The Duress of Stress: On Dutch Clippings › catalog › avt.14.21vij › fulltext ›...

The Duress of Stress: On Dutch Clippings

Ruben van de Vijver

0. Introduction

In Dutch it is possible to abbreviate names and nouns, adjectives and adverbials in order to form so-called clippings.1 Clippings are either disyllabic or monosyllabic. The set of possibilities is given in (l).2

(1) a two syllables Clipping Base brábo Brábant 'Brabantine' Bába Bàbangída name hérbo hérbeoordeling 're-evaluation' Àbdoe Àbdoesjáparov name Dèbóor Dèbóra name Pàndoor Pàndóra name

b one syllable Clipping Base stuud stù.déer 'geek, nerd' Dáaf David name

At first sight these clippings seem to offer a great variety in terms of syllable structures (open-open, closed-open, open-closed, closed-closed and open). Yet, the exact form of the clipping is determined by the stress pattern of the base and the

1 I wish to thank Geert Booij, Caroline Fery, Jan Kooi, Micheal Redford, Anthi Revithiadou and an anonymous reviewer for useful comments on an earlier version. Furthermore I want to thank everyone whose name I have (ab)used. The data come from spontaneously spoken Dutch gathered by the author. None of the above mentioned are responsible for errors of fact or interpretation. 2 In this paper Dutch orthograpy is followed. Tense vowels /a, e, o, y/ are written with two vowels symbols if only one consonant follows. For example, kaas /kas/ 'cheese', vrees /vres/ 'fear', Koos /kos/ 'name', Truus /trys/ 'name'. Except for /e/, word-final tense vowels are written with a single symbol: spa /spa/ 'spade', alpino (capitals denote lax vowels in this footnote) /Alpino/ 'baret', paraplu /paraply/ 'umbrella'. Word-final tense Id is written with two symbols while schwa is written with one symbol: zee 'sea' versus ze 'she, they'. Word-internally, tense vowels are written with a single letter and followed by a single consonant-letter. For example baden /bade/ 'to bathe, baths', vrede /vrede/ 'peace', molen /mole/ 'mill' and buren /byre/ 'neighbors'. Tense /i, u, øl are written ie, oe and eu always: bier 'beer', boer 'farmer' and beuk 'beech'. Lax vowels are spelled with a single letter. For example, bak /bAk/ 'bake', bek /bEk/ 'bill', bok /bOk/ '(he)goat' and buk /bUk/ 'stoop'. Diphthongs are spelled ei or ij, ui and au or ou.

Linguistics in the Netherlands 1997, 219–230. DOI 10.1075/avt.14.21vij ISSN 0929–7332 / E-ISSN 1569–9919 © Algemene Vereniging voor Taalwetenschap

220 RUBEN VAN DE VIJVER

phonology of Dutch. This will be shown by focusing on the issue: which base yields which clipping. The syllable structure of the base is retained in the clipping, only if this does not violate the certain properties of Dutch stress. These properties are: a) final open syllables may not be stressed and b) closed syllables are stressed. In other words, the stress pattern and the syllable structure of the base are found in the clipping, phonology permitting. Thus, the clipping may only deviate from the syllable structure of the base under the duress of stress.

There are a great many clippings that have a diachronic source and which do not reflect a productive pattern, about which nothing will be said in this paper. For example Kobus or Koos for Jakobus, or Nelis for Cornells, or Henk from Hendrik. Native speakers will not produce these clippings naturally. This can be seen if Koos is considered. It comes from Jakobus or Kobus. For some reason the second syllable has been deleted, except for its coda (Ko<bu>s). If this were a process that were active in the synchronic phonology of Dutch, we would expect that *Nees from Nelis would be wellformed. Unfortunately, it is not. The clipping Nees exists, but it comes from the base Neeskens and not from Nelis.

The paper is organized as follows. In section 1 I will present the data, which are analyzed in section 2. Section 3 discusses the default vowel in clipped forms. Section 4, finally, discusses some ancillary issues and contains the main conclusions of the paper.

1. Data

As has been said above, clippings consist of either one or two syllables. In (2), I will give examples of monosyllabic clippings.3

(2) Clipping Base gloss frust frùs.tréer 'frustrated person' símp sím.pel 'simpleton' stúud stù.déer 'geek, nerd' Dáaf Dà.víd name Sóof So.fíe name Klúif Klúi.vert name Wóut Wóu.ter name Réin Réi.nòud name

3 For clippings derived from verbs, I have taken the stem of the verb as the base for the clipping. I have done so, because all morphological alternations involving the verb use the stem as their basic form.

ON DUTCH CLIPPINGS 221

The clippings all have stress, and they all end in a consonant, even if this means that the syllable structure of the base is not respected. Another generalization which can be made is that all monosyllabic clippings have disyllabic bases. This generalization is not iron clad, however, given an exception such as bieb which has bibliotheek as base (see section 2.4).

The last three examples of (2) are especially striking, since the first sylable contains a diphthong, which are considered to be heavy (cf. van der Hulst 1984, Kager 1989, Booij 1995). If the size of clippings are simply feet which are somehow chopped of the base, then the clipping would have been *Klui, *Wou and *Rei. Since these are not the attested clippings it can be concluded that the size of clippings in Dutch are not determined by foot structure alone.

In (3) some examples of clippings consisting of two open syllables are given.

(3) Clipping Base gloss á.so à.so.ciáal 'antisocial' brá.bo Brá.bant 'Brabantine' chá.ri chá.ri.tàs 'charity' dé.pri dè.pre.ssíef 'depressed person' dé. pro dè.pre.ssíef 'depressed person' pá.ddo pá.dde.stòel '(psychedelic) mushroom' ré.fo ré. for. ma. tò. risch 'orthodox protestant (reformational)' Bá.ba Bà.ban.gída name Káu. ka Káu.ka.sùs name Mó.lo Mó.lo.tòv name Ná.vra Nà.vra.ti.lò.va name

In all words in (3) the first syllable of the base is stressed, and the syllable pattern of the base is kept in the clipping except for those clippings that have an unstressed second syllable, namely Ba.ban.gi.da which is clipped to Ba.ba and ré.for.ma.tô.risch which is clipped to ré.fo. This will be discussed in section 2.3. None of the forms in (3) end in a schwa, following a general restriction against schwa-final words (van Oostendorp 1995). Name-clippings do not allow for a default vowel, but in non-name clippings the final schwa may be replaced by the default vowel in clippings: [o]. For example, the word paddestoel is realized with a schwa as the nucleus of the second syllable, but it is shortened as paddo, with a tense round midvowel. Since this vowel has no apparent morphological function it can only be a default vowel.

In (4) I will present some examples of clippings consisting of a closed syllable followed by an open syllable.


(4) Clipping Base gloss áf.ko áf.kor.ting ' abbreviation' ál.to àl.ter.na.tief 'alternative' cóm.bi com.bi.ná.tie ' combination' dís.co dís.co.théek 'discotheque' hér.bo hèr.be.óor.de.ling 're-evaluation' hér.ko hèr.kon.tró.le 're-examination' lim.bo Lím.burg 'Limburger' ór.di or.di.náir 'vulgar' Gór.ba Gór.bat.sjöv name Àb.doe Àb.doe.sjá.pa. rov name

In the words in (4) the stress pattern of the base is retained in the clipping, just as in the words in (3). The second syllable of the clipping is open and unstressed, while the second syllable of the base is unstressed but may be closed. These forms further illustrate the absence of final schwas. The word herbeoordeling is realized with a schwa as the nucleus of the second syllable, but its clipping is herbo, avoiding a schwa as final vowel in the clipping.

In (5), clippings of the form 'open syllable followed by a closed syllable', are illustrated .

(5) Clipping Base gloss A.kós A.kóz.ba name Re.gfen Re.gi.na name Pa.trfes Pa.trí.ci.a name Fe.líes Fe.lí.ci.a name Dè.bóor Dè.bó.ra name Re.bék Re.bé.kka name A.níet A.ní.ta name A.léid A.léi.da name

These bases, which are always stressed on the second syllable, yield a clipping which always ends in a consonant.

In (6) a form consisting of two closed syllables is given.

(6) Clipping Base gloss Pàn.dóor Pan.dó.ra name

Again, the second syllable of the base is stressed and the second syllable of the clipping is stressed and closed.

In this section I have presented the relevant data, which will be analyzed in the next section.


2. Analysis

In this section I will first focus on the analysis of disyllabic clippings and then I will turn to monosyllabic clippings. There are two constraints that will be unviolated in clippings. The first is that the left edge of the clipping and the left edge of the base match. This means that the first segment of the base is also the first segment of the clipping. This constraint is called ANCHOR-LEFT (McCarthy and Prince 1994). The second constraint is that the order of the segments in the base is carried over to the order of the segments of the clipping. In other words, there is no intrusion of segments in the clipping, as compared to the base, nor is there any skipping of segments in the base, as compared to the clipping. This constraint is called CONTIGUITY (McCarthy and Prince 1994). Since these constraints are unviolated I will not include them in the discussion.

2.1. Disyllabic clippings; simple cases. Clippings are maximally two syllables. This follows from a constraint saying that some edge of every syllable of the clipping should be aligned to some edge of the PrWd (Kager 1995), which is given in (7).

(7) ALIGN(σ, P R W D ) 'Some edge of every syllable must be aligned to some edge of the PrWd'.

This constraint rules out clippings larger than two syllables. For example, a clipping of three syllables violates it, since the middle syllable is not aligned with some edge of the PrWd. If it is ranked above MAX-I-OS E G ('input segments must be retained in the output') this constraint gives the required disyllabic maximality. In order to see how this works consider the name Abdoesjaparov. The tableau in (8) shows the evaluation of four candidates against these constraints. All candidates, except for abdoe and ab, violate ALIGN(G, P R W D ) , and hence they fail. Evaluated against MAX-I-OSEG, the former candidate fares best.

(8) Tableau of Abdoesjaparov (part 1)

Abdoesjaparov ALIGN(G, PRWD) MAX-I-OSKG

Abdoesjapa *! rov

Abdoesja *! parov

Abdoe sjaparov

Ab doesjapalrov


2.2 The role of the syllable structure of the base in disyllabic clippings. There is a candidate which has been left out of the discussion of Abdoesjaparov. Namely a candidate with two closed syllables: abdoesj. By the logic of the ranking in (8), this candidate should win, since it incurs less violations of the constraint MAX-I-OSEG than the candidate abdoe. In order to see why abdoesj fails, it is necessary to introduce a constraint. This constraint is a kind of syllable integrity constraint which requires that the syllabic role of segments in the base is equal to the syllabic role of segments in the clipping (cf. Itô, Kitagawa and Mester 1995, who propose a similar constraint concerning foot-structure). In other words, if a segment is an onset in the base it must also be an onset in the clipping.4

(9) STROLE (McCarthy and Prince 1994) Segments in the clipping and their correspondents in the base should have identical syllabic roles.

If this constraint is ranked above MAX-I-OSEG, the candidate abdoesj is ruled out. This can be seen in (9), where the candidates abdoe and abdoesj are evaluated against STROLE and MAX-I-OSEG. In the candidate abdoe all segments are syllabified in the same way as in the base, while in the candidate abdoesj the segment sj is in the coda of the second syllable, whereas it is the onset of the third syllable in the base. This is laid down in (10), where the syllabification of the base is indicated by dots.

(10) Tableau of Abdoesjaparov (part 2).

Ab.doe.sja.pa.rov STROLE MAX-I-OSEG

ab.doe sjaparov

ab.doesj *! aparov

2.3 The role of the stress pattern of the base in disyllabic clippings. As has been noted in section 1, the syllable structure of the clipping may deviate from the syllable structure of the base, under the duress of stress. If the second syllable of the base is stressed, its corresponding syllable in the clipping is closed, irrespective of whether the syllable is closed in the base. This follows from the phonology of final stressed vowels in Dutch. Word-final vowels tend not to be stressed in Dutch (Kager 1989). Children as young as three or four year are aware of this generalization (Nouveau

4 The constraint *CODA may come to mind to deal with these cases. However, the constraint STROLE compares the syllabic role of a segment in the base with the syllabic role of the same segment in the clipping. Since the analysis of Dutch clippings presented here capitalizes on the relation between base and clipping, STROLE is preferred over *CODA.


1994). In an imitation test, children make most mistakes if they are asked to imitate words with final vowels which are stressed. Instead of faithfully imitating words with a stress on the final vowel, they tend to retract stress, so that it is not final anymore, or to add a consonant to the final stressed syllable, so that it is not open anymore. Van Oostendorp (1995) has observed that adjectives illustrate the same tendency. Adjectives take a schwa ending in indefinite DPs of the masculine/feminine gender. The ending only shows up when the adjective ends in a consonant (11a), not when it ends in an unstressed vowel (11b). This is illustrated in (11), where the symbol after the + is a schwa.

(11) a een ráuw+e vis 'a raw fish' b een míca(*+e) tafel 'a micaceous table'

een séxy(*+e) dame 'a sexy lady'

Adjectives which end in a stressed vowel pattern with the adjectives ending in a consonant. This can be seen in (12). An inserted glide is added between brackets.

(12) een blíje boodschap (*blij) 'a happy message' een continü(w)e stroom (*continu) 'a continuous stream' een gedwée(j)e volgeling (*gedwee) 'a meek follower'

I take these two effects (the relative rarity of words that end in a stressed vowel and the behavior of adjectives which end in a stressed vowel) to be caused by a constraint, which is unviolated in clippings. The constraint is given in (13).5

(13) FINALSTRESSCLOSED (FSC) Final stressed syllables should be closed.

The reason why (13) is unviolated in clippings (but violated in base forms) is because clippings do not have their own input form (i.e. they do not have an underlying form of their own). As a result constraints requiring identity between input and output (i.e. FAITHFULNESS Prince and Smolensky 1993) have nothing to

5 The constraint in (13) might reflect a more general tendency of stressed syllables to be closed. Micheal Redford (p.c) pointed out to me that in American English names can be clipped as well. Names like Julia, Cori, Dylan, Kathy and Lisa, which all have a stress on the first (open) syllable, are clipped as Jul, Cor, Dyl, Kath and Lis respectively. The clippings have closed syllables because otherwise the syllable of the clipping would remain unstressed.

An example of the desire of stressed syllables to be closed might be observed in the alternative realizations of the name Gorbatsjov. In Dutch, this name has two alternative stress patterns which are accompanied by two alternative syllable structures: Gór.ba.tsjov and Gor.bat.sjov. In the first alternative, the second unstressed syllable is open, while in the second alternative the second syllable is stressed and closed.


say about the form of clippings. Because of this, constraints that are normally hidden under FAITHFULNESS, are in control of the phonology of clippings. This situation is well-known and dubbed 'emergence of the unmarked' by McCarthy and Prince (1994). Moreover, there is a constraint saying that the stress pattern of the base should be retained in the clipping (14).

(14) MAX BASE-CLIPPING(STRESS) (MAX-BCSTRESS) Syllables that are stressed in the base are also stressed in the clipping.

Please note that the constraint in (14) does not require that the degree of stress in the base should correspond to the degree of stress in the clipping. The name Aléida serves to illustrate (14). The first candidate, Aléida, violates the constraints which ensures maximal disyllabicity ALIGN(G, PRWD). The second candidate, Aléid, violates STROLE but it complies with FSC. Since STROLE is dominated by FSC, the third candidate, Aléi, is worse than Aléid. The fourth candidate, Al, violates MAX-BCSTRESS.

(15) Tableau of Aléida

Aléida FSC ALIGN(C, P R W D ) MAX-BCSTOKSS STROLE MAX-I-O S E G

Aléida *! Aléid * a

Aléi *! da 1 Al * *! eida

Next, I will consider the case of' Bàbangída, which is truncated to Baba (words like afkorting, which have an unstressed closed second syllable, are handled likewise). The approach sketched so far would predict that Bában would be the correct outcome. However, Bában would contain a heavy unstressed syllable. Closed syllables are heavy in Dutch (van der Hulst 1984, Kager 1989, Booij 1995). The constraint requiring heavy syllables to be stressed is given in (16).

(16) WEIGHT-TO-STRESS (WSP)

Heavy syllables are prominent on the grid and in foot structure.

In the case of Bàbangida three clippings can be considered. The first, Bában, has an unstressed heavy syllable: the last syllable of the clipping is unstressed. The second candidate, Babán, violates MAX-BCSTRESS. The third candidate, Bába, violates neither constraint and hence wins. For reasons of succinctness, only three constraints are


given in tableau (17).

(17) Tableau of Babangida

Bàbangida M A X - B C ^ WSP MAX-I-OS H G

Bában *! gida

Babán *! gida

Bába ngida

2.4 Monosyllabic clippings. The next issue to consider are monosyllabic clippings which usually arise from disyllabic bases. Examples are names like David, which truncates to Dáaf and Sofíe which yields Soof The latter clipping is the only example which disrespects the stress pattern of the base. Other sources of monosyllabic clippings are trisyllabic words where the final two syllables are in hiatus: biblio-théek yields bieb (theek is arguably a bound morpheme, which should not be considered as part of the base of truncation; see below) and Bélla yields Beel. First I will discuss monosyllabic clippings from a disyllabic base. Apparently, in the pattern discussed in this paper, a clipping cannot include the final syllable of the base. I cannot think of a principled reason why this should be the case, except for the obvious fact that clippings are intended to be shorter than the base. This would not be the case if the final syllable were included in the clipping.6 There seems to be an analogy in stress, where the final syllable is extrametrical (Kager 1989). Perhaps the ban against base final syllables in the clipping again illustrates the 'unmarked' phonology of clippings. In the unmarked case the final syllable is invisible in stress, and since clippings represent unmarked prosody, the final syllable of the base is unavailable for truncation. But for now, I will just formulate the constraint in (18), noting that in an ideal phonology this constraint will have to be replaced by a more principled explanation.

(18) *FINALSYLLABLEBASECLIPPING (*FSBC) The final syllable of the base may not be present in the clipping.

This constraint will have to be ranked above MAX-BCSTRESS, as is shown by the example of Sofie, which is stressed on the final syllable. The clipping is Sóof which ignores the stress pattern of the base and thus violates MAX-BCSTRESS. The other active constraint is FSC, which says that stressed open final syllables are prohibited.

6 Please remember that the first syllable must be present due to the constraint ANCHOR-LEFT; see section 2 above.


This constraint ensures that the candidate Sóo will not be the optimal candidate. The ranking is laid down in (19) in which three candidates of the base Sofie are compared, namely Sóo, Sóof and Sófie?

(19) Tableau of Sofie

Sofie FSC *FSBC MAX-BCmiBi

Sôo *! * Sòof * Sòfie *! *

Let me now turn to truncations like Béel from Bélia and bìeb from bibliotheek. Again I have not found a solid explanation which makes clear why two syllables behave as if they were one final syllable.

In the word bibliotheek, the final syllable, theek, is a bound morpheme which is still productively used. This is witnessed by neologisms such as speel-o-theek 'toy library' and mediatheek '(multi) media resource center'. The truncation bieb is therefore probably formed from the base biblio. The final two syllables of both Bélia and bìblio are in hiatus; the two vowels are in separate syllables (Bélia is syllabified Bé.li a and bìblio is syllabified bì.bli.o). Perhaps the final onsetless syllable is treated as a kind of clitic to the previous syllable, which has an onset. If true, the behavior of bases like Bélia and biblio can be captured by the constraint ranking proposed for disyllabic bases.

3. The [o] as default vowel in clippings

Non-name clippings and name clippings differ in that non-name clippings can have a default vowel, while name clippings cannot. The default vowel in clippings is [o], unlike the rest of the Dutch phonology where the default vowel is schwa. The default vowel in clippings only emerges in final position, which explains why it cannot be a schwa, since default schwa's are prohibited word-finally (van Oostendorp 1995). A clear example of a default vowel is found in the clipping of paddo from

7 The candidate Fie is not included, for two reasons. The first is that clippings consisting of one open syllable do not form a productive pattern (see also section 4). This can be made clear if Fie from Sofie is compared to the clipping of Nine, which is Nien and never *Nie (Nine Elenbaas p.c). The second is that it violates ANCHOR-LEFT, which has been assumed to be unviolated in section 2.


paddestoel8 There seems to be no motivation for the schwa of padde to be realized as [o] in paddo, other than satisfying the disyllabic shape of clippings and at the same time avoiding a final schwa. It is important to know that it is assumed that a schwa is a root node with no feature material dependent on it (van Oostendorp 1995). The avoidance of word-final schwa is accounted for by the constraint (20) (cf. van Oostendorp 1995).

(20) *FINALEMPTYROOT (*FER) No empty roots are allowed finally.

The question now is: why would [o] be the next best default-vowel? As a complete formal analysis would lead to far astray, only a tentative answer will be given here. The answer is that [o] is close enough to schwa not to be too prominent, but yet distinct enough from schwa not to be confused with it. Kager (1989) gives a reduction hierarchy, which is, from easy reducible to hardly reducible: a) /e/, /a/, b) /o/ and /ø/, c) HI and finally, d) /y, u/. The grouping of /e/ and /a/ as vowels that regularly reduce, in formal and in informal Dutch (Kager 1989), can be interpreted as meaning that the behavior of these vowels is too close to schwa to be the next best default vowels. This would make both /o/ and /ø/ the optimal default vowels for the final position. The løl can, however, be excluded because its grouping with /o/, in this short reduction grammar, is more a theoretical extrapolation than an empirical observation (van Oostendorp 1995). The vowel løl is hardly ever stressed and its reduction behavior can therefore not be observed.

4. Ancillary issues and summary

There are a couple of names that consist of an open monosyllable, like Mo for Monique, Ru for Ruud, or Ka for Karin. These are however forms with their own underlying form and hence not generalizable to clippings. Although Ru is a possible name in Dutch, this does not give rise to clippings like Gu for Guus or Lu for Lutsen. Instead, Guus does not reduce at all and Lutsen reduces to Lut or Luts.

There are a few names for which the clipped form is productively an open monosyllable. These names have a stressed first syllable and an /h/ in the second syllable of the underlying form. This /h/ is, however, realized as a glide on the surface (Gussenhoven 1993). For example, the name Móhammed truncates to Mo. If it were to be clipped to Moha, the second syllable would be an unstressed syllable

8 Names are immune to default vowels as is shown by names like Jacqueline, or Koevermans, the surname of a male soccer trainer, which would both end in a schwa when truncated according to the constraints sketched above. If they were like paddo, the clippings would have to be realized as *Jaco and *Koevo, but instead we find Jac and Koef.


beginning with an /h/, an impossibility in Dutch (Gussenhoven 1993, van Oostendorp 1995). If it were clipped to Moh the /h/ would appear in coda position, which is also prohibited in Dutch (van Oostendorp 1995). In Dutch, /h/ can only appear in the onset of a stressed syllable. In order to be faithful to the stress of the base, the second syllable is not included in the base. The same reasoning applies to the name Jóhan, which is clipped to Jo. It cannot be clipped to Joha since that would entail that the second syllable would have an /h/ in the onset of an unstressed syllable, nor to Joh which has an /h/ in the coda.

In recent literature (McCarthy and Prince 1995 and references therein) the nature of templates has been investigated. The conclusion is that the shape of the template is the result of the interaction of constraints on the relation between the base and the templatic form on the one hand and the unmarked phonology of a language on the other hand. This paper confirms this for Dutch clippings.

In this paper I have discussed the shape of clippings in Dutch. Clippings adhere to the syllable structure and the stress pattern of the base as much as possible. In two contexts the mimicking of the clipping breaks down. Firstly, if copying the base exactly would result in a final open stressed syllable and secondly, if an exact copy of the base would result in an unstressed closed syllable. In those cases the syllable structure of the clipping deviates from the syllable structure of the base. The stress pattern of the base is adapted in the clipping, if the base ends in a stressed, open syllable.

References

Booij, Geert (1995) The Phonology of Dutch, Clarendon Press, Oxford. Gussenhoven, Carlos (1993) 'The Dutch Foot and the Chanted Call', Journal of Linguistics 29, 37-63. Hulst, Harry van der (1984) Syllable Structure and Stress in Dutch, Foris, Dordrecht. Itô, Junko, Yoshihisa Kitagawa and Armin Mester (1995) 'Prosodic Faithfulness and Correspondence:

Evidence from a Japanese Argot', Linguistics Research Center, UCSC. Kager, René (1989) A Metrical Theory of Stress and Destressing in English and Dutch, Linguistic Models

14, Foris, Dordrecht. Kager, René (1995) 'Foot Templates and Root Templates', in M. Den Dikken and K. Hengeveld, eds.,

Linguistics in the Netherlands 12, John Benjamins, Amsterdam, Philadelphia. McCarthy, John and Alan Prince (1994) 'Emergence of the Unmarked: Optimality in Prosodic

Morphology', ms., Umass Amherst and Rutgers University. McCarthy, John and Alan Prince (1995) 'Faithfulness and reduplicative identity', in J. Beekman et al., eds.,

University of Massachusetts Occasional papers in Linguistics, 18, GSLA, Amherst, Mass., 249-384. Nouveau, Dominique (1994) Language Acquisition, Metrical Theory, and Optimality. A Study of Dutch

Word Stress. PhD dissertation, Utrecht University. Oostendorp, Marc van (1995) Vowel Quality and Syllable Projection. PhD dissertation, Tilburg University.

The Duress of Stress: On Dutch Clippings › catalog › avt.14.21vij › fulltext ›...

Documents

Transcript of The Duress of Stress: On Dutch Clippings › catalog › avt.14.21vij › fulltext ›...