Publishing Partner: Cambridge University Press CUP Extra Publisher Login

New from Cambridge University Press!


Revitalizing Endangered Languages

Edited by Justyna Olko & Julia Sallabank

Revitalizing Endangered Languages "This guidebook provides ideas and strategies, as well as some background, to help with the effective revitalization of endangered languages. It covers a broad scope of themes including effective planning, benefits, wellbeing, economic aspects, attitudes and ideologies."

New from Wiley!


We Have a New Site!

With the help of your donations we have been making good progress on designing and launching our new website! Check it out at!
***We are still in our beta stages for the new site--if you have any feedback, be sure to let us know at***

Review of  Friction between phonetics and phonology

Reviewer: Michael C. Cahill
Book Title: Friction between phonetics and phonology
Book Author: Janine Berns
Publisher: Netherlands Graduate School of Linguistics / Landelijke (LOT)
Linguistic Field(s): Phonology
Subject Language(s): French
Issue Number: 26.3728

Discuss this Review
Help on Posting
Reviews Editor: Helen Aristar-Dry


“Friction between phonetics and phonology: The status of affricates” is Janine Berns’s Ph.D. thesis from Radboud University in Nijmegen. The distinction between strident and non-strident affricates is significant in this book, strident affricates being those with a noisy sibilant release (e.g. [ts, tʃ] and their voiced counterparts [dz, dʒ], and non-strident affricates being others (e.g. [pf, kx]).

Chapter 1 introduces the basic phonetics of affricates, distinguishing them from stops and fricatives. Berns helpfully reviews historical views of affricates as one vs. two phonemes, from Trubetzkoy’s (1969) criteria, to representations in non-linear phonology. She also notes that the disagreement on the nature of affricates – a special type of stop vs. a stop-fricative complex – remains, and suggests that this is due to an insufficiently broad and typologically balanced sample of languages. While referring to typology of affricates in the UPSID (UCLA Phonological Segment Inventory Database) of phonological inventories, for processes, both historic and synchronic, she also includes a major study on French. The obvious rejoinder to this approach is that this is yet another single-language study, but she asserts that French has occupied a special place in the debate, and by comparing results to the UPSID data, useful conclusions can be reached.

Chapter 2 provides phonological background to the study. It gives a detailed review of the debate on the nature of affricates, starting from affricates as strident stops (Jakobson et al 1952) and stops with delayed release feature (Chomsky and Halle 1968), to the more complex feature geometries and contour segment structure of Sagey (1990), to the view of Hualde (1988) and others that affricates are complex rather than contour segments (because of Basque, where the edge effects predicted by a contour analysis do not work). These have different geometries in terms of the specific nodes the features must be linked to. Steriade’s (1994) proposal in aperture theory and Schafer’s (1995) approach in terms of headedness are also noted. In the 1990s, the view of affricates as stops with various modifications returned, with Shaw’s radical underspecification approach to Tahltan, Rubach’s (1994) return to a stridency analysis similar to Jakobson’s, and Clements’ 1999 proposal resembling this, but also including features that would include non-strident affricates. Berns spends a fair amount of time sympathetically discussing the proposal of Kehrein (2002), based on a number of languages, that affricates do not exist as a distinct phonological category, but that relevant features are added to produce phonetic-level affricates. Finally, she notes that Jacobs and van Gerwen (2006) return to a contour segment analysis based on historical processes in French and Spanish, and raise the important theoretical question of whether affricates must have a uniform structure cross-linguistically, or whether languages may vary in their underlying representations of affricates. The issues to be resolved, then, are what features are involved in affricates as opposed to simple stops, whether these features are ordered or not, and where they attach in a feature geometry. This is a good historical review of what different approaches have been in the last several decades, and illustrates well that the same issues continue to be debated, but largely in terms of “Language X behaves this way, so I propose a universal approach to affricates.”

Chapters 3-5 present a typological overview. Chapter 3 taps into the genetically balanced 1992 UPSID to extract typological characteristics of affricates. The few previous cross-linguistic studies were based partially or totally on the earlier 1984 UPSID, and the updated version provides checks on these. Three hundred and two languages, about two-thirds of the languages in UPSID, contain at least one affricate. The affricates [tʃ, ts, dʒ, dz] occur in the most languages, in that order, and for languages with only one affricate, the order is [tʃ, dʒ, ts, dz]. Archi, a Caucasian language, has 24 affricates, among 81 consonants, and an examination of its inventory shows that affricates may share some properties with plosives and or fricatives (e.g. place, secondary articulations, length, voicing), but may also be different from, even complementary to, these groups. Bern spends several pages detailing the cross-linguistic patterns of these. Except for bilabials, affricates are attested at all places of articulation (for bilabials, the fricative release is always labiodental, e.g. [pf], in UPSID). Affricates may occur where plosives do not, or vice versa. As may be inferred by the most frequent affricates listed above, there is a cross-linguistic preference for voiceless rather than voiced affricates, which parallels the cross-linguistic preference for voiceless plosives and fricatives. For manner, there is a strong preference for sibilant affricates, e.g. [tʃ, ts] rather than non-sibilant, e.g. [pf, kx]. Also, if a language has only one sibilant fricative and one sibilant affricate, they are most often in different places of articulation, with the most common pattern being [s, tʃ]. If a language has a number of fricatives and affricates, they do not have only fricatives or only sibilants; they have both. Secondary articulations are also common in affricates, in 188 of the 302 languages. Only four languages have phonologically long affricates (generally with the plosive part being lengthened: [ttʃ], not [tʃʃ]). A number of other typological observations are made, too many to list in this brief review.

Chapter 4 discusses theoretical implications of the typological patterns extracted from UPSID in Chapter 3. Berns starts by discussing two phonological views. First is the claim by van de Weijer and others that affricates and fricatives typically occur at the same place of articulation, leading to a geometry in which Place features depend on [continuant]. The second view is that affricates can be regarded as modified stops, because affricates pattern with stops as far as segmental modification. These are shown to be far from universal in UPSID 1992 (recall the most common [s, tʃ] pattern, and occasional patterning of affricates with fricatives), and so are typologically questionable. She then proposes a simple stop phonological characterization of affricates. In this account, all stops and affricates are [-continuant], while fricatives are [+continuant]. Affricates are distinguished from stops in having an additional feature: [+strident] or [+lateral], the appropriate release for affricates such as [tʃ, kL]. Affricates such as [tθ] and [pf] are analyzed as having a different place than the pure stop. Thus stops and affricates differ by Manner or Place features, not [continuant], largely following Kehrein’s (2002) framework. Using these features can account for why affricates pattern with fricatives at times. Berns ends the chapter by showing how this would work out in a Feature Geometry framework, with one benefit being that such representations entail that a geminated affricate would always have the stop part ([-continuant]) lengthened, not the fricative portion, since there is no [+continuant] feature present. She concludes that a pure stop description of affricates fits the data better than any sort of complex stop representation.

Chapter 5 examines affricates in the historic phonology of French. Neither Classical Latin nor modern French has affricates, but the development and disappearance of these in historic French was expected to be relevant to the current study. Affricates arose through intrusion of stops, through multiple processes of palatalization of stops followed by affrication, and through glide strengthening. Deaffrication in the twelfth and thirteenth centuries A.D. took place by losing the stop component, leaving the sound as a fricative; this was evidently a part of a wider process of spirantizing a number of stops as well as affricates. Berns presents other researchers’ feature geometry accounts of all these historical processes of affricate formation, as well as a more strictly phonetically based account. The palatalization process (ki > tʃ) is somewhat problematic for either a strictly articulatory-based or a perception-based account, and Berns presents a proposal that combines both assimilation (articulatory) and a misperception of narrow constriction with turbulence as [strident]. A perhaps unexpected and unusual process in historical French is that velar stops palatalized before /a/ as well as before front vowels. In detailed discussion, she shows that all the proposed accounts of affricate formation by glide strengthening have various problems, and proposes an alternative that combines articulatory and phonetic factors. The presentation on deaffrication is briefer, and is based on Berns’s characterization of affricates as strident stops; a first historical stage of spirantization targeted plain stops, and a second stage targeted strident stops (fricatives).

Chapter 6 and the following two chapters comprise Part III: A Corpus Study of Modern French. Chapter 6 begins with a sketch of the two processes producing phonetic (not phonemic) affricates in modern French. First, velar stops palatalize before front vowels in modern French, and in some northern varieties, before /a/ as well, recalling the development of affricates in earlier French. Second, assibilation of coronal stops before front vowels (e.g. ti -> tsi) occurs not only in Canadian French, where it is well-known, but also in some varieties of continental French. The bulk of the chapter describes the PFC (Phonologie du français contemporain) corpus (website, which includes hours of orthographically transcribed wordlists and conversations, from which Berns draws her data. She selected subsets of the data containing 84 speakers from nine regions.

Chapter 7 focuses on the first relevant process in modern French, the fronting of velars. F2 is related to the place of a consonant. With the PFC corpus described above, acoustic measures of F2 at the onset of periodicity of the vowel and F2 at the following vowel nucleus were plotted, with different values of the resulting slope, depending on the identity of that following vowel. In previous studies, velars were shown to have two loci, one for front vowels and one for back vowels. Bern also extracted these two values of F2. The place of articulation of the velars, as expected, moved forward before front vowels, and back before back vowels. Statistical analysis confirms this, and also shows that regional location of has no effect on the results. The results are more complex with /a/, which for historical reasons may be pronounced with either a more front or a more back position. But still, the more anterior the /a/, the more anterior the velar.

Chapter 8, “Plosive releases,” is the longest chapter, and examines in detail the issues associated with plosive releases in modern French. Traditional descriptions have characterized French as having “smooth” (i.e. unaspirated, non-noisy) plosive releases. Again using the PFC corpus with much spectrographic analysis, Berns first examines the wordlist recordings, and shows that labial plosives are indeed largely unaspirated (though there is variation between speakers), but that preceding /i/, as in <piquet> and <épier>, there is some frication release, with more frication and variability in the second type of word. For coronals, there is little aspiration or other noisy release (this varies between voiced and voiceless stops), except in the case before /i/ as in <petit>, where there is oral channel noise that approaches making the stop into a phonetic affricate, e.g. [ts]. Canadian French (unlike Parisian French) does this so systematically that it can be considered phonologized. For velars, all have some degree of noise upon release, as phonetically and typologically expected. The Parisian sequence [ki] has enough noise to be a phonetic affricate [kx], but interestingly, not in Canadian French, and the [ka] sequence has the most inter-speaker variability. The portions of the PFC corpus that were sentences read in context generally showed the same patterns as the wordlists. The variability in these is attributed to variability in the shape of the vocal tract, influenced by the consonant itself, the vowel following, and the individual speaker. No broad effects due to sociolinguistic status of the speaker were found; differences were purely on the individual level.

Chapter 9 summarizes and concludes. To answer the first main research question, on the phonological status of affricates, Berns maintains that affricates are a kind of stop; affricates and plain stops form a natural class. Affricates are specified for Manner, e.g. if the only affricates are strident, then “strident stops” are also a natural class, a subset of “stops.” So the theoretical question raised in Chapter Two, whether affricates must have a uniform structure cross-linguistically or whether languages may vary in their underlying representations is addressed indirectly, with the answer being that there is a uniform structure, albeit with different terminal features depending on the exact affricate under consideration. The second research question was on related processes in modern French. Fronting of velars depended on the following vowel, but also exhibited wide variation. Release of plosives was quite variable also, but often resulted in phonetic affricates. Both are phonetically, not phonologically, conditioned.

Five Appendices are included: a list of languages in the UPSID database which contain affricates, the readings from the PFC corpus, and details of Berns’s measurements of the modern French language data.


This book gives a valuable overview of the literature and decades-long debates on the nature of affricates, and Berns’s proposal of affricates as a subclass of stops is quite reasonable on the basis of the typological evidence she presents. Her study of contemporary French does show in great phonetic detail how phonetic environments can be favorable to the development of phonetic affricates, which in turn may in the future be phonologized. These are both quite noteworthy contributions, independent of their relation to each other.

The relevance of the contemporary French to the debate on the phonological status of affricates is not as tight as could be desired, since today’s French does not contain phonological affricates. In essence, the thesis is two major studies in one, with a loose connection between the two. To tighten the gap between the detailed phonetic study of contemporary French and the phonological status of affricates is, to understate it, a challenge, since one would have to look to a future state of French in which affricates were phonological. It does not appear that French offers any crucial evidence bearing on the proposed feature tree geometries that goes further than the typological evidence presented in earlier chapters.

One question I had was how Berns’s proposal dealt with the phenomena in other publications which advocated a different perspective. One example would be Jacobs and van Gerwen (2006), who proposed a contour segment geometry that Berns noted in Chapter Two. Admittedly, to respond to all of the other theoretical approaches thoroughly would significantly expand the book, but some summaries of possible approaches would have been appropriate.

I have few quibbles with typos and other editorial matters. This is in general a well laid out volume, and well-edited. One apparent mistake: on page 2, Bern characterizes nasals as sonorants but defines them as having “a constriction in the oral cavity which is sufficiently wide to allow the airstream to pass through without yielding friction,” i.e. it sounds like nasals have an oral airflow. Typographically, some of the phonetic symbols are in a distinctly different, apparently monospaced font. In these days when phonetic fonts are more widely available, this aesthetic glitch could easily have been avoided. And some of the spectrograms are so dark that it is difficult to see details of the formant structure.

In keeping with the two major areas of this study, I see two major groups of people who may be interested. Both phonological theorists and students of sociophonetics will benefit from this study.


Chomsky, N. and M. Halle. 1968. The Sound Pattern of English. New York: Harper and Row.

Clements, G.N. 1999. Affricates as noncontoured stops. Proceedings of LP ‘98, ed. by O. Fujimora et al. Prague: Karolinum Press. PP. 271-299.

Hualde, J.I. 1988. Affricates are not contour segments. West Coast Conference on Formal Linguistics 6: 77-89.

Jacobs, H. and R. van Gerwen. 2006. Glide strengthening in French and Spanish and the formal representation of affricates. In D.L. Arteaga (ed.), Historical Romance Linguistics: Retrospective and Perspectives. Amsterdam: John Benjamins Publishing Company, pp. 77-95.

Jakobson, R., G. Fant, and M. Halle. 1952. Preliminaries to Speech Analysis. The Distinctive Features and their Correlates. Massachusetts: MIT Press.

Kehrein, W. 2002. Phonological Representation and Phonetic Phasing: Affricates and Laryngeals. Tübingen: Niemeyer.

Rubach, J. 1994. Affricates as strident stops in Polish. Linguistic Inquiry 25:119-143.

Sagey, Elizabeth. 1990. The Representation of Features in Non-linear Phonology. Garland Press. (revision of 1986 Ph.D. dissertation, MIT).

Schafer, R. 1995. Headedness in the representation of affricates. the Linguistic Review 12:61-87.

Steriade, D. 1994. Complex onsets as single segments: the Mazateco pattern. Perspectives in Phonology, ed. by J. Cole and C. Kisseberth. Sanford, California: CLSI, pp. 203-291.

Trubetzkoy, N.S. 1969. Principles of PHonology (translated by A. M. Baltaxe). Berkeley/Los Angeles: University of California Press.
Michael Cahill (Ph.D. Ohio State University) has a long interest in phonology, especially of tone, of African languages and labialvelars in whatever location. He served in a field project in Ghana for several years with the Konni language. He is currently the Orthography Services Coordinator of SIL International.