Exposure can’t rule out disasters

post by Chipmonk · 2024-08-15T17:03:37.259Z · LW · GW · 19 comments

This is a link post for https://chrislakin.blog/p/exposure-therapy-cant-rule-out-disasters

Contents

  Is there anything you avoid that exposure still hasn’t fixed?
None
19 comments

Is there anything you avoid that exposure still hasn’t fixed?

Why is that? 

In my experience, people who avoid X are usually not afraid of the average X… 

For example, I spoke to someone who is afraid of starting conversations with women he’s attracted to. But as far as I can tell, he has no fear of how this will usually go (probably just fine). 

Instead, he’s afraid of something like: he goes up to a girl, she finds him creepy and ridicules him, he ends up on social media and the whole world hates him. Exactly like this 19-second short parodies.

(I sent him this and he said, “LOL. They made a skit about my worst fear.”)

I suspect that people who are afraid of something, even after ample exposure, are afraid of the rare, worst case scenarios. The (subjective) disasters.

Why doesn’t exposure fix it?

Because exposure cannot disprove that something terrible might, at some point, happen. Update: See Kaj Sotala's comment below.

Need something stronger…

The general tool I like is memory reconsolidation [? · GW]. Here's an example of how I use it.

19 comments

Comments sorted by top scores.

comment by Kaj_Sotala · 2024-08-15T19:26:58.209Z · LW(p) · GW(p)

I think this might also have been mentioned in Unlocking the Emotional Brain [LW · GW], though I'm not entirely sure if it was or if I just figured it out after reading it. Namely, that if one's fear is "doing X will inevitably cause Y to happen", then finding counterevidence can be relatively easy. You just need to find a single instance where it doesn't happen. But if the fear is "doing X might cause Y to happen", then Y simply failing to happen won't trigger reconsolidation.

comment by ChristianKl · 2024-08-16T13:21:20.213Z · LW(p) · GW(p)

To me, this post seems to present a story about an empiric question that might be true or false but the post provides no good evidence that the story is true. It doesn't even offer anecdotal evidence. 

This is not the kind of reasoning that should get anyone on LessWrong to change their mind about the issue at hand.

Exposure therapy is well-studied in academia and the post makes no attempt to engage with the existing literature on it and whether it supports or challenges the claim of the post.

Replies from: habryka4, Chipmonk
comment by habryka (habryka4) · 2024-08-19T21:56:47.414Z · LW(p) · GW(p)

Eh, I don't buy it. I think it's making a locally valid point, and making an IMO reasonable inference that the cause of your phobias is probably not completely disconnected from good inference rules.

Agree that it would be nice to have a section in the post that's like "OK, but can we find empirical validation of this prediction (which seems reasonably likely based on these premises)?", but I don't think that's necessary for a good post. 

I personally hadn't super considered the important difference between stochastic risks and guaranteed risks and how those would interface with exposure therapy. The point is simple, so as soon as I would have come into more contact with the territory I would have probably noticed it, but I also wouldn't be surprised if someone had succeeded at making an invalid argument that I would nod along to if I was having a casual conversation about this topic, that makes the error this post is arguing against.

Replies from: ChristianKl
comment by ChristianKl · 2024-08-20T13:51:40.775Z · LW(p) · GW(p)

I didn't start by voicing an alternative narrative because I think the issue of epidemics is important enough to stand on its own. 

But let's write an alternative narrative:

People frequently have thoughts to justify their actions and their emotions. For me, it's easy to imagine that the causation goes from having a phobia to having thoughts about possible disaster. It's easy to imagine that people think more about hypotheticals that might happen in situation X than the usually things that happen in situation X when they are never exposed to situation X.

When fear started evolving in mammals or their predecessors, those animals were not capable of having cognitive concepts of low-probability risks. Neither for the freeze/shutdown mechanism that as far as I know are evolutionary older nor for fight/flight/fawn which are newer. 

We humans didn't get a new emotional fear system after using our precortex more. Our precortex wants to make sense of the fears that we have and might produce scenarios like the one described in the OP that we think about, but even if those scenarios are there and even if explicitly thinking about them feels scary, the core cause of the fear might still be in the amygdala. 

Do I know whether the narrative I wrote is more or less likely to be true than the one writing the OP? No, I don't it would need empiric evidence. It's the kind of empiric evidence that's either available or the thesis is so wrong that the professionals in the field just reject the thesis without finding it valuable to study it.

When it comes to self-help material on LessWrong, I don't want people to write about empiric claims for which they have neither empirical evidence from their own life experience nor empirical evidence from published sources. 

If these kinds of arguments that the OP is making would be the norm, I would expect it to lead to an adoption of a lot of wrong beliefs because they are convincingly argued but lack empirical backing. Especially, people who themselves don't have much contact with the territory are likely going to adopt those arguments over something that's more empirically grounded.

If you generally look at psychology research you find that ideas which can be argued in the strength the OP argued for his thesis often turn out to be wrong when researchers set out to study them. Plenty of research findings don't replicate. Building stable knowledge in that field is hard. 

Replies from: habryka4, Chipmonk
comment by habryka (habryka4) · 2024-08-20T22:00:09.646Z · LW(p) · GW(p)

FWIW, I get a feeling you are getting the basics of psychology reasoning basically backwards. Indeed, I think trying to make locally valid arguments from the basis of assumptions of rationality has a much better track record than macro-scale behavior arguments that rely on empirical data collection. 

Overall, I think if you are trying to understand how you and other people think, arguments of the type "here is a very common observation, or straightforward hypothetical, let's use common sense to figure out what people will do in this situation" perform much better than "here is a counterintuitive result which I have backed up with a bunch of experiments". 

comment by Chipmonk · 2024-08-20T21:46:46.474Z · LW(p) · GW(p)

empiric claims for which they have neither empirical evidence from their own life experience

what

comment by Chipmonk · 2024-08-17T06:05:38.076Z · LW(p) · GW(p)

I'm also just confused. What type of argument are you expecting for the claim "exposure cannot disprove that something terrible might, at some point, happen"? 

Replies from: ChristianKl
comment by ChristianKl · 2024-08-17T08:39:17.031Z · LW(p) · GW(p)

That's not the only claim you are making. You make a claim about how exposure therapy interacts with fear. Fear is not something that just happens via a rational analysis of the situation. It's an emotional process. 

There's something called evidence-based medicine. It's based on the idea that when people form their ideas about what's true about medical interventions the way you propose, they end up wrong. While evidence-based medicine isn't perfect, it's useful to understand why it exists.

If you reject evidence-based medicine as one of your unspoken assumptions when writing a post on LessWrong, I think you should be more conscious about doing so and have some justification for doing so that you can articulate. 

If those people who have fears that are about something terrible that happens in few instances don't get helped by exposure therapy, that would be quite useful knowledge for CBT therapists who want to decide whether or not to tell a given person to do exposure therapy. The study setup of asking people about their fears and then giving them exposure therapy and seeing whether you can predict treatment outcomes seems like it would be relatively easy to run. CBT has a knowledge community that likes running studies. 

If nobody run that study, that would mean that the experts in the field don't consider your thesis likely enough to be worth running a study. If someone however did run the study, it would be worth looking at the outcome of the study. 

Replies from: Chipmonk
comment by Chipmonk · 2024-08-17T16:16:04.628Z · LW(p) · GW(p)

Hi Christian, you seem frustrated. I would like to respond to your comment, but you haven't asked any specific questions I can respond to. Is there anything in specific you would like me to address?

Replies from: ChristianKl
comment by ChristianKl · 2024-08-17T16:28:54.413Z · LW(p) · GW(p)

You asked "What kind of argument are you expecting?"

I said "Either, the studies they did on the topic support my claim, see reference X" or "I think it's true despite them not having done studies on it, and I think the reason for them not having done studies is X". Or "I don't believe in the notion of evidence-based reasoning, empiricism is not valuable in a case like this because of X". 

Replies from: Chipmonk
comment by Chipmonk · 2024-08-17T16:30:28.631Z · LW(p) · GW(p)

Oh ok, I mistakenly thought that you wanted me to respond to what you said

Replies from: ChristianKl
comment by ChristianKl · 2024-08-18T11:45:50.573Z · LW(p) · GW(p)

I do would appreciate that you would respond and update in the direction of valuing knowing what's true. In practice that would likely involve engaging with the academic literature. Alternatively, you could also argue why you think why that isn't what would follow from it.

Replies from: Chipmonk
comment by Chipmonk · 2024-08-18T15:34:34.964Z · LW(p) · GW(p)

Sorry, I'm still trying to figure out what you're asking. I would love to respond to your comment but I need specific questions to respond to.  What specific questions do you have for me?

Replies from: ChristianKl, abandon
comment by ChristianKl · 2024-08-18T20:13:21.309Z · LW(p) · GW(p)

I care about truth and substance and not about debating.

If you want a specific question: "Why don't you put more effort into finding out that the thesis on which you wrote a post is true?"

Replies from: Chipmonk
comment by Chipmonk · 2024-08-19T00:57:38.420Z · LW(p) · GW(p)

I care about truth and substance and not about debating.

The trouble is that I don't know what you'll be receptive to. I don't want to talk randomly in your direction and have it not address your cruxes.

"Why don't you put more effort into finding out that the thesis on which you wrote a post is true?"

Thanks!

Though, you haven't specified which thesis, so I'll assume you meant the title:

Exposure therapy can’t rule out disasters

First, I'll explain what I said in the post, which I originally thought would've been sufficient. 

Afterward, I'll explain related context which I'm operating from which I did not include in the post.

First:

As I say in the post,

In my experience, 

There is something like I say in the post that definitely seems to be the case in my experience helping myself and others grow. I have seen many people vaguely try more exposure therapy and make little progress. 

You can consider this for yourself:

Is there anything you avoid that exposure still hasn’t fixed?

Do any examples come to mind in your life? Please let me know if yes or no.

Frankly I thought this would be self-evident for each reader to find things like this in themselves. I expected the evidence to come from you. (But I'm happy to explain this now that you've asked.)

Note that nowhere in the post do I say that exposure can't work or never works. Obviously it works sometimes. But not always. Again:

I suspect that people who are afraid of something, even after ample exposure, are afraid of the rare, worst case scenarios. The (subjective) disasters.

Maybe that could've been avoided if instead I added one word: "Exposure therapy can’t rule out all disasters"? 

But the reason I was okay with the title as-is was this:

Because exposure cannot disprove that something terrible might, at some point, happen.

This seems obvious so I'm not going to explain why this is true. Same reason that math proofs require actual proofs, rather than trying lots of numbers.

But I will clarify one thing:

My model of exposure is that if it updates unconscious predictions [LW · GW], it can only update predictions relating to stuff that actually happened. If average-case exposures made you unlearn your fears, then surely you weren't afraid of any worse case scenarios. (Maybe you happened to unlearn the fears simultaneously through other means, but that would be mere correlation.)

To unlearn fears to scenario X, you must be exposed to scenario X. However sometimes scenario X is so rare/bad that exposure is not workable.

(If that wasn't the case, then this seems like accepting the argument that "living" is exposure therapy for fear, and so simply by living you will unlearn all of your fears.)

Also:

I'll let you in on some context to my state writing this post. I didn't include this before but it seems helpful to say now.

I have helped people -- people who have tried intentional exposure for months and sometimes years before talking to me -- overcome much or all their aversions in just a couple of hours, no exposure necessary. (I will be sharing one of these case studies in another post soon.)

In general, I've learned much more about minds -- real results both for myself and others -- simply by iterating on my own, just like the post above. 

I recently had the chance to counsel a cofounder of Coherence Therapy and he said he was surprisingly impressed and made progress on an issue he wasn't able to make on his own.

The first person I counseled with my new method told me that our one conversation was "significantly more productive than my last 6 months of CBT and talk therapy I did". 

If mainstream growth theory was good, surely there would not be $100 bills lying on the ground like this?

I've also vaguely had the experience of trying to read mainstream academic material (with the exception of memory reconsolidation) and simply not understanding it because I disagree too hard with the assumptions. 

This has led me to somewhat avoid "the academic literature" out of fear of thought-contamination.


Does that make sense? Please let me know if you have any specific questions.

Replies from: Chipmonk
comment by Chipmonk · 2024-08-19T01:12:18.583Z · LW(p) · GW(p)

I should also add that I'm fortunate to be "sheltered" from the literature by mentors (one is a therapist, one Coherence Therapy, also others). They will just call me out if I say something wrong. I did not discuss this exact point with them in specific though but I think they would vaguely agree, I wouldn't be surprised if they helped me qualify my statement slightly more, but that's also why I'm posting this here to get feedback

comment by dirk (abandon) · 2024-08-18T20:17:26.129Z · LW(p) · GW(p)

I think ChristianKI might not be trying to ask questions, but rather to suggest that you engage with the literature that discusses whether or not exposure therapy works, in order to improve your post with a more detailed argument for why it doesn't (or alternatively to change your mind about whether it does).

comment by philip_b (crabman) · 2024-08-16T10:31:18.164Z · LW(p) · GW(p)

In my personal experience, exposure therapy did help me with the fear of such "extreme" risks.

Replies from: Chipmonk
comment by Chipmonk · 2024-08-19T19:31:25.575Z · LW(p) · GW(p)

This may sound tautological, but how do you know you that 1) you had the extreme fears; 2) it was exposure (as opposed to anything else or stochasticity) that fixed it?