Phonologically Disordered German-Speaking Children

Research

Phonologically Disordered German-Speaking Children Annette V. Fox Barbara Dodd University of Newcastle upon Tyne, United Kingdom

Speech sound disorders affect more children than any other developmental communication disorder and are associated with longterm social and academic difficulties. The diversity of presenting symptoms has resulted in the need for classifying subgroups of speech disorders. Research on English-speaking children suggests that there are four types of surface speech error patterns (B. Dodd, 1995): articulation disorder (e.g., lisp); delay (i.e., normal developmental patterns that are inappropriate for chronological age); consistent use of atypical error patterns (e.g., deletion of all initial consonants); and inconsistent pronunciation of the same lexical items. Classification typologies should be language independent. This study investigated whether the same four subgroups, in similar proportions, would be found in German-speaking children who had

B

etween 3% and 10% of 4-year-old children have disordered speech (National Institute on Deafness and other Communication Disorders, 1994; Gierut, 1998). Children with speech sound disorders constitute 70% of patients referred for treatment by pediatric speechlanguage pathologists (Weiss, Gordon, & Lillywhite, 1987), and ASHA reports phonological disorders to be the most prevalent pediatric communication disability (www.asha.org, 2001). These children often produce so many pronunciation errors that their speech is difficult or impossible to understand. These intelligibility problems can lead to difficulties in forming social relationships and to low academic achievement. Difficulties in producing speech can be persistent; the consequences are often observable in adulthood (Lewis & Fairbairn, 1992). Mispronunciations by these children do not reflect young children’s typical acquisition of speech in terms of rate of acquisition or types of errors. Most importantly, children with speech disorders are a heterogeneous group, differing in types of surface speech error patterns (Dodd & McCormack, 1995), severity (Garrett & Moran, 1992), etiology (Shriberg,

disordered speech. A total of 110 monolingual German-speaking children, aged 2 years 7 months to 7 years 7 months, participated in the study. They had been referred for assessment of a suspected speech disorder. The results supported the subgroup classification, providing evidence for the universal character of speech disorders. One significant difference was the relatively high proportion of children classified as having an articulation disorder. This was explained by the uncertainty regarding a lisp as a disorder in German, since it is also found in up to 40% of normally developing children of the same age. The theoretical and clinical implications of the findings are discussed. Key Words: phonological disorders, speech disorders, German, classification system, crosslinguistics

1982), and other associated language difficulties. In addition, different children respond to different intervention approaches (Dodd & Bradford, 2000). Because no single treatment technique is appropriate for such a diverse population, there is a need for research to evaluate methods of differentially diagnosing subgroups of children with speech disorders. The development of costeffective intervention programs is dependent on the validation of classification systems for speech disorder (Dodd, 1993; Shriberg, 1982). Three classification methods have been predominant.

Medical-Etiological Approach Medical diagnoses associated with speech and language disorders include cerebral palsy, hearing loss, cleft palate, and a number of syndromes such as Down syndrome (Crystal & Varley, 1993). Although these conditions play an important role in differential diagnosis, most children with speech disorders do not present with a medical history of this kind. Shriberg (1994) proposed a classification

American Journal of Speech-Language Pathology • Vol. 10 • 291–307 • August 2001 • © American Speech-Language-Hearing Association

Fox & Dodd: Differential Diagnosis of Phonological Disorders

1058-0360/01/1003-0291

291

system based on etiological factors that are often less obvious: positive family history of speech, language, and/ or literacy difficulties; frequent middle ear infections; developmental apraxia of speech; psychological involvement; and residual errors (isolated phone-distortions). Unfortunately parents often report either none or more than one of these conditions. Therefore, groups of children remain unclassifiable. Further, lack of normative data (Shriberg, 1993) raises the question of whether these factors accurately differentiate children with speech disorders from normally developing children.

Clinical-Inferential/Linguistic Perspective Approach A more successful approach to diagnosis has been the clinical-inferential analysis of speech data, including an examination of phonological error patterns. The classification approach relying on severity, the Percentage Consonants Correct measure (PCC; Shriberg & Kwiatkowski, 1982), is included in this category. Other researchers have identified subgroups of children in terms of the characteristics of their surface speech error patterns. Three phonologically disordered subgroups have emerged: delay (Fletcher, 1990), consistent but unusual (nondevelopmental) errors (Leonard, 1985), and inconsistent errors (Dodd & Leahy, 1989). Although assessment procedures should “provide some framework for the identification of different types of disordered pronunciation patterns in children” (Grunwell, 1985, p. 3), they fail to identify the deficit underlying different types of disorder (Stackhouse & Wells, 1997).

Psycholinguistic Approach Cognitive neuropsychological and psycholinguistic models “view a child’s speech problem in terms of breakdowns in aspects of input processing, output processing and internal representation” (Lambert & Waters, 1995, p. 97). These models (e.g., Dodd, 1995; Stackhouse and Wells, 1997) promote hypotheses about the nature of the deficits in perceptual, cognitive-linguistic, and motor skills that result in different types of speech disorders (Holm, 1998). Psycholinguistic approaches to classification have the advantage of identifying a deficit that can be targeted in treatment. A study of 50 Australian-English-speaking children (Bradford & Dodd, 1996) combined the clinical inferential/ linguistic and psycholinguistic approach. This study suggested that speech disorders can be classified into four subgroups in terms of surface error patterns that reflect differing underlying deficits in the speech processing chain: 1. Articulation Disorder: The consistent mispronunciation or distortion of a phoneme in isolation and in all phonetic contexts resulting from an impairment of the “processes involved in the planning and execution of smooth sequences of highly overlapping gestures of the speech organs” (Fey, 1992, p. 225) is classified as an articulation disorder. 2. Delayed Phonological Development: “A classification of delayed phonological acquisition is warranted when 292

all phonological processes derived to describe a child’s speech errors occur during normal development but are typical of a younger chronological age level” (Dodd, 1995, p. 55). A delay of 6 months has been suggested to be significant (Crystal, Fletcher, & Garman, 1989). 3. Deviant-Consistent Phonological Disorder: Children should be classified as having a deviant-consistent disorder if at least one of the error patterns they use consistently is nondevelopmental (i.e., an error pattern not observed during normal phonological development) or occurs on phonemes that are not affected by this pattern in normally developing children. For example, in German, the error pattern final consonant deletion (FCD) occurs in normally developing children only on /t k l/ (Fox & Dodd, 1999). FCD is idiosyncratic if it occurs on any other final consonant. “Most children who make nondevelopmental errors also use some developmental error patterns that may, or may not, be appropriate for their chronological age. They should nevertheless be classified as having a ‘deviantconsistent disorder’, since the presence of unusual error patterns signals an impaired understanding of their native phonological system” (Dodd, 1995, p. 56). 4. Deviant-Inconsistent Phonological Disorder: Children who do not consistently pronounce the same lexical item in the same way in one-word elicited utterances are classified as having an inconsistent phonological disorder. Children are included in this subgroup if their inconsistency rate is greater than 40% on a specific test of the same 25 lexical items produced on three separate trials in one assessment session (Dodd, 1995). This is an arbitrary criterion selected because normally developing children (who have a vocabulary of more than 50 words) show inconsistency of less than 10% (Teitzel & Ozanne, 1999). Children with phonological delay and deviantconsistent disorder show inconsistency of less than 30% (Zhu & Dodd, 2000). Three sources of evidence can be used to evaluate this classification system. Subgroups must be able to be differentiated by tasks evaluating perceptual, cognitivelinguistic, or motor skills. There should be evidence that a specific intervention approach benefits a specific subgroup. Finally, data should be sought to determine whether the classification system has application across speechdisordered populations, including those that speak languages other than English. A brief overview of available evidence follows.

Identifying Deficits In the Speech Processing Chain Experimental studies have compared subgroup performance on measures associated with speech processing. In an initial validation of the three phonologically impaired subgroups (Dodd, Hambly, & Leahy, 1989), the consistent use of nondevelopmental error patterns in children was attributed to an impaired ability to master the phonotactic constraints of the phonological system to be acquired. This subgroup of children performs more poorly than the other subgroups on tasks assessing rhyme and alliteration awareness, awareness of phonological legality, and literacy

American Journal of Speech-Language Pathology • Vol. 10 • 291–307 • August 2001

measures (Dodd et al, 1989, 1995). This would seem to reflect a cognitive-linguistic difficulty that lies at the organizational level of the speech chain (Grundy, 1989). This classification system has been further validated by Leitão, Hogben, and Fletcher (1997, who observed the same subgroups in their group of English-speaking subjects with speech disorders. Children in the deviantconsistent subgroup were most likely to experience difficulties in the acquisition of literacy. In contrast, children who make inconsistent errors seem to have an intact phonological system. A series of experiments indicated that children classified as inconsistent perform as well as age-matched, normally speaking controls on tasks assessing phonological awareness (Dodd et al., 1989), yet they perform poorly on tasks assessing expressive lexical measures (Dodd & McCormack, 1995) and phonological assembly tasks (Bradford-Heit, 1996). No deficit in the speech processing chain has yet been identified for the articulation-disordered and delayed subgroups in the preschool years. They performed within normal limits on most experimental tasks, albeit at the lower end of the normal range (Dodd & McCormack, 1995).

Intervention Specificity Experimental studies allow hypotheses concerning the aspect of the phonological processing chain that should be the focus of intervention for two subgroups: the metaphonological skills of children who consistently use atypical phonological error patterns, and phonological assembly for children who make inconsistent errors. The results of treatment case studies have provided initial evidence that supports these hypotheses. An alternative treatments clinical research design revealed that children with consistent nondevelopmental errors benefited most from treatment targeting phonological contrasts, whereas use of a core vocabulary increased both speech accuracy and consistency of children making inconsistent errors (Dodd & Bradford, 2000). Two other studies evaluated the efficacy and generalization of specific treatment approaches, in English, for speech disorder in bilingual children in order to test hypotheses about the nature of the underlying deficits (Holm & Dodd, 1999; Holm, Ozanne, & Dodd, 1997). It was predicted that treatment targeting phonological contrasts should not generalize because phonological contrasts are language specific, whereas a core vocabulary approach to inconsistency should generalize because phonological assembly is a mental process common to both languages. Although phonological contrast treatment resulted in remediation of cluster reduction in English, there was no generalization to [k(h)w] in Cantonese. In contrast, core vocabulary treatment in English generalized to Punjabi in a child who was inconsistent in both languages.

Crosslinguistic Studies Most research on phonological disorder has focused on monolingual English-speaking children. Crosslinguistic

description should allow evaluation of classification systems for speech disorder as well as data concerning the effect of the ambient phonology on disorder. Studies of speech disorders have been conducted for children speaking other languages including Italian (Bortolini & Leonard, 1991), Portuguese (Yavas & Lamprecht, 1988), Swedish (Nettelbladt, 1983), Cantonese (So & Dodd, 1994), Turkish (Topbas & Konrot, 1996), Spanish (Goldstein, 1996), and Putonghua (Zhu & Dodd, 2000). In general these studies suggest that some error patterns occur in most languages whereas others are languagespecific. Only the last four studies evaluated whether the classification system for English, proposed by Dodd (1995), is applicable to other languages. All four studies provided support for the subgroup classification. The four subgroups were identified in Cantonese, Putonghua, and Turkish. Goldstein (1996) identified only three of the groups, finding no children who made inconsistent errors. However, he did not test for consistency of production of the same lexical items. Given that previous studies have focused on small numbers of children (4–33), there is a need for larger studies of speech disorder in languages other than English. The primary purpose of the study reported here was to investigate whether the four hypothesized subgroups of speech disorders (Dodd, 1995) could also be found in German, as a second Germanic language, in a large-scale study. These data would provide evidence concerning the validity of the classification approach as well as the universal nature of developmental speech disorders. Of particular interest was whether the proportions of children classified as having articulation disorder, delay, consistent use of nondevelopmental error patterns, and inconsistent production of the same lexical items were similar across languages.

German Phonology and Variations in Northern German German belongs to the Germanic languages and is the first language of approximately 119 million people in 15 countries, with the largest communities in western and central Europe. German is the official state language in Germany, Austria, and parts of Switzerland, Liechtenstein, and Luxembourg (Barbour & Stevenson, 1990; Durell, 1992; Lyovin, 1997). There are regional variations in pronunciation, from accents to dialects, of the official language (Hochdeutsch), that is based on a North German pronunciation of the written language (Barbour & Stevenson, 1990; Durell, 1922; Goltz & Walker, 1961). The spelling of standard German generally gives a clear guide to pronunciation. High German contains 23 consonant phonemes /p b t d k g f v s z S x C h m n N l ‰ j / pf ts/. The glottal stop /// appears before syllable initial vowels and is compulsory. The sounds /pf/ and /ts/ will be treated as affricates following Ternes (1987), and not two-element clusters despite disagreement in the literature (Kohler, 1995; Ternes, 1987; Wiese, 1996). Whether /C/ and /x/ should be considered as two phonemes or as allophones is also Fox & Dodd: Differential Diagnosis of Phonological Disorders

293

controversial (Kohler, 1995; Si-Taek 1992; Wiese, 1996), but for ease of presentation they will be classified as phonemes. In word-final position the contrast between voiced and voiceless consonants is neutralized, with all consonants being voiceless. In word-initial position there are 22 two-element clusters and 2 three-element clusters. There are also many word-medial and word-final clusters. The vowel system contains 15 monophthongs (see Table 1) and 3 diphthongs, which, in the North, have a tendency to be over-stressed (Lenisierung). The shortest possible syllables are the structures CV, e.g. /ku…/ = Kuh (cow), and VC, e.g. /ap/ = ab (off). Structures of the combinations C1-3VC1-5 in monosyllabic words are possible: C1-3, e.g., /St‰/ in Strumpf (sock); C1-5, e.g., /mpfst/ in du schrumpfst (you shrink). German is an agglutinating language where, by addition of nouns, words can consist of about eight syllables or more, e.g.,

TABLE 1. German and English phonology. Englisha

German

pbtdkg mn fvzS‰h j l ts pf pbtdkg mnN f v s z S C xb ‰ l ts pf

Initial Consonants pbtdkg mn TDfvszSZh wj lr ‰¸ Medial Consonants pbtdkg mnN TDfvszSZ lr ‰¸ Final Consonants pbtdkg mnN TDfvszSZ lr ‰¸

ptk mnN fsSCx l ts pf b p g k f pf + l bpdtgkf+‰ k+n/v ts + v S+lmnp‰vt S+p/t+‰

Word Initial Clusters p g k f+l bpdtgkfST+r b p d t g k m n f v ‰ T+j d t g k T+w s+ptk m f l wj s+p+l s+ptk+r j s+k+w

iye„ouIY”ŒaÁE aI au OI (eå Oå uå)

Vowels i I ” œ Ø a Å O u Á ∏E Diphthongs eI oÁ aI aÁ OI IE ”E OE ÁE

Syllable Structure (Monosyllables) [C 0-3] - V- [C 0-3] nouns [C 0-3] - V- [C 0-4] [C 0-3] - V- [C 0-5] verbs a

Data concerning British English are from Modern English Structure by Strang (1969). b /C/ and /x/ are treated as two phonemes.

294

Hallenhandballweltmeisterschaft (indoor handball world championship) (Meinhold & Stock, 1980). All children taking part in this study grew up in the region of Southern Schleswig-Holstein and Hamburg. None of the parents or the children spoke North Saxon or any other dialect, but they all displayed influences of the North German accent. Variations in the use of consonants in this region have been described by Goltz & Walker (1961), Durell (1992), Barbour & Stevenson (1990), and Kohler (1995). The sound /r/ is always realized as [‰]. In all positions other than before vowels, /‰/ is realized as a vocalic Ersatzlaut (substitution sound) [å]. This creates several additional long or short diphthongs as in Wurst [vu…åst] (sausage) or Berg [be…åk] (mountain). The phoneme /N/ in word-final position can be pronounced as either [N] or [Nk]. The consonant /l/ before /C/, as in Milch /milC/ (milk), may be deleted and the vowel changed to [e]. The final letter “g” can be pronounced as either [k] or [C]. In word-initial position the affricate /pf/ may be replaced by [f] and the cluster /pfl/ may be replaced by [fl]. The replacement of /pf/ by [f] has been recorded as a phenomenon that is very common in colloquial speech (Barbour & Stevenson, 1990) and in the North (Durell, 1992). Further, the deaffrication of /ts/ to [s] in word initial position has been described as common. Replacement is unlikely to cause problems in understanding, since /s/ stays contrastive with /z/ in word initial position and no homonyms are created. The wordfinal unstressed syllables /En/ and /El/ can be reduced and assimilated, i.e., [ge…bm] instead of /ge…bEn/ geben (to give), and [fo…gl] instead of /fo…gEl/ Vogel (bird).

Speech Treatment Provision and Phonological Disorders In the German literature, speech disorders are referred to as “articulation disorder” or “dyslalia” and sometimes even “Stammeln” (stammer). The distinction between phonological and phonetic disorders was initially described by Scholz (1974), and more recently by Dickmann et al. (1994), Böhme (1998), and Hacker and Wilgerstein (1999). However, speech disorders are still classified primarily according to severity, etiology, or co-existing symptomatology (Wirth, 1990; Biesalski & Frank, 1994; Böhme, 1997). Treatment is typically based on articulation treatment approaches (e.g., Van Riper, 1963) and more recently approaches based on translations of English work (e.g., Metaphon: Jahn, 2000). Consequently, research needs to establish whether diagnostic categories developed from research on English-speaking children are applicable to monolingual German-speaking children. Two studies have assessed the distribution of phonological processes in children with speech disorders. Hacker and Weiss (1986) studied fifteen 5- to 7-year-old children with speech disorders. Most errors were substitutions, 9% were deletions, and 4% were assimilation errors. The processes most frequently identified were cluster reduction, fronting, stopping, and backing. Although this study sought to identify delayed or deviant acquisition, this could not be achieved since normative data were limited.


Romonath (1991) assessed 35 children aged 5;3–7;2 (years;months) and compared the number and types of phonological processes of children with speech disorders with those of an age-matched, normally developing control group. Children with speech disorders used a greater number of processes, and 25% of their processes did not occur in the speech of normally developing children of the same age. Processes occurring frequently were velar fronting, backing of consonants, prevocalic voicing, alveolar assimilation, stopping of liquids, obstruent devoicing, cluster reduction, and final consonant deletion. Möhring (1938) investigated the vulnerability to error of each phoneme of German in 1000 children aged 6–10 years. He described a hierarchy of difficulty according to their percentage of incorrect production. Three groups of different difficulty level were identified: 1. 1.5%–11.1%: m, n, b, d, p, l, t, f, v 2. 17.9%–28%: x, j, ‰, N, k, g 3. 33.5%–54.5%: C, S, s/z There are three hypotheses for this study: 1. German-speaking children with speech disorders can be classified according to four subtypes based on their surface error patterns: articulation disorder, delay, atypical error patterns, and inconsistent errors. This hypothesis is based on the assumption of the universal nature of speech development and its disorders. The data will provide evidence concerning the validity of the classification procedure and the assumption of universality. 2. The proportion of children classified as falling into each subgroup will be similar to results for other languages (Cantonese, English, and Putonghua). 3. Differences in the use of phonological processes among German, English, and other languages should be found, since the phonetic inventory and phonological constraints of the languages differ. In particular, since there is no contrast between /T/, /D/ and /s/, /z/ in German, German-speaking children might show a different error pattern from English-speaking children. Since English and German belong to the same language family, other differences should be minimal.

Method Subjects A native German speech-language pathologist (AF) tested 110 children between the ages of 2;7 and 7;7. These children had been referred to speech and language treatment because of concerns about their speech. Most children (N = 79) were on the waiting lists at two private practices in Northern Germany and were randomly chosen for an appointment for this study by the speech-language pathologists of the practices. Preschool teachers referred 31 children to the study during the collection of data on normal speech development in kindergartens. Criteria for participation in the study were: (a) age between 2;6 and 8;0; (b) referral for assessment of suspected speech disorder; (c) no previous treatment; (d) no sensory

impairment, organic motor disorder, cranio-facial anatomical anomaly, or intellectual impairment; (e) monolingual German speaker; and (f) no hearing loss detected at assessment. Based on these criteria, 100 children proved to be suitable for the study. A total of 10 children were excluded: 1 child did not speak, 2 children named too few pictures to obtain an adequate speech sample, and 7 children no longer evidenced speech problems. There were 63 boys and 37 girls in the population. This proportion of around 2:1 is reported to be commonly found in children with speech and language difficulties (Romonath, 1991).

Procedure The children were assessed individually in a single session at the private practice to which they had been referred. The assessment took place in a quiet room with only the tester, the child, and the mother (or parents) present. The session consisted of three parts: assessment tasks, free play, and parental advice. The whole session was recorded on audiotape using a Sony Professional Micro Stereo Recorder for a second phonetic transcription of the assessment. During the assessment the mothers were asked to complete a questionnaire about the child’s developmental and medical history and to give their permission for anonymous use of all data in the study. At the end of the session the mothers were informed whether treatment was indicated and were advised about how to deal supportively with the speech and language problem in daily communication. All children underwent the following assessments: 1. Single-Word Test: A picture-naming assessment procedure (Single-Word Test) was used to elicit data. The assessment included 99 items assessing all German phonemes in all possible word positions, as well as most word initial clusters and a sample of word medial and final clusters (see Appendix A). The task was identical to that used for our normative study (Fox & Dodd, 1999). The objective was to investigate the child’s phonetic and phonemic inventory and to derive the phonological processes used. The children were asked to name the pictures presented and were offered a sentence completion task in case of difficulty. If this help was insufficient, they were offered a choice of possible answers. Direct imitation was avoided. 2. The 25-Word Consistency Test: A German version of the 25-Word Consistency Test (Dodd, 1995) was created. It is a picture-naming task containing words of up to five syllables, with many consonant clusters or words that are, from clinical experience, difficult for German-speaking children to produce. Children were asked to name the pictures on three separate occasions within one assessment session, each occasion being separated by another activity. The words of this assessment can all be found in the Single-Word Test. Each child was asked to repeat these 25 words twice more throughout the session, either as a straightforward picture-naming task or integrated in a game, depending on the child’s age and cooperation. The word list can be found in Appendix B. This task was carried out to determine whether the child was consistent Fox & Dodd: Differential Diagnosis of Phonological Disorders

295

in his pronunciation of phonemes in an identical phonetic context, when producing single words.

Analyses Using the International Phonetic Alphabet Revised (1993), the examiner transcribed all utterances made by the child. The picture-naming tasks were transcribed online during the session, and again later from tape. Intrarater (point-to-point) reliability on data from 10% of the children randomly chosen for transcription reliability assessment (the picture naming task, the inconsistency task) was 95%. Additionally, 10% of the data were chosen randomly to assess reliability of phonetic transcriptions of two examiners; both spoke German as a first language, and one was a phonetician. The interrater reliability of broad phonetic transcription for consonants and vowels (as suggested by Shriberg, Lewis, McSweeny, & Wilson, 1997) was 96.2%. The criteria for classification into subgroups were strict and based on the Single-Word Test and the 25-Word Consistency Test. To be classified as being delayed, all of the child’s error patterns had to be identified as being used by at least 10% of children in the normative sample (Fox & Dodd, 1999) sometime during development between 18 months and 6 years. At least one of those error patterns had to be inappropriate for the child’s chronological age, being typical of an earlier stage of phonological development. To be classified as belonging to the deviant-consistent subgroup, a child had to consistently use at least one error pattern that was not used by more than 10% of children in the normative sample (Fox & Dodd, 1999) at any age. To be classified as inconsistent, a child had to realize at least 40% of words in the 25-Word Consistency Test differently over three trials (i.e., where all three productions differed, or one production differed from the other two that were the same). Two speech-language pathologists, a native German fairly unfamiliar with the procedure and an Australian who regularly uses the classification system in clinics, were asked to independently classify 20% of the children randomly chosen out of the 100 subjects. The classification reliability was 94.7%. Each child’s transcript was analysed to derive the following measures: Phonetic Repertoire: A phone was accepted as being part of a child’s phonetic inventory if it was correctly realized at least twice within the data elicited from the Single-Word Test, irrespective of whether the sound uttered was phonologically correct. Phoneme Repertoire: A phoneme was accepted as being part of a child’s phonemic inventory if it was produced correctly at least 66.7% of the time (on two out of three occasions) within the picture-naming task data. Percent Phonemes Incorrect (PPI): The percentage of incorrect phonemes was calculated by multiplying the number of incorrect phonemes by 100 and dividing by the total number of phonemes produced within the SingleWord Test. Inconsistency Score: An inconsistency score was derived by calculating the number of trials where a word 296

was not produced identically on all three opportunities, multiplying by 100, and dividing by the number of trials (out of 25) where a word was attempted on all three opportunities. Only spontaneous productions of the target word were included in the analysis. For example, if the target Hund was produced once correctly, once with initial consonant deletion, and once with cluster reduction, it would count as an inconsistent trial, as would an example where Hund was produced once with initial consonant deletion and twice with cluster reduction. Number of Processes: A process was counted as present if it occurred more than twice in different lexical items of the Single-Word Test. All measurements taken for this analysis were identical to those outlined in our study of normal phonological development in 180 monolingual children acquiring German at ages 1;6–6;0 (see Appendices C1 and C2; Fox & Dodd, 1999). This study was based on the same SingleWord Test. Children were classified into the four subgroups with reference to the normative data for phonological process use and their score on the 25-Word Consistency Test.

Results The data were inspected to determine whether the subgroups—articulation disorder, delay, deviant-consistent, and inconsistent—were apparent. As shown in Table 2, 20 children (20%) were classified as having an articulation disorder, distorting the sounds /s/, /z/, /C/, and /S/; 51 children (51%) were classified as having delayed phonological development; 17 children (17%) were classified as having a deviant-consistent phonological disorder; and 12 children (12%) were classified as having an inconsistent phonological disorder. Individual data for all children using consistent error patterns can be found in Appendix D. Table 3 gives an overview of the children in each subgroup, their error pattern usage, mean inconsistency percentage, percent phonemes incorrect (PPI), and their mean number and types of missing phones and phonemes.

Articulation Disorder Children who consistently distorted one or more particular phonemes in all phonetic environments, but made no other errors, were classified as having an articulation disorder. There were 7 girls and 13 boys, aged between 4;8 and 7;8 (mean age 5;9), in the articulationdisordered group. An incomplete phonetic inventory is the main feature of this disorder since children consistently produced a speech sound to mark a particular phoneme, but the sound produced was not a perceptually acceptable version of the target sound. Of these 20 children, 19 replaced /s/ and /z/ with [T] and [D] and /ts/ and [tT], 1 child replaced these phones by a lateral [s], and 4 children lateralized the phone /S/ (see Appendix D, Table D1). The distortions identified were interdental and lateral production of sibilants. The percentage of distorted phonemes (when counted as incorrect) was between one and two standard deviations above the mean for normally developing age-matched


TABLE 2. Subject information. Articulation

Delay

Consistent

Inconsistent

Total

20 4 (4.8%)

51 51 (61%)

17 17 (20.2%)

12 12 (14.3%)

100 84 (100%)

No. of children additionally classified as articulation disordered

N/A

29

10

6

No. of children additionally classified as articulation disordered excluding interdentalitya as a symptom for articulation disorder

N/A

5

1

0

No. of boys

13

31

12

7

No. and % of children No. and % of children when isolated /s/ and /z/ distortions are excluded

63

No. of girls

7

20

5

5

37

Mean age

5;9

5;1

4;9

4;2

5;0

a Because of findings by Fox & Dodd (1999), interdentality may be a variation of the norm rather than an indication of articulation disorder. Therefore, numbers of children with and without interdentality are presented.

children (see Figure 1). No children in this group showed an incomplete phonemic inventory (since the phoneme was marked consistently, although distorted). Their inconsistency rating was 0%. Interdental production of sibilants does not necessarily merit a classification of articulation disorder since the normative sample (Fox & Dodd, 1999) indicated that interdental production might be considered allophonic (see Discussion). Tables 2 and 5 therefore present data for the articulation-disordered group in two parts: including all children initially identified as being articulation disordered and excluding the 16 children whose only speech difficulty was interdental production of sibilants. Excluding these latter 16 children reduces the percentage of articulation disorder to 5%.

Delay A total of 51 children (22 boys and 19 girls), aged between 3;3 and 7;9 (mean age 5;1), were classified as delayed. The percentage of incorrect phonemes was between one and two standard deviations above the mean for normally developing age-matched children, but for children older than 5;3 years, it increased to two to four standard deviations above the age-matched appropriate mean (see Figure 2). When the phones /s/, /z/, and /ts/ were included, 37 children (73%) showed an incomplete phonetic inventory, but only 26 (51%) did so when these phones were excluded. Apart from /s/, /z/, and /ts/, the main missing phones were identical to the missing phonemes. Thirty-five of the delayed children (69%) showed

TABLE 3. Information about subgroup results.

Articulatory error patterns

Articulation

Delay

Consistent

★

★

★

★

★

★

N/A

★

N/A 59%

Developmental error patterns Idiosyncratic error patterns

Inconsistent

Mean inconsistency

0%

13%

19%

Mean no. of error patterns

1.15

2.55

5.06

N/A

Range of error patterns

1–2

1–7

3–10

N/A 22%

Mean PPI

7%

9%

19%

Mean z-score PPI

1.89

0.74

3.98

3.97

Mean PCI

10%

14%

29%

35%

Mean no. missing phones

2.05

1.9

3.6

2.25

s/z ts S

s/z ts S g k N

s/z ts S C ‰ t d n pf

s/z ts S ‰ pf vx

Mean no. missing phonemes

0

1.72

4.29

4.75

Standard deviation of missing phonemes

0

1.64

2.7

3.7

SgkN

S ts s/z C f v pf d t n k g ‰

All but mnNpbC

Most frequently missing phones

Most frequently missing phonemes

Note. PPI = percentage phonemes Incorrect; PCI = percentage consonants incorrect.


297

TABLE 4. Comparison of two studies ranking phonemes according to the percentage of children producing a specific phoneme incorrectly. Most correct Möhringa This study a

m n h b d p l t f v m b p l h n j x d t v pf

Least correct x j ‰ N k g f N C ‰ g k S

C S s/z s/z* ts*

Möhring (1938) does not present data for /ts/, /pf/ and /h/.

* /T/ and /D/ were accepted as correct allophonic variations of /s/ and /z/.

an age-inappropriate phonemic inventory, with a mean of 1.72 (SD 1.6) missing phonemes per child (usually /k/, /g/, /N/, /C/, or /S/). The mean inconsistency rate for this group was 13%. The error patterns of the children classified as delayed had to reflect normal development but inappropriateness for chronological age. The most common delayed processes were cluster reduction, fronting of plosives and sibilants, and interdentality as an additional articulatory distortion (see Table D2).

Phonologically Disordered—Deviant-Consistent A total of 17 children, 12 boys and 5 girls aged 3;5–6;11 years (mean age 4;9), were classified as having a deviantconsistent disorder. Their mean inconsistency rate was 19% and the mean PPI was 19%. Although this mean percentage was high, there was considerable individual variation: although some children’s PPI was only one or two standard deviation above the mean for normally developing children,

FIGURE 1. Percentage phonemes incorrect (PPI) for articulation disorder.

FIGURE 2. PPI for delayed phonological development.

298


TABLE 5. Crosslinguistic comparison of the distribution of subgroups according to Dodd’s classification system (1995).

No. of children Articulation Delay Deviant Inconsistent

English

Cantonese

Putonghua

Spanish

55 14% 58% 12% 16%

17 12% 47% 29% 12%

33 3% 55% 24% 18%

20 10% 65% 25% —b

Germana 100 20% 51% 17% 12%

84 5% 61% 20% 14%

Note. English (Dodd, 1995); Cantonese (So & Dodd, 1994); Putonghua (Zhu & Dodd, 2000b); Spanish (Goldstein, 1996). a

The left column indicates the number and percentage of children including children with an isolated lisp, whereas the right column presents data excluding these children. b Not assessed

others reached a PPI of six to eight standard deviations above the mean (see Figure 3). Fourteen children (82%) showed an incomplete phonetic inventory, and 15 children (88%) showed an incomplete phonemic inventory. Unlike results for the delayed group, there were no phones or phonemes (excluding /s/, /z/, and /ts/) that were particularly affected. Missing phonemes (that should have been acquired) in four to seven of the children were /d/, /t/, /n/; /f/, /v/, /pf/; /s/, /ts/, /C/; k/, /g/, and /‰/. All children showed developmental error patterns as well as idiosyncratic ones. Furthermore, some children used a very unusual pattern of cluster reduction and were therefore classified as deviantconsistent disordered. The most common developmental and idiosyncratic processes were: fronting, backing of plosives, /f/ → [s] or [T], and cluster reduction. Interdentality as an articulatory phenomenon was also common (see Table D3).

Inconsistent A total of 12 children, 7 boys and 5 girls aged 2;7–5;8 (mean 4;2), were classified as inconsistent. The mean

inconsistency rate was 59%. A one-way ANOVA comparing the rates of inconsistency between the four subgroups demonstrated that significant differences could be found: F(3, 96) = 52.336, p < .001. Post-hoc analysis using Student-Newman-Keuls showed that the inconsistent disordered group was significantly different from all other subgroups. The same was found for the articulationdisordered group, which was highly consistent. No significant differences were found between the delayed and the deviant-consistent phonologically disordered subgroups. The mean PPI in the inconsistent disordered group was 22%, the highest of all groups (see Figure 4). Nine children showed incomplete phonetic and 10 showed incomplete phonemic inventories. However, if the phones /s/, /z/, and /ts/ were excluded, only 7 children (58%) showed incomplete inventories, with fewer phones than phonemes missing. The most vulnerable phonemes were /k/, /‰/, /f/, /v/, /x/, and /pf/. Since the main feature of these children is inconsistency, error pattern descriptions will reflect only that one assessment and will differ on reassessment. Ball (1994) therefore argues that it is inappropriate to analyze inconsistent errors for phonological processes.

FIGURE 3. PPI for deviant-consistent phonological disorder.


299

FIGURE 4. PPI for inconsistent phonological disorder.

Group Comparisons Figures 1–4 show the distribution of the percent phonemes incorrect (PPI) for each child in each subgroup. Each child’s PPI was compared to the mean PPI z-scores obtained from the normative study (Fox & Dodd, 1999). A child’s PPI was compared to the z-scores of the appropriate normative age group. Two comparison data points are shown: the normative mean PPI z-scores, and that point plus two standard deviations. A one-way ANOVA comparing the PPIs of the four subgroups revealed significant differences: F(3, 96) = 31.772, p < .001. Post-hoc Student-Newman-Keuls showed no significant differences between the articulation disordered and delayed groups or between the deviant-consistent phonologically disordered and inconsistent phonologically disordered subgroups. The articulation disordered and delayed subgroups were significantly different from both the phonologically disordered subgroups.

Comparisons with Previous German Language Research Comparison of current results with those of previous studies (Hacker & Weiss, 1986; Romonath, 1991) reveals general agreement. The common error patterns identified were initial and final consonant deletion, weak syllable deletion, cluster reduction, assimilation, metathesis, fronting, backing, stopping, voicing, and devoicing. Other error patterns described were less frequent and idiosyncratic. The hierarchy of phoneme/phone difficulty as described by Möhring (1938) was also supported apart from the positions of the fricatives /C/ and /x/ and the approximant /j/ that appeared to be less vulnerable in our study (see Table 4) and the fricative /f/, which appeared more vulnerable in our study.

Crosslinguistic Comparison of Classification The distribution pattern (percentages per subgroup) for English, Putonghua, and Cantonese was similar across all 300

languages assessed, as shown in Table 5. The small number of Turkish-speaking children assessed (N = 10) did not allow the presentation of distribution percentages (Topas & Konrot, 1996). For German, two types of percentages are presented. The left column shows the percentages of all 100 children assessed, while the right column shows the distribution percentage when 16 children with an isolated interdental /s/ production are excluded. Speech error types across languages are compared in Table 6, which lists descriptive characteristics and most common error patterns for subgroups in German, English, Putonghua, and Cantonese. The most frequent individual error patterns found in each language depend on the phonological system of that language, but there are similarities. The most common delayed error patterns reported were cluster reduction, velar fronting, and stopping, whereas the most frequently reported nondevelopmental error patterns were backing and intrusive consonants. The data from all four languages indicate the following hierarchy for segmental errors across the four subgroups: articulation disorder < delay < deviant-consistent < inconsistent.

Discussion Crosslinguistic research provides evidence concerning the validity of classification systems for speech disorders. As previous studies have shown for Cantonese, Turkish, Spanish, and Putonghua, this study of German-speaking children has further supported Dodd’s (1995) classification of speech disorders based on English-speaking children. The four subgroups: articulation disorder, delay, and deviant-consistent and inconsistent phonological disorder, were apparent in 100 German-speaking children referred to speech and language treatment for assessment of a suspected speech disorder. Twenty children (20%) were classified as having a specific articulation disorder, which is high in comparison to other languages studied. There are two possible reasons for the finding. Data for this study were collected only in private practices and kindergartens. Children with an


TABLE 6. Crosslinguistic comparison of the subgroups. German

English

Cantonese

Putonghua

Stopping Deaspiration Final consonant deletion Final glide deletion Cluster reduction

Stopping Gliding Fronting of sibilants Affrication

Most frequent developmental error pattern

Interdentality Fronting of plosives Fronting of sibilants Cluster reduction Fronting of /N/ Cluster assimilation

Final consonant deletion Cluster reduction Weak syllable deletion Reduplication Fronting of velars Stopping

Most frequent idiosyncratic error pattern

Backing of alveolars /f/ → /s/ or /T/ Metathesis Intrusive consonants Consonant deletion Allophonic fricatives Favorite sound Cluster changes

Backing Backing Initial consonant Vowel rule deletion Medial consonant deletion Intrusive consonants Medial consonant substitution Denasalization Favorite sound

Backing and stopping Final consonant addition Vowel change

Mean no. of missing phones: articulation Mean no. of missing phones: delay Mean no. of missing phones: deviant-consistent Mean no. of missing phones: inconsistent

2 1.9 3.6 2.25

no data no data no data no data

1 3.5 2.4 1

1 1.6 2 6.3

Mean no. of missing phonemes: articulation Mean no. of missing phonemes: delay Mean no. of missing phonemes: deviant-consistent Mean no. of missing phonemes: inconsistent

0 1.72 4.29 4.75


0 5 3 6

0 3.3 3.8 10.7

Mean PCI: articulation Mean PCI: delay Mean PCI: deviant-consistent Mean PCI: inconsistent

10% 14% 29% 35%

no data 23% 46% 69%

10% 22% 24% 29%

7% 15% 26% 39%

Mean no. of developmental error patterns: delay Mean no. of idiosyncratic error patterns: delay Mean no of developmental error patterns: deviant-consistent Mean no. of idiosyncratic error patterns: deviant-consistent

2.55 0 1.6 2.2

2.7 0 3.5 1.9

3 0 0.4 2.8

1.6 0 1.25 1.5

Mean inconsistency: articulation Mean inconsistency: delay Mean inconsistency: deviant-consistent Mean inconsistency: inconsistent

0% 13% 19% 59%

no data 20% 24% 62%


14% 23% 25% 58%

Note. English: Bradford & Dodd (1996); Cantonese: So & Dodd (1994); Putonghua: Zu & Dodd (2000).

articulation disorder are usually referred to a private practice because the disorder does not have priority for assessment or treatment in public clinics. Data for the other languages were collected in hospitals or outpatient clinics. Thus, sampling may explain the difference in distribution of articulation disorders. An alternative explanation is that parents are more likely to seek treatment for a lisp in Germany than in other cultures. However, the phones /T/ and /D/ are not part of the German phonetic inventory and the study of typically developing children up to the age of 6;0 showed that 35% of the children in the oldest age group used /T/ and /D/ consistently as allophones of /s/ and /z/ (Fox & Dodd, 1999). The question then arises as to whether the interdental

production of the phonemes /s/ and /z/ really describes an articulation disorder, or whether this kind of replacement now needs to be accepted as a normal variation. If the latter is true, 16 of these 20 children would need to be excluded from our study, leaving only 5% of children with an articulation disorder, which is similar to the findings for English (14%), Putonghua (3%), Cantonese (12%), and Spanish (10%). A total of 51 children (61% after exclusion of the 16 articulation-disordered children with an interdental lisp) were classified as showing delayed phonological acquisition. This percentage is similar to findings for all other languages reported. These children used developmental error patterns typical of a younger child. Some children’s Fox & Dodd: Differential Diagnosis of Phonological Disorders

301

speech error patterns were chronologically mismatched, a finding also reported by Grunwell (1987) and So and Dodd (1994). The majority of children in this category presented a speech delay of 6 to 9 months, although some children showed delays of more than 18 months. This finding raises a question concerning the clinical implications of the extent of delay. Given that delay is the largest subgroup, research is needed to address this question. Seventeen children (20.2% after exclusions) were classified as having a deviant-consistent phonological disorder, a percentage consistent with other languages. As expected, all children showed three types of processes: ageappropriate developmental processes, age-inappropriate developmental processes, and nondevelopmental, idiosyncratic processes (see Tables 3 and 6). Although the presence of some normal developmental error patterns is positive, the consistent use of idiosyncratic error patterns that restrict syllable structure (all final consonants deleted, not just /t k l/) and reduce a class of consonants (e.g., plosives) to one phoneme (e.g., [k]) cannot be adequately described in terms of severity or delay. These nondevelopmental speech error patterns indicate phonological processing difficulties that impair the ability to deduce the constraints of spoken phonology and the orthographic representation of speech sounds (Leitão et al., 1997). Twelve children (14.3% after exclusions) were classified as inconsistent, similar to findings for other languages. These children represent the youngest age mean of all subgroups, a finding also made for Putonghua (Zhu & Dodd, 2000). There are two possible explanations. Ingram (1989) argues that very young children (vocabularies of up to 50 words) show an inconsistent pattern of word production, and that inconsistency is therefore a normal part of the acquisition process. Since these children represent the youngest age group, their inconsistency might reflect very severe delay. However, children who make inconsistent errors do not have small phonetic repertoires typical of very young children. Further, Teitzel and Ozanne (1999) found that even very young children (aged 20–24 months) show a very consistent pattern of word production. Alternatively, the younger mean age of the inconsistent group might reflect their unintelligibility. These children are often unintelligible even to their parents, who cannot acquire knowledge of how particular words are pronounced, and might therefore refer children who make inconsistent errors earlier. Data from the other subgroups provide some support for the suggestion that degree of unintelligibility affects the age at which children are referred. Children with an articulation disorder are usually intelligible, and this subgroup had the highest mean age (5.9 years, with the youngest child aged 4.8 years). Similarly, children in the delayed subgroup, who make fewer errors than the two disordered groups, had a mean age of 5.1 years. The argument that it is the type of errors that is important, rather than the absolute number, is emphasized by comparison of the two disordered groups. Although these two groups had similar PPI z-scores (3.98 and 3.97), the inconsistent group had a younger mean age (4.2 years). Parents with children who consistently make 302

the same errors (mean age in the current study, 4.9 years) learn to translate. Inconsistency makes such translation impossible and is likely to increase parental anxiety. As mentioned in earlier studies, some children show an articulation disorder in addition to a phonological delay or disorder (Dodd & Bradford, 2000). So and Dodd (1994) described two children with deviant-consistent disorder who also had an articulation disorder. In this study, five of the children who were classified as being delayed and one child classified as deviant-consistent disordered, were additionally classified as having an articulation disorder in that they were unable to produce acceptable versions of particular phonemes in any phonetic context. All distorted the phones /s/ and /z/ consistently as [Ò]. This is not surprising, since the co-occurrence of phonological and articulation (or phonetic) disorder has already been recognized (Fey, 1992; Kamhi, 1992). In all three phonological subgroups about half of all children showed an interdental production of /s/ and /z/. Given that up to 40% of children in the normative study also evidenced interdentality, this is more likely to reflect allophonic variation than an articulation disorder. The other subgroups—delay and deviant-consistent and inconsistent phonological disorder—are mutually exclusive by definition. For most children, classification into the four subgroups was obvious, given the strict criteria. There were only a few cases where children made errors that raised doubts. Two children from the delayed subgroup made up to three vowel errors. Vowel errors are atypical of normally developing children and if consistently used would indicate disorder. Six children made errors on up to three lexical items that were atypical of normal development. In the current study, these children were classified as delayed because these few errors were the only sign of phonological disorder as opposed to delay. In a clinical situation, any ambiguity in data from a standardized assessment procedure could be explored in depth to ensure correct diagnosis. Classification of subgroups of phonological disorder is currently a controversial topic. Although the use of a medical model (e.g., Shriberg, 1994, 1997) provides important information concerning factors influencing causal and maintenance factors, it has limited clinical applicability. By the time a child is referred for assessment, a number of causal and maintenance factors may be apparent (e.g., history of otitis media, older sibling who interprets for the child with a speech disorder, and family history of speech disorder). Given that a clinician’s task is to remediate the presenting problem, classification in terms of the surface speech error patterns provides more relevant information for determining treatment targets. Previous research has indicated that specific speech error pattern profiles are associated with differing deficits in the speech processing chain. Consistent use of nondevelopmental error patterns is associated with deficits in phonological awareness (Leitão et al., 1997); inconsistency is associated with deficits in phonological assembly (Bradford & Dodd, 1994, 1996). Further, clinical research trials suggest that specific treatment approaches targeting different aspects of the speech processing chain are most cost effective (e.g., Bradford & Dodd, 1998; Dodd & Bradford, 2000). The


finding that, irrespective of the language being learned, the same predominant surface error patterns are evident in approximately the same proportions of children with speech disorders emphasizes the clinical importance of using a behavioral as opposed to a medical model in the classification of speech disorders.

Conclusion It was hypothesized that the nature of speech development and its disorders is of universal character. Therefore, classifications of speech disorders found in speakers of one language can be validated by crosslinguistic research. The results of this crosslinguistic study of 100 monolingual German-speaking children support the existence of the four hypothesized subgroups of speech disorders found in English-speaking children: articulation disorder, delay, and deviant-consistent and inconsistent phonological disorder.

Acknowledgments We would like to thank all children, parents, and preschool teachers for their participation and interest. A special thank you goes to the private practices of speech-language pathologists Karen Grosstück and Holger Schultze for making this study possible. Thanks to Nicole Boheim for her transcription and Sharon Crosbie and Carola Hofmann for help with classification reliability. Further, we are grateful for the financial support given by the Economic and Social Research Council.

References Ball, M. (1994). Using dependency phonology in the analysis of disordered speech. Australian Journal of Human Communication Disorders, 22, 22–30. Barbour, S., & Stevenson, P. (1990). Variation in German. Cambridge, U.K.: Cambridge University Press. Biesalski, P., & Frank, F. (1994). Phoniatrie-Pädaudiologie (2nd ed.). Stuttgart, Germany: Thieme Verlag. Böhme, G. (1997). Sprach-, Sprech-, Stimm- und Schluckstörungen: Klinik (3rd ed., Vol. 1). Stuttgart, Germany: Gustav Fischer Verlag. Böhme, G. (1998). Sprach-, Sprech-, Stimm- und Schluckstörungen: Therapie (2nd ed., Vol. 2). Stuttgart, Germany: Gustav Fischer Verlag. Bortolini, U., & Leonard, L. (1991). The speech of phonologically disordered children acquiring Italian. Clinical Linguistics and Phonetics, 5(1), 1–12. Bradford, A., & Dodd, B. (1996). Do all speech disordered children have motor deficits? Clinical Linguistics and Phonetics, 10(2), 77–101. Bradford, B., & Dodd, B. (1994). The motor planning abilities of phonologically disordered children. European Journal of Disorders of Communication, 29, 349–369. Bradford-Heit, A. L. (1996). Subgroups of children with developmental speech disorders—Identification and remediation. Unpublished doctoral dissertation, University of Queensland, Australia. Crystal, D., Fletcher, P., & Graman, M. (1989). The grammatical analysis of language disability. New York: Elsevier. Crystal, D., & Varley, R. (1993). Introduction to language pathology. London: Whurr Publisher. Dickmann, C., Flossmann, I., Klasen, R., Schrey-Dern, D., Stiller, U., & Tockuss, C. (1994). Logopädische Diagnostik

von Sprachentwicklungsverzögerungen. Stuttgart, Germany: Thieme. Dodd, B. (1993). Speech disordered children. In G. Blanken, H. Dittmann, H. Grimm, J. Marshall, & C.-W. Wallesch (Eds.), Linguistic disorders and pathologies (pp. 825–834). Berlin, Germany: De Gruyter. Dodd, B. (1995). Differential diagnosis and treatment of children with speech disorder. London: Whurr Publisher. Dodd, B., & Bradford, A. (2000). A comparison of three therapy methods for children with different types of developmental speech disorders. International Journal of Communication Disorders, 35,189–209. Dodd, B., Gillon, G., Oerlemans, M., Russle, T., Syrmis, M., & Wilson, H. (1995). Phonological disorder and the acquisition of literacy. In B. Dodd (Ed.), Differential diagnosis and treatment of children with speech disorder (pp. 125–147). London: Whurr Publisher. Dodd, B., Hambley, G., & Leahy, J. (1989). Phonological disorders in children: Underlying cognitive deficits. British Journal of Developmental Psychology, 7, 55–71. Dodd, B., & Leahy, J. (1989). Disordered phonology and mental handicap. In I. Beveridge, G. Conti-Ramsden, & I. Leudar (Eds.), Language and communication in mentally handicapped people (pp. 33–56). London: Chapman & Hall. Dodd, B., & McCormack, P. (1995). A model of speech processing in differential diagnosis of phonological disorders. In B. Dodd (Ed.), Differential diagnosis and treatment of children with speech disorders (pp. 65–89). London: Whurr Publishers. Durell, M. (1992). Using German. Cambridge, U.K.: Cambridge University Press. Fey, M. E. (1992). Articulation and phonology: Inextricable constructs in speech pathology. Language, Speech, and Hearing Services in Schools, 23, 225–232. Fletcher, P. (1990). The breakdown of language: Language pathology and therapy. In N. Collinge (Ed.), Encyclopaedia of language (pp. 424–457). London: Rutledge. Fox, A. V., & Dodd, B. J. (1999). Der Erwerb des phonologischen Systems in der deutschen Sprache. Sprache-StimmeGehör, 23, 183–191. Garrett, K., & Moran, M. (1992) A comparison of phonological severity measures. Language, Speech, and Hearing Services in Schools, 23, 48–51. Gierut, J. A. (1998). Treatment efficacy: Functional phonological disorders in children. Journal of Speech, Language, and Hearing Research, 41, S85–S100. Goldstein, B. (1996). Error groups in Spanish-speaking children with phonological disorders. In T. W. Powell (Ed.), Pathologies of speech and language: Contributions of clinical phonetics and linguistics (pp. 171–177). New Orleans, LA: International Clinical Phonetics and Linguistics Association. Goltz, R., & Walker, A. (1961). North Saxon. In R. E. Keller (Ed.), German dialects (pp. 31–57). Manchester, U.K.: Manchester University Press. Grundy, K. (1989). Linguistics in clinical practice. London: Taylor and Francis. Grundy, K., & Harding, A. (1989). Developmental speech disorders, In K. Grundy, Linguistics in clinical practice (pp. 3–25). London: Taylor and Francis. Grunwell, P. (1985). PACS—Phonological Assessment of Child Speech. Windsor, U.K.: NFER-NELSON. Grunwell, P. (1987). Clinical phonology (2nd ed.). London: Croom Helm. Hacker, D., & Weiss, K. H. (1986). Zur phonemischen Struktur funktioneller Dyslalien. Oldenburg, Germany: Arbeiter Wohlfahrt Verlag. Fox & Dodd: Differential Diagnosis of Phonological Disorders

303

Hacker, D., & Wilgermein, H. (1999). Aussprachestörungen bei Kindern. München-Basel, Germany: Ernst Reinhardt Verlag. Holm, A. (1998). Speech development and disorders in bilingual children. Unpublished doctoral dissertation, Newcastle University. Holm, A., Ozanne, A., & Dodd, B. (1997). Efficacy of intervention for a bilingual child making articulation and phonological errors. International Journal of Bilingualism, 1, 55–69. Holm, A., & Dodd, B. (1999). An intervention case study of a bilingual child with phonological disorder. Child Language Teaching and Therapy, 15, 139–158. Ingram, D. (1989). First language acquisition: Method description and explanation. Cambridge, U.K.: Cambridge University Press. Jahn, T. (2000). Phonologische Störungen bei Kindern-Diagnostik und Therapie. Stuttgart, Germany: Thieme Publisher. Kamhi, A. (1992). The need for a broad-based model of phonological disorders. Language, Speech, and Hearing Services in Schools, 23, 261–268. Kohler, K. (1995). Einführung in die Phonetik des Deutschen (2nd ed.). Berlin, Germany: Erich Schmidt Verlag. Lambert, J., & Waters, D. (1995). Childhood phonological disorders. In M. Leahy (Ed.), Disorders of communication— The science of intervention (pp. 95–120). London: Whurr Publishers. Leitão, S., Hogben, J., & Fletcher, J. (1997). Phonological skills in speech and language impaired children. European Journal of Disorders of Communication, 32, 73–93. Leonard, L. (1985). Unusual and subtle phonological behavior in the speech of phonologically impaired children. Journal of Speech and Hearing Disorders, 50, 4–13. Lewis, B. A., & Freebairn, L. (1992). Residual effects of preschool phonology disorders in grade school, adolescence, and adulthood. Journal of Speech and Hearing Research, 35, 819–831. Lyovin, A. V. (1997). An introduction to the languages of the world. Oxford, U.K.: Oxford University Press. Meinhold, G., & Stock, E. (1980). Phonologie der deutschen Gegenwartssprache. Leipzig, Germany: VEB Bibliographisches Institut. Möhring, H. (1938). Lautbildungsschwierigkeit im Deutschen. Zeitschrift für Kinderforschung, 47, 185–235. National Institute of Deafness and other Communication Disorders. (1994). National strategic research plan. Bethesda, MD: Department of Health and Human Communication Services. Nettelblad, U. (1983). Developmental studies of dysphonology in children. Lund, Sweden: CWK Gleerup. Romonath, R. (1991). Phonologische Prozesse an sprachauffälligen Kindern, eine vergleichende Untersuchung an sprachauffälligen und nichtsprachauffälligen Vorschulkindern. Berlin, Germany: Edition Marhold. Scholz, H.-J. (1974). Zum Phonologischen Aspekt des Spracherwerbs und dessen Bedeutung für die Dyslalie. Die Sprachheilarbeit, 19, 145–152. Shriberg, L. D. (1982). Towards a classification of developmental phonological disorders. In N. Lass (Ed.), Speech and language—Advances in basic research and practice (Vol. 8, pp. 1–18). New York: Academic Press.

304

Shriberg, L. D. (1993). Four new speech and prosody-voice measures for genetic research and other studies in developmental phonological disorders. Journal of Speech and Hearing Research, 36, 105–140. Shriberg, L. D. (1994). Five subtypes of developmental phonological disorders. Clinics in Communication Disorders, 4, 38–53. Shriberg, L. D. (1997). Developmental phonological disorders: One or many? In B. Hodson & M. L. Edwards (Eds.), Perspectives in applied phonology (pp. 105–131). Gaithersburg, MD: Aspen Publishers. Shriberg, L. D., & Kwiatkowski, J. (1982). Phonological disorders I: A diagnostic classification system. Journal of Speech and Hearing Disorders, 47, 226–241. Shriberg, L. D., Lewis, B. A., McSweeny, J. L., & Wilson, D. L. (1997). The percentage of consonant correct (PCC) metric: Extensions and reliability data. Journal of Speech, Language, and Hearing Research, 40, 708–722. Si-Taek, Y. (1992). Unterspezifikation in der Phonologie des Deutschen. Tübingen, Germany: Max Niemeyer Verlag. So, L., & Dodd, B. (1994). Phonologically disordered Cantonesespeaking children. Clinical Linguistics and Phonetics, 8, 235–255. Stackhouse, J., & Wells, B. (1997). Children’s speech and literacy difficulties. London: Whurr Publisher. Teitzel, T., & Ozanne, A. (1999, July). Variability in single word productions of typically developing toddlers. Paper presented at the 20th Child Phonology Conference, Bangor, ME. Ternes, E. (1987). Einführung in die Phonologie. Darmstadt, Germany: Wissenschaftliche Buchgesellschaft. Topbas, S., & Konrot, A. (1996). Variability in phonological disorders: A search for systematicity? Evidence from Turkishspeaking children, Paper presented at the 5th Annual Conference of the International Clinical Phonetics and Linguistics Association, Munich, Germany. Van Riper, C. (1963). Speech correction: Principles and methods (4th ed.). Englewood Cliffs, NJ: Prentice Hall. Weiss, C. E., Gordon, M. E., & Lillywhite, H. S. (1987). Articulatory and phonological disorders. Baltimore, MD: Williams & Wilkins. Wiese, R. (1996). The phonology of German. Oxford, U.K.: Clarendon Press. Wirth, G. (1990). Sprachstörungen, Sprechstörungen, kindliche Hörstörungen (3rd ed.). Köln, Germany: Deutscher Ärzte Verlag. Yavas, M., & Lamprecht, R. (1988). Process and intelligibility in disordered phonology. Clinical Linguistics and Phonetics, 2, 329–345. Zhu, H., & Dodd, B. (2000). Putonghua (modern standard Chinese)-speaking children with speech disorders. Clinical Linguistics and Phonetics, 14, 165–191. Received November 29, 2000 Accepted February 5, 2001 DOI: 10.1044/1058-0360(2001/026) Contact author: Annette V. Fox, PhD, Kaiser-Friedrich-Ufer 24, 20253 Hamburg, Germany. E-mail: [email protected]


Appendix A

Appendix C

List of Test-Items

Developmental Error and Acquisition Patterns in Normally Developing German-Speaking Children

Mond Eimer Baum Ball Gabel Blume Brille Brief Pilz Wippe Korb Pferd Apfel Topf Pflaster Vogel Marienkäfer Schiff Flasche Frosch Wurst Löwe Lampe Teller Ball Nuß Kanne Telefon Dusche Feder Rad Drachen Tasse Auto Bett Traktor

[m] [m] [m] [b] [b] [bl] [b‰] [b‰] [p] [p] [p] [pf] [pf] [pf] [pfl] [f] [f] [f] [fl] [f‰] [v] [v] [l] [l] [l] [n] [n] [n] [d] [d] [t] [d‰] [t] [t] [t] [t‰]

Zitrone Jäger Eichhörnchen Milch Taucher Buch Roller Schere Gießkanne Nagel Berg Glas Gras Grün Schlange Anker Kuh Jacke Sack Kleid Krokodil Knopf Quak Sonne Hase Haus Hexe Zwerg Zange Katze Pilz Schuh Tasche Fisch Schlüssel Schmetterling

[t‰] [j] [C] [ç] [x] [x] [‰] [‰] [g] [g] [k,ç] [gl] [g‰] [g‰] [N] [Nk] [k] [k] [k] [k] [k‰] [kn] [kv] [z] [h], [z] [s] [ks] [tsv] [ts] [ts] [ts] [S] [S] [S] [Sl] [Sm]

Schnecke Schrank Schwein Spinne Spritze Stuhl Kiste Nest Strumpf Rutsche Fenster Heizung Gespenst Schornstein Zebra Bild Punkt Bank Arzt Hund Gitarre Tiger Erdbeere kaputt Unfall Elefant springt

[Sn] [S‰] [Sv] [Sp] [Sp‰] [St] [st] [st] [St‰] [tS] [nst] [N] [Sp], [nst] [nSt] [b‰] [lt] [Nkt] [Nk] [tst] [nt]

TABLE C1. Developmental error pattern.

Note. From Fox, A. V., & Dodd, B. Der Erwerb des phonologischen Systems in der deutschen Sprache. Sprache-Stimme-Gehör, Thieme, Stuttgart, Germany. Reprinted by permission.

Table C2. Acquisition patterns for phonetic and phonemic inventory.

Appendix B 25-Word Inconsistency Test–List of Items Träcker Flasche Fisch Unfall Knöpfe

Marienkäfer Eichhörnchen Strumpf Rutsche Gitarre

Elefant Schwein Springt Schlüssel Spritze

Krokodil Gespenst Brief Drachen Frosch

Schiff Zwerg kaputt Glas Tiger

Age Group

Age

1 2 3 4 5 6 7

1;6–1;11 2;0–2;5 2;6–2;11 3;0–3;5 3;6–3;11 4;0–4;5 4;6–4;11

Phonetic Inventory mbdtn pfvl x g k h ‰ pf jN C S

Phonemic Inventory mpd bn v f l t N x h k s/z* j ‰ g pf ts* C S

Note. General acquisition criterion: 90% of the children of an agegroup had acquired the phone or phoneme (for individual criteria, see the Analysis section). * /T/, /D/, and /tT/ were accepted as phonemically correct.


305

Appendix D Error Patterns in Children With Speech Disorders TABLE D1. Articulation-disordered subgroup. Age

56

58

58

62

64

65

66

67

67

67

67

71

74

75

75

75

75

76

78

80

Int Lat

✩

✩

✩

✩

✩

✩

✩ ★

✩

✩

✩ ★

✩

✩

✩

✩

✩

✩

✩

✩

✩ ★

★

Note. Age = age presented in months; Int = interdentality; Lat = lateral production of sibilants (→ Ò); ✩ = age-appropriate process; ★ = ageinappropriate process.

TABLE D2. Delay subgroup. Developmental Error Patterns Age 39 40 43 44 47 47 48 50 50 52 52 53 53 53 53 55 55 55 56 56 58 58 59 59 59 62

Asm ClR ✩

✩ ✩ ✩ ✩

✩ ✩ ✩

CD WSD Fro ✩ ★ ★

★ ★

★

★ ✩ ★

✩ ✩ ✩ ✩

★

★

✩

Voi BkS GlR Daf

★

✩

✩

Sto

Developmental Error Patterns

★ ★ ★ ★ ★ ★ ★

★

✩

★

★ ✩ ✩ ★

★ ★

★ ★ ★ ★ ★ ★ ★ ★ ★

★ ★ ★ ★

★ ✩

✩ ★

Age 62 62 63 63 63 64 64 65 65 65 66 66 67 67 67 71 72 73 73 74 78 79 80 80 81

Asm ClR

CD WSD Fro

Sto

Voi BkS GlR Daf

★ ★

★

★

★ ★

★ ★ ★ ★ ★

★

★

★ ★

★ ★ ★ ★

★ ★ ★ ★ ★

★ ★ ★

★

★ ★ ★ ★ ★

★ ★ ★

★

Note. These data do not include articulatory patterns. Age = age presented in months; ✩ = age-appropriate process; ★ = age-inappropriate process. Abbreviations: Asm = assimilation; CIR = cluster reduction; CD = initial or final constant deletion; WSD = weak syllable deletion; Fro = fronting of plosures/sibilants; Sto = stopping; Voi = voicing, devoicing, or cluster devoicing; BkS = backing of plosives; GlR = glottal replacement; Daf = deaffrication.

306


TABLE D3. Deviant-consistent subgroup. Developmental Error Pattern

Nondevelopmental Error Pattern

Age Asm ClR CD WSD Fro Sto Voi BkS GlR Daf 41 49 49 50 51 54 53 53 54 58 60 63 65 69 70 71 91

★

✩ ✩ ✩ ✩ ✩ ✩ ✩ ✩ ★

★

★

★ ★ ★

★

★

★

★

★ ★ ★

✩ ★ ✩ ✩ ✩ ✩ ★ ★ ★ ★ ★

Vow Met Affr f-T

s-f CD IntC BkP UnD Gl/l/ ‰/l-j ‰-s FavS AlloClCh ND/k ★

★ ★

★ ★

★

★ ★ ★

★

★

★

★ ★

★

★ ★ ★

★

★

★

★ ★ ★

★ ★ ★

★ ★ ★

★

★

★

★ ★

★ ★ ★

★

★ ★

★ ★

★

★ ★

★

Note. These data do not include articulatory patterns. Age = age presented in months; ✩ = age-appropriate process; ★ = age-inappropriate process. Abbreviations: Asm = assimilation; ClR = cluster reduction; CD = initial or final consonant deletion; WSD = weak syllable deletion; Fro = fronting of plosives/sibilants; Sto = stopping; Voi = voicing, devoicing, cluster devoicing; BkS = backing sibilants; GlR = glottal replacement; Daf = deaffrication; Voc = vocalization of /l/; Nas = nasality; Vow = vowel errors; Met = metathesis; Affr = affrication; f-T = /f/→[T]; s-f = /s/→[f]; CD = unusual consonant deletion; intC = intrusive consonant; BkP = backing of plosives; UnD = unusual developmental processes; Gl/l/ = glottal replacement of /l/; ‰/l-j = /‰/, /l/→[j]; ‰-s = /‰/→[s]; FavS = favourite sound; Allo = allophonic use of a sound class (i.e., nasals, fricatives); ClCh = cluster changes or assimilation; ND/k = /N/ deletion before /k/.


307