Principles of Clinical Phonology by Ball Martin J

Author:Ball, Martin J. [Martin J. Ball]
Language: eng
Format: epub
Publisher: Taylor & Francis (CAM)

7.1.5 Phonological Encoding

Of most interest to this book is the area of phonological encoding (together with phonetic implementation /articulation). Apart from the debate about two- versus one-lexicon models described earlier, different models of phonological encoding have also been debated in the literature. Harley (2001) describes a ‘frame and slots’ approach as one of the early proposals (for example, the ‘scan–copier’ model of Shattuck-Hufnagel, 1979). This approach assumes that linguistic structures comprise a syllabic frame with empty slots, and that the relevant phonological units are retrieved from a phonological lexicon and slotted into the frame. Such an approach can account for anticipatory and perseveratory speech errors (Harley, 2001: 371, gives the examples ‘heft hemisphere’ and ‘left lemisphere’).

A second approach is that of Levelt, Roelofs, and colleagues (e.g., Levelt, Roelofs & Meyer, 1999, the WEAVER++ model) where, as described earlier, there is separation between lemma activation and phonological encoding. However, this model posits a separate syllabary. Harley (2001) notes that the syllabary stores commonly occurring syllables, together with their stress patterns, and thus allows the emergence of phonotactic constraints to inhibit errors with non-native phonotactics. Such a store also allows processing time to be speeded up, as speakers do not have to phonologically encode every lemma from scratch. In this model, then, phonological encoding involves both segmental and prosodic information; the overt recognition of prosodic encoding is clearly to be welcomed by clinical phonologists (see chapter 9). We will look in more detail at this model in section 7.3.

Finally, we can consider the connectionist approach of Dell and colleagues (e.g., Dell, 1988). Connectionist models allow for the emergence of certain features following on from a bundle of similar connections. Thus, Dell’s spreading activation theory (SAT, also called ‘the Interactive Lexical Network’) suggests that recurrent networks allow the emergence of connections between words and their phonological forms without the need for explicit frames. In this way, the division between structure and content can be avoided. Again, because this is one of the dominant models of speech production and its assumptions may be of use to clinical phonology, we will devote more space to it in section 7.2.

It should be noted that the WEAVER++ model came out of research on the speeds at which normal speakers process various stimuli (e.g., visual), whereas the SAT grew directly out of attempts to account for speech as produced by aphasic speakers (which we return to later). Although the models are distinct, they have been moving closer to each other, as described in Levelt (1999a). Models designed especially with child speech acquisition and disorders in mind, such as the Stackhouse and Wells (1997) model, will be looked at in section 7.4 and returned to in more detail in chapter 8 when we consider models of speech perception.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.