Tonal word accents

University Press Scholarship Online

Oxford Scholarship Online

The Phonology of Swedish

Tomas Riad

Print publication date: 2013

Print ISBN-13: 9780199543571
Published to Oxford Scholarship Online: January 2014
DOI: 10.1093/acprof:oso/9780199543571.001.0001

Tonal word accents

Tomas Riad

DOI:10.1093/acprof:oso/9780199543571.003.0009

Abstract and Keywords

This chapter lays out the basic information regarding the tonal
accent distinction in Swedish. This lexical distinction is due to
the presence of a lexical tone in some morphemes, especially
in some very frequent syllabic suffixes. In intonation, the tonal
cases come out as what is called accent 2. Accent 1 is fully
made up of intonation tones. Beside the lexical accent 2 tones,
there is a postlexical source for accent 2, namely in the
context of two stresses (e.g. as in compounds). A few
characteristic tonal contours are given in the chapter, to
illustrate the way accents 1 and 2 come out in two intonational
conditions: word accented and focus accented, where the
latter is a higher prominence degree.

Keywords: accent, intonation, accent 1, accent 2, word accented, focus

accented

Several of the Germanic languages of Scandinavia exhibit

tonal word accent systems beside their stress systems. The

Page 1 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

accent systems are tonal in most varieties of Swedish and

Norwegian. In Danish varieties there is a corresponding
system of stød vs. no-stød (Basbøll 2005).1 The phonetic,
phonological, and distributional properties are quite different
between tonal varieties and stød varieties, but there is no
doubt that they are historically related, where all indications
point at stød being secondary to a tonal configuration
(Ringgaard 1983; Riad 2000a, b; see Fischer-Jørgensen 1989:
17ff. for a discussion of different theories).

By ‘tonal accent system’ we mean that there is a phonological

distinction between two tonal configurations, of which
minimally one must have lexically represented tokens.
Traditionally, as well as in modern analyses, one talks about
two word accents, ‘accent 1’ and ‘accent 2’, or (nowadays less
frequently) ‘acute’ and ‘grave’, and these terms refer to the
full tonal contour assigned in intonation, where lexical tones
have been integrated. Phonologically, the most recent
analyses tend to defend a representationally privative
distinction, whereby either of the tonal configurations is taken
to be marked and lexically represented (in the crucial sets of
forms), the other being either assigned by default, or simply
being the empty set. The analysis presented in this chapter is
of the latter kind, where accent 2 is taken to be the marked
category, by virtue of having a specific tone as part of the
lexical representation of a number of morphemes.2 Accent 1 is
taken to be intonation, pure and simple, and hence it is
invariably the result of postlexical processes. Proposals that by
and large adhere to this view include Sweet (1877), Rischel
(1963), Haugen (1967), Elert (1970), Engstrand (1995, 1997),
Kristoffersen (2000), and Riad (1998b, 2003a, 2006, 2009a).
Other analyses have proposed a so-called equipollent
representation (Bruce 1977, however with accent 2 considered
as marked; Gussenhoven 2004), or a reverse privative (p.182)

distinction, where accent 1 is taken to be lexically specified

and accent 2 is assigned by default (Lahiri, Wetterlin, and
Jönsson-Steiner 2005a; Kristoffersen 2006, 2007; Wetterlin
2007, 2010).3 A very useful discussion and comparison of the
various proposals made in the literature is given in Naydenov
(2011).

Page 2 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

9.1 Basic facts

The terms ‘accent 1’ and ‘accent 2’ are mostly used to refer
holistically to the entire tonal contour of words, that is,
including the accents assigned by intonation. The distinction is
typically illustrated with word pairs like those given in (182a).
Accent is indicated by a raised digit before primary stress.

(182) The tonal accent distinction

Page 3 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

a. 1 ‘the duck’ 2 ‘the spirit’

ˈand-en ˈande-n
1 ‘the steps’ 2 ‘the ladder’
ˈsteg-en ˈstege-n

b. 1 ‘angry’ 2 ‘mean’
ˈsyrak ˈelak
1 ‘id.’ 2 ‘mustard’
ˈketchup ˈsenap

Page 4 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an
individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see [Link] Subscriber: University of
Colorado at Boulder; date: 11 March 2017
Tonal word accents

The examples in (182a) are the closest one gets to minimal

pairs within the same word class, but they are defective as
minimal pairs in view of the fact that the words have different
morphological make-up, the accent 1 forms typically being
monosyllabic stems, and accent 2 typically being represented
by disyllabic stems. In larger sets of accent pairs one finds
other concomitant differences, like word class. There are
about 350 such pairs in Swedish (Elert 1972), and more than
3,000 in Norwegian (Leira 1998), due mostly to later vowel
neutralization between historical /ɑ/ and /e/ in poststress
syllables. The contention of this is that the functional load of
the tonal distinction is virtually zero, and several dialects also
lack it entirely.

The fact that there is a lexical distinction between forms is not

in question, and to show this it is better to look at forms like
those in (182b). These forms are not segmentally identical, but
they are monomorphemic and belong to the same word class,
indeed are semantically close to each other. The intonation in
citation form is different between the members of each pair,
and this means that there must be a lexically represented
distinction between them, causing the tonal difference (for the
representation, see (185) and chapters 11 and 14).

There are some basic conditionings. There is no requirement

of sonority in the syllable to which the lexical tone associates.
This is a difference vis-à-vis Danish (p.183) stød which can
only be realized in syllables with two sonorant moras (Basbøll
2005: 272ff.). Similar restrictions obtain also in tonal accent
systems like Central Franconian and Lithuanian, but not in any
of the Scandinavian tonal varieties.

The lexical (or postlexical, in compounds) tone of accent 2

must associate to the primary stressed syllable: 2ˈsommar
‘summer’, 2ˈhögˌsommar ‘high summer’ (*ˈsom 2 mar, *ˈhög 2
ˌsommar). This tone has precedence to the stressed syllable,
and thereby pushes the intonation tones that make up the rest
of the contour to the right. In accent 1, where there is no
lexical tone, the accentual tones are supplied by intonation
and may hence associate directly to the primary stress syllable
(1ˈketchup).

Page 5 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Accent 2 only occurs in words that have a poststress syllable,

whereas accent 1 has no such limitation. The extra syllable
may be stressed or unstressed, but it must be there: 2ˈnunnor /
nʉnμ-ur2/ ‘nuns’, 2ˈsenap /senɑp2/ ‘mustard’, 2ˈbarn-ˌdom /bɑrn
+dumμ/ ‘childhood’, 2ˈvärn-ˌplikt /vɛrn+plikt/ ‘military service’.
In a few dialects, apocope has given rise to what looks like
accent 2 in monosyllables (usually called circumflex, Lorentz
2008). A representational fact that correlates with the space
requirement on the part of accent 2 is the fact that it generally
contains one extra tone (the lexical tone) in comparison with
accent 1.4 Monosyllables and final stressed forms invariably
exhibit accent 1: 1ˈbil ‘car’, ortocera 1ˈtit ‘orthoceratite’ (*2ˈbil,
*ortocera 2ˈtit).

There are two sources for accent 2. One is the lexical marking
of morphemes with a lexical tone, be it on roots or—more
regularly—on suffixes. The other source is prosodic/
postlexical. In Central Swedish and many other dialects,
accent 2 is regular in any word structure that contains two
stresses (compounds, some derivations, see (126)). The tonal
configuration is the same in both realizations of accent 2, but
there is a difference in association pattern, due to the
presence of more than one stress in e.g. compounds. Examples
of forms containing more than one stress are given in (183).

(183) Compounds and compound-like forms

2
Compounds: ˈsommarˌdag ‘summer day’, ba2ˈnan
ˌskal ‘banana peel’, 2ˈbyxˌkjol ‘pant-
skirt’
2
Formal ˈävenˌtyr ‘adventure’, 2ˈparaˌdis
compounds: ‘paradise’, 2ˈarˌbete ‘work’
2
(Some) ˈsjukˌdom ‘illness’, 2ˈkraftˌfull
derivations: ‘forceful’, 2ˈunderˌbar ‘wonderful’

The tone‐bearing unit in Central Swedish is the stressed

syllable, namely the foot. Equivalently, one could say that the
minimal prosodic word is the tone‐bearing (p.184) unit and
that association takes place to the head, in an order of

Page 6 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

priority, where a lexical tone has precedence over an

intonation tone (Riad 1998b).

9.2 Phonological representation

Accents are very often realized in the intonation contour, such
that it is fair to say that the regular pattern is for every
primary stress to exhibit word accent.5 There are, however,
two prominence levels at which accent is realized (Bruce
1977; Heldner 2001). Thus, the presence of an accent on a
syllable means that a given prosodic word has received a
certain level of prominence. The first notch up from simply
stressed is known as ‘word accented’, and the next higher
prominence level is called ‘focus accented’. In Myrberg
(2010), these are called prominence levels 1 (word accented)
and 2 (focus accented). Bruce (1977) uses the terms ‘word
accent’ and ‘sentence accent’. Later the terms ‘word accent’
and ‘focus accent’ have come to be used, and those are the
terms used here (Heldner 2001; Bruce 2007). The realization
of each of accent 1 and accent 2 is different in the two
prominence conditions, as laid out in (184). The (lexical or
postlexical) tone that distinguishes accent 2 from accent 1 is
bolded.

(184) Accents and prominence levels

Page 7 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Prominence level Accent 1 Accent 2 Accent 2 in compounds Typical functions

Focus accent L*H H*LH H*L*H focus, contrastive topic

Word accent HL* H*L H*L given material, second

occurrence focus (post
focally), new material
(non-final in the
phrase)

Page 8 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an
individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see [Link] Subscriber: University of
Colorado at Boulder; date: 11 March 2017
Tonal word accents

The fact that there are two different accent ‘melodies’ (accent
1 and accent 2) is of no importance to the level of prominence
as such. The accents have distinct melodic properties and
different association patterns at both levels, leading to a
characteristic timing difference between the identical
(intonational) parts of the contours (Bruce 1977, 2007a). This
timing difference recurs in pretty much every dialect.

(p.185) 9.2.1 Tonal contours

The realization of the tonal accents in the focus accented
condition are given in Fig. 9.1, 9.2, and 9.3. These represent
the same realization as one gets in citation form.

Fig. 9.1. Accent 1, simplex, focus accent,

ano 1ˈnyma ‘anonymous’. The tonal
contour is L*H, followed by a L%
boundary tone.

Radio Sweden (SR)

Page 9 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Fig. 9.2. Accent 2, simplex, focus accent,

2
ˈmänniskor ‘people’. The tonal contour is
H*LH, followed by a L% boundary tone.

Vetandets värld, SR

Fig. 9.3. Accent 2, compound, focus

accent, 2ˈsebraˌfink ‘zebra finch’. The
tonal contour is H*L*H, followed by a L%
boundary tone.

Vetandets värld, SR

The compound has two association points. One is the first

stress, where the accent 2 tone (H*) associates, and one at the
last stress of the compound, where the prominence tone
associates (L*H). The associating L* spreads backwards to
(p.186) the H*, instantiating a clear fall right after the H*, in

CSw.6 This is clearly visible in long compounds (see chapter 12

and Fig. 9.6). In diagrams like (186) this is marked with an
arrow.

Page 10 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Let us next look at the realization of the accents in the word

accented condition, the lower degree of tonal prominence. The
panel in Fig. 9.4 begins (p.187) with two focus accents (the
first is accent 1, fa1, the second accent 2, fa2), but all the
remaining accents are word accents, of both types (wa1, wa2).
This is seen in the variable timing of the HL contour between
the two accents (HL* and H*L, respectively).

Fig. 9.4. Accent 1 and accent 2 at both

prominence levels. The second tier from
the bottom indicates tonal contours. The
bottom tier indicates prominence level,
where ‘fa’ is focus accented and ‘wa’ is
word accented, the following digit
indicating accent 1 or 2.

Människor och tro, SR

The panel in Fig. 9.5 illustrates the lower prominence level in

long compounds.

Page 11 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Fig. 9.5. The two compounds 2ˈtvärˌveten

ˌskapligt
‘interdisciplinary’ and 2ˈforsknings-pro
ˌjekt ‘research project’ occur with word
accents in the lower prominence level.
The bottom tier indicates prominence
level, where ‘fa’ is focus accented and
‘wa’ is word accented, the following digit
indicating accent 1 or 2.

Människor och tro, SR

9.2.2 Association patterns

A given tone‐bearing unit—a stressed syllable/a foot—can only
accommodate a single tone. While several tones are contour
tones, it is always only one segment of such a contour tone
that actually associates, while the other tone either leads into
the association point (i.e. HL*) or trails after it (i.e. L*H).
Bruce (1987) shows that a trailing tone has variable timing in
utterances of different length, a fact that indicates that it is
not associated.

The accent 2 H* tone always has precedence in association,

and this holds for all varieties of Swedish and Norwegian.
When there is no lexical tone present, the prominence L* tone
associates. This is illustrated with stylized examples in (185).
We add the boundary tone (which is aligned rather than
associated) that occurs at the end of citation forms.

Page 12 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

(185) Simplex accent 2 and accent 1, representation of

focus accent (fa)

(p.188) Thus, a large portion of the tonal contour (i.e. LHL) is

identical, but the timing differs due to the precedence of the
initial lexical tone to the stressed syllable (Bruce 1977).

In compounds and prosodically similar words, there are

several minimal prosodic words, hence several stressed
syllables, hence several tone‐bearing units. This allows for
both the accent 2 H* tone (postlexical) and the prominence
L*H tone (=accent 1) to associate. These tones associate to
the first and last tone‐bearing unit of a compound, thereby
creating what is sometimes called a ‘connective’ prosodic
structure.7

(186) Compound accent 2, representation of focus accent

(fa)

This is the compound accent generalization of Central Swedish

(see 5.2.1). As described in Riad (1998b), there are a number
of constraints involved that create this structure. Beside the
obligatory association of the accent 2 tone to the primary
stress, there is right-alignment of the prominence tone, and
the association of the prominence tone (L*H). Also, there is
left-alignment of the prominence tone, creating the spreading
of the L* tone segment to the preceding H*. Fig. 9.6 repeats
the example of a long compound given in Fig. 5.1. It contains
five stresses, the first and last of which are tonally marked.

Page 13 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Fig. 9.6. Long compound showing the

connective tonal pattern of accent 2

Vetandets värld, SR

(p.189) 9.3 A note on typology

When we look across the Scandinavian dialects for patterns in
their tonal grammar, we find that they lend themselves to
uniform description and analysis. Several properties are
shared in all dialects, while other properties vary in small and
orderly ways. This means that it is indeed meaningful to talk of
a coherent dialect area based on very similar grammatical
properties.

The lexical pitch accent opposition can be uniformly described

as privative (i.e. accent 2 is instantiated as a lexical tone,
while accent 1 has no lexical specification). The central tones
obey the Obligatory contour principle (Leben 1973) in a
strikingly uniform manner across dialects. Thus, if the first
tone is H, the next tone will be L or begin with L, and vice
versa. Part of the tonal contour is also functionally identical
between dialects and the crucial tones appear in the same
order. Thus, the lexical tone (accent 2) is always initial in the
primary stressed syllable. The prominence tone exhibits some
variation in how it holds itself with respect to the lexical tone
to the left, the boundary tone to the right, and whether or not
it spreads. Other areas of variation concern the extent of tonal
association, where Central Swedish has two associated tones
in compounds, whereas e.g. South Swedish only has one. Also
the orientation of unassociated tones varies (left/right, see
Riad 1998b, 2006, 2008).

Page 14 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

When we turn to the lexical distribution of accents, there is

also a strong core of shared patterns. The strongest shared
pattern is that of syllabic suffixes, which induce accent 2 in
dialect after dialect, in the canonical conditions (see 11.4.2).
Bruce (1998: 50) notes that it is the accent of root morphemes
that is the least stable across dialects, and that is also a much
smaller number of token forms. The bottom line is, however,
that the North Germanic tonal systems are quite
homogeneous, both regarding tonal behaviour and tonal
distribution. Several of the areas of divergence can be related
to representational differences. For instance, the major divide
between dialects that have uniform accent 2 in compounds
and those that permit either accent in compounds is directly
related to whether or not there is mandatory association of the
prominence tone to a secondary stress (shown in Fig. 9.7).
Double association entails uniform accent in forms that have
more than one stress (stressed syllables being the tbu),
whereas single association admits tonal variation in
compounds just as in simplex forms. In such dialects, lexical
tone and morphological structure also influence the resultant
accent, as well as prosodic factors.

Page 15 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

Fig. 9.7. Map of the major tonal dialects

of North Germanic

Dark areas have a L lexical tone in accent 2, light areas

have a H lexical tone in accent 2. The isogloss marks the
boundary between connective and non-connective
dialects. East and north of the isogloss, compounds
uniformly get accent 2 (two association points), whereas
west and south of the isogloss, compounds vary in accent
(one association point).

We assume that tonal accent is an obligatory property of the

maximal prosodic word, following Myrberg and Riad (2013).
While there are certainly deaccented words in phrasal
phonology, we take those instances to involve the

Page 16 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

incorporation of deaccented words into a maximal prosodic

word nearby. Previous scholarship (p.190) has often assumed
that there is a higher category, variably referred to as the
Accentual phrase (Kristoffersen 2000; Abrahamsen 2003;
Hognestad 2012), the Tonal foot (Fretheim and Nilsen 1989;
Nilsen 1992), or the Prosodic word in a different sense from
that used here (Bruce 1998; Hansson 2003). However, while
these models may fairly describe the tonal behaviour, the
prosodic categories do not themselves align with morpho
syntactic structure in a principled way, and (p.191) hence do
not organically connect with the prosodic hierarchy as
conceived of in phonology in general.

The unitary domain of the tonal accents as the maximal

prosodic word can be related to the culminative prominence
function of accent, as word accent and focus accent (see 12.2).
In each instance, only one tonal accent will be assigned to the
maximal prosodic word. Various intonational and information
structural factors influence whether a word accent or a focus
accent is assigned, while which of accent 1 or accent 2 is
assigned is determined by morphological information, prosodic
information at the word level, or else by default, see (131).

Notes:
(1) Stød is a laryngeal feature, realized as creaky voice or full
glottal stop, which has a prosodic distribution. It occurs in
several Danish varieties, with a grammatical distribution, but
also shows up as a facultative feature in varieties of Swedish,
west of Stockholm, where it is known as ‘Eskilstuna curl’ (Sw
Eskilstunaknorr or -knarr).

(2) The idea that tones are represented directly in the lexicon,
rather than via some diacritic, is pursued in the analysis of a
variety of languages, e.g. Japanese (Meeussen 1972; Poser
1985), Bantu (Hyman 1982; Pulleyblank 1986; Downing 2006),
Serbo-Croatian (Inkelas and Zec 1988), Lithuanian (Blevins
1993), Latvian (Karinş 1996).

(3) There is also a proposal that assumes no underlying tones

(Morén-Duolljá 2007, 2013) and which derives the tonal
variation from the prosodic structure of words. The

Page 17 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017
Tonal word accents

assumptions regarding the prosodic word are, however,

radically different in this proposal, making it difficult to
evaluate with respect to other proposals.

(4) Ultimately, there is a diachronic reason for this, see Riad

(1998a, 2005).

(5) There are some systematic cases where this does not hold,
listed in (267).

(6) In otherwise closely related North Swedish varieties, the

initial fall is typically less pronounced, sometimes leading to
the perception of a lesser prominence at the initial stress of
compounds, and primary stress at the last.

(7) The term ‘connective’ is also used with a more general

meaning, referring to the fact that accent 2 in nearly all
instances occurs on a morphologically complex form, even if
only by plural inflection (e.g. Elert 1981).

Access brought to you by: University of Colorado at Boulder

Page 18 of 18

PRINTED FROM OXFORD SCHOLARSHIP ONLINE ([Link]). (c) Copyright Oxford University Press, 2017. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see [Link] Subscriber:
University of Colorado at Boulder; date: 11 March 2017