Proceedings Fonetik 2009 - Institutionen för lingvistik

Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm Universitymodel, published from 1973 on and now supersededby Lexikon 2000. The only Swedish encyclopediawhere solely IPA transcriptions inbrackets are used appears to be Respons (1997—8), a minor work of c. 30,000 entries, whichis an adaptation of the Finnish Studia, aimed atyoung people. Its pronunciation system is, however,conceived in Sweden.It ought to be mentioned that SAOB (SvenskaAkademiens ordbok), the vast dictionary ofthe Swedish language, which began to be publishedin 1898 (sic!) and is still under edition,uses a system of its own. The above exampleswould be represented as follows: nåbäl 3 ,maŋ 4 kel, ram 3 lø 2 sa, ʃama4n. The digits 1—4represent different degrees of stress and areplaced in the same way as the stress marks insystem (c) above, their position thus denotingquantity, from which the quality of the a’scould, in turn, be derived. The digits also expressaccent 1 (in Mankell) and accent 2 (inRamlösa). Being complex, this system has notbeen used in any encyclopedia.Notation of segmentsFor showing the pronunciation of segments,there was a strong bias, 80%, in favour of theIPA, possibly with some modifications, whereasthe remaining 20% only wanted letters of theSwedish alphabet to be used. Two questionsconcerned the narrowness of transcriptions.Half of the subjects wanted transcriptions to beas narrow as in a textbook of the language inquestion, 31% narrow enough for a word to beidentified by a native speaker if pronounced inaccordance with the transcription. The remaining19% meant that narrowness should beallowed to vary from language to language.Those who were of this opinion had thefollowing motives for making a more narrowtranscription for a certain language: the languageis widely studied in Swedish schools(e.g., English, French, German, Spanish), 47%;the language is culturally and geographicallyclose to Sweden, e.g., Danish, Finnish), 29%;the pronunciation of the language is judged tobe easy for speakers of Swedish without knowledgeof the language in question, (e.g., Italian,Spanish, Greek), 24%. More than one optionhad often been marked.What pronunciation to present?One section dealt with the kinds of pronunciationto present. An important dimension isswedicized—foreign, another one standard—local. Like loanwords, many foreign geographicalnames, e.g., Hamburg, London, Paris,Barcelona, have obtained a standard, swedicizedpronunciation, whereas other ones, sometimes—butnot always—less well-known, e.g.,Bordeaux, Newcastle, Katowice, have not. Thepanel was asked how to treat the two types ofnames. A majority, 69% wanted a swedicizedpronunciation, if established, to be given, otherwisethe original pronunciation. However, theremaining 31% would even permit the editorsthemselves to invent a pronunciation consideredeasier for speakers of Swedish in‘difficult’ cases where no established swedificationsexist, like Łódź and Poznań. Threesubjects commented that they wanted both theoriginal and swedicized pronunciation to begiven for Paris, Hamburg, etc.In most of Sweden /r/ + dentals areamalgamated into retroflex sounds, [ ʂ], [ ʈ ], [ɖ]etc. In Finland, however, and in southernSweden, where /r/ is always realized as [ ʁ ] or[ ʀ ], the /r/ and the dentals are pronouncedseparately. One question put to the panel waswhether etc. should be transcribed asretroflex sounds—as in the recently publishedNorstedts svenska uttalsordbok (a Swedishpronunciation dictionary)—or as sequences of[r] and dentals—as in most encyclopedias. Thescores were 44% and 50% respectively, with anadditional 6% answering by an option of theirown: the local pronunciation of a geographicalname should decide. No one in the panel wasfrom Finland, but 71% of those members withSwedish as their first language were speakers ofdialects lacking retroflex sounds.Particularly for geographical names, twodifferent pronunciations often exist side byside: one used by the local population, andanother, a so-called reading pronunciation, usedby people from outside, and sometimes by theinhabitants when speaking to strangers. Thelatter could be described as the result ofsomebody—who has never heard the namepronounced—reading it and making a guess atits pronunciation. Often the reading pronunciationhas become some sort of nationalstandard. A Swedish example is the ancienttown of Vadstena, on site being pronounced[ˈvasˌsteːna], elsewhere mostly [ˈvɑːdˌsteːna].The reading pronunciation was preferred by62% of the subjects, the local one by 22%. Theremainder also opted for local pronunciation,provided it did not contain any phoneticfeatures alien to speakers of standard Swedish.216