Skip to main content
Log in

What basic emotions really are: modularity, motivation, and behavioral variability

  • Published:
Biology & Philosophy Aims and scope Submit manuscript

Abstract

While there is ongoing debate about the existence of basic emotions and about their status as natural kinds, these debates usually carry on under the assumption that basic emotions are modular and therefore cannot account for behavioral variability in emotional situations. Moreover, both sides of the debate have assumed that these putative features of basic emotions distinguish them as products of evolution rather than products of culture and experience. I argue that these assumptions are unwarranted, that there is empirical evidence against them, and that evolutionary theory itself should not lead us to expect that behavioral invariability and modularity mark the distinction between evolved emotions and higher cognitive emotions. I further suggest that claims about behavioral invariability and modularity have functioned as defeasible conjectures aimed at helping test basic emotion theory. Finally, I draw out the implications of these claims for debates about the existence of basic emotions in humans.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

© Elsevier. Originally published in Physiology and Behavior

Fig. 3

Similar content being viewed by others

Notes

  1. However, conscious recognition is obviously not the only route to achieve adaptive effects of ritualized signals.

  2. For my purposes, innateness can be thought of roughly in terms of phylogenetic information (see, e.g., Browne, 2005), where this is understood as “fit” with the environment that is can only be explained by evolutionary adaptation. As Griffiths (2020ate and Acquired Ch, n.d.) notes, this notion is closely connected with the evidential role of deprivation experiments in classic poverty of the stimulus arguments. (See, e.g., Khalidi, 2002; KhAlidi, 2007; Stich, 1975; O’Neill, 2015).

  3. Coltheart (1999) argues that modularity is best defined in terms of domain specificity. For a recent discussion of domain specificity in neuroscience, see also Spunt and Adolphs (2017). Kurzban and Barrett (2006) argue convincingly that modularity is primarily about functional specialization. Rather than committing myself to any particular way of defining modularity, I will merely use the word as a shorthand for the characteristics of emotion most relevant to behavioral variability, encapsulation and opacity, but those characteristics may only be imperfect indicators of modularity qua functional specialization (or domain specificity, or whatever else).

  4. Of course, massive modularity theorists argue that central processes are also modular in some sense (e.g., Carruthers, 2006). However, their notion of modularity is not closely tied to encapsulation or opacity. Moreover, these accounts need not deny that there are important, observable differences between processes that are more automatic and mandatory and those that are more deliberate and controlled. Even on a massively modular view of cognition, it is still possible to articulate a relationship between emotion and cognition on which the one is encapsulated and opaque to the other.

  5. It is beyond the scope of this paper to consider the neuroscience of emotions any further, given the vastness of that literature and given my focus on behavioral variability and modularity. As I understand it, much of the neuroscientific data on emotions is only tangentially relevant to these issues. This is especially true of the primary tool for investigating the neural basis of human emotions: fMRI. Celeghin et al (2017) capture some of the difficulties in using these neuroimaging methods to make any claims about basic emotions in humans, in particular claims about the speed or automaticity of affect programs:

    What inferences can be made on the basis of fMRI studies? What are the limitations of the methodology with respect to the debate on the existence of basic emotions?

    A first limit is epistemological. That is, imaging studies are good at revealing which neural structures are involved in the processing of basic emotions, but are silent with respect to what structures are necessary to recognize or express such emotions. In this sense, they offer a type of ‘weak’ or correlational evidence and should be interpreted in the light of other data, such as lesion studies, in which the correlational nature of fMRI data is elevated to a causal inference (Krakauer et al., 2017). Others and we tend to believe that the starting point to understand the neurobiology of emotions is the analysis of behavior, as we cannot rely solely on the correlational approach of neuroimaging data devoid of relation with behavioral outcomes. Indeed, the causal-mechanistic explanations are qualitatively different from understanding how component modules perform the computations that then combine to produce behavior (Krakauer et al., 2017)…

    To sum up, neuroimaging data and recent meta-analyses do not seem to us to provide sufficiently solid ground for rejecting the existence of basic emotions at the neurobiological level. In addition, other features considered typical of basic emotions, like automaticity or early onset during sensory processing (Ekman, 1999; see above), are not considered in these studies… (p. 4).

    More specific to the issues at hand, it is difficult to see how neural data could possibly help to elucidate whether an emotion can control flexible patterns of behavior or only rigid ones (as the strong invariability assumption states). At present it simply is not possible to use neural data to determine whether a behavior is caused by a basic emotion or not. To do so would involve a reverse inference from patterns of brain activation to the involvement of a basic emotion, and we would only have that if we could resolve one of the most heated debates in current neuroscience of emotion (see, e.g., Celeghin et al., 2017; Lindquist et al., 2012). Other lines of research in the neuroscience of emotion are no more relevant to the issues of behavioral variability and modularity: EEG studies tend to focus on perceptual processing of emotional facial expressions rather than emotional responses as such, and lesion studies provide only ambiguous and indirect evidence concerning changes in emotional behavior. So, while neuroscientific research is important for assessing certain claims about the existence of basic emotions, it does little to help assess the invariability assumptions that are my focus here.

  6. For a critical perspective, see, Celeghin et al (2017).

  7. Ekman apparently retains this view in more recent writings (2003). He writes the following in an endnote: “Frijda’s description of the actions that characterize each emotion includes what I have said and quite a bit more. I believe it is only these rudimentary initial postural moves [e.g. looking down on an object of contempt, fixed attention on the object of surprise, movement toward the source of sensory pleasure, and slumping posture and loss of muscle tone in sadness] that are inbuilt, automatic, and universal.” (p. 268) Moreover, a more recent piece by Ekman and Cordaro (Ekman & Cordaro, 2011) suggest that the view remains largely the same (see esp. pp. 367 and 368).

  8. At least, similar outcomes with some reasonable exceptions, such as when the emotion is suppressed or regulated in some other way. Ekman does appear to appreciate this point, and makes some comments that suggest he is skeptical that anger has any “central direction.” Nevertheless, once we distinguish between the direct motivational effects of anger on the one hand, and the individually variable effects of emotion regulation on the other (a distinction that Ekman does not consistently draw), I believe it becomes possible to see the outlines of a central direction. This makes the assessment of an emotion’s central direction less empirically tractable. Nevertheless, the problems with the central direction of hunger are exactly analogous, but clearly not insurmountable. The fact that I am hungry may not necessarily manifest itself in my behavior in any straightforward way (e.g., if I am feverishly working on an essay that I hope to submit in short order).

  9. Scarantino (2015b) proposes a motivational theory of emotions that makes motivational states essential to emotions, basic or otherwise. See also, Scarantino (2017), Scarantino & Nielsen (2015), and Wiegman (2020).

  10. See, e.g., Daly and Wilson (1988, Chapter 10) and McCullough et al. (2013)

  11. Anger need not always lead to such actions any more than hunger always leads to pursuit of food, but that is just to say that like hunger, the behavioral dispositions of anger can be suppressed by countervailing desires or by systems for executive control.

  12. Wiegman (2020) argues that some emotional actions do not have goals, in a sense roughly identical to Dickinson’s. These arguments undermine any simple distinction between “emitted, flexible behaviors and elicited, fixed ones,” where the category of “action” and “flexible behaviors” are subsumable to “goal-directed behaviors.”.

  13. Though he confesses that he and his associates never tested this hypothesis quantitatively “because it seemed so obvious” (Ekman, 2006, p. 208).

  14. While there is some experimental evidence of this capacity (see e.g. LALAND & READER, 1999), this kind of evidence appears to be surprisingly scant, perhaps because it seems obvious enough to take for granted.

  15. Though there may be distinct motivations for hunting and feeding. See Berg and Baenninger (1974).

  16. For an integrated theoretical treatment of the many influences on foraging behavior, see Nonacs (2001).

  17. The idea of integration/modularity of motivational states needs sharpening. Were there space, a more thorough way to do this is to start with a causal or computational characterization of encapsulation and opacity, perhaps along manipulationist lines (e.g. Woodward, 2005). In both cases, the limitation is in the ability of other systems to manipulate (or cause changes in) the target system at intermediate levels of processing in a way that influences its performance of the relevant function.

  18. Another thing worth noting here is that there may actually be some variability in certain signaling functions, which could also make evolved facial signals highly sensitive to context. See Griffiths (Griffiths, 2003, 2004) for detailed discussions.

  19. I think this justification ultimately fails on independent grounds: Griffiths is running together two things that should be kept separate. One is a question about existence: are there any psychological states in the ballpark of basic level emotions categories (e.g. anger, fear, disgust, interest) that are evolutionary adaptations? Another is a question about essence: are there any “psychological and neurological principles” that capture the category defining features of the class of basic emotions? While both questions are important, the existence of special purpose adaptations in the ballpark of anger and fear (etc.) simply does not depend on whether there is some important similarity among the states in those (possibly very separate) ballparks. Much less does the question of existence depend on whether the relevant states are contextually invariable.

  20. Caroline and Robert Blanchard call them offense and defense, respectively (Blanchard et al., 2009; Blanchard & Blanchard, 1984, 2003).

  21. See, e.g., Wiegman (2019)

  22. Recall from Sect. 2 that there is no reason to suppose that basic emotions are unique to humans or that the set of basic emotions in humans is the same as in other species.

  23. Perhaps a key difference between this research and research on basic emotions in humans is that no evidence has been presented that the facial expressions and physiological responses are specific to the relevant emotional state. There is even some evidence against this claim regarding rats and mice: The facial expressions appear to have a protective function rather than a communicative function. See Defensor et al (2012). Nevertheless, lack of specificity does nothing to negate the evidence that these distinct behavior suites exist and are adapted to solve basic life problems of sociality.

  24. In connection with this point, Adams and Schoel (1982) have pointed out that captive macaques also attempt to bite the backs of “intruders.”.

  25. For instance, in some resource competition scenarios, signaling aggressive intent is not an adaptive strategy (Caryl, 1979). As mentioned above, the facial expression associated with the confrontation in rats is not actually specific to confrontation. For instance, the rat will make a similar expression if its vibrissae are repeatedly stroked, suggesting that it may instead be a protective response to potential contact with the snout (Defensor et al., 2012). So if these expressions have any signaling function it appears to be secondary to their protective function, and perhaps also dissociable from it.

  26. See Gaulin and McBurney (2001, p. 266) for a similar argument.

  27. Lisa Barrett’s (2017) alternative to BET also predicts contextual variability. So contextual variability will not tell us which of these competing theories is most likely to be true. While this discussion has focused on behavioral invariability and modularity, other kinds of invariability assumptions have also come under scrutiny recently, especially invariability of neural responses, or the assumption that there is a one-to-one mapping between brain regions and emotions (cf. Lindquist et al., 2012). Similar to the case I have made concerning behavioral variability, it is unlikely that this commitment is part of the theoretical core of BET, because no one has shown how it follows from the central commitments of the theory.

  28. Frank’s (Frank, 1988) notion of a commitment device comes close to providing such a rationale, but many emotions are clearly not commitment devices (e.g., disgust, happiness) or at least that is not their first and foremost function. For a discussion of disgust’s signaling function that considers and rejects its role as a commitment device, see Kelly (2011).

  29. See Scarantino (2015a) for an argument along similar lines. One reply to this line of argument is just that BET should include the commitments that its proponents consider essential to the theory. In other words, BET is falsified by the contextual variability of emotions because Ekman and others articulate these assumptions as central to the theory. I am happy to grant this in exchange for a concession: the falsity of BET does not prove there are no special purpose adaptations in the ballpark of basic level emotion categories like anger, fear, etc. (I have yet to see Barrett and others make this kind of concession. On the contrary, they appear to take the falsity of BET as a reason to reject domain-specific theories of emotion quite broadly.) If the arguments above are correct, then special purpose adaptations may very well be cognitively integrated in a way that supports contextual variability. If they exist, some would lack so-called “behavioral signatures,” but they might resemble basic level emotion categories. For example, even if confrontation and avoidance systems in rats should not be called “basic emotions” because of their contextual variability, they are still special purpose adaptations that resembles human emotions in various respects (e.g., as irruptive, impulsive, or reactive motivational states). And if there are motivational adaptations like these in rats, they may very well be conserved from our common ancestors. Thus, there may very well be echoes of these special purpose adaptations in the human lineage as well (see e.g., Wiegman, 2016).

References

Download references

Acknowledgements

I owe an enormous debt of gratitude to Carl Craver, John Doris, Reiko Graham, Joseph McCaffrey, Ron Mallon, Olivia Odoffin, and Brooke Robb for their support and feedback on this paper.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Isaac Wiegman.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wiegman, I. What basic emotions really are: modularity, motivation, and behavioral variability. Biol Philos 36, 42 (2021). https://doi.org/10.1007/s10539-021-09816-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s10539-021-09816-z

Keywords

Navigation