Stephen Stich and Daniel Kelley

Stephen Stich (Rutgers) and Daniel Kelley (Rutgers), "Two Theories about the Cognitive Architecture Underlying Morality," with commentary by Michael Cholbi & Peter Ross (Cal State Polytechnic).  Both the paper and the commentary can be found here.

Comments

[Malachai M. Nilsai]

What impact (if any) does the S&S argument have on meta-ethical questions like whether or not values supervene upon certain worldly facts?

What is the difference, if any, is there between "independent normativity" and notions of "intrinsically rewarding" (in the sense opposed to "extrinsically rewarding"), and the notion of integrity? Is a norm necessarily characterized in these terms, or only contingently so? Is a "moral rule", in the sense used in this paper, just whatever has been dubbed as intrinsically rewarding, or does it also include deference to an actual moral system (i.e., utilitarianism)?

Is a "norm" understood on the psychological level, i.e., as being an internally consistent system of rule-like ideas within the mind, and/or at a sociological level, i.e. as a regular pattern of thoughts and behaviors displayed by an aggregate of people?

Is it possible for these cognitive elements to arise independently of human interaction? If so, why is the discussion being framed in terms of "cognitive architechture" and not simply in terms of human psycho-social development? (Indeed, since the entire distinction between moral and conventional rules in the M/C model is based on other-directedness, one seems tempted to adopt the thesis that morality, in whatever sense, is dependent on human interaction.) Not to say that the search is fruitless, since recent research shows that even monkeys may have some (presumably innate) sense of justice. The question is whether this is best explained in terms of interaction, mechanisms, or emergence, and to what degree of each.

"Moreover, this motivation does not depend on the agent’s beliefs about the social or personal consequences of compliance or non-compliance." (p.3) In what sense of "belief"? The consequences of an action may be imprinted upon the mind independently of the ability of the agent to have any say on the matter. I.E., Imagine a man, Alan, who has arachnophobia. Alan's disorder may be caused by a bad experience with a spider in childhood, such as being bitten by one. Alan's mind may be driven (irrationally) to believe that a similar bad experience shall occur when confronted with a rubber spider, despite the fact that he knows it is silly, and ridiculous, to have this fear. But this doesn't say anything against the consequentialist thesis of mind: rather, it says something in favor of the conscious-subconscious division of mind. Alan's subconscious is driven towards an associated understanding of consequences of meeting a spider, while his conscious mind recognizes this to be a proposterous worry.

In what sense is the word "mechanism" being used? As a product of complex mental causes, or rather in the sense of some Fodorian module? If the former, then this causes taxonomical worries: I may say I have a "pencil-seeing" mechanism in my mind, yet really it's just the faculties and modules of the mind which involve seeing, paired with the Kantian-schematic module, which do the work. In which case the postulation of "mechanisms" seems misleading on the face of it; since we're really just speaking of the mental products of underlying mechanisms.

"Once a normative rule is acquired, it gives rise to reliable and robust intrinsic motivation to comply with the norm and to punish those who violate it. It is worth emphasizing that this pair of motivations sharply distinguish norms from other rules or information that may be mentally represented elsewhere in an agent’s cognitive system." But how does one derive the second proposition from the first? Just because a thing is reliable and robust, does not mean it is distinctive.

"On the S&S account, the acquisition mechanism operates automatically – a person does not decide to turn it on and cannot decide to turn it off, though it may be the case that the acquisition mechanism gradually turns itself off starting at some point late in adolescence." If I am reading this correctly, it gives the empirical prediction that no child will ever reject a norm once it has been internalized. This sits roughly parallel to stage 1 of Larry Kohlberg's model. At stage 2, though, Kohlberg would predict that the child may treat norms as negotiable rules, depending on the desires/goals of self and other. Could you clear up exactly how S&S compares with Kohlberg in this respect? He is barely mentioned. In any case, what is the role of moral reasoning in either model (if any)?

Folkways are a species of norm that are akin to etiquette. The paper reads that "Of course, people also accept and follow many behavior governing rules that they do not treat as norms", and postulates that they are satisfied by separate mechanisms. But if you're speaking of "mechanisms" in the non-superficial sense (outlined above), as in modules, this seems like an arbitrary conclusion. At the very least the distinction between a folkways-mechanism (which delivers what the M/C model calls "conventional rules") and a norms-mechanism (which delivers "moral rules") needs to be explained.

With respect to the M/C models, I am particularly confused about the difference with respect to the "victim/lack of victim" criterion (presented in discussion of Turiel and the M/C models). If I am rude to someone, or play a practical joke, I am breaking a conventional rule, but presumably not a moral one. Yet there is a victim involved.

[David Leon Gil]

Kelly and Stich’s paper is quite convincing on a number of points. Alas, on its most interesting claim—that (C-2a) is false—it falters. The results from the Kelly et al. study just don’t seem to falsify any version of (C-2a); at least any version that someone could plausibly hold. And there’s reason to doubt that the study even gets at moral rules.

1. I think that there’s a pretty simple alternative explanation for Kelly et al.’s results. The survey respondents may break down like this:

Some people implement a norm that prohibits spanking. This is, let’s say, a moral norm in the sense of the M/C distinction, so they take it to be general and authority independent. These people are the 52% who dislike spanking in both examples.

Some people have no norm that prohibits spanking. These are the 8% who think it’s fine in (A) and—we may presume—also think it’s fine in (B).

Some people have no norm that prohibits spanking as such, but have a norm that prohibits disobedience to authority. So in (B), they think that spanking is okay because spanking is okay and Ms. Williams isn’t disobeying authority. But in (A), they think that spanking is wrong, even though spanking is generally okay, because Ms. Williams is disobeying her principal. These could be the 40% who switch sides.

This norm prohibiting disobedience to authority could be moral, in the sense of the M/C distinction, and the rule could be general and authority-independent and all that. And the study provides no reason to think that this isn’t the case—though querying people’s explanations for their moral judgments could be a helpful follow-up. But as it stands, it seems plausible that the switchers have a moral rule that prohibits disobedience.

If this is what’s going on, I don’t see how Kelly and Stich’s results are a challenge for (C-2a) at all. Unless they claim (C-2a) to mean that if something is judged wrong that involves harm, that judgment must rely on a moral rule involving harm. I suspect that the authors do claim this: they take, for example, conversation about the treatment of detainees at Guantanamo to “suggest[] that a significant number of people do not consider rules prohibiting harmful treatment in such cases to hold independently of authority.” (16-17) This in the course of discussing falsification of (C-2a).

I am not much one for charity. But Turiel and his followers wouldn’t just have to be wrong to hold this strong claim. They’d have to be idiots. It’s obvious that people consider some rules prohibiting harmful treatment to not hold independently of authority. Indeed, some of those rules seemed to be at issue in the Guantanamo and related cases. Rules called laws.

Law is characterized by authority-dependence and location-specificity. What’s the law in the United States is not the law in France and is not the law in Augustus’s Rome. And legal norms remain authority-dependent even in cases involving harm. (Is it illegal to murder if there’s no law prohibiting murder? No, seems the sure intuition.)

So I don’t think that Kelly and Stich can maintain that (C-2a) makes such a strong claim. If it doesn’t, then I don’t see how their examples falsify (C-2a) in the first place.

2. And a final note: the Kelly et al. study questions are ambiguous. “Is it OK for Ms. Williams to spank the boy?” (18) Okay how? Morally okay is one reading. But legally okay is another. The example, after all, begins with the statement “It is against the law for teachers to spank students.” (18) (Similarly for the whipping example, in which the example notes that there were no laws against whipping in the historical case. 19) It be that some people interpret the question “Is it OK?” as just meaning “Is it legal?” And those people’s answer is that it’s illegal when it’s against the law (as in (A)) and legal when it’s not against the law (as in (B)). As I noted above, it wouldn’t be particularly interesting to discover that people think that law is authority-dependent.

[Justin Fisher]

I'd like to raise questions regarding the "nativist" aspects of the S&S model, though this also connects with concerns that the commentators C&R raise.

K&S embrace "the existence of innate mechanisms dedicated to norm acquisition" and they say that "[t]he function of these mechanisms is to locate and internalize the norms prevailing in the surrounding society" (pg 4). Perhaps mistakenly (?), I read this as claiming that we have mechanisms that evolved, through a process of genetic evolution, to be as they are *because*of* their enabling us to internalize norms. My concerns revolve around the question of how present-day phenotypic evidence could support these historical claims about genetic evolution in a context where significant cultural evolution is also acknowledged to have occurred.

K&S are quite sympathetic to the idea that norms vary significantly across cultures, and that which norms one ends up internalizing will depend significantly upon which culture one grows up in. They also cite much of Shaun Nichol's work, and are surely aware of work of his that looks at factors (like whether a norm prohibits an elicitor of core disgust) that affect which norms will remain present in a culture over long periods of time. Hence, it seems, K&S must acknowledge that it is not an entirely arbitrary matter which norms are present in our cultures - instead we would expect our cultures to contain only those norms that are strongly and stably internalized, e.g., by prohibiting behaviors that fairly often elicit disgust-reactions anyway.

But, if cultural evolution can do a lot of the work of ensuring that present-day norms will be robustly internalized, it's not clear how such robustness could be evidence for the S&S claim that there was a *genetic* evolutionary pressure favoring innate mechanisms geared towards internalizing these norms. Instead, an alternative historical hypothesis seems equally well, if not better, supported.

On this alternative hypothesis, people have the innate capacity to come to associate strong emotional reactions (like anger, resentment, and disgust) with various sorts of things in their environment. This innate capacity might arise merely from associationist features of our cognitive architecture, or it might be promoted in part by genetic evolutionary pressures which tend (for example) to facilitate social learning of which local foodstuffs are poisonous. Given that people have the capacity to associate various things with various emotions, and to socially transmit this association, a process of cultural evolution then leads to the cultural evolution of norms that lead us to associate strong emotional reactions with various classes of behavior.

Notice that this alternative hypothesis doesn't posit any innate mechanisms whose (genetic evolutionary) raison d'etre is to internalize norms -- instead norms constitute just one category of cultural replicator which would naturally arise in a population of hosts with much more general innate mechanisms.

This alternative hypothesis also fits nicely with the likelihood (raised at the end of C&R's commentary) that different culturally-stable norms might secure their long-term persistence via different sorts of emotional associations -- e.g., norms against inter-racial marriage might draw upon disgust, while norms against eating with the wrong fork might instead draw upon mere sympathetic embarassment.

In sum, I've proposed a cultural evolutionary hypothesis involving only very general innate capacities. This alternative hypothesis seems to fit the present data at least as well as S&S's much more strongly nativist hypothesis, and it also seems to offer better prospects for accomodating the apparent diversity of culturally-transmitted norms.