• No results found

CHAPTER 2: SPEECH SOUND ACQUISITION IN PERSPECTIVE

2.3 IMPLICATIONS FOR FUTURE STUDIES OF PHONOLOGICAL ACQUISITION

2.3.3 CONCLUSIONS

Although there are quite a number of differences between these studies in terms of methodology and criteria, there are several trends that can be detected. In the first instance, there is a gradual development of speech sound acquisition across languages. Secondly, all studies which have examined the age of vowel acquisition have shown that vowels are mastered earlier than consonants. Thirdly, it has been documented that consonant acquisition tends to follow a similar course in different languages, which suggests that speech sound acquisition takes place along universal principles. Especially, the manner of articulation provides some clear evidence of such universal patterns. In many languages, nasals and plosives are mastered before fricatives, and fricatives before affricates. This has been confirmed in studies on English (Templin 1957, Prather et al. 1975, Dodd et al. 2003) and is also attested in several other languages, including Zulu (Naidoo 2003), Cantonese (So & Dodd 1995),

Xhosa (Mowrer & Burger 1990), German (Fox 2000), Putonghua (Zhu 2002) and Jordanian (Amayreh 1994). This finding can be explained by the fact that some speech sounds tend to occur more universally than others. Specifically, plosives occur in all the languages of the world, while 96.45% have nasals (UCLA Phonological Segment Inventory Database (UPSID). 96.23% have approximants: all these sounds are acquired relatively early by children across languages. The percentage of fricatives and affricates in the languages of the world is 93.13% and 66.52 %.( UCLA Phonological Segment Inventory Database (UPSID)) , respectively, and these sounds are typically acquired late. The percentage of fricatives raises problems for the earlier claim that fricatives are nearly as common as nasals, but nevertheless fricatives are acquired later than nasals. Locke (1983) and Stokes &

Surendran (2005) suggested that articulatory complexity has to be taken into account to explain why some speech sounds tend to be acquired relatively late although they tend to occur more universally than others.

Some researchers make a further connection between the frequency of sounds in the o ld s la guages a d the atu al ess of sou ds Vih a , as the spee h sou ds hi h a e a e i la guages ould e ega ded as u atu al a d the efo e tend to develop later.

In addition to the universal tendencies which have been observed in previous

studies, language-specific influences have also been noticed. Each study has reported particular patterns, both in the order and in the age of speech sound acquisition, as well as in the error patterns. For example, the speech sound /v/

shows considerable variation in acquisition rate amongst learners of different languages. It appears to be acquired earlier in German, Turkish and Xhosa than in English. Another example is the acquisition age of /ʃ/, which for Turkish children was found to be 1:11 (Topbas & Yavas 2006), while for English children it emerged at the age of 5:11 (Dodd et al. 2003). Some researchers have attributed these differences in acquisition rate to differences in the phonology of the various languages. Ingram (1989), for example, pointed out that a sound will be pronounced correctly earlier in a language in which it occurs more frequently and in which it has a greater functional load. Although the idea of functional load refers to the relative importance of each speech sound within a specific phonological system (Zhu 2002:

15), it has been widely argued that functional load is difficult to calculate across languages.

Phonological error patterns also provide evidence for the universality of acquisition as a great deal of similarity has been observed across languages. For example, Fox (2000), and Mowrer & Burger (1990) demonstrated that most error patterns in English also occur in German and Xhosa. Such errors include fronting, stopping, cluster reduction, weak-syllable deletion and final consonant deletion. Yet,

differences in error patterns across languages also provide evidence of the inherent variability of phonological acquisition across languages. One example is that Amayreh (1994) found that most of the errors in the speech of Jordanian children acquiring Arabic related to emphatic consonants, which reflects a language-specific issue. Another example comes from Mowrer & Burger (1991), who reported that Xhosa children made fewer errors on plosives and fricatives than English-speaking children. A third example comes from Zhu (2002), who indicated that there is a difference in the error patterns produced by children speaking Modern Standard Chinese and children speaking other languages. Specifically, syllable-initial consonant deletion and backing are the second most frequent error patterns in Chinese children, yet they are considered atypical error patterns in English children.

Considerable variation in speech sound acquisition between different languages has been reported in the literature, but differences between studies which have investigated the same language have also been noted. For example, there is disagreement about the age of speech sound acquisition in studies which have examined English children. Templin (1957), for example, found that the speech sound /r/ was mastered around the age of 4, while Smit et al. (1990) reported that this speech sound was not acquired until the age of 8:0. Yet another example of differing results from studies of the same language comes from Saleh et al. (2007),

sounds correctly in final position, which runs counter to the findings reported in Ammar & Morsi (2006) and Amayreh (1994) for speakers of Arabic. Other data f o “aleh et al. e ealed that the o so a ts / , h , ʃ, / a e pa t of Egyptian phonemic inventory before age 2:6 while Omar( 1973) indicated that these sounds were not acquired by Egyptian children until age 4:6.

2.4 ARTICULATION TESTS

Most studies of hild e s spee h de elop e t a e ased o a ti ulatio tests.

Some studies used a test which had previously been designed, while others designed their own diagnostic instrument. Before proceeding, it will be useful to clarify the difference between an articulation test and phonological assessment. In a phonological assessment, the speech and language therapist attempts to evaluate the phonological system (i.e. the rules of the sound system of a language including the combination of these sounds into intelligible speech) and the purpose of phonological testing is to determine and categorize error patterns which enable an understanding of the client's phonological system. In an articulation test, the SLT assesses the production of consonants and vowels. An articulation test can be a phonological test depending on the manner in which the researcher analyses the results. For example, if he/she looks at speech sound classes and attempts to find