Taal en Tongval. Jaargang 29
(1977)– [tijdschrift] Taal en Tongval–
Auteursrechtelijk beschermd
[pagina 77]
| |||||||||||||||||||||||||||||
Besprekingsartikel‘Atlas Linguarum Europae: Introduction’, A. Weijnen et al.Ga naar voetnoot1.
The Introduction to the Atlas Linguarum Europae, which consists of a detailed description of this project for a European linguistic atlas, contains the following sections, which we will discuss one by one:
The first chapter is very short, consisting of not much more than three pages, and deals with the history of international linguistic geography as a linguistic discipline. The authors are at pains to stress that this is no new discipline, and that the concept of a European linguistic atlas is nearly as venerable, dating from before 1930. In the second chapter an attempt is made to justify the atlas as a worthwhile object within linguistics. Weinreich's Languages in Contact is seen as a kind of stepping-off point for what will be intended. The aim of the atlas is to discover more about the influences languages have had on each other in as wide as possible a range of linguistic phenomena, from phonetics to the study of idiom, and semantics. The problems the atlas is supposed to attempt to clarify are to be not only those of a typological nature but also those connected with the theory of universals. As far as the last is concerned it seems to us that a purely European atlas can only produce definitive evidence of a purely negative character. On the positive side any evidence cannot be more than suggestive. The intention is to work on ‘local’ dialects and not on regional varieties of standard languages. Rather than the justification here given - that of Rudolf Grosse ‘that dialect vocabulary reveals how its speakers work and live, how they think and feel, and that it provides us with a valuable basis for cultural and historical research’ - which seems to us to be equally applicable to all ‘mothertongues’, whether they represent traditional dialects or varieties of the modern | |||||||||||||||||||||||||||||
[pagina 78]
| |||||||||||||||||||||||||||||
standard languages, unless of course it is to be assumed that the life and work, thoughts and feelings, etc. of speakers of traditional dialects are to be more highly valued, we might suggest that traditional local dialects are worthy as a field of investigation for two reasons: - firstly that they are the end-product and final remnant of a process of largely gradual change over a long period, which process in some places has been, in most places is in the process of being, and in nearly all will be broken off by the more or less sudden replacement of these forms of speech by a relatively small number of standard languages, and secondly - and this is intimately connected with the first factor - the sociolinguistic changes now in progress are of such a violence as to effectively obliterate many of the interlingual and interdialectal influences which have left their trace in the past on the traditional dialects. In other words the old disparity is being replaced with a new uniformity, with a consequent vast loss of information - including of course precisely the sort of information relevant for the problems the atlas is trying to elucidate. In view of these facts the necessity to seek out the most archaic forms of local dialect is obvious. It seems to us that the atlas collaborators, and indeed much of modern dialectology, exhibit the tendency that, since Labov, it is necessary to have a sociolinguistic dimension in all dialectological work. While it is quite obvious that sociolinguists have contributed to a new and revealing understanding of the processes of linguistic change and dialectal and linguistic replacement, it is just not true that all language study must have a sociolinguistic dimension. This is no less true of dialectology than other linguistic fields. Dialectologists must make the decision to what extent their work must be sociolinguistically oriented in terms of what it is they want to find out. They must not feel that unless they pay at least lip service to sociolinguistics they are necessarily old-fashioned. Clearly in terms of the aim of the atlas the recording at the one location of typical examples of different class, occupational or age dialects may reveal important trends. These will however by no means be restricted to those involved in the rivalry between dialect and standard language as might seem to be suggested in the Introduction. In short, the collaborators on the atlas do not, we think, make it clear enough that a sociolinguistics-oriented approach is only useful if and inasmuch as it furthers the aims of the atlas. To say that the atlas ‘must include the sociolinguistic dimension, even if only to emphasize the fact that the use of dialect involves sociolinguistic implications’ seems very odd. No service at all is performed by the dragging in of the Saussurian dichotomy between langue and parole in this connection. The different linguistic systems coexisting in one location or in one speaker are no less aspects of langue than any individual linguistic system. When different systems are used is of course a question of parole but would seem to be of little relevance for the atlas. The idea that ‘the ALE can hope to make a contribution to sociolinguistics’ by helping ‘the reader understand the general tendencies in the use of different linguistic systems’ would involve a great deal of research with little spin-off for the stated aims of the atlas. It seems better to restrict the work of the atlas to the things which will be productive in terms of these aims and not to waste time trying to be all things to all men. The third chapter is very brief and concerns the organization of the various workers associated with the project. We mention only one point. It is stated that ‘each scientific investigator may have full access to the materials’. If this | |||||||||||||||||||||||||||||
[pagina 79]
| |||||||||||||||||||||||||||||
is intended to mean that full access to the materials is to be restricted to those who have participated in the research, then this seems a wrong restriction. At least after the publication of the atlas the materials should be accessible for all linguists who are interested. The fourth chapter concerns the questionnaires. The first is to be lexicological because it is believed that this will provide the most promising results. One restriction we find strange is that the questionnaire be restricted to items used in the Slavic dialect atlas, the OLA, because, it is said, it would be impossible to repeat the surveys carried out in ‘the huge Slavic world’. The authors say that this is ‘easily understandable’. We find it difficult to see why such work should be less easily repeatable there than anywhere else. In the second questionnaire there will also be a lexicological portion, dealing with restricted domains of vocabulary (agriculture, fishing, etc.). A discussion is provided on the problems involved in the selection and phrasing of the questions. How other types of information which is also to be gathered in the second questionnaire - phonological/phonetic, etc. - is to be dealt with has apparently still to be discussed. It is intended to limit the project to two surveys for principally financial reasons. As the authors say, the very fact of the existence of the present atlas will make future projects along the same lines easier to carry out. The fifth chapter concerns the choice of base map and localities for the survey. The editorial board had difficulty deciding on the eastern boundary of Europe and have arbitrarily chosen to extend the map as far as Kuibyshev. One wonders why the conventional eastern boundary of Europe - the Urals - was not chosen. It is regarded as a problem whether to indicate on the base map to what linguistic area each point belongs. While this may sometimes be difficult (more than one language spoken at one point) it is obviously essential to know which language one is dealing with. As is said there are problems in deciding linguistic boundaries. They give the example of the Dutch-German language boundary in SE. Limburg. Here the dialects of Kerkrade (Netherlands) and Aachen (Germany) are very close. Clearly some arbitrary decision must be made - say in terms of some isogloss(es). At any rate national pride or prejudice should play no role. If the dialect of Kerkrade is closer to clearly German dialects than to clearly Dutch dialects, which seems in fact to be the case, then to call it German would seem to be the logical answer. To adopt Meillet's criterion - ‘the feeling and desire of belonging to one and the same linguistic community’ - as is here decided, would probably lead to inconsistent and inconclusive results. The density of the network chosen - that neighbouring points should be not less than 30 km. apart - would seem to create problems. As is well-known relative linguistic diversity varies considerably in different parts of Europe. Account is taken of this by allowing an increase of 30% in the density in areas of great diversity, but we wonder whether this is sufficient. 5 points may well be sufficient for Iceland, but are 9 (or even 12, allowing 30% extra) sufficient for the diversity of German-speaking Switzerland? In calculating the number of points of enquiry for each country no account seems to been taken of the shape of a country. One would have imagined for instance that Greece with its irregular coastline and numerous islands would qualify for extra points, but this does not seem to be the case. | |||||||||||||||||||||||||||||
[pagina 80]
| |||||||||||||||||||||||||||||
There are also some strange deviations from the average density. Four countries showing fairly normal ratios are the following:
Inexplicably large deviations occur in at least the following two cases:
In other words England and Wales have only about a third of the necessary number of points, while the Netherlands has nearly double the normal number. A further problem, which is not considered, is constituted by those areas in Europe which are bilingual in the sense that local dialects of more than one language are present, eg. Cyprus, large parts of Eastern Europe, parts of Ireland, etc. No mention is made of recently displaced groups of dialect-speakers, of which the most significant group is formed by the former German-speaking inhabitants of much of Pomerania, parts of Silesia, East Prussia, etc. It would be unfortunate if the valuable information which might be provided by the dialects of these groups were to be ignored for political or other reasons. Noteworthy is the exclusion of Scottish Gaelic from consideration. No mention is made of the Isle of Man, or the Channel Islands, although these might be intended to be subsumed under England. No mention is made either of the problems associated with the mapping of widely dispersed groups such as Yiddish, and Romany. It is to be hoped that the exclusion is only apparent, since both these languages or groups of languages illustrate well precisely the kind of phenomena the atlas is setting out to discover. The sixth chapter concerns the methods of information storage rather than the methods of information gathering. As is said, the advantages and disadvantages of the various methods of the latter are well-known. We have no comments on this section. The seventh chapter concerns which system of phonetic transcription to use. The chapter opens with a discussion on whether the transcription should be made in terms of phonetic or phonological units. The collaborators choose a phonetic notation for two main reasons, the lack of knowledge of the phonemic systems applying, and the generative phonological attack on the phoneme. While we accept their conclusion it would have been useful at this stage to provide a statement of the surface oppositions occurring at each point. Presumably this information will appear later in the second part of the survey anyway, however. They state that the intention is to provide a transcriptional system which will reveal all ‘systematic phonetic distinctions’ (ie. all distinctions that would be made in the output of a generative phonological description). This seems to be a very uncertain solution since generative phonology has in fact never been so clear about just how much phonetic detail it was intended to specify. They say that systematic phonetics represents the ‘perceived sounds of speech’ (mistranslated in the English version as ‘the sounds of perceived speech’). This definition smacks rather of performance, | |||||||||||||||||||||||||||||
[pagina 81]
| |||||||||||||||||||||||||||||
but even if we were to alter it to ‘the perceivable sounds of language’ it would not be automatically acceptable, since there might be systematic aspects of the sound-pattern of a language which are not directly perceivable, but need to be described. In any case it is unclear how this will enable us ‘to recover the phonological “distinctive features” to which Roman Jakobson, and many others, attach such great importance’. How are these features, which have both phonetic and phonological functions, to be recovered solely from phonetic data? This would form, in our opinion, an insoluble problem. Unsatisfactory is the justification for the rejection of the most widely known system of phonetic transcription in Europe, the IPA phonetic alphabet. The fact that Zwaardemaker and Eijkman, two Dutch phoneticians, found it insufficient for use with non-European dialects, seems hardly relevant for a European project, and in any case this is by no means a universally shared view. That ‘specialists in European dialects might (our italics, NS) have encountered similar difficulties and made the same objections’ seems equally irrelevant. Instead of a system where the vowel symbols are associated with fixed values and are not defined in terms of the vowels of some language or other, we are offered a system where the vowel symbols are defined in terms of the vowels of French, English, Russian and even Dutch. The usual problems associated with such systems are of course present. Which type of English is meant here? Presumably R.P., although we are not told. It seems further very unlikely that many European dialectologists are familiar with the precise phonetic value of the vowel in the Dutch word dom. One difference from the IPA and most other systems is that five basic height distinctions are recognized for vowels instead of four. Any greater precision will, we fear, be only apparent, judging from the definitions of these vowels. For front vowels they provide the five symbols i e E ɛ a. The arrangement of these symbols is such as to indicate that the difference between these vowels concerns the parameter of height. This may well be the intention but cannot follow from the definitions of the symbols. e is defined as the vowel of French été, while E is defined as the vowel of English pit. Ignoring the fact that this latter vowel is centralized, and thus not a fully front vowel, there is still a serious problem. According to Westermann and Ward (1933), among others, the latter vowel is actually higher than the former. Why are no symbols provided for non-low back unrounded vowels, which certainly occur in European dialects, or central rounded vowels? The consonant symbols provided represent a small subset of the IPA consonants, the main difference being the dropping of various symbols useful for European languages, eg. symbols for palatals (apart from j), alveolo-palatals (as distinct from palatoalveolars), uvulars and pharyngeals, all of which occur distinctively in Caucasian languages, symbols for retroflex consonants, which are required for dialects of many European languages - English, Gaelic, Scandinavian languages, Caucasian languages, etc., symbols for more than one type of l or r, etc. They justify the adoption of this simplified system on the basis that they are more concerned with ‘lexicological types which can in most cases be identified with reference forms taken from standard languages’, rather than with ‘the exact phonetic notation of each form’. How this links up with the fact that their transcriptional system is intended to be able to reveal all systematic phonetic dis- | |||||||||||||||||||||||||||||
[pagina 82]
| |||||||||||||||||||||||||||||
tinctions is unclear. It is quite obvious that this system is in fact incapable of expressing even some phonological oppositions, such as the frequent opposition between velars and uvulars in Caucasian languages. In fact a serious problem is bound to arise with Caucasian languages. In the case of the generally relatively short morphemes of these languages, and the tendency towards simple vowel systems, a high functional load is thrown onto the consonants, and the consequence of the use of the atlas transcription system would be a high degree of apparent homonymity. It is interesting that the phonetic categories that are recognised are precisely those that are sufficient for transcribing phonologically the major European languages. This suggests a certain glotto-centrism. McIntosh (1961) is quoted, who doubts if phonetic transcriptions are always necessary in lexical research. McIntosh may be right, but this is hardly a justification for using a simplified phonetic system. As for the ease of identifying lexicological types with ‘reference forms taken from standard languages’, this may be so with English, or German, or Russian. We doubt if it will be so easy in Andi, Bzyb, or Bats, especially with an underdifferentiated phonetic notation in languages where what are small differences to W.-European ears may be of major significance. It is stated ‘that degrees of nasalization and length cannot be considered as phonemic; hence there is no need to distinguish them’. While we have never seen any evidence to the contrary from Europe, this cannot be assumed a priori. Ladefoged (1971) gives examples of languages with four ‘phonemic’ degrees of length, and three degrees of nasalization (zero, light, and heavy). In short the approach to transcription in the atlas does not inspire confidence. The eight chapter deals with the question of mapping the lexicological data. The plan is to use symbols on the map. The question as to whether these symbols should represent phonetically similar forms, or etymologically related forms is then discussed. Because of the desire to associate under one symbol forms like Italian ruota, French roue and Rumanian roată (‘wheel’) they have chosen the second method. This method appears to have dangers, however. For instance, if two neighbouring languages have for a particular meaning cognate terms, then they will be represented by the same symbol. If border dialects of the one language have replaced the inherited form with that of the other language then this would be obscured under the strict etymological method. Apparently the authors are prepared to make an exception if a borrowing has crossed national boundaries but not otherwise. This is completely ridiculous - the atlas is concerned not with nations but with languages. A borrowing from French into Breton is surely more significant than a borrowing from standard French into Walloon. In the final section of this chapter it is made evident that the authors do not understand the purpose of generative phonology. They say ‘One might perhaps ask whether a final solution cannot be found, with the generative phonologists, by writing in each case the underlying form in terms of systematic phonemes. But then such questions arise as: How far back should we go in determining underlying forms? For German Tag, is the underlying form dag; or, older, is it dagaz; or, still older, is it dhoghos? Further: there is no way of generating, in any regular way, many of the underlying forms that would be needed eg. for ombre and lommer, or for the various Brabantish terms for the sill of a stable door’. The identification of generative rules with historical processes and of underlying forms | |||||||||||||||||||||||||||||
[pagina 83]
| |||||||||||||||||||||||||||||
with earlier historical forms indicates a confusion between synchronic and diachronic processes. The underlying forms that would be set up in generative phonology would be those that resulted in the simplest grammar capable of explaining the synchronic alternations and patterns. What the oldest recoverable ancestral form is, is completely irrelevant. A generative phonological description of a Brabantish dialect would of course provide an underlying form for the word for the sill of a stable door, from which the surface form would be generated according to the phonological rules of that dialect. On the two appendices we have nothing to say. In conclusion let us say that we are entirely sympathetic with the aims of the atlas, which we see as a potentially valuable source of evidence on the processes of linguistic interinfluence. The many criticisms we have made here have been intended in a constructive sense, and involve not the fundamental aims of the atlas, but concern the theoretical justifications for various aspects of the atlas, and also a number of methodological points. N.S.H. Smith, Instituut voor Algemene Taalwetenschap Spui 21 Amsterdam. | |||||||||||||||||||||||||||||
Bibliography
|
|