About Me

My photo
Australian philosopher, literary critic, legal scholar, and professional writer. Based in Newcastle, NSW. My latest books are THE TYRANNY OF OPINION: CONFORMITY AND THE FUTURE OF LIBERALISM (2019); AT THE DAWN OF A GREAT TRANSITION: THE QUESTION OF RADICAL ENHANCEMENT (2021); and HOW WE BECAME POST-LIBERAL: THE RISE AND FALL OF TOLERATION (2024).

Friday, February 04, 2011

The measurement, or metric, problem

Some of my disagreements with Sam Harris are rather theoretical and rarefied, though I find them no less interesting for that. In particular, I'm not actually that worried about the use of an expression such as "well-being" or the idea that well-being and morality are connected in some way. Although well-being seems to me to be a fuzzy concept, standing in for more precise concepts that have not really been nailed down, I don't think there's anything inherently wrong with using such concepts. Other examples include "flourishing", "harm", and, very likely, "health". I'm not planning to abandon the words that I've put in inverted commas in this paragraph. In many contexts, I would not dispute the claim that morality is (somehow) about requiring individuals to restrain their actions for the sake of collective well-being. This is a bit vague, but the general idea that it conveys seems about right.

Where I part company is when Harris seems to think that well-being is something we can measure, maximise, and at least in principle, compare in an objective way. In his writings he often refers to the practical difficulty of counting in situations such as the following: how many breaths did I take on a certain day? how many blades of grass are there in my lawn? how many grains of sand are there on a particular beach? If we define the question a bit more precisely, it will have an actual quantitative answer in each case, but there are practical difficulties in arriving at the answer. If the problem of measuring well-being is like that, the implication is that there are, indeed, units of well-being, and there is an actual quantitative answer to the question of whether one set of laws and customs (say) produces more well-being than another. It's just that there may be practical difficulties in making the measurement. Harris even contemplates the theoretical possibility that there could be a tie between two societies with different legal and customary arrangements.

Talking in this way does suggest that well-being is something with a metric. Despite the practical difficulties of measurement, there will be objective, quantifiable answers as to which set of social arrangements, or which course of action, produces more well-being than others. If we could get over the practical difficulties, we could say that course of action X produces more well-being units than course of action Y and is in that sense "morally better" than course of action Y.

Much in The Moral Landscape suggests this interpretation. In particular, when the book discusses various criticisms of utilitarianism, it seems to accept that well-being is - by analogy with "utility" in these discussions - something that can be quantified and maximised, despite the practical difficulties of measurement.

The Moral Landscape could probably be trimmed of some of this language, or some clarifications could be made to the effect that well-being can't actually be quantified, and I'd have less problems with it. However, I do need to insist that there is a difference between lacking a precise and practical method of counting (as with the grains of sand on the beach) and lacking a metric at all (perhaps because the concept is fuzzy and is really a placeholder for more precise concepts). In the former case, objective comparisons can be made and there is something to maximise. In the latter case, we have a more complex situation. If the concept used is a placeholder for other things, and if different people can quite rationally weigh those things differently, there will be cases where there is no objective answer as to which course of action actually maximises "well-being". That's not to deny that there can be extreme and clear-cut cases; nor is it to deny that perfectly rational and non-arbitrary judgments can be made. However, it does leave room for a class of disagreements where no one is objectively wrong.

We usually accept this with value judgments outside the area of morality, but there seems to be a psychological pressure to think that morality must be different, perhaps because what is at stake seems so important, and perhaps because societies, families, and religious traditions tend to indoctrinate children with the idea that their established morality is "just true" or absolutely authoritative or inescapably binding. Perhaps there are other reasons why we are tempted to treat morality as different from other value judgments, but in any event the temptation is surely there.

Harris talks about the measurement problem like this:
Even if we did agree to grant "well-being" primacy in any discussion of morality, it is difficult or impossible to define it with rigor. It is, therefore, impossible to measure well-being scientifically. Thus, there can be no science of morality. (The Measurement Problem)
But that is not the problem that I see. First, Harris does seem to "get" here that we are not talking merely about practical difficulties in measuring well-being (as with the grains of sand on the beach) but about whether we even have a rigorous definition of well-being. If we have only a fuzzy definition of well-being we may still be able to have perfectly rational conversations about it, but they will leave areas of rational disagreement, areas where there is no fact of the matter as to which party is right and which party is wrong. Even in principle, this will be enough to prevent us adopting an overriding policy of maximising well-being.

Whether or not that prevents us having "a science of morality" is another thing. I don't see why we can't still at least have a scientifically informed practice of morality. Harris may think that doesn't go far enough, but I have happily said in various places that I think he is approximately right in many of his views. I think it's not quite correct to talk about a science of morality, but this may be partly a matter of semantics. For example, I don't consider the design of automobiles to be a science, but I certainly acknowledge that it is informed by science - and just as well. I don't consider the practice of medicine to be a science, but again it is informed by biomedical science - and again, just as well.

So, okay, if Harris is using a concept of science broad enough to include the practice of medicine (by a GP, say) as a science, along with the practice of automobile design, then nothing I've said so far rules out a "science of morality". Indeed, it may well be appropriate to talk about a "science of morality" even from my viewpoint, but in rather a different sense. If anthropology is a science, then that part of anthropology that examines the social phenomenon of morality can be a science or a scientific specialisation.

My problem when I raise the idea of a metric is not so much with the possibility of a science of morality in some sense. It's that we can't simultaneously think that well-being is something which lacks a metric while also thinking that morality is ultimately about producing the maximum amount of well-being. Or that there will, in principle, despite the practical difficulties of measurement, always be an objectively correct answer as to which of two practices produces more well-being than another (or whether we have a tie).

I don't know how far Harris is really committed to this idea. There are passages in the book where he seems to use it. Even the idea of the moral landscape is developed in such a way as to involve tied quantities of well-being. Since he has a broad conception of science, however, he could pursue his idea of a "science" of morality while also accepting as a scientific outcome that there are cases where we don't have a  mere tie. Rather, because well-being is a fuzzy concept we have a situation where fully informed rational people can disagree on a course of action without either being objectively wrong.

That is certainly the case in the practice of medicine. Biomedical science finds out all sorts of things about the functioning of the human body and how this can be affected by various decisions that we might make as well as by events that are not our choice (such as invasion by infectious micro-organisms). When a GP or a specialist actually practises medicine, and when a patient actually makes decisions that affect the functioning of her body, there is very good reason to take into account the findings of biomedical science. But that doesn't mean we can simply read off from biomedical science how a particular patient should act in a particular case.

We could, for example, have a situation where a patient may be able to get some significant reduction of the pain that she's experiencing, but only at the expense of dying sooner. Or she may be able to gain a greater capacity to get around and do things, but only at the expense of being in greater pain. Two patients may be in the same situation but have different values - one prefers less pain even at the expense of dying sooner, while the other is prepared to put up with greater pain in order to get that extra period of life - so they act differently. In such a situation, we don't try to say that one decision more than the other maximises "health" and is therefore objectively superior. Rather, we accept that different people can make different decisions based on the same information without either behaving irrationally or being simply wrong.

The word "health" is a useful one in many contexts. It is not simply arbitrary to prefer not to be suffering from a debilitating disease or, say, a broken leg. We use the word "health" with a great deal of confidence and fluency, and I don't suggest for a moment that we abandon it. But it would be an illusion if we came to think of it as something that can be quantified in itself in some objectively correct way so that there is always as an objectively correct answer as to which possible course of action is correct from the point of view of "maximising health".

At one point in his recent response to critics, Harris actually acknowledges this point with respect to health (and also well-being). He says:


What if there are trade-offs with respect to human performance that we just can't get around -- what if, for instance, an ability to jump high always comes at the cost of flexibility? Will there be disagreements between orthopedists who specialize in basketball and those who specialize in yoga? Sure. So what? We will still be talking about very small deviations from a common standard of "health" -- one which does not include anencephaly or a raging case of smallpox.
The thing is, I don't disagree with this. In a very large class of cases it will be obvious how to act. We wish to cure people who have debilitating and potentially fatal diseases, and we mourn the birth of a baby with anencephaly. We do, indeed, have a great commonality of values that leads to consensus on these things. Given the sorts of beings that we are, our agreement to try to prevent (as we've now successfully done) or cure cases of smallpox is not something that we reached arbitrarily. But that's just my point.

We do not have to choose between (1) some kind of crude relativism in which anything at all counts as "health" or "well-being" or as a "good" car or knife or legal system, and (2) a kind of naive objectivism in which there is always, at least in principle, a correct answer as to what maximises some governing value, so that there is no room for legitimate disagreement by well-informed rational people. We have more sophisticated alternatives to either picture, and to the extent that the folk (or philosophers) think that picture (2) is part of the very definition of morality ... they are in error when they make moral judgments. But I agree with Harris that people who have adopted picture (1) are also in error.

It may be that Harris really agrees with me to some extent, judging by some of his comments in his reply to critics, but there is much in the book (and even in his reply) that suggests that he still has picture (2) in mind.

The problem with a metric is only a problem for someone who is thinking in terms of maximising something. Whatever is to be maximised must actually be quantifiable. Someone could have an objectivist picture in mind but be working with, say, a deontological system, and the metric problem will not arise. There are, however, deeper problems with objectivism which I won't address today.

Meanwhile, I don't think the metric problem prevents a scientifically-informed practice of medicine or even a scientifically-informed critique of laws and customs. It need not necessarily stand in the way of Harris's overall goals, but it can create a difficulty for the way he sometimes talks and writes. I'm reasonably happy to accept those ways of talking and writing as thought experiments and approximations, but it does seem that Harris really does believe two inconsistent things - that well-being is not the sort of thing that can have a metric, but at the same time it is an overarching value that we are to maximise in some quantitative sense. I think it's the second of these that has to go. Of course, Harris could fall back to some other value that is a more plausible candidate for being the sort of thing that can be maximised - pleasure, for example - but this will have its own problems.

I'm quite happy for him to go on talking about well-being, with the understanding that this is a fuzzy concept and to some extent a placeholder. That is a perfectly reasonable way of talking, and I do it myself, but it does impose certain limitations on the kinds of theoretical conclusions you can draw. In particular, it becomes difficult to talk like a utilitarian as Harris often does. But talking like a utilitarian has other problems (as Harris acknowledges in his brief discussion of Parfit), and it is not required for rational, reasonable, non-arbitrary discussion of laws, cultures, customs, and the other things that Harris and I both want to subject to rational critique.

34 comments:

Alex said...

Hmm, making this an option (1) vs. (2) seems to be setting up straw-men easily dismissed. I think you actually agree more than what this reply seems to indicate, but I'm sure you've got more to say.

I'd like to chuck in that this seemingly "problem of measurement" shouldn't get the "Philosophy doesn't work that way" treatment from the get go; there's a very interesting debate to be made there, and so far I'm quite with Sam's argument that even if we can't measure stuff accurately doesn't mean we cannot do so. How good was that movie? Impossible to tell, but I know it was better than X and worse than Y. We humans work tremendously good at this level, dipping straight into cognitive sciences. A measure is a measure even if it is fuzzy.

Charles Sullivan said...

I had a philosophy prof who, when discussing the measurement problem of pleasure. jokingly called a unit of pleasure "a hedon."

In that same spirit, I propose that a unit of well-being be called "a bene."

Pronounced "bay-nay" (or ben-nay, if you like) kind of like one would say it in Spanish, Italian or Latin.

Russell Blackford said...

A measure may be a measure even if it's fuzzy, but it's not an objective measure. It leaves room for rational disagreement.

I actually think that Sam and I agree on quite a bit. He actually concedes that there are at least some cases where there is no determinate answer (or I think he does that in his latest discussion of health). He could probably apply that more rigorously and he'd end up changing a few passages in the book, but it wouldn't be fatal to his overall project.

This is metric problem is actually the least of his worries, I think. Although it's an easily demonstrable point, and one that I think he should accept - and apparently does to an extent - by itself it leaves the overall project in pretty good shape if he takes it on board. There are deeper problems, such as "Even if 'well-being' were a less fuzzy concept why expect others to use it?" And, "Why shouldn't they give greater weight to their own well-being or that of their loved ones, if that's what they want to do?" But those issues are not the subject of this particular post.

Brian said...

I do not agree with Harris' approach to morality or your critique of it here.

"If we have only a fuzzy definition of well-being we may still be able to have perfectly rational conversations about it, but they will leave areas of rational disagreement, areas where there is no fact of the matter as to which party is right and which party is wrong."

How does having a definition that is non-computable leave areas of rational disagreement?

"It's that we can't simultaneously think that well-being is something which lacks a metric while also thinking that morality is ultimately about producing the maximum amount of well-being."

It seems to me you are ignoring the concept of mathematically non-computable functions. Mathematics-more precise than science-acknowledges the existence of unique real numbers that can't be computed even in theory, even though they quantify real measurements. I admit that I only poorly understand this concept myself.

"We could, for example, have a situation where a patient may be able to get some significant reduction of the pain that she's experiencing, but only at the expense of dying sooner. Or she may be able to gain a greater capacity to get around and do things, but only at the expense of being in greater pain. Two patients may be in the same situation but have different values - one prefers less pain even at the expense of dying sooner, while the other is prepared to put up with greater pain in order to get that extra period of life - so they act differently."

So there are objective, scientifically discoverable differences between the patients-their values. How would a third party selecting the pain averse one for the pain decreasing treatment be making anything other than a correct decision from the health perspective? How can a declaration that patients should be treated without regard for their preferences be considered anything but wrong from a health perspective?

"In such a situation, we don't try to say that one decision more than the other maximises "health" and is therefore objectively superior. Rather, we accept that different people can make different decisions based on the same information without either behaving irrationally or being simply wrong."

I do say that each patient has the treatment that maximizes her health, and cumulatively making such decisions maximizes health. People in different circumstances make different choices with the same information: I, stranded in the desert, choose to have my nanoreplicator make me water, my clone with my same knowledge and values who is drowning in the ocean has his nanoreplicator make him a boat. Similarly, if the difference in our circumstances is not physical facts about the world outside of our skulls but facts about how our neurons align such that we value different things, health for me will mean muscular power and flexibility as I enjoy rock climbing, whereas my clone seeks endurance and lower upper body weight so he can run marathons, as he desires.

Thanny said...

Your main problem is that you persist in defining "science" narrowly. I suppose that's understandable for a philosopher, where there's a visceral rejection of the notion that philosophy done right is just a form of science.

But it's wrong. Your example of looking out the window to check the weather being a way of knowing other than science is just incorrect. It's thoroughly scientific, unless you make an arbitrary choice to construct a whitelist of which forms of rational thought are "science", and which are some other form of rationality.

As for a metric of well-being, return to the health example that Harris seems (rightfully) quite fond of. There is no unit called "health", or even any other unit capable of converting "health" into a mass noun. But there are a host of metrics, which can be said to correspond with good health or bad health. High cholesterol? Tends to be bad health. High blood sugar after fasting? Ditto. I'm sure you get that point.

Well-being has similar kinds of metrics, applied at the social level. High murder rate? Well-being is not so great. High degree of health care coverage? Does not bode ill for well-being.

As with health, no single metric gives a complete answer, and there will be exceptions to the norm (such as that population of people with cholesterol levels over 1000, who have no cardiovascular disease as a result).

And trotting out "Brave New World" is no rebuttal - unless you think the scientific basis of health is entirely nullified by the idea that someone in perpetual suspended animation is in maximal health, which no sane person would agree with, no matter how many otherwise-applicable metrics contradict them.

It's also entirely unreasonable to expect a metric to be purely objective. The amount of science (narrowly construed) that depends on data that can be rationally disagreed about is staggering. How far away is that star? The ways we have of measuring are objective, in that anyone can do the raw measurements, but also open to rational disagreements about what those raw numbers actually mean.

Bottom line, as morality is nothing more or less than an emergent property of physical brains, it's impossible for it to *not* be measurable (in principle) and consequently subject to rational investigation (i.e. science).

Jambe said...

Mr Blackford, thanks for this. I had several objections to Harris' book which I couldn't quite solidify in my mind. You've elucidated those reactions and have done much more. My understanding of and interest in this wider subject has been greatly increased by you two.

I naturally doubt that "issues of morality" can have objectively-correct answers at all. Specific events or scenarios have solutions that are more concrete and fleshed-out, but they're not necessarily any more objectively correct or "proper". Whether you're talking broadly with vague words or specifically with concise ones, trying to determine the "correctness" of possible behaviors always seems to involve approximation, trade-off, context & circumstance, etc.

I find this interesting, and liberating. I suppose I see it as liberating because of my upbringing, when right and wrong were black and white and derived of an authoritarian god. Also, that's not say I don't have strong feelings about certain issues, but rather that I tend to ask myself "why" before I act on them much more than I once did.

Unknown said...

Russell, I disagree with your comment that without a metric something cannot be maximized. Sam's point about health is one interesting example. But in terms of engineering, this sort of thing is done all the time and quite successfully. Quality of a product is something that does not have a metric and yet engineers have found ways of maximizing it. I think the best examples of how this is done come from Taguchi who developed the "loss function" for maximizing quality.

Sarge said...

Mr. Blackford,

I really enjoyed this article, and personally think that "quantification" has its uses and such, but people seem to get bogged down in process, and the meaning, on the "quantifier"'s end always seems to be an abstract, and take on his meaning rather than the subjects.

As per Mr. Johannesen, I can see it: "A kick in the crotch as opposed to a poke in the eye with a sharp stick...let's crunch the numbers...hmmmm"

It is "anecdotal", I know, but what really got my head up on the article was something that happened to me last year.
I had some facial reconsstructive surgery and afterwards encountered pain that was very, very bad.
The highly educated medical doctor seemed to be rather off-hand about it. I inquired as to why so little was done about it and was informed that pain was "unquantifiable" and, therefore, wasn't an issue.
(a nurse told me a bit later that I might have noticed that he stayed well out of reach of me physically as well as emotionally. At least one other patient with whom he shared this tid-bit reached out with a cane and allowed him to examine the subjective/objective differences of the perceptive experience a tad closer than he liked.)

Also, I confess myself to rally enjoy your commentary on the "Gnu" atheists.

Although I am an atheist I am a musician and play in many churches, and when I hear Harris, Hitchens, and Dawkins mentioned, the religious have adopted the philosophy: "No 'gmus' is good 'gnus'".

Russell Blackford said...

EJ, I'm open to learning more, although I'm sceptical about whether we're talking about the same thing. Still, I'd be interested to see in what sense, or in what circumstances, you can objectively maximise something without any unit of measurement for what you're trying to maximise.

Thanny, at the moment I'm confused. Are you saying that well-being is something that can be measured objectively, like the raw measurements of distance to a star (e.g. 60 light-years)? Or are you saying that is not like that but we can't expect it to be?

Brian, my point here is that the "correct" decision for patient A will differ from that for patient B because they value different things or have different desires. Because of this, they are not seeking the same outcome, and neither is making a mistake.

Now you can reply that there's a sense in which they are seeking the same thing. I.e., each is seeking satisfaction of her own desires. But if you want to say that they are therefore seeking the same thing, namely "health", and that what we call "health" for A depends on what A values, while "health" for B depends on what B values, you have turned health into something that is relative to the desires possessed by individuals.

Perhaps it is, at least to some extent. But I doubt that Sam thinks that it's okay for you to try to maximise whatever, given your desires, you regard as "well-being", while I try to maximise what I regard as "well-being". He seems to want well-being to be something that is the same whoever is measuring it.

Brian said...

"...you have turned health into something that is relative to the desires possessed by individuals."

Which of the following are primary sources, (as the term is used in the study of history): A Bridge Too Far, Company Commander, WWII for Dummies, Rise and Fall of the Third Reich, Day of Infamy, War as I Knew It, With the Old Breed: At Peleliu and Okinawa?

I will address the second point next to give anyone reading this time to seriously think about that and settle upon his or her conclusions.

"But I doubt that Sam thinks that it's okay for you to try to maximise whatever, given your desires, you regard as "well-being", while I try to maximise what I regard as "well-being"."

You mean his ideal society might require cooperation? My friend Jack Sprat told me that while Harris desires the maximization of well-being, he wouldn't endorse an anarchy in which each individual tries to maximize his or her own desires on the smallest timescale and without a social contract. Rather, Harris would want people to work together to jointly do the best they can to fulfill desires.

"He seems to want well-being to be something that is the same whoever is measuring it."

I think insisting this same thing can't be desire fulfillment is like insisting one must give all hungry people the same type of food, regardless of their allergies and constitutions, or that one must always chop up living people as organ donors (to be a true utilitarian). Is failure to distinguish between relevantly different cases an endemic feature of utilitarianism (even in theory), critics of utilitarianism, both or neither? That I do not have the answer to.

I do have the answer to my first question: all of those books are primary sources.

Some are secondary sources for studying WWII, others are primary sources for studying WWII, and those that are secondary sources for studying WWII are primary sources for other purposes, such as what the For Dummies series is like.

Roughly speaking we use "subjective" to refer to things that depend on values and beliefs and "objective" to refer to things that are true regardless of epiphenomenal minds.

Declaring that any outlook or moral system that considers objective facts about brain states are "subjectivist" probably both occludes understanding and fails to respect the intent of people advocating for the viability of "objective" moral systems. At best you have a linguistic quibble with them. I do not think that facts about brain states are commonly called "subjective facts" nor do I think they should be interpreted as such absent strong evidence of idiosyncratic intent, particularly in material written for popularization of ideas like Harris' is.

(The metaphor is that everything is fair game for objectivity just as everything is a primary source. Some books are also about distant times or places just as some facts are subjectively true, but every book yields information about the circumstances of its creation.)

One final note: statements about what Harris would or would not think are testable!

Brian said...

"I'd be interested to see in what sense, or in what circumstances, you can objectively maximise something without any unit of measurement for what you're trying to maximise."

http://en.wikipedia.org/wiki/Kolmogorov_complexity

"The theorem says that among algorithms that decode strings from their descriptions (codes) there exists an optimal one...Kolmogorov used this theorem to define several functions of strings: complexity, randomness, and information...there is no way to effectively compute K."

@Thanny

"The amount of science (narrowly construed) that depends on data that can be rationally disagreed about is staggering. How far away is that star? The ways we have of measuring are objective, in that anyone can do the raw measurements, but also open to rational disagreements about what those raw numbers actually mean."

How could there be a rational disagreement if all parties have all relevant information, with "relevant information" very broadly construed: i.e. including scientist A having noticed over the years that multiple independent studies at .05% confidence level are more reliable than single studies at .001% confidence level, and scientist B noticing that his robotic hand thrashes about independently when his subconscious has correct insights that constitute brilliant interpretations of evidence that support theories he already tentatively favors?

The example of star distance seems like a mere issue of measuring...do you mean to show that difficulty of measurement is not good evidence of non-objectivity? Russell gets that, which (among other reasons) is why this is the best place to debate Sam Harris' ideas.

March Hare said...

Well being is used as a placeholder for the aggregation of other metrics that actually ARE (potentially) quantifiable so why the heck are we using well being?

By using well-being we are allowing people to have self-contradictory values (metrics) that, if explained, might cancel each other out whereas by using an aggregate they can be gloriously cognitively dissonant.

Even if you give Sam everything he is wishing for, grant him a metric for well being, a method for measuring it both on an individual level and a societal level and you'll still find that the vast majority of moral decisions are uncertain because they are based on trade-offs between the present and the future, either for people or between generations. This is a judgement call based on your particular values and there is not even in principle a way to work out the moral answer to that conundrum.

Obviously there are some that are just wrong e.g. a scorched earth policy that harms present and future, or some that are definitely moral [but not maximally] such as when there is a benefit to present AND future. I actually think that this lack of a 'discount rate' is at least as bad for any idea of scientifically objective morality as the lack of a metric.

Jimbo said...

Thanks Russell, I really appreciate your addition to this discussion. I would characterize the moral landscape this way: there are absolute answers to moral questions on a scale of morality with undetermined units. We are confident about what makes a peak with great certainty and likewise what determines and characterizes a valley. As we approach the inflection point, call it 'sea level', confidence reduces to zero i.e. we cannot with any certainty say A is better than B when they are close in moral equivalency. Again, so what? There is still moral certainty--very strong confidence at the extreme--which is good enough for action, NOW. Maybe our metrics will grow better in the future and our confidence about deciding which action is more moral on solid footing.

People are criticizing Harris for advocating that all people head for the peaks in a scientifically totalitarian way. I don't think he is. Rather, I think he's advocating an "all valleys should be eradicated" strategy. Can anyone deny that if individuals/societies occupying a valley (the Taliban) were raised to sea level that human prospering by any measure would be improved? Can't we (whether at sea level or on a peak) demand the Taliban change for the better? Absolutely.

I believe there are moral truths and absolute answers, it's just that sea level is confusing because we see personal taste and preference and equivalent outcomes clouding the data. And it might be that since the moral outcome at sea level is equivalent, the free exercise of choice is the most important thing.

J. Abraham said...

Russell, you are not analysing this carefully enough, in my opinion. For instance, consider your example of the patient who could live longer at an increased level of pain. If there are no logical and/or empirical criteria by which we can arrive at an answer, then certainly there is NO ANSWER AT ALL to this particular question. Your own "opinion" on the correct answer to such a question would be completely devoid of content, because you have already admitted that there isn't an answer in any meaningful sense.

All this is perfectly compatible with Harris' philosophy. He has compared equally valid moral preferences to different tastes in food. It's a bad idea to eat poison, even if we can't say whether banana ice-cream is better than vanilla.

Unknown said...

Scientists (at least in my field) present their measurements along with a “confidence interval”. Because measures—even “objective” measures—are inevitably imperfect, we can never be 100% confident that two measurements differ; however, sometimes the differences are large enough relative to the measurement uncertainty (and concomitant confidence interval) that we can, for all practical purposes, conclude that the difference is real. If measurements of well-being contain any information at all, then, even if they come along with large confidence intervals, the same would apply to them. That is, we could state that “this option is better” at a certain level of confidence (e.g., 25% or 75% or 95%), and at a certain agreed upon level of confidence (which may vary depending on an appreciation of the costs and benefits of making a mistake) we would be empowered to deem, for the purpose of taking action, the better option more "moral”.

Brian said...

@March Hare

"...self-contradictory values (metrics) that, if explained, might cancel each other out..."

Please demonstrate that some values, any values, contradict each other.

In my experience actions can be mutually exclusive, but I have never known any values to be.

"This is a judgement call based on your particular values..."

Is what people have as particular values eternally opaque to science?

"...there is not even in principle a way to work out the moral answer to that conundrum."

Therefore one doesn't exist? You know better than to imply that.

And how do you know that there is no argument in principle that can work out the moral answer? I've thought of what I think is an excellent answer already. Have you shut off your monitor, faced a blank wall, and spent five minutes trying to think of a good answer? Two minutes?

Russell Blackford said...

Okay, folks, I can't possibly respond to all these very diverse answers. I do thank you for them all. Maybe I should let you argue among each other because your views differ more radically among each other than mine differ from Harris's.

But I'll pick up one of two here and there as the discussion goes on.

J. Abraham says: "He [Harris] has compared equally valid moral preferences to different tastes in food. It's a bad idea to eat poison, even if we can't say whether banana ice-cream is better than vanilla."

I don't think this is correct. Harris imagines that there can be a tie between two moral choices in the sense that both deliver the same amount of well-being. That's the only kind of "equally valid" moral choices that he seems to acknowledge. But there can only be a tie if there really is a truth as to what amount of well-being is delivered in each case. There may of course be practical difficulties in measuring the amounts, but there is some truth of the matter as to how much well-being is generated.

But there won't be any such truth of the matter unless we agree on what well-being actually is. Is it pleasure, is it absence of pain (this is not necessarily the same thing or on the same scale), is it the ability to do certain things, is it actually doing certain things (such as having children), is it an already-moralised concept such that well-being is really being virtuous, is it preference satisfaction, or some mixture of some or all of these things, or what?

Until we know what it is, we can't even think about measuring it - or so it seems to me - so it's not possible, strictly speaking, to set about maximising it. That doesn't entail that well-being talk is totally useless. In a particular context we do know roughly what is meant. Person A may have a slightly different concept of well-being from person B, but they may have enough in common to communicate and reach agreement. These contexts are ubiquitous, so I'm not trying to stamp out well-being talk.

The upshot is that, as I said somewhere else above, I actually think that this is the least of Sam's problems - though still a real one. Sam knows all this, I think, but he doesn't consistently defend his position in this way. Sometimes, as in his discussions of Parfit and Nozick he writes as if there really is something identifiable and specific that can be measured and compared.

I think he can abandon that idea without doing too much damage to his system.

Russell Blackford said...

Brian, yes there are objective facts about brain states. But since I've never denied this, that doesn't impress me much. :)

The argument was never about that, and you can't settle the actual argument by defining terms in the way you want (even if you do so with rhetorical flourishes).

More generally, it seems that you're supporting preference utilitarianism, but that isn't Sam's position. He wants to maximise something called "well-being", not preference satisfaction.

However, if you want to support preference utilitarianism you should go back and have a look at the "Why Be Moral?" chapter of Singer's Practical Ethics. This chapter is very good and should be a touchstone in any discussion of objective prescriptivity. Singer basically concludes that we are not objectively required to be "moral" as he defines it. Rather, it's a contingent matter of psychology whether or not we will, as individuals, find acting in this way satisfying. He does urge us to act like preference utilitarians, and paints a picture of its satisfactions, but he concedes that this is not going to be satisfying for everyone.

Singer explicitly doesn't think that people who refuse to act like preference utilitarians are "just wrong" or somehow irrational. I think that Singer is essentially right about this. You can have a moral naturalism like Singer's brand of preference utilitarianism, and you may be able to get objectively correct answers as to what counts as "moral" by the definition used in your theory. But what you won't get (according to me and Singer) will be objective prescriptivity.

In one place you seem to be denying that Sam believes in objective prescriptivity, in which case my objections miss their mark. It's still possible that you're right on that, I suppose, in which case Sam and I basically agree after all. But it's not as if my review, to which he responded, didn't raise the issue at length.

Brian said...

"More generally, it seems that you're supporting preference utilitarianism, but that isn't Sam's position. He wants to maximise something called "well-being", not preference satisfaction."

This is true. Harris hasn't adequately delineated what he means by "well-being", but has demarcated limits to what he means by it. I think that a form of preference utilitarianism in which (roughly speaking, with caveats such as that all beings who will exist also count) fulfilling desires constitutes creating well-being fits (awkwardly) within his system. What I think is true is on the borders of what Harris thinks is possible; it's a definition for well-being that he seems to have not excluded, though his arguments are clearly intended to house a different sort of animal.

"Singer basically concludes that we are not objectively required to be "moral" as he defines it. Rather, it's a contingent matter of psychology whether or not we will, as individuals, find acting in this way satisfying."

Even two year olds know not to expect that the right action will always be the most satisfying or fun one.

"But what you won't get (according to me and Singer) will be objective prescriptivity."

Prescriptivity? Intrinsic values don't exist, so I wouldn't try to claim they did. Are you claiming that "morality" is inappropriately used if not intending intrinsic value? Values do exist and are reasons to act, just not intrinsic ones.

"I doubt that there is a metric for well-being...even if it existed, would not be binding on us in such a way that we are required (perhaps as matter of rationality) to maximise the sum of well-being."

I agree and don't think either is necessary for "objective morality" to be the best two word description of a certain collection of true facts.

I sometimes play chess according to the official rules as I know them, which is also what I did when I was seven. However, when I was seven and playing my brother, neither of us knew about not being able to castle out of check, taking pawns en passant, etc. We weren't truly playing official chess, but we were doing our best in our ignorance. Furthermore, we believed certain moves with the knight (or king, moving into check) to be legal when they weren't because we didn't envision and implement their movement well.

If I am losing a match, I am capable of taking my opponent's queen off the board and using my sledgehammer to obliterate her. I am also capable of using the hammer against my opponent directly. However, to do so would be to cease playing chess according to the rules as I know them (and as they actually are).

Sometimes I play a game called "Act to satisfy desires fairly" (ASDF) (An easy acronym to type!). If I thought that aliens would immediately torture five billion humans to death if I didn't go outside naked and make a snow angel, and would leave us alone if I did, the only act that would be ASDF compatible would be naked snow-angeling, just as the only act that is chess compatible when in check is to get out of check.

There are no such aliens (as far as I know), so I can play ASDF (as far as I know) by staying inside for now.

Brian said...

As far as I can tell, many others around the world play ASDF frequently, but much as I played chess when seven, they don't know or execute rules well. Some believe there is an alien ghost who is also a Jewish zombie who is actually his own father but only has one identity who tortures people not just to death but *forever* if they don't believe in him, so it's really important to tell everyone about it so they don't get tortured on account of a woman who ate a fruit when a snake told her to. Others have never heard of Euthyphro's dilemma and believe in divine command morality. Others believe in dualism, others have expansive notions of free will...there are a lot of false beliefs that interfere with actually playing ASDF, even when these people want to, rather than an approximation of it.

Taking into account what people believe to be facts and reverse engineering what they call "acting morally", one learns that they think they are playing ASDF. They call ASDF “objective morality”. Often it's in people's interest according to a subset of nearly all their values not to try and play ASDF, and they sometimes stop playing, and sometimes continue playing. Sometimes they try and limit the extent of their deviation from the rules as they understand them.

A few rules of ASDF as I understand the game are notable. It's almost certainly fair to give some priority to one's own desires because one best knows them and others are failing to prioritize them (and in fact prioritize their own desires) or prioritize desires without regard for who has those desires. Another is that it's not fair to only have regard for one's own desires. ASDF sometimes prescribes different actions for different people in identical circumstances.

I don't see how there can really be multiple rule sets. Having one prayer book for holidays and another for weekdays is having one liturgy.

One feature of centering morality on desire satisfaction is that we can abandon a search for a metric. It's obvious (those are dangerous words for me to type) that we are in principle capable of adjudicating fairly among competing desires of our own when we are the only person whose desires we consider. Furthermore, some people have some desire to fulfill others' desires per se. Somewhere short of The Gift of the Magi there is a theoretical entity that could, with infinite knowledge, tell us which act of ours in a given situation is uniquely compliant with the real rules of ASDF. It could do so without quantifying anything, just as we do not quantify among our competing desires.

Another feature is that desires are good reasons to act. What's more, they actually exist, unlike "intrinsic value".

Another feature is that "ought" statements are revealed as a type of "is" statement. Under facts as they are, if you want to fulfill certain desires to certain extents, then you can only do certain acts. "Ought" means "If your action is to be legal under the rules of ASDF, then it must be".

There is considerable disagreement about fairness, but I think most of it is errors that game theory can replace.

March Hare said...

Brian:
Please demonstrate that some values, any values, contradict each other.

People massively value stability. Yet they also massively value growth.

People value high fat/sugar foods.
People value being slim.

People value great public services.
People value low taxes.

People value being looked after.
People value not having a big-brother government.

etc. etc.

Is what people have as particular values eternally opaque to science?

Far from it, but science is a long, long way from being able to work out the future which is what I was getting at. Also, there is no metric to show whether it is better (choose any metric you like) to objectively show whether getting $100 this year is better than getting $110 next year.

Therefore one doesn't exist? You know better than to imply that.

Not only doesn't one exist, it is impossible that one could.

Of course, if you have one I'm all ears (eyes).

J. Abraham said...

Russell, Sam Harris has not tied himself to the naive position that well-being is described by a real number, or even some additive quantity. It would be quite silly to believe that vanilla is only as good as chocolate if they generate quantities of well-being which are equal in all decimal places. Actually, a "category model" of well-being is quite compatible with Harris' expressed philosophy. "Moral solutions" would only lie in the same category if there are no objective criteria allowing us to determine which solution is better.

What's important is that in many cases there are objective criteria. If we're going to explain why it isn't a good idea to drink and drive, there's no dearth of empirically sound and logically well-motivated criteria to which we can appeal. Our argument against drunk driving would be just as "objective" as anything in science. Probably there are more foundational considerations which this analysis glosses over. These simply aren't important, any more than the problem of induction is important in a conference on physics.

To put philosophical waffle aside and speak candidly now, the truth of this affair is that the Grim Reaper of moral nihilism always looms large. He will surely descend upon us unless we can find some cogent scheme which furnishes a "for-all-practical-purposes" way of answering the increasingly daunting moral questions. Vague and demonstrably false happy talk, about how we will all get by fine without an objective system of morality, is not going to help anything.

Brian said...

"People value high fat/sugar foods.
People value being slim."


Person A only values high fat/sugar foods. A gets hungry and finds some junk food and eats it. Desire fulfillment! Then A can't find any for A's next meal and is forced to eat a peanut butter sandwich. Mild desire frustration. Repeat every day. Over the years, A gets fat but does not care.

Person B only values being slim. B gets hungry and finds some roasted vegetables with a bit of chicken in it. B eats it. Then B can't find anything really healthy for B's next meal and is forced to eat a peanut butter sandwich. Repeat every day. Over the years, B is basically thin. Desire fulfillment (largely)! Mild frustration in being a bit over B's ideal weight (because of the PB sandwiches).

Person C values eating high fat/sugar foods and being slim. Person C gets hungry and finds junk food and roasted vegetables with chicken. Person C eats the junk food, desire fulfilled! Someone else takes the healthy food so the next meal is a PB sandwich, desire mildly thwarted. Repeat every day. Over the years, C gets fat and hates it. Desire thwarted!

Person D values eating high fat/sugar foods and being slim. Person D gets hungry and finds junk food and roasted vegetables with chicken. Person D eats the healthy food, desire mildly thwarted. Someone else eats the junk food and D is left with a PB sandwich next meal, desire mildly thwarted. Repeat every day. Over the years, D is basically thin. Desire fulfillment (largely)! Mild frustration in being a bit over D's ideal weight (because of the PB sandwiches).

Every individual has expected and appropriate emotional responses to each of his or her actual desires. There is no contradiction.

"Also, there is no metric to show whether it is better (choose any metric you like) to objectively show whether getting $100 this year is better than getting $110 next year."

Would you bet your family's life that no solution to this problem will be accepted among philosophers within the next 20 years?

Please share with us your proofs that these things are impossible, I'm guessing that they're some sort of proof from contradiction that's eluding me? Maybe argument from relevant authorities, since you give no details? Otherwise let us know how much time you spent racking your brain for solutions. A simple declaration that something is impossible is not helpful.

March Hare said...

Brian,
Every individual has expected and appropriate emotional responses to each of his or her actual desires. There is no contradiction.

But they are in conflict. And if the person does not know that they are in conflict then there is a contradiction, which is the point I originally made about cognitive dissonance.

Would you bet your family's life that no solution to this problem will be accepted among philosophers within the next 20 years?

Please share with us your proofs that these things are impossible, I'm guessing that they're some sort of proof from contradiction that's eluding me?


I hardly think that the general acceptance of some philosophers on an issue makes it so. Although 20 years does appear to be incredibly short given the pace of progress in that field...

The 'proof', such as it is, is that person A when presented with the choice of $100 now or $110 next year has to have some rational discount rate (assuming all immediate needs are met) and that rate will be based on some factor of their brain now. In a year's time when a test is done to determine whether the person has benefited from their choice (whichever it may be) they are a different person. Their desires and values have changed, which may or may not have to do with their decision. Either way, you are comparing the well-being of person A against the well-being of person A+ at a later time.

To expand on this we simply have to look at the current economy, is it right that this generation was living large on the back of debt that will have to be repayed by future generations? Since some of the benefits that the debt-fuelled spending will ultimately benefit future generations how do we measure current super-well-being against future reduced well-being?

Basically, I come from an economics background and we simply give the expected results of any given choice but refrain (in the main) from making recommendations. e.g. I can tell a factory owner the likely impacts on risk, profits and growth of investing in new machinery, but whether he should is determined by his relative weightings of these values.

So why can't we just measure people's discount rates, after all it's simply a function of their brain? True, but it measures current brains not future ones. It also exists within a very suggestible part of the brain and we can hugely alter our decisions based on how things are phrased - i.e. we are not rational when it comes to these things. If we are not rational then I don't see hwo we can be objective.

March Hare said...

Brian, if you don't like the discount rate then how about the choice between maximising well-being, minimising suffering (not the same), maximising the minimum well-being, maximising the median well-being.

These are all valid aims that appear to be brushed aside by Sam Harris, but in a land of moral truths it must be possible to choose between them objectively. Care to tell me how it is possible to do so?

Sam may well say that these all represent different peaks on his moral landscape, but so what? We should still be able to work out which is higher and that would make it more moral (in his framework.)

J. Abraham said...

March Hare, have you heard of an interest calculator? In plenty of cases applicable to real life, the "interest metric" is enough to determine whether $100 now is better than $110 next year.

Alternatively, if you're living in a failed state in Africa and haven't a penny to your name, the metric of "common sense" might be sufficient. I think all of us save crackpots can agree that the extra $10 just isn't worth a year's starvation.

Thanny said...

What I'm saying, regarding objective scientific measurements, is that they are rarely cut and dried.

I chose the star example because there is no way to actually measure the distance - you can't roll out a tape measure and read off a number. All measurements to distant stars rely on a model, into which we plug objectively measured values. Those models - ranging from simple trigonometry to best guesses about intrinsic brightness - are the best we can do, and we are still, today, revising the "correct" answers (even for the simplest of them, based on something as solid as the Pythagorean theorem).

The point is, most metrics are fuzzy and indirect, even the ones about uncontroversial facts.

Anyone looking for a simple measure of something as vastly complex as well-being is being entirely unreasonable.

I already gave examples of what numbers we have today to estimate the well-being of whole societies. Some of them are easily extrapolated to metrics for the well-being of an individual person. Income, friends, family, health - all these things affect a person's well-being, and none of us has any difficulty deciding when a person we know is doing well or not well based on that kind of information.

Really, we're surrounded by metrics of well-being, both for groups of people and individuals, but we haven't pulled them together into a framework where they can be readily quantified and interpreted by an outsider. This is something that can be done, but won't be done in a blog comment.

But it's clear that anyone bemoaning a lack of a metric for well-being just isn't thinking clearly about the question.

Tom Clark said...

J. Abraham:

"Vague and demonstrably false happy talk, about how we will all get by fine without an objective system of morality, is not going to help anything."

Does the following, from another thread on this blog, help at all?

Moral realist intuitions may well be part of our genetically transmitted moral sense: it strongly seems to us that some things are *just wrong*. And in fact there's widespread, culturally invariant agreement about the intrinsic wrongness of certain acts, such as torturing innocents and killing babies for one's own amusement. We wouldn't be here as a successful species without such aversions, and that they are nearly universal (sociopaths are the exception) is as much objectivity as we're going to get about morality.

Our shared, biologically transmitted moral norms about harm, fairness, and reciprocity set the ground rules for arguments about differences in cultural practices, since most individuals, even in the most regressive societies, operate according to such norms with respect to their peers. It's arguing that those norms be extended to *all* individuals whatever their status that the success of the liberal democratic project depends on.

Since our moral adversaries can't deny they share those norms with respect to their peers, they have to say why out-groups shouldn't be granted the same consideration. This usually involves some religious, political or economic ideology about human ends and social hierarchy as decreed by God, Hitler, Pol Pot, or other non-scientific authority. So part of the liberal democratic project has to include challenges to the flawed epistemology that supports such ideologies, and of course direct appeals to the moral intuitions about harm and fairness that we know exist in our opponents.

So as Russell says, arguments about morality and values aren’t irrational, unreasonable, useless, hopeless or arbitrary, even though we can’t, as Harris supposes, look directly to science to validate our side of the argument.

Russell Blackford said...

Everyone remember here that we're not just talking about the practical difficulties of measuring something. We all know that there are many things that it is difficult or impossible to measure with any accuracy, or can only be measured with theoretical assumptions made (though theories can be well-evidenced). We're talking about whether we even know what it is that we're supposed to measure. Is it preference satisfaction, is it certain feelings (pleasure? joy? ecstasy? abence of pain? which are not necessarily on the same scale), something already moralised like living a life of "virtue", having certain capabilities for action, actually living life according to a certain template, or some weighted mix of these, or what? Sam himself isn't sure, though when pressed he tends to talk about certain kinds of feelings (but not straight-up pleasure).

Again, it's not that we can't suggest one of these; it's that we don't currently have any agreement on which it is so there's no way we can know what we're trying to maximise. Or if it's a mixture, we don't have agreement on how to weight its components. Arguably, we actually value a number of these things, but different people weight them differently - and arguably don't even agree fully on which ones (e.g. I'm not interested in maximising something that is already moralised, like "virtue"; I also see no reason to accept any template of how a life should go, and I expect Sam would agree with me here).

And again, you can propose one of these things as what it is that we're trying to maximise. Perhaps you'll be able to state it in a way that doesn't involve the same problem at a new level (though I'd be sceptical). But even so, then you run into the point that Sean Carroll has been making on one of our threads and at his place: that your choice is not binding on someone else who'd make a different choice.

Again, I'm not denying that we do have a great deal of overlap in what we value and that "well-being" can be a useful way of talking about it in many real-life social contexts. It's only when we push very hard on the idea that we expose its theoretical limitations. If Sam doesn't push too hard on it, I'm actually happy for him to say, "Such and such a moral system is irrational (or some such word- 'irrational' may not be quite right) because it has lost touch with human well-being." Though there's some fuzziness in this sort of claim, as with many of our evaluations, it may be about right in many specific cases.

I will go on using fuzzy concepts such as "well-being", "harm", "flourishing", etc., myself. I'll be using such language in my own new book, as it's largely unavoidable in practice. As I said, I'm not opposed to using this language, even in philosophy. But I do note that it has limitations, and I don't think Sam is entirely consistent in acknowledging this and its implications for normative theory. At times, he seems to want to push the language beyond its limitations.

J. Abraham said...

Tom Clarke, I'm afraid it doesn't help, because close inspection reveals that you are reverting to happy talk. You seem to suppose that everything about life will be as right as rain if only we can get rid of false epistemology and the blasted tendency to create in-group/out-group dichotomies. In my opinion, moral questions get hugely more complicated than that, and I see no reason to prematurely throw in the towel. (If that is indeed what moral philosophers are doing, then it's hard to imagine how they stay busy.)

Also you repeat the fallacious statement that morality is non-arbitrary even if it is not objective. That is a contradiction: non-arbitrary means the SAME as objective.

Anonymous said...

Russell, I think you are basically mistaken when you say that a fuzzy measure is not an objective measure.

(And I'm not talking about measurement errors, but intrinsic fuzziness in the measure themselves.)

For example, in discussing genetics, we can talk about degrees of genetic relatedness, but there's no single right measure of genetic relatedness.

(Are you more interested in details of nucleotide sequences across all sequences, or more specifically in genes that are actually transcribed? Are you more interested in differences that have structural and function effects, e.g., mutations that affect how strongly transcribing a gene affects the transcription rates of other things? Are you more interested in important genes such as hox genes, changes to which are likely to be especially significant for broader-scale issues?)

There simply isn't a single metric definition of genetic relatedness, but it's even worse than that---the best definition for a certain purpose is context-sensitive. For example, it may matter differently for different purposes whether you're talking about mitochondrial DNA vs. nuclear DNA, or sex chromosome DNA, and so on. There many dimensions in which "genetic relatedness" is poorly defined.

Nonetheless that does not mean that geneticist can't or don't quite often say objectively true things about degrees of relatedness. They do it every day.

For example, under any measure of genetic relatedness, I'm more closely related to you than to my dog, and my dog is more closely related to me than to my goldfish, and all of us are far more genetically related to each other than to a petunia, and even more remotely related to a bacterium.

Those are objective facts. They're just flat true. Facts don't get much more objective than that.

The intrinsic, multidimensional fuzz in "genetic relatedness" simply does not prevent there from being vast numbers of objectively true and objectively false statements about degrees of relatedness.

It also means that there are vast numbers of statements that are ambiguous. (And that's objectively true, too.)

For example, I may be more genetically related to one cousin than another, even though they have the same parents. There may not be an unambiguous, objective truth as to which is a closer relative---I may share more genes with one cousin, but more important genes with another, etc.

Most scientific concepts about interesting things are like that to some extent. It doesn't necessarily make statements using those concepts subjective, or "not objective"---and it usually doesn't.

Some things are objectively "close enough for rock and roll," and are either true or false, and other things are objectively not close enough for rock and roll, and require more refined and specific concepts for statements to have a clear truth values.

I think Harris is quite right to take the same scientific approach to morality, though it's a very hairy project.

You're applying a higher standard of "objectivity"---and, I think, a mistaken one---to morality than we apply to everyday scientific concepts.

If Harris has failed in his project, that's not where he has failed. He is right to think that there can be objective measures that are also fuzzy, in any number of dimensions.

Scientists use such concepts every day---and for the most part, in objectively correct ways.

You seem to want something out of the word "objective" that just isn't there.

Paul W.

Russell Blackford said...

Sorry, anonymous, but you're not understanding the criticism. What you are actually stating is more or less my position, and the one that I want Sam to take.

I am arguing something very like what you are arguing. Yes, there are things around in the vicinity that can be measured objectively (though it's going to be harder when some of those things are themselves fuzzy). Yes, there will be some cases where all the measurements that anyone considers relevant point in the same direction. These will be clear-cut cases. But we can't just maximise a variable and expect that, apart from practical problems, there will always be an unequivocal answer to its size. Moral evaluation is not like that. (We don't really need the word "objectively" here, the word often seems to cause more confusion than it helps).

There will also be cases that are not merely
(1) A tie

or

(2) too hard to measure in practice


But

(3) Open to legitimate disagreement because there is no further truth as to which measure we use. This gives us indeterminate, ambiguous answers even in principle. Our evaluations of things are typically like that.

Now, you say that science acknowledges situations like (3) all the time. Okay, perhaps it does. Therefore you can say that the study of morality could be "scientific", if hairy, even if contains situations like (3).

Well, I agree it's likely to be hairy, because such situations are likely to be more typical than not. Could it still be scientific? Perhaps. But as I said in the post that is not my criticism. Generally speaking, when cases like (3) are ubiquitous and very important we are into areas studied by the humanities rather than the sciences, but I don't think there's a clear dividing line.

My point isn't that the practice of criticising laws and customs, etc., cannot be scientific in some sense or other. It's that Sam writes in places as if he acknowledges (1) and (2) type situations but doesn't realise that there will be (3) type situations. He certainly never makes a big thing of it. In fact the book scarcely mentions it, if at all. Even in his reply he seems to suggest that they will be very marginal cases.

My point is that when you have (3) type situations, especially in an area like, say, literary evaluation - or evaluating anything else beyond the simplest designed objects - where these situations are ubiquitous and important, the idea of maximising a quantity becomes very problematic. We end up not knowing what we are trying to maximise. Yes, we can talk in all the more complex ways that you mention - though it will be much more slippery even than you describe with genetics - but we won't be able to talk like utilitarians who talk about maximising some value.

So Sam can have his "science" or at least a practice informed by science (and scientific in a broad sense if you like). But he needs to take this stuff into consideration. If he takes it seriously, the book will change, or the next book will be different. He will sound a lot less like a utilitarian.

The question of whether morality is "objective" in the metaethical sense is a largely separate one not dealt with in this post.

Brian said...

"We're talking about whether we even know what it is that we're supposed to measure."

We're talking about whether there is something we are supposed to measure, so much the better if we can know that there is and even better if we can know what it is. Those last two aren't strictly required-there could be something we are supposed to maximize that we don't know to, or it could be unprovable either way whether objective morality based on maximization fully describes our world even if it is true.

"Is it preference satisfaction, is it certain feelings...living a life of "virtue"...or what?"

One of these things is not like the others.

One way of considering the project is that we are guessing which physical pattern (a type of action, type of effect, type of life lived) possesses the magical quality of intrinsic value that only interacts with the material world by infusing patterns of some kind with "good" so they are objectively moral. It is not contradictory to speak of preferences this way-it is as (un)likely that fulfillment of desires has this pixie dust as it is that honesty or pleasure do. Considered in this light, it seems arbitrary to select any specific property or amalgam of properties to weigh and maximize.

However, even without magic, desires are legitimate reasons to act. So I don't say that "preference satisfaction" doesn't belong in the above list-it does-but it is also and instead a thing unto itself.

"...your choice is not binding on someone else who'd make a different choice."

What does this mean?

"There will also be cases that are...[o]pen to legitimate disagreement because there is no further truth as to which measure we use. This gives us indeterminate, ambiguous answers even in principle."

If there are, within a range, only ambiguous answers, how is there room for legitimate disagreement? Is this importantly different than a case in which there are quantified but equal answers, and individuals have divergent preferences? What is the statement that person A can truthfully say about the dilemma between indeterminately equal X and Y that person B can disagree with?
A: There is no way to determine which activity is more moral.
B: I agree.
A: I prefer action X over action Y.
B: You sure do.
A: We should do X instead of Z, since Z, though also of indeterminate value, is demonstrably of lower value.
B: Yup.
A: We should do X instead of Y.
B: That's not a true statement. "Should" means that X is more moral, but it demonstrably isn't in this case.
A: We should do X.
B: That doesn't mean anything. X instead of Y, or X instead of Z?

Whether there is no theoretical way to decide between two mutually exclusive activities because they are either of demonstrated equal value or demonstrably of the same order of magnitude and indeterminate, there are facts of the matter, and human activities change the facts on the ground. If there is no way for a couple to choose between getting a cat and a dog, if one person wakes up early and builds a doghouse from lumber, the choice of animal can become objectively decidable as right or wrong from that investment.

Brian said...

"But they are in conflict. And if the person does not know that they are in conflict then there is a contradiction"

It is impossible to fully fulfill all desires at once. If the person does not know this, how is that importantly different than if they don't know any other fact, for purposes of discussing well-being?

I hardly think that the general acceptance of some philosophers on an issue makes it so.

That's obvious. However, one cannot directly observe the impossibility of something. It would make little sense to test your confidence that it is impossible by asking you to bet against the possibility that it is possible if you could simply reject it being shown to be possible.

I picked consensus among philosophers because that is not likely even upon something being adequately proven true (to the satisfaction of all who live fifty years after it is proven, even without any additional evidence beyond the original proof).

"Either way, you are comparing the well-being of person A against the well-being of person A+ at a later time."

So your claim is that in addition to having to show itself capable of adjudicating between A and B, a moral theory like Harris' has the additional and more difficult burden of adjudicating between A and A+?

"So why can't we just measure people's discount rates, after all it's simply a function of their brain?"

Because A and A+ are different people, just as A and B are. We can't refrain from giving 100 units of well being to B and instead give 2 units of well being to A simply because A only gets 1 unit of well being from knowing B gets 100 units of well being.

"It also exists within a very suggestible part of the brain and we can hugely alter our decisions based on how things are phrased - i.e. we are not rational when it comes to these things."

Most people, A do not follow the Golden Rule/play ASDF all the time. Their greatest bias is to favor A over all others (A+, B, B+...N, N+), their second greatest is often to favor A over A+, and A+ over B/B+. Telling A of a dilemma in dividing benefits between A and A+ where those two people have different utility functions, different phrasing makes A more biased for himself and more deviant from ASDF/The Golden Rule than another phrasing would. That does not mean that there is not a fair allocation, considering all desires.

"...maximising well-being, minimising suffering (not the same), maximising the minimum well-being, maximising the median well-being...brushed aside by Sam Harris, but in a land of moral truths it must be possible to choose between them objectively...how...?"

By abandoning the hunt for intrinsic value and accepting that actual values beings have are good reasons to act.

It is not good that desires be fulfilled unless a being wants desires to be fulfilled. There is no inherent value in maximizing the number of fulfilled desires, nor minimizing the number of desires thwarted, nor balancing those things, except for as actual beings desire. It happens to be that beings desire a number of things, which is good reason to fulfill those.

As you pointed out, even within one real person it is impossible to maximize fulfillment of all his desires even in principle. Nonetheless, we do not call arbitrary a person's choice of chocolate ice cream over cookie dough ice cream when he greatly desires the taste of chocolate, slightly desires the texture of cookie dough, and greatly desires to avoid the stomach ache he would have if he ate both. It is the correctness of such judgements that I think can be extrapolated from within one person to among several people.