A procedure for phonetic transcription by consensus

Methodological questions in studying consonant acquisition

Consonant mastery is one of the most widely used metrics of typical phonological acquisition and of phonological disorder. Two fundamental methodological questions concerning research on consonant acquisition are (1) how to elicit a representative sample of productions and (2) how to analyse this sample once it has been collected. This paper address these two questions by reviewing relevant aspects of experience in evaluating word-initial consonant accuracy from transcriptions of isolated-word productions elicited from 2- and 3-year-olds learning four different first languages representing a telling range of consonant systems (English, Cantonese, Greek, Japanese). It is suggested that both researchers and clinicians should consider a number of different item-related factors, such as phonotactic probability and word length, when constructing word lists to elicit consonant productions from young children. This study also proposes that transcription should be supplemented by acoustic analysis and the perceptual judgements of naive listeners.

Clicks produced as compensatory articulations in two adolescents with velocardiofacial syndrome

OBJECTIVE

To report perceptual and dynamic articulatory electropalatography data on clicks produced as compensatory articulations by two adolescents (S1 and S2) with velocardiofacial syndrome and velopharyngeal dysfunction.

RESULTS

The perceptual analysis showed that both speakers produced click substitutions for English targets /t/, /d/, /k/, /g/, and S2 additionally produced clicks for /p/, /b/, and /tintegral/, //. The adolescents produced a range of clicks, which varied in placement (bilabial, dental, alveolar, palatal), voicing (voiced, voiceless), and nasality (nasal, nonnasal). Measurements from the electropalatography data for lingual clicks revealed two articulatory closures, one in the anterior and another in the posterior region of the hard palate. The data revealed how the two closures were timed precisely to produce the click sound. The clicks involved a complex and highly coordinated sequence of tongue maneuvers similar to clicks in some southern African languages.

CONCLUSIONS

Clicks are interpreted as compensatory articulations enabling some speakers with velopharyngeal dysfunction to produce plosive and affricate sounds with perceptually salient acoustic bursts in the oral cavity. Clicks as compensatory articulations merit further systematic investigation, and the procedures reported in this study are considered appropriate for such research.

Phonological processing skills of adolescents with residual speech sound errors

PURPOSE

Research has shown that young children with speech sound disorders may have weaknesses in phonological processing. However, such skills have not been thoroughly examined in adolescents with residual speech sound errors. Therefore, this study compared the phonological processing abilities of adolescents with residual speech sound errors to those of normally speaking peers.

METHOD

Two nonword repetition tasks, multisyllabic word repetition, spoonerisms, phoneme reversals, and an elision task were used to compare the phonological processing skills of 10-14-year-olds with residual speech sound errors that include rhotic phonemes (RE, n = 13) to those of normally speaking (NS, n = 14) adolescents of similar age and receptive vocabulary abilities.

RESULTS

The 2 groups were found to differ on 5 of the 6 phonological processing tasks. Discriminant analysis showed that 85% of the participants could be correctly classified into the RE and NS groups based solely on phonological processing skills.

CONCLUSION

The possible nature of the phonological processing impairment is discussed in the context of current theoretical understanding. It is recommended that when planning assessment and intervention for adolescents with residual speech sound errors, clinicians be cognizant of the fact that the adolescents may also have weaknesses in phonological processing.

A progressive consonant-substitution pattern in a typically developing child

The purpose of this paper was to describe and explain an unusual consonant substitution (default) pattern in the early lexical repertoire of a typically developing boy. Ben was audiotape recorded monthly from age 2;2-2;8 (years; months), to follow an unusual substitution pattern, initially observed between 1;6 and 1;7 (diary notations). Data from each recording period were analysed with reference to (a) consonant repertoire, and (b) default and non-default features present. Data are presented for three periods of [l] substitution: initially, for a glide /w/ (age 1;6-1;7), next, as a substitution for word-initial fricatives /s/, /z/ and /[Formula: see text]/ (age 2;2-2;3), and finally, in place of /r/ in consonant + /r/ clusters, in place of /j/ in initial position, and as a between-word liaison (age 2;7-2;8). It is shown that preferences for specific features, consonant vowel interactions, and syllable positions may be implicated in motivating this [l] substitution at specific points over time. The observed pattern is explained with reference to universal and child-specific factors and adds further evidence for variability that exists among typically developing children.

Universal Parameters for Reporting Speech Outcomes in Individuals with Cleft Palate

Objective:

To achieve consistency and uniformity in reporting speech outcomes in individuals born with cleft palate with or without cleft lip using perceptual parameters that characterize their speech production behavior regardless of the language or languages spoken.

Design:

A working group of six individuals experienced in speech and cleft palate was formed to develop a system of universal parameters for reporting speech outcomes in individuals born with cleft palate. The system was adopted in conjunction with a workshop held in Washington, D.C., that was devoted to developing the universal system. The system, which was refined further following the workshop, involves a three-stage plan consisting of (1) evaluation, (2) mapping, and (3) reporting. The current report focuses primarily on the third stage, reporting speech outcomes.

Results:

A set of five universal speech parameters has been devised for the reporting stage. These consist of (1) hypernasality, (2) hyponasality, (3) audible nasal air emission and/or nasal turbulence, (4) consonant production errors, and (5) voice disorder. Also included are speech understandability and speech acceptability, global parameters that can be reported for any type of speech disorder. The parameters are described in detail, and guidelines for speech-sampling content and scoring procedures in relation to the parameters are presented.

Conclusion:

A plan has been developed to document speech outcomes in individuals with cleft palate, regardless of the spoken language, using a set of five universal reporting parameters and two global speech parameters.

Vocalizations of infants with hearing loss compared with infants with normal hearing: Part I--phonetic development

OBJECTIVE

Infants with hearing loss are known to be slower to develop spoken vocabulary than peers with normal hearing. Previous research demonstrates that they differ from normal-hearing children in several aspects of prelinguistic vocal development. Less is known about the vocalizations of early-identified infants with access to current hearing technologies. This longitudinal study documents changes in prelinguistic vocalizations in early-identified infants with varying degrees of hearing loss, compared with a group of infants with normal hearing. It was hypothesized that infants with hearing loss would demonstrate phonetic delays and that selected aspects of phonetic learning may be differentially affected by restricted auditory access.

DESIGN

The vocalizations and early verbalizations of 21 infants with normal hearing and 12 early-identified infants with hearing loss were compared over a period of 14 mo (from 10 to 24 mo of age). Thirty-minute mother-child interaction sessions were video recorded at 6- to 8-wk intervals in a laboratory playroom setting. Syllable complexity changes and consonantal development were quantified from vocalizations and early verbalizations. Early behaviors were related to speech production measures at 36 mo of age. Participants with hearing loss were recruited from local audiology clinics and early intervention programs. Participants with normal hearing were recruited through day care centers and pediatrician offices.

RESULTS

Relative to age-matched, normal-hearing peers, children with hearing loss were delayed in the onset of consistent canonical babble. However, certain children with moderately-severe losses babbled on time, and infants with cochlear implants babbled within 2 to 6 mo of implantation. The infants with hearing loss had smaller consonantal inventories and were slower to increase syllable shape complexity than age-matched normal-hearing peers. The overall pattern of results suggested that consonant development in infants with hearing loss was delayed but not qualitatively different from children with normal hearing. Delays appeared to be less pronounced than suggested by previous research. However, fricative/affricate development progressed slowly in infants with hearing loss and divergence from the patterns of normal-hearing children was observed. Six children (2 with normal hearing; 4 with hearing loss) were identified as atypical, based on their rates of development. At 24 mo of age, these children persisted in producing a high proportion (0.59) of vocalizations lacking consonants, which was negatively correlated with Goldman-Fristoe scores at 36 mo (r = -0.60).

CONCLUSIONS

Results suggest that early-identified children are delayed in consonant and syllable structure development, which may influence early word learning rates. Fricative/affricate development appears to be challenging for some infants with hearing loss. This may be related to the effects of sensorineural hearing loss on high-frequency information, restricted bandwidth provided by amplification, and reduced audibility in contexts of noise and reverberation. Delayed fricative use may have implications for morphological development. Atypically slow rates of change in syllable development may indicate that a child is at risk for delayed speech development.

Vocalizations of Infants with Hearing Loss Compared with Infants with Normal Hearing: Part II – Transition to Words

OBJECTIVE

By 24 mo of age, most typically developing infants with normal hearing successfully transition to the production of words that can be understood about 50% of the time. This study compares early phonological development in children with and without hearing loss to gain a clearer understanding of the effects of hearing loss in early-identified children. A secondary goal was to identify measures of early phonetic development that are predictors of later speech production outcomes.

DESIGN

The vocalizations and early words of 21 infants with normal hearing and 12 early-identified infants with hearing loss were followed longitudinally over a period of 14 mo (from 10 to 24 mo of age). Thirty-minute mother-child interaction samples were video recorded at 6- to 8-wk intervals in a laboratory playroom setting. Vocalizations produced at 16 and 24 mo were categorized according to communicative intent and recognizable words versus other types. Groups were compared on the structural complexity of words produced at 24 mo of age. Parent report measures of vocabulary development were collected from 10 to 30 mo of age, and Goldman-Fristoe Test of Articulation scores at 36 mo were used in regression analyses.

RESULTS

Both groups increased the purposeful use of voice between 16 and 24 mo of age. However, at 24 mo of age, the toddlers with hearing loss produced significantly fewer words that could be recognized by their mothers. Their samples were dominated by unintelligible communicative attempts at this age. In contrast, the samples from normal hearing children were dominated by words and phrases. At 24 mo of age, toddlers with normal hearing were more advanced than those with hearing loss on seven measures of the structural complexity of words. The children with normal hearing attempted more complex words and productions were more accurate than those of children with hearing loss. At 10 to 16 mo of age, the groups did not differ significantly on parent-report measures of receptive vocabulary. However, the hearing loss group was much slower to develop expressive vocabulary and demonstrated larger individual differences than the normal hearing group. Six children identified as atypical differed from all other children in vowel accuracy and complexity of word attempts. However, both atypical infants and typical infants with hearing loss were significantly less accurate than normal hearing infants in consonant and word production. Early measures of syllable production predicted unique variance in later speech production and vocabulary outcomes.

CONCLUSIONS

The transition from babble to words in infants with hearing loss appears to be delayed but parallel to that of infants with normal hearing. These delays appear to exert significant influences on expressive vocabulary development. Parents may appreciate knowing that some children with hearing loss may develop early vocabulary at a slower rate than children with normal hearing. Clinicians should monitor landmarks from babble onset through transitions to words. Indicators of atypical development were delayed and/or limited use of syllables with consonants, vowel errors and limited production of recognizable words.

Consonant Profile of Arabic-Speaking School-Age Children in Jordan

The paper provides a detailed consonant profile of 100 Jordanian children at the onset of formal schooling. The data were elicited through a modified version of Amayreh's (1994) articulation test. The findings showed that all consonants of Jordanian Spoken Arabic were acquired. The six consonants that were not acquired in Standard Arabic have dialectal forms. The accuracy rates of these consonants were discussed in the light of frequency of occurrence of consonants and diglossia. The study recommended an earlier targeting of consonants that have dialectal variants. It also suggested that knowledge of the diglossic nature of Arabic is important for literacy development as well as for diagnosis and treatment of articulation problems.

Phonological skills and disfluency levels in preschool children who stutter

The relation between stuttering and aspects of language, including phonology, has been investigated for many years. Whereas past literature reported that the incidence of phonological difficulties is higher for children who stutter when compared to normally fluent children, the suggestion of association between the two disorders also drew several critical evaluations. Nevertheless, only a limited amount of information exists concerning the manner and extent to which the speech sound errors exhibited by young children who stutter, close to stuttering onset, is related to the characteristics of their stuttering, such as its severity. Conversely, information is limited regarding the effects a child's phonological skills may have on his/her stuttering severity. The current study investigated the mutual relations between these two factors in 28 carefully selected preschool children near the onset of their stuttering. The children, 20 boys and 8 girls, ranged in age from 25 to 38 months, with a mean of 32.2 months. The phonological skills of two groups with different ratings of stuttering were compared. Similarly, the stuttering severities of two groups with different levels of phonological skills (minimal deviations-moderate deviations) were compared. No statistically significant differences were found for either of the two factors. Inspection of the data revealed interesting individual differences.

LEARNING OUTCOMES

The reader will be able to list: (1) differences in the phonological skills of preschool children whose stuttering is severe as compared to children whose stuttering is mild and (2) differences in stuttering severity in preschool children with minimal phonological deviations as compared to children with moderate phonological deviations.

Reduplication revisited: functions, constraints, repairs, and clinical implications

This case study considers the phonological forms of early lexical items produced by 1 normally developing boy, from 19 to 22 months of age, who began to produce all monosyllabic words as bisyllabic. In order to link this empirical data (the apparent creation of increased complexity) with universal tendencies (motivated by the reduction of complexity), the functions of reduplication were revisited. Phonological processes (i.e., reduplication and final consonant deletion) are viewed as repairs motivated by 2 interacting constraints (i.e., constraints on monosyllabic words and on word-final consonants). These longitudinal case study data provide further evidence for a relationship between final consonant deletion and reduplication. A possible treatment approach for similar patterns demonstrated clinically is recommended.

The influence of multiple presentations on judgments of children's phonetic accuracy

Two experiments examined whether listening to multiple presentations of recorded speech stimuli influences the reliability and accuracy of judgments of children's speech production accuracy. In Experiment 1, 10 listeners phonetically transcribed words produced by children with phonological impairments after a single presentation and after the word was played 7 times. Inter- and intratranscriber reliability in the single- and multiple-presentation conditions did not differ significantly. In Experiment 2, 18 listeners provided binary correct/incorrect judgments of /s/ accuracy in single- and multiple-presentation conditions. There was no systematic effect of presentation condition on either accuracy or intrarater reliability. However, greater interrater reliability was noted in the multiple-presentation condition, particularly for tokens of /s/ that were incorrect or acoustically intermediate between an incorrect and a correct /s/. Taken together, the results suggest that multiple presentations have no measurable effect on the accuracy and intrarater reliability of judgments of children's phonetic accuracy, but that they do have a small effect on interrater reliability. Clinical implications are discussed.

Compensatory articulatory behaviours in adolescents with cleft palate: comparing the perceptual and instrumental evidence

A combination of perceptual and electropalatographic (EPG) analysis is used to investigate speech production in three adolescent speakers with a history of cleft palate. All the subjects still sound markedly atypical. Their speech output is analysed in three conditions: diadochokinetic tasks; single word production; connected speech. Comparison of the EPG and perceptual data reveals important lingual behaviours which were not identified by perceptual analysis alone. Lingualpalatal contact patterns previously noted in the speech of younger speakers with cleft palate are identified, as well as some lingual behaviours not previously reported in the literature. Similarities and differences between the individual speakers' profiles are discussed, as is the considerable inter- and intra-speaker variability identified in the data, and the speech output is discussed with reference to the notion of compensatory articulation.

Risk factors for speech delay of unknown origin in 3-year-old children

One hundred 3-year-olds with speech delay of unknown origin and 539 same-age peers were compared with respect to 6 variables linked to speech disorders: male sex, family history of developmental communication disorder, low maternal education, low socioeconomic status (indexed by Medicaid health insurance), African American race, and prolonged otitis media. Abnormal hearing was also examined in a subset of 279 children who had at least 2 hearing evaluations between 6 and 18 months of age. Significant odds ratios were found only for low maternal education, male sex, and positive family history; a child with all 3 factors was 7.71 times as likely to have a speech delay as a child without any of these factors.

Cochlear implants and infants: expectations and outcomes

This study followed the communication and motor development of 3 profoundly deaf infants who underwent cochlear implantation. All infants were receiving auditory habilitation at the Sydney Cochlear Implant Centre, Australia, with the view to acquiring oral language skills and had English-speaking, hearing parents. Their progress was tracked with a battery of measurement tools before and after cochlear implantation. The results indicated noticeable growth in receptive language and motor skills after operation. Although expressive language improved, the rate of growth was much slower than changes in other areas. Although some areas of the subjects' communication skills remained delayed, the importance of tracking their progress according to their hearing age, as well as the chronological age, was evident.

Longitudinal changes in articulation rate and phonetic phrase length in children with speech delay

This study examined long-term changes in articulation rate (the pace at which speech segments are produced) and phonetic phrase length in the conversational speech of two groups of children with speech delay (SD) of unknown origin. Initial testing for both groups occurred at preschool age, with follow-up testing conducted for the Early Follow-Up Group (n = 17) at age 9 years and for the Late Follow-Up Group (n = 36) at age 12-16 years. At follow-up testing both groups produced significantly faster articulation rates (measured in both syllables per second and phones per second) and significantly longer phonetic phrases (measured in both syllables and phones) than at initial testing. Articulation rates at both test times were also judged to be similar to published values from typically developing children of similar ages when measured in syllables per second. However, findings for rate in phones per second suggested that at least at initial testing the children were articulating speech at a slower rate than their typically developing peers. This latter finding, however, may have been an artifact of the high frequency of errors--such as cluster reduction and final consonant deletion--observed in the initial samples. It would appear, therefore, that children with SD of unknown origin may start out with slower than normal articulation rates but eventually catch up to their typically developing peers.

Obtaining phonetic transcriptions: a comparison between expert listeners and a continuous speech recognizer

In this article, we address the issue of using a continuous speech recognition tool to obtain phonetic or phonological representations of speech. Two experiments were carried out in which the performance of a continuous speech recognizer (CSR) was compared to the performance of expert listeners in a task of judging whether a number of prespecified phones had been realized in an utterance. In the first experiment, nine expert listeners and the CSR carried out exactly the same task: deciding whether a segment was present or not in 467 cases. In the second experiment, we expanded on the first experiment by focusing on two phonological processes: schwa-deletion and schwa-insertion. The results of these experiments show that significant differences in performance were found between the CSR and the listeners, but also between individual listeners. Although some of these differences appeared to be statistically significant, their magnitude is such that they may very well be acceptable depending on what the transcriptions are needed for. In other words, although the CSR is not infallible, it makes it possible to explore large datasets, which might outweigh the errors introduced by the mistakes the CSR makes. For these reasons, we can conclude that the CSR can be used instead of a listener to carry out this type of task: deciding whether a phone is present or not.

Individual and consensus judgments of disfluency types in the speech of persons who stutter

Previous research has suggested that the reliability with which judges identify individual disfluency types, such as repetitions or prolongations of speech sounds, may be very poor. The use of disfluency types judgments in research and clinical applications is also complicated by important differences among the several disfluency-based characterizations of stuttered speech. In an attempt to address these problems, this study arranged for 30 judges to identify all disfluency types that they perceived to be present in 5-s audiovisually recorded speech stimuli, each in an Individual task and then with a partner in a Consensus task. Intrapair agreement and interpair agreement for occurrences of disfluency types (from Consensus conditions) were significantly higher than intrajudge and interjudge agreement for occurrences (from Individual conditions). Despite being higher than individual values, however, intrapair and interpair agreement for occurrences both averaged less than 50%. Results also showed that disfluency types judgments, interpreted in terms of three common disfluency-based definitions of stuttering, were not strongly related to previous assessments of whether these speech tokens contained or did not contain stuttering. When combined with previously available data, the present findings suggest caution in the use of disfluency types to describe or define stuttered speech.

Language, speech sound production, and cognition in three-year-old children in relation to otitis media in their first three years of life

OBJECTIVE

As part of a prospective study of possible effects of early-life otitis media on speech, language, cognitive, and psychosocial development, we tested relationships between children's cumulative duration of middle ear effusion (MEE) in their first 3 years of life and their scores on measures of language, speech sound production, and cognition at 3 years of age.

METHODS

We enrolled 6350 healthy infants by 2 months of age who presented for primary care at 1 of 2 urban hospitals or 1 of 2 small town/rural and 4 suburban private pediatric practices. We intensively monitored the children's middle ear status by pneumatic otoscopy, supplemented by tympanometry, throughout their first 3 years of life; we monitored the validity of the otoscopic observations on an ongoing basis; and we treated children for otitis media according to specified guidelines. Children who met specified minimum criteria regarding the persistence of MEE became eligible for a clinical trial in which they were assigned randomly to undergo tympanostomy tube placement either promptly or after a defined extended period if MEE remained present. From among those remaining, we selected randomly, within sociodemographic strata, a sample of 241 children who represented a spectrum of MEE experience from having no MEE to having MEE whose cumulative duration fell just short of meeting randomization criteria. In subjects so selected, the estimated duration of MEE ranged from none to 65.6% of the first year of life and 44.8% of the first 3 years of life. In these 241 children we assessed language development, speech sound production, and cognition at 3 years of age, using both formal tests and conversational samples.

RESULTS

We found weak to moderate, statistically significant negative correlations between children's cumulative durations of MEE in their first year of life or in age periods that included their first year of life, and their scores on formal tests of receptive vocabulary and verbal aspects of cognition at 3 years of age. However, the percent of variance in these scores explained by time with MEE in the first year of life beyond that explained by sociodemographic variables ranged only from 1.2% to 2.9%, and the negative correlations were concentrated in the subgroup of children whose families had private health insurance (rather than Medicaid). We found no significant correlations in the study population as a whole or in any subgroup between time with MEE during antecedent periods and children's scores on measures of spontaneous expressive language, speech sound production, or other measured aspects of cognition. In contrast, by wide margins, scores on all measures were consistently highest among the most socioeconomically advantaged children and lowest among the most socioeconomically disadvantaged children.

CONCLUSIONS

Our findings suggest either that persistent early-life MEE actually causes later small, circumscribed impairments of receptive language and verbal aspects of cognition in certain groups of children or that unidentified, confounding factors predispose children both to early-life otitis media and to certain types of developmental impairment. Findings in the randomized clinical trial component of the larger study should help distinguish between causality and confounding as explanations for our findings.language, speech, cognition, development, otitis media, middle ear effusion.

Risk for speech disorder associated with early recurrent otitis media with effusion: two retrospective studies

The goals of this two-part series on children with histories of early recurrent otitis media with effusion (OME) were to assess the risk for speech disorder with and without hearing loss and to develop a preliminary descriptive-explanatory model for the findings. Recently available speech analysis programs, lifespan reference data, and statistical techniques were implemented with three cohorts of children with OME and their controls originally assessed in the 1980s: 35 typically developing 3-year-old children followed since infancy in a university-affiliated pediatrics clinic, 50 typically developing children of Native American background followed since infancy in a tribal health clinic, and (in the second paper) 70 children followed prospectively from 2 months of age to 3 years of age and older. Dependent variables included information from a suite of 10 metrics of speech production (Shriberg, Austin, Lewis, McSweeny, & Wilson, 1997a, 1 997b). Constraints on available sociodemographic and hearing status information limit generalizations from the comparative findings for each database, particularly data from the two retrospective studies. The present paper reports findings from risk analysis of conversational speech data from the first two cohorts, each of which included retrospective study of children for whom data on hearing loss were not available. Early recurrent OME was not associated with increased risk for speech disorder in the pediatrics sample but was associated with approximately 4.6 (CI = 1.10-20.20) increased risk for subclinical or clinical speech disorder in the children of Native American background. Discussion underscores the appropriateness of multifactorial risk models for this subtype of child speech disorder.

Otitis media, fluctuant hearing loss, and speech-language outcomes: a preliminary structural equation model

The goals of this study were to estimate the risk for lowered speech-language outcomes associated with early recurrent otitis media with effusion (OME) with and without hearing loss and to develop a preliminary descriptive-explanatory model for the findings. Three statistical approaches were used to assess associations among OME, hearing loss, and speech-language outcomes. Participants were a subsample of 70 children followed prospectively in the Dallas Cooperative Project on Early Hearing and Language Development (Friel-Patti & Finitzo, 1990). Findings indicated that hearing levels at 12-18 months were significantly associated with speech delay and low language outcomes at 3 years of age. The risk for subclinical or clinical speech delay at 3 years of age was 2% for children with less than 20 dB average hearing levels at 12-18 months and 33% for children with greater than 20 dB average hearing levels at 12-18 months. A structural equation model (Jöreskog & Sörbom, 1993) indicated that the significant and substantial effects of hearing levels at 12-18 months on speech status at 3 years were significantly mediated by language status at 3 years. Discussion includes implications of these findings for alternative speech perception models linking early OME and hearing loss to later speech-language disorder.

Maternal education and measures of early speech and language

The present study was designed to determine whether 4 measures of children's spontaneous speech and language differed according to the educational level of the children's mothers. Spontaneous language samples from 240 three-year-old children were analyzed to determine mean length of utterance in morphemes (MLUm), number of different words (NDW), total number of words (TNW), and percentage of consonants correct (PCC). A norm-referenced, knowledge-dependent measure of language comprehension, the Peabody Picture Vocabulary Test-Revised (PPVT-R), was also included for purposes of comparison with the spontaneous measures. Three levels of maternal education were compared: less than high school graduate, high school graduate, and college graduate. Trend analyses showed statistically significant linear trends across educational levels for MLUm, NDW, TNW, and PPVT-R; the trend for PCC was not significant. The relationship of maternal education and other sociodemographic variables to measures of children's language should be examined before using such measures to identify children with language disorders.

Early childhood stuttering II: initial status of phonological abilities

Research on the relation between stuttering and phonological/articulation deficits has been reported in the literature over several decades. Yet virtually none of these investigations has taken into account that "children who stutter" includes a large number who spontaneously recover within a few months or years after onset. Thus, little attention has been given to differences between the phonological abilities of children whose stuttering persists and those who recover. This investigation compares these two groups soon after stuttering onset, before it was possible to classify them as members of either group, on a number of phonological characteristics, including mean percentage of error, relative levels of severity of phonological impairment, error on specific phonological patterns, progress in development of key patterns, and the children's strategies for coping with unmastered patterns. The results indicate that the children whose stuttering would be persistent had poorer mean scores on each of our measures than did the children who would recover from stuttering. Both groups, however, showed progression in phonological development that followed the expected order, and they used typical strategies when patterns had not yet been acquired. The persistent group was moving more slowly, however, so phonological development was more delayed than in the children who would recover from stuttering. Our findings support the assumption that most previous studies probably have compared children with persistent stuttering to normally fluent children, and that those who recovered early were not considered differentially.

Effects of time-interval judgement training on real-time measurement of stuttering

The purpose of this study was to investigate whether a previously developed interval-based training program could improve judges' stuttering event judgments. Two groups of judges made real-time stuttering event judgments (computer-mouse button presses) in 3 to 6 trials before the response-contingent judgment training program and in another 3 to 6 trials after training, for recordings of 9 adults who stuttered. Their judgments were analyzed in terms of number of stuttering events, duration of stuttering, and 5-s intervals of speech that could be categorized as judged (or not judged) to contain stuttering. Results showed (a) changes in the amount of stuttering identified by the judges; (b) improved correspondence between the judges' identifications of stuttering events and interval-based standards previously developed from judgments made by experienced, authoritative judges; (c) improved correspondence between interval-based analyses of the judges' stuttering judgments and the previously developed standards; (d) improved intrajudge agreement; (e) improved interjudge agreement; and (f) convergence between the 2 judge groups, for samples and speakers used during training tasks and also for other speakers. Some implications of these findings for developing standardized procedures for the real-time measurement of stuttering are discussed.

Probability estimates and paths to consonant normalization in children with speech delay

Probable ages of normalization were calculated for 24 children with speech delay, using Kaplan-Meier analysis and a threshold score of 85% on the Percent of Consonants Correct, Percent of Consonants Correct-Adjusted, and Percent of Consonants Correct-Revised metrics. Simple formulas are provided that permit calculation of the likelihood that individual children with speech delay will normalize by a given age. The sex of a child was found to have no significant influence on age of normalization. Analysis revealed two different paths to normalization. In Path A, errors of deletion, substitution, and omission declined as correct productions increased. In Path B, common clinical distortions increased as deletions and substitutions decreased. The ages of more and less rapid phonological gain correspond to and partially explain the findings of Shriberg, Gruber, and Kwiatkowski (1994), who studied more severely involved children. Children who follow Path B are those who retain residual errors in their speech.

The acquisition of Arabic consonants

This normative study of the acquisition of consonants of Arabic as spoken in Jordan answered 4 questions: (1) What percentage of children at each of 9 age levels produced each consonant correctly? (2) What are the ages of customary production, mastery, and acquisition for each phoneme? (3) Does accuracy of consonants within sound classes vary by position in the word? (4) What are the differences in ages of acquisition between Arabic and English? Samples were collected from 180 normally developing children between the ages of 2:0 and 6:4. The percentages of accuracy of both standard and acceptable consonants were plotted and showed clear developmental trends. Medial consonants were significantly more accurate than initial and final consonants. The ages of customary production, acquisition, and mastery of Arabic consonants were similar to those for English but with notable exceptions that have implications for description of phonological acquisition. Support for previously proposed universal sound acquisition sequences was found, but some language-specific effects were also seen.

The percentage of consonants correct (PCC) metric: extensions and reliability data

Research in normal and disordered phonology requires measures of speech production that are biolinguistically appropriate and psychometrically robust. Their conceptual and numeric properties must be well characterized, particularly because speech measures are increasingly appearing in large-scale epidemiologic, genetic, and other descriptive-explanatory database studies. This work provides a rationale for extensions to an articulation competence metric titled the Percentage of Consonants Correct [PCC; Shriberg & Kwiatkowski, 1982; Shriberg, Kwiatkowski, Best, Hengst, & Terselic-Weber, 1986], which is computed from a 5- to 10-minute conversational speech sample. Reliability and standard error of measurement estimates are provided for 9 of a set of 10 speech metric including the PCC. Discussion includes rationale for selecting one or more of the 10 metrics for specific clinical and research needs.

The speech disorders classification system (SDCS): extensions and lifespan reference data

A companion paper includes rationale for the use of 10 metrics of articulation competence in conversational speech (Shriberg, Austin Lewis, McSweeny, & Wilson, 1997). The present paper reports lifespan reference data for these measures using records from a total of 836 3- to 40(+)-year-old speakers with normal and disordered speech. The reference data are subdivided by diagnostic classification based on extensions to an instrument titled the Speech Disorders Classification System (SDCS; Shriberg, 1993). Appendices provide procedural information on the SDCS and statistical rationale for the reference data.

Developmental apraxia of speech: II. Toward a diagnostic marker

This second paper in a series on developmental apraxia of speech (DAS) (Shriberg, Aram, & Kwiatkowski, 1997a) reports findings from two studies. Study I compares speech and prosody-voice profiles of a group of 14 children with suspected DAS to profiles of 73 children with speech delay (SD). Results suggest that the only linguistic domain that differentiates some children with suspected DAS from those with SD is inappropriate stress. Study II cross-validates these findings, using retrospective data from a sample of 20 children with suspected DAS evaluated in a university phonology clinic over a 10-year period. Discussion considers methodological and conceptual issues in the measurement of linguistic stress. Theoretical issues and implications for research and clinical practice are deferred for synthesis of the present findings with those from a multi-site cross-validation project (Shriberg, Aram, & Kwiatkowski, 1997b).

Realizations of consonant clusters by children with phonological impairment

Children with phonological impairment frequently have difficulty producing consonant clusters. Speech pathologists often use phonological processes to describe children's productions of consonant clusters, a commonly used description being cluster reduction. However, this description does not adequately address children's differing realizations of consonant clusters. The purpose of this paper is to develop and refine methods for the characterization of realizations of consonant clusters. The work of Greenlee (1974) and Chin and Dinnsen (1992) has been extended by examining the effect of syllable position, number of elements and constituents on children's realizations of consonant clusters. Specifically, word-initial fricative clusters, stop clusters, three element fricative + stop clusters, and word-final nasal clusters and fricative + stop clusters were examined. The results for 40 phonologically impaired children between the ages of 3;6 and 5;0 years are compared with those of Chin and Dinnsen (1992). The relationships found between child and adult representations of clusters suggest the importance of considering syllable position, number of elements and the constituents of consonant clusters when analysing phonologically impaired children's speech sounds.

Phonological characteristics of children whose stuttering persisted or recovered

Among children who stutter, more will be identified with articulation/phonological deficiencies than among normally fluent children of the same ages. Most current literature has focused on phonological differences between those children who stutter and those who do not. The present study examines early phonological differences between young children whose stuttering persisted and those who recovered from early stuttering. Thirty-six children, 12 whose stuttering persisted, 12 who recovered early, and 12 who recovered later, had been assessed by means of the Assessment of Phonological Processes--Revised (Hodson, 1986) soon after they were identified as exhibiting stuttering. After many months of longitudinal evaluation of their stuttering that led to their classification into the three groups, the early phonological assessments of these children were re-examined to identify differences. Overall mean percentage of error scores as well as error scores on specific phonological patterns showed that the persistent group differed significantly from normally fluent control subjects matched by age and sex. Scores of the two groups who recovered and their matched controls, however, did not differ significantly. Although poor phonological ability in the early stage of stuttering appears to be a contributing factor to the differentiation of persistence and recovery, the wide individual variations in scores within groups suggest that additional factors are necessary for reliable prediction.

Factors associated with speech-sound stimulability

Speech-sound stimulability as a clinical procedure has been used in speech-language pathology for more than 60 years although review of the archival literature on this clinical construct provides little information on what factors may be associated with it. In this study, the factors found to relate to stimulability were articulation visibility, the child's age, the family's socioeconomic status, and the child's overall imitative abilities. Perception, severity, otitis media history, language abilities, consistency of productions, amount of stimulations provided, and willingness to self-correct were not associated with stimulability. Potential explanations are provided on the basis of prior research.

Time-interval measurement of stuttering: establishing and modifying judgment accuracy

The purpose of this study was to determine whether accuracy training for interval judgments of stuttering might generalize to increased accuracy and/or interjudge agreement for intervals other than those used during training. Ten upper-division speech-language pathology students judged 5-s audiovisually recorded speech intervals as stuttered or nonstuttered in a series of group and single-subject experiments. Judgment accuracy was determined with respect to judgments provided previously by 10 recognized authorities on stuttering and its treatment. Training occurred within single-subject experiments that used multiple baselines across speakers and repeated generalization probes to assess training effects. Results showed that judgment accuracy tended to increase after training for speakers used during the training process as well as for unfamiliar speakers. Results also replicated previous findings of slight increases in interjudge and intrajudge agreement after interval-judgment training. The implications of these results for developing a valid and reliable stuttering measurement system are discussed.

Traditional and phonological treatment for teaching English fricatives and affricates to Koreans

Non-native speakers of English are being seen for accent reduction in speech-language pathology clinics although there is little evidence that treatment is effective. Two types of treatment were evaluated. Traditional articulation-based treatment was successful for 2 native Korean speakers in improving English fricatives and affricate production. A second treatment using phonologically based contrasts was also successful for the same sounds with 2 other Korean speakers. Some issues in foreign accent reduction are discussed.

Novel word learning in children with hearing impairment

This study examined novel word-learning abilities in young school-age children with mild-to-moderate hearing losses. We questioned whether degree of hearing loss or measures of language and phonological processing abilities were more likely to be related to novel word-learning ability. Subjects were 20 children with hearing impairment (M = 9:0) and 20 children with normal hearing (M = 6:5) matched for receptive vocabulary knowledge. Children were administered measures of language and phonological processing. The novel word-learning task consisted of an acquisition and retention phase in which children received a series of trials to learn to produce four novel words. Half of the children with hearing impairment performed comparably to the children with normal hearing on all of the measures obtained, whereas the the other 10 children with hearing impairment performed more poorly than the higher functioning children with hearing impairment and all of the children with normal hearing on most of the measures of language, phonological processing, and novel word learning. Degree of hearing loss was not related to language or word-learning abilities. These findings suggest that the population of children with mild-to-moderate hearing loss may contain two distinct groups: a group of normally developing children who have a hearing loss and a group of children with language impairment who have a hearing loss. The implications of this categorization will be discussed.

Developmental phonological disorders. I: A clinical profile

Detailed information on the speech, language, prosody, and voice characteristics of children with developmental phonological disorders is central to diverse research questions. The present study provides a clinical profile of 178 children with developmental phonological disorders. It includes information from prior reports (Shriberg & Kwiatkowski, 1982a; Shriberg, Kwiatkowski, Best, Hengst, & Terselic-Weber, 1986) and from several new measures on a sample of 64 children. The speech, prosody-voice, and causal-correlates profiles for the most recent sample are consistent with prior findings, providing a descriptive profile for forthcoming subgroup research and companion studies addressing short-term (Shriberg, Kwiatkowski, & Gruber, 1994) and long-term (Shriberg, Gruber, & Kwiatkowski, 1994) speech-sound normalization.

Developmental phonological disorders. III: Long-term speech-sound normalization

Prior articles in this series provide a descriptive profile of 178 children with developmental phonological disorders (Shriberg & Kwiatkowski, 1994) and predictive correlates of short-term speech-sound normalization in 54 children (Shriberg, Kwiatkowski, & Gruber, 1994). The present article reports findings from a study of 10 children with developmental phonological disorders whose progress was followed at least once yearly for 7 years. Analyses characterize the sequence, rates, and error patterns of long-term speech-sound normalization in relation to developmental perspectives on the nature of children's phonological disorders. Findings are interpreted to support the hypothesis of a critical period for speech-sound development, with long-term normalization of significant speech delay reaching a chronological age boundary at approximately 8.5 years.

The effect of sampling condition on children's productions of consonant clusters

An investigation was conducted to compare the effects of single word and connected speech sampling conditions on the production of consonant clusters. Speech samples were obtained from 40 children with speech sound impairments who were aged 3 years: 6 months to 5 years. The children's productions of 36 commonly occurring consonant clusters were compared across the two sampling conditions. Overall children's productions were more similar than different. Differences between the sampling conditions were apparent for three of the eight phonological processes studied, namely, cluster reduction, final consonant deletion, and epenthesis. Of 12 fine phonetic variations, only aspirated stops showed a significant difference between the sampling conditions. There was a wide range of individual variation.

The reliability of observational data: I. Theories and methods for speech-language pathology

Much research and clinical work in speech-language pathology depends on the validity and reliability of data gathered through the direct observation of human behavior. This paper reviews several definitions of reliability, concluding that behavior observation data are reliable if they, and the experimental conclusions drawn from them, are not affected by differences among observers or by other variations in the recording context. The theoretical bases of several methods commonly used to estimate reliability for observational data are reviewed, with examples of the use of these methods drawn from a recent volume of the Journal of Speech and Hearing Research (35, 1992). Although most recent research publications in speech-language pathology have addressed the issue of reliability for their observational data to some extent, most reliability estimates do not clearly establish that the data or the experimental conclusions were replicable or unaffected by differences among observers. Suggestions are provided for improving the usefulness of the reliability estimates published in speech-language pathology research.

Four new speech and prosody-voice measures for genetics research and other studies in developmental phonological disorders

Research in developmental phonological disorders, particularly emerging subgroup studies using behavioral and molecular genetics, requires qualitative and continuous measurement systems that meet a variety of substantive and psychometric assumptions. This paper reviews relevant issues underlying such needs and presents four measurement proposals developed expressly for causal-correlates research. The primary qualitative system is the Speech Disorders Classification System (SDCS), a 10-category nosology for dichotomous and hierarchical polychotomous classification of speech disorders from 2 years of age through adulthood. The three quantitative measures for segmental and suprasegmental analyses are (a) the Articulation Competence Index (ACI), an interval-level severity index that adjusts a subject's Percentage of Consonants Correct (PCC) score for the relative percentage of distortion errors; (b) Speech Profiles, a series of graphic-numeric displays that profile a subject's or group's severity-adjusted consonant and vowel-diphthong mastery and error patterns; and (c) the Prosody-Voice Profile, a graphic-numeric display that profiles a subject's or group's status on six suprasegmental domains divided into 31 types of inappropriate prosody-voice codes. All data for the four measures are derived from one sample of conversational speech, which obviates the limitations of citation-form testing; enables speech assessment as a qualitative, semi-continuous, and continuous trait over the life span; and provides a context for univariate and multivariate statistical analyses of phonetic, phonologic, prosodic, and language variables in multiage, multidialectal, and multicultural populations. Rationale, procedures, validity data, and examples of uses for each measure are presented.

Variability in speech dyspraxia

In a volume dedicated to Professor Crystal, this study takes up some remarks of his (1982) regarding definitions of disorders and use of day-to-day clinical data. The topic chosen relates to the notion of variability in speech dyspraxia. Several different uses of the term are noted, but the one used in this paper focuses on variability of errors in a given word on a repeated-trials task. This task was employed to investigate whether, as previously claimed, speech dyspraxics differ from other pronunciation-disordered groups in terms of error variability. Results indicate that as a group they differ from spastic dysarthrics, but not from phonemic paraphasic speakers. On an individual basis, divisions are not as clearcut. Implications for assessment and speech production are considered in the light of this.

Speech and prosody characteristics of adults with mental retardation

Audio-recorded continuous speech samples from forty 20-50-year-old noninstitutionalized persons with mental retardation were selected from a database of 192 samples. Descriptive data on segmental and suprasegmental characteristics were obtained using close phonetic transcription as input to linguistic analyses software. For this sample of adults with mental retardation, speech and prosody status were not statistically associated with gender or gross level of mental retardation, but were associated with estimated probability of independent living. Speech and prosody analyses and content analyses of transcribers' comments yielded diacritic-level profiles of these speakers' linguistic and paralinguistic behaviors in continuous speech. Additional analyses of the error data tested alternative sources of processing involvement within a four-stage speech production model. A cognitive capacity constraint, which limits the speaker's ability to allocate resources to phonological encoding, is proposed as a sufficient explanation for the obtained pattern of token-to-token inconsistency of articulation. An additional sociolinguistic constraint is hypothesized to account for reduced prosodic and paralinguistic competence in continuous discourse. Both constraints are amenable to intervention programming. Findings fail to support the view that the potential for long-term speech prosody competence in individuals with mental retardation is limited by speech-motor constraints. Discussion includes intervention considerations in the context of current trends in special education.

The Iowa Articulation Norms Project and its Nebraska replication

The purpose of the Iowa Articulation Norms Project and its Nebraska replication was to provide normative information about speech sound acquisition in these two states. An assessment instrument consisting of photographs and a checklist form for narrow phonetic transcription was administered by school-based speech-language pathologists to stratified samples of children in the age range 3-9 years. The resulting data were not influenced by the demographic variables of population density (rural/urban), SES (based on parental education), or state of residence (Iowa/Nebraska); however, sex of the child exerted a significant influence in some of the preschool age groups. The criteria used to determine acceptability of a production appeared to influence outcomes for some speech sounds. Acquisition curves were plotted for individual phoneme targets or groups of targets. These curves were used to develop recommended ages of acquisition for the tested speech sounds, with recommendations based generally on a 90% level of acquisition. Special considerations were required for the phonemes /ng s z/.

Isolated and sequenced oral motor posture production under different input modalities by left-hemisphere damaged adults

The purpose of this investigation was to use a well-refined system of oral movement notation to examine whether the quality of performances for production of elemental oral motor postures as isolated targets was similar to performances of the same postures produced as members of a sequence by left-hemisphere damaged (LHD) and normal adults. The effect of input modality was also examined using three conditions--imitation, photograph and command. Result suggested that "additional movements" and "errors of spatial alignment" prominently characterized both isolated and sequenced postural production errors made by both LHD and normal adults. The frequency of occurrence of these behaviours, however, was dramatically higher among the LHD subjects. Moreover, the quality of productions of oral postures deteriorated in sequences of increasing length among the LHD subjects but not among normal adults. No differential effect for mode of input was found for the LHD subjects but was for the normal adults. Results suggested that the inferior performances of LHD subjects on isolated gesture production are due to exaggerated augmentations and spatial targeting deficits. Sequence production was affected negatively by the same behaviours as well as a disability for "phasing" subcomponents of discrete movements as reflected by a significant increase in "complex" errors on the individual gestures comprising the sequence.