The Mind Science of Practice

There is a version of this material that circulates as prose, a few paragraphs on why belief matters, an anecdote about a ritual that worked. It is pleasant and it is almost useless, because it does not tell you which parts of the claim are supported and which are not. The alternative is a table. Four mechanisms have been studied well enough to be worth putting in a column, and for each of them you can say what it covers, what it does not cover, and roughly how large the effect is. That last column is the one everybody skips and it is the one that decides how much weight the row can bear. A table like this is not scepticism. It is the minimum equipment for making your own claims honestly, and once you have it, the arguments with other people get much shorter because both of you are looking at the same cells. The rows below are the ones that matter for this field, in the order they tend to be reached for, and the exercise at the end is to build the same table for a claim of your own.

What the table is for

A table of mechanisms is not a table of truths. It is a way of keeping four different questions from collapsing into one another, which is the failure mode of almost every discussion of this subject. The questions are: is there a physiological change, is there a change in reported experience, is there a change in behaviour, and is there a change in the world outside the person. Mechanisms in this field mostly serve the first three. Almost nothing anyone has ever done serves the fourth in a way that survived checking. Keeping those four apart in writing, one per line, is most of what a reference sheet is for, and it makes the entire conversation considerably more honest with very little effort.

The second use is that it forces you to be specific about size. Most claims about ritual are stated without magnitude, which makes them impossible to assess and easy to defend. A ritual that reliably reduces your heart rate variability before a difficult conversation is doing something modest and real. A ritual that reliably brings you money is doing something that has never been shown to happen. Stating the size in the same units for both is what makes the comparison visible, and the comparison is the whole argument. Nobody needs to think this is a scam or a religion to see the difference.

Open-label placebo, in one row

Open-label studies are the ones where participants are told what they are receiving. The finding that matters is that knowing the intervention is a placebo does not always remove the effect. In some conditions, telling people they are receiving a placebo and that it may still help produces measurable physiological change, particularly in conditions where the outcome is subjective, like pain or nausea. This complicates the simplest version of the placebo account, which says belief in the treatment is what does the work. The open-label data suggests the mechanism is not belief in the treatment but something more like the act of being enrolled in something that will be attended to.

The row's honest limits are important. This is a real effect and it is small to moderate, and it operates most reliably on outcomes that are measured by the person experiencing them. It does not produce changes in objective external measures such as infection markers or tumour size. So it covers a genuine mechanism for a real class of effect, which is more than most practitioners will admit and less than the people selling things will claim. What it does not cover is the claim that an outcome is being produced by something acting outside the person, which is a separate claim with a separate evidentiary history.

Predictive processing, in one row

Predictive processing is a family of accounts in cognitive science that treats perception as hypothesis management. The brain is described as generating predictions about incoming sensory data and then correcting itself when the data does not match. This is a well-developed and actively debated framework, and the version relevant here is what it implies about expectation: if perception is partly a matter of which hypothesis is being tested, then changing what you expect will change what you notice, without requiring anyone to be lying to themselves. That is a mechanical account of why expectancy changes experience, and it is a much more modest claim than the one that expectancy changes the world.

The row's honest limits are that predictive processing is a framework rather than a finding, and that its application to any particular ritual is an inference. It explains a mechanism by which expectation could colour experience, which is genuinely useful, and it does not explain why any particular symbol or sequence would carry a particular expectation. Practitioners who use this row well use it to explain why their practice works on them, and use it as a reason to be careful about their claims about anyone else. Practitioners who use it badly use it to explain away the fact that they were wrong.

The autonomic change, in one row

This is the best-evidenced row in the table, and the one least often used, which is a strange state of affairs. Repeating a fixed sequence of actions, with attention on the breath and the body, reliably produces measurable autonomic change: reduced heart rate variability, slower and more even breathing, changes in skin conductance. This has been demonstrated in controlled settings and it is not in dispute. The significance for this field is that when someone reports that a ritual settles them before a difficult situation, something has happened, and it is not a metaphor.

The honest limits are that this is a change in your own nervous system, that it is well within the range of what breathing exercises achieve, and that it has nothing to say about anything external. It is also the row most vulnerable to overreach in the other direction, because a practitioner who has established that their practice reliably changes their autonomic state can easily start describing that as the practice working, when the defensible statement is that the practice reliably changes their state. The distinction sounds pedantic and it is the difference between a claim you can repeat for years and one that unravels the first time somebody measures it.

The card deck, in one row

The Zener cards from the mid-century research programme into extra-sensory perception are five symbols, designed so that a guess has a one-in-five chance of being right. That design is the entire point and it is the part that gets forgotten. With five options and no information, you hit twenty per cent of guesses, and a group guessing for a whole session will land close to that. The dataset from those studies, once the scoring and selection procedures are examined, has not held up as evidence of anything beyond chance, and the reason is generally described as methodological rather than fraudulent, which is a more instructive kind of failure.

What the deck actually teaches is arithmetic, and the lesson transfers to every other claim in the field. Once you have done the twenty per cent calculation by hand, you have a reference point for what chance looks like in a guessing task. Then any claim that someone hit thirty-four per cent on forty calls looks different from what it did before. The card deck is the cheapest piece of equipment in this whole area and it does more for your judgement than most of the books. Keep it. Use it once a year on your own practice to check whether your hit rate has moved, and expect the answer to be no.

What each row does not cover

Read the coverage column of each row together and a pattern appears. The open-label row covers subjective outcomes and requires enrolment. The predictive processing row covers the colouring of perception and is a framework, not a finding. The autonomic row covers your own nervous system and does not require any belief at all. The card deck row covers the arithmetic of guessing and covers nothing about the practice itself. Three of the four rows are about changes in the practitioner, and the fourth is a measuring instrument. None of them covers the claim that a practice reaches outside the person.

That last claim is not in the table because it is not supported, and the honest way to present the table is with a fifth row reading not supported and nothing to size. This is where most tables of this kind get quietly dishonest, by omission rather than by lie. A reader given four well-sourced rows and no fifth will infer that the fifth is forthcoming. Putting the empty row in, with the reason it is empty, converts the table from a list of supports into an actual account of what is being claimed, and it is the single change that makes the reference worth carrying.

Building the table for a claim of your own

The exercise is the point of the whole reference. Take a claim about your own practice, something you actually believe, and put it in the first column. Then walk the rest of the columns. What is the mechanism, named? What outcome does it cover, subjective or objective, and in whose body? How large is the effect, in what units? What would falsify it? If you cannot answer all five without hedging, you have found the part of your belief that is doing no work, and that part is usually the largest part. Most practitioner claims fail this exercise at column five, and the failure is informative rather than discouraging, because a claim that cannot be falsified is one you will never learn anything from.

Two columns are worth adding for your own use. The first is time: how long has this claim been operating, and what is the base rate for the outcome if you had done nothing? Without a base rate you cannot tell improvement from fluctuation, and this is the single most common error in practitioners' own records. The second is cost: what does the claim require of you, in hours, and is that cost proportional to the size of the effect you have measured? A practice that takes four hours a week to produce a small autonomic change is a hobby with a health benefit, and calling it anything more is a description error rather than a judgement.

The size of the effect, column by column

Read the sizes across and you notice something about the distribution. The autonomic effect is measurable within a single session and is in the range that ordinary breathing work achieves. The open-label effect is measurable in trials over weeks, on subjective outcomes, and is in the range that clinical placebo effects occupy. The predictive processing row has no size, because it is an account of a mechanism rather than a measured intervention. The card deck row has a size and it is exactly one in five. There is no row where the effect is large, and there is no row where the effect is zero. That distribution is the honest shape of the field, and it is much less dramatic than either the people who sell things or the people who are waiting to expose them.

The practical consequence is about how you spend your effort. If you want a large effect, this table tells you plainly that none of the mechanisms on offer produces one, and the only large effects available in this general area come from ordinary things: sleep, movement, other people, and getting a real problem treated. If you want a reliable modest effect, the autonomic row is the one to build on, and it costs almost nothing and requires no belief. Practitioners who organise their practice around the modest reliable effect tend to still be doing it a decade later, and they are usually the ones other people come to for advice, because they are the ones who know what they are doing.

Where the table cannot help you

The table is a tool for claims that are stated as effects, and there is a whole category of practice it does not address. If your practice is a discipline rather than an intervention, if the value is in the attention rather than the outcome, then there is nothing to measure in the columns above and the table will tell you nothing. That is a legitimate position and plenty of long-term practitioners hold it. What they will usually say is that the discipline changed them rather than that it did something for them, and that phrasing fits the framework better.

The table also cannot help with questions about meaning, community, or the experience of a practice that is not reducible to an effect. Those are real considerations and they are the reason people keep traditions going. The error is not valuing them. The error is letting a claim about meaning slide into a claim about effect without a table, because in the absence of one the two will be treated as the same kind of statement and will be believed and disputed accordingly. Keeping them separate is not a coldness. It is what allows a tradition that means something to survive contact with people who will not.

The row that keeps being misread

The autonomic row is the one that goes wrong in both directions, which is unusual and worth examining. In one direction, practitioners hear that their practice reliably changes their autonomic state and conclude that this is the mechanism by which it affects the world, which is a non sequitur with no support in either column. In the other direction, sceptics read the same row as evidence that the whole practice is harmless, which is also wrong, because a reliable reduction in heart rate variability is not nothing and people who have never measured it underrate it badly. Both errors come from the same place, which is the assumption that one row has to carry the whole case.

The fix is to keep the row narrow and let it be narrow. It says: this practice reliably changes my nervous system, in this direction, by approximately this much, and I can show you. It does not say: therefore the practice works. It does not say: therefore the practice is a placebo. Both of those are additions made by a reader who needs the row to be bigger than it is. The discipline of leaving it at the width it actually is will make you more credible on every other claim you make, which is a strange kind of return on doing less.

Using the table with a sceptic

The table is genuinely useful here, which surprises people who think that a table of partial supports is bad public relations. It is good public relations because it removes the reason for the argument. When you present four mechanisms with their coverage and their size, and one empty row, you have taken the disputed ground away. What remains to discuss is which claims you are making, and you have already answered that. A sceptic's remaining options are to dispute the sourcing of a row, which is a task you can both do in an afternoon, or to assert a mechanism you have not named, which is a much harder thing to do out loud.

It also protects you from the specific bad outcome, which is being asked to defend a claim you never made. Most public disputes about this subject go wrong because a practitioner says something careful, a listener paraphrases it as something sweeping, and the practitioner spends the next year defending the paraphrase. A table prevents this because it is specific enough to be quoted accurately. Bring it out before you are asked, not after, and use the empty row without embarrassment. The empty row is the part that earns the credibility for the other four.

The reference worth keeping

One page, four rows, five columns, plus the empty row. That is the whole artefact, and the reason it is worth the effort is that it is the only piece of equipment in this subject that survives contact with an intelligent hostile reader. Most of what circulates in this field is written to be enjoyed rather than to be checked, and there is nothing wrong with that except that it cannot do any work when work is needed. A table is dull and it is portable and it can be handed to someone who is sceptical, which is more than can be said for almost any other document you will find in this area.

Keep it current by adding a row rather than editing one, so the history of what you believed and when is visible. Review it once a year against your own practice, using the last two columns from the exercise above. Most years you will add nothing and delete nothing, and that is the correct result, because the table is not supposed to be exciting. It is supposed to be true as of the last time you checked, which is the best any reference in this field can manage.

Frequently Asked Questions

Does the mind science of practice prove that magic works?

No, and the reference exists partly to prevent that reading. Three of the four well-supported mechanisms are about changes in the practitioner, and the fourth is a measuring instrument. None of them supports the claim that a practice reaches outside the person, which is why the table has an empty row in it. The table is an account of what has been measured, not a verdict on the practice.

What is the most reliably demonstrated effect in this area?

Autonomic change from a fixed, attended sequence: reduced heart rate variability, slower and more even breathing, changes in skin conductance. This has been shown in controlled settings and it does not require belief to produce. It is also within the range that ordinary breathing work achieves, which makes it modest and very reliable, and those two properties together are rarer in this field than they are anywhere else.

Why is the Zener card deck so important?

Because the arithmetic is the cheapest calibration available. Five symbols means a guess is right one time in five, so a group guessing over a session lands near twenty per cent, and any claim of thirty-four per cent on forty calls becomes legible in a way it was not before. The deck does not measure anything about your practice. It measures your judgement, and it does that in ten minutes with no equipment.

Is predictive processing proof that expectation shapes reality?

No. It is a framework in cognitive science, actively debated, and its application to any particular ritual is an inference rather than a finding. What it offers is a mechanical account of why changing what you expect could change what you notice, without requiring anyone to deceive themselves. That is a modest and useful claim. It says nothing about the world outside the person, and it is frequently stretched to say so.

Is it scepticism to build this kind of table?

It is accounting. The table separates physiological change, reported change, behavioural change and external change, and it puts an empty row in the last column because nothing in the well-studied mechanisms supports it. That is not a position on whether the practice is worth doing. It is a position on what you are claiming, which you can hold alongside a genuine commitment to the practice itself.