Chapter 10 Directions and reflections In this final chapter, Benjamin Munson, Suzanne Purdy and colleagues, Suze Leitão and Joan Rosenthal traverse a breadth of topics, including a rarely considered one in the context of child speech. It concerns sociophonetics, gender stereotyping and social indexing, and Munson (A51) approaches it with enthusiasm and empathy. Purdy, Fairgray and Asad (A52) provide an account of the important links between hearing and SSD, delivering expert guidance from an Audiology perspective. Leitão (A53) reflects on the art and science of clinical thinking, and the issues that can arise. And finally, and inspirationally, Rosenthal (A54) presents her key components of a practitioners’ survival kit.
In the world of phoneticians, the burgeoning field of sociophonetics resides at the intersection of sociolinguistics and phonetics. Most of its work has involved descriptive accounts of phonetic and phonological variation within regional dialects, speech styles, or (social) speech groups and attempts to explore the relationship between phonetics and phonology (Ohala, 1990). By comparison, there has been scant exploration of the relationship between phonetic and phonological variation and how speech is perceived. Roberts (2002) provides a summary of available data, which suggest that children acquire knowledge of sociolinguistic variation from the earliest stages. Precisely how variation comes to be learned in the course of language acquisition is poorly understood. However, we do know that many social factors systematically shape variation in speech production, including individual differences such as age, gender, ethnicity and socio-economic status (Labov, 1994–2001), and the influence of social groups and networks with which speakers are associated (Eckert, 2000; Milroy, 1987). Sociophonetics has applications in pedagogy, foreign language teaching, forensic phonetics and multi-layered transcription (Müller, 2006). In SLP/SLT, it has undeniable implications for understanding of child-directed speech (parentese), therapy discourse, style shifting, speaker- and listener-oriented articulatory control, register, code switching and for deepening cultural and linguistic sensitivity.
Dr. Benjamin Munson is a Professor in Speech Language Hearing Sciences at the University of Minnesota, Minneapolis. His many research interests include relationships among phonology, metaphonology, and the lexicon; speech production in phonological impairment; the cognitive and linguistic bases of phonological development and disorders in children; gender typicality in children’s speech, including when and how children learn to express gender through speech, with a particular focus on how this learning interacts with more general aspects of language learning; and sociophonetics.
Q51. Benjamin Munson: Sociophonetics and child speech practice Quite inadvertently, Van Borsel, Van Rentergem and Verhaeghe (2007) pointed to the importance of SLPs/SLTs having informed views of linguistic variation, enabling them to distinguish genuine pathology from natural non-standard variation, and this is clearly an area where sociophonetics can help. What are the methods of enquiry in this non-traditional area of study? Can you explore for the interested clinician or clinical researcher the likely impact of, and clinically relevant research areas in children’s SSDs for, sociophonetics as its literature base mushrooms and interfaces with clinical phonology (Müller & Ball, A3)?
A51. Benjamin Munson: Sociolinguistic variation and speech sound disorders As practicing SLPs/SLTs know, the articulatory and perceptual characteristics of speech sounds vary from talker to talker, and within talkers, from utterance to utterance. For instance, phonetic detail can vary across talkers due to anatomic and dialectal differences; and within talkers, as a function of ambient noise (Lane & Tranel, 1971), or the presumed language abilities of the person being addressed (e.g., Bradlow, 2002).
Determining whether a variation reflects pathology, warranting treatment, or whether it is normal, is a challenge faced whenever we differentiate between language impairment and first-language interference in children from culturally and linguistically diverse backgrounds (Goldstein, A19; Zajdó, A20). Understanding of, and sensitivity to, the sources of variation simplify the task of forming these judgements.
Imagine two girls growing up in North America who demonstrate superficially equivalent pronunciation patterns, apparently omitting within word /.ɹ/ as in every , substituting /f/ for /θ/ word finally as in bath , and omitting final /t/ and /d/ as in hat and bad, respectively. One girl has these errors because of a problem in phonological acquisition, and requires intervention. The other does not have errors per se , but rather, sound patterns that indicate successful acquisition of a variant of English, African American English, in which these are the speech community’s pronunciations (for a review see Thomas, 2007). The second girl requires no intervention, except perhaps to say that if she were to interact with people in dialectally diverse speech communities, she might benefit from explicit instruction in appropriate code-switching.
Assessing whether variation is pathological or not can be complex, and certainly not always as straightforward, for US clinicians at least, as the comparison above would indicate. Take for example the labiodental variants of /.ɹ/, transcribed as [ʋ], in some dialects of English in the United Kingdom. Superficially, they sound like /.ɹ/ misarticulations that occur in typical acquisition by younger children and in older children with misarticulations. An improbable interpretation of this variant is that it represents a widespread, persistent speech error, but as Foulkes and Docherty (2000) show, rates of use of [ʋ] are highly linked to social stratification. Indeed, its use might signal, intentionally or unintentionally, membership of different social groups, rather than social-group differences in the incidence of misarticulation. SLPs/SLTs cannot determine whether a [ʋ] for /.ɹ/, pattern is an error without knowing its social function in a speech community.
Sociophonetics Sociophonetics melds methodologies and theoretical constructs from several disciplines, including experimental phonetics, psycholinguistics and sociolinguistics. Foulkes (2005) summarises how sociophoneticians catalogue variation in the sound structure of language echoing social-group membership, in production and perception, and how this interacts with other linguistically based phonetic variation: segmental and prosodic.
Perceptual studies in this sub-field reveal that listeners readily associate different pronunciation variants with social categories, often in ways contrary to the actual use of these variants in a population. Niedzielski (1999) illustrates this in an influential study of vowel perception by people in Detroit, Michigan. Participants were presented with synthesised vowels in a speaker identification task, and told that the vowels were modelled on the productions of either Detroiters, or residents of nearby Windsor, Ontario, who speak a different English dialect. Labelling of the Windsor vowels, by the Detroit participants, showed tactic knowledge of the ways that people within that dialect region speak. Interestingly, the labels listeners gave for vowels presumed to be produced by Detroiters exposed social stereotypes of the speech of Detroiters that did not match their actual vowel productions.
A qualitatively similar case comes from Mack and Munson (2012). They examined listeners’ perception of men’s sexual orientation according to how /s/-initial words were produced. A popular-culture stereotype in North America and in much of the Commonwealth of Nations holds that gay men lisp. Though the term ‘lisp’ has fallen out of scientific use among SLP/SLTs, it clearly connotes a misarticulation. Published studies on /s/ variation and sexual orientation in men show that individuals’ production of /s/ is associated with both actual and perceived sexual orientation (Linville, 1998; Munson, McDonald, DeBoe & White, 2006). The distinctive /s/ associated with gay- and gay-sounding men’s speech, however, is arguably a hypercorrect /s/, and not a lisp, as its acoustic characteristics serve to better differentiate it from the acoustically similar sounds /ʃ/ and /θ/ than the heterosexual and heterosexual-sounding men’s /s/ (Jongman, Wayland & Wong, 2000). Munson and Zimmerman (2006) found that listeners label a talker as gayer-sounding when presented with stimuli containing a hypercorrect /s/ than when presented with stimuli containing /s/ with average acoustic characteristics. Nearly identical scores were elicited when listeners rated tokens containing a frontally misarticulated /s/, even though its acoustic characteristics differed markedly from those of hypercorrect /s/.
Other research demonstrating that listener expectations affect speech perception reinforces these findings. For example, expectations about talker gender and social class affect the categorisation of speech sounds (Hay, Warren & Drager, 2006; Munson, 2011; Strand & Johnson, 1996). Strand and Johnson, and Munson, showed that acoustically equivalent American-English lingual fricatives are labelled differently depending whether listeners believed they are listening to a man (favouring a /s/ response) or to a woman (favouring a /ʃ/ response), perhaps signifying tacit knowledge of sex differences in production of these sounds. Hay et al. (2006) showed that listeners in New Zealand label the acoustically ambiguous diphthongs in hair and here differently depending on whether they are led to believe they are produced by a woman or a man, and by a working-class or a middle-class person.
The cases of sexual orientation and /s/, and /.ɹ/ variation in the United Kingdom, are particularly interesting, illustrating that considerable variation in pronunciation can occur within a speech community, without appearing to be due to obvious anatomic or physiologic differences. Moreover, their origins appear to be different from those for regional dialects, the formation of which may be related to factors such as migration and language contact (Trudgill, 2004). But surely labiodental /.ɹ/([ʋ]), hyperarticulated /s/ and very local phonetic variants within high school cliques (Eckert, 2000; Mendoza-Denton, 2007), cannot result from such factors. Rather, they appear to be instances of groups of individuals exploiting permissible variation in speech to convey social categories, alongside propositional linguistic information.
In addition to understanding the causes of variation, SLPs/SLTs must understand its consequences . Consider the fairly robust finding that: English-speaking women hyperarticulate vowels more than men (Bradlow, Torretta & Pisoni, 1996). Perceptual studies reviewed in Munson and Babel (2007) show that many listeners make tacit associations between hyperarticulation and sex typicality of speech. What if children held these stereotypes, too? If they did, they might judge less-articulate male peers as more masculine sounding, and more-articulate female peers as more feminine sounding. This in turn might promote a powerful social motivation for some children, particularly young boys, to resist speech and language therapy aimed at improving intelligibility, because ‘success’ might manifest as a boy sounding less boy-like! Then again, imagine a child with a [t] for /s/ substitution being taught /s/ in therapy. One likely and reasonable instructional strategy would be for the clinician to model a hyperarticulate /s/. The social meaning associated with that phonetic variant in some English-speaking contexts might make boys in particular averse to learning it.
A child who is taught only one variant of /s/ in therapy is ill-equipped to manipulate its characteristics to convey different social registers, unless therapy promotes spontaneous learning of the full range of /s/ variants through encoding and emulation of different models in the population. To this end, peer modelling might be incorporated into therapy.
When clinical SLPs/SLTs are proactive in incorporating ethnographic analysis into their practice, especially with culturally and linguistically diverse populations, they examine the range of phonetic variation throughout the communities in which a child communicates. They develop both taxonomies of phonetic variants and observations of the communicative functions of these variants, much as Eckert (2000) and Mendoza-Denton (2007) did when researching sociophonetic variation in high school students’ speech. Ethnographic analysis holds promise for a rich and detailed picture, more complex, more informative and more culturally apt than traditional descriptive approaches to child speech, a suggestion that is consistent with many of the works assembled by Müller (2006).
When prescriptive standards rather than ethnographic analysis are used, the boundary between ‘error’ and ‘normal variation’ might not be clear. This is illustrated by a series of studies of the incidence of different articulations of /s/ by young adults in Belgium. Van Borsel et al. (2007) examined an almost 23% incidence of what they characterised as dentally misarticulated /s/ in Belgian university students aged 18 to 22, reported to be ‘native speakers of Dutch’. Their incidence fluctuated as a function of some variables rarely cited as being associated with misarticulation rates, such as university field of study. The lowest rates of interdental /s/ were among humanities students, with higher values for natural sciences and social sciences students, and a significant majority of those identified as lisping were unaware that they were assessed as such. Carefully indicating that their finding might not be new, they cite a palatographic study (Dart, 1991, see also Dart, 1998) that revealed dental articulation of /s/ and /z/ by French-speaking (42.1%, p. 48) and English-speaking (22.8%, p. 50) adults with no obvious speech, language or hearing impairments. Moreover, Van Borsel et al. (2007) concede that no definitive interpretation of their findings exists. They speculate, however, that they might reflect increased social tolerance to imprecision in articulation, or to the influence, on Dutch pronunciation, of English in which /θ/ and /s/ and their voiced cognates are phonemic.
How might these authors have incorporated insights from sociophonetics into their study? First, by examining more incisively the distribution of variants relative to actual or perceived social categories, especially in view of their intriguing finding that these categories differed as a function of university course. Were the students marking their affiliation to humanities or the sciences with distinctive patterns of phonetic variation? Then, analysis of listener perceptions of the participants’ /s/ production might have yielded surprising insights. For example, the dental sound might have been associated more strongly with affiliation with a particular social group than with a judgement that the person produced speech less accurately. It is interesting to speculate how such a finding might help explain why the variant is present. Consider, for example, that this research took place in Belgium, where many languages, including Belgian French and Flemish (the Belgian variant of Dutch), are spoken. As shown by Dart, French has a higher rate of dental fricative productions than English. Perhaps the higher use of dental fricative in certain groups relates to their exposure to or social identification with the French-speaking population in that country. That, of course, is mere speculation on this author’s part, but it shows how sociophonetic methods could have been used to flesh out Van Borsel et al.’s findings. If, in the analysis, these variants actually indicated pathology , then it might be reasonable to suggest that Belgian logopedists consider treating them more aggressively in children and adolescents.
But if these are indeed normal sociophonetic variants, then they do not warrant treatment in the traditional sense, although they might legitimately be the subjects of a regimen to increase talkers’ linguistic flexibility. That is, SLPs/SLTs should not be blind to the fact that non-pathological variation may be associated with negative judgements by some listeners, especially where they index membership in a group that is itself stigmatised. In this regard Van Borsel et al. (2007) cite references supporting an argument that frontal lisping can be associated with negative evaluative judgements.
An individual’s communicative effectiveness, broadly speaking, resides in part on their ability to fluently switch among different phonetic variants in socially appropriate contexts, and SLPs/SLTs are best positioned, in terms of their knowledge and skill bases, to help people who find this problematic. But it must be emphasised that a population that speaks a non-standard variant is not a disordered population, and their presenting ‘condition’ is not a disorder. By carefully assessing whether productions are deviant, as opposed to normal, socially stratified variants, SLPs/SLTs can ensure that they do not improperly treat normal variation as pathology.
SLPs/SLTs should also be aware that a variant perceived negatively in one context or by one group might be perceived positively by another group or in another context. The association between /s/ and men’s sexuality in the many English-speaking countries is a case in point. While this variant is associated with both actual and perceived sexual orientation, it is also associated with hyperarticulate speech. A man whose habitual /s/ demonstrates these characteristics would be ill-advised to change his /s/ characteristics in all communicative contexts, as doing so would prevent him from projecting the positive characteristics that are associated with clear-sounding speech.
In the industrialised world at least, an Audiologist also assesses almost every child with SSD who is assessed by an SLP/SLT. Apart from the resultant audiogram and tympanogram being read and carefully filed, there is often little overt appreciation of this essential input. SLP/SLT clinicians who are not dually qualified in Audiology may have a poor grasp of hearing issues relative to this population. For one significant example, the high incidence of conductive hearing loss in children with cleft palate is well known, but the generalist SLP/SLT may not know how hearing acuity should be monitored.
Professor Suzanne Purdy heads Speech Science in the School of Psychology at the University of Auckland. Her clinical background is in audiology and she has received service awards from the Australian and New Zealand audiological societies. She completed her PhD in Speech Pathology and Audiology at the University of Iowa in 1990 and has published widely in the area of communication disorders, with more than 100 published articles and book chapters. She has a particular interest in hearing loss, speech perception and auditory processing in children.
Ms. Liz Fairgray is a Speech-Language Therapist who has practised in the area of paediatric communication difficulties since 1985. Her MSc is in speech pathology and audiology. Liz has a strong interest in oral communication for children with moderate to profound hearing loss and in 2001, was the first New Zealander to become a Certified Auditory Verbal Therapist. Liz was the founding therapist for The Hearing House, a centre for oral communication for children with profound hearing loss and cochlear implants (CIs). Liz joined The University of Auckland’s Listening and Language Clinic, in 2007, seeing children with a wide range of communication difficulties. She uses a family centred approach, ensuring that parents attend sessions and collaboratively develop goals and practice strategies to develop spoken communication. Liz also provides supervision to SLT students and gives lectures to masters level SLT and Audiology students.
Ms. Areej Asad is a PhD candidate in Speech Science at The University of Auckland. She previously worked as a clinician and clinical tutor in the Center for Phonetics Research, Speech and Hearing Clinic at the University of Jordan. Her research interests are in narrative language, speech outcomes of monolingual and bilingual children with hearing loss and evidence-based speech therapy.
Q52. Suzanne C. Purdy, Liz Fairgray and Areej Asad: Audiology and speech pathology What would an Audiologist like to be able to tell SLPs/SLTs working with child speech regarding the speech spectrum and audibility of speech sounds with different types of hearing loss? Are there particular screening, referral and management consideration to be taken into account with indigenous, low SES, culturally and linguistically diverse, and other special populations; what are the research needs and directions; what communication and collaboration would you like to see between SLPs/SLTs and Audiologists; and are there any good news stories?
A52. Suzanne C. Purdy, Liz Fairgray and Areej Asad: Hearing and children’s speech sound disorders Hearing loss degrades perceived acoustic characteristics of speech, and speech perception and speech production are closely linked. In pre-schoolers, more severe hearing loss is associated with more severe SSD (Schonweiler, Ptok & Radu, 1998). Infants with hearing loss have delayed onset of consistent canonical babbling and delayed consonant development (Moeller et al., 2007). Moeller, et al. found that infants with hearing loss, aged 10–24 months produced fewer fricatives and affricates [f, v, θ, ð, s, z, ʃ, ʒ, tʃ, ʤ] than controls. There was variability in speech development for children with similar hearing losses, so the audiogram does not completely explain speech outcomes for children with hearing loss.
