Cullity's system‐building

Cullity has written a splendid and extremely ambitious book, full of novel ideas and interesting arguments. It offers nothing less than a complete foundation of almost the whole of morality, and from this foundation Cullity derives crucial norms and values that have many illuminating applications. The number of issues covered is breath-taking, and the knowledge of the field extremely impressive. What sticks out in particular is the combination of particularism (or holism) and foundationalist system-building. Usually system-building of this kind is seen as anathema to particularism, since it eschews any general moral principles. Undeterred by this potential tension, Cullity sets out to combine the two traditions. In this short commentary, I shall focus on the architectural elements and structures of Cullity's system. I will start with some comments on his methodology, and then move on to questions about the architecture of the system, especially its support structure: the crucial derivation relations. Along the way, I will contrast Ross's views with Cullity's and make a few points that I think speak in favour of the Rossian approach and against Cullity's. However, my main conclusion is that it is not clear that Cully has managed to ease the tension between particularism and foundationalist system-building. Cullity lists truth, illumination, and usefulness as important general constraints on any plausible moral theory. Parsimony does not make it to this list, however, which is a bit odd, since it seems obvious that a more parsimonious moral theory has something going for it, especially if it satisfies all other constraints to the same degree as do the relevant alternative theories. How does Cullity's theory compare to Ross’ in terms of parsimony? Well, Cullity beats Ross on the number of foundational principles, but not with a great margin. Ross has five: promotion of the good, non-maleficence, fidelity, gratitude, and reparation (27).1 And Cullity has three, possibly four: concern, respect, cooperation, and, possibly, reverence for what is precious.2 But Ross beats Cullity, with a large margin, on the number of primitive derivation relations. Ross has only one, deductive subsumption, while Cullity has nine! They divide into two main classes, homogenous derivations (derivations of one value from another, or of one norm from another that involves the same normative relations), and heterogenous derivations (derivations across different normative relations, of norms from values, or of values from norms) (68). Cullity posits three main homogenous derivation relations (subsumptive, enabling, and responsive), but with subdivisions of each category there are 2 + 2 + 3 = 7 primitive categories! (70-79) The heterogenous derivations divide into two more subcategories (to do with content-undermining and context-undermining, respectively) (chs. 5–6). So, nine in total. Surely, this substantial difference in the number of primitive derivation relations should count for something. Of course, Cullity could claim that by invoking these relations he can explain so much more with his theory. But he needs to show that it cannot be done with a fewer number of derivation relations. After all, one should not multiply primitive derivation relations beyond necessity. Furthermore, the greater the number of different derivation relations is, the more disjoined and less unified the explanations will be, and a theory with more disjoined explanations is, in that respect, worse than one with less. Each of the general methodological constraints—truth, illumination, and usefulness—are subdivided into more substantive principles. Cullity claims that these can be seen as ‘criteria of betterness’ to judge theories against (1). The problem is that they are stated in a very vague and often ambiguous manner, so that it is not easy to see how they could function as independent criteria for theory choice. He claims that there is no need to make them more precise before his own theory is presented (3). I can see that this would be a difficult and time-consuming task. But this perhaps shows there are too many of them, or that they are not all equally important. The main problem is that if the yardsticks are allowed to adapt to the theory under consideration, it does not seem easy to use them to do theory comparisons fairly. This should be contrasted with the axiomatic approach in other kinds of system-building enterprises, such as logic, formal linguistics, formal ethics, and social choice. Here we have precise substantive and conceptual constraints, and it is always possible to check if a theory satisfies it, (except when logical reasons prevent it). The reason I go on about this issue is that I think there is too little of theory comparison in philosophy. Usually, one satisfies oneself in just providing a critique of another theory, defending one's own theory against certain objections, or presenting one's own theory in detail. There is room for this too, of course. But we cannot avoid the important comparative questions. Any theory worth considering will have both weaknesses and virtues. So, the important question is not whether a theory has a certain weakness or a certain virtue, but how it compares to others in terms of weaknesses and virtues. Of course, my concern about the adaptable yardstick is not an issue, if the methodological principles are only meant to ‘indicate the spirit’ in which one is ‘setting out’, which Cullity also claims (3). But then they provide very little help with theory choice; they mainly help us with theory presentation. According to Cullity, on the other hand, subsumption is a kind of primitive, non-deductive, relation of making the case, which is supposed to be ontologically innocent and can be signified by ‘because’ (13). It may be true that actions of tormenting a child are prevention-worthy and that killing the child is a way of preventing him from being tormented, but this surely does not entail that the child is worthy of being killed. (71). There is a problem with Cullity's use of this example, however; the example is underdescribed. If the only way of preventing the torment is to kill the child, the torment is really horrendous and will eventually lead to her death, and the child begs to be killed and the parents approve, then it may very well be true that the child is worthy of mercy killing. In a footnote (fn. 7, 71), Cullity concedes that we can have reason to kill the child, but that does not show that the child is worthy of being killed (according to Cullity, reasons can be based on worthiness facts, but they do not have to be). I agree that it sounds odd to say without qualification that the child is worthy of being killed. But to say that the child in this tragic situation is worthy of mercy killing is not strange. Indeed, it can be true for some very tragic cases. My welfare calls for promotion. Stealing all my friends' money and secretly putting it in my wallet it is a way to promote my wellbeing. My welfare calls for stealing all my friends' money and secretly putting it in my wallet. Parking in the street is permissible. Leaving my car in the street with its front sticking out is a way of parking my car in the street. Leaving my car in the street with its front sticking out is permissible. Promoting my welfare is permissible. Stealing all my friends’ money and secretly putting it in my wallet is a way to promote my wellbeing. Stealing all my friends’ money and secretly putting it in my wallet is permissible. Here the explanation why the derivations fail is straightforward. When we are saying that parking is permissible, we are not saying that all ways of parking are permissible, only that some ways are. When we say that promoting my welfare is permissible, we are not saying that all ways of promoting my welfare are permissible, only that some ways are. Exactly which ways are permissible depends on other normatively relevant factors. The same explanation is applicable to Cullity's call-for cases too. When we say that my welfare calls for promotion, we are not saying that it calls for any way of promoting it, only some ways; which ways depends on other normatively relevant factors. Note that there can be serious disagreements about what those factors are and which ways of acting they rule out. For example, two people can both agree on all non-moral facts and also agree that suffering calls for prevention (so we agree that prevention is not undermined by other factors in this situation), but nevertheless disagree about whether a person's severe suffering calls for mercy killing. Indeed, this might be the situation Cullity and I are in when we think about his example about the child's torment. This points to an important further difference between Ross and Cullity. Ross puts all his cards on the table, as it were, when he lists his foundational principles about prima facie rightness. (Or at least he puts all his claims about prima facie rightness on the table; Ross is sceptical about general principles that take us from prima facie rightness to overall rightness). This is what you get, nothing more, nothing less. The derivation is precisely defined (deduction) and can be applied in a normatively neutral way by anyone with elementary logical skills. No matter which moral views you harbour, you can see what follows from what. (...) explains one of the main things that moral disagreement can be about. It can be about what does derive by subsumption, enabling, and responsiveness from fundamental norms of concern, respect, and cooperation (82). But if this is one of the main things we can disagree about, it surely looks like a fundamental moral disagreement about the fundamental moral importance of various factors, like in the mercy killing case above. But these factors should show up in the fundamentals of morality and not hidden away as parts of some auxiliary factor which is necessary for a successful derivation of one norm from another. But the aim is not to provide a set of necessary and sufficient conditions for successful derivations, nor to supply an additional set of theoretical primitives from which, in combination with the foundational norms, further norms can be deduced (70). But this means that we can't decide whether Cullity's project succeeds until we know exactly which factors he wants to supply to make the derivations successful. This will make friends of foundationalist system-building disappointed. Cullity promised to show that we only need to posit three foundational principles in order to derive (almost) the whole of morality. But then it turns out that we in fact need an indefinite supply of other normatively relevant background factors to succeed in these derivations. It thus a bit of an overstatement when Cullity says that his account is ‘genuinely explanatory’ since ‘it shows us how observed complexity is generated by a relatively simple underlying structure’ (vii). and that he is tempted to understand morality ‘as consisting in a relatively simple set of elements and compositional principles, with its limitless complexity being generated by their repeated combination’ (3). These statements suggest that, as in the case of syntax and semantics, we start with some basic elements and by some compositional principle generates more complex wholes. But in fact, what happens in Cullity's account is that for each application of a compositional principle—a combination of two different derivations, say—we need to rely on other normatively relevant factors in order to have a successful combined derivation. The set of successful derivations we start with will require additional normative factors, and for each combination of successful derivation possibly even further normatively relevant considerations have to be imported. But morality can be systematic without providing algorithms, I can hear Cullity say. I agree that a demand for algorithms would be to go too far. After all, not even dyadic predicate logic has an algorithm that determines when a formula is logically valid. But the logic still has a clear account of the conditions that are sufficient and necessary for a formula to be valid. As seen, Cullity's account does not give us such conditions for successful derivations, which will please the particularists but not the traditional foundationalist system-builders. In fact, Cullity himself seems to be tempted to look for necessary and sufficient conditions for some derivation relations. For heterogenous derivations (from presumptive fitness to actual fitness), but not for homogeneous ones, he does provide both necessary and sufficient conditions (or at least attempts to find such conditions). (See, for instance, the principle stated on p. 112). This is a curious difference which needs to be explained. Why should we expect that heterogenous and homogenous derivations differ in this respect? A further question is why the principles that provide the necessary and sufficient conditions for the derivation of actual fitness and reason should not themselves count as foundational principles. After all, they give us some idea about how we should determine overall fitness and reason on the basis of normatively relevant inputs. In this respect, they are similar to aggregation principles in normative ethics, for example, deontological principles that take us from claims and prima facie rightness and wrongness to overall rightness claims, or consequentialist principles that take us from claims about outcome-values to claims about overall rightness. Surely, these deontological and consequentialist aggregation principles are as fundamental as the principles about prima facie rightness and wrongness and the principles about outcome-values. More traditional particularists would say that both kinds of derivations would be unprincipled and no sufficient and necessary conditions can be provided in either case. Rather, we have to make do with default principles that tell us what has presumptive normative implications or significance. Cullity wants to position himself somewhere between the traditional system-builders who provide foundational general principles and clearly defined derivations relations and the mainstream particularists who eschew all general exceptionless principles. If I am right in my comments above, that position might not be stable

Epistemology · Field (mathematics · Focus (optics · Foundation (evidence · Foundationalism · Holism · Metatheory · Morality · Nothing · Political science · Sociology · Computer Science · Law · Mathematics · Philosophical Ethics and Theory · Philosophy · Political Philosophy and Ethics · War, Ethics, and Justification

• Concern, Respect, and Cooperation

  Garrett Cullity•Concern, Respect, and Cooperation•2018

  Add to list