Homographic and heterographic homophones in speech production: Does orthography matter?

This paper investigates homophone naming performance in an individual with impaired word retrieval. The aim of the study is to investigate the status of homophone representations using treatment of homophone picture naming in aphasia. The focus of this paper is the representation of heterographic homophones (words which sound the same but are spelled differently, e.g., ‘knight’ vs. ‘night’). Additionally, we replicate and expand previous findings regarding homographic homophones of Biedermann and Nickels (2008) in English and Biedermann et al. (2002), in German. eoretical positions about the mental representation of homophones are tested. First, do homophones – regardless of whether they are spelled the same or differently – share a phonological word form (e.g., Levelt et al., 1999; Dell, 1990)? Or second, do they have independent phonological word forms? (e.g., Caramazza et al., 2001; Miozzo and Caramazza, 2005)? In addition, might it be the case that homographic and heterographic homophones behave differently in word production reflecting different word form representations? These theoretical accounts are put to the test by looking at the generalisation of improvement following the treatment of homophone naming in aphasia, in particular, whether picture naming improves for both homophone meanings if only one is treated using a phonological cueing hierarchy. d and untreated homophones improved significantly, regardless of their spelling. Homographic and heterographic homophones showed the same pattern of generalisation. There was no generalisation for phonologically related controls. The pattern of generalisation extends our previous findings (Biedermann et al., 2002; Biedermann and Nickels, 2008) by showing evidence that heterographic homophones benefit to the same extent as homographic homophones. These results are interpreted as favouring a theory where both homographic and heterographic homophones share a single phonological representation. It is inferred that facilitation of naming takes place at the level of phonological representations, where orthography seems to have no influence.