Sdsasdasd

Child phoneme errors are not
substitutions*


Peter Richtsmeier
University of Arizona


In this paper, I argue there is no reliable evidence for the concept of a “substitution” error in young children’s speech. First, I review research on child errors,including stopping, word-final devoicing, word-initial devoicing, fronting, and gliding. There are several general trends that can be seen in the data. First, the phonological analyses rely exclusively on transcription-based data. Second, phonetic analyses have repeatedly found that apparent cases of phoneme substitutions actually involve covert contrasts, or measurable phonetic differences in thequality of two productions that are indistinguishable to the adult ear. This leads to an apparent confound in the phonological literature—evidence for substitutions appears to be based solely on what adult listeners hear. I will take the strong and possibly equivocal position that substitutions are perceptual illusions and that the theoretical concept should be discarded.

1. IntroductionChild speech errors have been a focus of language acquisition research since the middle of the 20th century (e.g., Jakobson 1941; Velten 1943). Speech errors are of interest because they provide an important metric of the child’s phonological development (e.g., Peña-Brooks & Hegde 2000; Smit 1993), and they provide researchers with an easily obtainable empirical tool for testing theories of languagedevelopment (Hayes & Steriade 2004). This paper focuses on substitution errors, a type of error which is made when a child produces one or more component sounds of a word with a different sound or sounds. I will review a number of reports of substitutions in the literature, and then I will argue that these do not represent literal substitutions but cases of perceptual illusions on the part of theadult listener.
Before beginning, I would like to qualify the claim. The evidence and analysis of substitutions has largely been conducted within the field of phonology, whereas the counter-evidence comes from phonetic studies. In my view, certain facts about substitution errors are regularly left out of published phonological work, but I am not arguing that the core principles of phonologyshould be rejected, or that evidence of phonological organization is not to be found in child speech. Rather, I am focused solely on the claim that children substitute one sound for another. Where possible, I will offer an alternative explanation for the proposed substitution, but most of these explanations are speculative. Therefore, the primary purpose of the paper is to offer evidence againstsubstitutions being a viable description of child speech data.
The structure of the paper is as follows. I review published data on five purported substitution errors: stopping, word-final devoicing, word-initial devoicing, fronting, and gliding. The discussion of each type of error includes some data from a large corpus-study, The Iowa Articulation Norms Project (Smit, Hand, Freilinger, Bernthal &Bird 1990, also discussed by Smit 1993). I also review in general terms some phonetic and phonological positions for describing or explaining the data, and why, in each case, the phonological description is not to be preferred. I conclude by summarizing the evidence against substitutions and make suggestions for how the study of child speech might proceed.

2. Stopping

Stopping refers to thesubstitution of a fricative for a stop. N. V. Smith (1973) gives the following examples of stopping from two-year-old Amahl, a child acquiring English. For present purposes, patterns other than stopping can be ignored.

|(1) |Word |Child’s | |
| | |Production | |
| |knob |[nap] |(1;10) |
|...