Phonological impairments in Hindi aphasics: Error analyses and cross-linguistic comparisons

ABSTRACT

We assessed phonological and apraxic impairments in Hindi persons with aphasia (PwA) and compared them to Italian PwA reported in previous studies. Overall, we found strong similarities. Phonological errors were present across production tasks (repetition, reading and naming), most errors were non-lexical and, among those, a majority involved individual phonemes. There were significant effects of length, but not frequency. Hindi PwA, like the Italian PwA, showed strong effects of syllabic structure, with most errors occurring on consonants and weak syllabic positions, preserving syllable structure and simplifying phonemes or syllabic templates. These similarities were modulated by some language-specific patterns. Vowel insertions were more common in Hindi, possibly due to the presence of a central vowel, and segmental simplifications concentrated on marked aspiration and retroflection features. We hope our study will encourage further research in Hindi and other Indian languages. This will improve clinical diagnosis and our understanding of cross-linguistic differences.

KEYWORDS:

• Hindi
• aphasia
• speech errors
• phonological errors

Acknowledgements

We are deeply grateful to the participants who took part in this study for their generosity with their time and effort. We would like to thank Ranjan Raj and his colleagues at Dr Baba Saheb Ambedkar Hospital for providing access to the participants. We are also grateful to Bipin Kumar for providing us with contacts at the University of Delhi. Finally, we are grateful to Mr and Mrs Sinha and Mr and Mrs Khanna for their hospitality during our work in India.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 From now on we will refer to phonological non-lexical errors as to phonological errors for short; we will refer to phonological lexical errors as to formal errors for short.

2 But see Romani, Galluzzi, Bureca, et al. (2011) for the argument that an increased decay of activation or a buffer impairment should produce exaggerated or exponential effects of length.

3 In this paper, we use articulatory planning/articulation to refer to a stage where phonology is unpacked into a set of commands ready to be sent to the muscle effectors. We do not make a distinction between articulatory planning and programming as made in other models (see DIVA/GODIVA model; Miller & Guenther, 2021). However, we do distinguish articulatory planning from motor execution with dysarthria being a deficit distinct from apraxia of speech and involving difficulties in sending messages to or receiving messages from the articulators.

4 Rice (1999) has outlined a useful list of non-phonological and phonological properties which characterize marked and unmarked phonological segments. Non-phonological properties include: (1) intra- and inter-language frequency (marked segments are rare), (2) acquisition (marked segments are acquired later), (3) implicational relations (a marked segment necessarily implies the presence of its unmarked compliment), (4) difficulty in articulation (marked segments require more effort), and (5) perceptual salience (marked segments are more perceptually distinct). Phonological properties include (1) Neutralisation: when a phonological contrast is neutralized in a linguistic environment this is generally to the unmarked counterparts and (2) epenthesis, assimilation, and deletion; these processes eliminate marked segments more than corresponding unmarked counterparts.

5 We assume that lexical representations play an important role in repetition as demonstrated by performance normally being better with words than nonwords (Romani, Galluzzi, & Olson, 2011).

Additional information

Funding

This research was funded by Universitas21 and the University of Birmingham Graduate School travel bursary to the first author.