Rhotic relationships: variationist and phonological perspectives on Dutch r variation

Research output: Contribution to conferencePaperOther research output


The cross-linguistic diversity of r-sounds is well-known (Lindau 1985; Ladefoged and Maddieson 1996), and a large amount of variation has also been reported within a number of languages, e.g. Spanish (Lipski 1994; Penny 2000), German (Ulbrich 1972; Schiller 1998), and Swedish (Muminovic and Engstrand 2002; Engstrand et al. 2007). Dutch is another example of a language with a particularly high number of r- variants. Even Standard Dutch permits a large amount of variation; drawing ever- finer boundaries, recent studies have distinguished between 5 (Voortman 1994), 6 (Booij 1995), 10 (Van de Velde 1996) or 24 (Smakman 2006) variants in Standard Dutch alone. This paper reports on the analysis of a large-scale corpus of urban colloquial Dutch (close to Standard Dutch), specifically compiled to study r-variation. It comprises of recordings of over 20,000 r-tokens from over 400 speakers from six Dutch and four Flemish cities, which were analysed perceptually and acoustically, revealing intricate layers of variation. The variation in place of articulation ranges from alveolar to palatal to uvular, manner of articulation from trills and fricatives to approximants and vowels. There is both between-speaker and within-speaker variation, and some individual speakers display almost the full range of variants. The variation is regulated by both linguistic and social factors, and in the first part of my talk I will give a brief overview of the patterns of variation that emerge from statistical analysis. A fuller overview is provided by Van de Velde and van Hout’s proposed talk for this conference.
While for the (socio)phonetician it is this diversity that is of the greatest interest, a phonologist might focus on the opposite, i.e. what is it that unites all these r-sounds? This question, which I consider in the second part of the talk, is especially important for the generative phonologist, who will want to assign a shared featural or structural representation to all rhotics, in order to explain why they are able to function in such a similar way across languages despite their phonetic diversity. Since there seem to be no phonetic features that are shared by all rhotics (Lindau 1985; Ladefoged and Maddieson 1996:244), phonologists have attempted to characterise rhotics in more abstract ways, by assigning them a shared structure (Walsh Dickey 1997), abstract feature [rhotic] (Hall 1997) or place on the sonority hierarchy (Wiese 2001). Such top-down approaches, however, are inevitably arbitrary and/or circular. Lindau’s (1985) solution (see Magnuson 2007 for an expanded version) is to characterise rhotic unity in terms of “family resemblance”. That is, while it is not true that every r- sound shares a phonetic property with every other r-sound, it is true that every r- sound shares something with at least one other: alveolar and uvular trills are both trills, uvular trills and uvular approximants are both uvular, uvular approximants and retroflex approximants are both approximants, etc. In all, this creates a network of speech sounds that resemble each other in at least a single aspect. However, while this is a lot less arbitrary than the top-down phonology approach, it is equally unrestrictive. There are many other speech sounds that share properties with some rhotics, and yet these do not function as rhotics themselves. Also, while the “family resemblance” approach may be enough for characterising the cross-linguistic unity of rhotics, a more stringent approach is needed to explain the variation r displays in Dutch: if several variants appear within the speech of a single speaker, a simple network of similarities cannot be enough explanation.
I propose to characterise r-variants not in terms of family resemblances, but family relationships. This makes a stronger claim: while resemblance among variants is established on the basis of common phonetic properties only, a relationship necessitates a diachronic link between two variants. Such links can be established on

the basis of age-graded and geographical variation in the data; and it is by inspecting very closely the phonetic detail of r-sounds and their linguistic distribution that we can in fact locate their origin in other variants. Specifically, this origin often lies in what happens to certain r-sounds in casual speech processes, such as lenition. For example, in casual speech, an alveolar trill may fail to be realised as such, given that it requires very precise articulatory settings in combination with narrow aerodynamic margins (Solé 2002). The result can be an alveolar approximant or fricative, which may consequently come to function as an r-variant itself, in the absence of social restrictions. This approach follows up on a tentative proposal from Barry (1997) and Schiller’s (1998) account for variants of German uvular r, and is in line with what appears to be Ladefoged and Maddieson’s (1996:245) interpretation of Lindau’s original suggestion.
The unity of r-variants, then, is expressed by tracing the diachronic relationships between them. This anchors it within a theory of sound change, but locates it outside the phonology. I suggest that this is where it should be, and that the top-down featural/structural approach to r is best replaced by a bottom-up one. Instead of requiring, a priori, that there be a category “r” (or /r/), such a category is formed by speakers on the basis of multi-level evidence, which may come from phonological alternations, phonetic identity and sociolinguistic variation. These various sources of evidence lead to rich phonological representation which itself includes variation, as e.g. in Exemplar-based models (Pierrehumbert 2001; Johnson 2006). For example, when speakers encounter two apparently phonetically unrelated phones, such as [ʁ] and [ɻ], in an allophonic relationship, as is the case for many Dutch speakers in our corpus (see also Scobbie and Sebregts 2010), they may set up representations at multiple levels in a hierarchical system (in the sense proposed by Ladd 2006), reflecting both their phonetic differences and their phonological identity. In other words, while the phonetic unity between such different r-sounds is one that can be traced diachronically (as a family relationship), their phonological unity (in representation) need not refer to this relationship, but is forged synchronically by speakers.
Original languageEnglish
Publication statusUnpublished - 29 Oct 2013
Event'r-atics4 (Grenoble) -
Duration: 31 Jan 2031 → …


Seminar'r-atics4 (Grenoble)
Period31/01/31 → …


Dive into the research topics of 'Rhotic relationships: variationist and phonological perspectives on Dutch r variation'. Together they form a unique fingerprint.

Cite this