There is No Akrasia

I don’t think akra­sia ex­ists.

This is a fairly strong claim. I’m also not go­ing to try and ar­gue it.

What I’m re­ally here to ar­gue are the two weaker claims that:

a) Akra­sia is of­ten treated as a “thing” by peo­ple in the ra­tio­nal­ity com­mu­nity, and this can lead to prob­lems, even though akra­sia a sorta-co­her­ent con­cept.

b) If we want to move for­ward and solve the prob­lems that fall un­der the akra­sia-um­brella, it’s bet­ter to taboo the term akra­sia al­to­gether and in­stead em­ploy a more re­duc­tion­ist ap­proach that fa­vors speci­fic­ity

But that’s a lot less catchy, and I think we can 8020 it with the state­ment that “akra­sia doesn’t ex­ist”, hence the ti­tle and the open­ing sen­tence.

First off, I do think that akra­sia is a term that res­onates with a lot of peo­ple. When I’ve de­scribed this con­cept to friends (n = 3), they’ve all had vary­ing de­grees of re­ac­tions along the lines of “Aha! This term perfectly en­cap­su­lates some­thing I feel!” On LW, it seems to have gar­nered ac­cep­tance as a con­cept, ev­i­denced by the posts /​ wiki on it.

It does seem, then, that this con­cept of “want-want vs want” or “be­ing un­able to do what you ‘want’ to do” seems to point at a phe­nomenolog­i­cally real group of things in the world.

How­ever, I think that this is ac­tu­ally bad.

Once peo­ple learn the term akra­sia and what it rep­re­sents, they can now pat­tern-match it to their own as­so­ci­ated ex­pe­riences. I think that, once you’ve reified akra­sia, i.e. turned it into a “thing” in­side your on­tol­ogy, prob­lems oc­cur:

First off, treat­ing akra­sia as a real thing gives it ad­di­tional weight and power over you:

Once you start to no­tice the pat­terns, it’s harder to see things again as mere ap­par­ent chaos. In the case of akra­sia, I think this means that peo­ple may try less hard be­cause they sud­denly re­al­ize they’re in the grip of this ter­rible mon­ster called akra­sia.

I think this sort of wor­ld­view ends up re­in­forc­ing some un­helpful at­ti­tudes to­wards solv­ing the prob­lems akra­sia rep­re­sents. As an ex­am­ple, here are two para­phrased things I’ve over­heard about akra­sia which I think illus­trate this. (Happy to re­move these if you would pre­fer not to be men­tioned.)

“Akra­sia has mu­tant heal­ing pow­ers…Thus you can’t fight it, you can only keep switch­ing tac­tics for a time un­til they stop work­ing…”

“I have mas­sive akra­sia…so if you could just give me some more high-pow­ered tools to defeat it, that’d be great…”

Both of these quotes seem to have taken the akra­sia hy­poth­e­sis a lit­tle too far. As I’ll later ar­gue, “akra­sia” seems to be dealt with bet­ter when you see the prob­lem as a col­lec­tion of more iso­lated dis­parate failures of differ­ent parts of your abil­ity to get things done, rather than as an um­brella term.

I think that the cur­rent akra­sia fram­ing ac­tu­ally makes the prob­lem more in­tractable.

I see po­ten­tial failure modes where peo­ple come into the com­mu­nity, hear about akra­sia (and all the re­lated scary sto­ries of how hard it is to defeat), and end up us­ing it as an ex­cuse (per­haps not an ex­plicit be­lief, but as an alief) that im­pacts their abil­ity to do work.

This was cer­tainly the case for me, where im­proved in­tro­spec­tion and metacog­ni­tion on cer­tain pat­terns in my men­tal be­hav­iors ac­tu­ally re­moved a lot of my willpower which had served me well in the past. I may be get­ting slightly tan­gen­tial here, but my point is that giv­ing peo­ple mod­els, use­ful as they might be for things like clas­sifi­ca­tion, may not always be net-pos­i­tive.

Hav­ing new things in your on­tol­ogy can harm you.

So just giv­ing peo­ple some of these pat­terns and say­ing, “Hey, all these pieces rep­re­sent a Thing called akra­sia that’s hard to defeat,” doesn’t seem like the best idea.

How can we make the akra­sia prob­lem more tractable, then?

I claimed ear­lier that akra­sia does seem to be a real thing, as it seems to be re­lat­able to many peo­ple. I think this may ac­tu­ally be­cause akra­sia maps onto too many things. It’s an um­brella term for lots of differ­ent prob­lems in mo­ti­va­tion and effi­cacy that could be quite dis­parate prob­lems. The typ­i­cal akra­sia fram­ing lumps prob­lems like tem­po­ral dis­count­ing with mo­ti­va­tion prob­lems like in­ter­nal dis­agree­ments or ugh fields, and more.

Those are all very differ­ent prob­lems with very differ­ent-look­ing solu­tions!

In the above quotes about akra­sia, I think that they’re an ex­am­ple of hav­ing mixed up the class with its mem­bers. In­stead of treat­ing akra­sia as an ab­strac­tion that unifies a class of self-im­posed prob­lems that share the prop­erty of act­ing as ob­sta­cles to­wards our goals, we treat it as a prob­lem onto it­self.

Say­ing you want to “solve akra­sia” makes about as much sense as di­rectly ask­ing for ways to “solve cog­ni­tive bias”. Clearly, cog­ni­tive bi­ases are merely a class for a wide range of er­rors our brains make in our think­ing. The ex­er­cises you’d go through to solve over­con­fi­dence look very differ­ent than the ones you might use to solve scope ne­glect, for ex­am­ple.

Un­der this fram­ing, I think we can be less sur­prised when there is no di­rect solu­tion to fight­ing akra­sia—be­cause there isn’t one.

I think the solu­tion here is to be spe­cific about the prob­lem you are cur­rently fac­ing. It’s easy to just say you “have akra­sia” and feel the smooth com­fort of a catch-all term that doesn’t provide much in the way of in­sight. It’s an­other thing to go deep into your ugly prob­lem and ac­tu­ally, hon­estly say what the prob­lem is.

The im­por­tant thing here is to iden­tify which sub­set of the huge akra­sia-um­brella your in­di­vi­d­ual prob­lem falls un­der and try to solve that spe­cific thing in­stead of throw­ing gen­er­al­ized “anti-akra­sia” weapons at it.

Is your prob­lem one of re­mem­ber­ing to do tasks? Then set up a Get­ting Things Done sys­tem.

Is your prob­lem one of hy­per­bolic dis­count­ing, of fa­vor­ing short-term gains? Then figure out a way to re­cal­ibrate the way you weigh out­comes. Maybe look into pre­com­mit­ting to cer­tain courses of ac­tion.

Is your prob­lem one of in­suffi­cient mo­ti­va­tion to pur­sue things in the first place? Then look into why you care in the first place. If it turns out you re­ally don’t care, then don’t worry about it. Else, find ways to source more mo­ti­va­tion.

The ba­sic (and ob­vi­ous) tech­nique I pro­pose, then, looks like:

  1. Iden­tify the akratic thing.

  2. Figure out what’s hap­pen­ing when this thing hap­pens. Break it down into mov­ing parts and how you’re re­act­ing to the situ­a­tion.

  3. Think of ways to solve those in­di­vi­d­ual parts.

  4. Try solv­ing them. See what happens

  5. Iterate

Po­ten­tial ques­tions to be ask­ing your­self through­out this pro­cess:

  • What is caus­ing your prob­lem? (EX: Do you have the de­sire but just aren’t re­mem­ber­ing? Are you lack­ing mo­ti­va­tion?)

  • How does this akratic prob­lem feel? (EX: What parts of your­self is your cur­rent ap­proach do­ing a good job of satis­fy­ing? Which parts are not be­ing satis­fied?)

  • Is this re­ally a prob­lem? (EX: Do you ac­tu­ally want to do bet­ter? How re­al­is­tic would it be to see the im­prove­ments you’re ex­pect­ing? How much bet­ter do you think could be do­ing?)

Here’s an ex­am­ple of a re­duc­tion­ist ap­proach I did:

“I suffer from akra­sia.

More speci­fi­cally, though, I suffer from a prob­lem where I end up not ac­tu­ally hav­ing planned things out in ad­vance. This leads me to do things like browse the in­ter­net with­out hav­ing a con­crete plan of what I’d like to do next. In some ways, this feels good be­cause I ac­tu­ally like hav­ing the nov­elty of a lit­tle un­pre­dictabil­ity in life.

How­ever, at the end of the day when I’m look­ing back at what I’ve done, I have a lot of re­gret over hav­ing not taken key op­por­tu­ni­ties to ac­tu­ally act on my goals. So it looks like I do care (or meta-care) about the things I do ev­ery­day, but, in the mo­ment, it can be hard to re­mem­ber.”

Now that I’ve far more clearly laid out the prob­lem above, it seems eas­ier to see that the prob­lem I need to deal with is a com­bi­na­tion of:

  • Re­mind­ing my­self the stuff I would like to do (maybe via a sched­ule or to-do list).

  • Find­ing a way to shift my in-the-mo­ment prefer­ences a lit­tle more to­wards the things I’ve laid out (per­haps with a break that al­lows for some med­i­ta­tion).

I think that once you ap­ply a re­duc­tion­ist view­point and speci­fi­cally say ex­actly what it is that is caus­ing your prob­lems, the prob­lem is already half-solved. (Hav­ing well-speci­fied prob­lems seems to be half the bat­tle.)

Re­mem­ber, there is no akra­sia! There are only prob­lems that have yet to be un­packed and solved!