the phonology of icelandic and faroese

The Phonology of Icelandic and Faroese
THE PHONOLOGY OF THE WORLD’S LANGUAGES
General Editor: Jacques Durand
Published
The
Phonology of Icelandic and Faroese
Kristja
´
nA
´
rnason
The Phonology of Danish
Hans Basbøll
The Phonology of Dutch
Geert Booij
The Phonology of Standard Chinese, second edition
San Duanmu
The Phonology of Polish
Edmund Gussmann
The Phonology of English
Michael Hammond
The Phonology of Italian
Martin Kra
¨
mer
The Phonology of Norwegian
Gjert Kristoffersen
The Phonology of Portuguese
Maria Helena Mateus and Ernesto d’Andrade
The Phonology and Morphology of Kimatuumbi

David Odden
The Lexical Phonology of Slovak
Jerzy Rubach
The Phonology of Hungarian
Pe
´
ter Sipta
´
r and Miklo
´
sTo
¨
rkenczy
The Phonology of Mongolian
Jan-Olof Svantesson, Anna Tsendina, Anastasia Karlsson, and Vivan Franze
´
n
The Phonology of Armenian
Bert Vaux
The Phonology and Morphology of Arabic
Janet Watson
The Phonology of Catalan
Max Wheeler
The Phonology of German
Richard Wiese
In preparation
The Phonology of Tamil
Prathima Christdas
The Phonology of Welsh
S. J. Hannahs

The Phonology of Turkish
Baris¸ Kabak
The Phonology of Japanese
Laurence Labrune
The Phonology of Spanish
Iggy Roca
The Phonology of Greek
Anthi Revithiadou
The Phonology of Swedish
Tomas Riad
The Phonology of Washo
Alan C. L. Yu
THE
PHONOLOGY
OF
ICELANDIC AND
FAROESE
KristjanA

rnason
1
3
Great Clarendon Street, Oxford ox2 6dp
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto

With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
# Kristja
´
nA
´
rnason 2011
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2011
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by SPI Publisher Services, Pondicherry, India

Printed in Great Britain
on acid-free paper by
MPG Books Group, Bodmin and King’s Lynn
ISBN 978–0–19–9229 31–4
13579108642
To the memory of Edmund Gussmann
This page intentionally left blank
CONTENTS
Acknowledgements xiii
Abbreviations xv
Part I The historical and theoretical setting
1 THE TWO LANGUAGES AND THEIR HISTORICAL RELATION 3
1.1 The genetic relation: ‘Proto-West Nordic’ 3
1.2 West Nordic obstruents 4
1.3 West Nordic sonorants 6
1.4 West Nordic vowels 7
1.5 Diphthongs and semivowels 8
1.6 Prosodic structure 8
2 THE HISTORICAL DEVELOPMENT 11
2.1 Quantity and prosodic structure 12
2.2 Overlong (superheavy) syllables and their development 15
2.3 The components of the quantity shift 17
2.4 Quality changes in the Icelandic vowel system 20
2.5 Faroese vowel developments 23
2.6 The short diphthongs 25
2.7 The West Nordic consonant shift 26
2.8 New postvocalic stops 28
2.9 The skerping and hiatus 31
2.10 Systemic arrangements and types of syllables 33
3 THEORETICAL PRELIMINARIES TO THE SYNCHRONIC ANALYSIS 35

3.1 Phonological levels of representation 35
3.2 Alphabets for phonological representation 39
3.2.1 Segments and distinctions 40
3.2.2 Phonological primes 41
3.2.3 Analysing diphthongization 43
3.3 The representation of time and precedence 47
3.4 Saturation and fission in West Nordic diphthongs 48
3.5 The modern diphthongal systems 51
Part II The modern sound systems
4 THE ICELANDIC VOWEL COLOURS AND DIPHTHONGS 57
4.1 The Icelandic vowel system 57
4.1.1 The monophthongal units 57
4.1.2 The phonetic description of the vowel sounds 60
4.1.3 Analysing the monophthongal colours 61
4.1.4 The diphthongs as branching phonemic units 62
4.1.5 Diphthongs and hiatus 65
4.2 The vowels of non-initial syllables 66
5 FAROESE VOWELS AND DIPHTHONGS 68
5.1 An overview 68
5.1.1 The inventory of vowels 68
5.1.2 The ‘long’ and ‘short’ systems 70
5.2 The phonetic and phonological analysis of the Faroese monophthongs
and diphthongs 74
5.2.1 The vowel qualities 74
5.2.2 Analysing the Faroese diphthongs 76
5.3 More on dialect variation and vowel systems 79
5.4 Hiatus phenomena in Faroese 80
5.4.1 Intervocalic glides 80
5.4.2 Raising in hiatus 82
5.4.3 One or two syllables 85

5.5 The unstressed vowels of Faroese 86
5.5.1 The vowel qualities 86
5.5.2 The syllable types 88
5.5.3 Dialects and morphology in unstressed vowel merger 90
5.5.4 The postlexical reduction and syncope 94
5.5.5 Coping with the variation: a folk-linguistic anecdote
from To
´
rshavn 96
6 ICELANDIC CONSONANTS 98
6.1 An overview 98
6.2 The stops 99
6.2.1 Places of articulation 99
6.2.2 On palatals and velars 100
6.2.3 The fortis–lenis opposition 103
6.3 The fricatives 106
6.4 The sonorants 109
6.5 Summary: the classes of consonants and their element analysis 111
viii contents
7 FAROESE CONSONANT SEGMENTS 114
7.1 An overview 114
7.2 The fortis and lenis plosives 118
7.2.1 The phonological correlation 118
7.2.2 ‘Hard’ and ‘soft’ dialects in Faroese 118
7.3 The fricatives 122
7.3.1 Initial fricatives 122
7.3.2 Word-internal fricatives and glides 123
7.4 Sonorants 124
7.5 The element analysis of the Faroese system 124
Part III Systemic relations and syllabic structure

8 SYSTEMIC RELATIONS IN VOWELS 129
8.1 Trends towards a diasystem in Icelandic 129
8.1.1 The ‘new’ diphthongs 129
8.1.2 A context-free merger in the ‘long’ system 132
8.1.3 The interplay of merger and diphthongization 133
8.1.4 The ‘long/open’ and ‘short/closed’ correspondence 134
8.2 The Faroese vowel systems 135
8.2.1 The polysystemic structure 135
8.2.2 On prominence 138
8.3 The element analysis of reduction: limits on information in restricted
environments 140
8.4 Conclusion: systemic relations in vowel systems 142
9 SYLLABLE STRUCTURE AND PHONOTACTICS 144
9.1 Syllable structure in Icelandic 144
9.1.1 Motivating the syllable 144
9.1.2 The subsyllabic constituents 145
9.1.3 The Icelandic length rule 148
9.1.4 Half length and overlength 149
9.1.5 Emphatic stress on non-initial syllables 151
9.2 Faroese syllables 152
9.2.1 The length rule in Faroese 152
9.2.2 The template for full syllables in Faroese 155
9.2.3 Half length and overlength in Faroese 156
9.2.4 Full syllables and restricted syllables in the phonological
hierarchy 157
contents ix
9.3 The consonantal phonotactics of Icelandic 160
9.3.1 Initial onsets in Icelandic 161
9.3.2 Consonants in the Icelandic coda 165
9.3.3 Internal onsets in Icelandic 167

9.3.4 More complicated interludes in Icelandic 170
9.4 The consonantal phonotactics of Faroese 173
9.4.1 Monosegmental onsets 173
9.4.2 Initial clusters in Faroese 175
9.4.3 Faroese coda consonants 176
9.4.4 Internal onsets in Faroese 178
9.5 Gemination of glides and consonants 180
9.6 Conclusion: remarks on systemic structure and prominence 182
10 LENGTH AND QUANTITY IN ACCENTUATION AND
PHONOTACTICS 185
10.1 Length and quantity in Icelandic 186
10.1.1 Some distinctions 186
10.1.2 Segmental length and syllabic structure 189
10.1.3 Length in postlexical accentuation 191
10.1.4 Representing length and quantity 195
Excursus: A brief comparison with Finland Swedish 202
10.2 The length rule on lexical and phonological levels in Icelandic 203
10.2.1 Length and syllabification 203
10.2.2 Two versions of the rule 205
10.3 The prosodic character of Faroese vowels 208
10.4 Vowel shortness and the scale of prominence 210
Part IV Segments and syllables on phonological levels
11 ASPIRATION IN SYLLABIC AND SEGMENTAL STRUCTURE 215
11.1 Aspiration and the character of the fortis–lenis opposition 216
11.1.1 Representing the opposition 216
11.1.2 Phonotactic or phonetic neutralization of the
fortis–lenis opposition? 217
11.2 Preaspiration in Icelandic 219
11.2.1 The phonetic character 219
11.2.2 The distribution of preaspiration in Icelandic 220

11.2.3 Phonological analysis 222
11.2.4 Related phenomena 225
11.3 Preaspiration in Faroese 228
11.4 Representing the difference 230
11.5 Preaspiration in morphophonemics 231
x contents
12 ALLOMORPHY, MORPHOPHONEMICS, AND PHONOLOGICAL
LEVELS 234
12.1 Introduction 234
12.2 Lexical and postlexical relations in paradigms 235
12.3 Vocalic patterns in Icelandic 238
12.3.1 The ablaut series 238
12.3.2 I-umlaut 239
12.3.3 U-umlaut and breaking 243
12.3.4 Morphology and phonotactics in vocalic alternation 246
12.4 Faroese vowel morphophonemics 247
12.4.1 Ablaut 247
12.4.2 Umlaut 248
12.5 Consonantal patterns in Icelandic morphophonemics 250
12.6 Consonantal patterns in Faroese 252
12.7 Vowel deletion in paradigms 253
12.8 Intersyllabic glides and fricatives in allomorphy 257
12.9 Morphosyntax and phonology 259
12.9.1 Lexical and postlexical principles 259
12.9.2 Clitics 263
12.9.3 Fossilized and non-fossilized patterns in Faroese 264
12.10 Conclusion 265
Part V Rhythmic structure
13 WORD STRESS PATTERNS IN ICELANDIC AND FAROESE 271
13.1 Icelandic word stress patterns 271

13.1.1 Native patterns 271
13.1.2 Foreign patterns in recent loans 274
13.2 Word stress patterns in Faroese 275
13.2.1 Native words 275
13.2.2 Loanwords in Faroese 278
13.3 The accommodation of foreign stress patterns 279
13.3.1 Right-strong forms in Icelandic 279
13.3.2 The Faroese patterns 280
13.4 Morphological considerations: Faroese pseudo-morphology? 282
14 PHRASING AND POSTLEXICAL PHONOLOGY 285
14.1 Phonological phrasing 285
14.2 Systematic exceptions 286
14.2.1 Deaccenting of weak morphosyntactic categories 286
14.2.2 Deaccenting in Icelandic definite noun phrases 289
contents xi
14.3 Pragmatically motivated exceptions 290
14.3.1 Emphatic rephrasing 290
14.3.2 Contrast, focus, and given information 291
14.4 The phonological consequences of phrasing 292
14.4.1 Cohesive laws or sandhi-rules in Icelandic 292
14.4.2 Sandhi in Faroese 297
14.4.3 Rhythmic rearrangement 300
14.5 Demarcative signals 300
14.5.1 Stress and glottal onset 301
14.5.2 Right hand signals in Icelandic 302
14.5.3 Demarcative signals in Faroese 302
14.6 Constituency and prominence 303
14.6.1 Boundaries or dependencies? 303
14.6.2 Domains, directionality, and prominence 305
15 RHYTHM AND INTONATION 308

15.1 Rhythm and constituency 308
15.2 Icelandic intonation 313
15.2.1 The tonal inventory 314
15.2.2 Downstep and upstep 318
15.2.3 Functional considerations 320
15.3 Faroese intonation 324
15.4 The absence of word tones 326
A note on phonetic data 328
References 329
Index 343
xii contents
ACKNOWLEDGEMENTS
Writing this book has given me a unique opportunity to look, in the context of a
monograph, into several problems and phenomena that have been close to my heart
for a number of years. I am grateful to Oxford University Press and the editors of the
series The Phonology of the World’s Languages for this opportunity and for their
excellent cooperation in the publication process. The experience has turned out to be
rewarding for myself in spite of the fact, pointed out by several colleagues, that
there is no such thing as the phonology of two languages, even closely related
ones like Icelandic and Faroese, so that this book should perhaps never have been
written. But the necessarily comparative approach, rather than a purely descriptive
or analytical one, appealed to me from the very beginning. It is for the reader to judge
whether this has turned out to be a fruitful enterprise, and I’ll allow the text to
speak for itself.
During the writing of the book I have been fortunate in receiving generous help
from a number of people, in particular specialists on Faroese. During my stay in
To
´
rshavn in the autumn of 2009, I had several sessions with people from the team o f
linguists working in the Fro

´
ðskaparsetur Føroya: Jo
´
gvan ı
´
Lon Jacobsen, Anfinnur
Johansen, Zakaris Svabo Hansen, and last but not least Eivind Weyhe. Their generous
advice h as undoubtedly helped to eliminate an endless number of errors in my repo rt,
although the remaining one s are of course my responsibility. I am also gra teful to the
Nordic Language Variation Network for inviting me to join in on the marvellous
fieldwork excursion (organized by Øyste in Vangsnes, Ho
¨
skuldur Þra
´
insson and
others) in August 2008. This gave me an opportunity to record the speech of
informants from various locations in the Faeroes. These data turned out to be an
invaluable tool in the phonological description of Faroese on which the present book
is partly based. I am also grateful to the Arnamagnæan Section of the Depar tment of
Scandinavian Research at the University of Copenhagen for hospitality in Septem-
ber–October 20 08, and to So
´
ley Hammer a nd all the other Faroese informants, for
their patience and cooperative spirit.
I thank the audiences at conferences like the Manchester Phonology Meeting in
May 2008 and at the phonological seminar in Flaam, Norway in June 2008 for various
comments and discussion. Many colleagues have made helpful suggestions regarding
parts of the draft or oral presentations, among them Nicole Dehe
´
, Edmund Gussmann,

Gert Kristofferesen, Curt Rice, Helge Sandøy, and Michael Scha
¨
fer. Special thanks
go to Hjalmar Petersen for his comments, corrections and ready answers in endless
e-mail messages, and for checking the Faroese data in the final draft, although again,
he should not be held responsible for their (lack of) accuracy.
Special thanks go to John Davey, Kim Allen and Elmandi du Toit at Oxford
University Press for their cooperation and friendly assistance and Geoff Sampson
for excellent help with the proofs. Lastly, I must thank my wife Arna Emilı
´
a and
daughter Nanna for their good work in keeping up the family morale during the time
of writing.
xiv acknowledgements
ABBREVIATIONS
A appendix
ACC accusative
ADJ adjective
ADV adverb
AP adverbial phrase
C consonant
COMP comparative
CWN Common West Nordic
DAT dative
DEF definite
DTE designated terminal element
FEM feminine
Fn. footnote
FTA Focus to Accent
FVD Final Vowel Deletion

GEN genitive
IMP imperative
INF infinitive
INTR intransitive
IO Input–Output
IP Intonational Phrase
ip Intermediate Phrase
IPA International Phonetic Association
Lv. occasional stanza (lausavısa)
MASC masculine
ME Middle English
MED Mediopassive
MF Modern Faroese
MG Modern German
MI Modern Icelandic
N, NEUT neuter
NOM nominative
NP noun phrase
O Onset
OBL Oblique (case)
OCP Obligatory Contour Principle
OE Old English
OHG Old High German
OI Old Icelandic
ON Old Norse
OT Optimality Theory
OWN Old West Nordic
P person
PART participle
PAST past tense

PL plural
POS positive
rd. rounded
R rhyme
ROTB ‘Richness of the Base’
ó syllable
SG singular
SUBJ subjunctive
SUPERL superlative
unrd. unrounded
V vowel
VC vowel – consonant
VOT voice onset time
xvi abbreviations
Part I
The historical and theoretical setting
This page intentionally left blank
1
THE TWO LANGUAGES AND THEIR
HISTORICAL RELATION
The two languages treated in this book have, along with some west Norwegian
dialects, been categorized historically as forming the western branch of the Nordic
stem of the Germanic Stammbaum (see Haugen 1982 for a comparative overview).
They are presently spoken and function as official languages and languages of culture
on islands in the North Atlantic, belonging to two societies, the Republic of Iceland
with about 320,000 inhabitants, and the Faroes with about 48,000 inhabitants, which
form an autonomous region within the Danish Union (Rigsfællesskab). Iceland has a
history of monolingualism and a very old written norm, whereas until the twentieth
century Danish was the official language in the Faroes, and written Faroese stems only
from the eighteenth century (see e.g. Barnes 1978, Svabo Hansen 2003, and

Thr

ainsson et al. 2004: 369–465 on the history of Faroese). The Danish influence
on Faroese is obvious, particularly in a great number of loanwords, and it is some-
times said that syntactically speaking, Faroese lies somewhere between ‘Mainland
Scandinavian’, that is the Scandinavian languages proper, and ‘Insular Scandinavian’,
that is prototypically Icelandic (see Rischel 1992 for a typological evaluation, cf. also
Barnes and Weyhe 1994, Thr

ainsson 2000, Sandøy 2001; see Petersen 2010a for
discussion of the sociolinguistic relation between Danish and Faroese). The Faroese
orthography used today is very archaic and heavily influenced by the Old and Modern
Icelandic standard.
The third branch of the West Nordic tree was cut off or ‘pruned’, so to speak,
when the Norwegian chancery adopted Danish (cf. A
´
rnason 2003). However, several
of the phonological features of Faroese and Icelandic are found in southern and
western Norwegian dialects (cf. e.g. Hægstad 1907–42; Chapman 1962; Christiansen
1946–48; Sandøy 2003; Ottosson 2003).
1.1 THE GENETIC RELATION: ‘PROTO-WEST NORDIC’
It is well known that in spite of its usefulness as a model of genetic relations between
languages, the Stammbaum-metaphor and the methods of reconstruction, assuming
regular sound changes, that go with it tend to lead to abstractions which often have
less to do with reality than intended. And more often than not paradoxes appear,
entangling the branches and making the picture messier than it was set out to be. The
genetic relation between Icelandic and Faroese is in fact obvious, and it is easy to
correlate many phonological, morphological, and even syntactic characteristics of the
languages by assuming a common ancestor, some sort of ‘Common West Nordic’
(CWN), from which the present-day idioms have developed. (We will leave the

complicated issue of the relationship with Norwegian out for the moment.) Further-
more, this ancestor is (or should be, according to the simplest model of this relation),
amply documented in the classical Icelandic–Norwegian literature.
In fact, establishing a historical relation between the modern languages and classi-
cal Old Icelandic (OI)/Old Norse (ON) is often quite straightforward, for example by
saying that the morphological or phonological structure of the present-day languages
have ‘lost’ or ‘added’, perhaps borrowed from elsewhere, features that were or were
not present in the original language. This is evident when we look at morphosyntactic
categories such as, for example, the case system, where in Modern Faroese (MF) the
genitive has been more or less lost as a thematic case (see e.g. Thr

ainsson et al. 2004:
248ff.). In general, the Faroese system has undergone important changes, compared to
the Old Norse standard, and also compared to Modern Icelandic (MI), which is more
conservative. Thus it can be said that the inflectional system of MI is basically the
same as that of classical Old Icelandic. But in spite of the conservatism, interesting
changes have taken place in Icelandic morphosyntax. Thus the use of dummies like
það ‘it’ and the verb vera ‘to be’ has increased over the course of time (cf. e.g.
Sigurj

onsd

ottir and Maling 2001; Benediktsson 2002; Ro
¨
gnvaldsson 2005), and the
use of prepositions in possessive constructions is on the increase, as in hendin

am

er

‘my hand’ and to¨lvan hj

am

er ‘my computer’ beside ho¨nd mı´n and to¨lvan mı´n.
When it comes to phonological structure, which obviously is our main concern
here, the testimony of standard Old Icelandic is in some ways quite clear, and in fact
the phonological history is more interesting, since more has happened over time in the
phonology than in the morphosyntax. We have ample documentation of the phonemic
inventory of Old Icelandic, based on the evidence of texts (both prose and poetry)
from the twelfth century onwards, and no less importantly in works such as the First
Grammatical Treatise, where a detailed phonological analysis is presented. This good
documentation makes it possible to compare the older and later stages in quite a bit of
detail, and, for Icelandic in particular, the continuous production of written texts
allows the historical development to be plotted quite closely. (For an investigation of
the phonological history of Icelandic, see e.g. Benediktsson 1959, 1970a). A detailed
historical documentation of the phonological development of Faroese is not so easy,
since written sources are relatively rare, but comparative evidence from Old and
Modern Icelandic helps to create a reasonably clear picture. (See Hansson 1983;
Thr

ainsson et al. 2004: 369–465 on the phonological history.)
1.2 WEST NORDIC OBSTRUENTS
But although much is known about Common West Nordic, many things are unclear
about ‘phonetic detail’, and even some fundamental properties of the system. One
such point of ignorance regards the phonetic character of the opposition between
the fortis and lenis, hard and soft, plosives, p,t,k vs b,d,g, in ON forms like pungr
4 1 the historical and theoretical setting
‘bag’, taka ‘to take’, korn ‘corn’, vs borg ‘town’, dagr ‘day’, garðr ‘garden, fence’.
According to most traditional handbooks the opposition in Old Icelandic was based on

voicing (cf. Benediktsson 1972: 165–74; Iversen 1973: 11–12), but as shown by
scholars like Steblin-Kamenskij (1960, 1974) and Hansson (2001), relating the
modern West Nordic system—where aspiration distinguishes between the respective
series—to this alleged proto-structure and the more eastern or central European
system—where voicing is more prevalent—is by no means a simple matter. In fact,
although commonly assumed, it is not clear that the voiced–voiceless opposition was
true of the whole North Germanic area; it is conceivable that in at least some varieties
of Proto-Nordic, the opposition was based on aspiration. In any case, if voicing was
the distinctive feature in Old West Nordic (OWN), the fact that both Icelandic and
Faroese use aspiration to distinguish the systems tells us that a consonant shift has
taken place in both languages, although dating this change is not easy.
This detail regarding the interpretation of the historical relations, and more impor-
tantly, the phonological ‘foundations’ of the languages treated here is all the more
important because of the likelihood, often noted, that the consonant shift was some
sort of chain shift, that is, that we are dealing not with individual sounds, but with a
system of sounds. In fact the development is a chain shift in two respects. On the one
hand, the fortis–lenis opposition was maintained in initial position (in pairs like punga
‘bags-ACC.PL’ and bunga ‘hill’), which means that the change did not affect
individual segments, but the (sub)system as a whole, and, on the other hand, in medial
position the development of preaspiration in vakna [vahkna] ‘to wake up’ as a part of
the shift, was most likely somehow related to the development of a stop in vagna
[vakna] ‘wagons’- GEN. So word internally, what was once an opposition between a
stop and a voiced obstruent (originally a fricative) became an opposition between two
voiceless stops, distinguished by aspiration. Comparative reasoning points to earlier
[*vakna] vs [*vaɣna], and this consonant shift is something which both languages
share, and so by the logic of historical comparative linguistics should be assigned to
the mother language, that is Proto-West Nordic. This would then mean that the
traditional interpretation, that the fortis–lenis opposition was based on voicing is
not justified.
Whatever the phonetic character of the fortis–lenis opposition and its geographic

distribution, the structure of the reconstructed obstruent system can be pictured as
in (1.1):
(1.1) Labial Dental Velar Laryngeal
Stops Fortis p t k
Lenis b d g
Fricatives Voiceless f þ, s h
⇕⇕
Voiced β (v) ð ɣ
As shown in (1.2) there was a three-way opposition in initial position with regard to
the manner of articulation between fortis (voiceless) plosives, lenis (voiced) plosives,
and voiceless fricatives and a sibilant s. (No voiced fricatives occurred word initially.)
1 the two languages and their historical relation 5
(1.2) Proto-Nordic OWN
Fortis plosives
p *pungaR pungr ‘bag’
t *tungo¯(n) tunga ‘tongue’
k *kurna korn ‘corn’
Lenis plosives
b *burgo¯R borgar ‘town-GEN’
d *dagaR dagr ‘day’
g gastiR gestr ‘guest’
Voiceless fricatives
f faihido¯ f

aða ‘painted’
þ þar þar ‘there’
h horna (<*hurna) horn ‘horn’
s satido¯ setta ‘put-PAST’
In word internal position after a vowel, the opposition between fortis and lenis stops
was neutralized, since only fortis (i.e. voiceless) ones occurred, as in drepa ‘to kill’,

gata ‘road, path’, and taka ‘to take’.
No voiced fricatives occurred in initial position, but voiced and voiceless fricatives
were in complementary distribution in the labial and dental series, as partly reflected
by the spelling: OI fara [fara] ‘to go’ ho˛fuð [hɔvuð] ‘head’, þar [
Ł
ar] ‘there’ taða
[taða] ‘hay’ (also spelled taþa). In the velar region, there is a gap, left by the
disappearance of Proto Nordic voiceless [x] in word internal position in *fahan > f

a
[fɑ
ː
] ‘to get’. A part of this development was the change of the initial /h/ in horn(a)
‘horn’ from a velar fricative [x] to a laryngeal [h]. (A remnant of the fricative
articulation is to be found in an Icelandic dialect pronunciation of historical /hv/
with a fricative as in hvalur ‘whale’ as [x
w
a
ː
lYr
˚
]; the majority dialect in Iceland, and
all Faroese dialects, have developed stops in these forms, giving MI [k
h
va
ː
lYr
˚
]/
[kfa

ː
l
Yr
˚
], MF [k
h
v
ɛ
a
ː
lʊɹ]/[kf
ɛ
a
ː
lʊɹ].)
1.3 WEST NORDIC SONORANTS
Classical Old Norse had a labial nasal /m/, as in maðr ‘man’ koma ‘to come’, and a
dental /n/, as in niðr ‘descendant’ and skı´na ‘to shine’. It is likely that velar (and perhaps
palatal) variants occurred before velars and palatals, as in langr ‘long ADJ-M’ and lengi
‘long ADV’. The standard written representation of the nasal was the grapheme < n >,
but the Fist Grammarian notes the assimilation of /n/ to a following /g/ and suggests a
special character, which he calls eng to represent the cluster /ng/ in hringr ‘ring’
(see Benediktsson 1972: 236). We cannot exclude the possibility that the First
Grammarian’s suggestion of using one letter was motivated by a stopless pronuncia-
tion, that is something like [hriŋr]. But this is unlikely, since both modern languages
have stops in forms like MI fingri [fiŋkrɪ] ‘finger-DAT.SG’ and MF fingrar [fiŋkɹəɹ]
‘finger-NOM.PL’. In any case we can assume that a principle of place assimilation for
6 1 the historical and theoretical setting
nasals to following a stop was valid for OWN. And this is also confirmed by the fact
that the phonotactics of the modern languages provides for forms like lamb ‘lamb’,

land ‘land’ with place agreement, but no underived forms like *lanb or *lamd.
As regards liquids, the lateral /l/, as in land ‘land’ and tala ‘to speak’ was most
likely dental, although there may well have been some variation, giving retroflex,
palatal, or velar variants. It is also possible that there was some variation in the
articulation of the trill /r/. Here retroflexion may well have been an option and it is
conceivable that uvular pronunciation may have been dialectal or even widespread at
some stage. An overview of the West Nordic sonorants is given in (1.3):
(1.3) Labial Dental Velar
Nasal m n ŋ
Lateral l
Trill r
1.4 WEST NORDIC VOWELS
Based on the testimony of Old Icelandic writing and phonological analysis, the West
Nordic vowel system can be reconstructed as a diasystem, containing (in principle)
isomorphic systems of long and short stressed vowels and a smaller set of unstressed
vowels (cf. Benediktsson 1959). The effect of the proto-Nordic processes of umlaut
and breaking was to multiply the set of vowel qualities so that for the earliest stage of
classical Old Icelandic the system was relatively rich, containing nine vowel qualities
as shown in (1.4):
(1.4)
front unrd. front rd. back unrd. back rd.
high i [i] y u [u]
mid e [e] ø o [o]
low e¸[
ɛ
]ao˛[ɔ]
In the ideal ‘proto-lect’, all of these qualities occurred as long or short, based on a
phonemic length correlation, although the actual testimony for this state of affairs is
indirect. The clearest evidence comes from the twelfth-century First Grammatical
Treatise (see Haugen 1950; Benediktsson 1972), which presents an analysis based on

nine vowel qualities, and a correlation of length (in addition to nasalization). How-
ever, the picture presented in the literature bears certain marks of being a reconstruc-
tion, since we have no direct evidence of an idiom where the system is preserved in its
original form, that is with total isomorphism between the long and the short system,
and there are some ambiguities as to the actual quality of the vowels.
For unstressed syllables, where there is no length contrast, a system of three vowel
qualities has been assumed, as in (1.5):
1 the two languages and their historical relation 7
(1.5) i u
a
These are the typical vowels for non-initial syllables, grammatical morphemes, and
inflectional endings, as in gestir ‘guests-NOM.PL’, hestar ‘horses-NOM.PL’, hestum
‘horse-DAT.PL’. The fact that both the length and the quality of these unstressed
ending vowels is limited suggests that there was a systematic difference in Old
Icelandic between what may be called full, stressable syllables and reduced or
restricted ones, which could not take the word stress.
1.5 DIPHTHONGS AND SEMIVOWELS
In addition to these monophthongal qualities, the following closing and falling
diphthongs occurred: /au/ as in bauð ‘ordered, invited’, /ei/ as in leit ‘looked’, and
/øy/ (written < ey >) as in reykr ‘smoke’. This last diphthong developed from /au/ by
i-umlaut (cf. rauk, the past tense of rj

uka ‘to reek, to smoke’). One way to analyse
these diphthongs is to see them as combinations of the vowel qualities [a], [e], and [ø]
with [i] and [u] as semi-vowels. And a similar interpretation seems to be possible of
the falling diphthongs in kjo˛lr ‘keel’ and kjalar ‘keel-GEN’. These short diphthongs
came about by the ‘breaking’ of older /e/, in certain environments. This breaking
involved the epenthesis of a low vocalic quality after the /e/, which then lost its
syllabicity and became a glide [i ̯]. This type of analysis gets support from the fact that
the initial sound of jo˛rð ‘earth’ alliterated with vowels in eddic and skaldic poetry.

And the same goes for the initial sound of j

arn ‘iron’ in spite of the fact that it has a
different origin (cf. Harðarson 2011). According to the First Grammarian, /v/ was
a consonantal variant of /u/, and there are some instances of such a /v/ apparently
alliterating with a vowel in early poetry.
1.6 PROSODIC STRUCTURE
Obviously we do not know much about intonation and tonal contours in Old Norse
(nor indeed the details of the contours, which created the basis for the present day
Eastern Nordic word tones and stød (see Section 15.4 below). But some information
on phrase phonology can be drawn from the testimony of the older poetry (e.g. the
workings of the so-called Kuhn’s Laws and Craigie’s Law in eddic and skaldic poetry,
see Kristj

an A
´
rnason 2002, 2009a).
When it comes to word phonology, we are in a better position, and in many respects
we are relatively well informed about the segmental and suprasegmental structure of
words in Old West Nordic. We know, for example, that the distinction between long
and short vowels had an effect on syllabic quantity. The testimony, both of the
8 1 the historical and theoretical setting