پیش بینی لکنت زبان از عوامل زبانی برای سخنرانان آلمان در دو گروه سنی

Brown’s factors [J. Speech Disorders 10 (1945) 181] predict the likely loci of disfluency in English-speaking adults who stutter. A word is more likely to be stuttered for these speakers if it is a content word, starts with a consonant, is positioned at the beginning of a sentence, and if it is a long word. These same factors were examined in native German-speaking children and adults who stutter. Speech data of 15 German adults and 17 children were coded according to Brown’s factors. For the adult group, it was predicted that words starting with consonants would not lead to as much of an increase in disfluencies compared with English samples, because of cross-linguistic differences in syllable onset properties. It was predicted that stuttering would be more likely in later sentence positions in German because in German the verb is usually near the end of a sentence. There were no obvious reasons to expect differences on the two remaining factors, content words and word length. With children, it was hypothesised that Brown’s factors that specify level of linguistic difficulty would not be such a good predictor of stuttering rate. Specifically, it was predicted that the difference in stuttering rate between function and content words would be lower in children. For the adults both word type (content/function) and word length increased stuttering rate significantly, whereas changes in stuttering rate for the other two factors were non-significant. It was also found that when word difficulty (based on a combined measure of all factors) increased, stuttering rate rose. With children, only the word-length factor was significant, and stuttering rate was not governed to the same extent by overall word difficulty. Conclusions are drawn as to the effect of linguistic and motor influences on stuttering.

The original impetus for research into linguistic variables associated with speech disfluency was triggered by Brown (1945). In this, the final article of a series of papers, he summarised his findings and identified four basic factors that determined whether words will be spoken disfluently by adults who stutter. The factors are: (1) word class (this has subsequently been interpreted as showing that content words are more prone to stuttering than function words); (2) word length (long words are more difficult); (3) sentence position (words that appear in early positions are more likely to be stuttered); (4) phone the word starts with (words starting with consonants are more difficult than those that start with vowels).2 The investigation of stuttering events in general (such as Brown’s) was criticised recently by Smith (1999). She pointed out that this approach misled researchers into thinking that stuttering is a static, rather than a dynamic, process, using the analogy of researchers investigating volcanoes by studying only the shape of the landform and type of eruptive material. The research reported here does not dispute the fact that stuttering is a multifactorial phenomenon, and Smith’s criticism highlights the need to approach it from many angles, such as the study of disfluent events. One reason to carry out cross-linguistic, or comparative research is to find out whether stuttering occurs in linguistic structures irrespective of their motor form; or whether difficult motor outputs lead to stuttering independent of the linguistic unit in which they occur. Some dissociation between motor and linguistic aspects can be achieved because the levels of motor complexity on different linguistic units differs between languages. Even though use of other languages allows scope for separating motor properties from the linguistic units in which they occur in English, no previous studies have made such comparisons. The main concern of the present study is why disfluency occurs on certain words, and in particular the degree to which linguistic and motor factors affect disfluency. Bloodstein (1995) pointed out that little work on the subject of Brown’s factors has been done in languages other than English apart from the factors which have been found to operate in Norwegian (Preus, Gullikstad, Grøtterød, Erlandsen, & Halland, 1970 as cited in Bloodstein, 1995) and Kanada — a Dravidian language (Jayaram, 1981). Although there is debate about confounding variables in Brown’s factors (see for instance Bloodstein, 1995, and Wingate, 1988), and the factors can be viewed as relatively crude measures, the current analysis considers them as linguistic determinants of disfluencies and focuses on a comparison between languages. One of the aims of the current investigation is to clarify how these factors influence stuttering rates in German. Although German and English are close in origin (both stemming from the West-Germanic branch of the Indo-European language tree), there are a number of important differences relevant to the operation of Brown’s factors in the two languages. One of these differences concerns syllable onset. Syllable onsets are constituents of the subsyllabic structure of German (Wiese, 1996) as shown in the following diagram (σ stands for syllable and C and V stand for consonant and vowel slots): equation(1) A syllable comprises the onset (start of the syllable) and rhyme (the syllable ending). The rhyme is generally further subdivided. Syllables have to have a phone that functions as the peak (usually a vowel), which is called the nucleus of the syllable. The coda is the sequence of one or more final consonants. All prevocalic consonants belong to the onset. This is where the main difference between English and German lies. In German an onset is obligatory. As such, syllables that start with a vowel (which would be the nucleus) have to have an onset added. In these cases a glottal stop is inserted as an onset to that syllable (Rogers, 2000). Although Wiese (1996) indicates that this onset would be a consonantal feature, he states, at the same time, that the glottal stop should not be analysed as a phoneme of Modern Standard German (p. 59). This means that a case can be made that there is no separate consonantal segment and the vowel gets the feature [+onset] which is then realised phonetically as a glottal constriction with a tendency to be located at the beginning of the vowel. This leads to stronger vowel onsets in German than in English. Previous work has shown that in both Dutch and Afrikaans (languages where there is also more articulatory tension in initial vowels) more words with initial vowels are stuttered (Uys, 1970 as cited in Bloodstein, 1995 and Vaane & Janssen, 1978). It would be predicted that this would also be the case for German. A second issue is that syntactical differences between English and German could influence the sentence position factor. (Brown, 1945 found that in English words earlier in a sentence were more likely to be stuttered.) Many learners of German consider the language to be fairly inflexible in its word order. This is mainly due to the fact that the finite verb and the past participle have rigid positions in German sentences. All finite verbs must be in second position in independent clauses, and non-finite ones in final position (with the exception of cases of extraposition of adjuncts or arguments). In dependent (subordinate) clauses, all verbs are in the final position. This inflexibility can, however, be contrasted with a much greater flexibility with regards to the position of the subject, direct object, and indirect object in German. The reason for this is that in German the case endings will always indicate how the constituents fit together syntactically (examples of this can be found in Johnson, 1998). According to Rommel, Häge, Johannsen, and Schulze (1997) and Rommel, Häge, Kalehne, and Johannsen (1999) disfluencies in a group of German stuttering children (of preschool age) appeared increasingly in the middle of sentences and words. It could be argued that the syntactical differences suggest that there would be more planning involved for words in later sentence positions in German compared with English. Thus it would be predicted that earlier sentence positions would not be as subject to increases in stuttering rate as they are in English. This is investigated below in German speakers and compared to Brown’s original results on English. In Bock and Levelt’s (1994) review of the planning of grammatical processes, this influence would fall into the stage of positional encoding. This stage determines the serial order of the lexical elements in an utterance. For function and content words, the predictions, following Brown (1945), are that content words (nouns, main verbs, adverbs, and adjectives that are an open class of words) will be stuttered more than function words (pronouns, articles, prepositions, conjunctions and auxiliary verbs that are a closed class of words) (Hartmann & Stork, 1972; Quirk, Greenbaum, Leech, & Svartvik, 1985). Function words display different phonological properties from content words (Kanerva, 1989, Kanerva, 1990, Nespor & Vogel, 1986, Selkirk, 1984 and Selkirk, 1996). The class of function words contains a set of highly practised high frequency words that are short (predominantly monosyllabic) in both English and German. Differences in the processing of content and function words are now increasingly found in fluent speakers both for the activation of their neurological substrates, using EEG studies (Bastiaansen, Van Der Linden, ter Keurs, Dijkstra, & Hagoort, 2002; Brown, Hagoort, & Keurs, 1999; Osterhout, Allen, & McLaughlin, 2002), in terms of lexical access (Segalowitz & Lane, 2000) and their role in a cross-linguistic examination of speech errors (Wells-Jensen, 2000). The function/content word distinction has also been investigated in an EEG study with individuals who stutter (Weber-Fox, 2001). Weber-Fox’s results indicated that the ERPs of people who stutter were characterised by reduced negative amplitudes for closed-class words, open-class words, and semantic anomalies in a temporal window of approximately 200–400 ms after word onsets. Differences between adults and children who stutter have been observed with respect to the function/content word distinction. An exchange of function and content words over ages has been observed in English people who stutter (Howell, Au-Yeung, & Sackin, 1999) and in Spanish speakers who stutter (Au-Yeung, Vallejo Gomez, & Howell, in press). In the exchange pattern, children are more disfluent on function than content words, whereas the opposite holds for adults giving something like a cross-over pattern for disfluency rate across these word types. For German an increased rate of function word disfluencies was also reported by Rommel (2001) in preschool children who stutter. For the word-length factor, stuttering rate would be expected to be higher on long words than short words, as Brown (1945) reported for English. The current study investigated Brown’s factors in spontaneous speech samples of German adults and children who stutter. The following research questions were addressed. At a general level, the paper investigated how Brown’s factors operate between English and German. For adults it was predicted that the stuttering rate of words starting with vowels would be high in German compared with English. Early sentence positions would have a lower level of stuttering than in English. The content word, and word length, factors are expected to operate in German in a similar way to English. Furthermore, in line with previous research it was predicted here that children who stutter would not be affected in the same way as adults by the factors associated with articulatory difficulty (Bloodstein & Gantwerk, 1967 and Bloodstein & Grossman, 1981). Specifically with respect to the function/content word distinction, it was expected that for children there would be less of a difference in disfluency rates on these word types compared with adults. Results will be discussed with reference to the EXPLAN theory of fluency failure (Howell & Au-Yeung, 2002). Central to this model is the assumption that function and content word disfluencies are different responses to the “planning” for some property of the content word not being ready for output. This emphasis on timing contrasts with other models, such as the covert repair hypothesis (CRH) (Kolk & Postma, 1997 and Postma & Kolk, 1993) where the speaker’s phonological system is viewed as prone to error. In the EXPLAN model, disfluencies on function and content words are in complementary distribution (repeating function words prevents content word disfluencies, and content words disfluencies ensue when speakers do not repeat function words). This would then account for age-related differences in word type disfluencies.