Socrates, weakness of the will, and addiction

Socrates“People are dying because we misunderstand how those with addiction think,” says the title of a recent article in Vox by philosopher Brendan de Kenessey, who argues that addiction is not a moral failure, and that it is the moralistic attitude of a number of politicians and a significant portion of the public that makes the problem more difficult to deal with. Addicts are not bad people who need to be punished, he says, they are sick and need help.

And he is completely right, I think. And yet, I also suggest that the bulk of the article is based on the wrong philosophical criticism. de Kenessey blames Socrates for the moralistic attitude, while he should blame certain brands of Christianity instead. Here I will not make the positive case against Christian moralism (which is well known among certain politicians of a certain party in the US), nor will I unpack the idea that addicts are sick, not bad, people, as de Kenessey does a very fine job of that in his article. But I will defend Socrates and use the occasion to talk a bit not just about addiction, but in general the phenomenon of what the Greeks called akrasia, or weakness of the will, and which Socrates thought simply does not exist.

The starting point of de Kenessey’s analysis of the Socratic problem is the Platonic dialogue known as the Protagoras, in which the discussion between the Athenian sage and one of the most famous sophists turns to the topic of akrasia. Let’s contrast two instances of alleged akrasia, brought up by de Kenessey to make his point against Socrates, and which I think, on the contrary, show pretty clearly why Socrates was correct (once we add an hidden premise to the Socratic position, a premise not discussed by de Kenessey).

Imagine yourself in front of the television, intending to binge watch a season of Black Mirror (or whatever your favorite show happens to be). You think, when you reflect on it, that this isn’t really the best use of your time, and that you should instead pick yourself up and go to the gym, as lately you’ve let yourself go a little, and you don’t feel good, both physically and psychologically. You mull it over a bit, but in the end decide to stay and watch television, with munchies to accompany the experience.

Now imagine, says de Kenessey, an addict who is driving down the projects, thinking that he really ought to stop what he is doing, turn his life around, clean up, get a job, and take care of his family. Nevertheless, he keeps driving to the corner where he regularly meets his dealer, and buys some cocaine instead.

The two cases appear to have a similar structure, like this:

Subject A has two courses of action available to him, X and Y.

A thinks that he should do X, even though he is very tempted by Y.

A ends up doing Y, rather than X.

Socrates has this to say, in the Protagoras, about this kind of situation:

“No one who knows or believes there is something else better than what he is doing, something possible, will go on doing what he had been doing when he could be doing what is better.”

This seems paradoxical, in the original meaning of the term (para doxan = uncommon opinion), as it is a straightforward observation that people, like both our hypothetical television binger and drug addict, very often don’t do what they believe to be the best thing for them. And yet, Socrates is not alone in taking this position. Modern economists such as Paul Samuelson have proposed an approach in behavioral economics known as “revealed preference,” according to which people show what they really like by what they do, not by what they say. Similarly, modern psychology has accumulated a pretty good amount of evidence that we often confabulate about the reasons why we do things, i.e., we make up reasons to justify our actions because we often don’t really have a good understanding of our own motivations.

How does Socrates defend his “paradoxical” position, which seems to fly so clearly in the face of the evidence? He thinks that people in these cases do not suffer from akrasia, i.e., weakness of the will, thus acting against their best judgment. He thinks instead that people are doing exactly what they want to do, but are doing it because of bad judgment. Doing bad things is, therefore, a matter of ignorance, not malice.

Ignorance my ass, one might easily retort. The television watcher is not ignorant, and neither is the drug addict. They don’t luck the pertinent information, they don’t need to be educated about what is going on. True, but the word used in the Platonic dialogues in this context is amathia, which although usually translated as ignorance actually means something closer to un-wisdom, the opposite of sophia, one of the roots of the word philosophy. Socrates is arguing that apparent cases of weakness of the will are actually cases of lack of wisdom — not of factual or empirical knowledge, but of the proper way to arrive at judgments given certain factual or empirical knowledge.

Ever since discovering the Socratic idea of replacing akrasia (and, more importantly, actual “evil”) with amathia I found myself to be significantly more prone to understand others’ motivations and actions, to sympathize with their manifest lack of wisdom even when I cannot possibly condone their actions, and to generally cultivate an attitude of sorrow rather than anger when people do bad things. I find this new approach liberating and far more constructive than either the akratic or, much worse, the moralistic one.

Still, isn’t de Kenessey right that Socrates ends up blaming the victim here, and that it is this sort of blame that justifies the kind of draconian measures implemented by politicians, and supported by the public, that made the so-called war on drugs a total disaster with a high cost to society, both in human and financial terms?

I don’t think so, and the reason is that if we want to read Socrates charitably we need to see that the two cases above are actually distinct, and they are distinct because of a hidden premise in the Socratic approach. That premise is that we are talking about a normally functioning human mind, not a diseased one. It was well known even in the ancient world that human beings have a tendency to reason very poorly when they are under the influence of a number of external conditions, particularly drugs (including wine). A good deal of Greek tragedy is built on that premise, such as Euripides’ The Bacchantes. That is why Diogenes Laertius, commenting on the Stoics — which were explicit followers of Socrates — says that “they will take wine, but not get drunk.” (VII.118) Getting drunk artificially impairs one’s judgment, so when one is under the influence, as we say today, one is not suffering from lack of wisdom, he’s suffering from a temporarily dysfunctional mind.

If this is a reasonable and charitable interpretation of Socrates’ take, then the two cases of the television binger and the drug addict are very different. The first is an actual case of what Socrates is arguing against Protagoras: the binger — in accordance with modern behavioral economics theory — really does prefer to stay at home to watch Black Mirror rather than going to the gym. Yes, of course he knows that in the long run he would be better off taking the second course of action, but he judges that for him, right here and right now, binging is better. His future self be damned. He is, of course, mistaken in such judgment, just like Socrates maintained.

The same reasoning, by contrast, does not apply to the drug addict, precisely because he is an addict, and therefore his judgment is impaired. He is not suffering from amathia, he is suffering from a chemical addiction. And that is why the moralist attitude criticized by de Kenessey is pernicious, because it does not recognize that the person in question is sick, not evil (or unwise, as Socrates would put it).

There is, of course, a wrinkle in all this, which de Kenessey must be aware of, and yet never mentions in his article: on the first occasion that the soon-to-be drug addict decided to take cocaine his judgment was not impaired by being sick, yet. Which means he is still responsible for the initial decision to go down that road. Now we only have two ways of looking at the onset of the addiction, then: either the person is morally bad (the moralist view), or he lacks wisdom (the Socratic view). Not only the second view is more humane, it also makes much more sense than invoking akrasia: the future drug user had not yet had the experience of being on drugs, so he couldn’t possibly have yielded to the temptation of temporary pleasure promised by the drug. More likely, he made the unwise judgment that the drug wasn’t as bad as people say, or that he will have the willpower to resist the addiction, or something along similar lines and to the same effect.

de Kenessey points out that several modern philosophers have attempted to come up with an anti-Socratic account, but they can’t agree on what’s going on: for Harry Frankfurt the desires that represent our true self are those desires that we want ourselves to have (Harry Frankfurt); for Gary Watson they are the desires that align with our judgments of what is valuable; for Michael Bratman they are the desires that cohere with our stable life plans; and for Susan Wolf they are the desires that are supported by rational deliberation (Susan Wolf).

This business of a “true self” is, however, a red herring. As de Kenessey argues, modern psychology has done away with that notion (so did David Hume, two a half century before modern psychology). But the fact remains that “we” do make decisions in response to our desires and as a function of our capacity to arrive at judgments. Whether “we” are made of a unitary self, a bundle of perceptions, or whatever, doesn’t matter. Our judgments are either made by a functional human mind (in which case we are responsible for them) or by a non-functional one (in which case we are sick and need help). The difference between the moralist and Socratic view pertains to the first, not the second case. And there one has a choice of blaming people for the evil doing, or pity them for their lack of wisdom. I find the latter course of action to be far more preferable.

139 thoughts on “Socrates, weakness of the will, and addiction

  1. wtc48

    Much of the debate on this subject revolves around the question of agency, which could be regarded as the central problem of humanity. We are the only animals to whom this concept applies, because the rest of them derive their behavior from instincts developed through eons of natural selection. For humans, the relation between our culture and the vestigial instincts that we still possess must certainly be very complicated, as is well expressed by Massimo:

    “Similarly, modern psychology has accumulated a pretty good amount of evidence that we often confabulate about the reasons why we do things, i.e., we make up reasons to justify our actions because we often don’t really have a good understanding of our own motivations.”

    Further, in Lisa Bortolotti’s concluding remark: “Confabulation compromises our understanding of reality and of ourselves, but, when it comes to supporting agency, it often fares better than a well-grounded explanation, or even the accurate one.”

    Confabulation seems to be an essential element of human culture, which is not to say that it is not also potentially a kind of dancing with the devil. Ibsen dealt with it at length in “The Wild Duck’ through the character of Dr. Relling, whose concept of the “life-illusion” acts as a foil to the misguided idealism of some of the other characters:

    “Rob the average man of his life-illusion, and you rob him of his happiness at the same stroke.”


  2. Massimo Post author


    category mistake: coq is a theorem verifier, it has nothing whatsoever to do with whether mathematical theorems needempirical evidence to be verified. Nothing.


  3. brodix


    ““Reason tends to be linear”

    No. This is far too broad of a statement. Some types of reasoning are linear, many are not. And at any rate, my answer to your question remains the same: reason. What else?”

    Which is why I prefaced it with “tends.”

    The topic of discussion is the power of the will, versus habits and/or addictions. So far as I understand it, conscious will is linear. It is goal oriented. Loose weight, stop drugs, climb the mountain, run the marathon, etc. While habits and addictions are self-reenforcing feedback loops, i.e., circular.

    So my point is that possibly it might be useful to give some consideration as to how linear and circular logic interact and relate to one another. It might inform everything from our understanding of personal and cultural pathologies, to whether time is eternalist, or presentist.


  4. Massimo Post author


    there is no begging the question. It’s just that you seem under the impression that if one checks a mathematical theorem using a computer that somehow means that empirical evidence is needed to prove the theorem. I don’t know how else you put it, that’s just plain wrong.


  5. Massimo Post author


    again, compute-assisted proof just means that mthematicians are incapable (at the moment, ever) to provide an analytic justification of a theorem. It has nothing to do with empirical evidence as understood in the natural sciences. It’s not about facts “out there.” At all.


  6. Disagreeable Me (@Disagreeable_I)

    Hi Massimo,

    So perhaps it is you who are missing something.

    Certainly. My attitude is of one seeking clarification of something that doesn’t seem quite right, not of a self-proclaimed visionary insisting that everyone else is wrong.

    two views even have different empirical implications in terms of behavioral economics

    I don’t get this bit. Behavioral economics shows that people have a revealed preference. That is compatible with akrasism as I understand it — it’s just that the akrasists wou’dn’t call your revealed preference “what you really want”, or at least that’s not what they’re talking about when they talk about what you really want.

    You also mention confabulation. Again, that is also compatible with akrasism. Confabulation shows that we don’t understand our innermost drives. It does not show that people don’t believe they have certain desires, or desires they wish they had, or desires they believe they should have, and these are the desires that the akrasists are talking about when they discuss what you really want. There is no fact of the matter regarding whether this is actually “what you really want”, since “what you really want” is just a common phrase, and it means whatever it is that people mean when they say it. Different people may mean different things, as in the case of the akrasists and the amathists.

    Unless you want to deny that addicts are sick, or you think that everyone is sick.

    I would say that pretty much everyone is at least mildly unhealthy in some way or another.

    As I said, I’m not married to the position that the difference is qualitative.

    Sure. My problems is that I don’t see why it is important to draw the distinction you are drawing in this case. If one wanted to treat the two cases differently, sure. The inference I (it seems incorrectly) drew from my initial reading of the OP was that one should not be judgmental of addicts because they are sick, unlike couch potatoes — the obvious corollary being that being judgmental of couch potatoes is fine. You even say that the point of drawing the distinction is so that we see we shouldn’t “blame the victim” in the case of addicts, which clearly implies that we should indeed blame lazy people. If you don’t want to be judgmental of couch potatoes, I don’t see what work the distinction is doing.


  7. Disagreeable Me (@Disagreeable_I)

    Hi Dan,

    I actually agree with a lot of what you say here. I agree that there is no principle by which we can clearly and consistently ascribe a defect”. I’m not looking for one. I’m trying to figure out why Massimo sees the case of the addict and the couch potato as being so different.

    The example you gave is simply one in which we are asked to imagine a non-lazy couch potato; i.e. one who is sitting there not out of laziness, but as a result of clinical depression.

    No, that’s not it. This person is simply lazy. But I see laziness as a mental defect. I don’t think in terms of drawing a distinction between “failures of character” and “mental illness”. To me, it’s all the same sort of thing, and which defect gets put into which category by society is somewhat (if not entirely) arbitrary. There may be practical reasons for splitting the categories (e.g. laziness may not be serious enough to warrant medical insurance coverage), but I don’t see the point of making the distinction in the context of Massimo’s analysis of this issue unless it is to justify self-righteous blaming and disapproval of lazy people while arguing against the same attitude regarding addicts.


  8. Daniel Kaufman

    Philip; the truth of anything that is confirmed empirically can only ever be probabilistic. You’re not seriously suggesting that the truth of , say, propositions in geometry are probabilistic are you?

    Liked by 2 people

  9. davidlduffy

    “[L]ogic studies, from an objective point of view, our pieces of knowledge as they are organized in demonstrative science, or, if you think about it from the act point of view, it studies our acts of judging, or knowing, and how they are interrelated.”

    Per Martin-Lof

    Click to access article.pdf

    Liked by 2 people

  10. Massimo Post author


    I’m not dismissive of experimental mathematics (though, frankly, Zeilberger’s analogy is outrageously stupid). I just don’t think “experimental” here means empirical in the sense of the natural sciences.

    Liked by 1 person

  11. Daniel Kaufman

    Philip: that zeilberger quote strikes me as both facile and demagogic.

    The question of whether mathematics is ultimately empirical or a priori was hashed out a long time ago, between Mill and Frege and empiricism lost. So much so that the logical empiricists themselves didn’t think mathematics was empirical.


  12. Robin Herbert


    Ultimately there is no physical system that is computable because physical systems are modelled with uncomputable numbers. So chlorophyll is not any kind of exception here. But I don’t understand why you think that any of the classical logics would have any problem with this.

    Also, time is not a component of any classical logic system and so it has no problems with things that would take a very long or even infinite time, thus we can write an algorithm for a quantum computer without needing anything more that the classical logics.

    Things don’t need to be physically possible in order that they can be described by classical logic. I can define a Zeno Machine using classical logic even though a Zeno Machine is physically impossible. A Zeno Machine is just like any computer except that each step after the first takes half the time of the previous step. So if you want to know if an algorithm halts and you know the first step on the Zeno machine takes one second then if you haven’t received any output after 2 seconds then you know the algorithm never halts. So a Zeno Machine has no problem computing the binary expansion of a Chaitin Constant to any number of digits and so is more powerful than a Turing machine.

    If we give the Zeno Machine some Zeno Registers (where each subsequent bit takes up half the space of the previous bit) we can generate an entire Chaitin Constant (which is uncomputable and has infinitely many bits). We could even add two Chaitin constants together.

    The problem would happen when we try to use the Rev(A,B) operation which takes the first bit of the A Zeno Register and pushes it onto the first bit of the B Zeno register and then the next bit of the A register and pushes it onto the B register so that it pushes the next bit along and continues this while there are more bits in the A register (remembering that the A Zeno Register contains infinitely many bits).

    But the process is run on a Zeno machine and so completes in a finite amount of time. The question is, what will be the first bit in the B Zeno Register? I have always wanted to know what the last digit of Pi is.

    The point is, the Zeno Machine is, as I said, physically impossible but we have no problem whatsoever talking about it using the classical logics because (and this is the point many don’t seem to grasp) the classical logics are not tied to any physical assumptions.

    I think the problem is that we have this term “classical physics” and another term “classical logic” and people think they are related. They aren’t.

    Liked by 2 people

  13. Robin Herbert

    Incidentally, take some time to read the original paper that defined fuzzy logic. There are a number of theorems in that paper and they are all proved using … wait for it … classical logic!

    Liked by 2 people

  14. Robin Herbert

    Classical logics cannot handle logics which dispense with the identity and non-contradiction tautologies, but of course that means that it is also true the classical logics can handle logics which dispense with the identity and non-contradiction tautologies with ease.


  15. Robin Herbert


    “We like to think our conceptually-conscious minds can solve problems rationally and determine the future we want, but both the ‘infinite regress’ that defines our past, and the ‘infinite progress’ that defines our future can hardly be handled rationally by the conceptually-conscious mind in ‘the now’ …”.

    Well, I am conscious and if we could not understand concepts then we could not be having this discussion.

    We can solve some problems rationally, I do so all the time at work. But I don’t think that anyone thinks that we can solve any problem.

    I can determine my future to a certain extent. I can plan to have a certain thing for lunch, go to a certain place for a holiday and make it happen. Again, I doubt that there is anyone so naive that they think they can make any future happen.

    Do you think that you are capable of coming to rational conclusions about anything? Do you think that you are capable of making a rational statement about the concepts of the infinite regress of our past and the infinite progress of our future?

    If you can, then why don’t you think anyone else can? If you think that you can’t do this then anything you say on the subject must be either completely wrong or meaningless.


  16. Robin Herbert

    I don’t like the “what you did is what you really wanted” approach.

    Some people would say that if I said I wanted to do my homework and I just sat there the entire time staring at a wall, then the thing I really wanted to do was stare at a wall and not to do my homework.

    But that is like saying that if I am behind a tall wall and I say I want to be on the other side of that wall, then what I really want is to be on this side of the wall, because that is what I did. But if I had a ladder then I would have climbed over the wall.

    With the homework, if I had help to get past that block then I would have done my homework.

    Liked by 1 person

  17. Daniel Kaufman

    DM: Sounds to me like you don’t think there are any defects of character — i.e. vices. That would also mean that there are no virtues. Both positions strike me as non-starters. As flat-out not just denying the phenomena, but refusing to play the language game that everyone else is playing. These things are not decided individually. They are matters of social practice.

    Liked by 3 people

  18. Patrice Ayme

    Creating may sound like equivocating. I read logic books at the research level for several decades, and papers too. I have pondered the subject quite a bit. Graduate texts, in nearly all fields speak with authority, but aren’t always correct (see the Molecular Biology textbooks of the 1960s, which made drastic, overarching claims now known to be false; same in cosmology, paleontology, anthropology, or even geology!)

    There are cartesian closed categories interpreting classical logic, or intuitionistic logic. I (implicitly) generalized this to generalized logic: say any category is a logic. The arrows are implications, the objects, propositions. What I say is that implications can be produced by a machine, and it does not have to be just a Turing machine (as it is in classical logic), it just has to be systematic. It could be a Quantum computer (see chlorophyll), or a DNA computer.

    Topological dynamics of knots can duplicate known logics (this is the underlying reason for Microsoft Q program for using the Quantum Hall effect to make Quantum computers; the project head has a Fields Medal in differential topology). A fortiori neurochemistry can make knots in brains among other things. No I am not “making this up”. Consider:

    This general set-up, applied to neurochemistry as a computer, makes it into a logic, which I chose to call “emotional logic” it is a meta controller onto classical (neuronal) logics in the brain… So not only do knowledge bases vary, but so do logics. Even in the same individual, mood dependent.

    Liked by 1 person

  19. Daniel Kaufman

    Philip, your reply is a dodge. You claimed that mathematics is empirical. I pointed out that this would entail that mathematical statements are probabilistic, which they clearly are not. Simple modus tollens. To which you reply “it’s random.”

    No one ever died from admitting they were wrong about something. Why not give it a try?

    Liked by 3 people

  20. Disagreeable Me (@Disagreeable_I)

    Hi Dan,

    DM: Sounds to me like you don’t think there are any defects of character — i.e. vices.

    No, that’s not true, I just think that vices are a kind of mental illness (or at least the same sort of thing as mental illness, i.e. something suboptimal or undesirable in cognition) and don’t really need to be analysed differently in the specific context of Massimo’s point.

    I also don’t think that there’s much philosophically wrong with calling addiction or clinical depression a vice, except of course that it connotes a lack of sympathy, understanding or compassion. But what I am advocating is compassion and understanding of all mental conditions. It’s only when you want to cling to judgmentalism or moralistic thinking that the distinction becomes important.


  21. Disagreeable Me (@Disagreeable_I)

    OK, let me put it this way.

    It seems to me that part of the reason depression or addiction are now treated as diseases rather than vices is at least in part because we have a better understanding of their causes and how to treat them. Similarly, pyschopaths used to be just evil, now they’re mentally ill.

    If we had a deep understanding of the neurochemistry underlying laziness, and in particular if we found an effective side-effect free drug or therapy which “cured” laziness, then we would have less trouble regarding laziness as a disorder. This is something that might actually happen in future. I think we see this kind of transition all the time. Over-eating/obesity was long regarded as a vice, but I think we’re now moving towards regarding it as a psychological disorder.


Comments are closed.