Chapter 3. Mental simulation and the spontaneous evaluation of moral events
3.4. Simulation hypothesis: an alternative account?
The previous section advanced two broad empirical claims about the nature of spontaneous moral judgment:
(1) the spontaneous processing of moral stimuli involves sensory and motor systems, and
(2) moral judgments exhibit the typical pattern of projection, i.e., third-party moral attitudes are causally dependent on personal aversions and preferences. Now we might ask: What sort of unconscious processes would recruit sensory and motor systems to visualize the action under evaluation, and cause third-party evaluations to be fashioned after own preferences and values? I suggest that processes of mental simulation are fit to accomplish just this. In the previous chapter, I introduced mental simulation as the imaginative ‗run-through‘ of another‘s behavior using shared neurocognitive systems. These shared systems are activated either endogenously or exogenously, such that consideration of others‘ intentional actions, emotions, or sensations activates internal replicas of, among other things, the motor plans and affective states that lie behind the behavior (Gallese & Goldman, 1998; Gordon, 2004). I will refer to this account as the simulation hypothesis:
Third-party moral judgments involve a (spontaneous) sensory and motor simulation, which elicits the evaluator‘s personal affective response. This affective response shapes the moral evaluation of the third-party behavior.
Consider again the example of the footbridge problem. According to the above hypothesis, the presentation of the footbridge vignette yields – via mental simulation – an internal replica of the behavior, i.e., a first-person, imaginative run-through of the experience (in more or less detail). In turn, this process sets off an aversive, affective response which motivates the third-party judgment. This view would explain the involvement of motor systems, as suggested by Greene and colleagues‘ (2009) study, and of sensory systems as demonstrated by Amit and Greene (2011), and by Inbar and colleagues (2012).10 In addition, the simulation hypothesis is equipped to explain why we find projection in moral evaluation of others (Lieberman & Lobel, 2007): the question whether another‘s behavior is right or wrong is answered by simulating own performance of the behavior, and morally evaluating another‘s action on the basis of the affective response that stems from the simulation.
Finally, I have argued that mental simulation may be the principal cognitive process in third-party moral judgment and that simulation (at least the simulation of emotionally-charged behavior) often results in an affective response. Therefore, a simulation-based account seems to favor a constitutive role of affect in moral judgment.
10
It is worth also reviewing how this account would explain the correlational and experimental studies reviewed earlier in this chapter. First, consider studies that index a correlation between affect and third-party judgment. In the case of purity violations, for example, participants‘ own disgust sensitivity (arising through simulation) shapes condemnation of the third-party behavior. Similarly, the moral difference between
doing a harmful action and allowing comparable harm to take place might arise from
differences in the aversion to performing harmful actions versus performing harmful omissions (see ‗omission bias‘ in Baron & Ritov, 1990; Ritov & Baron, 1995). Finally, on this view, the manipulations of affective state through hypnosis (Wheatley & Haidt, 2005) and watching comical videos (Valdesolo & DeSteno, 2008) influence the evaluator‘s actual affective state, interfering with the shared system‘s normal representation of the third-party agent‘s affective state. That is, these studies render the simulation mechanism less accurate at reflecting how the evaluator would feel when performing these third-party moral violations, thereby influencing the severity of their moral judgments.
That is, (if the account is right) we ought to conclude that moral judgments are constituted by affective responses to processes of third-party simulation.
3.4.1. The role of VMPFC in moral judgment
The proposed account faces a difficulty, prompted by the examination of moral judgment patterns in acquired sociopathy. Koenigs and colleagues (2007) tested the moral judgments of a group of patients with bilateral damage in the ventromedial prefrontal cortex (VMPFC) against a group of matched controls. They found that, compared to psychotypical participants in the control group, patients with VMPFC damage made more permissive judgments about harmful action in personal moral dilemmas. By contrast, the VMPFC patients made normal judgments about impersonal and non-moral dilemmas (see also Ciaramelli et al., 2007). VMPFC patients are known to have profound affective deficits, despite retained general intelligence. Remember that, on Greene‘s view, moral condemnation depends on affect-laden processes and rational processes support more permissive moral judgments about these cases. Therefore, Greene‘s theory straightforwardly predicts the result that is obtained. Meanwhile, on the account I have developed both kinds of judgments recruit affective processes, so it is not obvious how my account can explain these data.
A somatic marker view
A distinct perspective on the role of VMPFC, based in the work of Damasio and colleagues, suggests an alternative explanation for their moral judgments. According to the somatic marker hypothesis, psychotypical individuals have the ability to bring to bear affective cues in complex decision-making contexts, through a repository of unconscious associations between behavioral options and the outcomes that are associated through experience to each option (Damasio, 1996).
This perspective derives its strongest support from a series of experiments employing the Iowa Gambling Task. The IGT is a single player game where participants
select cards from a set of decks seeking to maximize pay-off. Each card has a different pay-off but some decks have better average pay-off rates than others. Within a few rounds of the game, participants learn to select from the advantageous deck, before being able to consciously report which deck yields the highest rewards. Moreover, physiological evidence indicates that this process relies on anticipatory aversions that deter the selection from low-yield decks (Bechara et al., 1997). The VMPFC appears to play a crucial role in this process: Participants with lesions to the VMPFC fail (a) to learn to select from advantageous decks, and (b) to exhibit these same signs of anticipatory aversion to a selection from low-yield decks (Bechara et al., 1994). More recently, neuroscientific evidence points toward heightened activation in the VMPFC during the IGT (Li, Lu, D‘Argembeau, Ng, & Bechara, 2010). Therefore, the VMPFC plausibly houses the somatic markers which unconsciously guide advantageous decision-making.
An analogous somatic marker account could be told about moral decision- making in trolley-type moral dilemmas. Somatic markers may also encode anticipatory aversions to immoral behavior, such as the violent action of pushing an innocent person. On this explanation, the personal (but not the impersonal) harmful action is associated through prior knowledge to ensuing victim harm, and therefore a fast and advantageous decision to condemn personal harm (but not impersonal harm) can be produced by psychotypical controls, owing to the influence of somatic markers.
Empathy and quantitative trade-offs
Still, it remains the case that VMPFC damage is associated with impaired affect, and moreover, impaired empathic abilities (Adolphs, 2002), and according to my proposal in Chapter 2 the endorsement of harm for the greater good recruits empathic abilities. So, why then would utilitarian judgment not also be impaired following
VMPFC damage (on my account)? Moreover, why would patients be more likely to endorse harmful actions for the greater good?
My account does indeed predict that the utilitarian concern with others‘ welfare should be subdued as a result of VMPFC damage. Reduced empathic concern should result in less motivation to actively ―search‖ for moral outcomes and enact welfare- maximizing courses of action. This does appear to be true of their real-life behavior, as documented by the tendency toward antisocial personality disorder accompanying damage (Grafman et al., 1996) or cortical thinning (Narayan et al., 2007) of the VMPFC.
But then why does their reduced empathic concern go unnoticed in judgments about trolley-type dilemmas? Recall that in Chapter 2 we discussed the components of utilitarian judgment – probability and magnitude – and showed that it is determining the
magnitude of moral outcomes which requires empathic abilities, but only (or primarily)
when the relevant moral outcomes vary qualitatively. In addition, empathic concern may guide the prior ―search‖ for morally relevant outcomes to consider in the welfare calculus. Yet, in the classic one vs. five moral dilemmas in the experimental literature, these morally relevant outcomes are (i) provided by the scenario description, and (ii) of a single kind. Even severely limited empathic concern might suffice to prefer the utilitarian course of action in these quantitative trade-offs. To see why, consider the following example:
I care deeply about my library. Therefore, if someone threatened to burn many of my books unless I shredded a single book from my library, I would adamantly comply. My CDs I don‘t like nearly as much; yet if someone posed the same dilemma, I reckon I would act the same (though not with the same drive, perhaps, with which I would try to save my library). I still prefer to save the larger number of my CDs, simply because I like CDs at all. This I think is roughly the case with VMPFC patients: they
care less about human lives than do psychotypical utilitarians, but they still (in these particular circumstances) prefer welfare-maximizing outcomes. Their empathic deficit just means they won‘t pursue them with equal zeal.
In sum, VMPFC regulates a range of affective processes, including the influence of somatic markers in behavior regulation and the ability to empathize with others. Consequently, both the condemnation of harm and the promotion of the greater good are impaired following VMPFC damage. Guided by an experimental literature that has emphasized quantitative (one-versus-five) trade-offs, for which only minimal empathic concern is required, we have devised an incomplete picture of the role of VMPFC in moral judgment. Yet it is clear, outside lab settings, that VMPFC patients are not utilitarian in any broader sense. To the contrary, their empathic deficit renders them less concerned with the greater good, and less motivated to defend it.
3.5. Conclusion
A vast experimental literature demonstrates the involvement of affective processes in moral judgment. In this chapter, I differentiated three views about the psychological character of moral judgment, and argued that the empirical literature – and, in particular, the numerous studies employing experimental manipulations of affect – are incompatible with the epiphenomenal view, according to which affect merely results from the process of making moral judgments. Next, I addressed a related question concerning the cognitive processes that normally take place in the wake of third-party moral behavior, which yield this affective response. I characterized the prevailing account as one according to which the evaluator computes a moral intuition over different components of moral behavior such as intention, causation and so on. Instead, I argued that a growing body of literature points towards the involvement of a sensory and motor simulation of the behavior. A consequence of this is that we find a
tendency to project one‘s own moral values and aversions in third-party moral judgment, as seen in experimental studies examining the moralization of protected values and the moralization of third-party incest among Kibbutz members. Therefore, I defend a view of the processing of moral stimuli according to which the evaluator spontaneously simulates the behavior under evaluation, and this process triggers (in some cases) aversive reactions that guide judgment and decision-making. If this account is correct (awaiting empirical confirmation), moral affect is the core mental state constituting spontaneous moral judgment.