Line Kruse, Department of Linguistics, Cognitive Science and Semiotics, Aarhus University
Roberta Rocca, Psychoinformatics Lab, Department of Psychology, University of Texas at Austin
Mikkel Wallentin, Department of Linguistics, Cognitive Science and Semiotics, Aarhus University
Background: Spatial demonstratives are highly frequent linguistic universals (Levinson, 2018; Diessel, 1999), with at least two contrastive expressions (“this” vs. “that”) indicating physical, emotional, or functional proximity of the speaker to the referent (Kemmerer, 1999). Recent evidence indicates that even in the absence of a communicative context, demonstrative choices are highly consistent across individuals, suggesting that demonstratives in such cases reflect experienced or emotional proximity to the self in a mental space (Rocca et al., 2019). Further, these representations appear to be related to the semantic features of the referent, such as valence and manipulability (Rocca et al., 2019). Depression is a disorder consistently characterized by a maladaptive focus on the self, manifest in language-use such as increased use of first-person pronouns (Holtzman, 2017). Further, neural evidence has suggested impaired self-referent emotional processing (Miskowiak et al., 2018) as well as altered processing of emotional valence stimuli (Groenewald et al., 2013). The current project aimed to extend previous findings on demonstrative choice, and address whether individuals with depression can be detected from demonstrative choice behavior along specific semantic dimensions.
Methods: 775 adult native English speakers completed a 290-nouns Demonstrative Choice Task (Rocca & Wallentin 2020), in which they were presented with one noun at a time and were to match it with either “this” or “that”. Depression was defined as a sum score above 10 on the 9-item Patient Health Questionnaire (PHQ-9) (Kroenke et al., 2001). PCA was performed on the 290 response items. 100 logistic regression models were trained with repeated k-fold cross-validation, predicting PHQ-9 class (healthy vs. depressed) from 1-100 of the PCs. All models were evaluated on a hold-out test set (30% of the data).
Results: The best model exhibited an out-of-sample classification accuracy of 0.65 (p<0.001), with an F1-score of 0.65. The nouns with strongest negative effect (indicating that depressed individuals chose “this” more often than healthy individuals) included: crap, burden, poverty, distraction, avoidance, complaint, darkness, excuse, perjury, and woe. The nouns with strongest positive effect (indicating that healthy individuals chose “this” more often than depressed individuals) included: faith, wealth, meeting, sense, use, carriage, package, belief, beer and motive.
Conclusion: Individuals with depression could be classified from demonstrative choices with an accuracy of 65%. Further, depressed individuals were more likely to use a proximal demonstrative for highly negatively valanced nouns than healthy individuals, while healthy individuals were more likely to use proximal demonstratives for positively valanced nouns and nouns related to social interaction. These findings indicate that demonstrative choices may capture important aspects of self-representation related to depressive mental states and could possibly serve as non-reflective markers of depression.
References
Diessel, H. (1999). Demonstratives: Form, function and grammaticalization (Vol. 42). John Benjamins Publishing.
Groenewold, N. A., Opmeer, E. M., de Jonge, P., Aleman, A., & Costafreda, S. G. (2013). Emotional valence modulates brain functional abnormalities in depression: evidence from a meta-analysis of fMRI studies. Neuroscience & Biobehavioral Reviews, 37(2), 152-163.
Holtzman, N. S. (2017). A meta-analysis of correlations between depression and first person singular pronoun use. Journal of Research in Personality, 68, 63-68.
Kemmerer, D. (1999). “Near” and “far” in language and perception. Cognition, 73(1), 35-63.
Kroenke, K., Spitzer, R. L., & Williams, J. B. (2001). The PHQ‐9: validity of a brief depression severity measure. Journal of general internal medicine, 16(9), 606-613.
Levinson, S. C. (2018). Introduction: demonstratives: patterns in diversity. In Demonstratives in cross-linguistic perspective (pp. 1-42). Cambridge University Press.
Miskowiak, K. W., Larsen, J. E., Harmer, C. J., Siebner, H. R., Kessing, L. V., Macoveanu, J., & Vinberg, M. (2018). Is negative self-referent bias an endophenotype for depression? An fMRI study of emotional self-referent words in twins at high vs. low risk of depression. Journal of Affective Disorders, 226, 267-273.
Rocca, R., Tylén, K., & Wallentin, M. (2019). This shoe, that tiger: Semantic properties reflecting manual affordances of the referent modulate demonstrative use. PloS one, 14(1), e0210333.
Rocca, R. and M. Wallentin (2020). ”Demonstrative Reference and Semantic Space: A Large-Scale Demonstrative Choice Task Study.” Frontiers in Psychology 11(629).