Understanding understanding

How do you know that you’ve been understood?

This question is one I think about a fair bit, and part of what motivated me to write the jamming/honing blog post.

If I’m saying something something really simple and hard to misunderstand, all I basically need to know is that the message was received and the listener isn’t confused. for example “Hey Carla, I left the envelope outside your room.” If Carla says “OK” then I can be pretty sure she’s understood. (Unless of course she misheard me saying something else reasonable.) A slight modification of this would be a situation where the information is straightforward but detailed—and the details matter. In these situations, often the entire message is recited verbatim. A classic example would be when a number is spoken over the phone, and the listener echoes each set of 3-4 digits.

Gauging understanding via re-generating

But when communicating something more complicated or nuanced, it’s usually not enough for the speaker to just get a “K” in response. If I’m trying to convey a model to you, one common way for us to verify that you’ve understood the model is for you to say something that you would be unlikely to be able to say if you hadn’t. This could take the form of explaining the model in a new way: “ahh, so it’s kind of like Xing except you Y instead of Z” or it could involve generating an example of something the model applies to.

I think we do this intuitively. Responding to an explanation with “K” potentially implies a lack of having engaged with the details. More like “You’ve said some things and I’m not arguing with them.”

On the International Space Station, the American astronauts would speak to the Russian cosmonauts in Russian, and the Russians would reply back in English (source). The principle is that it’s much easier to tell if someone has your language confused than it is to tell if you’ve correctly interpreted something in a foreign language.

Similarly, it may be much easier for me to tell that what you said was generated from a different understanding than the one I wanted to convey, than for you to tell if you’ve understood the thing I was trying to say… because the thing being communicated is “native” for me in the same way that most Americans have a much richer model of English than Russian.

Three ways this understanding-gauging process breaks down

1. One relatively common issue encountered in communicating is that someone doesn’t really speak out of their actual model of the thing… instead they’re just trying to say a phrase that will cause the other person to think they’ve understood. I think that this dynamic gets reinforced in contexts like school, where teachers may chastise students who profess a lack of understanding, yet happily accept vague responses that sound acceptable, so they can move on. (Related: “guessing the teacher’s password“) So it makes sense that people would develop that habit.

But it totally gets in the way of having a conversation! If you don’t understand, we need to know that. Or if you do, that too. Or maybe you got an interesting insight from what I said that wasn’t what I was trying to say. We want to be communicating from our actual understandings of things. But this requires a certain trust that (a) if I misunderstand you, that doesn’t mean I fucked up, and (b) if I misunderstand you, that doesn’t mean you fucked up. In other words, it requires a shared sense of communication as involving co-responsibility.

2. On the other end of the spectrum: I’ve noticed a pattern in myself of interrupting people when I’m not sure quite what they’re saying. For instance, a person would say something with two distinct interpretations, and I would find myself wanting to cut in and say “I’m not sure what you’re getting at with that… did you mean this thing or that thing?”

Sometimes this is worthwhile and important, but I’ve realized recently that often the confusion becomes resolved (or the distinction clearly irrelevant) when I just keep my mouth shut and my ears open for another few sentences.

I think that something about the compulsive need to ask for this kind of clarification actually represents a kind of unsafety with the conversation being not totally clear to me. In that sense, it’s sort of the opposite of the guessing the teacher’s password thing. “I don’t understand and I’m going to make sure you know it.”

3. On a totally different note: I was having a conversation with someone and the models just didn’t seem to line up or something. The person is cool/smart/interesting etc, which I knew from reading their blog/newsletter, but for some reason our communication was really weird on this dimension. I would say a thing, and they would say a thing back and… it wasn’t like they didn’t understand. They’re model wasn’t wrong, it was just that the thing they were saying didn’t seem to speak to the point of what I was saying… so it was like “well yeah, I guess? But…” And I genuinely don’t know where this dynamic comes from.

My best lead is something like… in the jamming/honing model, honing refers to narrowing in on something; but this requires having a shared sense of what it is you’re attempting to narrow in on. Or jamming involves exploring a space, which still requires having a shared sense of what the center of the exploration is. In this conversation, it seems we lacked a shared sense of what we were attempting to talk about, so we were doing neither jamming nor honing.

Other error surfaces

Illusion of Transparency: too few words

“Illusion of Transparency” is the term for the general human phenomenon of assuming that others understand what we’re saying very clearly, independently of how well they actually do. That if we think about something and speak what we’re thinking, others will understand.

At the CFAR Instructor Training event I went to a year ago, many of us realized that during presentations we weren’t very good at tuning into the audience and gauging how well they were following us. I had the excellent opportunity to take part in an exercise that several people invented to illustrate (and potentially train) this, which went as follows:

  1. Stand up in front of a small group of people with a 3-minute talk to give.
  2. Attempt to give the entire talk without speaking. You have flipchart paper, and you try to use as few written words as possible too (a few may be necessary for context or complex terms). This basically involves a lot of drawing pictures, pointing, making funny sounds, and pantomiming gestures.
  3. After a couple of minutes, restart the presentation, but this time speaking. Pay attention to what shifts.

During the phase without speaking, most presenters noticed that they were naturally inclined to check in with the audience every 5-10 seconds or so. This ranged from simple eye contact to holding up a thumb and getting people to indicate how well they were following.

But when permitted to speak again, they suddenly found it very easy to get lost in their own heads again.

Aside: I imagine that people using a sign language would fall somewhere in between; checking in perhaps less than a non-signer who is gesturing, but probably more than someone speaking… since sign language *is* more visual, even though it does have a grammar and is relatively precise

It seems that this suggests something important about the illusion of transparency. Because most of us have some pretty clear mental mappings from words to meanings, we assume that when we say words, all we need to check is “do these words resonate for me with my intended meaning?” If so, we assume we are communicating the thing. But since gestures and drawings have much fuzzier meanings, it becomes clear that the aim is not so much to “articulate what you mean” but rather to cause the other person to generate their own meaning out of what you’ve said.

Of course, this is true when using words too, we just forget because they sound so precise. Some logical positivists actually basically held a stance that each word has a precise meaning, with no context-sensitivity or nebulosity. But even if this were true, you still couldn’t assume that the people you were talking to knew all of the words you were using, except maybe in very specific contexts like a lecture where all of the students have already learned the relevant jargon.

Methinks thou dost explain too much

But sometimes you do need fewer words than you think. Either because the person literally already knows whatever it is you’re attempting to communicate, or because they can infer it from a relatively small amount of information.

Prior to writing the blog post about it, I went to explain my concept of Ultimatums in the Territory to some friends of mine, and I started by saying the name of the concept: “Ultimatums in the Territory.” They were like “Whoa, huh…!” Then I started into my first example story to explain it, and they both stopped me: “No, we get it. You don’t need to explain.”

This was very cool, but also quite disorienting. I was left with this weird sense of “but surely you haven’t fully understood, so let me say more things.”

I suspect that something similar is part of what happens during mansplaining: if the speaker (in this case the proverbial man) isn’t tuned in to the extent to which the listener already understands (or has yet picked up on) what’s there to be explained, they’ll keep explaining long past when that is needed or wanted. There are of course power dynamics at play in relation to why the speaker isn’t tuned in.

So the listener can end up feeling patronized or bored because while they wanted to talk about say developmental psychology, they didn’t need a refresher course on DevPsych 101 at this particular party. So instead of:

“…Kegan stages…”
“Kegan stages?”
“Ahh yes, well, [long rant about developmental psychology in abstract]”
“Right, I knew all that stuff. Where does Kegan fit in?”

…you could have something more like:

“…Kegan stages…”
“Kegan stages?”
“Ahh, yeah. Are you familiar with other developmental frameworks?”
“Well, Piaget had one for children.”
“Right, so it’s like that, except it extends to adults, uses a particular structure called ‘subject-object shifts’, and also many adults never reach the last stages.”

And then you’re into the meat of the conversation… perhaps. Be wary of thinking that you’re done with this checking-in process once you’ve done it at the start. It applies to the whole thing, fractally.

I’ve also sometimes been in a conversation where I’m going back and forth with someone but it feels we both already understand everything each other is saying. One experiment I like to do in this case is to suggest that we both be silent for a minute and imagine that the other person already knows everything we might want to tell them. And then what feels worth saying then? This can be a great way to jump to the edge of one’s conceptual comfort zone.

Explaining things without live feedback

Mostly, in this article, I’ve been talking about 1-on-1, relatively realtime communication. And I’ve made a bunch of suggestions about getting more attuned as conversationalists so as to more precisely hone in on what’s worth discussing, and more rapidly and robustly reach shared understanding.

But how can you do that when communicating in the form of pre-written text? (Or in a radio or video show.) How do I know, at this point, that you, the reader, have understood my thesis? I basically don’t. I’m like the speaker who isn’t looking at the audience at all. The only advantage of these mediums is that the reader/listener/watcher usually can usually go back and check something out if they know they don’t understand. But there’s still the matters of ensuring that there’s enough detail for them to grasp if they spend enough time with it, and providing ways for people to ascertain if they’ve understood and to navigate towards greater understanding.

This means that writers need to do some combination of:

  • making the post really long and thorough to ensure all questions or confusions are addressed
  • aiming at a group of people who they can trust have enough shared lexicon and context that it represents a relatively simple point for them\
  • writing multiple versions of something, aimed at different audiences. This is part of what Arbital was designed for (a great example: see the tabs at the top for different levels of explanation)
  • using verification questions or puzzles. These don’t necessarily ensure understanding, but they at least make it less likely for a reader to walk away thinking they’ve understood when they haven’t. An example for this blog post might be: “if you’ve understood this post, then it should now be clear why kids who ask questions in class learn better (even controlling for other factors)”.

(Apparently people have been using my “ultimatums in the territory” term to refer to self-destructive/manipulative ultimatums too, which is… precisely not the point. So maybe that blog post could have used such a verification check.)

Get it?

I sometimes wonder if there are a class of sentences that are incomprehensible to anyone who doesn’t already know the information they contain. Consider this jargon-filled sentence from Wikipedia:

The α subunit of an inhibitory G protein – Gi dissociates from the G protein, and associates with adenylyl cyclase.

Is there anyone that that sentence is both meaningful and non-obvious to?

On a more serious note, the tao. You know, the one that can’t be spoken of.

If you found this thought-provoking, I invite you to subscribe:    
About Malcolm

Constantly consciously expanding the boundaries of thoughtspace and actionspace.

Creator of Complice, a system for achieving your important goals.

Personal Website


SquirrelInHell » 8 Dec 2017 » Reply

I love and applaud your choice of question to consider. +150 points

An for “incomprehensible to anyone who doesn’t already know”, explaining non-fake meditation is a recurring (and quite frustrating) example.

Btw, first paragraph seems broken?

Doug S. » 8 Dec 2017 » Reply

I can sort of make sense out of that jargon sentence, but it lacks a lot of context. Apparently the way this one protein inhibits something is that a piece breaks off and interacts with this other thing, and the protein, the thing interacted with, and the piece that breaks off all have names, but I don’t know what the function of the thing being inhibited is, or why it needs to be inhibited.

Have your say!

Have your say!


Name *

Email *