Phonetic preconditions for historical sound change - evidence from the dialects
Olle Engstranda, Sven Björstena, Gösta Bruceb and Anders Erikssonc
aDept. of Linguistics, Stockholm University, bDept. of Linguistics, Lund University,
cDept. of Phonetics, Umeå University
Universal perception and production constraints may constitute phonetic preconditions for historical sound change. It is pointed out that the traditional literature on the Swedish dialects provides a rich source of information in evaluating the diachronic role played by these constraints.
Historical sound change is known to take place as a consequence of factors such as language-contact and borrowing as well as through phonetically motivated, language-internal developments. As a well-known example of the latter, consider the high and low pitches on vowels which are considered to follow as fortuitous consequences of the production of voiceless and voiced obstruents, respectively, and which may lead to the historical development of tonal contrasts on vowels (e.g., Hombert et al., 1979; Svantesson, 1983). It is of a general phonetic interest to evaluate further the extent to which historical sound change can be explained in terms of universal perceptual and articulatory constraints. In addition, an improved knowledge of the phonetic mechanisms of change may provide historical linguistics with criteria for well-informed choices between alternative hypotheses (Ohala, 1993).
Dialects of ‘the same language’ provide a convenient basis for the study of sound change since dialectal variation can be observed against a relatively homogeneous linguistic background. Since the traditional literature on the Swedish dialects provides a wealth of careful and detailed phonetic observations, inferences can be made as to which sound changes must have taken place. It is suggested that these inferences can provide a reliable ‘answer key’ for evaluating the extent to which universal perception and production constraints form the phonetic preconditions for historical sound change. This is briefly exemplified below under the headings ‘auditory assimilation’, ‘articulatory assimilation’, ‘epenthesis’ and ‘quantity shift’.
From an auditory point of view, velar stops such as /k/ and labiovelar semivowels such as /w/ share the feature ‘grave’, whereas dentals/alveolars such as /t/ are specified as non-grave, or ‘acute’ (cf. Jakobson et al., 1965). We thus hypothesize that sound change will occasionally take place in terms of regressive assimilation of the grave feature such that, e.g., the sequence /tw/ turns into /kw/. This would provide auditory enhancement of the grave quality of the stop-semivowel cluster. Is there evidence that such an assimilation may result in historical sound change in the world’s languages? Judging from the dialect literature, the answer is yes since in many dialects the sequence /tw/ has changed to /kw/, sometimes with a subsequent change /kw/ > /kv/. For example, the Lappfjärd dialect (spoken in the northwestern coastal region of Finland) has kvättas ('be washed') for Standard Swedish tvättas (Ivars, 1996). This change has also taken place in other regions of the Swedish dialect area (see, e.g., Pamp, 1978).
It is conceivable that a related auditory enhancement effect also underlies the well known fact that labialization occurs much more frequently in velar segments than in dental or alveolar segments in the world's languages. This is reflected in table 1 which shows the number of the 451 UPSID languages (Maddieson, 1984; Maddieson & Precoda, 1989) with plain/labialized dentals/alveolars and plain/labialized velars, respectively. It can be seen that the incidence of plain dentals/alveolars and velars is comparable whereas labialization occurs much more frequently with the velars than with the dentals/alveolars.
Table 1. Number of UPSID languages with labialized vs. plain dentals/alveolars and labialized vs. plain velars, respectively.
From an articulatory point of view, there is no obvious motivation for this pattern since dentals/alveolars and velars, both being lingual, are equally compatible with the labialization gesture. From an auditory point of view, however, labialization and velar articulation have mutually enhancing auditory qualities (they are flat and grave, respectively, in the feature system of Jakobson et al., 1965); in contrast, labialized dentals/alveolars are specified as flat and non-grave (acute), i.e., a less unambiguous feature combination from an auditory point of view.
Some vowel-consonant combinations involve diametrically opposed tongue shapes, e.g., sequences of high front vowels and strongly retroflex (possibly subapical) consonants such as [} ] which is observed in many Swedish dialects. Since such VC sequences involve quite extensive articulatory transitions, they would be low-valued in an articulatorily based markedness theory and would tend to occur infrequently in the world’s languages (cf. Willerman, 1994). Clearly, there is also a substantial accompanying acoustic transition which will affect the auditory quality of the vowel. From the point of view of a general theory of markedness, it is reasonable to assume that gestures that are both articulatorily costly and auditorily detrimental are particularly low-valued and thus infrequent in the world’s languages. This assumption is corroborated by the Swedish dialects in which the retroflex [} ] is typically not preceded by high or high mid front vowels such as /i/, /y/ and /e/ (e.g., Eriksson, 1961, p. 77). Thus, with few exceptions, Swedish dialects with [} ] show the alternation [sa:} ] 'hall' vs. [si:l] 'strainer'. This articulatory constraint on vowel-retroflex sequences is probably related to the allophonic lowering of mid front vowels before /r/ and (moderately) retroflex consonants in Standard Swedish as well as to a general vowel lowering effect in connection with /r/ as observed in other languages (e.g., Lindau, 1985, p. 158).
Stop epenthesis (poststopping) in nasal and lateral consonants probably occurs fortuitously in many languages and is reported to have become fossilized in some languages such as French which has chambre from Latin camera (e.g., Ohala, 1992). The underlying articulatory mechanism, anticipatory denasalization/delateralization, is relatively simple. In the nasals, for example, raising the soft palate and thus blocking the velopharyngeal passage just prior to the oral release will result in a stop at the same place of articulation. In general, poststopping occurs as a consequence of a slight change in the relative timing of articulatory events. It can be speculated, however, that this kind of stop epenthesis is not exclusively an effect of bad timing; it might also serve the perceptual purpose of enhancing place of articulation. Specifically, this might be the case for the nasals since many languages have nasals at different places of articulation. Do nasals tend to be poststopped to a greater extent than laterals?
Is there dialectal evidence for stop epenthesis as a process of historical sound change? The answer is again yes. In many dialects, nasals and laterals are poststopped is some contexts, i.e., they are released by a full, homorganic stop rather than nasally or laterally. For example, the Lappfjärd (W. Finland) dialect has /mbl/, /mbr/, /ldr/ in word medial position where Standard Swedish has /ml/, /mr/, /lr/, e.g., /sa:mblast/ for /samlades/ 'gathered' (Ivars, 1996). In other dialects, this stop epenthesis has developed a step further and undergone additional vowel epenthesis such that, for example, Standard Swedish /mr/ corresponds to /mber/ in, e.g., /semre/ vs. /semberi/ 'worse'. This is the case in the Pernå dialect, which is also spoken in the western region of Finland (Ivars, 1996). This phenomenon, too, appears to have a phonetic motivation. In order to bring about an audible stop release as required in, e.g., the sequence /mbr/, a certain amount of air pressure must be present behind the lip closure. This can be achieved only given an unobstructed air passage between the glottis and the lips. Thus, in this example, the following /r/ can not be fully anticipated during the labial /b/ closure. In consequence, an epenthetic vowel will fortuitously follow the release; and occasionally this fortuitous epenthesis may be perceptually re-evaluated in some dialects giving rise to a sound change of the kind observed in the Pernå dialect.
The following observations are frequently cited as examples of universal tendencies in the world’s languages: a) Low vowels tend to be longer than non-low vowels, cet. par. This appears to be related to the relatively great amount of jaw opening required by low as compared to non-low vowels; specifically, an extensive opening-closing gesture is time-consuming since the jaw is a relatively massive, and thus inert structure (cf. Lindblom, 1967); and b) vowels preceding voiced stops tend to be longer than vowels preceding voiceless stops. The reason for this effect is, however, not completely clear (cf. Chen, 1970), but it can be noted that there is frequently a VC complementarity such that voiced stops tend to be shorter than voiceless stops. The reason for the latter tendency may be an aerodynamic one (see Ohala, 1983), and the complementary durational VC pattern might be related to a tendency to maintain a relatively even rhythmic structure.
Does the tendency to lengthen low vowels and vowels before voiced consonants constitute preconditions for historical sound change? As previously pointed out by Riad (1992), a classical study by Hesselman (1901) suggests that these factors have played a crucial role in the quantity development of many Swedish dialects. Most varieties of present-day Swedish have a two-way quantity contrast operating on lexically stressed vowel-consonant (VC) domains. This system has developed historically from a four-way contrast including short VC syllables. Hesselman concluded, among other things, that 1) in some dialects, low vowels (/a/ and /ae/) have lengthened with the consonant remaining short (V:C), while non-low vowels have remained short with lengthening of the consonant (VC:); 2) in other dialects, low vowels were always lenghthened while non-low vowels were lengthened when followed by a voiced stop, but not when followed by a voiceless stop; in the latter cases, the stop was lengthened. Again, then, there seems to be a clear parallelism between universal phonetic microvariation and patterns of historical sound change.
It is frequently worthwhile consulting the Swedish dialect literature for effects of phonetic constraints on historical sound change.
Work supported by The Bank of Sweden Tercentenary Foundation, grant 1997-5066:01-02 to the project Phonetics and phonology of the Swedish dialects around the year 2000 (SWEDIA 2000).
Chen, M. 1970. Vowel length variation as a function of the voicing of the consonant environment. Phonetica 22, 129-159.
Elert, C.-C. 1964. Phonologic studies of quantity in Swedish. Uppsala: Almqvist & Wiksell.
Eriksson, M. 1961. Svensk ljudskrift 1878-1960. Svenska landsmål och svenskt folkliv [Archives des traditions populaires suédoises], No. 280.
Hesselman, B. 1901. Skiss över nysvensk kvantitetsutveckling. [A sketch of the development of quantity in Modern Swedish.] Språk och stil 1, 10-25.
Hombert J.-M., Ohala J.J. and Ewan W. 1979. Phonetic explanations for the development of tones. Language 55, 37-58.
Ivars, A.-M. 1996. Stad och bygd. Finlandssvenska stadsmål i ett regionalt och socialt perspektiv. Folkmålsstudier 37.
Jakobson, R., Fant, G. and Halle, M. 1965 (6th printing). Preliminaries to speech analysis. The distinctive features and their correlates. Cambridge, Massachusetts: The MIT Press.
Lindau, M. 1985. The story of /r/. In Fromkin, V.A. (ed.), Phonetic linguistics, pp. 157-168. Orlando, Florida: Academic Press.
Lindblom, B. 1967. Vowel duration and a model of lip mandible coordination. Speech Transmission Laboratory, Quarterly Progress and Status Report, Royal Institute of Technology, Stockholm (STL-QPSR), 4, 1-29.
Maddieson, I. 1984. Patterns of sounds. Cambridge: Cambridge University Press.
Maddieson, I. and Precoda K. 1989. Updating UPSID. Journal of the Acoustical Society of America, Suppl. 1, Vol. 86, S19.
Ohala, J.J. 1983. The origin of sound patterns in vocal tract constraints. In MacNeilage, P.F. (ed.), The production of speech, 189-216. New York: Springer.
Ohala, J.J. 1992. What is the input to the speech production mechanism? Speech Communication 11, 369-378.
Ohala, J.J. 1993. The phonetics of sound change. In Jones C. (ed.), Historical linguistics: Problems and perspectives, 237-278. London: Longman.
Pamp, B. 1978. Svenska dialekter. Stockholm: Natur och Kultur.
Riad, T. 1992. Structures in Germanic prosody. A diachronic study with special reference to the Nordic languages. Stockholm University, Department of Scandinavian Languages.
Svantesson J.-O. 1983. Kammu phonology and morphology. Travaux de l'Institut de linguistique de Lund XVIII. Lund: Gleerup.
Willerman, R. 1994. The phonetics of pronouns: articulatory bases of markedness. PhD dissertation, University of Texas, Austin.