On nouns, verbs, lexemes, and lemmas: Evidence from the spontaneous speech of seven aphasic patients

Abstract

Background: Although disproportionate impairment of noun or verb retrieval has been described on the basis of the evidence from several aphasic cases since the mid 1980s, with different theoretical frames being proposed to account for noun–verb dissociation, very few studies have dealt with this dissociation in spontaneous speech.

Aims: The objectives of this study were to investigate (i) whether the dissociation also emerged in connected speech, and (ii) whether the analysis of patients' narratives could shed light on the functional damage underlying their grammatical-class-specific impairment.

Methods & Procedures: Two non-fluent verb-impaired patients, two fluent verb-impaired patients, and three fluent noun-impaired patients participated in this study. Their noun–verb dissociation was preliminarily assessed through a picture-naming task, following which their spontaneous speech was collected and analysed using a single-case approach, taking into consideration both lexical productivity (as indicated by the number of different tokens produced by the patients) and lexical diversity (as indicated by the number of different types and stems used by the patients).

Outcomes & Results: Non-fluent verb-impaired patients tended to produce a lower proportion of verb types than unimpaired control participants, as opposed to fluent verb-impaired patients, who produced a normal verb rate in their spontaneous speech on all counting procedures. One out of three fluent noun-impaired patients produced a lower proportion of noun tokens, types, and stems compared to normal speakers.

Conclusions: The data presented in this paper indicate that noun–verb dissociation as assessed in picture-naming tasks might not emerge in spontaneous speech and indicates the need for the inclusion of a lemma level in models of word production that aims at explaining grammatical-class-specific impairments in people with aphasia.

Keywords:

• Grammatical class
• Noun–verb dissociation
• Spontaneous speech
• Lexical diversity
• Type
• Token
• Stem
• Lemma
• Lexeme

Acknowledgments

The study was supported by research grant no. 2003107555/001 from the Ministero dell'Università e della Ricerca Scientifica (MIUR; Italian Ministry of Education, University and Scientific Research). Preliminary results of this study were presented at the 7th Science of Aphasia Conference, Alghero (Italy), 7–12 September 2006, and at the 44th Annual Meeting of the Academy of Aphasia, Victoria (Canada), 15–17 October 2006. We thank Mirella Frustaci, Graziella Ghirardi and Paola Saletta for their help in collecting the spontaneous speech samples.

Notes

¹The overall rate of self corrections was .04 (22/490); however, only 10 of the 22 repairs (2% of all responses) were produced within 3 seconds from the presentation of the stimulus and were thus considered as correct.

²The formula used to calculate the modified t value as suggested by Sokal and Rohlf (1995) is as follows:

where X ₁ represents the individual's score, X ₂ is the mean of the normative sample, s ₂ is the standard deviation of the normative sample, and N ₂ represents the normative sample size.

³The same verb stem generates more than 40 different inflected forms differing on three dimensions (person, tense, and aspect), while nouns can only appear either in their singular or their plural form.

⁴For evidence on the absence of priming effects when a piece of information is severely impaired, see for example Marslen-Wilson and Tyler (1997).

⁵Note that nouns do sub-categorise complements, but this happens only exceptionally, for example in the case of abstract deverbal nouns (e.g., destruction) whose argument structure percolates from the corresponding verb.