| |
Descriptive linguistic methods have long been used in the analysis of fully developed primary languages. These same methods are also well-suited to the study of language development, particularly the analysis of children's speech sound systems. Descriptive methods are a preferred analytic tool because they are designed to gather evidence that reveals the hallmark and defining characteristics of a sound system, independent of theoretical orientation, age, or population of study. The defining properties of descriptive linguistic analyses of children's sound systems are discussed in this article.
The Phonetic Inventory
A phonetic inventory comprises all sounds produced or used by a child, regardless of whether those sounds are correct relative to the intended (adult) target. In the acquisition literature, the conventional criterion for determining the phonetic status of sounds is a two-time occurrence independent of the target or context; that is, any sound produced twice is included in a child's phonetic repertoire (Stoel-Gammon, 1985). Children's phonetic inventories reflect the range of individual variability expected in development. As such, complementary methods have been designed to further depict developmental variation, including the phone tree methodology (Ferguson and Farwell, 1975) and the typology of phonetic complexity (Dinnsen, 1992). For children with speech sound disorders, the phonetic inventory may be quite large despite errors of production, and may consist of sounds that do not occur in the ambient language.
The Phonemic Inventory
Phonemes are used to signal meaning differences in a language. Phonemes are conventionally determined by the occurrence of minimal pairs. A minimal pair is defined as two words identical except for one sound, for example “pat” and “bat” or “cap” and “cab.” Here, the consonants /p/ and /b/ are the only point of difference in each pair of words; therefore, these would be said to function as phonemes in the differentiation of meaning. For children, the phonemic inventory is generally smaller than the phonetic inventory (Gierut, Simmerman, and Neumann, 1994). Gaps in the phonemic repertoire often affect the sound classes of fricatives, affricates, and liquids. From a linguistic perspective, the nonoccurrence of these sound classes in children's speech parallels markedness. Markedness defines lawful relationships among sound categories that have been found to hold universally across languages of the world. One type of markedness is implicational in nature, such that the occurrence of property X in a language implies property Y, but not vice versa. The implying property X is taken to be marked, and is presumably more difficult to acquire, whereas the implied property Y is unmarked and predictably easier to learn. In development, then, phonemic gaps in the inventory correspond to more marked (difficult) structures of language. In linguistic terminology, these gaps would be characterized as a type of phonotactic constraint (Dinnsen, 1984).
The Distribution of Sounds
Distribution refers to where sounds (phones or phonemes) occur in words and is determined by examining context. For children, sounds may be used in all word positions, initial, intervocalic, and final, or they may be limited to certain contexts. In development, obstruent stops commonly occur word-initially but not postvocalically; whereas fricatives and liquids commonly occur postvocalically but not word-initially (Smith, 1973). As with the phonemic inventory, restrictions on the distribution of sounds correspond to markedness, with children having a tendency toward unmarked as opposed to marked structure.
Rule-Governed Alternations
Asymmetries in the distribution of sounds may be further indicative of systematic rule-governed alternations in sound production (Kenstowicz, 1994). Rule-governed alternations occur when morphologically related words are produced in different ways, for example, “electric” but “electricity.” Alternations are typically sampled by adding either a prefix or suffix to a base word in order to change the context in which a sound occurs. There are two general types of rule-governed change: allophonic variation and neutralization. Allophonic variation occurs when a single phoneme has multiple corresponding phonetic outputs that vary by context. An example is /t/ produced as aspirated in word-initial position “tap,” as flap in intervocalic position “bitter,” and as unreleased in word-final position “it.” In each case, the target sound is /t/, but the phonetic characteristics of the output differ predictably by word position. Thus, there is a one-to-many mapping between phoneme and phones in allophonic variation. Neutralization occurs when two or more phonemes are merged into one phonetic output in a well-defined context. An example is /t/ and /d/ both produced as flap in intervocalic position “writer” and “rider.” In neutralization, the contrast between phonemes is no longer apparent at the phonetic (surface) level. Consequently, there is a many-to-one mapping between phonemes and phone. In children, the emergence of target-appropriate morphophonemics occurs later in language development. For children with speech sound disorders, nontarget allophonic variation and neutralization have been observed and parallel the rules of fully developed languages of the world (Camarata and Gandour, 1984).
| |
