Response to Coase’s Ghost’s Criticism of My Article on Particularism
Particularism remains implausible
0
A short while ago, I wrote an article criticizing moral particularism. One Coase’s ghost wrote a response to my article, featured here. In this article, I’ll respond to the criticisms of moral particularism. For background, one would be wise to read my original article criticizing particularism.
1
On my first objection about simplicity, Coase’s ghost (henceforth CG) writes
1) The particularist is not committed to the idea that there are no "uniting factor[s]" whatsoever. In fact, the particularist believes that it's typically wrong to humiliate people and to lie and so forth. She just recognizes, plausibly enough, contextual features that can accentuate, nullify or reverse these common moral tendencies. Just as a scientist recognizes that *most* metals are solid at room temperature, despite the belief being more complex than the silly view all metals are solid at room temperature.
CG claimed that particularists aren’t committed to there being no uniting factors and then expressed his commitment to the view that there are no uniting factors. Let me be clear about what I mean by a uniting factor. A uniting factor, in this context, would be a single underlying thing that all bad things have in common. The entire question is whether there’s a formula that one could design, even in theory, for determining which actions are choice-worthy. Particularism says no; it is thus not simple.
The analogy to metals is an error. Metals are not about the most fundamental feature of reality. They are instead a higher order emergent phenomena. Thus, having metals possess complex necessary and sufficient conditions—at least in the domain of temperature—isn’t a cost to the theory, for parsimony is only useful in the most fundamental account of things. The necessary and sufficient conditions of, for example, a chair are hard to pin down, but if fundamental physics were as complex as our descriptions of chairs, that would be a problem of immense significance.
Appeals to simplicity are highly problematic for other reasons. The simplest view is that there are no moral properties whatsoever. The generalist rejects this.
Simplicity is a theoretical virtue, and thus the fact that a theory is simple counts in favor of it. However, claiming that there are no true moral facts would be wildly implausible because it would lack explanatory power. There are some moral facts which we have reason to accept into our ontology—the foolishness of future Tuesday indifference, the wrongness of child torture, and so on—and positing that there aren’t moral facts requires denying these truths.
Thus, CG makes a fairly basic error. One can grant that simplicity favors a theory, ceteris paribus, while also holding that some theories which do better in simplicity will not be good theories.
We seem to be particularist about aesthetics: we don't think we can specify the simple set of principles that determine whether art is good or not. Nobody is thereby tempted to spell out the necessary and sufficient conditions for good art, which looks flat-out impossible.
There are several relevant distinctions between ethics and aesthetics in this regard. For one, I don’t think there are robustly objective facts about aesthetics, unlike about ethics. If everyone found a mountain beautiful in a twin earth, they would be making no error, even if that mountain is not something we currently regard as beautiful. I similarly don’t think that there’s a fact of the matter about whether one mountain is prettier than another.
Another relevant distinction is that aesthetics is reducible, unlike normativity. While I think it’s plausible that claims about what we should do are reducible to reasons, and reasons are reducible to value, there’s no reductionist story to be told about value. Thus, as in the response to CG’s first point, the error here is assuming that the facts about the deepest picture of reality will apply to higher order, reducible phenomena. The necessary and sufficient conditions for being a chair are more complex than the fundamental laws of physics, and that isn’t a problem.
Occam's Razor, in fact, is highly controverted and has been denied at length by, for example, Michael Huemer. It's justification is notoriously hard to pin down.
I’ll just quote my book in progress, where I respond to Huemer on this point.
Next they could argue like (Huemer, 2009) in his paper titled “When Is Parsimony a Virtue?” that simplicity isn’t necessarily a virtue. However, there are compelling arguments for simplicity as a virtue, particularly when applied to normative ethics. One such argument appeals to the sciences, in which simpler theories have been proposed to be better. Huemer worries (p.3) that this is circular, for the only way we can identify that science does best with simpler theories is by appealing to the simplest explanation of the data. However, this is not so. If simplicity were not virtuous, that would be because the world was, in reality, very complex fundamentally. If this were so, it would be unlikely that we’d be able to provide simple explanations of physics that adequately explain the data, the way we have successfully in physics.
Next, Huemer notes that this doesn’t explain why simplicity is a virtue. This is true, however, one can know that simplicity is a virtue without knowing why simplicity is a virtue. If an omniscient perfectly honest truth teller informed us that simplicity was a virtue, we’d know that it was a virtue, despite not knowing why it was a virtue.
Huemer objects to this being applied to philosophy (p.15), writing “No persuasive case for the use of Ockham’s Razor in philosophy can be made along these lines. The kind of evidence we have that science has been highly successful in identifying truths, we do not have in favor of philosophy’s efforts to attain the truth. The sort of consensus on substantial bodies of theory that we find in most sciences, particularly the natural sciences, is lacking in philosophy. Nor has philosophy produced the impressive sort of technology or successful predictions that modern science has.”
I would object to Huemer’s view here. As I shall argue later in the book, in a variety of philosophical domains, simpler theories like totalism about welfare have been shown to be true. However, the parallel between science and philosophy is not merely an analogy, at least, not in regards to normative ethics. This is because normative ethics attempts to identify the deepest underlying laws that explain what things are good and what things are bad. If the analogy to science establishes that the deepest underlying laws tend to be simple, the same would be likely to be true in the philosophical domain. Even if the analogy to science fails, there’s a plausible analogy to mathematics and rationality. It is reasonable to believe what one has most reason to believe, and mathematics is deductible from a small number of simple axioms. Ethics would be likely to be similar. (Brenner, 2017) provides additional arguments for the symmetry between science and philosophy, when it comes to the virtue of simplicity.
Huemer notes (p.6-7) that one can argue for simplicity in the following way. There are far more complex theories than simple theories that can explain data. It’s possible to posit a near infinite number of complex systems to explain phenomena. Thus, even if one thinks that the correct theory is equally likely to be simple or complex, any individual theory that is simple will be more probable on this basis than a complex theory. This is certainly true when it comes to ethics. There are only two or so very simple theories in ethics that are at all plausible, those being desire theory and hedonism, while there are lots of complex theories, such as the many different deontological accounts of rights, act omission distinctions, intent foresight distinctions, the vast array of possible virtues that virtue ethicists think matter, and many more.
Additionally, Huemer notes (p.7-9) that simpler theories have fewer adjustable parameters so they can’t accommodate as much data intrinsically. If we have three data points, the hypothesis Y=AX^2 + BX + C can accommodate any possible data, while Y=BX + C cannot. Thus, given that simpler theories limit the possible data more, data consistent with simpler theories provides stronger evidence for those theories.
This applies to normative ethics. If utilitarianism has a more narrow range of adjustable parameters than other theories like threshold deontology, the fact that it gets correct results across the cases is stronger evidence for utilitarianism than threshold deontology, because it’s more strongly predicted on the hypothesis that utilitarianism is correct that it would get correct results then it is on the hypothesis that threshold deontology is correct that it would get the correct results, based on the specific, finely adjusted parameters selected by the threshold deontologist.
An additional argument can be made for simplicity as a virtue, based on intuition. It is very intuitively plausible that the ultimate nature of reality is likely to be relatively simple. This argument would not persuade the skeptic, but it is hard to deny the intuitiveness of simplicity as a virtue.
One could object by claiming that utilitarianism is not simple. After all, it posits that a vast number of different mental states are desirable. In fact, there are plausibly infinite desirable mental states, so utilitarianism may have infinite stipulations. This response misses the point. There are similarly an infinite number of possible colors, but positing the existence of colors (at least as a feature of our mental states--not necessarily as a mind independent feature of the world), is not infinitely complicated. If there is a derivative property, then facts about that derivative property are not complex, any more than positing that color qualia are intrinsically good would be complex, merely on the basis of there being lots of color qualia. (Kringelbach & Berridge, 2009) argue that there is a satisfactory neuroscientific account of pleasure, meaning that positing the desirability of pleasure only requires positing a specific feature of one type of already existing thing. The authors note (p.481) “pleasure is never merely a sensation or a thought [29], but is instead an additional hedonic gloss generated by the brain via dedicated systems,” meaning that all the instances of pleasure share a common feature. Thus, positing the desirability of pleasure doesn’t violate the virtue of parsimony.
Next, CG says
Anyhow, simplicity must be weighed against explanatory power: the particularist has perfect power to explain our judgements, the generalist must, always, bent over backwards to explain away a whole class of highly dubious implications of their view.
The particularist theory is infinitely complex, which will outweigh explanatory power. Additionally, the generalist is able to better explain morality, as I argue at great length in my series defending utilitarianism. The fact that our moral views can reach reflective equilibrium and do so in accordance with the most virtuous theory is something that particularism can’t explain.
2
On my second point, arguing that there were true moral principles by appeal to examples, CG says the following
At least two problems here. Firstly, your examples are so complex we're likely describing a state-of-affairs, not a moral principle as standardly thought of. If you want to bake in every conceivably morally salient feature into a scenario and then come to a conclusion, that's not a rejection of particularism, that *is* particularism. That's no defense of "maximize net utility", or "follow the Golden Rule" or "never violate rights".
The examples don’t describe states of affairs—they instead describe principles that govern wide classes of cases. This is evident if one rereads the initial examples I gave, which I’ll quote here.
ere are several moral principles that seem true
1 If there is a large group of people who are
A) The only existing sentient beings
B) Have a very high quality of life
C) Don’t want to be tortured
D) No benefits to anyone else will accrue from torturing them to death
E) Are infants who have not harmed anyone at any point
F) Will not harm anyone at any future point
You should not torture them to death.
2 If some action would cause an infinite number of people who have not harmed anyone at any previous point and will not harm anyone at any future point to endure infinite suffering, when otherwise they would have lived good lives, you shouldn’t take the action
3 If some action violates rights, decreases desire fulfillment, decrease joy, decreases virtue, causes people’s well-being to be less proportional to their virtue, increases impure pleasures, decreases virtuous pleasures, decreases objective list fulfillment, that action should not be taken.
None of these describe particular cases. They are all wide descriptions of cases that encompass a vast swath of cases.
Perhaps most importantly, the particularist does not have to commit herself to the idea there are no defensible principles, only that the *rationality of morality does not depend on the provision of principles*. This is Dancy's own position. It won't do to point to things (really not principles) that seem universally applicable, the generalist claims *all* moral truths are fully reliant on principles. The generalist has set for himself a seemingly impossible burden.
This seems like an inadequate definition of particularist. The person who thinks utilitarianism gets the true result in all cases but doesn’t believe that it’s useful to reason in terms of principles would obviously not be a particularist. The SEP, written by Dancy, says
Moral Particularism, at its most trenchant, is the claim that there are no defensible moral principles, that moral thought does not consist in the application of moral principles to cases, and that the morally perfect person should not be conceived as the person of principle.
3
On my third point in which I claim there’s a parity between moral and scientific particularism. Why posit true scientific principles, rather than scientific particulars? CG says
3) I don't understand the supposed parity. Dancy's point is that we seem to, often, be in conflict about which set of values to follow. According to absolutism, values are never in conflict. In that way, absolutism is at odds with our moral experience. Nothing analogous emerges in science. We may be undecided between theories, but we don't feel like we've betrayed someone or something if we go for the Everett interpretation of quantum mechanics over the Bohm interpretation, at least not on scientific grounds. Aesthetics is the much better analogy, and it supports particularism.
Take the case of reconciling quantum mechanics with general relativity. In this case, positing particularism wherein there isn’t a general law that governs the interactions would seem to explain the data. It is, however, inadequate given its other costs. The burden is on the particularist to point out some salient distinction between the two. None have, in my view, been given.
Drawing an analogy between the alleged universality of moral principles and the universality of scientific laws begs the question and fails to answer Dancy's challenge of providing a justification for the universality. Scientific laws are at least partly justifiable via observation and induction. The particularist will insist moral observation strongly supports particularism (see how much trouble generalists have producing *any* intuitively-plausible principles). Presupposing otherwise just begs the question.
Given how vast the modal space is, it’s unsurprising that we’d find apparent counterexamples to any moral principle, even if they’re true. As I’ve argued in great detail in my utilitarianism wins outright series, reflective equilibrium does, in fact, converge on principles.
Normative ethics looks nothing like science, with millennia of stagnation over even the basic structure of morality, nothing that even resembles a consensus, no accepted methodologies, no equations and no widely popular candidates for moral laws. Again, aesthetics is clearly the superior analogy.
While there may have been millennia of stagnation about the structure of morality, something similar is true of science. Things can be understood while their basic structure remains obscure. There is no consensus about, for example, the correct view of quantum physics. Additionally, one finds similar types of agreement in morality and science. Both share agreement on an enormous range of issues—for example, when it comes to ethics, everyone agrees it’s generally bad to torture people, take people’s property, and much more. The method of reflective equilibrium is relatively accepted. There are popular candidates for moral laws, including those advocated by natural law theorists, the utilitarian law, and much more.
4
4) If our moral intuitions are often wrong, they're often wrong about moral principles. Worse, humans have a generalization bias: they seek, too quickly, to falsely simplify complex phenomena. See the witchcraft theory of disease, failing to account for bacteria, viruses, parasites, cancers, autoimmune disorders et cetera. This undermines generalism.
The reason I pointed out that our moral intuitions are wrong is to show that we’d expect the correct morality to diverge from our (often wrong) intuitions in a variety of cases. If this is true, then the apparent counterexamples to various theories are not a counterexample to generalism. If a theory perfectly matches our intuitions, then it’s just copying our intuitions—the correct theory would diverge in some cases from our moral intuitions.
The fact that the particularist *can* just dig in her heels hardly shows that it is plausible to think, e.g., skin colour is morally relevant in certain situations. Notice, of course, the generalist can do the exact same thing: proclaim a principle to the effect skin colour is important. As, for example, the Confederates did. At least the particularist is not committed to a universal "truth"!
But in the case of skin color, we could employ thought experiments of the following variety: imagine making a white person (which the racist thinks is morally salient) and then making their skin color change. In that case, would it be permissible to mistreat them? If they say yes, that would be a view that most racists would find hard to stomach. If they say no, then they’d have to appeal to other (false) descriptive claims about people of a particular skin color. However, the particularist can always double down and say that the features don’t spill over, because there are particular facts that make the reductio unsuccessful.
5
The particularist, in fact, is not committed to a posteriori moral principles. Why would that follow? Thinking about a case, on the part of a sufficiently sensitive person, in detail, may be sufficient for knowledge.
I’ll just quote my earlier article—for a more detailed account, see here.
According to moral particularism, moral facts can’t be deduced a priori, because they depend on the specific features of a case. Thus, we need to experience a case in order to know facts about it. However, this would mean we can’t make judgements about merely imagined cases that it seems like we can know a priori, like that an alien holocaust that inflicts massive suffering and benefits no one would be bad.
I say the following in the original article
Small children, for example, seem to learn that it’s bad to hurt others by being hurt by others and seeing the reactions of others from being hurt. However, if particularism is true, then they can’t extrapolate the general principle that it’s wrong to hurt people. However, if this is true, then it’s unclear how we can have moral knowledge that we can apply to other cases. What are we learning in moral cases if we’re not gaining knowledge of moral principles?
CG replies
On the child example: the particularist can just believe, plausibly enough, experience of suffering gets one to understand it's a thing to look out for, as Dancy put it. Other factors may override it, as the utilitarian himself believes, or it may be nullified or reversed in value, in other circumstances. What's mysterious about any of this?
However, my original article had several responses to this line of argument, which I’ll quote in full. This is another case where it seems like CG raises a point that was already extensively preempted.
This raises several problems.
1 It leaves us unable to determine when we should assume that like cases implications to spill over to other like cases. If there aren’t true moral principles, what reason do we have to think that harming someone is morally wrong in general, rather than the notion merely that harming someone is wrong in a few specific circumstances.
2 How does it inform us of anything if there are no principles that can be extracted? If, for example, scientific particularism were true, we couldn’t extrapolate trends from other cases, because there would be no static rules that govern all cases. Particularism holds the same is true.
3 If there are no true moral principles, what reason do we have to even think that like cases should be similar in their moral judgements. If, for example Z is a function of X^Y^A^B^C+BCA-LR^N, where each of them are variables, we couldn’t extrapolate trends based on the values for the values of any of the variables. All of them depend on the others. In the particularist case, it’s even more dire, for there are an infinite number of variables. Thus, how can we extract any general lessons about other cases if the permissibility of some action depends on an infinite number of specific features of the case?
Next, CG says
Tännsjö's argument proves too much in threatening aesthetics and there's actually no argument whatsoever for the coherentism. It's just asserted.
This is another case of CG raising a point that was preempted. As I said
Note, this claim about coherentism being the only appropriate method for ethics is consistent with the view that foundationalism is true broadly. This is not about our fundamental axioms, just about higher order ethical principles. When asked why I accept, for example, the claim about 10!!! dust specks being worse than torture , I can appeal to a series of other judgements that are more plausible. However, the particularist rejects any such judgements, so they can provide no deeper account of how they have moral knowledge.
Finally, CG says
It's, finally, not the case that particularist has no resources after rejecting the universality of principles (half true) or thought experiment (highly dubious, these could get us to consider factors we hadn't yet), she can *think about the case at hand*. What else would be better?
The first sentence was already addressed. Thinking about the cases at hand and not reaching reflective equilibrium doesn’t do much, given the poor track record of moral intuitions. This is not a reason to reject moral intuitions wholesale, it’s instead a reason we need to check them against other intuitions. Particularism eliminates nearly all of the theoretical resources to do this.
Conclusion
This article has responded to all of the points in CG’s response to my criticisms of particularism. However, the primary reason I reject particularism isn’t one presented in those articles. Instead, the case for utilitarianism is so overwhelming—as I’ve argued elsewhere at length—that it means we have good reason to accept one and only one moral principle.