11 • Attention and Insight
Dynamic, self-organizing feedback processes structure our attention. What can guide us to insight, can also lead us into self-deception.
Our intuitive understanding of attention is often based on a spotlight metaphor — shining a spotlight on something so that it stands out, and that’s what we are focused on. That model isn’t wrong, but leaves out some important details of how attention works.
Attention and self-deception
If I say pay attention to this finger, you can, and you can also choose to pay attention to some thing. […] There's two sides to attention:
You can direct your attention. For example, if I say, “your left big toe” you're paying attention to it, and suddenly it's salient to you. When you pay attention to something, it makes it more salient. It stands out for you.
But you know what else? Attention can not only be directed by you to make things more salient. Your attention can be caught. [claps hands] A sudden noise and you turn, and you attend to it. It was salient, and it captures your attention.
So not only can you direct your attention, your attention can be captured by what you find salient.
A positive feedback loop
Notice what this means you can do: You can direct your attention to something and make it more salient. And because it's more salient, it will tend to capture your attention. And because you're paying attention to it, you make it more salient, which means it will more likely capture your attention.
Do you see what's happening here? These two things feed on each other: I pay more attention to it, it becomes more salient. It becomes more salient, it gathers my attention. I pay more attention to it, I'm more likely to be attracted to it. And it spins on itself in a self-organizing manner until your attention is attached to something — it's super-salient to you, it's highly relevant to you. And you lose the capacity to notice other things.
That's how you bullshit yourself. The salience and the catchiness of the stimulus has overtaken any concern you have for whether or not it's true or represents reality. This is how you deceive yourself.
At the heart of attention lies a positive feedback loop: directing our attention makes things more salient, and salient things capture our attention. Feedback loops will accompany us throughout our journey of understanding insight. Feedback loops are the heart of self-organizing dynamic systems.
Such feedback loops can also be detrimental to our goals. Next time you’re endlessly scrolling on Facebook, Instagram, or TikTok, can you now understand a little bit better what’s happening? How your attention is constantly captured? And how the algorithms are tuned to become more and more successful at guessing what needs to be salient to you so that you keep scrolling?
Let’s deconstruct attention further.
A practical experiment
(For this one you might want to watch the video and follow along.)
It's an experiment you can sort of follow along with me. Let me describe it to you first. I need you to get some object, like a pencil or a pen. And we will call that your probe. Nothing untoward is meant by that, that's just what it's called in psychology. It doesn't involve any aliens doing graphic things to your body or anything like that.
[…] I'm going to ask you to find some object that you could put on a desk in front of you or hold in your hand. And then you're going to do the following — do not start yet because I want to describe it to you:
I'm going to ask you to tap on the object as if you were blind and you're trying to figure out what the object is. Its shape, its structure, its weight, its density. "Oh that's a cup.” […] You should close your eyes as you're doing this. I'm using touch because touch is slower than sight. And so you can become more aware of what's happening. Now, it's important while you do this, that you continue tapping.
I'm going to ask you in a moment to close your eyes, start the tapping, and then, while you're doing it, continue the tapping as you are following my instructions. And this will give you a sense of what you're doing.
Ok, so, what I want you to do is close your eyes. You start tapping on your object. Start tapping until you start to form an image of the object in your mind. Ok, so your eyes are closed, you're starting to get an image of what that object is in your mind. Ok. So right now you're aware, you're focally aware, what you're focusing your awareness on is the object.
I want you to keep tapping, but I want you to shift your awareness into your probe, feel how your pencil or your pen is moving around, shifting.
Keep tapping, and then I want you to shift your awareness into your fingers and feel how your fingers are moving around, shifting around. Some of you may be able to pick up on the individual feelings that are occurring in your fingers.
Now, go back, feel your fingers, your thumb and how they're moving.
Now, feel how the probe is moving.
And now, allow the tapping to reveal the object to you once again.
I've done this multiple, multiple times with people. And what's interesting is the following thing: most people find this very readily easy to do. And a couple of things, when you're initially tapping, for example, I was aware of my cup but then my awareness moves into my marker and then my awareness moves into my finger, and when my awareness is in my finger I'm not aware of the cup at all.
Then I was able to reverse it. I go from being aware of my fingers to being aware of the probe to being aware of the cup. And you're saying "What's all this about? What's going on?". Well, there's an important structure. Let's take a look at it step by step.
If you have ever tried mindfulness meditation or similar practices, what’s happening here may seem familiar. When focusing on your breath, for instance, this has a similar effect to shift your attention to something that usually happens automatically and is totally transparent to you, and bring it into your awareness to make it salient, make it opaque to you, so you can look at it, focus your attention on it, and investigate it.
Transparency-opacity shifting
Here's the cup (or whatever your object was), and I'm tapping on it with my probe. Now here's the interesting thing: It's not like I was completely unaware of my probe, because if I was completely unaware of it, I couldn't manipulate it. But I wasn't actually aware of it, I was aware through it. I was aware through my probe of the cup. I'm aware through this [probe] and I'm aware of this [cup]. It's like my probe is transparent to me. Let me give you an analogy right now, where this [probe] is opaque.
Here's the analogy — and we talked about this before, but let's do it again — and it's like my glasses are like my framing. My glasses are transparent to me in the sense that I'm looking through them, beyond them, by means of them. They're transparent to me. But what I can do is I can redirect my awareness, so that I'm now looking at my glasses rather than through them. My glasses have now become opaque to me. So I can do a transparency to opacity shift.
Now, what does that ability to shift indicate? This is part of [Michael] Polanyi's1 idea. Here's my probe, I'm aware through my probe. He has, what I call, a subsidiary or an implicit awareness because I'm aware through it — I'm not aware of it, I'm aware through it — of my focal object, for example my cup. And this I have a focal awareness or an explicit awareness.
Now his point, which is really quite good, is that attention is this kind of structuring phenomena. What it is, it's always attention as he says from-to — it’s an attention through subsidiary awareness into focal awareness. When I'm paying attention, I'm doing this. But here's the interesting thing: I was then able to step back and make this [probe] focal, and now […] I'm aware through my fingers of my probe. And then I can even step back and be aware of my feelings, what some people would call sensation.
I can keep stepping back and stepping back. So I'm looking at the cup through my probe. Now I'm looking through my fingers at the probe. And now I'm looking through my feelings at my fingers. And of course, the whole time I was actually looking at the cup I was doing all of that: I was looking through my feelings, through my fingers, through my probe, into the cup. And you see the [attention as] spotlight metaphor is missing all of that layered, recursive, dynamic structuring that's going on.
And notice you can move in both directions — you can do a transparency-opacity shift, in which I step back more and more into my mind, or I can go the opposite way. I can do an opacity to transparency shift. That's when you went the opposite way, that's when you go from looking at your fingers to looking through your fingers at your probe and going from looking at your probe to looking through your probe to the cup.
And your attention is doing that all the time, flowing in and out, doing transparency and opacity shifting. Now, that's very important because what you're seeing is how many different processes are being coordinated, integrated together to optimize and prioritize — to use an important term from [Sebastian] Watzl2 — this particular object or this particular scene or situation.
That's one way in which attention is operating. Now, for reasons I'm not quite sure of — I think it has to do something with we're using a visual metaphor in the way vision is oriented in our bodies — we tend to use an in-out metaphor for this. Like that's why I'm using stepping back and looking at as opposed to looking through.
Deep participatory integration
[…] Notice when I was, if you'll allow me, when I was knowing the cup through the probe, I'm indwelling the probe. I'm participating in how the probe is being with respect to the cup. I'm sort of indwelling it. I'm not knowing the probe, I'm knowing through the probe. I'm “inter-esse”. I'm so deeply interested that I'm actually right integrated with it and through it into the cup. The way my vision is integrated with these glass lenses so that I'm actually seeing through them and by means of them.
[…] This also works, not just with technology, but with psycho-technologies. […] You can so integrate literacy, for example, into your cognition that you don't look at literacy very much, you automatically look through it. […]
People talk about this metaphorically as moving in and out with their awareness. So one of the ways attention works is it moves in and out. You can look through a lot of processing deeply out into the world or you can step back and look at a lot of processing and withdraw towards the center of your mind.
We can have focal/explicit awareness and subsidiary/implicit awareness. And we can dynamically shift from transparency to opacity and back, moving in and out across several layers. If we pay attention, we can control what we look through and what we look at.
This includes tools and technologies. When you become one with a tool that you’re holding with your hand, you integrate it, indwell it, participate in its being, such that it becomes transparent to you, part of your subsidiary/implicit awareness.
And it includes psycho-technologies. You may not be focally/explicitly aware of reading these words, transparently using the psycho-technology of literacy to automatically make sense of this writing. But it is something that you had to learn before you got used to using it transparently all the time.
Feature-gestalt shifting
There's another important axis upon which your attention is working, and I can bring it out by a famous example: You give this [see video] to people and you ask them to read it. And they say, what does it say? And they'll say, "the cat". And they're like, "Oh yeah!" All right. And then you point out to them that they're reading this as an H, and they're reading this as an A, and these are exactly the same thing. Why are you reading one as an H and the other as an A? And so what they'll typically say to you is, "Well, because it fits in with this word as an H and it fits in with this word as an A."
Let's use language we've already developed. The letters are the features and the word is the gestalt, the overall structure. Now notice here: You've got a problem. It's almost a pseudo-Zen problem. In order to read the words, I must read each individual letter. But in order to disambiguate each letter, I must have read the whole word. Therefore reading is impossible.
Now of course reading isn't impossible, which means something else has to change. What has to change is your model of attention. The search light metaphor, the spotlight metaphor, can't address that problem.
Here's what your attention is actually doing: It's simultaneously going up from the features to the gestalt, the eidos, the structural-functional whole, and it's going down from the gestalt, the words, to the individual letters, the features. It's simultaneously doing that. Your attention is also doing this.
Not only is your attention flowing in and out, doing transparency-opacity shifting, it's also flowing up and down between feature and gestalt. Your attention is doing all of that, it's doing it right now. And the spotlight metaphor doesn't capture any of that.
And mindfulness has to do with making use of all of this complex, dynamical — remember what dynamical systems are — dynamical processing. These are dynamic, self-organizing processes, and they can be optimized. And mindfulness optimizes them in some way.
We are constantly and simultaneously shifting between features and gestalt, and between transparency and opacity, when we direct our attention, or when our attention is captured.
Attention is fundamentally dependent on dynamic, self-organizing processes which can be optimized. And manipulated.
Scaling your attention up and down
I'm going to put something up on the board. It looks like a graph but it's not a graph because it doesn't have absolute position. It's just a schema because it has relative position.
When I move this way [right to left] — like we were talking about when we're talking about Polanyi's work — I'm doing transparency to opacity shifting. And going [right to left] is to do transparency to opacity and to go [left to right] is to do opacity to transparency. It's not an absolute, no position is transparent and the other is opaque. It's always the direction that matters. The more I move [right to left], the more I'm stepping back and looking at. The more I go [left to right], the more I am indwelling and looking out into the world.
Then we have this: I can be going down from the gestalt to the features, using the word to decide the letters, for example. And I can be going up from the features to the gestalt. Nothing is inherently a feature. Look, the letters are a feature in the word but the word is a feature in the sentence. Nothing is absolutely a feature, it's always relative.
That's why I'm putting these double arrows. This is not a Cartesian graph. This is a schema. […] Although I can describe and you can understand these two axes independently, they're almost always operating in a highly dynamic, integrated fashion.
Very often, as I'm moving towards a gestalt — grabbing a bigger picture, I'm using that bigger pattern to look more deeply into the world. So, often I'm [moving up-right] — I'm grabbing bigger patterns and I'm using those deeper patterns to look deeper into the world.
This is what we do in science, for example. I find this and this and this, I get a pattern and then I find a way to integrate it together, and then I use that pattern to look more deeply in the world.
This is what [F = ma] is. I found a pattern and it allows me to look more deeply into the world. I'm no longer looking at these individual things — force, mass, and acceleration. I've integrated them together and that allows me to look more deeply into the world.
Often, when we're stepping back and looking at our minds, our awareness processes within attention, we're also often breaking up gestalt into features.
For example, you were breaking up your experience of your whole finger into individual sections of your finger when we were doing the experiment. You were breaking up the whole of the cup into individual moments of contact. So very often, these two come together. Let's call [up-right] scaling up of attention, and [down-left] scaling down of attention.
I find this model of attention incredibly useful, perhaps because I find so many parallels to what I do when I develop software.
Scaling down (zooming in) = focusing inward towards your self.
transparency → opacity: become aware of the tool/technology
gestalt → features: deconstruct the whole into its parts and their configuration
Scaling up (zooming out) = focusing outward towards the world.
opacity → transparency: become one with the tool/technology
features → gestalt: integrate parts into a coherent whole
As programmers, I’d argue, we explicitly train shifting from features to gestalt when we compose and abstract, and back from gestalt to features when we decompose and analyze. Many programmers also encounter a deep fluency of integrating certain tools and techniques such that they almost “become one” with them. That, however, often happens implicitly — we may not be aware when that happens. This model allows us to become aware of it and reflect on it.
Back to problem solving and insight
Remember that fact that you can mis-frame things? Let's do the 9-dot problem again! Join all 9 dots with four straight lines. And people find it difficult. Why? […] They automatically and unconsciously project a square there. And then they automatically take this to be a connect the dots problem and so no non-dot turns are possible. And therefore they can't get the solution. […]
The reason why people find that so difficult is, I have to break the square, and I have to not treat it as a typical connect the dot problem. I have to not treat it categorically, to use language you've heard already, because you don't do non-dot turns.
Step 1: break the inappropriate frame (scale down)
Now notice there's two moments to having an insight:
I have to break up an inappropriate frame.
What do I have to do? I have to break up the gestalt.And I also have to de-automatize my cognition.
I have to make it not operate unconsciously and automatically.How do I do that? I take stuff that's normally happening unconsciously, and I have to bring it back into consciousness. How do I do that? I do that by doing a transparency-opacity shift. Normally, I'm automatically sensing through my probe. But I can shift my awareness and become aware of my probe. I can bring things back into awareness. You de-automatize cognition by doing a transparency to opacity shift.
I break up the inappropriate frame, and I de-automatize my cognition by scaling down.
Chunk decomposition and constraint relaxation
Now interestingly enough there is lots of work by Knoblich and other people3, showing that you can improve people's ability to solve insight problems if you get them to do what's called chunk decomposition and constraint relaxation:
Chunk decomposition is just breaking up the gestalt. That's what chunk decomposition means.
Constraint relaxation is basically de-automatizing your cognition.
Scaling down helps you to break up the chunks, break up the gestalt, and helps you to de-automatize your cognition. But is that enough for insight? It's not enough.
Step 2: make a better frame (scale up)
Yes, I have to break up the inappropriate frame, but I have to make an alternative and better frame. I have to widen my field of awareness. I have to take stuff that was in the background and change its relevance. I have to look more deeply for deeper, broader patterns that I have not considered before. What do I have to do? In order to make a new frame, I have to scale up.
And we also have lots of independent evidence, having nothing to do with mindfulness meditation, that one of the ways you can improve people's ability to be insightful is if they have training or practice, or are naturally disposed to being able to scale up. If people can complete patterns in a kind of leaping that Ceci and Baker4 talked about, and other people. We can scale up in that way if we can take pictures that are out of focus and refocus them mentally, so we can suddenly see what the picture is. Again and again and again when people can scale up better, they're better at solving insight problems.
Beware of self-deception
So both make you better. But there's a problem: Because both also make you worse.
Because if I just scale up, if I just maximize, like tightening a string, then of course I immediately project the square and then I'm locked. Shouldn't I just scale down? Just meditate always, if I just keep breaking up gestalt, I'll never make the solution, I'll choke myself. That's what happens when people are choking.
If you're sparring with somebody, a way to get them off is to compliment them. "That was a really good right hook you just threw!" Because then the person will start stepping back and looking at it, and they'll get all screwed up. Because they'll break up the ability to generate the gestalt.
So notice what I'm saying, stick with me, because this is really sort of tricky:
[Scaling down] can improve your chances for insight by breaking up a bad frame. But it can also mess up your problem solving by causing you to choke.
[Scaling up] can improve your ability for insight by causing you to make a better frame. But this can also cause you to leap into an inappropriate frame and be locked in fixation.
So what should you do? You don't want the strings too tight. You don't want the strings too loose. And you don't want it just half way.
What you want to do is: You want to train people in both of these skills, and then train them to flow between them. It's called opponent processing. They're pulling and pushing on each other and so they're forced to coordinate and constantly get the right degree of attentional engagement that is most dynamically fitted to the world.
We can utilize our ability to direct our attention, to scale up and down, zoom in and out, to break inappropriate frames and make more appropriate ones.
We can look at both the problem in its context (the world) and our cognitive processing of it (our self) and then we can change our perspective on both of these things. And, perhaps, this will lead to insight.
But insight is not guaranteed. The same skill that can help us optimize for insight, can also get us stuck in a wrong frame or missing the forest for the trees. To apply it successfully, we need to cultivate this skill.
If you are new to this series, start here: A secular definition of sacredness
When we use a hammer to drive in a nail, we attend to both nail and hammer, but in a different way. We watch the effect of our strokes on the nail and try to wield the hammer so as to hit the nail most effectively. When we bring down the hammer we do not feel that its handle has struck our palm but that its head has struck the nail. Yet in a sense we are certainly alert to the feelings in our palm and the fingers that hold the hammer. They guide us in handling it effectively, and the degree of attention that we give to the nail is given to the same extent but in a different way to those feelings. The difference may be stated by saying that the latter are not, like the nail, objects of our attention, but instruments of it. They are not watched in themselves; we watch something else while keeping intensely aware of them. I have a subsidiary awareness of the feeling in the palm of my hand which is merged into a focal awareness of my driving the nail.
Polanyi, M., & Nye, M. J. (2015).
Personal Knowledge: Towards a Post-Critical Philosophy (Enlarged).
University of Chicago Press.
Watzl, S. (2017).
Structuring Mind: The Nature of Attention and how it Shapes Consciousness (1st ed.).
Oxford University Press.
Knoblich, G., Ohlsson, S., Haider, H., & Rhenius, D. (1999).
Constraint relaxation and chunk decomposition in insight problem solving.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 25(6), 1534–1555.
https://doi.org/10.1037/0278-7393.25.6.1534
Baker-Sennett, J., & Ceci, S. J. (1996).
Clue-Efficiency and Insight: Unveiling the Mystery of Inductive Leaps.
The Journal of Creative Behavior, 30(3), 153–172.
https://doi.org/10.1002/j.2162-6057.1996.tb00765.x