From phonological paraphasias to the structure of the phonological output lexicon

Abstract

This study examined how phonological information is represented and retrieved, using analysis of naming errors in anomia. We focused on questions that relate to metrical and segmental information, the types of information each of them contains, and whether they are organised in parallel or serially. Nine Hebrew-speaking individuals with anomia due to phonological output deficit named 200 pictures. A detailed analysis of the 208 phonological paraphasias they produced, at the group level and at the individual level, revealed errors preserving only segmental information, errors preserving only metrical information (number of syllables and stress pattern), and errors preserving partial information of both types. There were also errors in the order of segments. The pattern of errors indicates that metrical information and segmental information are accessed in parallel rather than serially, and are merged at a later stage in which the segments are inserted into the word form; and that segmental information includes consonants and vowels, and involves information about their identity as well as about their relative position. Information about the number of syllables seems to be retrieved together with information about the stress pattern. The analysis also showed preservation of phonological principles.

We thank Drik-Bart den Ouden, Outi Bat-El, Gary Dell, Argye Hillis, Aviah Gvion, Rama Novogrodsky, and Ronit Szterman for comments on an earlier version of the paper and for stimulating discussions of this study. The research was supported by the Joint German-Israeli Research Program grant in Neuroscience GR01791 (Friedmann).

Notes

¹This discussion of serial versus parallel access to metrical and segmental information should not be confused with the discussion over whether the slot-filling process occurs in a serial (segment-by-segment) or a parallel (all the segments of a phonological word or morpheme at once) fashion (see Den Ouden & Bastiaanse, 2004).

²There were also studies that examined what was the most robust phonological information in phonological errors. Valdois, Joanette, and Nespoulous (1989) for example, examined the phonological information—number of syllables, segments and CV structure—preserved in the approximations of three aphasic patients, and found that number of syllables was the most available information for all the participants and the information about the CV structure was the least available. Other studies carried out different analyses of the phonological information conveyed by the phonological errors (e.g., Gagnon, Schwartz, Martin, Dell, & Saffran, 1997; Lecours & Lhermitte, 1969; Whilshire, 2002.)

³As suggested by Butterworth (1989) and Caramazza and Hillis (1990), such semantic errors could stem from a deficit in accessing the phonological representation of the target word, rather than from a deficit at the semantic representation. When there is no access to the phonological representation of the target word, a word which is semantically related to it, and its phonological representation is available, might be produced instead.

⁴“No response” in phonological anomia might indicate that both metrical and segmental information are unavailable, making the participant unable to produce any response, or that the speaker only has metrical structure, and possibly metrically correct neologisms that include no segments of the target are perceived too far from the target and are therefore ‘edited out’ by speakers with good self-monitoring, leading to no response.

⁵Translations of the examples will be given only in case the error forms an existing word, to enable the non-Hebrew speakers to evaluate the semantic relations between the target and the errors. When the error is a phonemic error that forms a nonword, a translation is not given. The stressed syllable is marked in bold.

⁶A reviewer has suggested that a possible reason for the small number of responses that preserve only metrical information is that responses that preserve metrical information without segmental information might be screened out by the aphasic speakers, and cases in which only metrical information is available to them might result in no response (or possible circumlocutions and semantic errors). An interesting way to assess in these cases whether metrical information is available is to ask questions that relate to metrical information when no response is given, which could uncover metrical knowledge even when no response is offered.