Summary: Some forms of meditation may train key skills of metacognition, serving as powerful tools for applied rationality. I expect aspiring rationalists to advance more quickly with a regular practice of mindfulness meditation.

The state of scientific research on meditation isn't great. Although there's evidence that it does something good--probably something involving down-regulation of negative affect--there are many basic questions1 that either haven't been studied at all or haven't been studied well enough to let me update much. According to a meta-analysis by Sedlmeier et al., one problem with evaluating the research is that it's hard to pin down what meditation is, let alone what it does or why it does it. In their words,

...two of our main findings are that (a) meditation has a substantial impact on psychological variables, indicated by a medium-sized (e.g., Cohen, 1988) global effect, and (b) its effects might be somewhat stronger for negative emotional than for cognitive variables. Due to the lack of a comprehensive theoretical approach (and results from studies derived therefrom), it is still unclear how meditation works... Moreover, a closer look at the studies included in the meta-analysis revealed that they differed in many respects that might have affected the results.2

So I just want to be clear that I don't mean in this post to wholeheartedly recommend daily meditation as the best possible use of 1/24th of your time.

Nevertheless, my own experience and reports from several of my friends suggest a specific cognitive result from a certain flavor of meditation that will be very good news for rationality if we can reliably reproduce it.3 In a recent post, Julia Galef pointed out exactly what I consider to be far and away the greatest benefit I've reaped from my meditative practices over the years. She wrote,

Meditation seems to train you to stop automatically identifying with all of your thoughts, so that, for example, when the thought "John’s a jerk" pops into your head, you don’t assume that John necessarily is a jerk. You take the thought as something your brain produced, which may or may not be true, and may or may not be useful -- and this ability to take a step back from your thoughts and reflect on them is arguably one of the building blocks of rationality.

I'd like to delve more deeply into how and why this could work. There seem to be multiple paths to establishing the central rationality skills comprising metacognition--several highly advanced rationalist I know have no background in meditation--so meditation is by no means a necessary condition for successfully applied rationality. I think it may, however, have the highest signal to noise ratio among methods for developing foundational metacognitive abilities. At a minimum, I expect that regular practice of certain kinds of meditation would help aspiring rationalists to advance more quickly.

What kind of metacognitive skills am I talking about?

How about an example. When you read these words, you're probably hearing a little voice inside your head that's reading them to you aloud, so to speak. Your relationship to this imaginary voice (aka "subvocalization" or "inner speech") may be quite a bit more intimate than you realize. It's likely with you not only when you read, but when you ride the bus home and think, "Maybe I'll have steak for dinner"; it's with you when you've had an awkward interaction with someone you admire and you think, "God, I must have looked like such an idiot"; it's with you most of the time, in fact, during your waking hours and maybe even when you dream.5

Exercise One

This fact may be more salient for you if you try to turn it off for a while. Set a timer for one minute, and force yourself not to verbally narrate your experience. When the minute is up, jot down a brief note about how it felt. Three two one go.

No, really, don't read the next paragraph 'til you've done the exercise.6

Even if you managed to go the entire minute without subvocalizing, it probably didn't feel like the natural way of things. It probably took effort, and possibly a great deal of effort. But I predict that most of you didn't go through the whole minute in total subjective silence. (If you succeeded and it did feel like the natural way of things, I'd very much like to know.)

Exercise Two

Now set the timer for a minute again, but this time don't force yourself not to subvocalize. Simply notice when words arise in consciousness. Don't bother doing anything with them. Just be aware of them.

Again, when the minute is up, make a brief note about how it felt. In particular, how did it differ from the first exercise, and how did it differ from your usual experience?

Exercise Three

Finally, notice that your present increased awareness of subvocalizations lets you change things about them that you couldn't change if you weren't aware of them. For example, you're now reading this in the voice of Morgan Freeman. (You're welcome.)

Now pick some other aspect of subvocalization to change--perhaps the accent, or the speed, or the pitch--and read the next sentence in that way. Set a timer for one minute, and experiment with things you can change about your experience of inner speech.

What does this have to do with rationality?

In general terms, what have you done in the above exercises? You've become aware of a mental process that usually runs in the background whether you like it or not. You've gained and exercised some degree of control over it. You've come up with and tested, in real time, alternative ways of running your own cognitive software.

Now, this has merely been a simple illustration. My point is not that swapping Morgan Freeman for yourself as official narrator would itself improve your daily life. (Although that may well be.) Rather, my point is that these skills are central to rationality and are cultivated by meditation. Those of you with a strong background in meditation probably did not learn anything important from these exercises, and wouldn't have regardless of your rationality training. Stepping back from your experiences in a way that lets you examine them and modify them is so old hat, if you meditate a lot, that you may even have forgotten what it's like not to have that action available as primitive.

This is extraordinarily valuable! There are three abilities that together form the bridge between knowledge of rationality and the application thereof. They are

  1. the ability to introspect and promote a sought cognitive process into consciousness
  2. the ability to not identify with any particular cognitive process you become aware of
  3. the ability to make changes to cognitive processes you’re aware of in media res

For example, even if you understand how important it is to make beliefs pay rent in anticipated experiences, actually doing that can be really hard. Why is it so hard? Possibly for a few different reasons, but prominent among them is the following. If you've thought something lots of times without ever explicitly identifying it as something you're thinking, without putting much distance between yourself and the thought, your sense of self gets tangled up in it. It's not nice to let go of something that close to you, even if it's useless or harmful. It feels sort of like trying to kick out your own child when you know you can no longer afford to take care of her--and it feels distinctly unlike taking a broken blender to the dump, which is closer to what should really be going on.

Other directly related examples include noticing and tending to confusion, actually behaving as though you might be wrong when you think you might be wrong, and thinking about politics without your head exploding.

Note that merely willing problems solved is not a reliable way of solving them. Resolving to not identify with your thoughts isn't the same as causing yourself to not identify with your thoughts. There's a reason you identify with your thoughts in the first place, and it's not because you decided to. If you don't alter any of the mechanisms that actually give rise to the problem, nothing will change--which is why, I think, it's possible to possess oodles of declarative knowledge about rationality without making a single significant improvement to your life.

One way or another, you have to get some distance between yourself and your thoughts and feelings if you want to let go of them or change them. That's exactly what meditation teaches you to do.

What does this have to do with meditation?

There are many kinds of meditation. Some involve intense concentration on very specific sensations, like visualizations of geometric patterns, repeated phrases called mantra, the breath, or movements (not all meditation is done seated and motionless). There are interpersonal types of meditation that can involve maintaining eye contact with someone for extended periods, imagining someone hurting and nurturing the desire to help them, or sex. The kind I'm most familiar with is a form of Japanese Buddhist meditation called shikantaza, which translates roughly to "just sitting". Although it comes with basically no instructions, as the name suggests, in practice it's nearly identical to the most general form of "mindfulness meditation".

Mindfulness is one of the most popular meditative practices in the West, and of the types I know about, it's the one I expect to be most relevant to applied rationality. Though all of the above, in one way or another, teach the backward step7 that allows you to stop identifying with thoughts, mindfulness is only that. Exercise two above is a limited form of mindfulness meditation. Although there's a whole family of practices that fall under the heading of "mindfulness", what they have in common is the cultivation of awareness.

All I mean by "cultivation of awareness" is the power to broaden/focus attention to encompass more things, or more specific things. I've often heard practitioners describe it as "openness to the world". Ordinarily, we experience a lot of things on which we don't bother to turn our subjective spotlights of attention, sometimes because they're just not important, and sometimes because we actively avoid stimuli we perceive to be aversive.

Subvocalization is an example. Other examples are the sounds in your external environment, what you know about how those around you are feeling, the sound of your own breath and heartbeat, the sensation of flinching away from a painful thought, the temperature in the room, the colors and shapes that appear behind closed eyelids, and the sensation of confusion. I find it difficult to describe the most general form of this, because without analogy to more specific forms, all I've got is that it's experiencing... what you experience. Which really just sounds like the default mode of living, doesn't it? But in practice it can feel very different.

When you're well practiced at noticing these things, at welcoming them into your attention, you're acutely aware of not being them. And when you don't feel as though you are your thoughts and feelings, it becomes emotionally easier to let go of them or to modify them. Changing your mind feels less like losing a part of yourself.

Further Resources

  • Sam Harris recently posted an excellent introduction to mindfulness meditation in the form of two audio tracks (one nine minutes, the other twenty-six). I recommend them pretty highly. They each guide you through a meditation session without any annoying religious or new-agey distractions.
  • I find this meta-analysis by Northoff et al. of neuroimaging studies of self-referential cognitive processing to be fascinating for all sorts of reasons. Chief among them is the light it sheds on how and why including clear-cut self/other distinctions in models of human minds doesn't always work so well. (Link is to a PDF.)
  • If you want to take a soaring leap across the bridge between knowledge of rationality and the application thereof, you simply must try a CFAR workshop. I did one of these back in April, and it was every bit as fun as it was effective (which was very).

Notes

1. Off the top of my head: What aspects of particular forms of meditation cause the various purported benefits? If we pinpoint those aspects, can we harness their corresponding benefits individually without committing to meditation as a whole? Can we improve upon them? Does meditation have different effects when practiced in a religious context? What is the relationship between meditation and hypnosis? How do the effects differ among different age groups? Does learning to meditate while young have any effect on adult meditation?

2. Sedlmeier et al. (2012). The Psychological Effects of Meditation: A Meta-Analysis. Psychological Bulletin, 138(6) 1139-1171.

3. I don't consider lack of supporting double-blind studies much evidence against my thesis, largely because the result in question would only show up in tests of metacognitive techniques I expect not to occur to the vast majority of researchers just yet.

4. In case you're wondering about my relevant background: I did Vinyasa yoga throughout high school, taught it during college, trained at a residential Soto Zen temple for a summer, have maintained a fairly regular practice of Zen meditation for about five years now, practiced Tai Chi (very casually) off and on for most of my life, and have a degree in religious studies with a focus on East Asian Buddhism.

5. Fun fact: You internally simulate your voice in parallel with actual talking.

6. Yes, you're doing it right. If you're trying to do it at all, you're doing it right. The idea is to find out what it feels like to make the effort, not to beat the game. There is no game. Some of the comments below have me concerned that I may be contributing to the "meditation means being brain dead" misconception. This exercise isn't meant to teach you the One True Way of Meditation. It's just to point at certain kinds of movements your mind makes. Monks who have been meditating for multiple hours a day for decades don't have completely featureless minds when they meditate. That isn't even close to what meditation means to them. Beginners are given exercises along these lines because it's an easier entry point, like training wheels. Eventually, counting breaths simply becomes irrelevant.

7. The Japanese “Su sube[karaku] mochi[iyo] eko-hensho no taiho o mochi-iyo,” from Dogen Zengi’s instructions for meditation (1227) literally translates to English (character-by-character) as “Remember/employ of backward step turning light/consciousness reflecting/illuminating.” (Dogen loved wordplay, and the double meanings are intentional.) Though most translators render this something like, “Take the backward step that turns your light inwardly to illuminate the self,” the characters for “self” and “inward” do not in fact appear in this part of the text. Thus, his central instruction for meditation is to step consciousness backward so it can be generally reflective. This is why I say that in practice shikantaza and mindfulness amount to the same thing.

New to LessWrong?

New Comment
62 comments, sorted by Click to highlight new comments since: Today at 2:58 PM

For people who fail to generalize from the exercises:

I think the more general point with these kinds of exercises is that you can't stop your mind from working on something, only change the object of attention, and that requires that you're conscious of what your mind is doing in the first place. When you tried not to subvocalize, you probably tried to focus your attention on other objects. If you weren't subvocalizing in the first place, you were probably already focusing your attention on something else.

This is why meditation usually anchors your mind on some relatively boring object. You can't stop the mental chatter, but you can mostly replace it by intensely focusing on your breath, for example.

Music, video games and subvocalization are poor anchors for attention, because they can hijack it completely. Breathing is boring so you have to be conscious of your attention to focus on it, and it also happens to be a tool you always carry with you, and it happens by itself.

I didn't link to other posts on Lesswrong about meditation in the OP for three reasons. First, most of them aren't related in especially useful ways (but perhaps I'm wrong; I did skim parts). Second, some of them make claims and express perspectives I'd feel a strong need to respond to if I sent people over to them, and I wanted to stick as closely as possible to my main point rather than talking a lot about meditation in general, which I could do for a very, very long time once I got going. Finally, I didn't want to prime people who hadn't already read them. I attempted to divorce meditation from many of its stereotypes so I could make points that didn't depend on the reader taking seriously whole memplexes deeply entangled with religious doctrine.

But now that all of that is out of the way, if you're interested in how this relates to past LW meditation posts, you can find the first of a three part series on meditation here, a discussion thread on minfulness in particular here, a post on theory here, a solid description of Vipassana here plus a followup here, a discussion of a study on mindfulness here, and something I haven't yet checked out about meditation for brainstorming here.

I'm quite new to the site and expect I've left out some interesting posts, so please leave a comment if you know of more.

I recently stumbled across Shinzen Young's stuff on meditation (scroll to the bottom of this page; you can also search his site; example) and was pleased to discover that it seemed way more lucid, concrete, and sensible than anything I'd read on the topic previously. I'm curious what your take is, if any.

One thing I've wondered about meditation: it seems like most/all meditation resources invoke some notion of an authoritative "guru", which reminds me of science in the medieval era where because Aristotle said something, it was assumed to be true (e.g. women have fewer teeth than men). I'm curious what it would look like if someone started exploring their brain in the same general sort of way that meditators do, but without being primed by specific meditation instructions, and keeping careful logs/observations of everything they tried (akin to performing actual scientific experiments rather than just trusting the "authorities", who got their knowledge from other "authorities", etc. until you get to someone living in the ancient era who actually came up with something original to do with one's brain). Here's a hasty list of ideas from me, rot13'd to prevent priming for others who want to brainstorm from scratch:

Vqrnf/dhrfgvbaf:

  • Qb ybgf bs 1-zvahgr zrqvgngvba rkcrevzragf, naq fcraq 30 frpbaqf erpbeqvat lbhe bofreingvbaf nsgre rnpu 1-zvahgr rkcrevzrag... jung'f gur cbvag va tbvat sbe ybatre guna n zvahgr?

  • Unir n gvzre gung orrcf rirel 10 frpbaqf gb erzvaq lbh gb ersbphf ba jungrire lbhe rkcrevzrag vf.

  • Vs lbh'er ernyyl gelvat gb qrirybc n arj jnl bs guvaxvat unovghnyyl, jbhyqa'g vg or orggre gb qb vg va gval yvggyr puhaxf guebhtubhg gur qnl naq gura tenqhnyyl rkcnaq gur ahzore naq fvmr bs gur puhaxf? R.t. qrirybc n unovg bs zrqvgngvat rirel gvzr lbh jrag gb gur onguebbz.

  • Jung jbhyq qryvorengr cenpgvpr sbe zrqvgngvba ybbx yvxr? Pna gur fxvyyf vaibyirq va zrqvgngvba or tenahenyvmrq yhxrcebt fglyr?

  • Gjb engvbanyrf V'ir urneq sbe jul zrqvgngvba vf vagrerfgvat arhebybtvpnyyl: vg vapernfrf cnenflzcngurgvp areibhf flfgrz npgvingvba naq erqhprf gur npgvivgl bs gur qrsnhyg argjbex. Ner gurfr begubtbany gb bar nabgure, v.r. pbhyq bar or gubhtugyrffyl nebhfrq, be erynkrq va zbfg culfvbybtvpny erfcrpgf, ohg jvgu enpvat gubhtugf? Jung bgure arhebybtvpny engvbanyrf zvtug rkvfg?

  • POG frrzf hfrshy va bgure nernf bs yvsr; jul abg ercrngrqyl fhoibpnyvmr vafgehpgvbaf gb barfrys juvyr zrqvgngvat? R.t. fnl gur jbeq "pnyz" bire naq bire ntnva gb barfrys. Vs nqiregvfref pna trg lbh gb ohl gurve cebqhpg whfg ol ercrngrq rkcbfher gb gur fnzr fybtna, frrzf yvxr lbh bhtug gb or noyr gb qb fbzrguvat fvzvyne jvgu zrqvgngvba, be nal pbtavgvir punatr lbh jnagrq gb znxr ernyyl.

  • Jul fubhyqa'g fgnaqneq pynffvpny pbaqvgvbavat nccyl gb zrqvgngvba? Sbe rknzcyr, V zrqvgngrq va bar gvzr naq pbagrkg naq graqrq gb trg ernyyl tbbq erfhygf, naq zrqvgngrq va nabgure gvzr naq pbagrkg naq graqrq gb trg ernyyl onq erfhygf. Znlor sbe rnpu fbeg bs zrqvgngvba rkrepvfr gung lbh svaq hfrshy, lbh fubhyq gel gb znxr vg n qvfgvapg xvaq bs cenpgvpr? Sbe rknzcyr, jrne tybirf naq pybfr lbhe rlrf sbe bar xvaq bs zrqvgngvba, naq chg ba terra-gvagrq yrafrf naq xrrc lbhe rlrf bcra sbe fbzr bgure xvaq.

  • "Crbcyr bsgra gnxr hc zrqvgngvba orpnhfr gurl jnag gb npuvrir be tnva fbzrguvat; gur cnenqbk va gur cenpgvpr vf gung gur orfg jnl gb trg “gurer” gb or shyyl cerfrag “urer.”" BX... naq n ebbzzngr bs zvaq onfvpnyyl gbyq zr ur jnf n ybg zber qevira naq uneq-jbexvat orsber ur fgnegrq zrqvgngvat n gba. Jung vs V jnag gb cerfreir zl inyhrf?

  • Bsgra zrqvgngvba rkrepvfrf jvyy gryy lbh abg gb gel gb qb cnegvphyne guvatf jvgu lbhe oenva. Vg frrzf jbegu qvfgvathvfuvat orgjrra gur tbnyf lbh ner gelvat gb npuvrir naq gur zrgubqf lbh'er hfvat gb nggrzcg gb npuvrir gurz. Znlor anviryl nggrzcgvat gb npuvrir n tbny vf n onq jnl gb npghnyyl npuvrir vg, naq n zrnaf/raqf qvfgvapgvba pbhyq or hfrshy.

Fcrpvsvp rkrepvfr vqrnf (abg arprffnevyl gung bevtvany)

  • Gel gb pbzr hc jvgu n gnkbabzl bs gur fbeg bs gubhtugf lbh guvax, be bgure jnlf bs haqrefgnaqvat gurz (r.t. jurgure gurl fcnjarq bss n cerivbhf gubhtug be abg). Gura gel gb ynory gurz va erny-gvzr.

  • Pbzr hc jvgu n dhvpx jnl gb qb pbvasyvcf. Rirel gvzr lbh unir n gubhtug, syvc n pbva. Vs vg'f urnqf, gel gb yrg gur gubhtug cnff. Vs vg'f gnvyf, tb nurnq naq sbyybj vg hc. Nygreangviryl, gel gb qrpvqr ubj hfrshy n gubhtug vf evtug nf lbh'er guvaxvat vg, naq ersvar lbhe vaghvgvba sbe ubj hfrshy gubhtugf ner sbe shegure rkcybengvba.

  • Frg n gvzre gung evatf crevbqvpnyyl. Jura vg qbrf, gel gb guvax bs n cnggrea va lbhe pbtavgvba bire gur cnfg zvahgr, naq jevgr vg qbja.

  • Gel gb znkvzvmr gur crepragntr bs lbhe oenva gung'f guvaxvat nobhg fbzr cnegvphyne guvat.

  • Gel gb guvax nf abezny, ohg znvagnva rzbgvbany qvfgnapr sebz lbhe gubhtugf, naq bayl pubbfr gb srry gubfr rzbgvbaf gung frrz cyrnfnag be hfrshy.

...which reminds me of science in the medieval era where because Aristotle said something, it was assumed to be true (e.g. women have fewer teeth than men).

Although the Problemata (which contains speculations about the teeth of humans and other animals) was indeed part of the Corpus Aristotelicum, it is very doubtful that Aristotle actually wrote it. In fact, many copies of it list the author as Pseudo-Aristotle.

Fair. But I do think the larger thesis, that learning from authorities is inferior to learning through experimentation and that we've been profitably trending from the first to the second over the course of human history, is true.

Anyone have opinions on James H. Austin's Zen and the Brain and Zen-Brain Reflections? Austin is a neurologist who practices Zen Buddhist meditation, and writes some quite involved neurological speculations about what's going on with the meditation. I started reading Zen and the Brain, but once it got to the serious neuroscience part it started to look like I needed to learn a bunch of brain anatomy to keep being able to follow, so I haven't continued with it yet.

I've read his newest book, "Meditating Selflessly: Practical Neural Zen", that seems to be aimed more at a layperson than "Zen and the Brain". It also talks a bit about his speculations about what meditation does in the brain, along with some recommendations on meditation. It might be too speculative though.

He also has a third book, perhaps that is a happy medium? Depending on how motivated you are, you might even try one of those Human Brain coloring books...

Even if you managed to go the entire minute without subvocalizing, it probably didn't feel like the natural way of things. It probably took effort, and possibly a great deal of effort. But I predict that most of you didn't get through the whole minute in total subjective silence. (If you succeeded and it did feel like the natural way of things, I'd very much like to know.)

Does playing wordless sounds in your head count? I can't quite shut off my inner audio stream at will, but I can concentrate on, say, a clock ticking, although instrumental music is much easier to focus on. Another time it turns off is when I'm playing a video game and frantically reacting to things without stopping to "think" about what I'm doing.

total subjective silence

Non-verbal/linguistic auditory simulation does not qualify as silence, no.

However, being able to bypass thought is indeed a useful skill, especially in meditative contexts.

Often, if I'm not thinking about anything in particular, my brain defaults to playing a song rather than doing nothing.

[-][anonymous]10y00

I think the more general point of the exercises is that you can't stop your mind from working on something, only change the object of attention, and that requires that you're conscious of what your mind is doing in the first place.

This is why meditation usually anchors your mind on some relatively boring object. You can't stop the mental chatter, but you can mostly replace it by intensely focusing on your breath, for example.

Music, video games and subvocalization are poor anchors for attention, because they can hijack it completely. Breathing is boring so you have to be conscious of your attention to focus on it, and it also happens to be a tool you always carry with you.

[This comment is no longer endorsed by its author]Reply

I like this post a lot, but I would like it more without the images, especially the first one. I find them disruptive.

Same. I'd add that the font/formatting is non-standard as well.

Yeah.

My solution to this is here.

Others have said this in person; I'll fix both things. Thanks for the feedback!

(I'm used to blogging for a very different audience with short attention spans, a desire for constant entertainment, and a great fear of large blocks of text.)

Okay, this is weird, but the first thing that popped into my head when you mentioned that there were images that used to be from this article was an image of a pony, vaguely Pinkie Pie looking. (being aware of cognition is weird)

I don't even watch My Little Pony or participate in its community. Now I'm starting to wonder if it has evolved into some sort of toxic meme which is replacing itself into generic forms of things.

Alternative data point, i quite like articles having images.

What were the images about?

Thanks. :)

Now pick some other aspect of subvocalization to change--perhaps the accent, or the speed, or the pitch--and read the next sentence in that way.

When I imagine a higher- or lower-pitched voice, I can feel my larynx actually moving up or down, whereas I can't detect any movement of tongue or lips if my subvocalization is at normal speaking speed or faster. Anyone else finds the same?

I had the same experience with the pitch changes, but when I altered speed I could feel my pulse change. It didn't feel good to think quickly.

I also tried switching the gender of my subvocalizations (male to female). It was incredibly disorienting, which makes me curious about how my mind distinguishes between 'thinking like a male' and 'thinking like a female'.

So on the one hand, abstract thinking improves your self control: http://www.ncbi.nlm.nih.gov/pubmed/19493322

On the other hand, abstract thinking leads to procrastination: http://pss.sagepub.com/content/19/12/1308

And vice versa for concrete thinking (lower self control but no procrastination).

But according to Piers Steel, procrastination is caused by giving in to impulses! Higher control SHOULD lead to lower procrastination, shouldn't it?

So the findings seem to contradict each other. How can you have more self control AND procrastinate more? And conversely, how can you be more impulsive AND procrastinate less? Is analysis paralysis a real thing?

I'd really like to hear some opinions on this apparent contradiction.

As an exercise of meditation this was new for me as I have avoided meditation until now (mostly because of its lack of theory and risk of self modification into unwanted directions). But these exercises had a clear focus and goal - so I tried.

Exercise 1

I didn't subvocalize much. After a few seconds of concentration on sleepiness (as a method of inducing inner silence and with closed eyes) the first subvoc was

"Hm, this is interes..."

cut off by

"Oh, I'm subvocalizing"

cut off by concentration on sleepiness again.

I felt patterns about the experience not yet named concepts but of the same quality as the mental content that activates when I mentally switch to a new topic. The required background concepts that are needed to deal with the topic (esp. for knowledge heavy topics).

I didn't let these concepts subvocalize. At least mostly so. My thoughts never formed more than a few words of incomplete sentences.

I'm not clear where the border between subvocalizing (which I think I do not do much consciously anyway) and 'thinking' in general is.

I didn't set a timer but I usually can guess a minute to a few seconds and I stopped a bit before this.

Exercise 2

I also closed my eyes but let my thought wander about this topic.

It was the typical stream of consciousness with the inner moderator which keeps track of the topic and refocuses the thought streams on the topic. I thought things like in this paragraph (only less formulated).

I cut this exercise short when it became clear that it was basically my normal mode of pre-sleep letting-thought-wander-mode and I couldn't figure out what might be differnt without going back to the post.

Exercise 3

I'm not clear about this exercise. Reflecting on text read is normal, or?

Reading text doesn't feel like subvocalizing to me.

If the meaning of the text read isn't immediately clear I may reread it more slowly and that does feel like subvocalizing and I then can observe the sentence as if someone has spoken it.

Changing pitch, speaker or speed of the text read is not unusal to me. It is required if you read books to children, because you have to edit them on the fly. But this editing has to be done faster than subvocalizing so I don't think reading is subvocalizing. At least it is not if you are skimming or cross-reading or doing the pre-editing needed for e.g. chilrens book reading

But maybe I am not a usual candiate for this as I am a fast reader (once able to cross read 25 pages a minute).

Interestingly if I read text from someone I am familiar with I can hear their voice when I read their sentences including their timing, chuckles etc. (at least a felt equivalence to that).

Summary

This was interesting experience and you post has given me some more insights into what meditation is and that I am using something like meditation on a self-tought basis.

When you read these words, you're probably hearing a little voice inside your head that's reading them to you aloud, so to speak. Your relationship to this imaginary voice (aka "subvocalization" or "inner speech") may be quite a bit more intimate than you realize.

I am one of those who do not subvocalize at all. In fact, it takes me some effort to start subvocalizing what I read. I had to work to create a clear and discernable train of thought, sometimes accompanied by "subvocalization", when working on my PhD, so that I can write the arguments down. It is still an effort, and deliberately doing that at bedtime is a reliable way to fall asleep for me, probably because it's so taxing.

I agree with the general importance of attending to one's thought process, of course, however it is done.

Interesting! You're the first person I've heard of who never used inner speech while reading in the first place. I know people who believe they've learned to stop, but there's also evidence that their vocal chords continue moving when they read just like everyone else's so something fishy is going on. On the other hand, I really doubt Eliezer's vocal chords are doing that, because he reads at least three times more quickly than I could possibly speak no matter how many times I'd rehearsed the words, yet he does claim to hear a little voice in his head as he reads. (How he parses it I have no idea. It's frightening how quickly his mind can move.) I was about to say, "I wish we could know by third-person science rather than mere reporting on phenomenology whether vocal chord movement reliably tracks inner speech," but if we could do that then vocal chord movement wouldn't matter since what I'm really interested is knowing whether inner speech is happening.

Edit: Not the study I had in mind, but evidence that subvocalization can be detected via electrical signals sent to vocal muscles. http://www2.hawaii.edu/~binsted/papers/BinstedJorgensenHICSS2006.pdf

On the other hand, I really doubt Eliezer's vocal chords are doing that, because he reads at least three times more quickly than I could possibly speak no matter how many times I'd rehearsed the words

I bet most people read much faster than they could possibly speak, and most of them subvocalize. One possible reason is they skip all the filler that doesn't help comprehension. I don't think I subvocalize whole words, and I skip many. Then there's skimming which is a completely different matter.

You're the first person I've heard of who never used inner speech while reading in the first place.

Or maybe I don't notice it somehow.

he reads at least three times more quickly than I could possibly speak

I naturally (without learning to speed-read) read non-taxing fiction at slightly over a page a minute, which is not very fast, but still faster than I could possibly subvocalize it. I thought this rate is pretty average, but maybe others can subvocalize faster. I do notice that I subvocalize when writing or when formulating my thoughts into words for further expression, just not when reading or listening.

Or maybe I don't notice it somehow.

Yes, this is exactly why I want to be able to directly test for inner speech without having to go through the vocal muscles.

.... There's something that's confusing me here. I notice that I am confused and am pointing this out to remind me to get back to it.

It's to do with the fact that, while whenever I consciously read I subvocalize, when I really get entranced by a book my subjective experience looks more like watching a movie, and I hear things only when people speak them.

It's also to do with the fact that I'm not sure what reading without subvocalization would look like; no matter how you take in an input string you're going to have a temporary variable with bits of the string lying around somewhere.

I am one of those who does not subvocalize at all.

Is there any 'sensory' process which seems to take its place, like a stream of imaginary images instead, or does it seem like other people have this extra thing going on which you don't?

When I read a fiction book or listen to a sound recording it feels like there is a direct connection between the words and my brain, with relevant images, sounds, feelings and ideas popping up inside. This flow tends to be interrupted when something in the story or in its delivery stops making sense, syntactically, semantically or logically. Then I have to disengage the default direct access method and go through the delivery medium to make sense of it. I don't know how common this is.

Non-fiction is completely different, there the medium is never transparent. Except for that one time when I first read an intro linear algebra text and it just made sense, with the images of lines, planes, rotations and intersections just showing up in the right places for many theorems, lemmas and calculations. I take it that this is a much more common feeling for those with aptitude for math, if not totally in the Will Hunting way.

Sounds familiar, it used to feel like that when I read a lot of fiction. These days I read maybe one or two fiction books per year, and my imagination is a lot bleaker.

How much fiction do you read?

[-][anonymous]10y00

Are you a particularly fast reader? Speed readers claim to unlearn subvocalization.

[This comment is no longer endorsed by its author]Reply

I am one of those who do not subvocalize at all.

That's probably good for you, I seem to recall that one the sub-goals in speed reading is to stop the subvocalization as it slows you down considerably.

I appear to be very attached to my subvocalizations: while I can stop "mid-sentence" it's very hard not to start, and when I'm trying to just pay attention nothing comes to mind until "I" "intentionally" think something.

Same here. (This isn't the first time I try to meditate but it's still not yet a habit for me.)

Awesome post! Well written and well organized.

I actually got myself enrolled in a mindfulness meditation study for precisely this reason. I'm 7 weeks into the 12-week study, and plan to post about it once it's finished. Ostensibly it's a study for people with social anxiety, which I have to a mild degree, but not to the point that it really affects my life at all–I enrolled in the study mainly because having homework logs to hand in increases my motivation to meditate half an hour a day. Beeminder is also helpful.

I'm 7 weeks into the 12-week study, and plan to post about it once it's finished

Please do, I'm looking forward to that.

Thanks!

Fascinating; I know just which study you're talking about, and I applied for it myself. Alas, I did not make it through screening. I await your post with excited anticipation!

Thank you for this. I don't know why it's been so hard for me to find anything this clear about what meditation even is. But today in my junior efforts at mindfulness mediation, I realized that I've been confused and annoyed about Sam Harris's insistence that somehow consciousness is it's own thing. I didn't totally discredit it, after all he's a life long practitioner, but it makes no sense to me. It occurred to me while sitting this morning that maybe this mysterious consciousness is just metacognition, which word had not even occurred to me before, and which thought led to an internet search and your article. Mr. Harris seems like a reasonable guy, so maybe I'll get to understanding what the hell he's talking about, but I've just read two of his books, and this article was by far more helpful. Maybe he is assuming I already knew the things you are outlining here. I'll listen and read more of him to see if this all checks out. Anyway, in appreciation

Thank you for this. I don't know why it's been so hard for me to find anything this clear about what meditation even is. But today in my junior efforts at mindfulness mediation, I realized that I've been confused and annoyed about Sam Harris's insistence that somehow consciousness is it's own thing. I didn't totally discredit it, after all he's a life long practitioner, but it makes no sense to me. It occurred to me while sitting this morning that maybe this mysterious consciousness is just metacognition, which word had not even occurred to me before, and which thought led to an internet search and your article. Mr. Harris seems like a reasonable guy, so maybe I'll get to understanding what the hell he's talking about, but I've just read two of his books, and this article was by far more helpful. Maybe he is assuming I already knew the things you are outlining here. I'll listen and read more of him to see if this all checks out. Anyway, in appreciation.

Upvoted.

N=1: Time to stop self-identifying with thoughts was less than 5 total hours of meditation practice (scattered across months, but still). This was especially helpful in diminishing neurotic behavior - the thoughts are still there just not engaged with.

Even if you managed to go the entire minute without subvocalizing, it probably didn't feel like the natural way of things

You gave the example: "Maybe I'll have steak for dinner" ...that sort of thought doesn't usually happen to me. I just feel hunger, followed by imaginary eating of whatever I am craving (I imagine the taste, how it will feel to chew, how my stomach will feel after having eaten, and stuff like that). These aren't full-blown imaginations - just little pieces of imagination (the same way that when you think of someone you don't vividly imagine their entire face and personality, but little pieces of it)

That said. I did sub-vocalize during your mediation exercise, but largely because your telling me not to implanted the suggestion. "Don't sub-vocalize" in particular kept sub-vocalizing into my head. Just you try not to think of an elephant for one minute.

If I'm subvocalizing, it means one of the following things:

1) I'm communicating (reading, writing, etc), recalling a communication, constructing a communication, or imagining a communication.

2) I'm formally applying a counter-intuitive logical operation (Counter-intuitive means: I can catch myself subvocalizing for the Wason selection task, but I can't catch myself subvocalizing for the human-customized "drinking" version. I also have to sub-vocalize for sums and products, despite having lots of practice.)

3) I'm trying to put things into categories and/or pin down very fine distinctions (although I suppose this is a logical operation, it feels different, perhaps because it's used more frequently)

4) I'm dealing with a concept which is extremely abstract: "Statistically significant", "Eukaryota" are examples of this (it's really hard for me to think of these words). "Association", "socioeconomic", "exponential", correlation:, or "complexity" are not examples of this. I'm not sure what makes something abstract enough to bring language into it. I suspect that it may be related to the extent that I don't fully understand the concept and am using the word as a placeholder.

5) I'm trying to commit something to memory. Also, if I come to some important insight, I'll put the effort into verbalizing it because often, if I don't verbalize a thought, I'll have difficulty accessing it later. I'll have the memory of having the thought, but I won't always be able to get at the thought itself.

6) I'm engaging in self-analysis: Sort of how like writing down your thoughts can help make them more explicit or clear, the act of verbalizing stuff can make stuff more explicit and clear for me. (This isn't really separate from the others - the faculties I mentioned earlier which are strengthened by verbalizing are helpful)

Oftentimes when I do sub-vocalize, it's not a complete sentence. It's just a word or so, interspersed here and there among sensations and emotions.

Verbalizing is good, but I also find it helpful not to always verbalize things immediately. While I'm talking I'm much more verbal, and I find that when I go back over the conversation in non-verbal format stuff that seemed important before ends up seeming like a silly distinction and questions get dissolved, and these mistakes are not made as often when I'm not thinking verbally in the first place.

Side information: I have an unusual language history. Frequent moving in childhood caused me to rapidly learn and forget multiple languages. I am unsure as to whether this is related to sub-vocalization frequency.

I actually doubt the extent that I do not sub-vocalize is particularly unusual. I'm guessing other commentators will shortly verify that they too, do not sub-vocalize frequently. This might be an individual differences thing?

Yeah, I felt a little uncomfortable about my presentation of the exercises for precisely this reason. It was a choice between giving instructions that might be insufficient to cause the desired effect--which is noticing of subvocalization, not subvocalization itself--and causing subvocalization that might not have been there had I a way to communicate "notice subvocalizations" without mentioning subvocalizations.

Still, I'm claiming something whose positive and negative versions are extremely difficult to verify. I'm claiming that subvocalization happens when you aren't paying attention to it, which means you're very unlikely to remember that it ever happened. It's easy to design an experiment to test this, but it would require some expensive equipment few of us have access to. My claim comes from the many many reports of meditators who are almost universally astonished to discover how much narration is going on in their head when all the distractions are removed--and they often aren't told explicitly about subvocalization. Soto Zen is notorious for simply telling students so sit down and shut up, basically. And then they experience this anyway.

Obviously, that doesn't mean everyone experiences it. Just that lots of people do. I'm banking on "lots" being "almost all" for the exercises to work.

Maybe switch exercises two and one? The first exercise could be "just sit down, don't try to think or not think...just take notice of what arises in your mind"

meditators who are almost universally astonished to discover how much narration is going on in their head when all the distractions are removed--and they often aren't told explicitly about subvocalization.

I too, experience an endless stream of thought. It's just that it is largely nonverbal thought and therefore not described by the term sub-vocalization. But it's still thought, and it is similarly problematic in its capacity for distraction (if not more so), and would probably benefit from meditation.

Sam Harris's earlier, introductory post (with added resources) on meditation can be found here.

For those interested in the subject, I would recommend checking out Mindfulness in plain english. Even reading introduction and chapter 1 might provide strong motivation for taking the exercises seriously. It is written by highly regarded buddhist monk but stripped of almost all religious/mythical references.

Set a timer for one minute, and force yourself not to verbally narrate your experience.

I stopped and tried to do this, but couldn't. How exactly do you do this? I haven't read onwards yet.

You mean you couldn't set the timer, or you couldn't not narrate?

The latter.

Ah. The point is to try for a whole minute to do it. I don't expect you to succeed. I can't always manage it myself.

It helps to be relaxed. Do this when you don't worry about anything. Set a timer to ten minutes or as much as you are willing to spend trying, so you also turn off worrying about whether you are not already spending too much time.

For me one obstacle was getting some good ideas while meditating; and then I was like: okay, this is a very good idea, it's more worth to remember it than to continue with this meditation exercise I am most likely doing wrong anyway. But then I realized it's short-term thinking: if I predictably get good ideas while trying to meditate (as if my brain is bribing me to stop), then I lose nothing by ignoring them and continuing to meditate; I can do the same thing to recall them (or other, equivalently good ideas) later. -- The idea is to realize that whatever temptation there is to stop meditating and do something else right now, it's not rational and you can safely ignore it without further analysis.

At first, you will verbally narrate. Just notice it and do a mental gesture equivalent of "thanks, but not interested now". No anger, no feelings of failure. Notice that the verbal thought is there, and stop paying it attention. If you want some benchmark, if you can notice and ignore each line of thought in an interval shorter than your breath cycle (which feels like a long enough time), you are doing it right -- if you keep doing this, the narration will stop after a few exercises. If you don't develop the thoughts, your brain will stop generating them. Don't worry; when you end meditation, everything goes back to normal. You don't stop a pattern with a lifetime of reinforcement by doing a short exercise. You just learn the skill of stopping the narration consciously for a moment.

So the first level of this exercise is to learn not continuing the thoughts that spontaneously start. Instead always decide to turn your attention to your breath (or whatever you use).

It's like: breath in... breath out... breath in... "I wonder whether I am doing this correctly, because I don't have any feedback and maybe I misinterpreted the instructions; perhaps I should read a book about meditation and -- oh, now I am thinking, but I'm not suposed to, hell how can I stop these thoughts, they come automatically, it's impossible, maybe I am doing it all wrong and -- yeah, I'm supposed to focus on breathing"... breath out... breath in... "oh, I am doing it correctly now, I must totally write a comment on LessWrong that I succeeded to stop my narration while meditating, it will be like: dear fellow LessWrongians, you won't believe me but today I tried meditating and it finally -- oh, I am doing it wrong again, am I not? back to breathing"... breath out... breath in... "did I set a timer? yes, I did. back to breathing"... breath out... breath in... "I wonder what's for dinner -- not now, back to breathing"... breath out... breath in... "also I could -- no, not now"... breath out... breath in... breath out... breath in... "oh, I seem to get it -- oh no, I actually failed again, that's just abusurdly difficult, I should probably stop now -- okay, I will continue until the timer"... breath out... breat in... "I'd like to see a movie -- not now"... breath out... breath in... breath out... "how about a book? no, not now"... breath in... breath out... breath in... "a computer game? -- eh, not now"... breath out... breath in... breath out... breath in... breath out... breath in... TIMER!

The important thing is to notice that there is a small improvement here; the focused breathing to thoughts ratio is improving. For me the most difficult thoughts to turn off are the meta-thoughts of "am I meditating correctly?". You have to discontinue those, too. After a long exercise it clicked to me. My first success was when travelling in a bus, where I spent about 30 minutes trying this, because I had nothing else to do anyway. I got it right only during the last 5 minutes. But then I could replicate the success much faster. (I am still not sure I am doing it completely correctly, but I don' worry about that too much. It's already nice to do something that seemed impossible previously.)

It's okay, the point is that it's incredibly difficult and most people can't do it.

Added: I am generalising from one example.

I find it incredibly difficult, and can't do it for a minute.

I would have said “attempt” rather than “force yourself” -- but at least this way you actually tried rather than just trying to try.

I did something similar to what CronoDAS mentioned -- I internally made nonsense sounds. It's not clear to me to what extent that should really count, though...

I didn't mean for this to be a challenge to conquer. If you're noticing internal dialogue or what's happening in your head as you try to prevent it, you're doing it just fine. The only way to do this wrong would be to read or to play a video game and pay no attention for the minute instead.