Abstract
In this paper we discuss the simplification of the Dutch cluster /sxr/ and its relation to the variation found with /r/ and /x/. The place of articulation of the back fricative /x/ ranges from palatal to pharyngal in Dutch dialects (Goeman and Van de Velde 1999), the phonetic realisation of /r/ in Dutch is even more variable (Van de Velde 1996, Van de Velde and Van Hout 1999, Sebregts et al. 2003). Alongside many segmental variants including trills, fricatives and approximants at alveolar, retroflex and uvular place of articulation, a number of speakers have no segmental realisation of /r/ in certain contexts. This may lead to the disappearance of rhoticity altogether; in other words, there is r-deletion (Van den Heuvel and Cucchiarini 2001). However, speakers may also employ strategies to retain the contrastive potential of /r/ without its segmental realisation. That is, rhoticity within a lexical item is signalled by an alteration of surrounding speech sounds, such as consonants that form a cluster with /r/. For instance, a following coronal obstruent may be palatalised when an /r/ is lexically present but ‘deleted’, as in (1).
(1) bord /bçrt/ ‡[bçtj]
While /r/ can be said to be absent from the segment string, it leaves its mark on its surroundings by ‘merging’ with the following consonant (Plug and Ogden 2003).
In word-initial clusters, /r/ is also sometimes seen to ‘disappear’. This is the case for some speakers with uvular realisations of [R] when /r/ is in an onset cluster with the fricative /x/1. In schrift ‘notebook’ this leads to cluster simplification:
(2) schrift /sxrIft/‡[sXift]2.
The question is what exactly this simplification entails. There are three possible routes:
1. /x/ is deleted. In the Netherlands, a reduction of the cluster /sxr/ into [sr] is very often observed, also if /r/ is alveolar. This is one of the characteristic differences with
Flemish Dutch (Goossens 1973).
2. /r/ is deleted. Although /r/ is not an obvious reduction target as it is in coda position,
/r/-deletion in this cluster cannot be ruled out, as the result of the simplification is a
purely fricative cluster.
3. Merger of /x/ and /r/, comparable to that of /r/ and /t/ in codas. If this is the case, the
resulting sound should be one where the rhotic has altered the realisation of /x/ in such
a way that its contrastive potential (the opposition /sxr/ - /sx/) is left intact.
In order to determine which of the three possible routes of simplification is taken by the relevant speakers, we compare their realisations of /sxr/ and /sx/ clusters with respect to its auditory and acoustic properties. Our data come from three database sources:
• The Goeman-Taeldeman-Van Reenen-Project corpus (GTRP), containing local dialect data from 613 localities with 1 speaker each (Van den Berg 2003).
• The Hema corpus, containing urban dialect data from 10 major cities with 40 speakers each, socially stratified for sex and age (Sebregts et al. 2003).
• The VNL corpus, with Standard Dutch data from 8 regions with 20 speakers each, stratified for sex and age (Van Hout et al. 1999).
From each of these databases, we have selected one or more word pairs to be analysed auditorily and acoustically (see table 1). The results of these analyses will shed light on speakers’ simplification strategies, and on the nature of /r/ variability in standard and non- standard varieties of Dutch.
1 The place of articulation of the velar fricative is itself variable: it ranges from palatal to uvular.
2 Note that the onset cluster /sx/ is exceptional of itself: it is rare cross-linguistically, and many Dutch dialects have [sk] or [S] in the relevant lexical items instead.
(1) bord /bçrt/ ‡[bçtj]
While /r/ can be said to be absent from the segment string, it leaves its mark on its surroundings by ‘merging’ with the following consonant (Plug and Ogden 2003).
In word-initial clusters, /r/ is also sometimes seen to ‘disappear’. This is the case for some speakers with uvular realisations of [R] when /r/ is in an onset cluster with the fricative /x/1. In schrift ‘notebook’ this leads to cluster simplification:
(2) schrift /sxrIft/‡[sXift]2.
The question is what exactly this simplification entails. There are three possible routes:
1. /x/ is deleted. In the Netherlands, a reduction of the cluster /sxr/ into [sr] is very often observed, also if /r/ is alveolar. This is one of the characteristic differences with
Flemish Dutch (Goossens 1973).
2. /r/ is deleted. Although /r/ is not an obvious reduction target as it is in coda position,
/r/-deletion in this cluster cannot be ruled out, as the result of the simplification is a
purely fricative cluster.
3. Merger of /x/ and /r/, comparable to that of /r/ and /t/ in codas. If this is the case, the
resulting sound should be one where the rhotic has altered the realisation of /x/ in such
a way that its contrastive potential (the opposition /sxr/ - /sx/) is left intact.
In order to determine which of the three possible routes of simplification is taken by the relevant speakers, we compare their realisations of /sxr/ and /sx/ clusters with respect to its auditory and acoustic properties. Our data come from three database sources:
• The Goeman-Taeldeman-Van Reenen-Project corpus (GTRP), containing local dialect data from 613 localities with 1 speaker each (Van den Berg 2003).
• The Hema corpus, containing urban dialect data from 10 major cities with 40 speakers each, socially stratified for sex and age (Sebregts et al. 2003).
• The VNL corpus, with Standard Dutch data from 8 regions with 20 speakers each, stratified for sex and age (Van Hout et al. 1999).
From each of these databases, we have selected one or more word pairs to be analysed auditorily and acoustically (see table 1). The results of these analyses will shed light on speakers’ simplification strategies, and on the nature of /r/ variability in standard and non- standard varieties of Dutch.
1 The place of articulation of the velar fricative is itself variable: it ranges from palatal to uvular.
2 Note that the onset cluster /sx/ is exceptional of itself: it is rare cross-linguistically, and many Dutch dialects have [sk] or [S] in the relevant lexical items instead.
| Original language | English |
|---|---|
| Publication status | Unpublished - 7 Oct 2004 |
| Event | VarPhon (University of Potsdam) - Duration: 7 Jan 2004 → … |
Seminar
| Seminar | VarPhon (University of Potsdam) |
|---|---|
| Period | 7/01/04 → … |