This festschrift honours phonologist and acquisitionist Dan Dinnsen, who was a member of faculty in linguistics at Indiana University from 1972 to 2012, and whose collaborative NIH grants on learning, theory and treatments of children's phonological disorders were held continuously from 1985 to 2014.* His career spanned several decades of theoretical progress, reassessment and revolution: in phonological theory, in experimental methods, in clinical practices and in the relationships between each of these domains. For much of this time, to quote Stephen B. Chin's introduction (p. 6): ‘Dinnsen and his research assistants kept one eye on the [child disordered] data and the other on current theoretical literature, adapting their analyses to changing theoretical conceptions of representations, rules, and constraints’. Dinnsen's legacy is evident in this volume of eleven research overviews and original articles from his former students and colleagues, edited with comprehensive section introductions by two of his PhD advisees, Ashley W. Farris-Trimble and Jessica A. Barlow.
From my reading, the stand-out aspect of the volume is the excellent and comprehensive literature reviews that almost every chapter provides within its theoretical or conceptual area. The effect is rather like reading the proceedings of a wide-ranging, interdisciplinary seminar; in the areas where I am knowledgeable I was impressed by the depth achieved in a few pages, and in the areas where I am not I learned a lot quickly. This review therefore emphasises the scope and breadth of each chapter, along with some brief summaries of their findings and some caveats regarding interpretation. The volume is organised into four sections, although my summaries of the first three are somewhat reordered to highlight their empirical coverage; the last section's papers have more explicit theoretical goals and are discussed together, as the editors intended.
Three of the chapters tackle L1 phonological acquisition, each from a different empirical angle. Memet Yavaş’ chapter, entitled ‘What guides children's acquisition of #sC clusters? A cross-linguistic account’, provides an extensive and cross-linguistic introduction to sonority, in relation especially to consonant clusters, and to the special behaviour of /s/+consonant clusters. The acquisition data presented is impressive in its scope, analysing /sC/ productions from a total of 356 children learning six languages (Hebrew, Croatian and Polish, plus three Germanic languages) from both typical and atypical populations. With respect to cluster accuracy, the Germanic data showed better performance with better-sonority clusters, while the other three languages showed, if anything, the opposite trend. The chapter suggests that sonority preferences in this second language group are trumped by learners’ exposure to those languages’ larger range of clusters, including many with flat or falling sonority. (Another confound not discussed is the relative difficulty of the segments among the rising sonority clusters in the two language groups, such as the many stop–nasal clusters in the non-Germanic set.) With respect to the deletion site when clusters were reduced, the same pattern held for all six languages: /s/+[–continuant] clusters were more often reduced to their stop or nasal member, whereas /s/+[+continuant] clusters were more often reduced to [s]. The overall conclusion the study draws is that sonority is not the primary factor driving cluster reduction – although a hierarchy of onset sonority constraints interleaved with *Fricative, as in Pater & Barlow (Reference Pater and Barlow2003), would neatly capture these two patterns of deletion.
David Ingram & Virginia L. Dubasik's contribution, ‘Sibling rivalry: comparing phonological similarity between twin and non-twin siblings’, presents a method for capturing a multidimensional overview of children's phonological development (Ingram & Ingram Reference Ingram and Ingram2001), using measures of word length, preferred syllable shapes, and consonantal inventories and accuracy, which can be compared in order to measure the similarity between any two children's phonologies. The present paper applies this method to compare four pairs of 17–22 month olds: identical twins, fraternal twins and two pairs of siblings. The most general result is that higher degrees of relatedness are consistent with closer-matched phonological development, as the identical twins showed 88% overall similarity, while the siblings with the greatest age difference between them (five years) scored only 68%. When each measure is considered in turn, no clear connection emerges between dimensions of similarity and familial relatedness; much of the discussion emphasises the wide-ranging potential for individual variation.
A chapter by Judith A. Gierut, Michele L. Morrisette & Caitlin J. Younger introducing clinical data, ‘Abstracting phonological generalizations: evidence from children with disorders’, provides a non-technical overview of how children respond to treatments for phonological articulatory disorders, with respect to how they generalise from a single treated sound to new contexts, segments and natural classes. It reports on a longitudinal, retrospective study of 65 children, each given therapy on a single segment missing from their inventory (mostly drawn from velar stops, liquids and fricatives), and tracks their generalisations from that sound over time. The somewhat counterintuitive overall result is that progress is most often seen first in the production of sounds other than the treated one, and from a different manner class; given this, the authors point out that tracking children's abilities with untreated sounds might well be important even before they show progress on the target segment. The authors also raise implications for theories of error-driven learning derived from their clinical data, but claim (p. 84) that error-driven algorithms have not been tested with the kind of ‘behavioral evidence from children directly engaged in phonological learning’ in this chapter. This seems to dismiss the wealth of longitudinal behavioural data used even in some of the learning studies cited (see especially Levelt et al. Reference Levelt, Schiller and Levelt2000, and many other references in Jarosz Reference Jarosz2010) – perhaps because it was not associated with explicit error correction as used in therapy, though it is uncontroversial that most children engage directly with phonological errors and learn from them without intervention.
Another two chapters explore morphological interactions with children's phonological learning. The opening chapter, ‘Prosodic Licensing and the development of phonological and morphological representations’, by Katherine Demuth, is a balanced expert summary exploring the ways in which prosodic structures (syllables, words and phrases) interact with children's early use of morphology. One clear example comes from functional morphemes which emerge first with monosyllabic roots in multiple unrelated languages – including English and French determiners as well as Sesotho (Bantu) noun-class prefixes – consistent with an output prosodic maximum of a single foot. The contribution by Jessica A. Barlow & Sonja L. Pruitt-Lord, entitled ‘The role of phonological context in children's overt marking of ‘-s’ in two dialects of American English’, looks in more depth at the predictive variables for two homophonous English suffixes (plural and 3rd person singular), in two experiments focused on learners of African-American English and more standard English respectively. It begins with succinct but wide-ranging background on the many orthogonal factors influencing children's production of /-s/ suffixes, including the preceding and following segmental contexts, the suffix's morphological role and the increased rates of coda deletion in the target, African-American English. While the experiments showed some expected trends, whereby plurals were more frequent than verbal marking, each group also unexpectedly produced more [-s] on some set of C-final stems rather than V-final ones. One explanation the authors suggest is that the /-s/ suffixes following stem-final consonants are more often preserved because they can be prosodified as appendices, rather than as simple syllable codas.
Moving beyond child speech production, three chapters consider language development in other learning contexts. ‘Covert contrast in the acquisition of second language phonology’, by Fred R. Eckman, Gregory K. Iverson & Jae Yung Song, begins with a valuable overview on the intellectual history of covert contrasts in L1 acquisition. In the paper's study, this potential for less than target-like contrast is investigated among adults learning the L2 English phonemic status of [s] vs. [z], given that their L1 Spanish grammar treats these segments as allophones. In this regard, the study found that two of the 13 participants had been transcribed by native speakers as producing no [s~z] contrast, but showed significant differences in the acoustic overlap between frication and voicing between /s/ and /z/ targets.
Two other chapters deal with populations which will perhaps be more unfamiliar to phonologist readers. David B. Pisoni contributes a chapter entitled ‘Rapid phonological coding and working memory dynamics in children with cochlear implants: cognitive foundations of spoken language processing’, which argues that the wide range of individual differences in language-perception abilities among children who received cochlear implants can be better understood using ongoing measures of more domain-general cognitive processes. The chapter details numerous studies from Pisoni and colleagues, with the claim that low performance on post-implant spoken language measures is connected to deficits in information storage and retrieval, memory and executive function. I note that this paper is silent on some topics that may also strike other linguists as crucial. For example, it spends very little time on the relevance of input modality, as split between speech and sign, and it does not make explicit the assumption that such deficits are the product of all language input being disrupted via hearing loss, i.e. that they would not be expected among children who had received consistent sign input from birth.
A very different study of language users with unusual input is William D. Keel's ‘German settlement varieties in Kansas: some unusual phonological and morphological developments with the approach of language death’. Here the linguistic focus is on non-standard innovations reported in the German spoken by the final generations of such speakers in Kansas, such as post-nasal assimilation of stops (/nd/→[nn]) and lexically specific /f/-voicing, and the extent of their similarities with other German dialects.
The volume's last section includes three constraint-based analyses of phonological data. The first is Laura W. McGarrity's ‘The role of onsets in primary and secondary stress patterns’, which begins with an excellent cross-linguistic summary of how the presence and segmental qualities of onset consonants can be restricted or required in stressed (sometimes just primary stressed) syllables. The final goal of this chapter is to connect this adult typology with some exceptional onset choices and primary vs. secondary stress asymmetries from child truncation. For example: Gnanadesikan (Reference Gnanadesikan, Kager, Pater and Zonneveld2004) documents a pattern of pre-tonic syllable fi- overwriting ([fi-ˈbeja] umbrella, [fi-ˈdina] Christina), where sometimes the overwritten onset was shifted to the preserved syllable: [fi-ˈbun] balloon, not *[fi-ˈlin], [fi-ˈkala] koala, not *[fi-ˈwala]. McGarrity points out the similarity of this preference for stressed onset stops over glides and liquids, at the expense of a Contiguity violation, with the avoidance of stressed glide onsets in adult languages like Niuafo'ou (Tsukamoto Reference Tsukamoto1988, de Lacy Reference de Lacy2002).
Farris-Trimble's chapter, ‘A faithfulness conspiracy: the selection of unfaithful mappings in Amahl's grammar’, deals with Smith's (Reference Smith1973) classic data from a fresh perspective. The paper first uses sets of Amahl's segmental repairs to carefully illustrate the concept of ‘gang effects’ among faithfulness constraints: how the processes interact, how competing pressures act cumulatively to choose outputs, how classic OT formalism is not equipped to handle them and how either local conjunction or constraint weighting (as in Harmonic Grammar) can provide a working analysis. The chapter then takes a step into meta-analysis, introducing the notion of a faithfulness ‘conspiracy’ to characterise a child or adult grammar which avoids a brand of faithfulness across multiple contexts or processes.
Some similar data are found in the final chapter, ‘Superadditivity and limitations on syllable complexity in Bambara words’, by Christopher R. Green & Stuart Davis. The facts are from Colloquial Bambara (spoken in Mali), in which syncope and intervocalic [k]-deletion create marked syllable structures – codas and complex onsets – which are both avoided in Standard Bambara. In longer and multimorphemic words, the penalties for these more complex syllables can sometimes gang up together, and the authors also capture this additive constraint interaction within Harmonic Grammar. The chapter further argues that some restrictions on multiple reductions require more additivity than constraint weights alone can provide; the solution presented invokes the local conjunction of weighted constraints to derive these ‘superadditive’ effects. This is the most complicated theoretical analysis in the volume; perhaps excessively so, and possibly as the result of a suboptimal constraint set (for example, the location of deletable vowels is often described with reference to feet, but the syncope constraints are simply *Peak[+hi] and *Peak[–hi], without reference to foot-structure goals that could drive deletion; cf. Gouskova Reference Gouskova2003). Nevertheless, its wealth of well-organised facts is so tantalising that they invite any reader interested in additive interactions to have a crack at constructing alternatives.
This festschrift is a very fond intellectual tribute to Dan Dinnsen, as a pioneer of theoretically driven inquiry across speech-related disciplines whose frequently unique work applied phonological theories to novel data. In the present era, where so much phonology is being developed at the interface with other disciplines, its breadth may similarly spark new searches for the connections between disparate facets of speech, acquisition and grammar.
