Meditation, and Zen meditation in particular, is hard to teach in the sense that most people cannot learn to do it as intended by simply reading or hearing about it. Instead they need to attempt it, talk about what happened with someone more experienced, get feedback, and try again. I think this is because most models of how to meditate are implicit and complex, so written meditation instructions tend to leave out important subtle details.

Probably the most successful attempt to counter this dynamic and get all the necessary instructions in writing is The Mind Illuminated, which offers a book's worth of words and diagrams to teach you how to perform a particular style of Theravada concentration meditation. Aside from a lot of specific advice and answers to questions, it usefully contains a lot of explicit models built using fake frameworks to give the reader gears for understanding meditation.

Zen Training: Methods and Philosophy by Katsuki Sekida does something similar, but for Zen.

In this post I'll share some of Sekida's models that I think are especially useful to have if you're interested in Zen, Buddhism, or meditation more generally. I have to cover a couple sections of preliminaries to set those models up, though, so things don't really get going until the section on breathing.

Caveat lector:

Sekida's models as presented in the book have some limitations. In particular, I think he often gets the low-level details wrong when he tries to ground his models in science because much of his information was outdated even at the time of writing (1960s). Thankfully this ends up not being too much of an issue because the models don't actually rely on those details, which is to say Sekida is sometimes overzealous in trying to ground his models in scientific details that are not actually gears within his models.

Thus what follows is my take on his models. It's fairly faithful, but I've taken editorial license in places to either increase concision (this is a post rather than a book) or to cut out what I consider to be superfluous details. I've also added some commentary reflecting my understanding of how Sekida's models might connect with other ways of thinking about Zen practice.

Apologies if I end up misrepresenting Sekida.

Zen meditation

The primary meditative practice in Zen is called "zazen" which just means "seated meditation". Although Sekida touches on some other topics like koan practice and stages of maturity in Zen practice in the book, we'll focus only on his models as they relate to zazen.

The traditional, short instruction for how to do zazen is to "just sit" and do nothing else. It's a deceptively hard instruction to follow, but Sekida provides a nice way to see how it can work. He suggests a simple exercise of sitting still, holding your breath for as long as possible (he suggests trying for about 1 minute), and seeing what happens to your mind during this time when you invite it to rest. Most people, he posits, will notice that it becomes hard to maintain discursive thought and will find it replaced by a growing awareness of what one is experiencing. Assuming it works for you the way he hopes, this gives a taste of the kind of mental activity one is aiming to effect during zazen.

The practice of zazen is to develop samadhi. "Samadhi" is often translated as "concentration" but this is a potentially misleading translation since "concentration" here doesn't mean "applying effort to pay a lot of attention" so much as "reducing down into a concentrated form", even if some practices do require applying effort to pay attention. To avoid this confusion, I prefer translating "samadhi" as "gathering of attention" or "pulling the mind together". Translation aside, another way of describing samadhi is that it's what happens when the mind becomes unified around a single mental activity, which is definitely what Sekida means by samadhi as we'll see later.

He divides samadhi into two kinds: positive and absolute. Positive samadhi is like flow state and focused on something particular in the world. Absolute samadhi refers to a kind of relaxed but single-pointed attention we'll discuss in more detail below. For now it's just worth understanding that, while positive samadhi is useful and something to be cultivated for use when off the cushion, the primary activity of zazen is the generation of absolute samadhi.

Zen posture

Sekida views posture as essential for good meditation. We'll get more into why in the next section on breathing. He says the essence of good posture for zazen is to sit in a way that is stable, upright, and still.

A stable posture is balanced and is easy to maintain for a long period of time. The recommended forms are full lotus, "easy" Burmese posture (both legs on the ground), and seiza, although sitting on a chair also works for those who can't sit on a floor cushion. Notably, half lotus is recommended against because it pulls the body to one side, and sitting cross legged with the knees up is also recommended against but for reasons related to abdominal posture that we'll get to. What's important is to keep three points of contact with the ground/cushion via the butt and both knees forming a triangle of stability.

Speaking of cushions, stability requires sitting on a cushion, in Zen called a zafu. This is needed to get the hips above the knees, because if the hips are not above the knees one cannot sit upright.

What is upright? Something like sitting up straight, but in a way that is relaxed with the back naturally rested backwards, the butt outwards, the torso rotated forwards a bit over the hips, and the head tilted slightly down so the neck is a straight extension of the spine. This sort of posture is important because it causes the abdominal core to be held in an active position, which will become important when we talk about breathing.

Sitting stable and upright is also necessary to remain still during zazen. Sitting still is important because activity in the nervous system feeds information into the brain, and absolute samadhi only arises when brain activity become like a steady drum repeatedly beating out the same thought. We'll follow up more on this later after we talk about breathing.

Note: I've not talked about the full form for zazen here, just some essentials of posture. If you actually want to practice zazen, you'll probably want to learn the full form which includes many additional instructions.

Zen breathing

Besides stillness, which has its own value that we'll get to in the next section, the purpose of good posture for zazen is to enable meditative breathing that will generate samadhi. Thus, Sekida spends some time explaining how to perform this breathing once the right posture for it is attained.

Breathing for zazen starts from the hara, specifically the bottom of the abdominal cavity. Often this is described as located about 2 inches inside the body from the belly button, but because of slight variances in individual anatomy you might experience it more or less deep or higher or lower than the belly button. The actual physical location is less important than the location your proprioception reports.

It starts from there because meditative breathing in Zen uses the diaphragm to do all the work and as much as possible leaves other parts of the respiratory systems still. Notably this is as opposed to breathing by expanding the chest. The idea is to use the diaphragm to pull the lungs down to create negative pressure that pulls air in, then release the diaphragm to push air out.

This breathing should be steady and regular. At the start of meditation it will be "big" and will pull in a lot of air, but as meditation deepens it will naturally become "smaller" and pull less air. Beyond this, Sekida gives some recommendations about what steady and regular might look like, specifically recommending the bamboo breathing method. As he notes, though, ultimately everyone who meditates long enough develops their own personal style of breathing that arises naturally through long hours of zazen, so it's not so important to try to follow a specific method as it is to experiment and discover the form of breathing from the diaphragm that works best for the individual.

The reason for breathing this way, as Sekeida explains, is in order to regulate attention. He explains it with a diagram like this that gives us the first of two fake frameworks I'll talk about in this post:

The idea is that the nerves in the diaphragm are tightly connected to what he calls the "wakefulness center" of the brain (for a modern take on what the wakefulness center might be, I recommend this summary of Consciousness and the Brain). By breathing in a controlled way, using the diaphragm and focusing attention on proprioception of the hara, this causes the strong, regular pattern of the breath to nudge the wakefulness center into a steady pattern. This, he argues, causes attention more generally to adopt that pattern and causes thought to follow the same pattern. Thought can also influence breathing and can also regulate attention on its own, but because the brain is not already trained to maintain samadhi, the effect of breathing on attention is generally stronger.

What's interesting to me is that this model fits with some of the evidence we have about what happens during meditation and what it does to the brain. In particular, I'm thinking of what meditation looks like from the perspective of connectome-specific harmonic waves. It also suggests that what one does with the breath is in some ways more important than what one does with mental effort to control attention and that changes in attention to attain samadhi are primarily the result of regulated breathing creating a positive feedback loop between attention and breathing.

This also gives part of a possible explanation for why the common meditation instruction to let attention rest with the breath works. For the other part of the explanation, though, we need another fake framework about how thought arises in the mind.

Zen thinking

The word used to talk about mental activity in Zen is "nen", which is often translated as "thought impulse" but you might also describe it as a mental action or motion. It's not necessarily a fully formed thought, but rather the wisp of mental activity that might cohere into a thought or might pass without note or conscious attention.

Sekida classifies nen into three different types (first, second, and third) based on the way they relate to other nen. Let's consider them in turn.

The first nen is, as Sekida puts it, the outward looking mental action. By this he means the type of thought that arises from the senses, including the sense of memory, that is purely about its object and totally unaware of itself. The kind of thought that thinks "that's a cup" without any further awareness of who is seeing the cup, why it is there, or what it is doing. This is also the nen of felt sense prior to "having" an emotion.

The second nen is the reflective mental action that looks back on the first nen and contextualizes or evaluates it in some way. This is the nen that looks back and says "I see a cup" or "I like that cup" in response to "that's a cup". This is also the nen of having an emotion like anger and sadness or joy. The second nen is not, however, aware of itself, and in some ways function like a first-order logic over first nen.

Third nen, however, are aware of other nen, including other third nen. They reflect on thought and on the sequence of thought. They construct a narrative running through the mind. They say "I'm seeing a cup now and have been for the last few moments" or "I'm angry because of what he did".

Sekida gives some helpful visuals about how these nen types relate. I'll reproduce my own version of one of those here to illustrate how first, second, and third nen relate during normal, non-meditative mental activity.

As you can see, first nen feed into second nen into third nen, and third nen "look back" at previous first, second, and third nen with the strength of that look back decreasing with time. A more apt metaphor, though, would be that nen echo through attention for a while and overlay one another with older nen fading towards silence with each reverberation while new nen resound.

Sekida doesn't offer a neurological explanation to justify this model, only phenomenological reports that he and his students find that their minds appear this way to them under careful introspection. Despite this I think his model doesn't disagree with the neuroscience of attention and also fits surprisingly well with hierarchical predictive processing models of the brain. Again, as I suggest above, we'll primarily be using this as a fake framework to understand zazen and samadhi, so don't be surprised if it falls apart when trying to apply it to other domains.

Getting back to the topic of zazen, Sekida claims that placing attention on the breath causes third nen to become quite (to stop looking back so much), then for third nen to stop arising, and finally for second nen fall away leaving only a steady beat of first nen. The first part of the process where third nen become quite produces positive samadhi (flow state) focused on an object (proprioception of the process of breathing from the hara in zazen) and looks like this:

As breathing continues to regulate attention during zazen, third nen then stop being produced and the mind enters a kind of feed-forward state where only first and second nen arise. Then even the second nen fall away and there's only first nen beating out a steady rhythm as they sense the breath at the hara:

When this first-nen-only samadhi settles on a single nen being repeated continuously he calls this one-eon nen because subjectively it feels like the same thought stretches out indefinitely and all sense of time passing is lost. Zazen that produces one-eon nen is called "shikantaza" and is pointing to the same thing as shamatha (samatha, calm abiding) in Theravada traditions. Within the Soto Zen school, shikantaza is often viewed as the primary way one should practice Zen.


Reading the previous section, you might reasonably ask why you would want to practice shikantaza since you could interpret the above description as "zombification by self-hypnosis". Sekida says it's because shikantaza creates a build up of spiritual energy pressure that becomes like superheated water that will dramatically explode into a boil when perturbed and that explosion will push one to kensho.

Kensho means seeing into the nature of things as they are. Alternative terms include satori, insight, realization, awakening, and enlightenment. Note that kensho here is not necessarily a final kensho that, say, totally frees one from suffering (in Zen such a kensho is called "resolving the great matter"), and a person will likely experience kensho dozens of times if they persist in Zen practice.

Sekida's explanation of why this happens (a build up of spiritual energy) is obviously not a very useful model and at best a helpful metaphor for what it feels like to undergo kensho. Can we say anything more about what might be going on here?

Here's my take. Adopting a hierarchical predictive processing view of the brain, normally there's bottom-up evidence coming in from the senses and top-down models anticipating that evidence. Zazen works by shutting down feedback from top-down models, i.e. anticipation and judgement are suspended. Then, during zazen, bottom-up evidence has a chance to freely propagate up the network without being suppressed by feedback telling the senses they must be wrong because they are reporting something unanticipated. Then when the top-down models start anticipating evidence again there can be a moment of surprise (kensho) because while the top-down models were "off" they were able to update without dismissing minorly disconfirming evidence as they normally would.

An alternative explanation is offered via the theory of neural annealing. In terms of that model, I'd describe zazen as a process for "heating" the brain, and kensho as the moment when the brain "cools" and suddenly "hardens" into a new, more aligned state. Extending the metaphor, that would make Zen a forge for the mind.

Stuff I left out

Sekida has more to say about other of aspects of Zen practices. I've left discussion of what he says about practice with koans, the role of laughter, connections between Zen and the philosophy of Husserl and Heidegger, progression in Zen practice, and much besides.

I've also not talked a great deal about other models of meditation. In particular I've left out any mention of Kaj's great post offering a mechanistic model of meditation even though the aims are similar and his post proposes similar claims about what meditation does. Alas, I just couldn't find a way to naturally slip it into the body of the text above.

This post is also a very patchy guide to meditation and leaves out a lot of details, so if it seems confusing it's probably because I assumed you know more about meditation than you do. Any fault in that department is mine, as Sekida offers a more complete explanation in his book.

New Comment
10 comments, sorted by Click to highlight new comments since:

too much dullness: longer inhales shorter exhales, too much restlessness: invert.

Very nice! Thank you. I almost didn’t read it because from the title I expected it to be about something else. It would probably be helpful to explicitly say that this is a book summary in the post title.

I went back and forth on this. During the draft stage I originally had the title position the post as a book summary/review, but as I wrote the post it morphed into something that wasn't quite a book summary and made the decision to change the title because I thought the current one might better attract readers.

Based on your comments, though, I'll split the difference and update the title to be long with a mention that it's a book review.

I also considered suggesting a title that better captured the specific insights but couldn't come up with one (except for my specific takeaway, so I passed. 

Very interesting! I've been curious about Zen but hadn't heard of anything that would have anything like a gears-level model of it. 

The model of three different kinds of nens matches my own experience. I think that in my "mechanistic model" post that you reference, this bit 

Robot’s thought at time 1: It’s raining outside

Robot’s thought at time 2: Battery low

Robot’s thought at time 3: Technological unemployment protestors are outside

Robot’s thought at time 4: Battery low

Robot’s thought at time 5: I’m now recharging my battery

Robot’s thought at time 6: At time 1, there was the thought that [It’s raining outside]

would fit nicely into it. The thought at time 6 would be a third-level nen, and the rest first- or second-level nens depending on how exactly you interpret them.

Three characteristics: Impermanence also talks about nens, I believe:

That means that the content of your consciousness may be something like:

  • Time 1: The sight of a bird outside the window. [first-level nen]
  • Time 2: The thought “there’s a bird over there”. [second-level nen]
  • Time 3: The experience of typing on a keyboard. [first-level nen]
  • Time 4: The sound of a car outside. [first-level nen]
  • Time 5: A mental image of a car. [this could be first- or second-level I think]
  • Time 6: A sense of being someone who sees the bird and hears the car, while typing on a keyboard. [third-level nen]

… that is, normally you may experience there being a constant, permanent self which feels like what you really are. But in fact, during a large part of your conscious experience, that sense of self may simply not be there at all. Normally this might be impossible to detect due to what’s called the refrigerator light illusion: the light in a refrigerator turns on whenever you open the door, so it seems to you to always be on. Likewise, whenever you ask “do I experience a sense of self right now”, that question references and activates a self-schema, meaning that the answer is always “yes”. It is only by developing introspective awareness that records all mental content, without needing to make reference to a self, that you can come to notice the way in which your self constantly appears and disappears.

Rephrasing that quoted last paragraph a bit, it's saying that much of our consciousness consists of first- or second-level nens, but that we assume there to be a durable self which witnesses all experience (i.e. that our consciousness is all third-level nens); and that the existence of the lower-level nens can be difficult to notice because the mental motion of asking "what am I experiencing right now" intrinsically invokes a third-level nen. (But if the mind is quick enough in doing something like noting practice, it might start to notice that the nen that was returned in response to that query is not actually the nen of which the question was asking, and that the mental objects that the "response nens" are referring to, keep slipping away just a little too soon to be observed...)

Thanks, one thing I didn't quite understand in the post is Why making the attention more regular through diaghramatic breathing causes 2nd and 3rd nen to stop being produced (if indeed those two things are linked).

Does Sekida have a framework for why the other nen fall away?

To my mind jumps immidietly an explanation that higher tier nens need for the lower tier tiers to exist when they are formed. If the low tier nens are sproadic the "fertiliness" for higher tier nens stays constant. If they happen sycnhronised when they are plentiful they will overwhem the capability and skill to make higher nens and when they have all faded away simultanouesly they will starve high nen formation for having nothing to refer to.

Thinking this little bit over it also explains what kind of good is produced if the synronization state is actually pretty simple. Once you renable higher nen formation but have low amount of base nen to build from you are more likely to build every possible higher nen. In addition ordinarily it might be possible that the 3-nens are possible to form when 2-nens are active but have been on their way out for a long time already and the underlying nen-1s are already non-existent. Making the nen hierachy build fast might allow to have the nen-1 still around when highest nens form (ordinarily you can have a sensation or be aware that it happened but by shuffling things you can have the sensation and be aware of it, a type of lucid-wakedness).

He doesn't really.

As I recall he does make reference in the text that meditation decreases alpha waves and increases theta waves, but I'm not sure at the time there was any principled reason to think that was meaningful so if he's right it was in the absence of a model for why it would work.

My best guess though is that the theory of connecting specific harmonic waves might offer an explanation, but I think ultimately we really need better neuroscience to figure this out.

Insightful. As you anticipated, I didn't think that Shikantaza wouldn't be much different from a "zombi-state" and was very much positively surprised by your explanation of Kensho. I will probably experiment in that direction.

Thanks! Yeah, when I was reading this book I thought this was an important insight worth sharing from it and actually moved me to want to write this post because, without this model, it's hard to understand why I would want to get really good at pretending to be a rock.

My personal journey with this has been first to try to follow instructions like the above because I had so much restlessness it was hard to stay seated for an entire meditation period and so I used techniques as a way to fight myself into staying on the cushion. Then I had a kensho experience that left me able to effortlessly meditate, but this has lead to a lot of time on the cushion that wasn't bad but also wasn't very effortful and probably wasted a lot of time getting settled rather than going straight to the intended practice. Now I find I'm motivated again to apply effort since I have some reason to believe it will do something other than to simply do the practice being handed down to me.