9 ISBN 978-951-42-8983-5 ISSN 1796-4725 S u o m i , T o i v a n en Kari Suomi, Juhani Toivanen & Riikka Ylitalo & Y l i t a l o FINNISH SOUND STRUCTURE F I N N IS Phonetics, phonology, phonotactics and prosody H S O U N D S T R U C T U R E : P h o n e t ic s , p h o n o lo g y , p h o n o t a c t ic s K A a N n S d I R p A ro IM s O o d A y H O N E N STUDIA HUMANIORA OULUENSIA 9 KARI SUOMI, JUHANI TOIVANEN & RIIKKA YLITALO FINNISH SOUND STRUCTURE Phonetics, phonology, phonotactics and prosody UNIVERSITY OF OULU, OULU, FINLAND 2008 Copyright @ 2008 Studia humaniora ouluensia 9 Editor-in-chief: Prof. Olavi K. Fält Editorial secretary: Lect. Elise Kärkkäinen Editorial Board: Prof. Maija-Leena Huotari Prof. Sari Kunnari Prof. Anni-Siiri Länsman Lect. Taina Kinnunen Lect. Elise Kärkkäinen Lect. Jari Okkonen Publishing office and distribution: Faculty of Humanities Linnanmaa P.O.Box 1000 90014 University of Oulu Finland ISBN 978-951-42-8983-5 ISSN 1796-4725 (URL: http://herkules.oulu.fi/issn17964725) ALSO AVAILABLE IN ELECTRONIC FORMAT ISBN 978-951-42-8984-2 (URL: http://herkules.oulu.fi/isbn9789514289842) Typesetting: Heikki Keränen Cover Design: Raimo Ahonen OULU UNIVERSITY PRESS OULU 2008 Preface This book attempts at a description of the sound structure of Standard Spoken Finnish, intended for an international audience familiar with the basic concepts of phonetics, phonology and linguistics. No prior knowledge of Finnish, a Finno- Ugric (and ultimately Uralic) language, is presupposed. The book describes the phonemes and their allophones, the phonotactics and the prosodic system of the language, and it is based on the corresponding parts of our textbook in Finnish (Suomi, Toivanen & Ylitalo 2006), albeit updated and adapted in many ways to the intended readership. To our knowledge, no equally comprehensive description of Finnish sound structure is currently available. The description of the prosodic system is to a considerable extent based on our own recent research. Including phonotactics along with segmental phonetics and prosody may seem an odd decision. However, we feel that the inclusion of phonotactics is warranted for at least four reasons. Firstly, Finnish is a full-fledged quantity language in which both consonant and vowel durations are contrastive, independently of each other, and of word stress, and according to the standard phonological interpretation, the quantity opposition is a matter of phonotactics: in a given word position, there may be a contrast between one phoneme or a sequence of two identical phonemes. Secondly, given the standard interpretation of the quantity opposition, sequences of up to four vowel phonemes in a word are possible; across a word boundary, even longer sequences of vowels can occur. Thirdly, Finnish has vowel harmony, as a result of which only certain vowels can co-occur in a word; in our view, vowel harmony is best described as a phonotactic restriction, although it is sometimes treated as a prosodic property. Fourthly, in many descriptions of Finnish available in English, the phonotactics of especially word-initial consonants is described in a way that is clearly unrealistic in view of the situation in modern Standard Spoken Finnish. We therefore feel that excluding these phonotactic properties would result in an inadequate picture of the sound structure of the language. The book aims at a description of the Finnish sound structure, and the descriptive frameworks used in the various chapters of the book are mostly intentionally as theoretically shallow and neutral as has seemed possible. The intention is to provide primary data on Finnish, data that researchers with different theoretical inclinations hopefully find useful for their own purposes. 3 Segment durations serve both lexical and postlexical functions in Finnish, in contrast to many other languages, and we therefore often repeat rather detailed numerical results of the original papers. In doing this, we only report results whose statistical significance has been tested in the original papers. It has been our goal that the reader should get an adequate idea of the factors influencing segment durations without necessarily consulting the original papers. Although morphophonological alternations of many kinds are a characteristic property of Finnish, we do not describe them systematically, but offer a glimpse at them in Chapter 1. For example, Grade Alternation is a common type of morphophonological alternations in Finnish. The nominative singular form of ‘lamb’ is lammas, the genitive singular form is lampaan. That is, the weak grade /mm/ sequence alternates with the strong grade sequence /mp/. We do not describe such alternations systematically for two reasons. Firstly, morphophono- logical alternations are properties of morphemes and, phonologically and phonetically, segments that participate in alternations in their respective morphemes do not differ from those segments that do not participate in any alternation in their respective morphemes; at least, there is no phonological difference in models of phonology that are relatively surface-oriented. Thus e.g. the sequence /mm/ in lammas is not, at least phonetically, in any way different from the same sequence in tamma ‘mare’, a word that does not participate in grade alternation. Secondly, there are descriptions of Finnish morphophonological alternations available in English, albeit not exhaustive ones, in grammars such as Karlsson (1999) and Sulkala & Karjalainen (1992); readers interested in aspects of Finnish not dealt with in this book are advised to consult these sources. Besides offering a glimpse at morphophonological alternations, Chapter 1 also briefly exemplifies inflection and word formation by derivation; these often involve morphophonological alternations. Hopefully, Chapter 1 gives an adequate overall picture of the structure of Finnish words. We wish to thank Matthew Gordon (University of California, Santa Barbara) for useful comments on the manuscript of this book. Any errors and inconsistencies that remain are our own. Oulu, December 2008 Kari Suomi Juhani Toivanen Riikka Ylitalo 4 Contents Preface 3 Contents 5 1 . Standard Spoken Finnish: definition and some structural properties 7 2 Airstream mechanisms and phonation 17 3 Phonemes and allophones 19 3.1 Vowels ............................................................................................................ 20 3.2 Consonants ..................................................................................................... 23 4 On the phonological interpretation of the quantity opposition 39 5 Sandhi phenomena 43 5.1 Nasal assimilation .......................................................................................... 43 5.2 Boundary lengthening ................................................................................... 44 5.3 The glottal stop .............................................................................................. 46 6 Phonotactics 49 6.1 Vowel phonotactics ........................................................................................ 49 6.1.1 Vowel sequences................................................................................. 49 6.1.2 Vowel harmony ................................................................................... 51 6.2 Consonant phonotactics ................................................................................ 55 6.2.1 Word-initial consonant sequences ..................................................... 55 6.2.2 Word-internal consonant sequences .................................................. 56 6.2.3 Word-final consonant sequences ....................................................... 59 6.3 Restrictions on #CV, #VV and #(C)VVCC sequences ............................... 61 7 Syllable and mora structure 65 8 Canonical word structure, minimal word, minimal utterance 69 9 Word-level Prosody 75 9.1 Degrees of word stress and their phonetic realisation ................................. 75 9.2 The phonetic realisation of accent ................................................................ 78 9.3 Segment durations and moraic structure ...................................................... 85 9.4 Segment durations and a speech timing model ......................................... 101 10 Utterance-level prosody 111 10.1 Orientation ................................................................................................. 111 10.2 Accent and information structure ............................................................. 112 10.3 Intonation ................................................................................................... 114 10.4 Rising intonation ....................................................................................... 117 5 10.5 Descriptive frameworks for Finnish intonation ...................................... 120 10.6 Intonation range ........................................................................................ 125 10.7 Intonation of Finnish as a second language ............................................ 126 10.8 Emotional Finnish speech: research questions, data bases and tools ............................................................................................................ 130 10.9 Emotion in spoken Finnish: evidence from classification experiments ............................................................................................... 133 10.10 A summary of some things glottal ......................................................... 136 11 Sound structure and orthography 141 References 145 6 1. Standard Spoken Finnish: definition and some structural properties The book aims at describing the sound structure of Standard Spoken Finnish (SSF). This is a form of speech that is used in the educational system and in the media across the country. Originally, it was based on Standard Written Finnish, which in turn was consciously created, in the nineteenth century, as a compromise between the various dialects. In contrast to many other standard languages, then, Standard Finnish (written or spoken) is not based on the language spoken in the centre of power. Consequently, at first, SSF was nobody’s native dialect, and it was spoken, in formal situations, by a small number of educated people only; most of the educated and other upper class people at those times spoke Swedish as their native language. Later, SSF has been actively and successfully propagated through the educational system. Today, SSF may be the native dialect for a number of speakers, but most people learn a local dialect first, and then SSF. Most speakers today have command over two varieties of spoken Finnish: their local dialect and SSF. Usually, the former is used in informal speaking situations, the latter in formal ones; however, some speakers, especially elderly ones, do not necessarily speak SSF on any occasion — and even many younger people never have the chance or duty to speak in formal situations. Although SSF is spoken across the country, it is not a monolith: it has local colourings, especially as concerns prosody, notably segment durations in certain word positions and the way sentence accents are realised; this is clear from so far unpublished results of the third author. It is not expected in the Finnish society that an educated speaker should, in a formal speaking situation, speak SSF according to the strictest norms (recommended if not demanded by an advisory board funded by the state); instead, local colourings are both used and tolerated. That is, local varieties of SSF do not stigmatise the speaker as they do in many other countries, language areas and cultures; recall that we are talking about local varieties of SSF, and not about local vernaculars. Moreover, there is increasing variability in SSF due to register, so that increasingly informal forms of SSF are emerging, and their use in formal speaking situations is increasing; these informal forms include features such as deletions of certain segments in certain positions, or replacements of phonemes of foreign origin by fully native ones, as will be explained in more detail below. In fact, speech according to the strictest norms is nowadays used in highly formal 7 situations only, and not always so. Speeches by statesmen, and interviews between reporters and ordinary folk (including children), recorded a few decades ago, often sound ridiculous because of their excessive formality, which is very much signalled by the prosody. Speakers also differ among themselves with respect to the conditions of formality under which they switch from the vernacular to SSF and vice versa. For the colloquial spoken language, see also Chapter 22 in Karlsson (1999). While we aim at describing only the sound structure of SSF systematically, we make occasional references to differences between SSF and local dialects; we would gladly make more systematic references to this effect if there were more reliable data available on such differences. When we write that such and such circumstances obtain in Finnish, this means that, as far as we know, the circumstances obtain for Finnish in general. Morphophonological alternations are very common in Finnish, to the great delight of adult learners of Finnish as a second language. Here only a few examples are given. For now the reader should just take it that all forms written differently are pronounced differently, and that double letters stand for phonetically long segments that contrast with single letters that stand for phonetically short segments. The word talo ‘house’ exhibits no morphophonological alternation, while susi ‘wolf’ does. The forms just given are uninflected, Singular Nominative forms in which both Case and Number always have zero expression. That is, both word forms consist of a stem only. Below these and some inflected Case forms of these words are given, in both Singular (Sg) and Plural (Pl): Nominative Genitive Essive Partitive Illative Sg talo talon talona taloa taloon Pl talot talojen taloina taloja taloihin Sg susi suden sutena sutta suteen Pl sudet susien susina susia susiin No attempt has been made here at a morphological analysis of the word forms. In many forms the morphological structure is fully transparent (e.g. talo+n, talo+na, talo+t), in others it is more or less opaque and portmanteau morphs exist (e.g. in 8