Reading: “Moral outrage in the digital age”

“Moral outrage in the digital age” by M.J. Crockett is a short theoretical paper, drawing together from several lines of research a [model? hypothesis? theoretical framework? theory?] to explain how the operation of moral outrage is transformed by digital media.

I’m not particularly keen on the underlying view on moral outrage which seems to be based on basic emotion interpretation of anger and disgust (from Fig. 2: “For each immoral act, moral outrage was calculated by multiplying self-reported anger and disgust” – btw, why multiplying and not averaging or calculating a sum?), but otherwise it makes a nice and plausible case of the differences the digital media might make. I’m not familiar with most of the empirical research it refers to, so I can’t say much about how convincing the actual evidence is, but the overview fits my preconceptions.

The main points can be summarized (Fig. 1 is not immediately clear):

  • Humans have psychological processes to react with an emotional condemnation when they think a moral norm has been violated.
  • Digital media
    • gives us a greatly increased access (removes physical constraints) to information about moral violations than traditional social communication (like gossip)
    • lowers the costs (effort; the article talks about the possibility of physical retribution, but I’d generalize that as the risk of potentially wasting the social capital) for expressing outrage
    • lowers the inhibitions (no face-to-face feedback means we don’t have to deal with the fact of causing emotional distress in others, which is a negative experience for most) of expressing outrage
    • increases the potential benefits (reputational rewards of moral quality and trustworthiness; successful regulation of group behavior).
  • These factors drive more moral outrage in digital media, which increases social polarization, dehumanize the targets (and their groups?), and reduce societal trust.

The short paper does not suggest any interventions, but if these mechanisms hold, then it seems to me that potential ways to inhibit this process would be to increase the costs and inhibitions, as the access and potential benefits are more difficult to control (and latter perhaps should not be controlled?). Especially effort, but perhaps costs of social capital as well, could be increased via technological solutions. These are testable predictions to cut out the most low-effort outrage. It would be interesting to see what portion of the outrage would this influence. For instance:

  • Minimally increase the effort, by increasing the steps of, or introducing a small waiting period to sharing.
  • Introducing a way to incur a minimal social cost to sharing, e.g. a downvote, perhaps limited to the friends of the sharer only, so a downvote would actually carry a meaning of “people I care about think somewhat less of me” and maybe would not be constantly abused like on anonymous platforms?




Crockett, M. J. (2017). Moral outrage in the digital age. Nature Human Behavior. Retrieved from

Cuteness elicits play, not care

An article by Sherman and Haidt [1] made some good points that I had not thought before. Like many others (apparently, as S&H consider this the typical view on the topic), I have learned that the appearance of cuteness, or kindenschema (by Lorenz in 1950s), is an adaptation to foster care for a helpless baby. S&H point out that, empirically:

  • Newborns are not considered as cute as older babies – apparently the peak-cute is at 3 months and stays high until 10 to 36 months. If cuteness was supposed to elicit care, shouldn’t it be highest when the baby is most helpless?

An obvious objection would be that evolution does not create optimal adaptations, but adaptations that are better than the alternatives, in regard to cost-benefit. Perhaps reacting to short-lasting newborn features carries costs that are avoided by reacting to features that emerge later but also last longer? Still, I didn’t know this, and I wonder whether cuteness has a relevant relationship at all with the maternal instinct to care for the baby.

  • Perceived cuteness is reduced when the baby shows negative expressions – i.e. when it would probably need the care most – while positive expressions enhance cuteness.
  • Actual behaviors triggered by cuteness – baby-talk, petting, holding – are social behaviors, not caretaking behaviors. Caretaking is social behavior as well, but cuteness does not seem to trigger that any more than other social behaviors.

Thinking about cute puppies, it’s true that I’m more likely to pet them, hold them or play with them than take care of their physical needs per se.

The authors have an argument that cuteness is a mentalizing prime and related to expanding the moral circle, which I guess is ok, but did not create that TIL in me. However, the above points were news to me and made me realize (although the authors do not say it) that cuteness would be adaptation for other people than the primary caretaker(s). The tribe immediately comes to mind: by eliciting social interaction, the tribe creates a bond with the new member. And by eliciting social interaction with the baby (that has already developed some rudimentary ways to respond to social engagements), cuteness elicits first version of play – simple actions that are aimed at getting a positive reaction, which in turn elicits a positive reaction in the interactor. This is relevant for the idea of play that I’ve had [2], as I have primarily understood play (neurophysiologically: trying different things in a safe context in order to get positive reward) as a way to learn how the environment works. But, as is typical, in hindsight it seems obvious that play has the binding function as well – and this realization suggests the possibility that this binding function was first, and more complicated play may be an exaptation from that.


[1] Sherman, G. D., & Haidt, J. (2011). Cuteness and Disgust: The Humanizing and Dehumanizing Effects of Emotion. Emotion Review, 3(3), 245–251.

[2] Kivikangas, J. M. (2016). Affect channel model of evaluation and the game experience. In K. Karpouzis & G. Yannakakis (Eds.), Emotion in games: theory and praxis (pp. 21–37). Cham, Switzerland: Springer International Publishing. doi:10.1007/978-3-319-41316-7_2

Do ’emotions’ exist? Or, abstraction for a certain purpose

Last week, I was in a seminar that discussed emotion theory, and the discussion helped me change my mind.

Since 2014, when I really started reading emotion theories and developing my own idea of the affective system, I have doubted the existence of ’emotion’. I first encountered the idea in article by James Russell[1], who argued against the basic/discrete emotion theories that those are only a matter of perception, as the mind categorizes similar experiences under the same label, regardless of whether they are actually produced in the same way. The problem is, the different instances of a particular discrete emotion, say ‘fear’, do not show coherent patterns of emotion components (the signals by which we study emotions, such as psychophysiological responses and behavioral tendencies). If we cannot find coherent patterns, what basis do we have to call a particular set of experiences with the same name? We tend to categorize all kinds of things in a quick-and-dirty way that works as long as we don’t start philosophizing about what things really are: strawberries are not actually berries while bananas and eggplants are – but this does not matter the least unless we are academics specializing in botanics.

But I am an academic specializing in emotions (or something like it), so it makes a lot of difference what, exactly, I am investigating, how they should be categorized, and how do they work. Does categorizing certain experiences as emotions actually help the investigations, or does it simply lead astray, like the layperson’s conception of ‘berries’? Russell made a powerful argument:

As an analogy, consider the constellations we see in the sky. When we perceive the Big Dipper, we perceive real features: the stars comprising the Big Dipper are real and the pattern among the stars is real in the sense that the stars are really positioned in the universe such that they form a certain geometric pattern when viewed from earth. But stellar constellations, contrary to the beliefs held in many traditional cultures, are not interesting scientific entities. The Big Dipper does not explain the presence of the stars or the pattern among those stars. Constellations are not part of the causal story of astronomy. There is no use in asking for an astronomical account of why the Big Dipper exists or why it is structured the way it is. Astronomy long ago abandoned questions such as: What is a constellation? How many constellations are there? What are they? How were they generated? How does one constellation differ from another? What are their effects/functions?  ([1] p. 1276)

There is a difference between scientific entities and how the layperson calls things. If ‘fear’ and ’emotion’ are not useful for science, they should be discarded as scientific terms. Emotions do not really exist [2], and we are actually investigating something else (I call them ‘affective processes’, but more about that in some other post). A similar idea has been repeated by many scholars (e.g. Lisa Barrett [2], but also non-constructionists, like neuroscientist Joseph LeDoux [3] who is sometimes considered as a basic emotion theorist, and philosopher Peter Zachar [4]), and some emotions researchers have been trying to seriously fight for the idea that we should not call the whatever-we-are-studying ’emotions’. I agreed, until the seminar.

It was actually one of the “there are no emotions” -arguments that changed my mind. They were arguing that instead of emotions, we are actually studying the components themselves – the psychophysiological signals, the behavioral tendencies etc. – and how they relate to each other. True, but this seems unnecessarily cumbersome. For a researcher studying, for example, the link between emotions and moral judgment [5], it is not all psychophysiological signals and behavioral tendencies that are relevant, but only those that are related to emotion. Clearly there is some kind of clustering in what kind of psychophysiological signals and behavioral tendencies are considered emotional [6] – even if ’emotion’, in a strict sense, does not exist. In this, emotions are more like berries than constellations: for instance, culturally, strawberries are considered as berries so an abstraction that it is a berry is more useful for studying how, say, people use them in recipes.

This is what I understood – and it seems so obvious in hindsight, that I’m sure that most non-emotion researchers reading this will roll their eyes and say “are they this distanced from the reality?”, but while this is not exactly a groundbreaking philosophical invention, it is important to understand it intuitively. So, as the saying goes:

All models are wrong, but some are useful. (attributed to George Box)

The addition should be: useful for certain purposes. Of course emotions don’t exist, strictly speaking, but neither do memories, or attitudes, or attention, or any mental phenomena (or ‘behavioral tendencies’ or ‘psychophysiological responses’ for that matter) – they are all pragmatic categorizations that get too fuzzy on a more precise level. The “emotions don’t exist” crowd is correct, but don’t help, because they miss the larger point that ’emotion’ might still be scientifically useful for certain purposes, for investigations on certain level. And while their message is probably mostly correct – that in most cases where ’emotion’ is used, it probably would make more sense to talk about the components – the broader message should not be “emotions don’t exist”, but “make sure that the abstraction you are using is actually useful on your level of scrutiny”.

[1] Russell, J. A. (2009). Emotion, core affect, and psychological construction. Cognition & Emotion, 23(7), 1259–1283.

[2] Or, as Barrett says, they don’t hold any explanatory power – but in this case it’s the same thing. Barrett, L. F. (2006). Are Emotions Natural Kinds? Perspectives on Psychological Science, 1(1), 28–58.

[3] LeDoux, J. E. (2014). Comment: What’s Basic About the Brain Mechanisms of Emotion? Emotion Review, 6(4), 318–320.

[4] edit: Zachar, P. (2010). Has there been Conceptual Progress in The Science of Emotion ? Emotion Review, 2(4), 381–382.

originally I referred to Nesse, but the point is clearer with Zachar:

Nesse, R. M. (2014). Comment: A General “Theory of Emotion” Is Neither Necessary nor Possible. Emotion Review, 6(4), 320–322.

[5] And there are links, even if it’s not clear what emotions are. Avramova, Y. R., & Inbar, Y. (2013). Emotion and moral judgment. Wiley Interdisciplinary Reviews: Cognitive Science, 4(2), 169–178.

[6] Mauss, I. B., & Robinson, M. D. (2009). Measures of emotion: A review. Cognition & Emotion, 23(2), 209–237.


Published: Moral Foundations and political orientation

Hogrefe just informed me that my manuscript (together with Jan-Erik Lönnqvist and Niklas Ravaja; manuscript pdf, supplementary material) with the title “Relationship of Moral Foundations to Political Liberalism-Conservatism and Left-Right Orientation in a Finnish Representative Sample” is now available online before publication in Social Psychology. It’s my first published paper on my new topic after changing from game research to moral psych (and emotion psych, but that’s another thing), so I’m happy about that!

It is a short paper, with a rather straightforward point. Although Moral Foundations Theory has been linked to political orientation a lot, it has been based on the US conceptualization of political orientation, which collates the left-right with the liberal-conservative to much higher extent than rest of the (Western) world. The often-repeated original finding is that liberalism is associated mostly with Harm (also called Care) and Fairness (and Liberty, but that’s another story), while conservatism is associated as strongly with Harm and Fairness as it is with Loyalty, Authority, and Purity (also called Sanctity). (In terms of correlations, it means that correlations between MFs and lib-cons are low for Harm and Fairness, because they are associated with both conservatism and liberalism, but high for the other three MFs, as they are only associated with conservatism.) Researchers have used the knowledge of this finding in other countries too, without considering whether it is reasonable to assume that this is true in other political cultures that separate the two dimensions more. My paper uses a representative Finnish sample (instead of a typical student sample) to test the association* between (self-identified) liberalism-conservatism and the five moral foundations separately from the association between left-right orientation and the five moral foundations. Our findings are that while the typical MF/lib-cons -associations are replicated in our sample (see betas in the table below), the left-right dimension has different associations with the MF – namely, Harm and Fairness are mostly (but weakly) correlated with left-orientation, Loyalty and Authority with right-orientation, and Purity is not associated with either. So basically, the finding is that liberal-conservative orientation and left-right orientation are differently related to moral foundations.

Table 2. Standardized regression weights for model Mboth.
 Parameter β 99 % CIs p
Harm .012 [-.086, .122] .053
Fairness -.085 [-.198, .027] .663
Loyalty .409 [.305, .510] <.001
Authority .432 [.327, .525] <.001
Sanctity .581 [.496, .655] <.001
Harm -.208 [-.314, -.105] <.001
Fairness -.147 [-.105, -.045] <.001
Loyalty .197 [.084, .302] <.001
Authority .197 [.097, .293] <.001
Sanctity .061 [-.037, .157] .093
Note. Higher beta indicates higher association of conservative- or right-orientation to the moral foundation.


Practical observations

(In no particular order)

  • The correlation between the orientations was r = .26 [99 % CI = .18, .34]. So although the dimensions are not the same (as it is implicitly assumed in some US-centered discussions), they are not completely independent either (as e.g. political two-dimensional maps used by the media to portray party differences before elections would lead you to believe). Conservatism is still related to right-orientation, and some have hypothesized that this is because the “resistance to change” attitude (theoretically behind conservatism) is aligned with “resistance to equality” attitude (theoretically behind right-wing orientation) in cultures that are inequal.
  • That left-orientation is more related to Fairness should not come as a surprise, as the factor actually has an item explicitly related to the morality of large unearned inheritances. It is a methodological and theoretical question whether this is a problem or not (I think it is), but it means that you can’t make an argument that “see, leftists care more about fairness!” if fairness is partly defined as something that leftists care more about.
  • In the later version of MFT (Haidt 2012, The Righteous Mind), fairness is theoretically less about equality of outcomes and more about proportionality, or law of karma: that everyone gets what they deserve, be it good or bad. This is likely to make it more relevant for conservatives, but it’s not incorporated in the MFQ measure yet.
  • The theoretical addition of liberty – that nobody should not be forced to do anything – would separate between libertarians and tradition-related conservatives, but that’s not part of the measure yet, either.
  • Although the outdated journal publication system forces to focus on one thing and one thing alone, I think the methodological part, partly touched on in the supplementary materials, is also important. I’m still looking into it, but I’m having some doubts about the MFQ as a measure (partly related to what I said above). I should do another blog post about that later.


*) n = 874, which exceeds required sample size to detect a small effect (ρ = .1), assuming a power of 0.80 and an alpha of 0.05. The p-values are not corrected, because this study was a typical half-exploratory study people do in psychology while pretending that we are doing hard confirmatory research, and it would be really difficult to determine afterwards what the appropriate correction would be (I hadn’t awoken to the whole replication/methodology crisis thing at the time of writing this). However, because of the big enough effect sizes that follow what has been found before (in case of lib-cons), and p’s that are not simply bordering the arbitrary .05, I think the results are plausible. I’ll try to write a retrospective post about the researcher degrees of freedom regarding my old papers at some point.

To all who have waited for my answer

Great piece in New Yorker!

Sorry for the delayed response. I opened your e-mail on my phone while my date was in the bathroom, but then I saw that it required more than a “yes” or “no” reply, decided that was too much work, marked it as unread, and then forgot about it entirely until just now!


Sorry for the delay! I put off answering your e-mail until I had an even more tedious task that I wanted to avoid. Thanks!


You e-mailed asking for my opinion, and I wanted to give a really thorough, well-thought-out, articulate response, so I starred your e-mail, and over time it became a mascot for my illogical but oppressive sense of dread in the face of slightly annoying tasks. That little yellow star became a shining testament to the burden of modernity! Every day, it dared me to write a response worthy of the time I’ve made you wait, and every day I thought, Ugh, no. But today! Today I will respond! Rejoice, my patient friend! (I’m actually really busy, though, so this is going to be a vague, half-assed response that I could have easily written in the minute after I first read your e-mail, five months ago.) Sorry!

We probably all do this*, so it shouldn’t be such a pain, but again and again it feels like it is. For instance, I haven’t started this blog yet, because I really want the first one to be a good one, and it grows bigger and bigger the more I wait, so nothing feels good enough.

I aim at two things with this post: first, to half-ass the start of this blog so it does not seem so insurmountable anymore, and to tell everyone, this is how it is sometimes. Can we just skip the explanations and carry on?

*) Or maybe those weird people who seem to be far too productive don’t? How, tell me!?