Frequent violation of the sonority sequencing principle in hundreds of languages: how often and by which sequences?
-
Ruihua Yin
,
Jeroen van de Weijer
Abstract
The Sonority Sequencing Principle (SSP) is a fundamental governing principle of syllable structure; however, its details remain contested. This study aims to clarify the empirical status of the SSP in a cross-linguistic study of 496 languages. We adopt a phonetically-grounded definition of sonority – acoustic intensity – and examine how many languages contain SSP-violating clusters word-initially and word-finally. We consider the treatment of complex segments both as sonority units and as clusters. We find a significant proportion of languages violate the SSP: almost one half of the language sample. We examine which clusters cause the violations, and find a wide range: not only the notorious case of clusters with sibilants, but also with nasals, approximants and other obstruents. Violations in onsets and codas are not symmetrical, especially when complex segments are treated as units. We discuss where existing theoretical accounts of the SSP require further development to account for our crosslinguistic results.
1 Introduction
Ever since Sievers (1876) related the relative loudness of segments to their permitted linear arrangements within syllables and formulated this relation as the Sonority Sequencing Principle (SSP), the SSP has been a major explanatory tool in phonology. However, the SSP has been challenged by persistent disagreement over its definition, and a growing inventory of counterexamples reported from lesser-studied languages. In this paper, we re-visit the SSP in a study of nearly five hundred languages. By examining evidence for the SSP in a large cross-linguistic dataset, we aim to clarify the empirical status of the SSP – an essential step towards a more complete and theoretically satisfactory understanding of this fundamental principle of syllable structure (Clements 1990; Prince and Smolensky 1993/2004).
When we examine clusters, we adopt alternative analyses of complex segments (Round 2017b), to assess how findings about the SSP may depend upon assumptions about segmentation. Overall, we find that a significant proportion of languages (around two fifths to one half depending on assumptions made about complex segments) possess clusters that violate the SSP, that violations are due to a range of segment types, and that violations in onsets and codas are not symmetrical. The paper contains background in Section 2, materials and methods in Section 3, results in Section 4, discussion in Section 5 and conclusions in Section 6.
2 Background
Since the SSP was proposed, it has played a central role in phonological theory for its ability to account for the distribution of segments in the syllable, and relatedly to motivate phonological alternations that repair structures which otherwise would violate it (Blevins 1995; Clements 1990; Hooper 1972; Jespersen 1904/1913; McCarthy 2008; Parker 2002, 2008; Prince and Smolensky 1993/2004; Selkirk 1982, 1984; Sievers 1876; Smolensky 2006; Steriade 1982, 2002; Vennemann 1972, 1987; Zec 1995, 2007). According to the SSP, the most sonorous segment of a syllable is assigned to the nucleus position. Progressively less sonorous segments are assigned to progressively more marginal positions, moving out from the nucleus. For this procedure to be successful, the relative sonority of any two segments must be well defined. However, phonologists have disagreed extensively over the specifics of the sonority hierarchy, and thus over what constitutes an SSP violation, and therefore what is expected in the syllable structure of the world’s languages.
One influential sonority scale categorises segments into five sonority classes: vowels (highest sonority) > glides > liquids > nasals > obstruents (lowest sonority) (Clements 1990; Hooper 1972; Kenstowicz 1994; Smolensky 2006; van der Hulst 1984), but many studies diverge from this. The literature ranges greatly in the number of levels proposed within the sonority hierarchy, from the logical minimum of two sonority classes (1) to as many as seventeen (2).
| A sonority hierarchy with two sonority classes (Zec 2007): |
| sonorants > obstruents |
| A sonority hierarchy with seventeen classes (Parker 2008): |
| low vowels > medial vowels (except /ə/) > high vowels (except /ɨ/) > /ə/ > /ɨ/ > glides > rhotics > flaps > laterals > trills > nasals > voiced fricatives > voiced affricates > voiced stops > voiceless fricatives, /h/ > voiceless affricates > voiceless stops, /ʔ/ |
Disagreements also emerge concerning the relative rank of sonority classes, especially rhotics versus laterals; fricatives versus affricates versus plosives; and voiceless obstruents versus voiced ones. A few of these pivotal disagreements are listed below in (3).
| Scales differing in sonority of liquids, obstruents and their voicing: | |
| rhotics > laterals | (Hall 2002) |
| laterals > rhotics | (Hankamer and Aissen 1974) |
| fricatives = affricates > stops | (Nakajima et al. 2012) |
| fricatives > affricates > stops | (Orzechowska 2018) |
| vcl fric > vcd stops | (Gnanadesikan 1995b) |
| vcd stops > vcl fric | (Jespersen 1904/1913) |
| vcd aff > vcd stop > vcl fric | (Parker 2008) |
Sonority scales also differ in the placement of laryngeal segments: [h, ʔ] may be grouped with sonorants (Gnanadesikan 1995a); or obstruents (Orie and Bricker 2000); in between them (Brittain 2000); or be classed as both (Churma and Shi 1996).
One wellspring of disagreements about the sonority hierarchy is empirical in nature. Multiple studies, each examining a small set of languages, begin by assuming that the languages under investigation obey the SSP; empirically, however, the sets of languages pattern differently. The result is that researchers are led to contradictory conclusions about the SSP based on empirical differences in the languages studied. A second source of disagreement about the sonority hierarchy is theoretical, based on disagreements with respect to the definition of sonority. Although there is broad agreement that sonority relates to the loudness or perceptual prominence of a segment (Anderson 1986; Blevins 1995; Christman 1992; Clements 1990, 2009a, 2009b; Katamba 1989; Ladefoged 1975; Price 1980; Selkirk 1984), at least approximately, the details and the precise definition of sonority have failed to converge. Parker (2002) lists ninety-eight correlates of sonority that had been proposed by the start of the millennium.
In light of these obstacles, some have called for sonority theory to be abandoned altogether (Ball and Müller 2016; Harris 2006; Ohala and Kawasaki 1997), yet the majority view appears to be that sonority theory is still a viable concept, although it is not obvious how much explicit awareness there is about the depth of disagreement that it involves (see Parker 2012 for more discussion). One view, which we would share, is that the SSP cannot be rigidly exceptionless, and therefore, if it is to be retained as a major explanatory principle in phonology, then some idiosyncrasy must be admitted, alongside the generalisations that exist. Currently, however, owing to a history of applying different interpretations of sonority to different sets of languages, it is unclear just how much idiosyncrasy this view entails, and precisely what form it takes. Here we contribute to clarifying the empirical status of the SSP by conducting an examination of SSP violations, at a large cross-linguistic scale, using a uniform definition of sonority.
For the present study, we use one particular definition of sonority. Though it may be proven in future that we have not used the ‘right’ definition of sonority here, it will be impossible to even evaluate this question until studies such as the present one have been conducted, so for now, the most important desideratum is consistency across a large set of languages. As a second-order aim, we attempt to avoid overcommitment to any particular phonological theory, by using a widely-influential definition of sonority grounded in the phonetics of acoustic intensity, following Parker (2002). Acoustic intensity has been characterised as the closest acoustic correlate of sonority (Ladefoged and Johnson 2015; Parker 2002, 2008), and as the closest acoustic counterpart of perceptual loudness (Blevins 1995; Ladefoged 1975; Parker 2008). After testing five phonetic correlates – intensity, intraoral air pressure, F1 frequency, total air flow, and duration – Parker (2002) concludes that acoustic intensity has a near-perfect correlation with phonological sonority indices, and so proposes a universal sonority scale based on this property, given in (4).
|
Sonority scale when sonority is defined as acoustic intensity (Parker 2002) glide > rhotic > lateral > flap > trill > nasal > h > voiced fricative > voiced stop/affricate > voiceless fricative > voiceless stop/affricate/ʔ[1] |
To investigate the empirical status of the SSP, our method is to generate sonority contours from word-initial and word-final clusters by matching all consonants to their ordinal rank in the scale in (4), where glide = 11 and voiceless stop/affricate/ʔ = 1. SSP-violations were identified as sonority falls in word-initial clusters and sonority rises in word-final clusters.[2]
3 Materials and methods
We examine SSP violations in 496 languages from 58 language families. In this section we describe our data sources in Section 3.1, then discuss levels of analysis in Section 3.2, clusters are examined in Section 3.3, the treatment of complex segments in Section 3.4, and statistical procedure in Section 3.5.
3.1 Data sources
Our data come from two cross-linguistic databases of lexical citation forms: the Database of Cross-Linguistic Colexifications (CLICS2) (List et al. 2018) and AusPhon-Lexicon v.0-4 (Round 2017a). CLICS2 (List et al. 2018) aggregates and processes lexical data from a wide variety of database resources, in which consonant clusters are represented in IPA symbols in word lists. Not all CLICS2 datasets have IPA transcriptions. Here we use only those that do. CLICS2 has been used in cross-linguistic studies of polysemy (Rzymski et al. 2020), semantic shift (Zalizniak 2008), and lexical borrowings (List 2019), among others. AusPhon-Lexicon (Round 2017a) extends Bowern’s (2016) Chirila database, and contains phonologically normalised lexicons from Australian languages, in which IPA transcriptions are also phonemic. AusPhon-Lexicon has previously been used to study phoneme frequencies and phylogenetic signal in phonotactics and functional load (Macklin-Cordes et al. 2021; Macklin-Cordes and Round 2020; Round et al. 2022). From this phonemic data, we obtained the types of consonant clusters documented in each language. To study cross-linguistic permissible consonant clusters, we take the unique forms of consonant clusters (not tokens) in each language, e.g., /pr/ in /pra, pri, pro, pram/ is counted only once.
The language families in our study are shown in Table 1 (see the Supplementary Materials for a full itemisation of the languages.) Though the sample is large, it is geographically and genealogically patchy, due to the uneven availability of data in our database sources. Some gaps are significant: we have no data from Africa or North America, and this is a distinct limitation. Nevertheless, several factors motivated our study using this dataset. Firstly, notwithstanding its gaps, our sample extends well beyond most prior studies on sonority, which often have been heavily skewed towards Indo-European languages. Secondly, one of the contributions of this study is to the methodology of typologising using large digital datasets. In coming years, these datasets will only improve in coverage, and it is important for the field to be developing methods for their investigation now. Thirdly, our dataset makes possible not only the current study, but a series of follow-up studies which go into more detail (Yin 2021), and are largely possible only with the kind of digital dataset we use.
Language families, linguistic areas and database sources used in the study.
| Linguistic area & dataset | Families (languages) | Total |
|---|---|---|
| Australia: AusPhon-Lexicon | Bunaban (2), Eastern Daly (1), Garrwan (1), Giimbiyu (3), Gunwinyguan (9), Iwaidjan (3), Jarrakan (2), Laragia (2), Limilngan (1), Maran (4), Maningrida (3), Mirndi (3), Nyulnyulan (2), Pama-Nyungan (133), Southern Daly (1), Tangkic (1), Tiwi (2), Wadjiginy (1), Wageman (2), Western Daly (1), Worrorran (4), Yangmanic (1). | 182 |
| Eurasia: CLICS Allenbai, Beidasinitic, Northeuralex | Abkhaz-Adyge (2), Afro-Asiatic (2), Ainu (1), Basque (1), Burushaski (1), Chukotko-Kamchatkan (2), Dravidian (4), Indo-European (37), Japonic (1), Kartvelian (1), Koreanic (1), Mongolic-Khitan (3), Nakh-Daghestanian (6), Nivkh (1), Sino-Tibetan (28), Tungusic (3), Turkic (8), Uralic (26), Yeniseian (1), Yukaghir (2). | 131 |
| Papunesia: CLICS Robinsonap, Zgraggenmadang | Timor-Alor-Pantar (13), Nuclear Trans New Guinea (98). | 111 |
| South America: CLICS Hubercolumbian | Arawakan (9), Barbacoan (5), Boran (3), Camsá (2), Cariban (2), Chibchan (6), Chocoan (6), Eskimo-Aleut (3), Guahiboan (5), Huitotoan (4), Jodi-Saliban (1), Kakua-Nukak (2), Naduhup (1), Páez (1), Puinave (2), Quechuan (1), Tucanoan (19). | 72 |
| Total | 58 Families | 496 |
3.2 Level of analysis
Recent research in phonological typology has emphasised the importance of attending to levels of analysis and representation (Hyman 2008; Kiparsky 2018; Lass 1984; Round 2017b; van der Hulst 2017). The issue is relevant to the SSP and to our data.
Theories of phonological representation have changed repeatedly since the SSP’s first formulation in the nineteenth century, but the relevant level seems always to have been relatively superficial and contrastive, such as the phonemic level (in distinction to the morphophonemic or allophonic level) or the lexical representation (in distinction to underlying or post-lexical representation). At times, formal implementations have been more precise as to the level involved,[3] but notably, even when these proposals have disagreed about the level at which the SSP operates, they continue to agree on the general predictions of the SSP. The fact that phonologists have attributed essentially the same notion of the SSP to multiple levels of representation suggests that, conceptually, the notion of the SSP is somewhat ambiguous or ambivalent with respect to its appropriate level of representation.
A similar ambiguity characterises our data. Consider a language that predictably inserts a vowel V to break up a cluster C1C2. In our data, will a linguist have represented words in this language with C1C2 or C1VC2? If V is entirely predictable, some linguists will exclude it from the representation, but if V has similar phonetic properties to other contrastive vowels, it might be included instead. In this study, we rarely know whether, or how, such decisions were taken in the primary research that underlies our datasets; we see only the resulting strings like C1C2 or C1VC2. As it happens, the resulting ambiguity in the data matches the ambiguity present in the conceptualisation of the SSP itself: both refer to some kind of relatively superficial, phonemic-like level. Our data, comprised of phoneme level IPA transcriptions, ineluctably inherits this ambiguity about levels of representation. On the other hand, our assessment is that the data is appropriate to the concept under study. Moreover, our data is comparable to the kind of empirical evidence that has underpinned sonority theory for a century, so that our results can be informatively compared with prior research.
3.3 Word initial and final clusters of lexical citation forms
We limit our focus to word-initial and word-final consonant clusters. This is significant, because word peripheral consonant clusters are known to differ from those in word-internal positions, with word margins supporting more phonological and morphological complexity (Clements 1990; Dressler and Dziubalska-Kolaczyk 2006; Easterday 2019; Hayes 1982).[4] We make this methodological choice because in practical terms, our large-scale data does not allow us to observe syllable boundaries in word-medial consonant clusters. In the case of /…VCCCV…/ for example, the parsing of the medial cluster as /C.CC/, /CC.C/ or /.CCC/ is not directly stated in the database. Moreover, any extant parsing of these word-medial clusters into onsets or codas may be based in a circular fashion on assumptions about the SSP, as pointed out by Ohala and Kawasaki (1997). An investigation of medial cluster syllabification is clearly necessary but must be left for further study.
Studying clusters at word edges also has inherent limitations. Some clusters only occur at word-edges but not word-medial syllable edges, e.g., sC in many Indo-European languages. By studying only word edges, we do not uncover such details. Word-edges can host clusters that are hetero-morphemic, which in some languages may tolerate more SSP-reversals than homo-morphemic clusters (Dressler and Dziubalska-Kolaczyk 2006). In CLICS, morphemic boundaries are not systematically coded for some languages: for example, ‘night’ /nakt-s/ ‘night-nom.sg’ in Latvian (Glottocode latv1249) is simply represented as /nakts/. In our study, consonant clusters are comparable across languages insofar as they are all found in citation forms, but they may differ from language to language in terms of their morphological complexity. Thus, morphological effects on sonority sequencing cannot be directly evaluated from the current study.
3.4 Factorial analysis of complex segments
Typological research has perennially faced choices over its units of analysis (Round and Corbett 2020). In this respect, sonority theory is complicated by the issue of whether complex segments, which usually comprise multiple subparts (Shih and Inkelas 2018; Steriade 1993) constitute a single segment, and thus a single sonority value, or segment sequences and thus a sonority contour. The adoption of one answer as opposed to the other could potentially alter one’s conclusions about the typological prevalence of SSP violations. Moreover, it is not clear that analyses in primary sources should automatically be taken at face value. Round (2022) shows that comparable language facts have been accorded divergent analyses, as complex segments versus clusters, in Australian languages, and the same is likely true elsewhere in the world (Round and Macklin-Cordes 2015).
In cases like this, where a phenomenon under typological investigation is regularly accorded any one of several competing analyses by linguists, Round (2017b) proposes a method termed Factorial Analysis, in which each of the contending analyses of a dataset is adopted and examined in turn. Here we apply Factorial Analysis to the analysis of complex segments, by examining the consequences for our cross-linguistic results of making such multiple possible assumptions. For practicality’s sake, we apply this only to two of the most canonical complex segment types, affricates and homorganic nasal-stops. First, we assume that affricates and homorganic nasal-stop sequences are clusters and thus sonority contours; then we assume they are single segments, and we compare the results.[5] By doing so, we reveal how the typology of SSP violations changes as analytic assumptions change regarding the status of complex segments.
3.5 Statistical procedure
Our statistical analysis centres around the proportions of languages in the sample that have certain kinds of SSP violations. The language sample in this study is not a classical balanced sample (Dryer 1989), and thus some measure needs to be taken to account for the statistical non-independence of the languages. Macklin-Cordes and Round (2022) provide methods enabling the genealogical structure of related languages to be accounted for when calculating proportions in typological studies without requiring balanced sampling. The result is a proportion which is ‘genealogically sensitive’, and it is this kind of proportion that we report below. Calculations were made using the R packages glottoTrees and phyloWeights (Round 2021a, 2021b). Details appear in the Supplementary Materials, as do tables reporting more traditional proportions, which are more susceptible to distortion from language relatedness.
3.6 Under-representation of SSP violations in small samples
Our dataset potentially under-represents the prevalence of SSP violations, since a language could possess SSP-violating clusters which fail to appear in any of the words in a wordlist, simply by chance. To evaluate the likely strength of this effect, we applied a Monte Carlo test, taking the 100 languages with the longest wordlists in the dataset, and from these sampled a smaller number of words, matching the wordlist lengths of much shorter lists in the dataset and then checking if the shorter samples missed any satellite[6] types found in the full lists. For each language, the sampling procedure was repeated 1,000 times. We estimate that in the shortest 25% of our wordlists, a given satellite type in the onset or coda will be missing by chance from around one in 25 wordlists.[7] In the next-shortest 25%, this figure decreases to one in 35, and in the next-shortest 25%, it decreases to one in 80. This indicates that our results almost surely do understate the prevalence of SSP violations, although not at a frequency that will alter our main findings. More details on the Monte Carlo results appear in the Supplementary Materials.
4 Results
Here we present findings on the proportions of languages that exhibit SSP violations in Section 4.1, and then, among the languages which do exhibit SSP-violations, the proportions of these that have clusters that contain sibilants, obstruents, nasals or approximants (i.e., SSP violators) in Sections 4.2 and 4.3. Findings are stated in term of raw counts of languages, and genealogically sensitive proportions (similar to regular proportions, but weighted to account for language relatedness, cf. Section 3.5; for regular proportions, which is the number of languages in question divided by the total number of languages see Table 4 in the Supplementary Materials). We report all findings using both sets of assumptions about complex segments: that affricates and prenasalised stops are always a single unit (dubbed the ‘merged’ method) or always clusters (the ‘split’ method). The results show that languages with SSP-violations are common, occurring in around two fifths to one half of the languages in our sample, depending on the assumptions made on complex segments. Moreover, SSP violators take various shapes. Sibilants and nasals are the main causes of SSP violations in onsets, while approximants, nasal and sibilants are the main causes in codas.
4.1 Languages with SSP violations are common
We begin by examining whether all 496 languages exhibit SSP violations in their onsets (i.e., word-initial clusters), codas (i.e., word-final clusters), or in either or in both. These results appear in Table 2.
Numbers and proportions (genealogically sensitive) of languages with SSP violations.
| Total languages: 496 | Split method (ts, mb = clusters) | Merged method (ts, mb = single units) | Difference (split – merged) | |||
|---|---|---|---|---|---|---|
| N | Proportiona | N | Proportion | N | Proportion | |
| Have onset violations | 177 | 39.4% | 132 | 32.1% | 45 | 7.3% |
| Have coda violations | 145 | 37.4% | 121 | 29.6% | 24 | 7.8% |
| Have either | 224 | 51.7% | 178 | 43.1% | 46 | 8.6% |
| Have both | 98 | 25.1% | 75 | 18.6% | 23 | 6.5% |
-
aFor regular proportions, which is the number of languages in question divided by the total number of languages, see Table 4 in the Supplementary Materials.
Table 2 arranges results by the split method, merged method and the difference between them. The differences convey how the cross-linguistic picture of SSP violations changes when we change assumptions about the analysis of complex segments. Numbers for the merged method are necessarily lower than for the split method. This is because sequences like /ts/ or /mb/ count as clusters in the split method, and therefore can add to the tally of SSP violations (/ts/ in codas and /mb/ in onsets) whereas in the merged method they count as single segments and so cannot, in and of themselves, contribute to SSP violations.
Around 39% of languages (split method, weighted proportions) have violations in onsets, around 37% have violations in codas, and over 50% have violations in at least one of these, indicating that SSP violations are strikingly common cross-linguistically. (Nor are they limited to certain geographical areas or language families. For full lists of languages corresponding to the results here and throughout Section 4, see the Supplementary Materials). Additionally, we find a pronounced difference between the split and merged methods. The ‘difference’ columns in Table 2 indicate that around one fifth of all violations hinge upon one’s analysis of complex segments.
4.2 SSP-violating consonant clusters
Next, we investigate which kinds of consonant clusters cause SSP violations. In an onset or coda with a violation, there will by definition be some consonant which has a sonority rank that exceeds the rank of its neighbour closer to the nucleus. For instance, /l/ in /lba/ has higher sonority than /b/, and /s/ in /aks/ has higher sonority than /k/. We refer to these consonants, which constitute a local peak in sonority, as sonority satellites (connoting that although they are sonority peaks, they are not in the nucleus). It is often supposed that the main challenge to the SSP is posed by satellites that are sibilants, especially in Indo-European languages. However, the findings below demonstrate that once we cast a wider net, sonority satellites extend well beyond just sibilants. We examine sibilant satellites and other obstruent satellites in Section 4.2.1, and nasal and approximant (including liquid) satellites in Section 4.2.2. Sections 4.2.1 and 4.2.2 refer to results displayed in Table 3.
Numbers of different satellites in languages with SSP violations, given the scale glide > rhotic > lateral > flap > trill > nasal > h > vcd fric > vcd plo/aff > vcl fric > vcl plo/aff (Parker 2002).
| Split method (t s, m b = clusters) | Merged method (ts, mb = single units) | |||
|---|---|---|---|---|
| N | Proportiona | N | Proportion | |
| Onset | 177 | 132 | ||
| Sibilant satellites b | 97 | 65.2% | 95 | 76.9% |
| vcl sib + vcl plo | ||||
| vcd sib + [vcd plo, vcl fric, vcl plo] | ||||
| Non-sibilant obstruent satellites | 40 | 33.7% | 39 | 40.6% |
| vcl fricc + vcl plo | ||||
| vcd plo + [vcl fric, vcl plo] | ||||
| vcd fric + [vcd plo, vcl fric, vcl plo] | ||||
| h + [vcd fric, vcd plo, vcl fric, vcl plo] | ||||
| Nasal satellites | 88 | 44.6% | 29 | 26.6% |
| nas + obs | ||||
| Approximant satellites | 54 | 34.3% | 54 | 42.1% |
| trill + [nas, …, vcl plo] | ||||
| flap + [trill, …, vcl plo] | ||||
| lat + [flap, …, vcl plo] | ||||
| rho + [lat, …, vcl plo] | ||||
| gl + [rho, …, vcl plo] | ||||
| Coda d | 145 | 121 | ||
| Sibilant satellites | 98 | 76.5% | 37 | 40.5% |
| Non-sibilant obstruent satellites | 29 | 24.1% | 26 | 27.7% |
| Nasal satellites | 56 | 35.3% | 56 | 44.6% |
| Approximant satellites | 80 | 53.7% | 80 | 67.9% |
-
aFor raw proportions, the number of languages in question divided by the total 496; see Table 4 in the Supplementary Materials. bIn this table, languages with sibilant satellites refer to languages with any of clusters listed below, i.e., languages with vcl sib + vcl plo, or with vcd sib + [vcd plo, non-sibilant vcl fric, vcl plo]. For the merged method, vcl and vcd plo also contain vcl and vcd affricates. cVcl fric here refers to fricatives other than sibilants. dClusters falling in each category of satellite in the coda position are the mirror image of clusters listed in the onset position.
4.2.1 Sibilant and non-sibilant obstruent satellites as SSP-violators
Clusters containing sibilant satellites present a notorious problem for the SSP. Sibilants often stand further away from the nucleus than less sonorous plosives, an observation which has motivated many studies to assign sibilants special status, in terms of phonetic features (Liljencrants and Lindblom 1972; Maniwa et al. 2009) and/or formal phonological explanations (Blevins 1995; Steriade 1982). What’s more, much of the variety in sonority scales is caused by variation in the fine-grained sonority distinctions proposed among obstruents, especially issues around the ranking of vcl/vcd fricatives, vcl/vcd affricates, and vcl/vcd plosives. Among languages which violate the SSP, this section examines the proportions of these that permit sibilant or non-sibilant obstruent satellites.
To be a satellite, a sibilant must be more sonorous than its neighbour closer to the nucleus. In the current study, given the sonority scale in (4) above, this neighbour must be a voiceless plosive or affricate (if the sibilant is voiceless) or any plosive or voiceless fricative (if the sibilant is voiced). To be a satellite, a non-sibilant obstruent must be more sonorous than its neighbour closer to the nucleus, i.e., the neighbour must be a voiceless plosive or affricate (if the satellite is a voiceless fricative) or any voiceless obstruent (if the satellite is a voiced plosive or affricate), or any voiceless obstruent or voiced plosive/affricate (if the satellite is a voiced fricative).[8]
The results confirm that, except in codas under the merged method, if a language has SSP violations, it is more likely than not to have sibilant satellites. The figures for codas in the merged method are dramatically lower because the merged method counts sequences like /ts/ as single units, not as plosive-sibilant clusters. The result shows how starkly one’s assumptions about complex segments in terms of plosive-sibilant as clusters versus complex segments can affect conclusions about how the cross-linguistic frequencies of SSP violations are due to sibilant satellites. We find that non-sibilant obstruent satellites are responsible for violations in around 24%–41% of languages that have SSP violations. Thus, we find, despite the plethora of sonority scales that have been proposed in order to finesse the sonority ranking of obstruents, that such obstruents are responsible for relatively few violations cross-linguistically.
In sum, sibilant satellites, which are assumed to be a primary cause of SSP violations, are indeed found in a great proportion of languages with violations.
4.2.2 Nasal and approximant satellites as SSP-violators
Another common type of SSP violator is nasal-obstruent clusters, as observed in Pacific languages like Fijian (fiji1243) or Fula (Arnott 1970; Wiswall 1989), in Australian languages like Arrernte (east2379), Alawa (alaw1244), Yidiny (yidi1250) etc. (Dixon 1977; Evans 1995), and in African languages, such as /n-pene/ ‘class9-goat’ in Kilega (Niger-Congo) (Tak 2011); although whether nasal-stops are sequences or segments has been a controversial issue (see Anderson 1976; Herbert 1975, 1977; Hyman 1992; Ladefoged and Maddieson 1996; Maddieson 1989; Round 2022; van de Weijer 1996 for more discussion). Clusters containing approximant satellites have been documented in some Slavic languages, such as /lb-/ or /rt-/ in Russian, /jd- js-/ in Czech (czec1258), as well as in languages like Pashto (Eastern Iranian, cent1973) (/wr- wl-/) or Chatino (an Oto-Manguean language, Mexico) (/wn- yn-/) (data from Greenberg 1978). Although approximant satellites are reported in some languages, they are not as well studied, in terms of being accorded any special formal-phonological status, as more recognised satellites such as sibilants.
To be a satellite, a nasal must be more sonorous than its neighbour closer to the nucleus. Given the position of nasals just above the obstruents in the sonority hierarchy in (4), this means the neighbour must be an obstruent. To be a satellite, an approximant must be more sonorous than its neighbour closer to the nucleus which should be a nasal or obstruent (if the satellite is a trill); a nasal, obstruent or trill (if the satellite is a flap); a nasal, obstruent, trill or lateral (if the satellite is a rhotic); or any consonant at all (if the satellite is a glide).
In onsets among languages that have SSP violations, the results show that nasal satellites appear in close to half of the sample languages (split method). This proportion drops greatly, to around one quarter, if homorganic nasal+stop is treated as a single segment (merged method). This indicates, once again, how assumptions about complex segments can significantly affect conclusions about how often the SSP is violated. In codas, nasal satellites (i.e., stop+nasal sequences) are also rather common, appearing in 56 languages. In the split method, where the total number of languages with SSP violations is 145, this equates to around 35%, and in the merged method, where the total number of languages with SSP violations is 121, it equates to around 45%. In sum, nasal satellites appear in a substantial proportion of languages that exhibit SSP violations. This is true not only in onsets but also in codas.
The count of approximant satellites is unaffected by the choice of our split versus merged methods, which differ only in their treatment of affricates and nasal+stop sequences, neither of which contain approximants. Thus, we find approximant satellites in the onsets of 54 languages and in the codas of 80, irrespective of method. In onsets, these 54 languages equate to around one third of all languages with SSP violations. In codas, the 80 languages equate to closer to two thirds. These are quite substantial contributions for a satellite type which has received relatively little attention in the phonological literature.
4.3 Comparison of satellites by type and position
The summary of results from above is presented graphically in Figure 1 below, which compares the proportions of languages with each type of SSP-violators according to their position (onset/coda) and analysis method (merged/split). Comparing proportions for split versus merged methods, we find they are identical or nearly so with the notable exceptions of nasal satellites in onsets (due to the different treatment of homorganic nasal+stop) and sibilant satellites in codas (due to the different treatment of homorganic stop+fricative). From this, it follows that most of the differences in the proportions, between the two methods, are due solely to the differences in the total numbers of languages with satellites: for instance, the 80 languages with approximant satellites in codas comprise only 54% of the total of 145 languages (split method), but 68% of the total of 121 languages (merged method). These language totals, which cause the difference in proportions, are themselves due almost entirely to the differences in the counts of onset nasal satellites and coda sibilant satellites. Consequently, whenever we see a difference in the proportions in the split versus merged method, it should be recalled that the function of those differences is to reveal the consequences of choices in the analysis of complex segments—a matter which is clearly of significance, though seldom front and centre of debates about sonority hierarchies and the SSP.
Genealogically-weighted proportion of different kinds of SSP-violators in languages with SSP violations.
Comparing proportions in onsets, for both methods, sibilant satellites are the most common, occurring in 65%/77% (split/merged) of all languages with onset SSP violations. Nasal satellites are also very common, occurring in 45% (split). This figure drops to 27% in the merged method, though interestingly it does not drop to anywhere near zero. Other types of satellite are less common, though both approximant satellites and non-sibilant obstruent satellites appear in around one third of languages (split). These results reveal that sonority satellites are diverse. Consequently, even if a theory of sonority attempted to exclude certain classes of sonority satellite from consideration—e.g., by proposing that sibilants, or nasals, are exceptional in some fashion, or by proposing the existence of complex segments so as to remove some clusters from consideration—this would still leave very many languages with onset violations of the SSP.
Comparing proportions in codas, sibilants are the most common, occurring in 77% of all languages with coda SSP violations, but only in the split method which treats all affricates as clusters. Once affricates are analysed as single units, the dominance of sibilants disappears, slipping to just 41% of languages. Interestingly, it is approximants (including liquids) that are the next most dominant, appearing as satellites in 54% of languages (split) or 68% (merged). Nasal satellites are also common, in 35%/45% (split/merged). Non-sibilant obstruent satellites occur more rarely, though still in around one quarter of languages with SSP coda violations.
Comparing onsets against codas, a notable difference is the frequency of approximant satellites. These are rare in onsets but strikingly common in codas. Otherwise, in the split method, onsets and codas are broadly comparable: sibilant satellites are more frequent than nasal satellites, which in turn are more frequent than non-sibilant obstruent satellites in both onsets and codas. In the merged method, the treatment of affricates and nasal+stops as units results in a much more asymmetrical picture: sibilants dominate nasals in onsets but are on par in codas, while non-sibilant obstruent satellites outnumber nasals in onsets but not in codas. One interpretation of this finding is that if we expect onsets and codas to mirror each other in terms of the rates at which satellites of various kinds appear, then those expectations are not met if we treat affricates and homorganic nasal+stop as single units; if we treat them as clusters then those expectations are much more nearly met (though approximant satellites are still an exception).
5 Discussion
In this cross-linguistic study, we have investigated SSP violations in 496 languages from 58 language families. The language sample is not a classical balanced sample, and for this reason we have expressed our results in terms of genealogically sensitive proportions. We find that a substantial proportion of languages exhibit SSP violations in onsets (39% split, 32% merged), codas (37% split, 30% merged), and in either margin (52% split, 43% merged). The sonority sequences responsible for these violations are not restricted to clusters that contain sibilants, but instead come from a range of sonority sequencing types. Moreover, we have confirmed that choices in the analysis of complex segments have a pronounced impact on conclusions reached about the SSP and its violators.
How is this commonness of SSP violations to be explained? An immediate response might be to appeal to perceptual-grounding, rather than the SSP, as a source of explanation of segment sequencing. Sequences with strong perceptual cues are considered to be more likely to survive (Flemming 2004, 2005; Harris 2006; Lindblom 1983; Ohala and Kawasaki 1997; Wright 1996, 2001, 2004), which has also been captured in formal studies such as Côté (2000), Goldsmith (1990), Itô (1986), Jun (1995), among others. Acoustically, when individual segments have strong internal phonetic cues (e.g., in the parameters of F0, amplitude, periodicity and spectral shape), then segment sequences characterised by a large inter-segment modulation of spectral shapes are more likely to survive; and when strong cues exist external to segments, i.e., in the transition between them, this also promotes the survival of the sequence (Kawasaki and Ohala 1980; Ohala and Kawasaki 1997; Wright 2004). Under these accounts, sC clusters are common due to the robustness of their internal cues for place and manner of articulation for sibilants, which enable them to enjoy a freer distribution compared to other segments, and Cr clusters are common due to their strong transitional cues presented by the large magnitude of its acoustic modulation or long transitional trajectory. However, in the current data sample, clusters other than the obvious sC and Cr are also common. Heterorganic onset nasal-stops, coda stop-nasals, as well as obstruent clusters other than sC will require further investigations to establish whether they also admit of explanations based in perception. Furthermore, there are asymmetries between onsets and codas in terms of the common sequences they permit which would also require an account. Achieving a perceptually-based explanation of segment sequencing still requires convincing, precise accounts of much more than just the familiar example of sC and Cr.
Turning our attention to the SSP itself, does the commonness of SSP-violations present a genuine challenge to the SSP as a fundamental governing principle of syllable structure? Syllables are regarded as hierarchical in nature, and the SSP is proposed to govern the syllabic level or levels below it such as onsets, rhymes, or moras (Kaye 1992; van der Hulst 1984, 1994; Zec 1994, 1995). Apparent exceptions to the SSP—typically sibilants—are often assigned an exceptional prosodic status, such as extra-syllabicity or extra-prosodicity (Clements and Keyser 1983; Durand 1990; Goad 2011, 2012, 2016; Goldsmith 1976, 1990; Kiparsky 2003; Rialland 1994; Scheer 2004; Scheer and Cyran 2017; Selkirk 1984; Steriade 1982, 2001a, 2001b; van der Hulst 1984, 1994, 2020; Vaux and Wolfe 2009). Assigning this kind of exceptional status to ‘problematic’ segments in terms of the SSP is thus a common strategy within formal phonology for keeping the SSP intact in the face of apparent counterexamples. In the current study, we find that SSP-violators are not limited to sibilants. Nearly one third of onset violations (or one quarter if affricates or homorganic nasal-stop sequences are treated as one complex segment) are due to clusters that contain approximants, or obstruent clusters that do not contain any sibilants, while clusters that contain nasals are also responsible for around one third (heterorganic nasal-stop sequences) or one half (both homorganic and heterorganic nasal-stops) of violations. In the coda position, dominant SSP-violators, besides sibilant clusters, are approximant clusters, responsible for coda violations in nearly one half to 70 percent of languages that have coda violations; followed by nasal clusters, responsible in over one third to nearly one half of languages; and even non-sibilant obstruent clusters are responsible for coda violations in nearly one third of languages that have coda violations. Assigning a segment which is problematic for sonority sequencing to a structural position outside the syllable, or to a representational level which the SSP does not govern, is capable of handling apparent SSP violations per se; however, it is not clear that these strategies will offer a principled account for the commonness of violations, or for the full variety of SSP-violators that we have found, or for the asymmetric frequencies of them in onsets and codas. For instance, within the prosodic hierarchy, if sibilants are accorded extrasyllabic status to account for their apparent violations of the SSP, should this analysis be extended to all kinds of satellite found in the current study? If not, on what grounds? Or if so, how should we reflect the unequal frequencies of various satellites cross-linguistically, and the asymmetries between onsets and codas, if all cases are explained in terms of just one structural configuration? It should be emphasised that the assignment of extra-syllabic status to sibilants does not always depend solely on phonotactics: additional, language-specific phonological evidence such as reduplication (Steriade 1982; van de Weijer 1996), deletion (Côté 2004), epenthesis (Haddad 1984), and assimilation (Cser 2012) can also provide such motivation. Until such additional phonological evidence, beyond phonotactics, is found the various types of SSP-violators found in the current study, these violators do appear to present a standing prima facie challenge to the SSP and its role as a governing principle of syllable structure in the world’s languages.
Nonetheless, to return to the view we introduced earlier, the SSP clearly admits exceptions. Previously, the nature and extent of those exceptions has been unclear, and consequently theorists have not been pressed to account for them. The present paper has focussed on the exceptions to the SSP, but a vast weight of empirical data, which we have not focused on here, is still well accounted for by the SSP (Yin 2021). What remains is to piece together the exceptions and the generalisations. This can only be done successfully if the exceptions are known. Here we have attempted to progress along the path to a better sonority theory by clarifying the exceptions.
6 Conclusions
The SSP is widely regarded to be a fundamental principle governing syllable structures, yet even after a century the details of sonority theory have remained contested and unresolved. We have subjected the SSP to a large, cross-linguistic examination, asking in how many languages the SSP is violated when a phonetic operationalisation of sonority in terms of acoustic energy is employed. Our foremost finding is that SSP-violations are common, and SSP violators are diverse and are asymmetrical in onsets and codas. We situated existing theoretical attempts to accommodate apparent SSP violations within the updated context of our findings, and have found them broadly in need of additional theoretical justification if they are to accommodate the true diversity of sonority sequencing in the world’s languages.
Funding source: Australian Government Research Training Program Scholarship
Funding source: Ministry of Education, Guangdong Province
Award Identifier / Grant number: 2019WCXTD010
Funding source: British Academy
Award Identifier / Grant number: GP300169
Acknowledgments
A version of this research was presented at the 2020 Conference of the Australian Linguistic Society. Our thanks to the audience for valuable discussion and suggestions. Support for this research is gratefully acknowledged from an Australian Government Research Training Program Scholarship to RY; a Ministry of Education, Guangdong Province, grant number 2019WCXTD010 to JvdW; and a British Academy Global Professorship GP300169 to ER.
-
Data availability statement: Data supporting the results reported here can be found in the supplementary materials and archived at https://zenodo.org/record/6792036.
-
Research funding: This work was supported by Australian Government Research Training Program Scholarship; Ministry of Education, Guangdong Province (2019WCXTD010); and British Academy (GP300169).
References
Anderson, John. 1986. Suprasegmental dependencies. In Jacques Durand (ed.), Dependency and non-linear phonology, 55–133. London: Croom Helm.10.4324/9780429452918-4Search in Google Scholar
Anderson, Stephen. 1976. Nasal consonants and the internal structure of segments. Language 52(2). 326–344. https://doi.org/10.2307/412563.Search in Google Scholar
Arnott, David. 1970. The nominal and verbal system of Fula. Oxford: Clarendon Press.Search in Google Scholar
Ball, Martin & Nicole Müller. 2016. Challenging sonority: Cross-linguistic evidence. Sheffield: Equinox.Search in Google Scholar
Blevins, Juliette. 1995. The syllable in phonological theory. In John Goldsmith (ed.), The handbook of phonological theory, 206–244. New Jersey: Blackwell.10.1111/b.9780631201267.1996.00008.xSearch in Google Scholar
Bowern, Claire. 2016. Chirila: Contemporary and historical resources for the indigenous languages of Australia. Language Documentation & Conservation 10. 1–44.Search in Google Scholar
Brittain, Julie. 2000. A metrical analysis of primary stress placement in Southern East Cree. International Journal of American Linguistics 66(2). 181–217. https://doi.org/10.1086/466417.Search in Google Scholar
Christman, Sarah. 1992. Uncovering phonological regularity in neologisms: Contributions of sonority theory. Clinical Linguistics and Phonetics 6(3). 219–247. https://doi.org/10.3109/02699209208985532.Search in Google Scholar
Churma, Donald & Yili Shi. 1996. Glottal consonants and the sonority hierarchy. Paper presented at the Proceedings of the Eastern States Conference on Linguistics, 1995, Ithaca, NY.Search in Google Scholar
Clements, George. 1990. The role of the sonority cycle in core syllabification. In John Kingston & Mary Beckman (eds.), Papers in laboratory phonology I: Between the grammar and physics of speech, 283–333. Cambridge: Cambridge University Press.10.1017/CBO9780511627736.017Search in Google Scholar
Clements, George. 2009a. Does sonority have a phonetic basis? Comments on the chapter by Vaux. In Eric Raimy & Charles Cairns (eds.), Contemporary views on architecture and representations in phonological theory, 165–175. Cambridge: MIT Press.10.7551/mitpress/9780262182706.003.0007Search in Google Scholar
Clements, George. 2009b. The role of features in phonological inventories. In Eric Raimy & Charles Cairns (eds.), Contemporary views on architecture and representations in phonology, 19–68. Cambridge: MIT Press.10.7551/mitpress/7997.003.0005Search in Google Scholar
Clements, George & Jay Keyser. 1983. CV phonology: A generative theory of the syllable. Cambridge: MIT Press.Search in Google Scholar
Côté, Marie-Hélène. 2000. Consonant cluster phonotactics: A perceptual approach. Cambridge, MA: Massachusetts Institute of Technology dissertation.Search in Google Scholar
Côté, Marie-Hélène. 2004. On perception and sonority in cluster simplification: Consonant deletion in Icelandic preterites. Journal of German Linguistics 16(3). 203–204. https://doi.org/10.1017/s147054270400039x.Search in Google Scholar
Cser, András. 2012. The role of sonority in the phonology of Latin. In Steven Parker (ed.), The sonority controversy, 39–64. Berlin: De Gruyter Mouton.10.1515/9783110261523.39Search in Google Scholar
Dixon, R. M. W. 1977. A grammar of Yidiny. Cambridge: Cambridge University Press.Search in Google Scholar
Dressler, Wolfgang & Katarzyna Dziubalska-Kolaczyk. 2006. Proposing morphonotactis. Italian Journal of Linguistics 18(2). 249.Search in Google Scholar
Dryer, Matthew. 1989. Large linguistic areas and language sampling. Studies in Language 13(2). 257–292. https://doi.org/10.1075/sl.13.2.03dry.Search in Google Scholar
Durand, Jacques. 1990. Generative and non-linear phonology. London: Routledge.Search in Google Scholar
Easterday, Shelece. 2019. Highly complex syllable structure: A typological and diachronic study. Berlin: Language Science Press.Search in Google Scholar
Evans, Nicholas. 1995. Current issues in the phonology of Australian languages. In John Goldsmith (ed.), The handbook of phonological theory, 723–761. Cambridge, MA: Blackwell.Search in Google Scholar
Flemming, Edward. 2004. Contrast and perceptual distinctiveness. In Bruce Hayes, Robert Kirchner & Donca Steriade (eds.), Phonetically based phonology, 232–276. Cambridge: Cambridge University Press.10.1017/CBO9780511486401.008Search in Google Scholar
Flemming, Edward. 2005. Speech perception and phonological contrast. In David Pisoni & Robert Remez (eds.), The Handbook of speech perception, 156–181. Malden, MA: Blackwell.10.1002/9780470757024.ch7Search in Google Scholar
Gnanadesikan, Amalia. 1995a. Deriving the sonority hierarchy from ternary scales. Paper presented at the Linguistic Society of America, New Orleans.Search in Google Scholar
Gnanadesikan, Amalia. 1995b. Markedness and faithfulness constraints in child phonology. In René Kager, Joe Pater & Wim Zonneveld (eds.), Constraints in phonological acquisition, 73–108. Cambridge: Cambridge University Press.10.1017/CBO9780511486418.004Search in Google Scholar
Goad, Heather. 2011. The representation of sC clusters. In Marc van Oostendorp, Colin Ewen, Elizabeth Hume & Keren Rice (eds.), The Blackwell companion to phonology, 898–923. Oxford: Wiley.10.1002/9781444335262.wbctp0038Search in Google Scholar
Goad, Heather. 2012. /sC/ clusters are (almost always) coda-initial. The Linguistic Review 29(3). 335–373. https://doi.org/10.1515/tlr-2012-0013.Search in Google Scholar
Goad, Heather. 2016. Sonority and the unusual behavior of /s/. In Martin Ball & Nicole Müller (eds.), Challenging sonority: Cross-linguistic evidence, 21–44. Sheffield: Equinox.Search in Google Scholar
Goldsmith, John. 1976. Autosegmental phonology. Cambridge, MA: Massachusetts Institute of Technology dissertation.Search in Google Scholar
Goldsmith, John. 1990. Autosegmental and metrical phonology. New Jersey: Wiley.Search in Google Scholar
Greenberg, Joseph. 1978. Some generalizations concerning initial and final consonant clusters. In Joseph H. Greenberg, Charles A. Ferguson & Edith A. Moravcsik (eds.), Universals of human language, 243–279. Stanford: Stanford University Press.Search in Google Scholar
Haddad, Ghassan. 1984. Epenthesis and sonority in Lebanese Arabic. Studies in the Linguistic Sciences 14(1). 57–88.Search in Google Scholar
Hall, Tracy. 2002. Against extrasyllabic consonants in German and English. Phonology 19(1).33–75.10.1017/S0952675702004268Search in Google Scholar
Hankamer, Jorge & Judith Aissen. 1974. The sonority hierarchy. In Bruck Anthony, Robert Fox & Michael Galy (eds.), Papers from the Parasession on natural phonology. Chicago: Chicago Linguistic Society.Search in Google Scholar
Harris, John. 2006. The phonology of being understood: Further arguments against sonority. Lingua 116(10). 1483–1494. https://doi.org/10.1016/j.lingua.2005.07.009.Search in Google Scholar
Hayes, Bruce. 1982. Extrametricality and English stress. Linguistic Inquiry 13(2). 227–276.Search in Google Scholar
Herbert, Robert. 1975. Reanalyzing prenasalized consonants. Studies in African Linguistics 6(2). 105–123.Search in Google Scholar
Herbert, Robert. 1977. Phonetic analysis in phonological description: Prenasalized consonants and Meinhof’s Rule. Lingua 43(4). 339–373. https://doi.org/10.1016/0024-3841(77)90111-5.Search in Google Scholar
Hooper, Joan. 1972. The syllable in phonological theory. Language 48(3). 525–540. https://doi.org/10.2307/412031.Search in Google Scholar
van der Hulst, Harry. 1984. Syllable structure and stress in Dutch. Dordrecht: Foris.10.1515/9783112419984Search in Google Scholar
van der Hulst, Harry. 1994. Radical CV phonology: The locational gesture. UCL Working Papers in Linguistics 6. 439–478.Search in Google Scholar
van der Hulst, Harry. 2017. Phonological typology. In Alexandra Aikhenvald & Robert M. W. Dixon (eds.), The Cambridge handbook of linguistic typology, 39–77. Cambridge: Cambridge University Press.10.1017/9781316135716.002Search in Google Scholar
van der Hulst, Harry. 2020. Principles of radical CV phonology: A theory of segmental and syllabic structure. Edinburgh: Edinburgh University Press.10.3366/edinburgh/9781474454667.001.0001Search in Google Scholar
Hyman, Larry. 1992. Moraic mismatches in Bantu. Phonology 9(2). 255–265. https://doi.org/10.1017/s0952675700001603.Search in Google Scholar
Hyman, Larry. 2008. Universals in phonology. The Linguistic Review 25(1–2). 83–137. https://doi.org/10.1515/tlir.2008.003.Search in Google Scholar
Itô, Junko. 1986. Syllable theory in prosodic phonology. Amherst, MA: University of Massachusetts, Amherst dissertation.Search in Google Scholar
Jespersen, Otto. 1904/1913. Lehrbuch der Phonetik, 5th edn. Leipzig: B.G. Teubner.Search in Google Scholar
Jun, Jongho. 1995. Perceptual and articulatory factors in place assimilation: An Optimality Theoretic approach. Los Angeles, CA: University of California, Los Angeles dissertation.Search in Google Scholar
Katamba, Francis. 1989. An introduction to phonology. London & New York: Longman.Search in Google Scholar
Kawasaki, Haruko & John Ohala. 1980. Acoustic basis for universal constraints on sound sequences. Journal of the Acoustical Society of America 68(S1). S33. https://doi.org/10.1121/1.2004680.Search in Google Scholar
Kaye, Jonathan. 1992. Do you believe in magic? The story of s+C sequences. In Henryk Kardela & Bogdan Szymanek (eds.), A Festschrift for Edmund Gussmann from his friends and colleagues, 156–176. Lublin: University Press of the Catholic University of Lublin.Search in Google Scholar
Kenstowicz, Michael. 1994. Phonology in generative grammar. Cambridge, MA: Blackwell.Search in Google Scholar
Kiparsky, Paul. 2003. Syllables and moras in Arabic. In Caroline Féry & Ruben van de Vijver (eds.), The syllable in optimality theory, 147–182. Cambridge: Cambridge University Press.10.1017/CBO9780511497926.007Search in Google Scholar
Kiparsky, Paul. 2018. Formal and empirical issues in phonological typology. In Larry Hyman & Frans Plank (eds.), Phonological typology, 54–106. New York: De Gruyter Mouton.10.1515/9783110451931-003Search in Google Scholar
Ladefoged, Peter. 1975. A course in phonetics. New York: Harcourt Brace Jovanovich.Search in Google Scholar
Ladefoged, Peter & Keith Johnson. 2015. A course in phonetics. Stamford: Cengage Learning.Search in Google Scholar
Ladefoged, Peter & Ian Maddieson. 1996. The sounds of the world’s languages. Cambridge, MA: Blackwell.Search in Google Scholar
Lass, Roger. 1984. Phonology: An introduction to basic concepts. Cambridge: Cambridge University Press.Search in Google Scholar
Liljencrants, Johan & Björn Lindblom. 1972. Numerical simulation of vowel quality systems: The role of perceptual contrast. Language 48. 839–862. https://doi.org/10.2307/411991.Search in Google Scholar
Lindblom, Björn. 1983. Economy of speech gestures. In Peter F. MacNeilage (ed.), The production of speech, 217–246. New York: Springer.10.1007/978-1-4613-8202-7_10Search in Google Scholar
List, Johann-Mattis. 2019. Automated methods for the investigation of language contact, with a focus on lexical borrowing. Language and Linguistics Compass 13(10). e12355. https://doi.org/10.1111/lnc3.12355.Search in Google Scholar
List, Johann-Mattis, Simon Greenhill, Cormac Anderson, Thomas Mayer, Tiago Tresoldi & Robert Forkel. 2018. CLICS2: An improved database of cross-linguistic colexifications assembling lexical data with the help of cross-linguistic data formats. Linguistic Typology 22(2). 277–306. https://doi.org/10.1515/lingty-2018-0010.Search in Google Scholar
Macklin-Cordes, Jayden & Erich Round. 2020. Re-evaluating phoneme frequencies. Frontiers in Psychology 11. 3181. https://doi.org/10.3389/fpsyg.2020.570895.Search in Google Scholar
Macklin-Cordes, Jayden & Erich Round. 2022. Challenges of sampling and how phylogenetic methods help: With a case study of the Pama-Nyungan laminal contrast. Linguistic Typology 26(3). 533–572. https://doi.org/10.1515/lingty-2021-0025.Search in Google Scholar
Macklin-Cordes, Jayden, Claire Bowern & Erich Round. 2021. Phylogenetic signal in phonotactics. Diachronica 38(2). 210–258. https://doi.org/10.1075/dia.20004.mac.Search in Google Scholar
Maddieson, Ian. 1989. Prenasalised stops and speech timing. Journal of the International Phonetic Association 19(2). 57–66. https://doi.org/10.1017/s0025100300003856.Search in Google Scholar
Maniwa, Kazumi, Allard Jongman & Travis Wade. 2009. Acoustic characteristics of clearly spoken English fricatives. Journal of the Acoustical Society of America 125(6). 3962–3973. https://doi.org/10.1121/1.2990715.Search in Google Scholar
McCarthy, John. 2008. Doing optimality theory: Applying theory to data. New Jersey: Wiley-Blackwell.10.1002/9781444301182Search in Google Scholar
Nakajima, Yoshitaka, Kazuo Ueda, Shota Fujimaru, Hirotoshi Motomura & Yuki Ohsaka. 2012. Acoustic correlates of phonological sonority in British English. Proceedings of Fechner Day 28. 56–61.10.1121/1.4800164Search in Google Scholar
Ohala, John & Haruko Kawasaki. 1997. Alternatives to the sonority hierarchy for explaining segmental sequential constraints. In Stig Eliasson & Ernst Jahr (eds.), Language and its ecology, 343–366. Berlin & New York: De Gruyter Mouton.Search in Google Scholar
Orie, Olanike & Victoria Bricker. 2000. Placeless and historical laryngeals in Yucatec Maya. International Journal of American Linguistics 66(3). 283–317. https://doi.org/10.1086/466427.Search in Google Scholar
Orzechowska, Paula. 2018. Exponents of sonority in Slavic and Germanic languages. Paper presented at the Proceedings of the 9th international conference on speech prosody, Poznań.10.21437/SpeechProsody.2018-202Search in Google Scholar
Parker, Steve. 2002. Quantifying the sonority hierarchy. Amherst, MA: University of Massachusetts, Amherst dissertation.Search in Google Scholar
Parker, Steve. 2008. Sound level protrusions as physical correlates of sonority. Journal of Phonetics 36(1). 55–90. https://doi.org/10.1016/j.wocn.2007.09.003.Search in Google Scholar
Parker, Steve. 2012. The sonority controversy. Berlin: De Gruyter Mouton.10.1515/9783110261523Search in Google Scholar
Price, Patti. 1980. Sonority and syllabicity: Acoustic correlates of perception. Phonetica 37(5–6). 327–343. https://doi.org/10.1159/000260001.Search in Google Scholar
Prince, Alan & Paul Smolensky. 1993/2004. Optimality theory: Constraint interaction in generative grammar. Malden, MA & Oxford: Blackwell.10.1002/9780470759400Search in Google Scholar
Rialland, Annie. 1994. The phonology and phonetics of extrasyllabicity in French. In Patricia Keating (ed.), Phonological structure and phonetic form, 136–159. Cambridge: Cambridge University Press.10.1017/CBO9780511659461.009Search in Google Scholar
Round, Erich. 2017a. The AusPhon-Lexicon project: 2 million normalized segments across 300 Australian languages. Paper presented at the 47th Poznań Linguistic Meeting, Poznań.Search in Google Scholar
Round, Erich. 2017b. Review of Gordon, Matthew K. Phonological typology, OUP 2016. Folia Linguistica 51(3). 745–755.10.1515/flin-2017-0027Search in Google Scholar
Round, Erich R. 2021a. glottoTrees: Phylogenetic trees in Linguistics. R package version 0.1. Available at: https://github.com/erichround/glottoTrees.Search in Google Scholar
Round, Erich R. 2021b. phyloWeights: Calculation of genealogically-sensitive proportions and averages. R package version 0.3. Available at: https://github.com/erichround/phyloWeights.Search in Google Scholar
Round, Erich. 2022. Phonotactics in Australian languages. In Claire Bowern (ed.), Oxford guide to Australian languages. Oxford: Oxford University Press.10.1093/oso/9780198824978.003.0011Search in Google Scholar
Round, Erich & Greville Corbett. 2020. Comparability and measurement in typological science: The bright future for linguistics. Linguistic Typology 24(3). 489–525. https://doi.org/10.1515/lingty-2020-2060.Search in Google Scholar
Round, Erich & Jayden Macklin-Cordes. 2015. On the design, in practice, of typological microvariables. Paper presented at New developments in the quantitative study of languages, Helsinki.Search in Google Scholar
Round, Erich, Rikker Dockum & Robin J. Ryder. 2022. Evolution and trade-off dynamics of functional load. Entropy 24(4). https://doi.org/10.3390/e24040507.Search in Google Scholar
Rzymski, Christoph, Tiago Tresoldi, Simon Greenhill, Mei-Shin Wu, Nathanael Schweikhard, Maria Koptjevskaja-Tamm, Volker Gast, Timotheus Bodt, Abbie Hantgan & Gereon Kaiping. 2020. The database of cross-linguistic colexifications, reproducible analysis of cross-linguistic polysemies. Scientific Data 7(1). 1–12. https://doi.org/10.1038/s41597-019-0341-x.Search in Google Scholar
Scheer, Tobias. 2004. A lateral theory of phonology: What is CVCV, and why should it be? Berlin: Mouton de Gruyter.10.1515/9783110908336Search in Google Scholar
Scheer, Tobias & Eugeniusz Cyran. 2017. Syllable structure in government phonology. In Stephen J. Hannahs & Anna R. K. Bosch (eds.), The Routledge handbook of phonological theory, 262–292. London: Routledge.10.4324/9781315675428-10Search in Google Scholar
Selkirk, Elizabeth. 1982. The syllable. In Harry van der Hulst & Norval Smith (eds.), The structure of phonological representations: Part 2, 337–384. Dordrecht: Foris.Search in Google Scholar
Selkirk, Elizabeth. 1984. On the major class features and syllable theory. In Mark Aronoff & Richard Oehrle (eds.), Language sound structure: Studies in phonology, 107–136. Cambridge: MIT Press.Search in Google Scholar
Shih, Stephanie & Sharon Inkelas. 2018. Autosegmental aims in surface-optimising phonology. Linguistic Inquiry 50(1). 137–196. https://doi.org/10.1162/ling_a_00304.Search in Google Scholar
Sievers, Eduard. 1876. Grundzüge der lautphysiologie zur einlührung in das studium der lautlehre der indogermanischen sprachen. Charleston: Nabu Press.Search in Google Scholar
Smolensky, Paul. 2006. Optimality in phonology II: Harmonic completeness, local constraint conjunction, and feature domain markedness. In Smolensky Paul & Géraldine Legendre (eds.), The harmonic mind: From neural computation to optimality-theoretic grammar, 27–160. Cambridge, MA: MIT Press.Search in Google Scholar
Steriade, Donca. 1982. Greek prosodies and the nature of syllabification. Cambridge, MA: Massachusetts Institute of Technology dissertation.Search in Google Scholar
Steriade, Donca. 1993. Closure, release, and nasal contours. In Marie Huffman & Rena Krakow (eds.), Nasals, nasalization, and the velum, 401–470. Amsterdam: Elsevier.10.1016/B978-0-12-360380-7.50018-1Search in Google Scholar
Steriade, Donca. 2001a. Directional asymmetries in place assimilation: A perceptual account. In Elizabeth Hume & Keith Johnson (eds.), The role of speech perception in phonology, 219–250. Cambridge: Academic Press.10.1163/9789004454095_013Search in Google Scholar
Steriade, Donca. 2001b. The phonology of perceptibility effects: The P-map and its consequences for constraint organization. In Kristin Hanson & Sharon Inkelas (eds.), The nature of the word: Studies in honour of Paul Kiparsky, 151–179. Cambridge: MIT Press.10.7551/mitpress/7894.003.0011Search in Google Scholar
Steriade, Donca. 2002. The syllable. In William Frawley (ed.), Oxford Encyclopedia of linguistics. Cambridge: Cambridge University Press.Search in Google Scholar
Tak, Jin-young. 2011. Universals of prenasalized consonants: Phonemic or derived, single or complex? Journal of Universal Language 12(2). 127–158. https://doi.org/10.22425/jul.2011.12.2.127.Search in Google Scholar
Vaux, Bert & Andrew Wolfe. 2009. The appendix. In Eric Raimy & Charles Cairns (eds.), Contemporary views on architecture and representations in phonology, 75–97. Cambridge: MIT Press.10.7551/mitpress/9780262182706.003.0005Search in Google Scholar
Vennemann, Theo. 1972. On the theory of syllabic phonology. Linguistische Berichte 18(1). 1–18.Search in Google Scholar
Vennemann, Theo. 1987. Preference laws for syllable structure: And the explanation of sound change with special reference to German, Germanic, Italian, and Latin. Walter: De Gruyter Mouton.10.1515/9783110849608Search in Google Scholar
van de Weijer, Jeroen. 1996. Segmental structure and complex segments. Berlin: De Gruyter.10.1515/9783110955293Search in Google Scholar
Wiswall, Wendy. 1989. Fula consonant gradation: In support of radical underspecification. Paper presented at the Proceedings of the West Coast Conference on Formal Linguistics.Search in Google Scholar
Wright, Richard. 1996. Consonant clusters and cue preservation in Tsou. Los Angeles, CA: University of California, Los Angeles dissertation.Search in Google Scholar
Wright, Richard. 2001. Perceptual cues in contrast maintenance. In Elizabeth Hume & Keith Johnson (eds.), The role of speech perception in phonology, 252–277. New York: Academic Press.10.1163/9789004454095_014Search in Google Scholar
Wright, Richard. 2004. A review of perceptual cues and cue robustness. In Bruce Hayes, Robert Kirchner & Donca Steriade (eds.), Phonetically based phonology, 34–57. Cambridge: Cambridge University Press.10.1017/CBO9780511486401.002Search in Google Scholar
Yin, Ruihua. 2021. The sonority sequencing principle: A large-scale cross-linguistic investigation of phonotactics. Brisbane, QLD: The University of Queensland dissertation.Search in Google Scholar
Zalizniak, Anna. 2008. A catalogue of semantic shifts towards. In Martine Vanhove (ed.), From polysemy to semantic change: Towards a typology of lexical semantic associations, 217–232. Amsterdam: Benjamins.10.1075/slcs.106.10zalSearch in Google Scholar
Zec, Draga. 1994. Sonority constraints on prosodic structure. London: Routledge.Search in Google Scholar
Zec, Draga. 1995. Sonority constraints on syllable structure. Phonology 12(1). 85–129. https://doi.org/10.1017/s0952675700002396.Search in Google Scholar
Zec, Draga. 2007. The syllable. In Paul de Lacy (ed.), Handbook of phonology, 161–194. Cambridge: Cambridge University Press.10.1017/CBO9780511486371.009Search in Google Scholar
Supplementary Material
This article contains supplementary material (https://doi.org/10.1515/lingty-2022-0038).
© 2022 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
Articles in the same Issue
- Frontmatter
- Articles
- Current research in phonological typology
- Investigating the ‘what’, ‘where’ and ‘why’ of global phonological typology
- Canonical phonology and criterial conflicts: relating and resolving four dilemmas of phonological typology
- Refining explanation in Evolutionary Phonology: macro-typologies and targeted typologies in action
- The prosodic foot beyond prosodic prominence: a preliminary survey
- On the comparability of prosodic categories: why ‘stress’ is difficult
- Bootstrap co-occurrence networks of consonants and the Basic Consonant Inventory
- Frequent violation of the sonority sequencing principle in hundreds of languages: how often and by which sequences?
- Diachronic phonological typology: understanding inventory structure through sound change dynamics
- Place typology and evolution of implosives in Indo-Aryan languages
- Estimating areal effects in typology: a case study of African phoneme inventories
- Word prosody of African versus European-origin words in Afro-European creoles
- Towards a phonological typology of the Kalahari Basin Area languages
- The typological frequency of consonants is highly predictive of their order of acquisition in English
Articles in the same Issue
- Frontmatter
- Articles
- Current research in phonological typology
- Investigating the ‘what’, ‘where’ and ‘why’ of global phonological typology
- Canonical phonology and criterial conflicts: relating and resolving four dilemmas of phonological typology
- Refining explanation in Evolutionary Phonology: macro-typologies and targeted typologies in action
- The prosodic foot beyond prosodic prominence: a preliminary survey
- On the comparability of prosodic categories: why ‘stress’ is difficult
- Bootstrap co-occurrence networks of consonants and the Basic Consonant Inventory
- Frequent violation of the sonority sequencing principle in hundreds of languages: how often and by which sequences?
- Diachronic phonological typology: understanding inventory structure through sound change dynamics
- Place typology and evolution of implosives in Indo-Aryan languages
- Estimating areal effects in typology: a case study of African phoneme inventories
- Word prosody of African versus European-origin words in Afro-European creoles
- Towards a phonological typology of the Kalahari Basin Area languages
- The typological frequency of consonants is highly predictive of their order of acquisition in English
