Rationality, morals and Jonathan Haidt

Rationality, as it turns out, far from being the defining quality of humans, is very often used by humans only post hoc, to justify decisions we have already made based on unconscious, irrational processes.  In fact, rationality may actually have evolved, not so much to help us make effective plans in the physical world, as to help us out socially, by “explaining” our decisions to our peers.  This, at any rate, is one of the key lessons from Jonathan Haidt’s book The Righteous Mind: Why Good People are Divided by Politics and Religion.  At least, that is a main “take-away” for me, which fits in well with some of the other reading I’ve been doing over the past couple of years, including Gerd Gigerenzer’s Rationality for Mortals: How People Cope with Uncertainty, and Hilary Kornblith, et al, Naturalizing Epistemology.  Although, as a person who makes my living as an engineer, I may have more respect than psychologist Haidt for the role of rationality in helping us deal with our non-social reality. 

This particular argument about rationality, Haidt documents well with experimental data (although the  Gigerenzer book offers a caution in its chapter specifically on how statistical analysis is misused in most published psychology research).  Some of Haidt’s other themes are also very interesting, but raise more issues for me; at least, I have more questions about them. 

Haidt’s main interest, in the book is with moral psychology, including how moral systems actually function in our (conscious and unconscious minds), and how and why we evolved moral systems in the first place.  (The evolution of moral systems by natural selection is counter-intuitive to a naïve, “red in tooth and claw” view of nature and “survival of the fittest”.)  Haidt’s scientific work is mostly descriptive; i.e., he is trying to find out how moral psychology actually works, not how it ought to work.  His main foray into the prescriptive, in this part of the book, is to advise us that we should try to understand how our minds (including our moral systems) actually work, and not shrink away from the knowledge, even if it makes us uncomfortable.  This is a prescription I heartily endorse.  Later, he veers into some prescriptions I think less of, as I discuss below.

A main element in his taxonomy of moral systems is that they center around a number of separate “modules”, which function unconsciously, and provide the automatic moral reactions that our reason later tries to justify.  He identifies six of these: Care/harm, Liberty/oppression, Fairness/cheating, Loyalty/betrayal, Authority/subversion, and Sanctity/degradation.  He defines them all, and suggests reasons why it might have been beneficial for our ancestors, long ago, to evolve such centers.  He and his colleagues identified these modules experimentally (by questionnaires, interviews, and laboratory experiments).  When he gets around to discussing politics, he says liberals care mostly about the care/harm module, a little less about liberty/oppression and fairness/cheating, and not much at all about the other three.  Conservatives, on the other hand, care equally about all six, although they interpret some of them (care/harm and liberty/oppression) somewhat differently.  This makes it possible for conservatives to appeal to more people than liberals*, since they can reach them on more levels.  It can also make it very hard for one group to understand the other.  For instance, in interpreting some family law situation, liberals may focus strictly on the wellbeing of the child (care/harm), which conservatives care about, too, but conservatives also care about parental rights (authority/subversion), state interference (liberty/oppression), and perhaps others; liberals tend to think these other things are irrelevant.  He also says that his research has found that conservatives are better at predicting what liberals will think about a situation than the other way around, because the conservatives share the liberals’ concerns, but add to them, while the liberals don’t see the conservatives’ other concerns, at all.

Haidt’s assumption of six discrete modules bugs me.  For instance, it is pretty well established that people tend to accept traditional authority as long as it is associated with responsibility, i.e., as long as the authority figure carries out a traditionally determined social role seen as having benefits for the community as a whole.  It is not necessary for there to be any particular proportionality (let alone equality) in the distribution of benefits between the authority and the people.  But if the traditional responsibilities are seen as being ignored, then people are more willing to rebel against the authority.  (See, for example, Barrington-Moore, Injustice: The Social Bases of Obedience and Revolt.)  So, is there a seventh module, responsibility/irresponsibility?  Or is this subsumed under loyalty/betrayal, or fairness/cheating?  Or both?  Or all three?  For that matter, couldn’t loyalty/betrayal just be a special case of fairness/cheating?  And are liberty/oppression and authority/subversion really cognitively independent?  At the very best, there seems to be a lot of overlap between these categories.

Ultimately answering these questions may require a detailed mapping of the moral modules on the brain/somatic structures that implement or create them.  If this were done, then (for no particular reason but my gut feeling), I conjecture that:

1. There would be more than six modules, and,
2. Boundaries between modules would be fuzzy.  There would be physical overlap between the structures implementing different modules, with some sub-structures involved in two, three, many different modules.

I suspect that it is difficult to structure the type of research Haidt was doing so as to be free of question-begging.  You have to make decisions, consciously or unconsciously, as to what types of traits you are going to look for, and you are more likely to find the traits you identified than to stumble across ones you never thought of (although Haidt does describe how discrepancies in early results led him to expand his taxonomy from an original five modules to the final six).

A significant amount of Haidt’s book is spent discussing group selection and its role in evolution.  Many of his observations on this topic I have come across, elsewhere.  One that I have not is his belief that there can be, and has been, significant pro-social genetic evolution in historic times (or at least from late prehistoric, but definitely human, times).  Essentially, he believes that culturally determined group behaviors (such as religious practices) can create selective pressures that weed out individuals less willing to conform to group norms.  This he sees as a form of actual group selection, because groups whose members were selected to make them cohere more closely,  passed on more genes than groups who did not. 

I have no doubt that some kind of group selection for moral, cooperative behavior did occur (see previous blog posts, linked to at the end of this one).  I’m inclined to be dubious about Haidt’s very late time frame, though. He justifies it by comparison with intentional selection (by breeders) of domestic species, which can cause significant genetic change in a small number of generations.  But I think selective pressures in a rich, cultural environment are much more diffuse than in a controlled breeding situation.  He does give a nod to the time-frame difference, and admits that the genetic response he envisions would happen much more slowly than in domestic breeding.  But if the cultural pressures are sufficiently diffuse, they will not outweigh chance, and selection will not occur, at all.  I think there are just too many ways to escape conforming, in most social situations.  I think the level of late, culturally induced genetic change that Haidt envisions is unlikely.

Haidt’s ideas on group selection are intimately connected with his ideas on religion, social organization, generally, and politics.  It seemed to me that once he entered these areas, his scientist’s insistence on evidence tended to weaken.  He becomes much more willing to assert things without offering any (or at least, much) proof.

Haidt clear wants to believe that religion is a good thing.  He is critical of the so-called New Atheists and their tendency to see religious belief as the root of all evil.  I tend to be critical of the New Atheists, also, but I think the social utility of religion is much less than Haidt thinks it is.  Haidt makes a strong case for the utility of religion early in human history (or prehistory) as something that helped improve the coherence of groups.  I think group cohesion, partly helped by shared religion, almost certainly helped human groups to survive in early times, and to outperform (and perhaps out-breed) groups that were less cohesive.  (Dennet, Dawkins, and other New Atheist use these facts in constructing a theory of meme-level selection, which Haidt thinks feeds back into genetic-level change.)

Haidt makes a much less compelling case for the value of close group bonding in modern times, in a globalized world, with awesome weaponry of mass destruction.  He makes a claim that having a vast network of small, closely bonded groups (including religious ones), makes, in-and-of-itself, for a safer and more stable overall society.  He offers essentially no evidence for this claim.  In the interest of convincing us, he over-emphasizes the positives, and downplays the negatives.  By implication, we should think of groups like U.U. Social Action Committees, or the League of Women Voters, and not Daesh, Westboro Baptist Church, or the KKK.

He discusses some research on the survival of small insular groups, which he believes supports his position.  Groups like communes and religious communities are found to survive longer (as groups) if, (1) they demand costly sacrifices of their members, and (2) they “sacralize” the sacrifices (i.e., refer them to God’s will).  Surely, it is not obvious that this is a net good, at least for the members of the group.  What actual benefits does their membership in the group provide that outweigh what they are asked to give up?  Haidt doesn’t even raise that question, let alone try to answer it.  He seems to take for granted that survival of the group is a sufficient value, in-and-of itself.  It seems to me, on the other hand, that the members might well be better off if the group dissolved, and they were re-absorbed back into the mainstream.  (Survival of the group is not a precondition for survival of the members, after all.)  The study Haidt cites seems to support Dawkins’s idea of religion as a parasitic meme at least as well, if not better, than Haidt’s idea of religion as a positive force.  Another comparison that comes to mind is with drug addiction (i.e., a community of willing addicts).

Thinking of group survival as a “good” makes me think, re. morality, in general:  is survival, pure and simple, a sufficient justificatory basis for morality?  Or do other things matter?  How about knowledge and truth for their own sakes, and not just instrumentally?

Because, when it comes to religion, Haidt does not seem to be a believer.  The impression I got is that he is an atheist.  But religion, as a social phenomenon, appeals to him deeply, and he wants to believe it is good, and should be promoted.  But if religious beliefs are all wrong about matters of fact, does its appeal (to some people), and its (supposed) social utility justify our supporting and encouraging it?  This seems to skirt perilously close to Plato’s Noble Lie.  There is something elitist (and ignoble), it seems to me, about promoting beliefs you personally think are false, because of their social utility.  (Tolerating them may be another matter.)

When Haidt turns to political analysis, he leaves the scientific method even further in the dust.  He wants to prove that each of the different political constellations (which he identifies as “liberal”, “libertarian”, and “conservative”) has its own share of truth; therefore, we should learn to broaden our moral views, so that we can effectively listen to each other.  He gives examples.  For libertarianism, he gives a story entitled “Markets Are Miraculous”.  And a “story” is exactly what it is:  an entirely fictional, made-up scenario about an imaginary insurance scheme that proves exactly nothing.

Haidt’s approach to politics invites comparison with George Lakoff.  Lakoff has also been very concerned that people improve their understanding of the latest scientific theories of how cognition functions, and apply them to political work, and, like Haidt, he believes that conservatives tend to do a better job of this, here and now, than liberals do, and this concerns him.  But Lakoff unabashedly sides with liberals, and his interest is clearly that liberals should learn to better apply science in support of liberal moral views (although I know a lot of people who fail to understand that this is what he is saying).  Haidt, on the other hand, strikes me as a former liberal who has become more conservative with age.  He wants liberals to compromise on their principles, and he offers some fairly lame arguments for why they should do so.

Grains of salt are required.  But the insights into the psychological structure or rationality and morality that Haidt’s work provides make his book worth the reading.

*I use the word “liberal” in this essay in the conventional, modern American way, because that is the way Haidt uses it, and not in the classical way still more in favor, I think, in Europe, which is closer to what Americans call “libertarian”.

Related posts:

http://persistentwondering.blogspot.com/2009/08/natural-history-of-morals.html

http://persistentwondering.blogspot.com/2009/12/altruism-and-human-evolution.html

http://persistentwondering.blogspot.com/2012/09/irrational-atheists.html 

Pondering Truth

When we speak of “truth” do we describe some single entity or quality in the world?  Or is “truth” hypostasized to simplify a complex of relationships between the inner and outer worlds that we do not (cannot?) understand?

Is “truth” an operational concept?  There is no operational difference “now” between a justified belief and a justified true belief (a.k.a. “knowledge”, at least in many philosophers’ systems).  I make the same choices, take the same actions either way.  But there may be an effect on the outcome (or not).  Is “truth” operational in evaluating outcomes, and therefore, perhaps improving the reasons for future beliefs?

Even in evaluating the outcome of past choices, I still don’t “possess truth”.  I only formulate, and attempt to justify, further beliefs (e.g., beliefs about my past beliefs).  In Dewey’s terminology from Experience and Nature, truth does not seem to be something we can “have directly”.

There is a state of the world, and there is a state of my mind (or yours), which is, itself, part of the state of the world.  The state of my mind includes a simplified, impressionistic “image” (in some neural/synaptic medium) of the state of the world – to the extent it is accessible to my imagination.  In that image, the state of my mind figures under such rubrics as “reasons” and “beliefs”.  Based on my state of mind (at any given time), I will make certain choices, and take certain actions.  Partly because of such actions, at some future time the state of the world, and the state of my mind, will be different.  By comparing the current state of my mind (and especially its world-image) to my memory of its prior state, I make judgments about the truth of my previously held beliefs.  I formulate (reasoned) beliefs about my prior beliefs.

Is “truth”, then, just something we invent to explain our relative satisfaction or dissatisfaction with the outcomes of our past endeavors?  The trouble with that idea is that our feeling of satisfaction may be connected to aspects of the outcome which some mythical unbiased observer would be unwilling to call “truth”.  For instance, a racist may desire to join an organization of like-minded individuals.  His belief that African Americans are genetically inferior to “Aryans” may help him in that endeavor, and thereby lead him to be satisfied, but we would not want to call his belief “true” on such grounds.  Then there is the growing body of psychological research indicating that our brains have evolved so as to readily adopt certain beliefs which are adaptive, but not (necessarily) “true”.  An example is the “confirmation bias”, where by our minds tend to disproportionally accept data that confirms what we want to believe, and reject data that disconfirms it.  Another example is the “hyperactive agent detection” that Daniel Dennett discusses in Breaking the Spell (among other places).  Other examples can be found in various papers in Naturalizing Epistemology (Hilary Kornblith, ed.)

It seems there is something actual about the relationship between our mental states and the world state that we are trying to capture with the concept “truth”, which is related to, but not simply reducible to, our degree of satisfaction with outcomes.  I admit, at this point in my life, to be still wondering (persistently) about exactly what it is.

Postscript on Confirmation.  As with many of the musings on Persistent Wondering, this one pretty much starts “where I am” and doesn’t make much of an effort to relate to an audience that may not be starting at the same place.  I apologize for that… but after all, I am portraying myself as a “wonderer”, and making no claims to be a teacher.  (Lame excuse.)

In this essay in particular, though, it seems to me that many people may wonder why I would feel – at all – that “truth” is not directly accessible.  In many (most?) of our everyday interactions with the world, confirmation of our beliefs is direct and immediate, and seems incontrovertible.  I believe I left my keys on the kitchen counter.  I go downstairs – I either find them there, or I do not.

Other “facts” are not so easily confirmed or disconfirmed, though.  There are the challenges of philosophical skepticism.  How do I know I am not dreaming?  Or hallucinating?  Or a disembodied brain kept alive in a vat, with my neural inputs manipulated by alien scientists?  Then there is the question of “modeling”.  Complex physical or social systems cannot be grasped by our minds in their complete and detailed totality.  We need to abstract from them, simplify them, in order to understand them.  Do concepts like race, class, culture, the national income “truly” conform to some real world objects, and if so exactly what and how?  How do we indubitably confirm or disconfirm them?  Theoretical physics provides examples, also.  Do the objects of modern theoretical physics – quarks, bosons, photons – “really” exist, or are they just a convenient (not necessarily unique) way of mathematizing experimental results?  Are the relatively abstract and indirect confirmations of physics experiments really of the same class as our confirming (by looking) that our keys are on the counter?

But really, ALL of our knowledge involves some such modeling (abstraction and analysis).  All the objects we conceive involve some level of abstraction – focusing on certain aspects of experience and ignoring others.  Something of this is suggested by Heraclitus’s statement thousands of years ago that “You can never step into the same river twice.”  What exactly is a river?  Is it the specific water molecules?  But they start out in a glacier and end up in the ocean.  Is it the banks?  But they shift with time as soil particles are removed and deposited.  Is it some abstract (fractal?) pattern that encompasses changes over time?  What is the “truth” of the matter?

Our mental states (beliefs and so on) consist in synaptic patterns, roughly, stable-yet-changing patterns of chemical interactions between neurons.  The state of the world consists in the interplay of forces amongst patterned matter and energy, extending strongly or weakly between the various points of the entire universe.  It is not clear that some unique and transparent correspondence can be established between those two things and unambiguously labeled “truth”.  On a conceptual, theoretical level, the question of the truth of our beliefs, their confirmation or disconfirmation, is not at all a trivial one.  Although on the pragmatic level of day-to-day actions, it very often is.

Choice and Determination

If we choose, and our choosing is not pure, unanalyzable magic, then there must be some mechanism, however complicated, by which we make the choice.  Logically, it seems that, either this mechanism must be “deterministic” in the sense that if the mechanism is in exactly the same state, and receives exactly the same inputs*, then it will decide in exactly the same way; or there is some purely random component (e.g. quantum uncertainty) that affects the choice.

Neither of these options “sounds like” free will.  But if this is not free will, and free will is not simply a meaningless concept, or at least irrelevant to the choices we actual can and do make, then there must exist some entity whose choices are being constrained by this process.  That is, there must be some entity that is not “free”.   But if some entity is unfree, what entity is it?  If “you” are not exercising “free will” in these conditions, who are “you”?

We, the material, biologically constructed entities that we are, do make choices, and our choices are made in an effort to further the ends that we have conceived for ourselves.  This much is manifest.  It seems meaningless to me to say that we do not have free will in these choices based on some highly abstract argument that somehow we will never make the choices we don’t end up making.  What does it even mean for us to somehow be “able” to make the choice we do not, in fact, end up making?  I don’t see how that idea can be rendered coherent, except by reference to physical/mechanical potentialities, and epistemic possibilities that are fully compatible with the “deterministic” mechanisms of choice referred to above.    

In any case, other than these organic mechanisms, which make choices in the manner that we, in fact, make them, there is no entity here that can be “unfree”.

Of course, our choices are constrained, by the materials and opportunities available in our environment, by our abilities, physical and cognitive, by our histories, including things that have happened to us and previous choices we have made.  And of course the ends and goals that we conceive are formed by similar internal and external factors.  But for this to mean that our choices, from among the options that seem epistemically possible to us at the given moment, are “unfree”, there would have to be some self, other than the selves so constructed by history, biology, and prior choice, whose freedom was constrained, whose will was being denied, by the choices the biological “we” are making.  I cannot, for the life of me, conceive of what that other self might be.

I conclude that to say that people do NOT have free will makes no sense.  All that is left is to either accept that the way we actually choose represents free will, or to take the question of free will as an ultimately meaningless “pseudo-question”.  But at this point the choice is purely “semantic” in the colloquial sense of “just a matter of words”, and it makes no real, philosophical difference which wording you pick.

*The conditions of the mechanism being in “exactly the same” state and receiving “exactly the same” inputs  are considered only to focus on the logical argument.  They are both, in practical reasoning, impossible.  The universe is far too complex for it ever to be put back in the same state twice, and as for our “mechanism”, it is altered by every experience, and therefore can never be in the same state twice.  Even settling for “highly similar” is dubious.  Most significant human choices are, neurologically, a chain (actually a highly “parallel” web) of micro decisions, each of which potentially alters decisions further down the line, possibly in a binary manner (on/off – no shades of gray).  The end result, it seems to me, must show sensitive dependence on initial conditions, whether the component steps do, or not.  (They may.)  Thus even “similar” inputs to a brain in a “similar” state can lead to widely different outcomes.

Note:  I wrote most of the above text while I was reading Ch. 2 in Daniel C. Dennett’s book “Elbow Room: The Varieties of Free Will Worth Wanting”, a book which I have subsequently completed.  I think my views on this, as on many other topics, are broadly compatible with Prof. Dennett’s.