Proceedings Fonetik 2009 - Institutionen för lingvistik

Proceedings, FONETIK 2009, Dept. of Linguistics, Stockholm Universitynants, or it simply disappeared. Today, /ð/ hasvanished and the former /ð/ is represented by anumber of other consonants or by completeloss, and /d/ does not occur. But /d/ does occurin modern SSF as a result of conscious normativeattempts to promote “good speaking”. Thesecond author, for example, did not have /d/ inhis speech in the early childhood but learnt it atschool. In fully native words, /d/ occurs onlyword medially, e.g. sydän ‘heart’; in recentloanwords it is also found word-initially, e.g.demokraatti, desimaali, devalvaatio, diktaattori.Under the influence of foreign languages,nowadays most notably from English, /b/ and/g/ are entering Standard Spoken Finnish asseparate phonemes in recent loanwords, e.g.baari, bakteeri, baletti, banaani; gaala, galleria,gamma, gaselli. But such words are notyet pronounced with [b] and [g] by all speakers,nor in all speaking situations. On the whole, itcan be concluded that /d/ and especially /b/ and/g/ must be infrequent utterance initially in Finnishdiscourse, especially in informal registers,and consequently prevoicing is seldom heard inFinnish. Instead, utterance-initial stops predominantlyhave short-lag VOT. Even wordmediallyvoiced stops, with the exception of thesemiplosive /d/, are rather infrequent, becausethey only occur in recent loanwords and not forall speakers and not in all registers. Wordfinally— and thus also utterance-finally —voiced plosives do not occur at all becauseloanwords with a voiced final stop in the lendinglanguage are borrowed with an epenthetic/i/ in Finnish, e.g. blogi (< Engl. blog).Our Fenno-Swedish speakers’ /p t k/ hadshort positive VOTs very similar to those observedfor Finnish, assuming that the differencesbetween our utterance-initial /p t k/ andthe word-initial Finnish /p t k/ reported inSuomi (1980) are due to the difference in positionin the utterance. In utterance-initial position,the Fenno-Swedish /p t k/ are unaspiratedwhile the CS Swedish /p t k/ are aspirated. Wesuggest that the Fenno-Swedish /p t k/ havebeen influenced by the corresponding Finnishstops. Reuter (1977: 27) states that “the [Fenno-Swedish] voiceless stops p, t and k are whollyor partially unaspirated […]. Despite this, theyshould preferably be pronounced with astronger explosion than in Finnish, so that oneclearly hears a difference between the voicelessstops and the voiced b, d and g” (translation byKS). As pointed out by Leinonen (2004b), animplication of this normative exhortation is thatspeakers of Fenno-Swedish often pronouncethe voiceless stops in the same way as dospeakers of Finnish. Leinonen’s own measurementssuggest that this is the case.Many of our Fenno-Swedish speakers exhibitedinstability in the degree of voicing in /bd g/. We suggest that this, too, is due to influencefrom Finnish.The Fenno-Swedish speakers’ medial shortand long /d/ had considerably shorter closuredurations than did their medial /b/ and /g/. Inword-final position, this was not the case. TheFinnish semiplosive /d/ occurs word-medially,as does geminate /dd/ in a few recent loanwords(e.g. addikti ‘an addict’). But the Finnishsemiplosive does not occur word-finally. Thus,both short and long Fenno-Swedish /d/ have arelatively short duration in medial position, exactlywhere Finnish /d/ and /dd/ occur, but donot exhibit this typologically rare feature in finalposition where Finnish could not exert aninfluence. With respect to voicing, the Fenno-Swedish short medial /d/ behaved very muchlike Finnish /d/. The mean voiced proportion ofthe occlusion was 90%, and in Suomi (1980:103), all tokens of the medial Finnish /d/ werefully voiced. According to Kuronen and Leinonen(2000), /d/ is dentialveolar in CS Swedish,but alveolar in Fenno-Swedish. Finnish /d/is clearly alveolar and apical (Suomi 1998).Kuronen & Leinonen have confirmed (p. c.)that they mean that Fenno-Swedish /d/ is moreexactly apico-alveolar.Against a wider perspective, the suggestionthat the Fenno-Swedish /p t k/ have been influencedby the corresponding Finnish stops is notimplausible. First, it should be impressionisticallyapparent to anyone familiar with bothFenno-Swedish and CS Swedish that, on thewhole, they sound different, segmentally andprosodically; for empirical support for such animpression, see Kuronen and Leinonen (2000;2008). Second, it should also be apparent toanyone familiar with both Finnish and Swedishthat CS Swedish sounds more different fromFinnish than does Fenno-Swedish; in fact, apartfrom the Fenno-Swedish segments not found inFinnish, Fenno-Swedish sounds very much likeFinnish. Third, Leinonen (2004a) argues convincinglythat CS Swedish has no influence onFenno-Swedish pronunciation today. Leinonencompared what are three sibilants in CS Swedishwith what are two sibilants and an affricatein Fenno-Swedish. He observed clear differencesamong the varieties in each of these consonants,and found little support for an influ-63