Individual talks: Measurement and general theory

July 12th A2.09

Emotions: The Distinction between Function and Value

Dong An

The evolutionary function and the value of an emotion belong to two different topics. Function is a causal mechanism to promote survival and reproduction. It speaks to the question of explanation. “Value” is about “good”. It speaks to the question of justification. I argue that this distinction is neglected in the functionalist literature on emotions by giving evidences that scholars often overgeneralize the function claims to the value claims. I then argue that this is a mistake. I first contrast function with value by pointing out that emotions can have intrinsic or instrumental epistemic and moral value which are not necessarily related to the functions. I then contrast function with well-being by pointing out that all of the three common theories of well-being, hedonism, desire theory, and objective-list theory (eudaimonia theory), are not necessarily related to the functions. Lastly, I argue that confusing the functions and value is problematic. It would make one intentionally cultivate emotions that might cause harm. For example, anger is evolutionarily functional in helping the agent remove obstacles, but learning or allowing oneself to be easily angry can be bad. The confusion would also mislead people into downplaying the significance of certain emotions. For example, if we equate the value of gratitude with its function of facilitating future interpersonal cooperation, gratitude loses its value. My point is that we cannot derive that emotions are valuable from the claim that they are functional. Pace Hume, we need some premise to bridge the “is” and the “ought”.

The Structure Of Emotion

Kris Goffin

It seems to have become quite mainstream in emotion science to claim that an emotion consists of a number of components. Fear, for instance, should not be reduced to a representation of “danger”, but it should be considered as a complex phenomenon consisting out of a number of additional components such as physiological changes, motivational action tendencies and expressive behavior (such as facial expressions). This raises several questions for philosophical theories which define “emotion” in terms of mental representations. If it is true that the representation of danger is just a single component of a complex emotional experience, how should we think of the relation between mental representation and conscious emotional experience? It would be wrong to say that the conscious experience of fear solely supervenes on a representation of danger. Relying on Fodor’s work on the structure of representations, I will argue that we could think of a full-blown emotion as a structured whole of representational components. Different components of emotion are representations in their own right. Each word of a sentence has meaning, but the sentence as a whole has a meaning as well. The meaning of a word of a particular sentence partly depends on its relation to other the words of that sentence. Likewise, each emotional component has a content which partly depends on the relation to the contents of the other components. Together these components form a structured representation which is the representational basis of a conscious emotional experience.

Network Analysis as a Means of Assessing Translatability of Emotion Words

Katie Hoemann, Margherita De Luca & Lisa Feldman Barrett

Cross-cultural differences in emotion concepts are well-established (e.g., Russell, 1991; Wierzbicka, 1994). Purportedly ‘untranslatable’ emotion words are particularly salient examples: Tagalog speakers have a word, “gigil”, for “when something is so cute that you want to squeeze it”, and Italian does not have an easy translation for English “excitement”. But how should ‘untranslatability’ be defined? Previous work has employed a variety of methods for comparing emotion word meaning across languages (see Ogarkova, 2016 for a review). However, most methods are focused on describing the nature of differences in meaning, rather than measuring the similarity or dissimilarity of individual words. In this study, we use free association data from the Small World of Words project (De Deyne et al, 2008, 2013) to construct semantic networks for ‘untranslatable’ Dutch words (e.g., “gezellig”) and their possible English translations, as determined using synonyms and back translation. We selected networks with equivalent metrics, and evaluated the similarity of target words’ behavior within their respective networks using these same metrics. Using this data-driven approach, we found no (single) English word matches for Dutch targets. These observations are supported by pilot data comparing “gezellig”, its semantic neighbors, and possible English translations on the basis of valence and arousal, as well as features of elicited scenarios. To contextualize these results, we also generated networks based on lexical neighborhood data from Continuous Bag-of-Words (CBOW) models trained for Dutch and English (Mandera et al, 2017). Taken together, our results illustrate a novel means of understanding emotional meaning across cultures.

An Indirect Scaling Method for Testing Quantitative Cognitive Emotion Theories

Rainer Reisenzein & Martin Junge

We summarize research conducted during the past years on an improved method for measuring the subjective experience of emotions, an indirect scaling method based on graded pair comparisons. We have evidence that the scale values of emotion intensity obtained with this method have a much higher reliability than the usual rating scales (Junge & Reisenzein 2013; 2015) and even seem to achieve a metric scale level (Junge & Reisenzein, 2016). This makes this scaling method well-suited for testing quantitative cognitive emotion theories. Indeed, very good fits of quantitative cognition-emotion models on the individual subject level were obtained with this measurement method (Junge & Reisenzein, 2013). In recent work, the method was sucessfully extended beyond the measurement of emotions to the measurement of cognitions and action tendencies (Reisenzein & Franikowski, 2018). Junge, M., & Reisenzein, R. (2013). Indirect scaling methods for testing quantitative emotion theories. Cognition and Emotion, 27, 1247–1275. Junge, M., & Reisenzein, R. (2015). Maximum Likelihood Difference Scaling versus Ordinal Difference Scaling of emotion intensity: a comparison. Quality & Quantity, 49, 2169–2185. Junge, M., & Reisenzein, R. (2016). Metric scales for emotion measurement. Psychological Test and Assessment Modeling, 58, 497–530. Reisenzein, R., & Franikowski, P. (2018). Improved measurement of cognitions, feelings and action tendencies using Ordinal Difference Scaling. Manuscript in preparation, University of Greifswald.