01(~
LRT K
LoL.T
(~ (11 ~ (11 (iI
SDT LRT L
LoLT SDT L.RT
O
(~ (~ 0 (~ (~ (~ (~ 0 (~ (~ (~ (~ 0 (11 @
Lo L T
S.D.T, LRT. P3
0 (~ (~ (~ (11
LoLT. SD.T.
The purpose of this test was not to test the ability of Japanese subjects in English, but to test their ability of auditory discrimination between one syllable and another. So the following answers of a given subject whom we can call B, might be considered as partially correct discrimination of the voice of a subject whom we will call A. A: 1. lead and lead . 2. right and light 3. . . . 8. splint and sprint B: 1. R , R , s 2. L , R , d 3. . . . 8. R , L , d From this point of view, the answers limited to "s" and " d " were analysed separately ("Same or Different Test") and are shown on the third line (SDT) for each subject in Table 2. Among the mistakes in SDT of all the subjects there was scarcely any difference in frequency between the two types of misjudging, i.e. mistaking "'d" for "s" or mistaking "s" for "d". These results demonstrate that we Japanese have poor auditory discrimination ability as regards the sounds "'L" and " R " .
AUDITORY PERCEPTION BY NORMAL JAPANESE ADULTS OF THE SOUNDS " L " AND " R "
321
DISCUSSION There are many Japanese who can understand well books written in English, but usually, they are poor in English conversation. This is a common trait of Japanese people learning such foreign languages as English, French, or German, and is well known to Japanese teachers. Most Japanese from their childhood have never heard the sounds " L " , " R " , and "V". If we hear these sounds in our adolescence or adult age, it is probably too late for us to be able to distinguish them clearly. To be considered as having accurate discrimination of these sounds, the subject should achieve a score of 8/8, or at least 7/8. Conversely, a total inability to discriminate would be shown by a score of 0/8. The Japanese subjects had various degrees of difficulty in discrimination (from 6/8 to 1/8), though none completely failed. The scores in Table 2 show clearly that Japanese have poor auditory discriminatory ability even for difference between " L " and " R " sounds. The Japanese subjects with best performance (subjects I to N) could articulate " L " and " R " relatively well. It would seem that they articulated the sounds relying on the kinetic sensations of their own speech organs as deaf persons would do. That is to say, they must find it difficult to discriminate their own pronunciation of " L " and " R " by auditory sensation alone. It is possible that most Japanese people pronounce " L " and " R " without distinction because they do not have a feedback system with accurate auditory discriminator to control their articulation. In the tests reported here, different from usual conversation, there was the disadvantage of not being able to read the lips o f the speaker. The words in my test were in a determined order and they were pronounced carefully. Though the subjects had only to pay attention to one point of " L " or " R " in each word, they still got poor scores. It was found by checking the scores that some Japanese subjects got better scores on the voices of the Americans with whom they had had close contact. In the cases of Americans who hear Japanese for the first time, and likewise of Japanese who hear the voices of Americans for the first time, what they hear are not voices, but ambiguous noises. Even if the words were short, they could not repeat them. Only gradually did the author himself become able to catch the form and construction of English words by hearing. In such a state, it was found necessary to learn to make adjustment by depending on Japanese consonants similar to those of the English words. In an English conversation class it is noticable that some Japanese students consistently used "Japanese style" improvised sounds in English pronunciation, which was very different from the pronunciation of their American teachers. Besides " L " and " R " , there is a lesser difficulty for Japanese to clearly distinguish " B " and " V " as pronounced by Americans. The results of the tests described in this paper indicate that these phenomena are due not to immature articulation, but rather to auditory insensitivity to the difference of pronunciation between Americans and themselves. The author was also corrected by Americans on his pronunciation which he thought to be quite the same as that of Americans. But even if our auditory image of English words might be more approximate than that of Americans, we should be able to grasp most of their speech by hearing, even though we heard it in a distorted form. It is not so important that each Japanese consonant may be or may not be quite the same as English ones. But judging from the scores of the test on " L " and " R " , if there
322
HIROMUGoxo
were many fewer consonants in Japanese than English, we would then have greater difficulty in discriminating bv hearing those English syllables that had no close Japanese equivalent, even though at times we might be able to pronounce a sound close to the English consonant. Auditory discrimination of syllables must be acquired by daily practice from childhood. In cases of patients with "congenital aphasia", because there is no daily practice from early childhood, it seems impossible for them to develop auditory discrimination of voice. If there is a lack of auditory discrimination of voice in "congenital aphasiac patients", it might be said that this deficiency is not the cause, but rather the result of congenital aphasia. The author investigated a case of congenital aphasia in a boy eleven years old who could not pronounce " k a " , " t a " and " y a " distinctly due to palsy of the soft palate and the tongue, but he could pronounce " h a " , " s a " and " n a " quite correctly in comparison. A test of his auditory discrimination ability showed that he could not distinguish by hearing those same syllables, " k a " , " t a " , and " y a " which he could not pronounce. Some normal subjects who had little sensitivity to music were tested with several pairs of sounds made by striking glass vessels, for example, two similar tumblers one-half and three-quarters filled with water. When the vessels were struck successively or when one vessel was struck twice, the difference in sound could be identified easily. But when only one vessel was struck once, they could not indicate, even after much practice, which vessel the sound came from. Auditory perception may thus be more obscure than the visual perception of letters or figures, especially the auditory perception of complex sounds such as voice. In order to make difficult discriminations between poor auditory perceptions of voice, it is suggested that there is a need for some kind of resonance in the receptive system. In the case of hearing words, the author 5 has found that there is minimal speech movement, which however is more minimal and more automatic than that of silent reading which changes visual perception of letters into "inner speech" in order to fix them well in the memory. When, as a result of training, the young child becomes able to articulate syllables correctly, his speech movement and auditory perception (or image) of these syllables will become fixed as an accurate pattern or form in his brain mechanism. That such engrams of phonemes are important in the whole structure of speech is shown by the fact that Japanese frequently demonstrate the same confusion of " L " and " R " in written speech, e.g. writing " s u m m a l y " for " s u m m a r y " . F r o m our experiments, we consider the auditory discrimination of " L " and " R " not as a function of sensation or of perception, but of recognition. For example, the "leads" of "lead and lead" are not quite the same in pronunciation. Therefore, it is a problem of recognition to admit that those two "leads" are the same. But in usual cases of hearing voices, we do not need such a conscious recognition. Huskiness or nasal tone creates a special difficulty for Japanese in perception of a word with the sounds " L " or " R " . There was a difference in pronunciation by each American in the tests used by us, and it is likely that through our auditory training we Japanese are conditioned to distinguish by hearing the sounds " L " and " R " spoken by one particular speaker. With another speaker we are disturbed by the change of discriminative criteria in auditory sensation. O f course, in order to grasp a word as a unit, accent and intonation do have a role to play, but this paper is concerned only with the auditory discrimination of syllables, or to the sound of certain letters such a~ " L " and " R " .
Acknowledgements--Special words of thanks gotothe nine Americans priests andnuns andtomyten Japanese friends for their patient cooperation during the tests. I am very grateful to Prof. D. DENNY-BROWN(Harvard).
AUDITORY PERCEPTION BY NORMAL JAPANESE ADULTS OF THE SOUNDS " L " AND " R "
323
for his generous help and instructive advice. And finally I wish to thank the Rev. RONALD LAFRAMBOISE for his kind help with the translation and preparation of the manuscript. REFERENCES 1. BRAIN, R. The neurology of language. Proc. R. Soc. Med. 54, 433-441 1961. 2. BROWN, R. W. Language and categories. Appendix p.247. In A Stud), t~f Thinking, J. S. BRUN~R et al. (Editors)., John Wiley, New York, 1956. 3. WORSrER-DROtJOHT, C. and ALLEN, I. M. Congenital auditory imperception (Congenital Worddeafness) and its relation to idioglossia and other speech defects. J. Neurol. Psychopath. 10, 193-236, 1930. 4. L~NNF.B~RG, E. H. Case Report. Understanding language without ability to speak. J. Abnorm. Soc. PsychoL 65, 419-425, 1962. 5. Go'ro, H. Studies on "Inner speach". Part II: Hearing and speech movement. Folia Psychiat. Neurol. .lap. 22, 79-88, 1968. R~mn6---1. On fair gcouter des roots anglais enregistrgs sur bande magngtique par des Am6ricains et par des Japonais et on examine si ces m~mes sujets, 8 Amgricains et 11 Japonais, peuvent distinguer les roots qui comportent un " L " et un " R " . 2. Les sujets japonais faisaient des erreurs duns la discrimination auditive entre " L " et " R " , m~me lorsqu'ils entendaient la voix des amgricains. Aussi considgration qu'ils 6taient incapables de distinguer " L " et " R " par l'audition. 3. Les sujets Japonais dont la connaissance de l'anglais ~tait la meilleure avaient une prononciation du "'L" et du " R " relativement bonne, si on la compare ~t leur mauvaise discrimination auditive. 4. La discrimination auditive de syllabes est possible par la pratique, duns la parole et I'audition depuis la prime enfance. Sans une telle pratique, il existe de consid~rables difficult~s dans leur discrimination. Ces constatations souligent l'importance de la sensation kinesth~sique duns la propre parole du sujet quand il apprend/t articuler. Zusammenfassung--1. Amerikaner und Japaner wurden mit einem Tonband in englischer Sprache daraufhin getestet, ob sie unterscheiden konnten, welche W6rter L u n d R enthielten. An Versuchspersonen standen 8 Amerikaner und 11 Japaner zur Veff'tigung. 2. Japanische Probanden unterchieden L und R akustich ungenau, wenn sie amerikanische Stimmen vernahmen. Es wurde daraus geschlossen, dab sie unf/ihig sind, L and R vom Geht~r her zu unterscheiden. 3. Japaner, die die englische Sprache beherrschten, konnten die Buchstaben L und R relativ gut produzieren, gemessen an ihrem geringen akustischen Unterscheidungsverm6gen. 4. Akustische Unterscheidung von Silben wird von frOher Kindheit her durch Sprechtibungen und H~Sren erm6glicht. Wenn man es nicht praktisch iibt, so entsteht eine bemerkenswerte Unterscheidungserschwerung. Diese Ergebnisse zeigen die Bedeutung kinaesthetischer Empfindungen beim Erlarnen artikulatorischen Sprechens.