Typological predictions in developmental phonology.

Two common and seemingly independent error patterns, namely consonant harmony and gliding, are examined for their typological characteristics based on cross-sectional and longitudinal evidence from young children's developing phonologies. Data are drawn from the published literature and from the developmental phonology archives at Indiana University. An asymmetry is observed such that the occurrence of harmony is found to imply the occurrence of gliding, but not vice versa. While this finding would be unexpected within contemporary derivational theories, it can be shown to follow within optimality theory from a fixed universal ranking relationship among certain constraints. Optimality theory is also argued to offer a viable developmental account with clinical implications that can serve as a further test of the theory.

On the characterization of a chain shift in normal and delayed phonological acquisition.

Several theoretical and descriptive challenges are presented by children's phonological substitution errors which interact to yield the effect of a chain shift. Drawing on an archival study of the sound systems of five children (ages 3;5 to 4;0) with normal development and 47 children (ages 3;4 to 6;8) with phonological delay, one such chain shift, namely the replacement of target /theta/ by [f] and the replacement of /s/ by [theta], was identified in the speech of six children from the two subgroups. Different derivational and constraint-based accounts of the chain shift were formulated and evaluated against the facts of change and the children's presumed perceptual abilities. An adequate account in either framework was found to require the postulation of underspecified and, in some instances, nonadult-like underlying representations. Such representations were able to reconcile within a single-lexicon model the presumed production/perception dilemma commonly associated with acquisition. Continuity was also preserved by limiting underlying change to just those lexical items which exhibited a change phonetically.

Context-sensitive underspecification and the acquisition of phonemic contrasts.

Several competing proposals for the (under)specification of phonological representations are evaluated against the facts of phonemic acquisition. Longitudinal evidence relating to the emergence of a voice contrast in the well-documented study of Amahl (from age 2;2 to 3;11) is reconsidered. Neither contrastive specification nor context-free radical underspecification is capable of accounting for the facts. The problem is in the characterization of the change in the status of a feature from being noncontrastive and conditioned by context at one stage to being contrastive with phonetic effects that diffuse gradually through the lexicon. Both frameworks must treat as accidental the persistence of the early substitution pattern and require the postulation of wholesale changes in underlying representations, where these changes do not accord well with the observed phonetic changes or with the facts available to the learner. Context-sensitive radical underspecification provides a plausible account of each stage and the transition between stages with minimal grammar change.

Spectral analysis of target-appropriate /t/ and /k/ produced by phonologically disordered and normally articulating children.

Previous research (Forrest, Weismer, Hodge, Dinnsen and Elbert, 1990) has shown that some phonologically disordered children differentially mark seemingly homophonous phonemes; however, the resulting contrast may be spectrally distinct from that produced by normally articulating children of the same age. In the present investigation possible sources for these differences between normally articulating and phonologically disordered children's productions of target-appropriate phonemes were pursued. Spectral characteristics of seemingly correct productions of /t/ and /k/ in word-initial position were analysed for four normally articulating and seven phonologically disordered children to assess the effect of recency of acquisition, depth of knowledge of the contrast and/or the effect of a phonological disorder on accuracy and variability of production. Results revealed that children who had acquired the velar-alveolar contrast more recently, and who had incomplete knowledge of that contrast, produced target-appropriate /t/ and /k/ differently from their normally articulating peers and other phonologically disordered children with greater knowledge of the contrast. Further, the phonologically disordered children with incomplete knowledge of the velar-alveolar contrast were less variable than the other phonologically disordered or normally articulating children in the spectral characteristics across repeated productions. Analysis of the spectral characteristics of word-initial /t/ and /k/ at a later point in time indicated similarities between all speaker groups in the spectral parameters that distinguished the velar from the alveolar stop. However, the stability of these parameters across repeated productions decreased for the phonologically disordered children with greater knowledge of the contrast. These effects are related to motor skill development and found to be consistent with previously demonstrated patterns of skill acquisition.

Consonant clusters in disordered speech: constraints and correspondence patterns.

Comparison of patterns of cluster realization from 47 children ranging in age from 3;4 to 6;8 with functional (non-organic) speech disorders with those reported in the literature for normal acquisition reveals that these patterns are essentially the same for both groups. Using a two-level generative phonology for children's independent systems, further analysis of cluster realizations by means of feature geometry and underspecification theory reveals that there are systematic and principled relationships between adult representations of clusters and children's underlying representations and between children's underlying representations and their phonetic representations. With special emphasis on coalescence phenomena, it is suggested that the apparent diversity in children's cluster realizations can be reduced to four constraints on the form of underlying and phonetic representations.

Stimulability as a factor in the phonological generalization of misarticulating preschool children.

The relationship among six functionally misarticulating preschool children's phoneme-specific stimulability skills, the choice of treatment targets (i.e., stimulable or nonstimulable sounds), and generalization of correct sound production was explored in this prospective study. Each subject [age range of 4:11 (years:months) to 5:6] was taught to produce [r] and one other sound that was absent from his or her phonetic inventory using a contrasting-minimal-pairs production approach. A multiple baseline across behaviors single-subject research design provided experimental control. For 86% of the 28 monitored sounds, generalization was consistent with pretreatment stimulability skills; production of stimulable sounds tended to improve regardless of treatment target. These results suggest that nonstimulable sounds are likely to require direct treatment; thus, generalization probe responses may be maximized by treating nonstimulable sounds rather than stimulable sounds.

Generalization to conversational speech.

Although changes in children's phonological systems due to treatment have been documented in single-word testing, changes in conversational speech are less well known. Single-word and conversation samples were analyzed for 10 phonologically disordered children, before and after treatment and 3 months later. Results suggest that for most of the children, there were system changes in both single words and in conversational speech. It appears that many phonologically disordered children are able to extend their correct production to conversation without direct treatment on spontaneous speech.

Some constraints on functionally disordered phonologies: phonetic inventories and phonotactics.

The phonological systems of 40 functional misarticulators, ages 40 to 80 months were examined in terms of the nature and variation of phonetic inventories and phonotactic constraints. It was found that these properties of disordered systems were governed by severe constraints that yielded a typological characterization of such systems along with associated implicational laws. The principles governing disordered systems were also found to parallel closely the principles governing normal first language acquisition. The evidence suggests that at least these properties of disordered systems represent delays in the normal acquisition process and are not otherwise deviant. The assessment and treatment of functional disorders along with the projection of learning patterns can thus appeal to principles governing such phonological systems.

A functional analysis of phonological knowledge and generalization learning in misarticulating children.

It has been suggested that a child's productive phonological knowledge may be one factor that potentially accounts for individual differences in generalization learning observed among phonologically disordered children (Dinnsen & Elbert, 1984; Elbert, Dinnsen, & Powell, 1984). This paper evaluates the hypothesis that productive phonological knowledge influences generalization. Three related studies involving 6 functionally misarticulating children were conducted. In the first study, a description of each child's phonological system was developed using procedures of standard generative analysis. Based upon these descriptions, each child's productive phonological knowledge of his or her own sound system was determined and then ranked on a continuum ranging from "most" to "least" knowledge relative to the adult target. The second study implemented an experimental treatment program based upon each child's productive phonological knowledge, with treatment sounds selected directly from each child's continuum of knowledge. The third study reassessed each child's productive phonological knowledge following treatment. The results of these three studies indicated that a child's productive phonological knowledge of the sound system influenced the overall amount of generalization learning. However, the extent of generalization learning was associated with the point on the knowledge continuum at which treatment was initiated. These findings are discussed with reference to individual differences in generalization learning.

On the prediction of phonologic generalization learning patterns.

The purpose of this study was to obtain information on six children's misarticulation of consonant clusters in order to illustrate how certain factors influence generalization learning patterns. While all subjects evidenced generalization, individual differences in learning patterns did emerge. These patterns were explained in terms of the following three factors: (1) information about the children's unique knowledge about the phonologic system, (2) the linguistic relationships among sounds, and (3) the interaction of these factors with the treatment target. The implications of this explanation are discussed as they relate to the selection of treatment targets that will facilitate generalization.