@article{jbp:/content/journals/10.1075/msw.00007.tay, author = "Tay, Dennis", title = "Surveying views of metaphor vs. literal language in psychotherapy: A factor analysis", journal= "Metaphor and the Social World", year = "2020", volume = "10", number = "2", pages = "273-291", doi = "https://doi.org/10.1075/msw.00007.tay", url = "https://www.jbe-platform.com/content/journals/10.1075/msw.00007.tay", publisher = "John Benjamins", issn = "2210-4070", type = "Journal Article", keywords = "confirmatory factor analysis", keywords = "experimental survey", keywords = "metaphor functions", keywords = "exploratory factor analysis", abstract = "Abstract

Five key therapeutic functions of metaphors are often discussed by psychotherapists. They (i) help clients express emotions and experiences, (ii) help therapists and clients explain difficult concepts, (iii) introduce new frames of reference, (iv) help work through client resistance, and (v) build a collaborative relationship between therapists and clients. Research on how these functions are enacted in psychotherapy talk tends to assume that they are indeed perceived as such by clients, and that metaphorical language is preferred to comparable literal language in performing them. This paper reports a survey study (N = 84) to critically interrogate these assumptions. Participants read two constructed therapy dialogues, controlled and counterbalanced for presentation sequence, where therapist and client discuss an issue using metaphorical and literal language respectively. Each dialogue is followed by a 15-item questionnaire to rate how well the presumed functions were performed (e.g. the therapist and client can work effectively together, the therapist is able to explain difficult concepts). A combined Confirmatory (CFA) and Exploratory Factor Analysis (EFA) suggests that, instead of the five distinct functions proposed in the literature, participants discerned three functions which reflect a more holistic view of what metaphors can do. A second EFA conducted on literal responses yielded only two factors. This contrast in factor structure further suggests that (i) literal language is less functionally nuanced, and (ii) metaphors are not simply perceived as an ‘add-on’ to literal language, but are evaluated across an extended narrative in fundamentally different ways. Within-subjects metaphor vs. literal ratings of the items under the emergent three-factor structure were then compared. Metaphor ratings were significantly higher in all factors (p < 0.01), suggesting that metaphorical language is indeed perceived as more effective than literal language when discussing clients’ issues. Implications, limitations, and future directions are discussed.", }