Atlas of German Speech Sounds [Reprint 2022 ed.] 9783112679326

Table of contents :
CONTENTS
PREFACE TO THE ENGLISH EDITION
1. TEXT
2. PHOTOGRAPHIC PLATES

Citation preview

HANS.HEINRICH

WÄNGLER

ATLAS O F G E R M A N S P E E C H

SOUNDS

ATLAS OF GERMAN S P E E C H SOUNDS

by

Hans-Heinrich Wangler University of Colorado Phonetics Laboratory

Translated by

Keith O.Anderson

AKADE M I E • V E R L A G 1968

•

B E R L I N

Published by the Akademie-Verlag G m b H , 108 Berlin, Leipziger Straße 3—4 Copyright 1968 b y Akademie-Verlag GmbH Licence No. 100/236/68 Printed and bound b y : „Maxim Gorki", Altenburg Order Number 5316 E S 7 A, B, D

CONTENTS

1. T E X T 1.1 1.2 1.3 1.3.1 1.3.2 1.3.3 1.4 1.5 1.6 1.6.1 1.6.2 1.6.2.1 1.6.2.2 1.6.2.3 1.6.2.4 1.6.3 1.6.4 1.6.4.1 1.6.4.2 1.6.4.3 1.6.4.4 1.6.4.5 1.6.4.6 1.6.4.7 1.6.5 1.7

On the Significance of Speech Sounds Phonetic Methods of Speech Sound Description The Production of Speech Sounds Anatomical Aspects Physical Aspects Physiological Aspects The Classification of Speech Sounds Phonetic Transcription The German Speech Sounds Survey The Consonants The Stops The Nasals The Fricatives and Laterals The Trills An Aid for Classifying Articulations The Vowels a-Sounds e-Sounds i-Sounds o-Sounds u-Sounds o-Umlauts u-Umlauts The Diphthongs Bibliography 2. P H O T O G R A P H I C

2.1 2.1.1 2.1.2 2.1.3

Stops Voiced [b] — voiceless [p] Voiced [d] — voiceless [t] Voiced [g] — voiceless [k]

9 14 18 18 20 23 26 30 31 31 34 34 36 36 39 40 41 42 42 43 44 44 45 45 45 47

PLATES Plate 1 Plate 2 Plate 3

2.2

Nasals

2.2.1 2.2.2 2.2.3

Voiced [m] Voiced [n] Voiced [q]

2.3

Fricatives and Lateral

2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6

Voiced [v] — Voiceless [f] Voiced [z] - Voiceless [s] Voiceless [J] Voiced [j] - Voiceless [9] Voiceless [x J Voiced [1]

2.4

Trills

2.4.1 2.4.2

Voiced [r] Voiced [R]

2.5

a-Sounds

2.5.1 2.5.2

Front Short a-Sound [a] Back Long a-Sound [a:]

2.6

e-Sounds

2.6.1 2.6.2 2.6.3

Short, Open e-Sound [s]; Long, Open e-Sound [e:] Long, Close e-Sound [e:] Short, Unstressed e-Sound [a]

2.7

i-Sounds

2.7.1 2.7.2

Short, Open i-Sound [1] Long, Close i-Sound [i:]

2.8

o-Sounds

2.8.1 2.8.2

Short, Open o-Sound [0] Long, Close o-Sound [o:]

2.9

u-Sounds

2.9.1 2.9.2 2.10.1 2.10.2

Short, Open u-Sound [u] Long, Close u^Sound [u:] Short, Open o-Umlaut [ce] Long, Close o-Umlaut [0:]

2.11

u-Umlauts

2.11.1 Short, Open u-Umlaut [Y] 2.11.2 Long, Close u-Umlaut [y:]

Plate 4 Plate 5 Plate 6 Plate 7 Plate 8 Plate 9 Plate 10 Piateli Plate 12 Plate 13 Plate 14 Plate 15 Plate 16 Plate 17 Plate 18 Plate 19 Plate 20 Plate 21 Plate 22 Plate 23 Plate Plate Plate Plate

24 25 26 27

Plate 28 Plate 29

PREFACE TO THE ENGLISH EDITION

This book is a translation of the fourth edition of my Atlas deutscher Sprachlaute, which was first published in 1958 in Berlin. Although it has been reworked and expanded several times since then, it has remained what it set out to be, a purely phonetic contribution. A need existed for a more up-to-date description of the genetic features of German speech sounds. This fact prompted me to attempt to fill the perceptible gap by presenting these results obtained through the latest techniques of investigation. The basically practical purpose of this publication demanded, above all, striking clarity of presentation. To achieve this goal, I was forced to limit myself to the most important features. Thus it was possible to select but a small portion of the much more extensive experimental data obtained for presentation in the book. For practical purposes, the desire for larger photographic plates also could not be realized. (The originals show the entire head and extend down to the upper thoracic cavity.) However, certain principles were strictly observed. For example, every plate represents an instant, which under close technical control was "shot out" of the continuous natural speech process. Also there was no retouching of the original films. The final selection and presentation of the plates for individual sounds was preceded by extensive experimentation and careful comparison of the results. Prof. B. Schlosshauer, M. D., was kind enough to offer me professional advice and actual help in the preparation of the x-ray photographs. The photographic work was done by Mr. Th. Voigt, who was then the head of the photography service of the University of Hamburg. Without such support, this work could not have been completed. Therefore, I feel the need to thank the above-named persons also on this occasion. Every author is gratified to see his work translated into a foreign language. Thus I also feel obligated to express sincere thanks to the Akademie

Verlag for their support of this undertaking and for their always exemplary cooperation with the author. Last but not least, I wish to thank Mr. Keith Anderson for his competent translation and hope that the work will not merely add to the flood of new publications, but in its present form will offer a real contribution, especially for German language instruction in English-speaking countries. Boulder, May 1967

Hans-Heinrich Wàngler

TEXT

1.1 O N T H E

SIGNIFICANCE

OF SPEECH

SOUNDS

A sound, in the broadest sense of the word, is understood to be something audible, something t h a t "sounds." I n German, the physical term for loudness (Lautheit) also seems to be directly derived from the word for "sound" (Laut). I n phonetics we associate the term especially with speech sounds, with the articulated sounds of man, which represent an ordered system of signs for conveying information. These human communication signals are investigated by phonetics as physiological processes (genetic investigation) and as acoustic-physical phenomena (gennemic investigation). Even before the birth of Christ, the Indians 1 and Egyptians, followed by the philosophers of classical antiquity, 2 ' 3 had already gathered important insights in this area. Interest in the phenomenon of human language is of early origin. Due to this interest, effort was directed especially toward the analysis of its manifested form. The period between the first and fifteenth centuries A. D. contains no significant milestones in the history of phonetics, b u t by the seventeenth century a significant rebirth can be noted. The first advances were the results of various dissertations on phonetic topics. More and more the viewpoint of antiquity was readopted, t h a t the speech process is the most important source of information in scientific study of language. Closely linked with this development, which progressed from the renunciation of speculative etymology based on orthography to the exacting methods of phonology, are the names VON KEMPELEN, R A P P , Johs. MÜLLER, BRÜCKE, MERKEL and HELMHOLZ on the one side and RASK, GRIMM, BOPP, e t c . o n t h e o t h e r . 1

2

3

In the 7th century B. C., S A U N A K A had already investigated the phonological rules of Sanskrit phonetics in his work "Rig-Veda-Prätisäkya." Cf. G. PANCONCELLI-CALZIA, Geschichtszahlen der Phonetik (3000 Jahre Phonetik), Hamburg 1941. Cf. G. PANCONCELLI-CALZIA, Quellenatlas zur Geschichte der Phonetik, Hamburg 1940.

10

Oil the Significance of Speech Sounds

The ensuing schism between the physiological and linguistic branches of phonetics was then bridged by Abbé ROUSSELOT,4 who deliberately placed the study of the speech process in service to linguistics. The simultaneous development by B. DE C O U R T E N A Y , 5 J . W I N T E L E R , 6 F. de SAUS7 SURE and N. S . T R U B E T Z K O Y 8 of a scientific approach to the speech sound as an abstraction within the framework of linguistically relevant relationships (phonemics), and the more recent work of E. ZWIRNER,9 who has compiled speech sound norms in phonometry through the statistical analysis of their spoken variations, would seem to deny at first the close bond between linguistics and phonetics. However, it is exactly this, to a degree fundamental, difference in approach and in methods of investigation which is of greatest benefit to general linguistics. Here we are concerned with the description of genetic sound features as a fundamental task of phonetics. The child learns to speak by hearing and seeing others speak. I n this manner he not only develops his speech as the audible realization of an imaginary norm, but also acquires his language. This normative system of relevant symbols is also built up out of the multitude of apperceived speech forms through the reproduction of corresponding trial, (later) selective and (finally) goal-directed acts. If only for this reason, it is both scientifically and practically important to understand the nature of the articulatory process, since it gives us insight into the essence and development of speech and language. Gaining this understanding is one of the goals of phonetics when it resorts to use of appropriate instruments where our senses no longer suffice. Criticism has frequently been leveled against this method. The main argument is that instrumental aids are unnecessary. Everyone learns to speak without them. I t is possible that their foreign nature and inaccuracy will but project new sources of error into a completely natural process. I t is regretable, that it is still necessary to point to the multitude of children who are deaf (and therefore speechless) or have other speech impediments in order to justify an approach aimed at obtaining a knowledge of the speech process which extends beyond the niveau needed for many commonplace purposes. The person who chooses to devote his life to helping 4 5 6 7 8 8

P. J. ROUSSELOT, Les modifications de la parole, Diss., 1891. In his Russian inaugural lecture of 1870. J. WINTELER, Die Kerenzer Mundart des Kanton Glarus, Leipzig 1876. F. de SAUSSURE, Cours de linguistique générale, Lausanne 1916. N. S. TKUBETZKOY, Grundziige der Phonologie, Prag 1939. E. und K. ZWIRNER, Grundfragen der Phonometrie, Berlin 1936.

On the Significance of Speech Sounds

11

such unfortunate children needs, among other things, a specialized knowledge of phonetics. The success of his work is dependent upon his ability to use special means for teaching these children the processes of articulation. Much the same is true in foreign language learning. The listener can at first only hear and reproduce correctly those sounds which are familiar to him from his native language. Insufficient mastery of linguistically correct articulation is the primary cause of bad German, English, French, Russian, etc. spoken by foreigners. I t is said that continued listening to correct models of articulation is a sure w a y to overcome such faults. So formulated, this assertion is but a stubbornly persisting fable, which has recently been retold by those who have misunderstood the methodology of foreign language instruction. Everyone knows foreigners who have lived as long as twenty years or more in a new country and who still speak with such a prominent "accent" that they are recognized as foreigners as soon as they open their mouths. During this period they have doubtless had sufficient opportunity to hear what is "correct." A t any rate, they have listened longer than their children born after their immigration. Y e t the children speak the language perfectly.

The decisive factor is not the quantity of the auditory impression, but its quality. Hearing represents a very complex psycho-physiological process. I t depends partially on the quality of the sound, but even more on its correct interpretation. Thus we are speaking less of the quality of the sent signal than of the quality of its reception. That which is " h e a r d " must also be correctly apperceived. This is impossible without clear recognition of the relevant phonetic differences between the native and foreign languages. Although a simple phonetic description coupled with appropriate controlled drill may lead to the correct articulation of the speech sounds of closely related languages, this approach is often insufficient for the languages of more remote cultures. Our knowledge of the nature of many sounds in African languages, for example (clicks, ejectives with a secondary oral or laryngeal closure, etc.), has been obtained solely through specialized phonetic experimentation. Phonetic experiments, furthermore, offer substantial heuristic aids for freeing recent findings in theoretical linguistics from their present hypothetical state. According to these findings, language (in the broadest sense of the word) occupies the key position in our existence as humans. Only through it is knowledge possible. I t is our primary access to knowledge; it marks out its breadth and depth and determines its nature down to its minor details.

12

On the Significance of Speech Sounds

Language is, accordingly, everything else than merely a practical device for expressing immanent thoughts which have been ladled from some sort of a logical reservoir within man. I t is rather simultaneously the very point of departure for our thoughts. We think within the framework of language. It is the eyeglass through which our intellect views the things of this world. It is the individual languages that divide up nature and our knowledge of the world in different manners, that interpret the totality of given factors according to their own historically developed structural relationships. They both reflect and shape the diverse worlds of their speakers. Speech, then, the inherent object of phonetic investigation, consists clearly of two different components. The first is doubtless the realization of the normative demands of a particular language. Yet speech today still contains aspects of the original phonetic attempt to establish contact with another human being, which in the course of time gave rise to the creation of the linguistic norm. This creative component in the speech process is a vital element of language, not only in the sense that so-called "living" languages must still be spoken, but also in the sense that it is an essential prerequisite for their very existence. An indication that speech still fulfills this function is the continuing linguistic change initiated by speech. This line of reasoning not only places the study of our native language10 as well as the study of foreign languages11 in a new light, but also points once more to linguistic phonetics, especially when based on a genetic and experimental approach, as a fundamental supporting discipline to general linguistics. Finally, every professional speaker (pastor, teacher, singer, actor) can profit by studying the genetic aspects of speech sounds. This is true whether he wishes merely to avoid possible detrimental effects caused by abnormal demands on his voice, whether he wishes to become better equipped for his vocation by gaining a basic physical knowledge of the tools of his trade, or whether his goal is to enhance the power of his expression based on such knowledge. Within the framework of his native language, the average unquestioning layman with normal speech is largely unconcerned, even for practical purposes, as to what takes place in speech. Yet this lack of concern obviously 10 11

L. WEISGERBER, Von den Kräften der deutschen Sprache I—IV, Düsseldorf 1953. H.-H. WÄNGLER, Bemerkungen zur Methodik des Fremdsprachenunterrichts, Zs. f. Phon. 1964, pp. 347 ff. (Festschrift Häla).

On the Significance of Speech Sounds

13

cannot be shared by those who undertake the scientific study of language. Here, insights into the articulation of individual languages are an inexpendable prerequisite for well-grounded linguistic findings. Furthermore, every scientific study, as a matter of fundamental principle, must go considerably beyond the immediate demands of applied science.

1.2 P H O N E T I C M E T H O D S O F S P E E C H SOUND DESCRIPTION

Commensurate with the great general significance of speech sounds, efforts were directed early toward their description. The degree to which the intellectual legacy of antiquity in this area, as in others, was lost on the Middle Ages, is demonstrated by the curious attempt of VAN H E L M O N T to describe the Hebrew alphabet as graphic representations of the individual sound articulations because (!) God spoke Hebrew.12 The present-day phonetician who undertakes a description of speech sounds is confronted by a variety of theories. First, there are those who consider a graphic representation of the individual articulatory positions to be outdated. The genetic side of the sounds — according to this point of view — has long been known to us; new discoveries can be expected only from the physical analysis of the acoustic end product. As a matter of fact, the genetic description of individual positions is an old method.13 This is true even when, as here, more recent technological aids are employed. And also, many more important advances in the field of phonetics can still be expected from the electro-acoustic analysis of speech sounds. But to play off the two approaches against each other is completely unjustifiable.14,15 To regard in the name of progress the "old" method as superseded by the "new" method is to demonstrate a deficient understanding of the problem as a whole. The two approaches are equally valid; the one chosen depends completely upon the goal of the 12

13 14

15

FRANZ MERCUKITTS VAN HELMONT, Kurzer Entwurff des eigentlichen Natur-Alphabets

der Heiligen Sprache (1667), Reprint edited by W. Vietor, Hamburg 1916, Vox 1916, Appendix pp. 1—64. Cf. Speech sound tables by TECHMER, RATTSCH, SCHNELL, VIETOR and others. J . FORCHHAMMER, Einteilung der Sprachlaute. Akustisch oder artikulatorisch, Studia linguistica 8, 1954/1, pp. 34 ff. B. MALMBERG, Unhorbare Laute. Eine Erwiderung an J . FORCHHAMMER, Studia linguistica 8, 1954/1, pp. 54ff.

Phonetic Methods of Speech Sound Description

15

investigation. For example, questions dealing with the transmission of information would not primarily be dealt with genetically, while electroacoustic aids fail when insight into the production of sounds is needed as a prerequisite for work in speech therapy. Yet the primary task of phonetics is still the description and explanation of the process of phonation within the human organism. Another possible criticism is more serious. The reality of actual speech knows no static isolated articulatory positions; the continuity of the process cannot be shown with photographic still pictures. We can become reconciled to this problem if we adopt the standpoint of E D U A R D SIEVERS, 1 6 t h a t static sounds can be regarded as discrete unities which are bound closely together in spoken sentences by the on- and off-glide of transitional sounds. B u t experimental phonetics itself has clearly shown t h a t such a viewpoint is no longer supportable. Speech is not a series of articulatory positions, 17 but is a continuous formative process. The individual sounds are not only closely joined with their neighboring sounds (coarticulation), 18 but are, together with their coordinated voice components, the product of continuous articulatory movements. 19 These movements only transiently realize the phonetic distinctive sound features (place, manner and mode of articulation) and do not cease completely even in the pauses dividing sense groups where no speech sounds are produced. Thus we are dealing with an entity, with a true "Gestalt" in the psychological sense. The question then arises as to what can be expected from a description of individual positions. First of all: Even though the conventional series of letters used in our orthography was long able to support the illusion of the word as a sequence of discrete sounds, no one who scientifically is worthy of our attention ever imagined it to consist merely of the sum of clearly definable components. The introduction of the term "Gestalt", however, (which means to say t h a t the components are bound together to form a higher level entity by means of a power which obviously does not reside within the individual components themselves) has often served to turn our attention solely to the new entity and away from the individual building blocks. Yet this can only be considered valid when, under due consideration of the situation 16 17 18

19

Eduard SIEVERS, Grundzüge der Phonetik, 4th Edition, 1893, p. 32. P. de S A U S S U R E , Cours etc., 1916. P . MENZERATH and A . de LACERDA, Koartikulation, Steuerung und Lautabgrenzung, Berlin and Bonn 1933. Chr. W I N K L E R , Deutsche Sprechkunde und Sprecherziehung, Düsseldorf 1954, pp. 218ff.

16

Phonetic Methods of Speech Sound Description

as a whole, a new dissection is undertaken. An essential aspect of scientific inquiry is, after all, the pressing onward from the particular to that which is typologically general. In seeking out the unifying order by this method, a knowledge and understanding of the particulars is an inexpendable prerequisite. A whole without parts is equally as nonexistent as a whole conceived of only as a sequence of parts. Thus it probably is necessary to redefine the elements and to give special attention to their function in context. In this process it is more likely that the individual unit will gain rather than lose significance, especially through its more exact delimitation, within the framework of a new comprehensive perspective. Furthermore, the fact remains that our ear actually does perceive the spoken word as a sequence of differentiable sounds. 20,21 These sounds must correspond to a substratum of articulatory movements, since the linguistically relevant modifications of the sounds are produced only through them. Although the segmentation of sounds from their neighbors genetically and in the acoustic end product does not always appear to be easy, this difficulty frequently is due to the inadequacy of our experimental instrumentation. It is predominantly the case, at least in German, that the (relatively) stationary segment, the so-called "Klarphase,"22 which is also genetically easily recognizable, gives the listener unquestionable recognition of the speech sound. The transitional phases, which objectively are certainly present, are linguistically irrelevant and to a large degree meaningless in speech. This acoustically marked central phase of the sound is also the audible result of a segment of articulatory motion. If we are permitted to make a completely justifiable generalization, this segment can be regarded as the determinative moment in the production of the sound. It is not too short to be captured by still pictures using careful experimental procedures. That has been done here and has led in many instances to new findings above and beyond the systematic manner of presentation used in this book. Although all of the examples (which are not based on isolated sounds, but "extracted" from the spoken word by careful comparison of many photographs) involve "stationary" positions, it was also desirable to study in some detail the whole articulatory process and the conditioned variations 20 21 22

O. von ESSEN, Allgemeine und angewandte Phonetik, 3rd Edition, Berlin 1963, p. 96. 0 . von ESSEN, Über das Wesen der Assimilation, Vox 1953. E. DIETH, Vademekum der Phonetik, Bern 1950, p. 230.

Phonetic Methods of Speech Sound Description

17

that occur in different positions. Therefore, simultaneous sound motionpicture x-ray films were made of the entire speech process. These films, although phototechnically not surpassing the well-known material produced by ROBERT J ANKER (Bonn) in 1939, are clearly superior from the standpoint of phonetics. The motion-picture film confirms the results presented in this work and also shows that the method of portrayal used here has the advantage of giving the greatest clearness for the purpose intended.

2

Wangler, Atlas

1.3 T H E P R O D U C T I O N O F S P E E C H

1.3.1 Anatomical

SOUNDS

Aspects

The term "attached pipe" (Ansatzrohr) referring to the resonance cavities attached directly to the sound source has been borrowed from the wind instruments to designate the human vocal tract. Although the name is neither linguistically nor physically fully satisfactory, 23 ' 24 its use has become established. After all, sounds are not only modified in the human "Ansatzrohr," but are also formed and produced by it. This is, from a technical point of view, not implied in the original meaning. The coordination of all three components, i. e. respiration, voice production and resonant cavities, is by no means an unconditional prerequisite for speech. For clicks and ejective stops with secondary oral occlusion, for example, the tract alone is active. The human vocal tract (Ansatzrohr) extends from the vocal bands (exclusively) to the lips and nostrils and thus includes the pharyngeal cavities (hypo-, meso- and hyperpharynx), the oral cavity (cavum oris) and the nasal cavity (cavum nasale) divided into two halves by a vertical septum. The oral and nasal cavities are separated from each other by the palate. The entire vocal tract can undergo certain modifications in size and configuration through muscular action. I t is completely covered with mucus membrane. The parts of the vocal tract which are essential for articulation can be grouped according to whether their roles are active or passive. The active parts are the lower jaw, tongue, lips and soft palate (as well as the hyoid bone and the epiglottis), the passive ones are the teeth, gums and hard palate. 23

24

proposed the term "Lautrohr" (sound pipe) as a substitute. Die Sprachlaute etc., 1884. E. D I E T H also wishes to avoid the meaningless "häßliche Bild vom angesetzten Rohr" and offers the term "Lautgang" (sound passageway). Vademekum der Phonetik, Bern 1950, p. 124.

N . TRAUTMANN

Anatomical Aspects

19

The lower jaw is moved by the muscles of mastication (mm. masseter and temporalis). The base of the powerful lingual muscle (m. genioglossus) is the hyoid bone (os hyoideum), to which the hyoglossus muscle is also attached. Several other lingual muscles pass freely through the tongue. Others act as retractors (m. palatoglossus, m. styloglossus) and make the complex muscle system "tongue," which is the most mobile organ of our body, capable of fulfilling a wide variety of tasks. The lips are brought into the positions important for articulation mainly by the orbicularis oris muscle, a circular muscle group encompassing the periphery of the mouth, and by muscles of the upper (m. quadatus labii superior) and lower lips (m. quadratus labii inferior).

The great mobility of the soft palate (velum) also indicates that the action of several muscles is involved. Those to be noted are the m. tensor veli palatini, the m. levator veli palatini and the m. uvulae contained in the uvula, which is the terminaton of the soft palate. The action of these three muscles, together with the action of the uppermost of the three constrictors of the pharynx (mm. constrictores pharyngis), which has the function of forming the "cushion of PASSAVANT", permits complete closure of the nasopharyngeal passageway. The epiglottis (cartílago epiglottica), shaped like a shoehorn, extends upward from the incisura thyroidea (whose edge can be felt externally 2*

Wangler, Atlas

20

The Production of Speech Sounds

as the "Adam's apple") and lies against the root of the tongue, with which it is joined by a small ligament. In swallowing, it is lowered. Together with the sphincter-like constriction of the pharynx it helps to prevent food particles from entering the trachea. Otherwise — even during phonation — it remains more or less upright. Although the epiglottis and the hyoid bone do not remain stationary during speech (as proven by x-ray motion-picture films), their movement is not of essential importance for articulation. Therefore, the muscles which move them need not be mentioned here. We should note, however, that the vocal tract experiences significant modifications in size through the raising and lowering of the entire larynx (by the action of the mm. sternothyroidei and thyreohyoidei).

1.3.2 Physical Aspects

The so-called active parts of the vocal tract work essentially in two manners to produce speech sounds. First, by means of changes in spatial volume they form various resonators which modify the quality of the glottal tone (produced by the vibration of the vocal bands) in characteristic fashions (vowels); second, they act as obstacles for the breath stream either with or without accompanying voicing (consonants). The essence of the vowels and, in a certain sense, also of the consonants rests acoustically on the realization of certain formants. These are one or more groups of overtone regions which have been strengthened by resonance and which are almost completely independent of the frequency of the fundamental.25'26 The vowel [a], for example, always has a formant region around 600 — 900 cps if it is to be perceived as an a-sound. This is true whether it is spoken with a deep man's voice or with the high voice of a child. If this formant region is not present in the sound, then we do not hear the vowel [a]. It can be demonstrated with the proper laboratory instruments employing a system of condensors and coils, so-called band pass filters, that this formant region is an inexpendable prerequisite for the acoustic perception of the sound.27 Thus, if hearing loss extends down to the region around 800 — 900 cps, the vowels [i], [e] and [a] sound like an 25 26 27

L. HEBMANN, Phonophotographische Untersuchungen, Pflügers Archiv 1889—95. C. STUMPF, Die Sprachlaute, Berlin 1926. First done by BRUNNINGS at the Hamburger Kongreß Deutscher HNO-Ärzte, 1 9 2 6 .

21

Physical Aspects

indistinct [o] to the impaired ear because these vocoids have an additional lower formant which — like the main formant of [o] — lies below the upper limit of hearing. Many consonants have formant regions which are higher The

Formant

Regions

of German

F2

Fi

around 800 cps

a

Vowels

around 1400 cps

a:

„

750

„

„

1150

„

£

„

500

„

„

1900

„

e:

„

375

„

„

2100

„

a

500 325

„ „

„

1200

„

i

„ „

„

2200

„

i:

„

275

„

„

2400

„

0

„

500

„

900

„

o:

„

375

„

850

„

V

,,

325

,,

850

„

u:

„

275

„

750

„ „

oe

„

500

„

„

1550

0:

„

375

„

„

1700

„

Y

„

325

„

„

1800

„

y:

„

275

„

„

2000

„

lOOcps

/'••

y:

I

Y

e-

z œ a a w 3 oThe Formants F, and F2 of the German Vowels

u

u-

22

T h e P r o d u c t i o n of S p e e c h S o u n d s

than the vowels. Not infrequently a faulty s-articulation results because the speaker is unable to hear the relatively high partial-tone regions of this consonant. (Those which differentiate it from [f] lie in the vicinity of 4000 cps.) The speaker loses acoustic control of the linguistically correct articulation — defective pronunciation is the result. Yet he doesn't necessarily have a noticeable general hearing loss. In general, dispersions of ± 75 cps (up to ± 100 cps) with (up to ± 250 cps) with F 2 are possible.

Acoustic Quadrilateral

of the German Vowels

or ± 200 cps

Physiological Aspects

23

Entering the values of German vowel formants into a system so t h a t F 1 is represented on its vertical and F 2 on its horizontal axis, the similarity to t h e vowel quadrilateral representing t h e physiological characteristics of vowel articulations is striking (Cf. p. 33). The characteristic conformation of the vocal tract is also important for the strengthening of t h e fundamental. I n so-called forced resonance the effect is produced by direct coupling of the sound source to the resonator and every frequency is reproduced. Here the law applies: the greater the radiating surface, the louder the sound. I n the strengthening of phonation, however, we employ primarily t h e principle of free resonance. A hollow body which is open on one side will (theoretically) only sound when excited by a tone matching its own inherent resonant frequency. I t cannot be made t o oscillate by tones with other frequencies. 28 Smaller resonators in general amplify higher tones, larger resonators lower tones. The "ideal resonance" in speaking and singing therefore demands a continual change in t h e size of our resonator, of our vocal tract, coordinated with the changes in the frequency produced. Practically, this requirement is only more or less approximately met. More t h a n this is not necessary, since a soft-walled resonator (in contrast to a hard-walled resonator) counters the disadvantage of rapid damping with the unquestionable advantage of a certain broadness of resonance. Even with no change in dimensions, it can be excited not only by an isolated frequency, b u t also by sound energy in the adjacent frequency regions. 29

1.3.3 Physiological

Aspects

When direct observation is no longer adequate, phonetic experimentation can show us how the active parts of the vocal tract work together in articulation. The position of the mandible determines, first of all, the degree of "openness" of each sound. This can be measured and described by the angle 28 29

H. HELMHOLTZ, Die Lehre von den Tonempfindungen, First Edition, Braunschweig 1863. The presentation or even discussion of the numerous research findings recently made in relation to the physical structure of speech sounds must be omitted here. For the purposes of this short general introduction we are concerned only with the theories which were basic to all subsequent research and which remain valid today although they have been considerably expanded and modified by later investigators. For further details consult the bibliography given below.

24

The Production of Speech Sounds

which the (immovable) maxilla makes with the (movable) mandible (jaw angle). The lips passively follow this movement for some sounds ([a], [a], [e]), B u t when they are rounded or protruded ([o], [ce], [u], [Y] or [o], [0], [u]. [y], [i], [J]) or are spread ([e], [i], [s], [z]) they assume an active role in the production of the sound. Finally, they form the occlusion for the stops [b] and [p] by their contact with greater or lesser degree of pressure and they also play an essential part in the production of the labial nasal [m]. The tongue, because of its great mobility and capacity to change its shape, is the primary organ of articulation. B y raising or lowering, b y moving back or to the front, either as a whole or at certain points, it provides the resonating cavities necessary for vowel production; b y creating various types of constrictions or points of contact with passive articulatory structures it gives the consonants their distinctive features. The soft palate allows completely free passage of air through the nose only for the nasals ([m], [n], [rj]) in German; for all other sounds the breath stream is more or less hindered or blocked from entering the nasal passagew a y . I t was a known fact, that closure is not achieved by the uvula (which always hangs down), but by the raising of the velum to make contact with P A S S A V A N T ' S cushion. To m y knowledge, however, the observation based on the results presented here, that each sound has a specific degree of closure determined obviously by some rule, is new. The velopharyngeal closure seems directly dependent on the oral closure. If the oral passage is completely open (as for [a]), the tendency toward closure is the weakest; if in various w a y s the oral passage becomes more and more restricted ([e] - » [e] -> [i] or [o] [u] -> [y]) closure becomes greater until reaching a maximum for the fricatives and stops. Finally, complete blockage of the nasopharyngeal passage is reached. The soft palate moves (in respect to its function in sound production) according to the momentary articulatory necessity. I t does not simply open and close the nasal passage, but rather forms finely controlled degrees of closure which are always just sufficient to prevent nasality (rhinolalia aperta). If the danger of nasality is slight (because of an unimpeded oral passage), the velum expends a minimum of energy; not until the danger becomes acute (as with fricatives or stops) is the nasal passage hermetically sealed. Nevertheless, it is still completely correct to classify vowels as oral sounds; the degrees of nasopharyngeal opening determined on the basis

Physiological Aspects

25

of our observations are irrelevant in respect to both language and speech. 30 H E L M H O L T Z believed that the partial tones characteristic of the individual vowels were already produced through the mode of vibration of the vocal bands. Yet his work may be regarded as the real breakthrough in our understanding of vowel production. The resonance qualities of the supraglottal cavities are beyond doubt the determining factor. This was proven already by Johs. MÜLLER.31 If air is blown through the larynx of a cadaver after the vocal bands have been brought together with the proper manipulation, the result is nothing but a harsh buzzing sound. If a model vocal tract formed according to the phonetic features of a simple speech sound is then placed over it, the corresponding sound is produced. 32 The well known buzzers t h a t have been used by many people following a laryngectomy prove the same thing. When placed at the proper location against the vocal tract, approximately at the angle between the neck and lower jaw, clearly understandable vowels and voiced consonants can be produced. 30

H.-H. WÄNOLEB, Über die Funktion des weichen Gaumens beim Sprechen, Wiss. Zs. d.

31

Johs. MÜLLER, Handbuch der Physiologie des Menschen, Coblenz 1837, Vol. 2, Section 1. The author has duplicated the experiment many times.

M a r t i n - L u t h e r - U n i v e r s i t ä t H a l l e - W i t t e n b e r g 1962, p p . 1747FF. 32

1.4 T H E C L A S S I F I C A T I O N O F S P E E C H

SOUNDS

From the great number of different speech sounds which can be produced by man — theoretically the possibilities are unlimited — each speech community selects but a few. The principle behind this process of selection is not simply ease of articulation, even though certain articulatory features, which are conditioned by the structure of the speech mechanism and physiologically more accessible, are found in all parts of the world. Rather, various fundamental psychological principles play a decisive role here. The nature and regularity of these principles have not been satisfactorily determined to date. The phonetician also has little stake in them; in general his work is based on the given realities of speech. Certainly we do recognize our speech sounds essentially by their acoustic features.33 If they are clearly articulated we can positively identify them even from a record or magnetic tape or over the radio. Yet this is only possible because we know how they are produced in our native language. The acoustic form awakens within us the memory of familiar motor processes which — as experience reinforced by years of childhood experience has taught us — lead to similar results. We abstract, — and the abstraction is successful because we have built up for ourselves an empirical knowledge of the basic principle that acoustic sound quality is completely dependent on underlying physiological phenomena. This process is an essential underlying factor in the natural acquisition of our native language. At the very least, it is so inseparably bound to language development that we are no longer conscious of the original connections between the sound and its production. Thus in the scientific study of foreign languages, whoever believes that acoustic perception is sufficient for gaining new insights into the sounds, their production and their interrelationships is in error. That is completely 33

The written sign will not be taken into consideration here.

The Classification of Speech Sounds

27

out of the question. W h a t is heard, understood, recorded and classified is not the foreign speech sound but an impression of it conditioned more or less by the particular native language of the observer. Our ear is no microphone. E v e n if the complex psychophysical act of hearing were nothing but an objective registration of sound events, we would be no closer to the basis of the actual problem, for any sound product permits only superficial conclusions as to its production. Y e t speech sound production needs further study, since from it can be expected the most reliable insights into the nature and position-conditioned variations of the sounds themselves. In this respect, the great dependability of recent physical analysis and synthesis of speech sounds, which, for example, is of great practical value in problems of information transmission, can offer no substitute. I t is neither possible nor necessary to discover and describe all of the factors which work together to constitute the sounds. All, by far, are not necessary criteria of differentiation. I t is usually sufficient to give the phonetic specifications for: a) point of articulation, b) organ of articulation, c) manner of articulation. Frequently it is also necessary to indicate the presence or absence of voicing. The point of articulation can take in all parts of the oral cavity which remain essentially passive in articulation and are contacted or approached b y an organ of articulation. This includes especially the immobile palate (back to the velar region) together with the upper teeth and alveolar ridge, but also the pharyngeal cavity and the larynx. I t has also been proven expedient to include the upper lip and the uvula, a practice which is not entirely correct, since they could also be classified as organs of articulation. Thus we have the following system in designating the first criterion in speech sound classification: P o i n t of a r t i c u l a t i o n

Sound designation

upper lip upper teeth

labial dental a) addental b) interdental 3 4 alveolar

alveolar ridge 34

Does not exist in German.

28

The Classification of Speech Sounds

front central • hard palate back soft palate uvula pharynx larynx

pre medio • palatal post velar uvular pharyngeal laryngeal35

The active parts of the speech mechanism are primarily organs of articulation (except for the upper lip and the velum, cf. above). The second specification of speech sounds leads to the following classes: Organ of a r t i c u l a t i o n lower lip tongue tip front central surface of the tongue back vocal bands

Sound d e s i g n a t i o n labial apical 36 pre medio • dorsal post laryngeal

The manner of articulation describes the way in which the two determining factors above work together in producing speech sounds. Differentiation is made between the following characteristics of articulation: 1. open (breath stream unimpeded) 2. fricative (breath stream forced through a narrow opening causing audible friction) a) lateral fricative (breath stream exits over one or both sides of the tongue) 3. stop (breath stream stopped at some point in the vocal tract) a) intermittent stop (breath stream is interrupted by a rapid sequence of closing and opening) 4. nasal (breath stream allowed to pass through the nasal cavity) As a rule, an articulation can be characterized by giving one specification from each of the three criteria of classification plus a specification of the presence or absence of voicing. Thus a [p] is a voiceless bilabial 37 stop, [n] 35 36

37

Does not exist in German. The term "koronal" is sometimes used in German to designate a broader area along the front edge of the tongue. In this case "bilabial" designates both the point and organ of articulation (upper lip and lower lip).

29

The Classification of Speech Sounds

an apico-alveolar (voiced) nasal. Often, however, additional features such as release (of stops), energy, duration, etc. must also be given. The naming of essential articulatory features is, for practical purposes, a satisfactory method of classification for the consonants. The system corresponds in principle to the phenomenon. Although it can handle stops and fricatives or sounds with (noticeable) narrowing of the vocal tract, it is by nature much less satisfactory for classifying sounds, such as vowels, which are produced by certain configurations of the resonating cavities. In the production of so-called open sounds (Öffnungslaute) the tip of the tongue usually does not play an active role. Furthermore, no contact along the sagittal midline of the palate is made in articulating these sounds.38 In an attempt to find a simple system for depicting the vowels according to their physiological features, H E L L W A G in 178 1 39 conceived his wellknown vowel triangle. He oriented himself on the position of the tongue and took its extreme points ([a] = lowest position; [i] = highest front position; [u] = highest back position) for the vertices of his triangle. The other vowels then occupied intermediate positions. Among the numerous later vowel descriptions which have been attempted, the more significant are those of CHLADNI, DU B O I S - R E Y M O N D , W I N T E L E R - S I E V E R S , T E C H M E R , TRAUTMANN, B E L L , S W E E T , P A S S Y , J O N E S , FORCHHAMMER a n d STUMPF.

Later, the vowel quadrilateral,40 as an expansion of the vowel triangle, was found to be a more accurate and useful vowel diagram. However, all such figures doubtless have serious deficiencies, which are not appreciably reduced, even by the use of special supplementary notation. They fall far short of the clarity achieved in the classification of the consonants. 38

„Vokale sind öffnungslaute, zu deren Klangfarbengestaltung die Resonanz der Ansatzräume wesentlich ist, ohne Berührungsfläche in der Mittellinie des Gaumens, sofern sie prosodische Wortmerkmale tragen . . . " . Von Essen, Allgem. u. angewandte Phonetik, 3rd Ed., Berlin 1962, p. 72.

39

Chr. F r . HELLWAG, Dissertatio inauguralis physiolo-gicomedica de formatione loquelae, Tübingen 1781, Reprint (Vietor), Heilbronn 1886. Adopted by the Copenhagen Conference (1925).

40

e

u

u

o

o

a

Front

à a

Hellwag's vowel triangle (1781)

•a Vowel quadrilateral of the so-called cardinal vowels

1.5 P H O N E T I C

TRANSCRIPTION

Our conventional orthographic symbols are not reliable indicators of phonetic value. The degree of abstraction (the " e " in the stem syllable of "beten" is not only quantitatively, but also qualitatively clearly different than the " e " in "betten") is less disturbing than the fact that many sounds are represented by several different letters and that many letters can serve as symbols for different sounds. For example, the voiceless s-sound is represented in German orthography by s, ss or B, while the so-called front and back "ch"-sounds ("ich" and "ach") are represented by the same letters, despite their essential phonetic differences. Furthermore, the voice component sometimes is not indicated by the orthography (voiceless "s" — voiced "s"), while in other cases, such as in the voicing of the front "ch", the same difference is indicated by a completely new letter ("j"). Such inconsistency and inexactitude has made it necessary, for scientific as well as for practical purposes, to devise a system for the unambiguous representation of speech sounds. I n the course of time, a whole series of phonetic alphabets has been developed. Those of L E P S I U S , E L L I S , S W E E T , 41 J E S P E R S E N and F O R C H H A M M E R deserve special mention. Today, a single type of further development, the system of the International Phonetic Association (IPA), is most widespread. Although other systems (such as the diacritic system of the Copenhagen Conference, 1925) may be more suitable for some special purposes, the IPA system is well suited to German. A complete listing of the IPA symbols is given in The Principles of the International Phonetic Association, London 1949. As a preliminary orientation, the symbols which are essential for indicating German speech sounds will be presented and explained. 41

Cf. M. HEEPE, Lautzeichen und ihre Anwendung in verschiedenen Sprachgebieten, Berlin 1928.

1.6 T H E G E R M A N S P E E C H S O U N D S 1.6.1 Survey [b] — voiced, bilabial stop: bitte, Ebbe [p] [m] [d] [t] [n] [g] [k] hl [v] [f] [z] [s] [J] [j] [§] [x] [1] [r] [R] [h]

—

voiceless, bilabial stop: Peter, Mappe, knapp — voiced, bilabial nasal: Mutter, Amme, Lamm — voiced, apico-alveolar stop: Dieb, Ida — voiceless, apico-alveolar stop: Tag, Watte, Welt — voiced, apico-alveolar nasal: Name, Kanne, Mann — voiced, postdorso-velar stop: Garten, legen — voiceless, postdorso-velar stop: Karl, wecken, Sack — voiced, postdorso-velar nasal: Angel, eng — voiced, labio-dental fricative: Wiege, ewig — voiceless, labio-dental fricative: Finger, Yater, Affe, Haff — voiced, predorso-alveolar (dental) fricative: singen, langsam — voiceless, predorso-alveolar (dental) fricative: was, essen, Haß — voiceless, apico-prepalatal fricative with lip rounding: Schüler, Esche, rasch ~~ voiced, dorso-palatal fricative: ja, Ajax — voiceless, dorso-palatal fricative: China, ich — voiceless, postdorso-velar fricative: suchen, ach — voiced, apico-alveolar lateral: Laut, alle, Null — voiced, apico-alveolar trill ( = tongue tip or "front"-r): rot, Ehre, Herr) — voiced, postdorso-uvular trill ( = uvular or "back"-r): rot, Ehre, Herr — voiceless, breathed vowel onset: halt, Ahorn 4 2

The voicing of the sounds is generally subjected to a large degree of variation. This variation is conditioned especially by position and accentuation, but also is a typical feature of many dialects. Full voicing of stops 42

A breathed vowel release is only relevant in interjections such as "ah!"

32

The German Speech Sounds Table of German Speech Sounds La b.

Stops

vcd.

vcl.

[b]

[P]

Dent.-Alv. vcd.

[d]

centr. Fricatives

M

Lv]

P