Sam Harris' Moral Landscape objection

Question

I know that Harris' book, "The Moral Landscape" has been widely criticized by philosophers. And after listening to him talk with Alex O'Connor relatively recently here, I came to feel that Harris really isn't saying anything new at all. His claim of well-being objectively being an experience a person can subjectively have seems almost indisputable. Are the following critiques valid objections to his "moral landscape" being an objective framework of ethics?

Harris claims that there is a landscape of possible choices of action that will result in various values of well-being in a person. His axiom (though he often disagrees with the fact that he is defining a specific ethical framework) is that we must maximize well-being of people; we must find "peaks" in the moral landscape, as he describes. An immediate objection appears to me: whose well-being are you maximizing? Any action will generally affect different peoples well-being in a different direction or magnitude. Thus each person has an individual "well-being function". And it is meaningless to find a point that simultaneously optimizes a general set of functions; one can easily construct a set functions in which a maximum of one is a minimum of another. If Harris attempts to combine all individual well-being functions into a "universal well-being function", does he just arrive at a form of Utilitarianism? At which a very simple and extreme counter-example is available. Suppose we were capable of torturing one infant for all of eternity, and doing so would grant extreme well-being to everyone else alive or to be born. Torturing the infant would definitely be a peak on the moral landscape; the extreme well-being of everyone would far outweigh the suffering of a single infant. But that is intuitively incorrect.

Furthermore, the idea of even having a correct "universal well-being function" also seems meaningless. It's meaningless to even calculate a mean across everyone's well-being because peoples subjective experiences of well-being cannot be compared to each other without defining some sort of value to each individual's experience. I would say it is even meaningless to assign real numbers to each person's subjective experience in any modal scenario; I don't believe well-being can be ordered. Even if it could, two peoples' "5.0 well-being" are different subjective experiences. And any function that translates one persons 5.0 to another persons 5.0 would once again be assigning value to different people.

If we do not compare "well-being functions" and simply optimize our own well-being, we just get Egoism correct?

Even if there is a neurological way to objectively assign a real number to a persons subjective experience, we still have the problem of which function we optimize.

Answer 1

In Moral Landscape, Harris set out to breach the is/ought barrier. He argues that morality is a natural fact, and can be studied as a field of science.

This is an interesting premise. It is not obviously wrong. But unfortunately it was not well served by Harris's mismanagement of the discussion.

IF one wants to breach the is/ought barrier, AND wants to use science to do so, THEN there are two things one must do:

• Identify inescapably real phenomenon one can use as the basis of this moral science
• Demonstrate, with science data, that a science-based method provides the greatest utility and understanding in interpreting this data in moral terms.

Harris fails at both tasks.

What Harris starts with is welfare, which is a nice improvement on early utilitarianism, which focussed on the more basic but far less important pleasure/pain. BUT Harris limits himself to a neurologic reductionist framework, and welfare/harm is not even translatable in the terms he limits himself to. Harm and welfare are emergent concepts, not reductionist ones, and are not localized in time, nor can they be characterized neurologically. If Harris wanted to work just with pleasure/pain, he would at least have some neural analogs to point to. But WELFARE? HARM? He misses out on the first bullet above.

THEN, Harris provides NO rationale for his choice of a welfare/harm utilitarianism. Instead, he repeatedly asks a rhetorical question "what other alternative is there"? WELLLLL ... anyone even moderately versed in ethical theorizing, could identify a dozen! Just take a freshman level intro to philosophy, and at least a half dozen will be discussed!

Any actual "scientist" would have identified the alternate hypotheses that others have proposed (pleasure/pain utility, rights ethics, harm minimization, darwinian law of tooth and claw, eusociality, virtue ethics, teleological ethics, clean hands, deep ecology, etc), then would have proposed goodness metrics to evaluate between them, then would have cited experimental evidence that shows the preferability of his approach over his rivals. Yet Harris, despite being a vocal advocate for scientism, does not seem to actually understand what science consists of, and fails to do any of these steps.

I consider this failing of Harris's to be characteristic of the advocates of scientism, who are often as woefully unaware of what good science consists of as Harris appears to be. His use of rhetorical sleight of hand, rather than science, in his argument for scientizing morality, serves as a thorough condemnation of his project.

Along the way, he commits several other absurdities. He argues that the uniqueness of consciousness is what makes humans particularly morally valuable. BUT -- very many animals are almost certainty conscious, yet he does not address animal welfare/rights at all.

Also, he cites the need for non-scientists to accept science consensus, basically because civilians cannot understand the details of how an experiment is done well or badly and data analyzed. Then he advocates for NEUROSCIENTISTS to therefore be our moral theorists, and named several philosophers too as among the expert community who should do our moral thinking, in Daniel Dennett and the Churchland's. Philosophers DO do pretty complex thinking, but few understand neuroscience to the point of expertise, and neither Dennett nor the Churchland's likely qualify. But of particular note, is that the Churchland's are eliminative materialists about consciousness, and Dennett is the leading Delusionist about consciousness. Their views lead them to DENY the phenomenon he is using to even make a moral argument -- hence per his own thinking, they CANNOT be experts on morality, because they deny its object.

For a review making these points, see: https://www.amazon.com/gp/customer-reviews/R3SB3ZQ7Y8SX1Y?ref=pf_ov_at_pdctrvw_srp

How to do this right, rather than how Harris did it?

Start with admitting that most subjects are not actually sciences. Moral thinking is -- a branch of philosophy. It can and should be informed by evidence, but scientific studies are not going to be sufficiently clear and definitive to settle moral questions.

Then abandon a narrow ontology. DON'T rely upon materialism, dualism, or neutral monism to derive a valid morality. Our world's ontology is an open question, so don't lock the answer into a very likely wrong assumption.

Treat our moral intuitions as actual sense data about the world, but colored by our cultural upbringing. Some moral thinking is enmeshed with a culture's prejudices. But others, can be seen as moral across cultures. Look for the cross-cultural moral commonality, and treat those intuitions as sense data about the reality of morality.

We have three widely respected moral frameworks cross-culturally. Utilitairnaism, rights ethics, and love/truth virtue ethics. There are two others that I add to these three, to make a collection of moral thinking tools: Eusociality, and Deep Ecology. Use these 5 cross-cultural tools to do a moral evaluation. If they all agree, then you know what is moral! If 3/5 agree, and the two outliers are in regions of decision space they do poorly in, once more you know what is moral!

For an example of how to do this, see: https://philosophy.stackexchange.com/a/78826/29339

Answer 2

I believe Sam Harris' mistake is a kind of begging the question; what is at issue in discussions about value are if value exists and what kind of existence value(s) has, if any. It seems to me that Harris assumes that well-being is what is exclusively valuable, which is not an uncontroversial position (e.g. this SEP page on well-being goes through some challenges to this view).

If we take it as given that well-being is what is valuable then, even if we do not think we are obligated to maximize everyone's well-being, we are committed to the view that it is in fact better for overall well-being to be higher. That is, we might not be compelled to maximize well-being even if it is in fact better for well-being to be improved.

If it is not well-being per se that is valuable but one's own particular well-being I think you can avoid the entailment that higher overall well-being is better (independent of moral obligations to pursue that higher well-being). Such a view comes with its own difficulties, though, as an argument would need to be made that there is something special about the fact that any particular constituent experience of well-being is "one's own" that makes that well-being valuable but that fails to obtain for others.

Though the baby thought experiment you raise might seem to be an unacceptable trade-off, I think utilitarians of various sorts would bite the bullet and affirm that in such a hypothetical it would be better overall to torture the baby. However, utilitarians could counter that the world in which that thought experiment is possible is one very far from our own and so the unintuitiveness of that commitment should not be too worrying.

Regarding potential egoism, it could be that each individual is generally in a uniquely good epistemic position to assess their own well-being and so pursuing that is actually good for others inasmuch as many of the things that can help well-being require cooperating with others (community, kinship, love, and nowadays a specialized market economy pushing the technological envelope).

Answer 3

Harris claims that there is a landscape of possible choices of action that will result in various values of well-being in a person. His axiom (though he often disagrees with the fact that he is defining a specific ethical framework) is that we must maximize well-being of people; we must find "peaks" in the moral landscape, as he describes. An immediate objection appears to me: whose well-being are you maximizing? Any action will generally affect different peoples well-being in a different direction or magnitude. Thus each person has an individual "well-being function". And it is meaningless to find a point that simultaneously optimizes a general set of functions; one can easily construct a set functions in which a maximum of one is a minimum of another. If Harris attempts to combine all individual well-being functions into a "universal well-being function", does he just arrive at a form of Utilitarianism?

If you read his book, Harris deliberately refrains from defining a function.

One of the key arguments the book tries to make is that, much like medical science is preoccupied with studying the effect of real-world problems on human health to improve it in general (without defining a specific global health function), a science of ethics should similarly study the effect of real-world problems on human well-being to improve it in general.

Even medical science works with trade-offs, and it leaves the judgment of those trade-offs to health professionals and patients themselves. A drug is not going to make you +X healthier, but rather have a range of positive effects as well as a number of potential side-effects, and there's no single global function to evaluate those.

Harris seems to be pushing for studying the positive and negative impact of actions and policies on well-being in general and doesn't seem too concerned with finding a global utilitarian well-being function. While "the moral landscape" is a theoretical visualization of such a function, it's used simply to illustrate the point that well-being is worth optimizing and can be optimized.

At which a very simple and extreme counter-example is the available. Suppose we were capable of torturing one infant for all of eternity, and doing so would grant extreme well-being to everyone else alive or to be born. Torturing the infant would definitely be a peak on the moral landscape; the extreme well-being of everyone would far outweigh the suffering of a single infant. But that is intuitively incorrect.

Just to comment on this counter-example — you're assuming here that the universal well-being function would be additive. It could very well not be. If it was, say, multiplicative, a zero well-being value for the infant would make the global well-being function equal zero as well, regardless of how high the well-being of others would be.

Furthermore, the idea of even having a correct "universal well-being function" also seems meaningless. Its meaningless to even calculate a mean across everyone's well-being because peoples subjective experiences of well-being cannot be compared to each other without defining some sort of value to each individual's experience. I would say it is even meaningless to assign real numbers to each person's subjective experience in any modal scenario; I don't believe well-being can be ordered. Even if it could, two peoples' "5.0 well-being" are different subjective experiences. And any function that translates one persons 5.0 to another persons 5.0 would once again be assigning value to different people.

If we do not compare "well-being functions" and simply optimize our own well-being, we just get Egoism correct?

Even if there is a neurological way to objectively assign a real number to a persons subjective experience, we still have the problem of which function do we optimize.

The fact that there isn't a clear-cut way to design a universal well-being function does not mean that well-being cannot be optimized.

Perhaps the set of well-being outcomes is not a totally ordered set, but rather a partially ordered set (where only some elements can be compared). Having a cold or being yelled at by a stranger might not be directly comparable, but in either case, your well-being would likely be better without either of those things. And there are a large number of cases that can be similarly evaluated even without a specific global function.

Answer 4

Per your:

Furthermore, the idea of even having a correct "universal well-being function" also seems meaningless. Its meaningless to even calculate a mean across everyone's well-being because peoples subjective experiences of well-being cannot be compared to each other without defining some sort of value to each individual's experience. I would say it is even meaningless to assign real numbers to each person's subjective experience in any modal scenario; I don't believe well-being can be ordered. Even if it could, two peoples' "5.0 well-being" are different subjective experiences. And any function that translates one persons 5.0 to another persons 5.0 would once again be assigning value to different people.

... this all well-echoes Rawls (A Theory of Justice):

The significance of this question can be brought out by a comparison with utilitarianism. When applied to the basic structure this view requires us to maximize the algebraic sum of expected utilities taken over all relevant positions. (The classical principle weights these expectations by the number of persons in these positions, the average principle by the fraction of persons.) Leaving aside for the next section the question as to what defines a relevant position, it is clear that utilitarianism assumes some fairly accurate measure of utility. Not only is it necessary to have a cardinal measure for each representative individual but some method of correlating the scales of different persons is presupposed if we are to say that the gains of some are to outweigh the losses of others. It is unreasonable to demand great precision, yet these estimates cannot be left to our unguided intuition. Moreover, they may be based on ethical and other notions, not to mention bias and self-interest, which puts their validity in question. Simply because we do in fact make what we call interpersonal comparisons of well-being does not mean that we understand the basis of these comparisons or that we should accept them as sound. To settle these matters we need to give an account of these judgments, to set out the criteria that underlie them (§49). For questions of social justice we should try to find some objective grounds for these comparisons, ones that men can recognize and agree to. I believe that the real objection to utilitarianism lies elsewhere. Even if interpersonal comparisons can be made, these comparisons must reflect values which it makes sense to pursue. The controversy about interpersonal comparisons tends to obscure the real question, namely, whether the total (or average) happiness is to be maximized in the first place. (pg. 78)

Or then:

[Edgeworth's] conception of cardinal utility suffers from well-known difficulties. Leaving aside the obvious practical problems and the fact that the detection of a person’s discrimination levels depends upon the alternatives actually available, it seems impossible to justify the assumption that the social utility of a shift from one level to another is the same for all individuals. On the one hand, this procedure would weigh identically those changes involving the same number of discriminations that individuals felt differently about, some having stronger feelings than others; while on the other hand, it would count more heavily the changes experienced by those individuals who appear to make more discriminations. Surely it is unsatisfactory to discount the strength of attitudes, and especially to reward so highly the capacity for noting distinctions which may vary systematically with temperament and training.⁴⁵ Indeed, the whole procedure seems arbitrary. It has the merit, however, of illustrating the way in which the principle of utility is likely to contain implicit ethical assumptions in the method chosen for establishing the required measure of utility. The concept of happiness and well-being is not sufficiently determinate, and even to define a suitable cardinal measure we may have to look at the moral theory in which it will be used. (pgs. 282-83)

He goes on to address what is called the Neumann-Morgenstern definition, etc. In the limit, a general problem with aggregationism (we can refer to Harris' position as this, rather than consequentialism or utilitarianism, if we see fit to make such fine-grained distinctions) seems to be that the ideal of "maximizing the good" (whatever that is) can be manipulated too easily to justify too many options. Once the aggregationist admits a mismatch between an action's being right per the principle of summation, and an action's being obligatory per reasonable belief that the action is most in line with that principle, then it seems possible to end up "reasonably believing" that every decision one makes will have the same final impact on the sum of value anyway,^🌌 or then since no matter how much short-term negative utility an action generates, it will (with the exception of the permanent extinction of all life in existence) have its utility function smoothed over upward at some time in the future (kind of like how Hercules will indeed kill the Hydra, if he keeps trying long enough) (c.f. Pascal's mugging or the "repugnant conclusion" and its cohorts), then we have not quite so much that aggregationism is "false" as that it is trivial (like an explosion in deontic logic).^Y

---

^🌌According to Narlikar and Padmanabhan[01]:

Also Silk[00] reads, "Even if it is not infinite today, the universe is destined to become arbitrarily large in the future." However, to my understanding, Bostrom is primarily referring to the question of eternal inflation,^T so I'm not sure how to put together all these references.

^YAlternatively, let's take the equation of goodness and happiness (or utility or whatever) at face value. However, if what this is, is that aggregationists are stipulatively defining terms like "good" and "right" in such a way, then aggregationism isn't necessarily even a moral theory in the first place. For the more generic sense of "good" or "right" is something like "rational," but it is not automatic that prioritizing the maximization of happiness over everything else is more rational than trying to balance one's intentions with everyone else's so as to avoid aggressive behavior (Kant's so-called "analytic" principle of rightness), or more rational than focusing on one's mental fitness (like in virtue ethics; and a reliable critique of utilitarianism and Kantianism^E is that if they are taken too seriously, the result is serious damage to the believer's emotional and intellectual health).

---

---

^TBostrom does cite Tegmark, specifically this essay. Here, Tegmark specifically claims that the Level 1 multiverse (the symmetrical-inflation one) is "rather uncontroversial." More detailed:

Although the implications may seem crazy and counter-intuitive, this spatially infinite cosmological model is in fact the simplest and most popular one on the market today. It is part of the cosmological concordance model, which agrees with all current observational evidence and is used as the basis for most calculations and simulations presented at cosmology conferences. In contrast, alternatives such as a fractal universe, a closed universe and a multiply connected universe have been seriously challenged by observations. Yet the Level I multiverse idea has been controversial (indeed, an assertion along these lines was one of the heresies for which the Vatican had Giordano Bruno burned at the stake in 1600^†), so let us review the status of the two assumptions (infinite space and “sufficiently uniform” distribution).

So I myself don't know how true any of that is, it seems to run counter to what Dcleve is also claiming to be the mainstream picture of these issues for example. On the other hand, I've never heard Tegmark accused of falsifying the mainstream record, either, so who knows. Per this PhysicsSE question about the very topic, one of the comments reads: "I think this is on-topic, but understand it might result in a fractious, partisan discussion (which I discourage)." That is not very reassuring (even for a (closed!) 2014 question/comment).

^ESurprisingly, Kant's own views about mental health appear to have included an acknowledgement that mental disorders can be non-culpable interference/dampening factors when it comes to our use of abstract reason, see Thomason[21] for an exploration of this topic. Still, if one takes e.g. much of the casuistical material in The Metaphysics of Morals at face value, one wonders what might happen to a practitioner of such morals. There is the discouraging story of Kant's letters to-and-from a certain Maria von Herbert to consider, for example...