-> Reviews -> Pathological Altruism

Reasons Pathological Altruism caught my eye:

  1. provocative title
  2. name recognition (one of the editors, Barbara Oakley, created the very nice "Learning How To Learn" course)

It's a collection of 31 essays by various authors on the loose theme of altruism-gone-wrong.

This is a long review, you can skip around. I've only covered the parts I found most interesting, which are not necessarily representative of the book as a whole.

Defining "pathological altruism"

You can see a few different conceptions in the book:

One or two chapters, such as the one on unconditional cooperation discussed below, seem to regard as "pathological" what I would simply call "altruism": helping others, at cost to oneself, when it's not likely to pay off for you in the long run. But most chapters assume there is a reasonable form of selflessness, and are concerned with cases that are somehow excessive or misguided.

Everyone on the Internet needs an intervention

"Self-Addiction and Self-Righteousness" is the first of two chapters David Brin contributed. He speculates that many harmful behaviors may be the result of an addiction process that's enabled by the psychoactive chemicals our brains release when we are in certain states of mind. To establish the plausibility of this, he points to things such as fMRI studies comparing the effects of gambling and cocaine, and claims regarding the effect of prayer on brain chemistry (in the brain of the pray-er, not the prayed-for, I assume).

Brin is especially interested in whether an addiction process may be causing an intensification of "self-righteous indignation":

We all know self-righteous people. (And, if we are honest, many of us will admit having wallowed occasionally in self-righteousness ourselves.) It is a familiar and rather normal human condition, supported — even promulgated — by messages in mass media. Although there are many drawbacks, self-righteousness can also be heady, seductive, and even . . . well . . . addictive. Any truly honest person will admit that the state feels good. The pleasure of knowing, with subjective certainty, that you are right and your opponents are deeply, despicably wrong. Or, that your method of helping others is so purely motivated and correct that all criticism can be dismissed with a shrug, along with any contradicting evidence.


Indeed, one could look at our present-day political landscape and argue that a relentless addiction to indignation may be one of the chief drivers of obstinate dogmatism and an inability to negotiate pragmatic solutions to a myriad modern problems.

That was written in 2008 (or maybe earlier), back when Twitter was still just an app for telling your friends you were pooping. Subsequent developments have not exactly made the issue seem less important.

"Pathological Certitude", the subject of Robert A. Burton's chapter, seems related. Immense harm is often done by people who feel totally sure that they're doing the right thing. Consider suicide bombers: a later chapter on "Suicide Attack Martyrdoms", discussing what factors make someone more likely to become one, notes that being altruistic is "the only attribute for which there is full consensus between scholars who have approached the sociopsychology of suicide attacks". Burton uses the less dramatic example of a doctor who pressured patients into painful and useless treatments because he felt an unshakeable commitment to trying to prolong life at any cost, even when the probability of success was extremely low and the patient wanted to let go. Burton warns:

Although motivations will vary, the thread linking all altruistic behavior is a profound sense of conviction that one's actions are both morally correct and serve an ultimate good. Yet, despite the fact that a moral conviction feels like a deliberate, rational conclusion to a particular line of reasoning, it is neither a conscious choice nor even a thought process. Certainty and similar states of "knowing what we know" arise of out of primary brain mechanisms that, like love or anger, function independently of rationality or reason..."

Feeling sure is just that: a feeling, an emotion. "It helps," Burton says, "to see this feeling of knowing as analogous to other bodily sensations over which we have no direct control" - sensations which are vulnerable to malfunctions and illusions. I think this is an important thing to try to keep in mind when you find yourself totally convinced of anything. A sense of certainty, similar to feelings of anger or anxiety, can easily reflect our own ignorance and biases rather than reality.

Empathic distress vs compassion

Klimecki and Singer's chapter discusses the problem of compassion fatigue, which is widespread among those involved in healthcare ("prevalence rates ranging from 40% to 80%"). It's a kind of burnout from continuously dealing with suffering people.

The authors believe empathic distress fatigue would be a better term. Compassion can be thought of as "feeling for" someone while empathy involves "feeling with" them. A compassionate person cares about a suffering person; an empathizing person actually feels suffering that mimics that of the suffering person.

The chapter argues that when you encounter a situation that triggers empathy, the healthy response is for it to develop into compassion. If it instead develops into "empathic distress", not only will you suffer, but your ability to help will be compromised. This diagram from page 377 summarizes:

They cite studies indicating that when a person feels high "empathic concern" for someone suffering, there is an increased chance that they'll simply try to get away from the situation (in order to relieve their own distress) than try to actually help the person. This "empathic distress" response may be what underlies burnout among medical workers, not compassion, which the authors portray as pleasurable and sustainable.

If that's true, how can we help people who habitually respond with distress shift to a compassion-based response instead? The chapter has little to say about this, but mentions studies on the effectiveness of certain kinds of meditation. I wonder if there may also be some benefit simply in talking about the distinction and promoting compassion as the superior social norm. I've encountered people who seem to feel morally obligated to enter into the pain of others. This can be quite debilitating, given the overwhelming quantity of suffering happening on all around the world all the time. It's important to realize that such empathy is not the only, or best, way to be a caring person.

Forget pathological; does altruism exist at all?

Chapter 24 is subtitled "There Is No Such Thing as Altruism, Pathological or Otherwise". To help make this bold claim, the author (Satoshi Kanazawa) refers to a distinction between two types of altruism:

psychological altruism: "behavior that the actors believe would benefit others at the cost to themselves"
evolutionary altruism: "behavior that increases the reproductive fitness of others at the cost of decreasing the actors' own reproductive fitness"

He thinks the latter kind doesn't actually exist. Or, if it does, it's only an accident caused by the fact that our brains are evolved for an environment very different from the modern world; behaviors that nowadays hinder the spread of a person's genes, might have promoted that spread for a person living in our ancestral environment.

As a case study, he considers women who stay with violently abusive partners for prolonged periods of time. This represents psychological altruism because "[t]here is nothing that the abused women themselves (as individuals) gain by staying with their abusive partners." (I don't think he addresses the possibility that they may fear being subjected to even worse abuse if they leave.) But it may be beneficial for their genes; he reasons as follows:

  1. battered women have more sons than other women (he cites a previous study of his)
  2. violent men are likely to have violent sons
  3. violence helps men compete and have more opportunites to reproduce (or at least, it used to)

So, genes which predispose a woman to stay with a violent man might have evolved because they propagate efficiently via her violent sons. Kanazawa does inject one ray of sunshine into that disturbing story: there is data to suggest that violent tendencies have become a disadvantage for men in our era.

Bernard Berofsky also questions the existence of altruism from a somewhat different angle in chapter 20, "Is Pathological Altruism Altruism?" He defends psychological egoism. One of my pet peeves is when people defend a naive version of psychological egoism by saying things like: even when you help someone, you do it to fulfill your goal/desire of helping them, so it's ultimately selfish. Berofsky does a good job of summarizing what's wrong with that perspective:

...the egoist may be confusing motive as a psychological state, an internal thing, with the object of the motive, say, food, an external thing. I am moved by hunger, an internal state; but the object of my hunger is the food out there. ... Now, it may be true that we never act without some motive. I would not seek food if I were not hungry; I would not seek the good of others if I did not desire to help others. But it is still true that my desiring is different from what I desire, even if the satisfaction of my desire is a pleasant consequence of my realizing my desire ... Thus, if "we are always motivated by self-interest" means "we are always moved to act by some motive or desire," it may be true; but if "we are always motivated by self-interest" means "the object of every motivation is the self-interest of the person with the motive," it is false.

He then presents a refined form of psychological egoism. He allows that our conscious intentions may be altruistic, but thinks the important issue is whether our subconscious motives - which we may not even know ourselves - are ultimately self-serving. Let's imagine how this might apply (the scenarios are mine, don't blame them on the paper):

Scenario (A)

Berofsky wouldn't count (A) as altruism. That seems somewhat plausible, but what if we extend that thinking?

Scenario (B)

In Kanazawa's terms, in (B) you exhibit psychological altruism but not evolutionary altruism. If we think (A) isn't true altruism because the altruistic conscious thought had a non-altruistic cause, doesn't that suggest (B) isn't true altruism because the altruistic brain as a whole had a non-altruistic cause? What would count as true altruism? It seems like we're headed down a path that rules it out a priori. To highlight this, let's make the thought experiment a little more fanciful:

Scenario (C)

In (C) as in (B), the altruistic state of mind has a cause that is non-altruistic. Does that mean you're not really altruistic in (C) either?

Ultimately it's a question of definitions, and different definitions are useful in different contexts. I'm resistant to psychological egoism, even in the refined form, because it seems to suggest that conscious experience is inherently less important than the subconscious. For one thing, I'm not sure that gives sufficient weight to the fact that our conscious experiences are themselves causal inputs feeding back into our own minds and the minds of others. But also, there is a perspective from which conscious experience is the most important thing; the thing that gives everything else value and meaning. When I listen to music, the brain processing and interpreting the sound waves is the janky result of 3.5 billion years of accumulated accidents, but the phenomenal experience of the song can still be inexpressibly beautiful. A conscious mind experiencing a feeling of care for another is a wonderful thing too, even if the same wonder isn't present in all the layers of causal factors that brought it about.

Unconditional cooperation, or, What Would Jesus Do In Prison

Suppose you've got a species that does nothing but play the prisoner's dilemma against each other. Matchups among members of the population are random. What sort of behavior will become most prevalent? Chapter 23, "The Messianic Effect of Pathological Altruism", considers a particular mathematical model of this problem, in which defection always dominates. I.e., evolution weeds out all the nice people. But try adding even just one unconditional cooperator, i.e., someone who will always be nice even when everyone else is a jerk. Then cooperation becomes a viable strategy for others, and the model predicts that the population will ultimately contain a significant number of cooperators.

There's an enormous gap separating such clean and simple models from the complex realities of biology. But there's at least a little intuitive plausibility to it. Someone who sets an example of altruistic behavior over the course of a lifetime has a chance of gaining followers who will also behave altruistically. As the number of such followers grows, it becomes safer to be one of them, perhaps leading to further growth. From a consequentialist perspective, this is a point in favor of putting effort into good causes even if you know they won't come to fruition: the influence your attempts have on other people may add up to something significant.


Follow-up questions