ABSTRACT
This report presents evidence suggesting that the phoneme-based approach taken by Romani, Galuzzi, Guariglia, and Goslin (Comparing phoneme frequency, age of acquisition, and loss in aphasia: Implications for phonological universals. Cognitive Neuropsychology, this issue) falls short of capturing the complexity of articulation planning in patients with apraxia of speech. Empirical and modelling data are reported to demonstrate that the apraxic pathomechanism resides in the hierarchical architecture of phonological words rather than in the context-independent properties of phonemes. Because the factors determining complexity of articulation planning are interlaced between gestural, syllabic, and metrical levels, they cannot be captured by markedness rankings limited to any of these levels.
Acknowledgments
I am grateful to Ingrid Aichert and Anja Staiger for longstanding collaboration on this topic.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 The 13 samples were produced by 13 different patients. Any diacritics describing the gradual phoneme distortions and awkward phoneme transitions that were also present in most of them were omitted in the transcripts in (1), leaving only errors by which, on the surface, phonemes were added, deleted, or substituted by other phonemes.
2 The higher than predicted accuracies of longer words cannot be explained by lower markedness ranks of their consonants, because the empirical starting value of the consonant model was obtained from words whose consonants (all plosives) already had considerably low markedness ranks.