Thursday, October 27, 2016

Quantify Implicit Beliefs

Edit: I mostly don't endorse this post in retrospect. Sometimes I have to risk publishing posts I don't like/disagree with if I want to actually post anything at all. Edit edit: As gwern and others have pointed out, a great way to deal with this is (quoting gwern here): "elicit by relative rates or frequencies, and then normalize to probabilities. I think about all the possible options and say to myself 'the butler feels like he's twice as likely as the maid - but wait, the twin brother is super suspicious and must be at least three times as guilty as the butler!' then I write down 1/2/6, sum them to 9, and see that I think the maid must be 11%, the butler 22% and the twin brother is 66%."

It's sometimes said of murder mysteries that success is equal to the sum of the probabilities the reader intuitively assigns to the top suspects.

If the butler seems 20% likely to have done it, the cook 30% likely, and the maid 15% likely, and no other character is a plausible candidate for the crime, then your book has a score is a lousy 65%. But if the butler seems 60% likely, the cook 85%, and the maid 30%, then you knocked it out of the park with a score of 175%. You've made your readers hopelessly credulous, which is exactly your job as a storyteller.

I think the next step I should take toward reasoning with quantitative precision is to stop caring whether my probabilities sum to one.

There's this very familiar mental dance I go though when making predictions. First, I think, "what is the probability?" Then I think of alternative hypotheses, and assign probabilities to those. I notice the sum of my probabilities approaching one (or going past it), get uneasy, and check how many alternative hypotheses are left. When I find the answer is "a lot", I move a little toward revising the probabilities I've already stated, feel a combination of embarrassment and dissonance, and give up.

Or, occasionally, I go all the way through the dance, and end up with ad hoc probability assignments that make sense together but have nothing to do with the model I use to predict things in real life.

What I need is two separate procedures: one for gaining explicit knowledge of my implicit model, and another improving my model.

Gaining explicit knowledge of my implicit model should usually come first. The backtracking dance results from feeling that my automatic implicit models should already make sense. Which is absurd; I'm currently running on an ape.

Right now, it seems like there are two groups: 1) people who are good at this stuff and trot out their lovely, consistent numbers with confidence, and 2) people who are scared shitless of saying numbers out loud, because numbers lack plausible deniability, so if you say a dumb thing with numbers people will laugh at you.

I'd really like to see people publicly assigning greater than 100% on their first pass at probability assignments. Then they can walk us through what they think their brain is doing and why, and we can learn about the mistakes people are making and how to fix them.

"Right, so, I think a huge chunk of that 85% for the cook came from the scene where he seduces the pool boy. The pool boy scene was super salient, and I... might have gotten a little carried away. On further reflection, seducing the pool boy doesn't mean much for the case one way or the other. My new estimate is about 60% for the butler."

If I don't have a lot of training in quantitative probabilistic reasoning, and I end up giving about 100% probability to the top candidates in a really good murder mystery when asked to assign numbers, I probably need to ditch the pretentious formality and quantify the crazy mess I really feel.

Tuesday, September 6, 2016

How Studying Mnemonics Changed the Way I Learn

I have found myself repeating in conversation that studying mnemonics has changed my understanding of “learning”. I want to express what learning means to me now.

I’ll start with the basics, and lead you through my development. Say you want to remember the list “quail, Mars, roller coaster”.

The first mnemonic technique I learned was “linking”, and using that, I’d memorize the list thusly: First I imagine a quail. Then I link it to Mars by concocting an imagined scenario that includes both a quail and Mars. For example, perhaps the war god Mars, with his helmet and shield, impales the pet quail I’m holding and and waves it about on the end of his sword, leaving me scandalized, grief-stricken, and suddenly empty-handed.

Then I link “Mars” to “roller coaster” (sans quail). So now I and all the god-planets are lined up riding a roller coaster, and Mars has his arms in the air and is drunkenly bellowing Gustov Holst’s “Mars” at the top of his lungs as we go over the hills, to the dismay of the other planets (and myself).

After that, I might link “roller coaster” to “quail” so that I can start anywhere in the list and still retrieve all the items.

The linking technique is cute and sort of clever, and it’s handy for memorizing lists (which is in fact, on rare occasion, a worthwhile activity). And there are lots of other clever techniques that do similar things, all of which ultimately depend on linking.

But far more useful than linking per se proved my experience of failure to recall list items I’d attempted to link.

I failed a lot when I first started doing this.

Suppose that instead of the links I just described, I imagined a quail sitting beside the planet Mars. That’s still a link between “quail” and “Mars”, but it’s not a particularly memorable link, and there’s a good chance I’d drop it. I’d get to “quail” and go, “shit, I know the quail was sitting beside something, but what is it???” So then I'd go back, consult my written list, and make a new link between "quail" and "Mars", hoping this one would work better.

Just as knowing grammar and vocabulary isn’t enough to be an inspirational speaker, memorable thought is deeper than the tricks of the trade. It’s discovering that depth that has made all of this mnemonics stuff worthwhile for me.

As I’ve said in the past, I found I remembered more, and more easily, if my imagined scenarios were concrete, emotional, multi-sensory, vivid, dynamic, personally engaging, and story-like. “A quail beside the planet Mars” is at most one of those things. “Mars impaling my pet quail on a sword and waving it around while I cry about my loss” is all of them.

An even deeper mnemonic principle than the terms of System 1 language is conceptual confluence. Conceptual confluence is the way systems of concrete symbols can flow together with the structure and significance of the abstract concepts they represent.

If I’m memorizing a recipe, for example, rather than learning one long list of ingredients, it might be good to group all of the wet ingredients separately from the dry ingredients. (In case you’ve never baked from scratch: making a wet mixture and a dry mixture before combining wet and dry causes more even blending.)

I might also want to include the concepts “wet” and “dry” in the method of grouping. Perhaps the eggs, water, oil, and vanilla are all in a kiddie-sized blowup swimming pool, being sucked into a spinning vortex by a giant whisk. Meanwhile, the flour, baking soda, salt, and sugar are playing in a sandbox, when a spoon comes out of nowhere, buries them, and mixes them all together with the sand. Then a flood pours in from a kiddie pool and fills the sandbox with a sticky dough.

Now when I recall my recipe, rather than just having all the ingredients lined up on the table, it will be natural to put the wet and dry ingredients in separate bowls before mixing them together.

Additionally, the chunking of the information in my head corresponds to my external behaviors and experiences. As I go about my baking, associations with the relevant information lead me to access what I’ve stored exactly when I need it. “I seem to have laid out two bowls. Why two? Oh, because one is a pool, and the other’s a sandbox.” The mnemonic has a built-in trigger-action plan.

When I say that mnemonics has changed my understanding of learning, I mean that these deeper mnemonic principles have seeped into every educational thought I ever have. Any time I notice myself struggling even the slightest bit while trying to learn something - be it a skill or the point of a philosophical argument - I automatically reach for the structure and significance of the information, and try to express it in concrete, emotional, multi-sensory, vivid, dynamic, engaging, and story-like terms - just as if I had dropped a link.

For example, I was just reading about control theory, when I came upon this sentence in the fourth paragraph:

“The thing we're measuring is the input (to the controller), the level we want it to be at is the reference, the difference between those is the error, and the adjustment the control system makes is the output or feedback (sometimes we'll talk about the actuator as the physical means by which the controller emits its output).”

There are six central abstract concepts of control theory accompanied by seven terms, all of which are new to me, just in that one sentence. “That,” thought I, “is an Important Sentence. If I don’t make it a part of myself, the rest of this essay is going to be nonsense.”

Note that I am not interested in memorizing the terms. I don’t need to rattle them off in a hurry as a list. There will be no vocabulary test with bubbles darkened by a number two pencil. I do not need to memorize this sentence.

What matters is that I arrange some part of my mind into a coherent model of a “control system” - one that includes all the major components, their relationships to each other, and their impact on the system - with associations that will call the relevant parts to mind any time I see a word like “reference” in the text.

I paused, looked toward the meaning of the sentence, and began to associate.

I imagined a pot of heated water sitting atop a coal-powered stove. I’m holding a sensor that looks like a cross between a kazoo and a remote control for a television. I use an eye dropper to put a few drops into one end of the kazoo control, which happily gobbles up its input while making contented yum yum noises.

In my other hand, I am holding a large reference book, and I refer to a page on which is written a temperature that is the reference for this system. The control... defecates, I suppose, a bit of ticker tape out its other end, declaring the temperature of the input water. I compare the reference to the ticker tape number, and if they don’t match, I press a big red button on the kazoo control, which makes a grating “errrrrr, wrong!” buzzer sound, louder for greater differences between the numbers, indicating the error.

Startled by the sound of the button, a mechanical arm with an ax-like shovel on the end (the actuator) begins to act, shoveling coal, feeding it into the stove, grumbling as though it’s a bit put out by all this work it has to do (this behavior is the feedback or output). It shovels coal faster the louder the sound. The stove, exhilarated by its meal, burns hotter, returning the water's temperature to the reference written in my book.

Ok. At this point, I have explained control systems to the parts of my brain (and yours!) that actually matter for real learning.

I can tell because if that crucial sentence were suddenly deleted, I could compose something equivalent on my own, even if the lingo that comes out isn’t quite standard.

But I wouldn’t need to write down an equivalent sentence anyway, because my comprehension is beyond words now. The words have done their job, and I can leave them behind. They’re just triggers. I have built inside my mind a structure that directly supports further understanding of anything and everything about control systems.

If it turns out that there’s something wrong with my understanding of control systems, I’ll be able to notice because my control system will fail to behave the way it’s supposed to, and then I’ll adjust the structure. (…I’ll adjust it as much as is needed to align it with the reality of control systems, but will otherwise leave it be. Hey look, a control system! Yeah? Maybe? I guess I’ll find out when I read the rest of the essay.)

This is just how I think about things now, when my goals aren’t being effortlessly met. I think like this when I listen to people explain things, when I reason about problems, when I consider gaining new abilities, when I hear a poem that I want to fully experience, when I want to communicate with other people in a way that will help them hold onto the things that I tell them.

And yes, I can meet twenty people in five minutes and remember all of their names, if I want. Or recite a memorized speech. Or count cards. And that’s fun.

But it’s all silly party tricks compared to the deeper art of memorable thought.

Bonus problem:

To think memorably, you need to associate with abstract concepts in was that are concrete, emotional, multi-sensory, vivid, dynamic, personally engaging, and storylike; while encoding the structure and significance in your system of symbols; with links to experiences you’ll have when you most need to call on what you’ve learned.

What structure can you build in your mind that will support the application and development of your art of memorable thought?

Friday, August 12, 2016

How To Be My Guide Dog

There are a lot of things I need to be happy that feel unreasonable to me.

For example, it’s usually not possible for me to be comfortable in chaotic environments, like bus stations and sports bars. Or when the refrigerator is buzzing at the wrong pitch. Or when someone says words near me that require a response when I wasn’t prepared for language.

If it were just a few things, I imagine I’d be ok with it. But it’s a lot of things. They’re things that interact with nearly every aspect of daily life, imposing all sorts of constraints on my existence, and especially on socialization. Imagine being invited to meet friends you like a lot for lunch, but the place they’ve chosen has five car alarms going off inside. Or always having to speak a second language, any time you want to communicate in bodyspace, that you aren’t close to fluent in. Or a party where people whack each other with baseball bats at random. “What’s wrong, don’t you like parties???”

I’ve just spent a month doing a lot of traveling to visit family. As a result, I’ve had much less control over my experiences than I usually do. I’ve had little control over where I sleep, where I work, what I eat, who I talk to and when and how, what I hear, what I see, what I smell, etc. For a whole month. It's been... stressful.

So I’ve been thinking a lot about my responses to these things. By default, it seems I try to ignore or endure all this stuff. In the company of others especially, I try to behave as though everything is fine. “I’m going to run away because there is a light flashing in the window of the building next door” is not a thing I enjoy saying to people who don’t know me well. “Omg she’s such a prissy little princess.”

And I’ve noticed I’m really not being very agenty about any of this, partially because I’ve only recently recognized that it’s even a thing, and largely because I’m ashamed of my weaknesses. (And we all know that if you ignore something you’re ashamed of, it goes away.)

But I got to hang out with one of my friends who knows me well while I was visiting the Midwest, and he was proactive about helping me be comfortable. It was amazing. He did things like not talk to me while I was reading the menu, suggest that we go for a walk in a quiet park, and make simple decisions on my behalf when I was too overwhelmed to think.

I’ve come to think of it as “being my guide dog”, and I was immensely grateful.

I suspect I have a lot of friends who would be eager to help me in similar ways if only they knew how. One step toward a more agenty approach is to create affordances for those friends to make things easier for me when we’re together.

So I’ve compiled a list of behaviors that constitute “being my guide dog”. I’ll try to add to it over time as I learn more about myself.


Travel is the most stressful thing I do regularly. It tends to involve high-stakes decision making under time pressure in extremely chaotic environments while people try to talk to me and ask me questions. (Think of airport security, or a subway station with crucial announcements over a shitty loud speaker.) Intervening in this area is very high leverage for increasing my comfort.

  • Literally guide me around.
  • Meet me at the airport or bus station and tell me you'll take it from here.
  • Know how to get where we’re going.
  • Choose streets with fewer cars and less chaos.
  • Take my hand and decide when we cross the street.
  • Coordinate transportation for me.
  • Hail the Uber and identify it when it arrives.
  • Tell me where to be when, how to get there, and when to leave my house.
  • Prevent me from having to drive at night, or in the rain.
  • Prevent me from being rushed.
  • Offer to drive me.


I am almost literally never in an environment that makes me feel really safe. The safest environment I've ever been in was a Zen temple on top of a mountain in rural North Carolina during a silent retreat. "Make things more like a monastic retreat" is a good rule of thumb when crafting a Brienne-friendly environment.

  • Choose a calm and quiet meeting place.
  • Choose restaurants with no TVs or loud music.
  • Turn off the TVs.
  • Only play music that lacks lyrics.
  • Play rain and thunder sounds (you have no idea how soothing this is to me).
  • Take me to a lake, river, creek, or garden.
  • Pay the bill and send me a reimbursement request via Square.


Talking out loud is hard for me. I try to think the thoughts, and the feedback from my mouth and throat and ears distracts me. Parsing speech is also hard. So being asked a simple question can be a bit like, "Quick, what's 345 times 78?"

  • Speak slowly, clearly, and at a moderate volume.
  • Use a direct communication style (be blunt), to minimize the cognitive resources I have to spend on figuring out what you actually mean (odds are good I just won't spend them, and I'll fail to understand you entirely).
  • Avoid sarcasm and other humor that relies on awareness of deception.
  • Relax and be comfortable with silence (don’t talk just to fill silences).
  • Don’t interrupt me.
  • Don’t expect me to interrupt you when I want to get a word in. I'll get frustrated and just stop trying to talk.
  • If you want to show me text on your phone, read it to me instead of handing me your phone. That's probably a little counter-intuitive, but it's some kind of switching-modes thing that is really hard for me.
  • Ask me specific questions (less like “what have you been up to?”, more like “what poetry have you been learning?”).
  • Prevent me from having to make phone calls.
  • Don’t talk to me while I’m reading (as from a menu).
  • If I’m reading or have been silent for a long time, text me the first line of what you want to say (it gives me time to adjust, and gives me the option of responding via text). This is especially good for questions.

Group Contexts

People are chaotic. More people are more chaotic. I'm often debilitatingly overwhelmed in groups.

  • Speak to me only when nobody else is talking; don’t start a side conversation with me.
  • Don’t invite more people at the last minute.
  • Use hand signals to auto-moderate group conversations (ask me how if you’ve never done this).
  • Give me the end seat at the table so I hear fewer discussions at once.
  • If you're high status in the group or are otherwise the center of attention, set an example of conversing clearly and patiently.

Self Awareness

The more overwhelmed I am by external stimuli, the less likely I am to be aware of my own internal state. I'll also be less aware of your internal state, even though I'm perfectly capable of empathizing with you when I manage to devote resources to it.

  • Remind me I have noise-canceling earbuds.
  • Remind me I have night glasses for bright lights.
  • Remind me I can go somewhere else.
  • Ask me how I’m feeling.
  • Tell me how you're feeling and what you want, at the slightest provocation.
  • Ask me how I think you’re feeling. Connecting with you empathetically can actually be grounding, especially if you're feeling calm and confident.

To be clear, you don’t have to do any of these things when you’re hanging out with me. I'm not even making a request. I've gone my whole life without people catering to my pesky little preferences in every interaction, so I'm not depending on you to do this. I'm just providing an opportunity.

But if you’re actively looking for a way to make my life easier, doing any subset of these thins will help. If the spirit of the entire list appeals to you, you can tell me, “I want to be your guide dog” when you see me (or beforehand), and you will probably find that I’m much more relaxed around you.

Tuesday, August 2, 2016


It’s well past time for us to stop saying “irl” when we talk about the part of the world that our bodies occupy. Same for “in person”.

A few days ago, I was at a cafe, when the fifty-something stranger sitting beside me said, "Oh, you've got one of them fancy phones! ‘Smart phone’, right? I've been thinking about getting one, but I duuno if I'd be able to use it." I was a little startled to encounter someone who was unfamiliar with smart phones, but I didn’t think much of it.

Shortly thereafter, I had lunch with my brother and his girlfriend (both of whom are in their 20s). We were all visiting my hometown. She also had one of them fancy phones, and she was showing us how her followers had responded to the photos she’d posted of her visit.

I thought of the older man then, and the comparison filled me with warmth and transcendence. I became aware that there’s something wrong with the way I’ve been thinking of on-line interactions all this time.

“Birthform is not true shape. I am not some hairless ape,” as the saying goes. I’m information that happens to be encoded, for now, mostly in a squishy ape brain. But it’s the information that counts.

So this is me talking to you right now. Even though it's across time as well as space. Me. In real life. In person. Our togetherness is not somehow fake just because I'm not looking at you with my eyeballs and vibrating my vocal chords. I am with you more certainly than if our bodies silently shared space on the same bench while our minds moved elsewhere.

There are many people who see my body on a regular basis, but are far less familiar with the patterns of my mind than is someone who’s read a single Agenty Duck blog post. If you read my thoughts, then you know me, regardless of whether you’ve encountered my body in bodyspace, because I am those patterns.

And I can go so many places, and be together with so many people, while my body chills in an otherwise empty room. My keyboard is as much a part of my body as is my larynx, and Agenty Duck is as much a part of my home as is my kitchen.

When my friend took out her phone and showed us her Instagram photos, especially the one of the winery right by my mom’s house, I felt the presence of her followers in my little town. I felt the expansiveness of her augmented mind, how tremendously powerful she is compared to the man who dunno if he can use one of them fancy phones. She is something different. Something new.

So no more “irl”. No more “in person”.

We are bigger now, and our world is deeper. Let’s talk about “bodyspace”, denying neither the analog sensorium nor the digital realm. Let’s not slip into oppressive patterns of speech and thought that mask the extent of our reality.

Monday, July 25, 2016

On Becoming Poems

Committing a poem to memory is about becoming the poem. The words on the page are just blueprints. They aren’t the poem itself.

When I learn a poem, what I’m learning is a way of arranging my mind. The words suggest how to do that. If all goes well, the arrangement of my mind looks something like the mind of the poet.

I’ve been skimming articles on poetry memorization and recitation today, and I think everyone I’ve so far read is confused about what a “poem” is. They talk like a poem is a series of words, like it can go on a sheet of paper, and your brain is a xerox machine that makes a copy and files it away in a manila envelope labeled “poems”.

They advise that you to read the poem many times, write it out by hand, take it a couple lines a day, and recite it over and over to people and mirrors and dogs.

That is not learning poetry. That is copying blueprints.

Granted, I’m pretty sure that once you’ve recited it from memory enough times, the true form of the poem will begin to build itself in your mind, whether you like it or not. But humans make poor xerox machines, and this seems a terribly inefficient way to learn a poem.

Here is what it was like for me to learn “Sea Fever” the other day.

First of all, I read it, and halfway through realized I loved it and was going to commit it to memory.

So I reflected on what I’d just read, long enough to imagine myself a sea captain in love with the ocean. By “imagine myself a sea captain”, I mean that I invoked a character, in the same way I would if I were playing one on stage. If I reach to touch my face while preparing to recite this poem, I half expect to feel a tangled beard hanging from my chin.

Then I read the first phrase (in character): “I must go down to the sea again”. I look for the emotion behind those words, and once I’ve caught a glimpse, I amplify it. Craving, longing, adoration, desire, determination, resolve.

I feel the emotions with my body, in the way it makes me want to move and act, to position myself. In this case, I feel forward movement, reaching, the clenching of a fist, a tall firm stance, a nod of certainty.

When I speak the line, I fill my voice with those emotions, and I let the sound of the emotions resonate in my mind, amplifying them further.

I latch onto all the sensory information in the phrase, situating it in an imagined physical space. Here there is an image of the sea, and returning to it. I picture the ocean, feel the breeze, hear the crashing waves, and imagine myself walking toward it with eager steps of reunion and love.

I steep the scene in the emotions I identified before, tweaking it if it doesn’t quite mesh with them, until it seems a fitting illustration of the feeling.

Then I taste the music of the phrase, its rhythm and sounds, sweeping over the experience of the scene according to the cadence of the words.

This often modifies the bodily urges, as I’m drawn to gesture at the locations of the physical objects, or to imitate their movements, or emphasize a sound. If you watch me recite, you’ll see that I sort of dance my poems.

I think of all of this as “diving into” the phrase.

Then I read the next phrase, and read the first together with the second, and then the next, diving into each substructure until the string of phrases completes a coherent thought.

Once I have two coherent thoughts, I dive into the relationship between them, the transition points, the ways they fit together.

In Sea Fever, I learned the first stanza like so, diving into each structure in turn:

  1. I must go down to the sea again
  2. to the lonely sea and the sky
  3. I must go down to the sea again, to the lonely sea and the sky
  4. And all I ask
  5. And all I ask is a tall ship and a star to steer her by
  6. I must go down to the sea again, to the lonely sea and the sky/ and all I ask is a tall ship and a star to steer her by
  7. and the wheel’s kick
  8. and the wind’s song
  9. and the wheel’s kick and the wind’s song
  10. and the white sails shaking
  11. and the wheel’s kick and the wind’s song and the white sails shaking
  12. and a grey mist
  13. and a grey mist on the sea’s face
  14. and a grey dawn breaking
  15. and a grey mist on the sea’s face, and a grey dawn breaking
  16. and the wheel’s kick and the wind’s song and the white sails shaking/ and a grey mist on the sea’s face, and a grey dawn breaking
  17. I must go down to the sea again, to the lonely sea and the sky/ and all I ask is a tall ship and a star to steer her by;/ and the wheel’s kick [note: this is a transition point]
  18. I must go down to the sea again, to the lonely sea and the sky/ and all I ask is a tall ship, and a star to steer her by/ and the wheel’s kick and the wind’s song and the white sails shaking/ and a grey mist on the sea’s face, and a grey dawn breaking

How many times have I read the stanza by the time I’ve constructed its full form in my mind? How many repetitions does it take me to memorize a poem?

Well, it depends on how you count. In a sense, two: one pass to be inspired to learn it, and a second to actually learn it. But I’m not “reading” it in an ordinary way.

“How many repetitions?” is a little like asking “how many times do you have to read the manual to put together an Ikea desk?”

Just one, right? But you’re not reading it like a newspaper, you’re pausing at each step to follow the instructions, inserting tab A into slot b and so forth. When I construct a poem with my mind, I pause at each step to follow the instructions.

(Time-wise, Sea Fever took me between fifteen and twenty minutes to learn, while going for depth rather than speed.)

So by the time I’m done, I can walk through the poem as I’d walk through a building designed by an architect (…if buildings could experience themsleves). I’m not tracing out its blueprints. I’ve used the materials of my mind to construct the actual poem - in thoughts, emotions, sounds, smells, and movement - and now it’s part of who I am.

If you like poetry and you’ve never learned to recite a poem from memory by means other than rote memorization, I recommend trying this. It’s been revelatory for me.

If you’re looking for a place to start, Sea Fever is a great pick. It’s concrete multi-sensory imagery the whole way through, and it isn’t free verse.

I feel like I didn't understand what poetry was before I started doing this. I think I was looking at the blueprints of poems, and thinking some of them awfully pretty pictures.

Now I think poems are not things we read, but things we become.

Sea Fever by John Masefield

I must go down to the seas again, to the lonely sea and the sky,
And all I ask is a tall ship and a star to steer her by;
And the wheel’s kick and the wind’s song and the white sail’s shaking,
And a grey mist on the sea’s face, and a grey dawn breaking.

I must go down to the seas again, for the call of the running tide
Is a wild call and a clear call that may not be denied;
And all I ask is a windy day with the white clouds flying,
And the flung spray and the blown spume, and the sea-gulls crying.

I must go down to the seas again, to the vagrant gypsy life,
To the gull’s way and the whale’s way where the wind’s like a whetted knife;
And all I ask is a merry yarn from a laughing fellow-rover,
And quiet sleep and a sweet dream when the long trick’s over.

Sunday, June 5, 2016

Requests vs. Statements Of Desire

Me: I want a pony.

Eliezer: I’m not sure if we can afford a pony. How much does a pony cost? Maybe I can get you a pony if ponies are cheap on Craigslist.

Me: Thank you, but I wasn’t making a request. I was merely trying to create common knowledge about my desires.

Eliezer: I don’t see the distinction. When I respond to a request, I am responding to new information about what you want.

Me: No.

Eliezer: No? I think yes.

Me: No.

It’s important that we have different concepts for what I want independently of you, what you want independently of me, and what our joint extrapolated volition prescribes. If you respond to your belief about my belief about what our joint extrapolated volition prescribes, when I’ve intended only to provide evidence of my independent desires, then… problems.

Like, maybe I’ll say that I want a pony, because I wonder if you have insight into alternative ways of satisfying the need responsible for that desire, and you’ll try to get me a pony, when I don’t actually think it’s a good idea for us to have a pony.

Eliezer: Oh! Yes, that makes sense. That is a thing in Bayesian networks.

Me: Oh?

Eliezer: When stating your independent desires, you should begin by saying “lambda”. Refer to my independent desires with “pi”. And call the prescription of our joint extrapolated volition “BEL”.

Me: Do you want to explain to me where those terms come from?

Eliezer: *begins to explain belief propagation*

Me: Wait. That's not what I meant. Let me try that again.

Do you pi want to explain the thing?

Eliezer: *thinks* That sounds fun, so yes.

Me: Oh good! Then I BEL want you to explain the thing.

Eliezer: *explains belief propagation*

Tuesday, May 17, 2016

Beware the Bind

[Note: I’m not sure about my hypnosis terminology in this post. In fact I'm confident I'm using it weirdly. Almost all my experience is as subject, so I mostly just know what hypnotists do and how I respond, not what they call it or how they think about it.]

There's a suggestion technique called the hypnotic bind, which everyone heard a bunch when they were five. It looks something like, “Would you rather put away your toys now, or do you want to put them away after dinner?”

Consider what happens in a child's mind when they hear this.

They've been asked a question, so they're inclined to engage their attention in a search for an answer. But the search space for the answer is limited to the space of thoughts that assume they will clean up their toys at some point tonight.

Furthermore, the process of searching for an answer costs them attention, which limits their awareness of the broader desires they feel at the moment. (They want to keep coloring, and they don't want to put away their toys at all.)

So they say, "After dinner."

When this goes as planned, what they are aware of having just experienced is a weighing of options against their values, and then a decision among the options based on those values. When you experience the weighing of options followed by a decision based on your values, it feels a lot like you want whatever it is you've just chosen.

Used as a hypnotic technique, double binding is often about belief and perception of things besides choice. “Do you think you’ll fall deeply into trance now, or will you drift there more slowly as you listen to my words?” Either way, you’re attentive to whatever sensations are consistent with “going into trance”, which is over half of hypnosis right there.

(Wake up, hypnogeeks, that was just an example. I mean, unless you don’t want to. Would you rather enjoy my post from within trance, or is it just as fun to read from ordinary awareness? Or maybe you’ll love it most while mildly fractionated.)

Hypnotic binds don't have to take the either/or form, though. I often use single binding deliberately when I teach: When I pause for questions, I always ask, "What questions do you have?", and never "Are there any questions?"

Since students usually do have questions but often have trouble identifying them on command, directing their attention to the range of thoughts that assume they have questions saves them some work: It leaves more of their cognitive resources available for choosing among the questions that they have.

"Are there any questions?", by contrast, directs attention to the search space of "yes" and "no" - neither of which is itself a question! I always have trouble with this when someone asks me “any questions?”. “Welp, I see no questions in this search space, so I guess the answer is no.”

Binding is tricky. It's verbal sleight of hand. Sleight of mouth, if you will. And I've encountered it enough in hypnosis that I can sometimes pick out and notice the sensation of having just been hypnotically bound.

Sometimes this causes me to giggle unhelpfully in the middle of an induction. The hypnotist wants to create a floating arm effect, so they say, “As you relax more deeply, how much lighter does your arm begin to feel?”

And I think, “You crafty bastard! That directs my attention to sensations that are consistent with my arm already feeling light, decreasing my attention to sensations of heaviness!”

(Which doesn’t seem to prevent me from taking the suggestions, mostly.)

But it’s not just the verbal pattern I’m noticing when that happens. Among other effects of this suggestion is a feeling that presumably corresponds to my attention having been suddenly restricted to a smaller set of experiences, without an accompanying decision to focus my attention. It feels like something slipping, something incongruous, and there’s pressure in a direction, with a sense of unfamiliarity like the source of the pressure is external.

It’s very subtle, compared to the other things going on in my experience at that point. If I weren’t intensely curious about this sort of thing, I might never have noticed. But it’s there.

I’ve recently begun to notice inadvertent binding outside of the context of hypnosis, and I’m finding awareness of binding to be an important epistemic skill.

Which should not be surprising, in retrospect, because hypnotic binding is a way of deliberately inducing carefully crafted motivated cognition in another person, and I’ve long known “awareness of motivated cognition” to be an important epistemic skill. But “motivated cognition” comes in many forms; this is a special flavor of it, a non-central instance caused by someone else’s phrasing, and it’s usually extremely subtle.

Inadvertent binding has happened to me a few times in the past couple weeks, and it happened today.

I was talking on Facebook about the virtue of recklessness, and about how I approach difficult or dangerous things differently now than I used to, because three years ago, Eliezer observed that I was not failing often enough. So I updated.

Someone asked for concrete examples of things I've chosen to do because I made that update.

In response, I started listing things: Motivation characters, a week of “doing whatever I want”, formatting and publishing Eliezer’s novella, NaNoWriMo, trying to write a book on microrationality, falling in love with someone very dissimilar to me.

But as I listed, I felt a strange thing: Like something slipping, something incongruous, pressure in a direction with a sense of unfamiliarity as though its source is external. It felt like a hypnotist was messing with my perceptions through hypnotic binding.

The truth is that I don’t know which of my choices were caused by the update. It seems likely that I would have done a lot of the same things, or at least similar things, but my approach to trying things would have caused me to succeed more weakly when I did succeed, fail harder when I failed, and suffer more from my failures.

That answer - the truth - was not in the search space to which the question directed my attention.

The space of thoughts I was attending to was “things I’ve done in the past three years”. “I don’t know” is not a thing I’ve done in the past three years. Neither is “it’s more complicated than that”.

So I picked the least bad-sounding elements of the search space. It’s just like how “I don’t want to put my toys away” is neither “toys away now” nor “toys away later”, and “toys away later” is the least bad-sounding option in current awareness. “I just want to keep coloring” doesn’t cross the kid’s mind as a possible response.

Fortunately, in this case, I realized what had happened right after posting the comment, and was able to follow up with a correction. I’m sure I have failed at this many, many times in the past. It basically makes me lie, accidentally, in order to comply with the suggestion that I should have an answer of a certain type, or an answer at all.

I’ve sometimes felt a little worried when asking, “What are your questions?” while teaching a class. I’m worried about what I’m doing to the minds of people who don’t have any questions. Occasionally, I’ll respond to this discomfort by clumsily tacking on, “It’s ok if you don’t have any questions,” which explicitly suggests that they don’t have any questions! Which is the opposite of helpful for the people who struggle to identify the many important questions they do have.

On net, “What are your questions?” is probably best. I might even use the parental double bind, under some circumstances.

But if ever you find yourself listening to me, and I pause for questions, pay attention to what goes on in your head. See if you can feel yourself searching for the least bad element of the set of thoughts that might be questions, while neglecting all the other kinds of thoughts you could be having instead.

Even if it prevents you from identifying your questions, recognizing the sensation may empower you to escape inadvertent hypnotic binding later on.

So there was some stuff here you might not have encountered before - about hypnosis, or suggestion techniques, or phenomenology - and I’m sure I didn’t communicate all of it perfectly.

What are your questions?