Multiple Working Hypotheses: A Reminder from My Kids

By Amy Labenz @ 2024-09-23T20:33 (+92)

I like the epistemic norm of having multiple working hypotheses, and I think it is a valuable norm/skill for EAs to cultivate. Last night two of my kids gave me a nice reminder that even if we value an epistemic norm, we don’t always remember to employ it.

I was putting Earnie and Teddy to bed when they suddenly claimed they were hungry. This is after I made “yummy noodles”—the dinner they requested—and had a bit of a hard time getting them to actually sit down to eat it. We’d also just done baths and brushed teeth, and to be honest, I kind of exhausted and ready for bed. My first thought was, "This is definitely a stalling tactic." 

We had a back and forth about it, and ultimately I decided to go ahead and get them something, even though I wasn’t fully convinced they were hungry. So I went ahead and made a little "midnight feast" picnic of graham crackers.

I opened a new box of the same crackers we'd had the day before. The kids had one each, then asked for "yesterday's crackers" instead. That cracker box was empty, and this was the same brand and flavor. I was losing my patience, and at this point, I was sure this was just another delay tactic. I confidently told them, "These are exactly the same crackers as yesterday. They taste the same!"

Both boys were adamant that they were not the same. I persisted and argued with them, explaining my view over and over again for a few minutes, getting increasingly frustrated with them. But then I thought, “Ok, what is going on here, I’m actually quite confused.” So I decided to taste a cracker myself. 

Plot twist: The crackers were terrible. They tasted stale or maybe even expired. I genuinely spit it out and couldn’t believe they ate the first one! I was so embarrassed.

This made me reflect on a few things:

  1. Even when we're really confident about something, it's important to consider alternative explanations. My kids weren't trying to manipulate me;[1] they were giving me accurate information that I initially dismissed.
  2. Our prior experiences (like kids stalling at bedtime) can sometimes lead us to jump to conclusions too quickly. This can happen in much more important situations too.
  3. Practicing epistemic humility in small, low-stakes situations like this one can help build the habit for when it really matters.

I think holding multiple hypotheses in mind is a great norm to promote, and I didn’t immediately find a post about it; I’d love a link if someone has one! [2]

I also think it is a good norm to say “oops” and flag when we don’t hold ourselves to our own epistemic standards, so I’d be curious to hear other people share examples of when they’ve had a similar surprise! Any tips for making sure to practice this in daily life/in higher stakes situations? 

 

  1. ^

     Note: it is possible that they were trying to manipulate me to stall bedtime and also the crackers were stale. 

  2. ^

    Thanks Peter McClurskey for reminding me to include Split and Commit!

Joseph Lemien @ 2024-09-24T23:55 (+22)

I really love seeing anecdotes of these kinds of principles applied in everyday, concrete, relatable situations. It helps illustrate the principles nicely, and gives an example that can be linked to (rather than linking someone to a 20-minute abstract essay).

Amy Labenz @ 2024-10-02T14:37 (+2)

Thank you! I'm glad. :)

I was feeling a bit bashful after Peter reminded me I forgot to include Split and Commit, but when I pinged Duncan (the author) he said the more reinventions of it the idea the better! That seems like a great attitude, since different things work well for different people. 

In some cases, these kinds of stories tend to be better for me too (even though I came to EA through the rationality community!). I think this is especially true when I'm explaining something in conversations. 

JenBrown @ 2024-09-25T11:23 (+11)

Thank you so much for posting this! I had a similar experience with my daughter last night. Stalling bedtime, I was getting frustrated, she was constantly asking to go to the toilet, which I eventually decided was just a tactic (because nothing was happening when she went). Turns out she did have a poorly tummy during the night...and I felt a little shameful. As my frustration built up, I lost sight of what else could be going on (she had her flu vaccine a couple of days ago). Fast forward to this morning where I apologised and smothered her in cuddles. Sometimes we need to breathe, resolve to go with the flow, and see the bigger picture.

Amy Labenz @ 2024-10-02T14:41 (+2)

Aw, so relatable! Thanks for sharing. And I love how you're modeling how parents apologize to kids when we get it wrong, that is really important to me with my kids.

I hope she is feeling better! 

Rebecca @ 2024-09-25T13:42 (+2)

As someone without kids, I'm a bit confused by this - why would it not be the default to e.g. ask her why she needs to go again from the second time onwards?

Kirsten @ 2024-09-25T19:53 (+5)

Kids can't always easily explain what's going on. Even adults would have a tough time answering that question I think! Most would probably just say "... because I need to use the bathroom??"

Rebecca @ 2024-09-26T13:58 (+2)

I would have thought ‘my tummy hurts’ would be fairly easy to articulate, but possibly I’m overrating the relevant abilities

Amy Labenz @ 2024-10-02T14:49 (+2)

I'm not sure if this is helpful, but the other night, my 3-year-old son Teddy decided to start screaming for milk after we were in bed. He already had two cups of milk when we were downstairs, and there was no more milk. I explained and asked if he was thirsty. He said he was. I gave him a bottle of water. He screamed, "I want miiiiiilk" over and over and rejected the water. Eventually, I just cuddled him in bed. He cried for about a minute once I cuddled him up, and he passed out.  It seems like he was overtired but also wanted to go downstairs and delay sleep for some reason? Or maybe he was just desperate for milk for some reason, even though he had a bunch. I'm not totally sure what was going on. 

We now have a deal that I will ask him about milk right before bed, but once we are upstairs with teeth brushed, I won't be going down for milk. Wish me luck!

Ulrik Horn @ 2024-09-26T04:09 (+1)

I think there is a lot going on. For one, kids are equipped to deceive but are not equipped to learn when to deceive. So to learn how deceit/trust works in social settings they have to experiment on their parents - you often catch your kids silently raiding the candy cupboard in the kitchen. So that kind of makes you skeptical of them. Then I think there is a cultural thing where it is ok not to trust your kids that much. This I feel less sure about the validity of, but I think it is a thing. If your boss would have as little trust in you as the average parent has in their child, I think it would feel terrible (I am not insinuating anything about the author here, this is just general observations of young families). I think there is more to it too (including language as Kirsten pointed out) and kids seem to cope well even though they often do not get much trust - maybe kids are emotionally equipped to deal with not being trusted that much - they are pretty happy go lucky. And I am sure there are many more aspects I am missing, like the parents being exhausted, etc.

Amy Labenz @ 2024-10-02T15:04 (+7)

My five-year-old Earnie generally doesn't lie. For example, if I ask, "Why is Teddy crying?" he will say, "I took the bike from him". He sometimes does "tricking," but usually that's something very silly and obvious. (Note: I can imagine looking back at this in a few months and being surprised. Things often go in phases with kids.)

I think he learned about deceit and trust, at least in part because I make a point of not lying to my kids. When they are skeptical of something I've said, I'll ask them, "Do I lie?" and they will reflect and say, "Mommy doesn't lie." 

Our 3-year-old seems not quite on the no-lying page yet (he seems to enjoy it and giggles about it sometimes). Recently, he has seemed to treat pinky promises with more reverence, but we'll see. 

Rebecca @ 2024-09-26T14:10 (+5)

I’m curious if you remember how you felt about these things as a child, or where you get this model of kids’ psychology from? as this was basically the opposite of my experience as a kid.

Ulrik Horn @ 2024-09-27T03:56 (+2)

I should maybe have made it clearer that I am speaking to a large part from my own experience. That said I have done lay reading of psychology and combined with quite consistent observations of multiple kids I feel quite sure about some claims (like the need to develop skill in when to deceit/how to build trust). Other claims I feel less certain about, like that they are more equipped than adults to not be trusted. I should probably have made it clearer how certain I was about each of the claims and what, if any, research or observations underpinned each of my claims. Thanks for holding me accountable on my epistemics! I also recognize after reading your comment that there is a lot of diversity in kids and perhaps there are kids that suffer a great deal from not being trusted, and kids that might not actually choose to engage much in deceitful behavior. And perhaps it reflects on my own failing as a parent haha!

Rebecca @ 2024-09-27T12:07 (+5)

Yeah the point about needing to learn the relevant skills/norms makes sense to me. I just feel nervous about assuming that because a kid doesn't seem to be negatively affected by their parents being suspicious of them, that they aren't - knowing how much of a negative effect from another person on your wellbeing and sense of self you ought to tolerate is also a thing to be learned.

Ian Turner @ 2024-09-26T06:28 (+2)

Good points!

If your boss would have as little trust in you as the average parent has in their child, I think it would feel terrible

On the other hand, if you lied to your boss as frequently and blatantly as most kids lie to their parents, you should expect not to keep your job for long. In other words, the trust may be lower, but the consequences/expectations are also lower.

PeterMcCluskey @ 2024-09-23T22:58 (+5)

Split and Commit seems like the standard post on this topic.

Amy Labenz @ 2024-10-02T14:31 (+2)

Thank you! Added a footnote. 

Ulrik Horn @ 2024-09-25T08:30 (+3)

Perhaps, if you can get people around you to help you, I think that is helpful. If you can get your kids into not just sticking to their opinions, but themselves helping you create experiments to get at the truth, that could help. If they just said "mom, try this before we continue arguing" right out of the gate perhaps you would have realized it was true. I try to tell this to people close to me "please help me understand when I am wrong - I am often convinced by experiments or good comparisons". A recent lesson from my own life was when someone said they were tired of cleaning up after me. We then figured out that the person who cleaned up after me felt as annoyed cleaning 1 of my things as 10 of their own. This immediately changed my behavior (for me I care less about whose things I clean than how long time it takes me in total so their perspective was completely alien to me - I could not guess that someone could feel that way). Basically, just give people around you (family, colleagues) a recipe that is easy to follow on how to convince you.