Abstract
It has already been pointed out that the sound-wave emanating from the mouth of a speaker is physically a continuum. Both the wave form as such — in the form in which it appears when registered for instance on an oscillograph film (Fig.14) — and the picture given by a sound spectrograph (Fig.47) on the basis of a formant analysis, show no more than an incessant variation of the different parameters mentioned in Chap. III. Even in cases where typical segments of an apparently steady-state nature may be seen, there is hardly any possibility of indicating an indisputable point where one segment starts and the preceding one ends. The classical distinction established for instance by the phoneticians of last century between typical sound positions and transitional sounds (“glides”, “Gleitlaute”, “Übergangslaute”, etc.; Ellis, Merkel, Sievers) does not hold. Everything is transition. And even if one factor, e.g. voice, or nasality, may be said to cease at a given point (in the original complex curve, or on the spectrogram), this factor only rarely coincides with the other factors which together characterize a given segmental unit (e.g. a vocoid or a contoid). Even the early kymograms revealed the existence of transitional phases without any independent linguistic or communicative function. See e.g. Fig. 46, where the three phonetic features of voice, nasality, and degree of opening overlap in different directions. The apparent steady-state character of the vocalic portion on the kymogram is due to the defective registration of the vocal periodicity by the kymograph which normally only registers the fundamental. The same portion of speech is rendered in Fig.47 by a sound spectrograph.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Reference
It sometimes seems convenient to use text,in accordance with glossematic terminology, about any kind of linguistic utterance or message, long or short, spoken or written.
We have already stressed the fact that the result of an auditory segmentation always depends to a great extent on the previous experiences — i. e. mostly the phonemic background — of the listener. These play an important part and are, in the case of so-called naive listeners, decisive.
Malmberg, Structural linguistics In many languages, there are examples of a free variation between vowel and consonant, particularly between /i/ and /j/,/u/ and /w/. This is the case in French where /j/ in bien, /w/ in moi have often been interpreted as non-syllabic variants of the /i/- and lu/-phonemes, on account of the free variation in words like lier, louer,phonetically [lie] or [lie], [lue] or [lwe]. It seems preferable, however, to look upon such examples as so called word -variants, with variation in the phonemic make up of the morpheme (just as there is variation between [’aida] and [’i:ds],sc. either,in certain kinds of English). The definition of the “semivowels” in question as structural vowels leads to serious difficulties which cannot be discussed here. We find exactly the same problem e.g. in Spanish.
It is arbitrary because the point of departure chosen lies outside the object of analysis which in this case is a communication system, not a physical event.
The principle of any phonetic alphabet — of which that of the “International Phonetic Association” (dating from 1886) is the best known and the most widespread — is “one symbol for one sound”. A symbol supposes a class — not just an individual —, and as the classification of the speech events in the sound-wave into classes is necessarily arbitrary, as long as the substance structure is the only criterion, the choice of a given symbol for a given sound must always remain controversial. In reality, phonetic transcription has always been, and always must be, at least partly phonemic. When opposing two sounds by choosing different symbols and when identifying them by using the same symbol phonetic transcription has always to a certain extent taken into account linguistic function and phonemic points of view. Distinctions which are commonly used in the best known languages have got their symbolic counterparts in the international alphabet, but not rare distinctions.
Distribution alone is not sufficient to identify them, because other phonemes (e.g. /p/ and /k/) would be possible in the same position. Function alone does not permit to decide if initial /t/ is to be identified with final /t/ or with final /p/,etc.
Though it would for certain purposes be very useful to generalize a terminology referring to acoustic characteristics when talking about speech sounds, and a corresponding terminology with reference to physiological events when talking about articulations and articulatory possibilities and types, we prefer to use, for the sake of convenience, the current terms, though it will be clear from what has been said here that they are not very appropriate.
HARRIS uses /sy/ for /sj/.
In most parts of Sweden the (postvocalic) cluster rs is pronounced as one single fricative sound, a kind of IF,and as the/f/-phoneme (in many people’s speech identical phonetically with vs) is only initial, except for loan-words and onomatopoetica, ELERT has proposed to interpret any /f/ as a cluster /rs/ in any position and consequently transcribes 45 [fey:] as /rse:/,etc.
This is only a particular instance of a general phonemic law which implies a reversed proportion between the number of phonemes and the length of syntagms. A language with a small number of phonemes, in order to obtain the same number of signs as a richer language, has to make its syntagms longer (the length measured in number of phonemes). It is only natural that it should be valid also for the graphemic level.
Traditional phonetics taught that consonants are often articulated within a resonance room already formed for the following vowel, that e.g. the lips are rounded and protruded and the back of the tongue drawn backwards at the same time as, or even before, the tongue closes the air passage for [t] in [tu:],and that, in a corresponding way, the vocal tract remains in its vocalic position, as far as this is possible, when a postvocalic consonant is pronounced. We now have the necessary spectrographic evidence for these phenomena on the acoustic level. We know that vowel formants are to be seen in the consonant spectra, and we know that, in a certain sense, the syllable may be defined as an acoustically determined unit, dominated by the vowel and by its resonance, giving to the surrounding consonants the specific colour which indicates their participation in the acoustically determined syllabic unit. This particular vocalic colour of the consonants in a syllable is, of course, predictable from the vowel quality, and consequently non-distinctive. Such consonant colour may, however, in cases of reduced vowel length and/or intensity, be the essential, or sometimes the only manifestation of a vocalic, phonemic opposition. Thus, a palatalised [t’] with an i-colour in its explosion (and aspiration) phase, and distinct from an ordinary [t],may under certain structural conditions be regarded phonemically as a sequence /t/ +/i/. Consequently what is simultaneous on the substance level often has to be interpreted structurally as a linear sequence of two consecutive units in a determined order (cp. p. 84–86). If this interpretation is accepted, the difference between [t] and [t’] in the given system may be said to be distinctive but not phonemic (cp. p. 81), since the phonemic opposition is one between /i/ (or other vowel phonemes) and zero.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1963 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Malmberg, B. (1963). Segmentation. Forms of Expression. Oppositions and Distinctions. Paradigmatic Structures. In: Structural Linguistics and Human Communication. Kommunikation und Kybernetik in Einzeldarstellungen, vol 2. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-13066-7_5
Download citation
DOI: https://doi.org/10.1007/978-3-662-13066-7_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-662-13067-4
Online ISBN: 978-3-662-13066-7
eBook Packages: Springer Book Archive