Phonological Errors: What they tell us about aphasia, language processing and linguistic universals
Research Seminar at the University of Reading, 01/11/13
Speaker: Cristina Romani, Aston University (UK)
Diagnosing phonological impairment and apraxia of speech (AoS) is problematic. This is due to high co-occurrence of phonological and phonetic errors in aphasics.
Phonological impairments are difficulty in selecting and ordering phonological forms for words, which patients are sometimes unaware of. On the surface such errors are deletions, substitutions, transpositions and insertions of phonemes; there can also be a lack of muscle tone.
AoS is a motor planning speech disorder where patients access the correct phonological forms but cannot articulate them; as such it is characterized by inherently phonetic errors. These are distortion of sounds and floundering for phonemes with noticeable effort. Patients are aware of their speech impairments and muscle tone remains intact. In clinical practice it is important to determine the origin of the phonological error, whether it is intrinsically phonological or phonetic. The problem is that these errors can be mistaken for one another since both results in dysfluent speech.
Hypothesis (1) if phonological errors are all of one type then AoS does not exist and is a combination of phonological and articulatory impairment. This implies there is a single phonological deficit. Hypothesis (2) if phonological errors arise for different reasons then AoS is a separate impairment.
Using an independent measures design 24 Italian aphasics (apraxic = 11, phonological = 9, mixed= 4), where the mixed group leaned towards phonological impairment, were given tasks of repetition and reading of single words and picture naming.
Data was analysed using error analysis according to phonological and syllabic complexities in the types of errors, contrasts, tasks and the different methodology in all 3 groups.
Repetition and naming errors were similar across groups: Naming was more impaired than repetition, and the salient repetition error was substitution. The groups differed in rates of simplification. In all tasks apraxic patients used simpler syllable structures such as consonant-vowel (CV); simplified sonority such as devoicing voiced consonants; simplifying place and manner of articulation for example fricatives produced more errors than stop consonants.
No error pattern was found in phonologically impaired or mixed patients, suggesting the reason for their errors was due to selection. In AoS phonological errors were made systematically to make words simpler to produce. Evidence for linguistic universals was sought by comparing aphasia to phonological development, since simple constructions are the first to be acquired and the last to be lost in aphasia.
Secondly, the study compared child and aphasic errors in speech. Cluster reduction moved towards consonant and vowel equilibrium, cluster assimilation and simplifications were made in voice, manner and place of articulation; showing these linguistic features to be universal. What is preserved in error indicates the form of linguistic representation.
In all groups onset heads in hetero and homo-syllabic clusters and cluster structures were preserved. Thus syllable structure – CVCV – is stored with the lexical concept and is not a product of the articulatory level as is often thought. This study shows there are different types of phonological errors thus AoS exists and needs different treatment to phonological impairment.