The Emotional Verbal Fluency Task: A Close Examination of Verbal Productivity and Lexical-Semantic Properties

Boji P W Lam,Thomas P Marquardt

DOI: https://doi.org/10.1044/2020_JSLHR-19-00276

2020-07-20

Abstract:Purpose Emotional verbal fluency (Emo-VF) has the potential to expand neuropsychological assessment by providing information about affective memory retrieval. The usability of Emo-VF is limited, however, by significant variations in task administration and the lack of information about Emo-VF responses. This study investigated verbal productivity and the lexical-semantic properties of responses on positive and negative Emo-VF tasks. Comparing Emo-VF to non-Emo-VF tasks used regularly in neuropsychological assessment provided additional information about the basic characteristics of Emo-VF tasks. Method Twenty-five adult native speakers provided verbal responses to three Emo-VF ("happy," "sad," "negative emotions") and two non-Emo-VF categories ("animals," "things people do"). Verbal productivity was measured at the word and syllable levels. Multiple large-scale data corpora were used to estimate the lexical-semantic properties of the verbal responses. Results There was a robust positivity bias in verbal productivity within Emo-VF tasks. Emo-VF tasks tended to elicit longer words than "animals" and "things people do," which might impact the results of verbal productivity analyses, especially in comparisons with "things people do." Within Emo-VF tasks, negative Emo-VF elicited words from a wider range of valence than positive Emo-VF tasks. Similarities (e.g., word length and complexity) and differences (e.g., concreteness, age of acquisition) were found between positive and negative Emo-VF tasks. Conclusions The study provided information about the basic characteristics of Emo-VF tasks, which included evidence for a robust positivity bias, suggestions for analyses of verbal productivity (e.g., consideration of word length), and lexical-semantic properties associated with positive and negative Emo-VF tasks using corpora data.