Clarifying Sam Harris’ clarifications

by Dwayne Holmes

[Editor’s note: this essay is an expansion and follow up to the author’s submission to the contest organized by Sam Harris for the best criticism of his arguments on science and ethics, as laid out in The Moral Landscape.]

The semantics of “science” is important

In responding to Ryan Born’s essay [1] — which won the competition giving readers a chance to challenge the arguments Sam Harris made for merging science and ethics in his book The Moral Landscape (henceforth, TML) — Harris undermined most of the discussion concerning the “scientific” nature of his theory with this statement:

“The whole point of The Moral Landscape was to argue for the existence of moral truths … every bit as real as the truths of physics. If readers want to concede that point without calling the acquisition of such truths a “science,” that’s a semantic choice that has no bearing on my argument.”

Harris is right that the choice of calling TML theory a science (or not) is a semantic issue, which would not touch the validity or practical utility of his theory. However, that does not mean the decision is without serious consequence.

Harris’s expanded definition of “science” relies on a loosening of criteria that can become problematic for those in traditional scientific fields such as physics and biology. His most questionable claim is that the existence of answers in principle provides sufficient grounds for defining something as scientific. If that were true, Intelligent Design (ID) theory would become classified as a legitimate science, as there are answers in principle to the questions they ask. The difference between ID theories and traditional scientific theories is that the methodology underlying ID cannot generate answers in practice. For many that is a critical distinction (and Harris admits his approach may not meet that criterion).

If we decide to accept a broad definition of science (just to let Harris’s moral theory “in”), future court cases regarding science education may then hinge on being able to explain the difference between science (for people in lab coats) versus science (for everyone else) such that they shouldn’t be taught together in a “science” class. Why make the difficult job of protecting legitimate science education any harder than it already is?

Harris should concede that TML theory is not science as most people use the term, perhaps adopting “scientia” instead, as Massimo Pigliucci advocates, as a term covering the building of rational knowledge beyond strict empirical approaches.

Distinguishing between moral systems is important 

Harris argues that the concerns of different moral systems reduce to concerns about consequences:

“Similarly, if virtues such as generosity, wisdom, and honesty caused nothing but pain and chaos, no sane person could consider them good. In my view, deontologists and virtue ethicists smuggle the good consequences of their ethics into the conversation from the start.”

By using the most generic conceptions of “consequence” and “good,” it is possible to force deontology and virtue ethics to fit into the category of consequentialist theory. But that would not change the fact that traditional consequentialist theories (like utilitarianism or TML theory) are characterized by vastly different ideas regarding what consequences are desired (including for whom) and how to compare choices when making a moral judgment. In fact, outside the extreme end of avoiding absolute misery for everyone, Harris has not made a case that alternative systems compare practical consequences at all while rendering moral judgments, much less in the same way as consequentialist theories do.

Different moral systems can legitimately place importance on the way people conduct their actions, with a view toward perfecting the individual or society (in an esthetic sense) rather than toward overall gain (in a practical sense). For example, a well done action that saves no one (and costs one’s own life) can be viewed as holding greater moral merit than a devious or slipshod action that saves lives.

One real life example could be the martial code of Bushido in Japan, exemplified in the story of the 47 Ronin [2]. This is clearly not a story about calculating the maximization of some practical ‘good’ or ‘flourishing’ (beyond that of an individual’s character). The gravitas of the story comes from a commitment to one’s duties or beliefs that are held sacred (virtuous) for themselves, in spite of massive costs.

If “consequentialism” is broadened to such an extent that the interests, mechanics, and results of Bushido get lumped together with Mill’s utilitarianism, that term has lost much explanatory power. Eventually people will have to reinvent terms similar to the ones we already use to distinguish such systems, though now as subsets of consequentialism. Ultimately, accepting Harris’s argument simply shifts debate to what flavor of consequentialism is correct and so should be practiced.

“Ryan seems to believe that a person can coherently value something for reasons that have nothing to do with its actual or potential consequences… It is true that certain philosophers have claimed this… But I don’t find this claim psychologically credible or conceptually coherent.”

That an abstract principle might be chosen (credibly) over practical consequences can be seen with a simple hypothetical. Imagine that scientific evidence emerges that a false belief in wholly benign star fairies (who help when all natural/scientific measures have been exhausted, and require no other false beliefs or actions against others) leads to greater happiness, health, and longevity. According to traditional consequentialist theories (including Harris’) it would be right to maintain that false belief and promote it in others. More important, it would be wrong to promote doubt in others.

However, many people would find that an unacceptable moral conclusion. Those practicing “atheism” regarding these fictional beings, because they prefer honesty (or curiosity, truth seeking, etc), are coherently valuing something other than practical consequences.

There are many more hypotheticals that can be considered, such as refusing to engage in cannibalism, killing children, or forcing a woman to become pregnant even if one of these actions were required to save humanity from extinction. It is psychologically and conceptually valid to say that a world that requires that to maintain its existence (even if temporarily) is not a world worth saving.

It may be noted that such a concern (let’s take cannibalism) is still about consequences and does not reduce just to “cannibalism is bad.” Specifically, it is taking into account the psychological consequences one would face from such an experience. However, that move simply supports the overall argument being advanced. A traditional consequentialist theory (including Harris’s) would not, and theoretically could not, prioritize consequences from a specific act (or for a specific individual) over others. Traditional consequentialist theories are about maximizing a specific goal, in Harris’s case “well-being,” regardless of means.

If the fate of all humanity came down to one person having to choke down some man-flesh for a while, the choice would be crystal clear to a traditional consequentialist. In contrast, deontological theorists and virtue ethicists can take into account, and prioritize, specific methods or consequences to specific groups. This is true even in the face of extinction, making them very different moral systems indeed.

Distinguishing factual errors from moral errors is important 

“[T]he inner and outer consequences of our thoughts and actions seem to account for everything of value here. If you disagree, the burden is on you to come up with an action that is obviously right or wrong for reasons that are not fully accounted for by its (actual or potential) consequences… [and]… I don’t believe that any sane person is concerned with abstract principles and virtues — such as justice and loyalty — independent of the ways they affect our lives.”

Several examples of actions being judged right/wrong without appeal to practical consequence have already been given above. It seems especially hard to accept the label of insanity for those valuing truth over the beneficial delusion (placebo effect) of fictional beings.  Alternative challenges have been advanced by writers such as Massimo Pigliucci who flip the problem back to Harris. For instance, assuming that science found that cultural practices oppressing women actually resulted in net positives for societies, would Harris switch to accepting them [3]? To this I might add the question: in such cases, would it really require insanity to oppose them?

As it happens, (even without hypothetical benefits) the practice of female genital mutilation (FGM) does shift the burden of proof while revealing a critical flaw in Harris’s moral theory. This is because it shows how, by basing moral judgment on outcomes, TML theory loses the ability to distinguish mistakes (factual errors) from intentions (commonly considered the basis of moral errors).

TML contains scathing criticism of FGM, suggesting a comparison between FGM cultures and a sadist cutting up young girls for pleasure [4]. However, there is a clear difference in intent (mental states) between the two. The intention of parents practicing FGM is to help their child and their society (even if they are horribly mistaken about what they factually achieve). This is obvious when one considers that those practicing FGM have sought modern medicine to remove any similarity between the inadvertent outcomes of the procedure (physical suffering and danger) and the intended results of the sadist. Indeed, the only remaining “problem” would be the injustice of altering the physical features of a child without their consent (which, of course, is something commonly accepted for males in the West). Intriguingly, some people have moved to block access to such medical services (in the West and abroad), with the intent of preventing acceptance of FGM, despite the fact that their actions inherently lead to the very suffering and death of innocent girls that FGM practitioners were seeking to avoid.

A purely results-based consequentialist theory (which ignores mental states) cannot discriminate between these alternatives, treating them as roughly morally equivalent despite the vast differences in intent. It is reasonable to find such a conclusion mistaken.

Indeed, for many, conflating factual error with moral error would seem to be a major misfire during the initial test run of any moral theory.

Distinguishing between descriptive and prescriptive ethics is important (and shouldn’t his fans care?)

Harris certainly did clarify a mistaken impression with these statements:

“I also disagree with the distinction Ryan draws between “descriptive” and “prescriptive” enterprises. Ethics is prescriptive only because we tend to talk about it that way… We could just as well think about ethics descriptively… In my view, moralizing notions like “should” and “ought” are just ways of indicating that certain experiences and states of being are better than others… There need be no imperative to be good — just as there’s no imperative to be smart or even sane…. [and separately]… Ryan, Russell [Blackford], and many of my other critics think that I must add an extra term of obligation — a person should be committed to maximizing the well-being of all conscious creatures. But I see no need for this.”

While it is clear that Harris (in TML) equated statements of how to achieve well-being with “oughts,” I (and others) apparently misread that as elevating factual claims to the level of moral imperatives (that one ought to do it). It was not obvious that Harris had intended a wholesale assault on prescriptive ethics, by going the other direction and reducing oughts to mere shorthand descriptions.

On the contrary, TML appeared (again, to me) to be an opening shot against moral relativism and anti-realism. Perhaps this confusion arose from Harris’ claim of being a moral realist, while repeatedly attacking both moral relativists and anti-realists. According to moral realism right and wrong exist, and so do imperatives. Otherwise, how would this view differ in a practical sense from the antirealists who challenge the objective existence of prescriptive moral claims?

It is also hard to square the emotionally charged language found throughout TML, with the purely descriptive enterprise Harris now claims to be conducting:

“[p. 42, my emphasis] The difficulty of getting precise answers to certain moral questions does not mean we must hesitate to condemn the morality of the Taliban — not just personally, but from the point of view of science.”

If morality is about solving navigational problems, and judgments of “bad” are shorthand for failing to act intelligently or sanely, how exactly does “condemn” fit into the picture? Does one talk about condemning inadequate navigational charts? People with low IQs? People with neurological or psychological disorders? And what is the moral difference if one “hesitates” to condemn such people, given that the Taliban must be incapable of understanding the condemnation (analogous to lacking sufficient intelligence or sanity)? “Condemn” seems to carry a greater connotation than suggesting that they could improve their game (if they want), or that they lack sufficient ability to understand moral concepts.

And finally, TML showcases several direct instances of using prescriptive terms:

“[p. 49, emphasis in the original] We can think more clearly about the nature of moral truth and determine which patterns of thought and behavior we should follow in the name of morality.”

“[p. 80, italics are Harris, underline is mine] We have already begun to see that morality, like rationality, implies the existence of certain norms, that is, it does not merely describe how we tend to think and behave; it tells us how we should think and behave.”

Given statements such as these, people cannot be criticized for holding an impression that Harris had already built obligations into the structure of his moral theory. It seems hard to read “should” as meaning simply “if you happen to want to do X, then…”

But let’s take him on his word. This clarification means he just pulled the rug out from under all of his fans, who thought they could use scientific-sounding moral statements to promote or reject practices with some sort of moral force. It would be interesting to know how many people were surprised or disappointed by this clarification (or if neither, then actually understood the implications).

Truly, “throwing acid in the faces of young girls is bad” is now directly translatable to (according to Harris’ TML 2.0) “throwing acid in the faces of young girls is not the best thing you could do, but there’s no reason you have to do any better if you’re fine with being kind of like stupid or crazy.” And while Nazis were factually not maximizing the well-being of the undesirables they were stuffing into crematoria, and so descriptively being “bad,” there was no moral imperative to be doing anything other than stuffing undesirables into crematoria.

Of course, in his clarification, Harris makes a comment that tries to have it both ways:

“[my emphasis] What does it mean to say that a person should push this button? It means that making this choice would do a lot of good in the world without doing any harm. And a disposition to not push the button would say something very unflattering about him… I think our notions of “should” and “ought” can be derived from these facts and others like them. Pushing the button is better for everyone involved. What more do we need to motivate prescriptive judgments like “should” and “ought”?”

His apparent return to advocating a prescriptive ethical theory will be left for readers (and Harris) to work out.

Admittedly, moral skeptics (and other anti-moralists such as myself) could be quite comfortable with the kind of limited motivation and weight for “prescriptive judgments” Harris outlines here. But this still seems incongruous with a position of moral realism, and the level of charges he wants to make against specific practices and cultures.

Does Harris really hold (and do his fans accept) a moral judgment that Nazis simply exhibited “unflattering” dispositions?

Harris’s known unknowns are important (and troublesome)

“Following Hume, many philosophers think that “should” and “ought” can only be derived from our existing desires and goals — otherwise, there simply isn’t any moral sense to be made of what “is.” But this skirts the essential point: Some people don’t know what they’re missing. Thus, their existing desires and goals are not necessarily a guide to the moral landscape. In fact, it is perfectly coherent to say that all of us live, to one or another degree, in ignorance of our deepest possible interests. I am sure that there are experiences and modes of living available to me that I really would value over all others if I were only wise enough to value them.”

It seems true that experiences may generate novel interests, or allow us to discover interests we had not realized were within us. These may become important in ways we could not have predicted earlier. But this does not counter the fact that we can only derive oughts from our existing desires and goals. It simply suggests that our existing desires could be transient, not permanent.

But let’s take those last two sentences at face value. If indeed Harris can be ignorant of his deepest possible interests, that there are other modes of living that he might value more, is it not possible that he is ignorant of his ability to value things other than gradable (and so maximizable) single parameter consequences?

More importantly, if the argument he makes here is valid, how can any moral landscape map, much less conclusions from a moral landscape map, be made at this time? It seems implausible for anyone to have known all possible perspectives so as to discount any future changes in desires or goals. This is especially true for someone admitting ignorance on this topic, thereby undercutting all moral judgments Harris made against different practices in TML.

Maximizing well-being for everyone will create conflicts in the real world

“And unless one were to posit, against all evidence, that every person’s peak on this landscape is idiosyncratic and zero-sum (i.e., my greatest happiness will be unique to me and will come at the expense of everyone else’s), the best possible world for me seems very likely to be (nearly) the best possible world for everyone else.”

It only takes the inherent conflict/incongruity between one person’s peak and another’s to challenge the mechanics of TML. I have addressed the impact of inherently conflicting interests/values on TML in my full response to Harris (using his own analogy of health to the value of well-being) and so will not repeat them here [5]. Instead I will try a different tack on the same problem.

Some people require well-ordered environments, with near if not factual military regimen (i.e., externally driven order) to enjoy a sense of personal flourishing. This removes distractions allowing for increased productivity, contentment, and happiness. Others find such conditions suffocating. Similarly, some find extensive social networks useful, while others need solitude.  Some find value in conflict (even if nonviolent) while others demand calm cooperation. Some tradition, others novelty.

It is impossible for these people to maximize their personal concepts of well-being in the company of the other. However, in a world of finite space, time, and resources it is necessary for all these people to interact. Some sacrifice of maximized well-being (based on personal assessment) will be necessary. Harris says that his best possible world will likely be (nearly) the best possible world for everyone else. That “nearly,” even if accepted as the limit of sacrifice required, is crucial. Will he accept (nearly) the best world for himself, so that someone else can reach the actual best for themselves?

More importantly, what is the objective procedure by which we can decide who should sacrifice and how much? Harris’ silence on these matters, in TML and in his clarification, are holes in his theory and not merely frontiers to be explored.

The false analogy between Harris’ “science of well-being” and the factual “science of medicine”  

“Ryan writes that “Science cannot show empirically that health is good,” but he admits that, without this assumption, “the science of medicine would seem to defy conception.” I believe morality is also inconceivable without a concern for well-being and that wherever people talk about “good” and “evil” in ways that clearly have nothing to do with well-being they are misusing these terms. In fact, people have been confused about medicine, nutrition, exercise, and related topics for millennia. Even now, many of us harbor beliefs about human health that have nothing to do with biological reality.”

Speaking as a researcher in medicine I can confirm that one does not have to assume that “health” is “good.” All medicine requires is a desire to achieve or avoid specific physical effects, and an instrumental curiosity regarding how to reach those goals. This desire does not have to be shared (assumed “true”) by everyone involved in the process. Granted that there are many common or colloquial concepts of “health” that would be shared by most people. But the differences defy uniform consensus.

Harris has used this vagueness in defining “health” to suggest “well-being” holds some scientific merit. In the essay I submitted in response to Harris’ challenge [6] and in the full response [5], I explained why this is not the case.  Briefly, “health” provides no useful unit of measure for the scientist. Assuming for the sake of argument that the goal of scientists was to improve health because they are motivated by a belief that it is “good,” they would not be using some landscape map with “health” as the single unit of measure. Even something as clearcut as “death” can require multiple parameters to make a scientific judgment. If well-being truly is analogous to “health,” then a science of well-being would require specific, well-defined parameters that allow for relatively clear measurements. Improved well-being would be a cumulative moral “diagnosis” based on many different landscape maps keyed to different, well-defined parameters.

It is at this point that Harris’ mapping system for well-being falls apart. Moral parameters do not share an equivalent meaning across patients as those related to health can. Also, moral parameters effect more than one person, which is never the case for physical parameters in medicine (unless you are pregnant or a conjoined twin). My blood pressure levels, and what I must do to maintain them, do not effect you. An increase in concern for social order, however, will have a direct negative effect on everyone else’s personal autonomy.

It is true that many people have been confused for millennia about medicine, nutrition, et cetera.  One of the primary sources of confusion has been listening to people who talk in vague, ill-defined but important sounding terms, suggesting that they can be trusted to deliver solid results.  The solution has been listening to other people that, even if not promising the world (and getting a bit boring with technical details), provide the clear definitions and evidence required to connect actions to results. The former deal in “health” as “healers” and “health experts,” the latter deal in the “science of medicine” as researchers and medical professionals.

Since Harris has conceded he is not referring to people in lab coats (and the strict criteria used by such people) when using the term “science,” he needs to retract the false equivocation between his “science of well-being” and the factual “science of medicine.”  Modern medical science is driven exclusively by people in lab coats (except when wearing smocks, or in the office).

Empirical truth claims are not (necessarily) contingent on well-being

“I would argue that satisfying our curiosity is a component of our well-being, and when it isn’t — for instance, when certain forms of knowledge seem guaranteed to cause great harm — it is perfectly rational for us to decline to seek such knowledge… I’m not even sure that curiosity grounds most of our empirical truth-claims. Is my knowledge that fire is hot borne of curiosity, or of my memory of having once been burned and my inclination to avoid pain and injury in the future?”

The first sentence is challenged by the earlier example of maintaining atheism, despite benefits from false beliefs. While it can be rational to choose safety over curiosity (truth seeking), that does not mean it is irrational to choose the other way around. Different moral preferences can be equally rational and valid.

The second sentence only shows that truths can be discovered by happenstance, rather than exploration, or that curiosity can be prompted by external events rather than self-initiating. The kid who gets burned does not learn anything more about fire unless it stimulates his curiosity to discover more about fire.

In any case, this argument seems to ignore that many people choose to learn about fire without ever getting burned. At least for them, “I want to know because it is interesting” can be the motivating factor. How many have pursued such knowledge, despite the fact that they could end up being burned in the process?

Reading Hume is important

“We have certain logical and moral intuitions that we cannot help but rely upon to understand and judge the desirability of various states of the world. The limitations of some of these intuitions can be transcended by recourse to others that seem more fundamental. In the end, however, we must work with intuitions that strike us as non-negotiable.”

If this is Harris’s position on morality, he might be interested in picking up a copy of An Enquiry Concerning the Principles of Morals. It was published in 1751, and in it David Hume argues basically the same thing. He also went on to argue that the fundamental moral intuition of humans, that non-negotiable intuition that others collapse into, was promoting the interests of mankind. This is also known as utility and bears a striking resemblance to Harris’s concept of well-being.

Here are three quotes from Hume on this subject:

“In all determinations of morality, this circumstance of public utility is ever principally in view; and wherever disputes arise, either in philosophy or common life, concerning the bounds of duty, the question cannot, by any means, be decided with greater certainty, than by ascertaining, on any side, the true interests of mankind.”

“Upon the whole, then, it seems undeniable, that nothing can bestow more merit on any human creature than the sentiment of benevolence in an eminent degree; and that a part, at least, of its merit arises from its tendency to promote the interests of our species, and bestow happiness on human society.”

and just to show he also lumps virtues in with consequentialist theory…

“… public utility is the sole origin of justice, and… reflections on the beneficial consequences of this virtue are the sole foundation of its merit…”

The key areas of incompatibility are that for Hume: 1) moral judgments cannot be founded solely on logical intuition/empirical study (that is what the is/ought distinction means), and 2) in order to distinguish between moral error and factual error intentions rather than results must be the focus of moral judgment. These differences are of course critical, and could be used to fix some of the flaws in TML theory.

Distinguishing moral certainty from lapses in moral cognition is important

“The universe is whatever it is. To ask whether it is logical is simply to wonder whether we can understand it. Perhaps knowing all the laws of physics would leave us feeling that certain laws are contradictory. This wouldn’t be a problem with the universe; it would be a problem with human reasoning. Are there peaks of well-being that might strike us as morally objectionable? This wouldn’t be a problem with the universe; it would be a problem with our moral cognition.”

It is unclear how this admitted “problem with human reasoning/moral cognition” does not open the door to the very moral relativism Harris attacks in TML. For example, how is the Nazi attempt to maximize well-being for everyone left alive, by removing the undesirables (causing a permanent loss of well-being for a minority within a short window of time), not simply another “moral” peak that some find objectionable because they suffer from deficits in their moral cognition? Or at least how is this avenue of argument not open to Nazis?

Surely this kind of argument undercuts Harris’s critique of practical relativists (such as anthropologists) who actively assume the possibility their moral cognition is wrong in order to better study and so understand the moral systems of others.

So, in summary

There seems to be a great discrepancy between the heated, objective sounding judgments pouring forth from Harris on the attack, and the dispassionate, purely descriptive, potentially accommodating moral system Harris claims to use when pressed to defend his theory.

Given all the above, my hope is this:

1) Harris will retract his claim that TML theory is science, and that answers in principle are sufficient grounds for a scientific theory. Broadening the definition of science is unnecessary, and potentially damaging to traditional science education.

2) He will no longer use the false equivalence/analogy linking his moral theory of well-being to the science of medicine. Just because people are motivated by similarly vague terms does not mean they use equally precise methods.

3) He will actually read Hume’s works on moral philosophy to understand what he is trying to criticize. His criticisms have consistently been straw men, and in reality he might find less impediments and some fixes for his theory.

4) He will stop claiming that moral judgments can be made regarding cultural practices using TML theory (or science) at this time. His mapping system based on polar extremes of well-being/misery is clearly incomplete and lacking relevant data.

_____

Dwayne Holmes is a PhD student in neuroscience at Amsterdam’s VU University Medical Center, with prior degrees in philosophy and molecular biology. He is particularly interested in how science and philosophy impact our understanding of ethics (from molecules to social norms), and when he gets a chance he runs a website (using an increasingly pointless pseudonym) devoted to a form of moral skepticism/antirealism (thegooddelusion.blogspot.com).

[1] The Moral Landscape Challenge: The Winning Essay.

[2] Forty-seven Ronin.

[3] About Sam Harris’ claim that science can answer moral questions, by Massimo Pigliucci, Rationally Speaking, 6 April 2010

[4] The Moral Landscape (2010), by Sam Harris, Free Press, p.46.

[5] Against a moral landscape: full response, The Good Delusion.

[6] Challenge essay, The Good Delusion.