Probability
Risk, Rationality and Reality
(A summary of the transcript)
What appears over the page are reflections in my words (broadly speaking) on a talk by David Deutsch conducted in 2015 and titled “Physics without Probability”. The original talk can be found here: https://www.youtube.com/watch?v=wfzSE4Hoxbc and my words also appear in a podcast (ToKCast No. 111) for those wanting to explore what is said immediately below any further. What is below is very brief by comparison and serves as an introduction to the argument rather than a more complete fleshing out of the explanation. Those with objections may wish to consult the more complete transcript over the page where their misconceptions might be cleared up.
It is useful to keep in mind that as with all of David’s work, and what sets him apart from many other speakers and writers coming from a scientific background and who author books and deliver public lectures - he is often not only summarising our best explanations known at any given time but rather he is taking them seriously as statements about reality and taking them further. This talk takes physics seriously (and by implication “Constructor Theory” although CT is not needed for the argument he makes herein) and in doing so shows what logically follows about probability. David states, and I emphasise, the probability calculus - rather like Newtonian physics - is a useful fiction. Insofar as it is used in a sort of “instrumentalist” fashion (as a tool) - then it can “work” in much the same way as “Newtonian Physics” can work if one is just flying rockets to the moon (for example). But there is no actual force of gravity that permeates all of space instantly as required by that explanation. So too with probability. A “stochastic process” is a random process which would obey the probability calculus precisely. The assertion herein is that no such physical process exists. They only “seem to exist approximately” - which is not the same thing and again puts us in a class of phenomena like “Newton’s force of gravity”.
Probability theory originally came from Cardano and others in the 15th century who were trying to understand how to win at games of chance. They never imagined it actually would apply to the real world (let alone be applied everywhere from philosophy to high finance).
There are many interpretations of probability. One entails interpreting probabilities as “frequencies”. Frequencies are how often events occur. In my summary I use coin tosses. The fact is we can only toss a coin a finite number of times and in reality the frequency almost never matches the exact probability. This means that theory is, strictly, falsified (as compared to actual quantum theory which does govern the universe to the best of our knowledge). The “frequentist” can say at this point that the frequency will match the probability for an infinitely long sequence. Which, again, places probability well outside of physics because in physics - the real world - we do not have infinitely long sequences of things like coin tosses, dice rolls or anything else. One might argue that we should be interpreting probabilities are what they are “a priori” - i.e: without referencing the real world because they must be what they are by definition. And, again, we have left the real world and entered into a domain of pure mathematics. If the coin has a probability of “heads” that is defined to be “1/2” then we assume it is an ideal mathematical object no longer governed by literal physics that can cause it to deviate from that precise 1/2. Sure, it might be “approximately 1/2” but our business here is to consider what reality literally consists of. Not what rules it approximately conforms to. Finally it is sometimes said we should be interpreting probability as being a purely subjective thing - it is about what we should believe about the world given the limited information we have. This then turns probability from a theory of physics into a theory of psychology about “objects in the mind” rather than objects our there in objective reality. We should believe (set our credences to) what the probability calculus says about things even if it does not apply strictly. We are not told why exactly - and this leads David to compare that kind of denial about reality and probability to a kind of mysticism. A thing is “likely” on that account in the same way as it has “favourable omens”.
David does not need constructor theory to make his case because even absent constructor theory in the universe there is a simple heuristic I continue to come back to:
Things happen. They do not probably happen.
That’s simply physical realism as we know it. But constructor theory does build a mode of explanation out of what can possibly happen and not possibly happen. There are many modes of explanation (David goes through in his talk) and probability is just one - and it is false. Our purpose here, again, is to consider reality. Not merely approximations to reality that have been superseded - no matter how useful those approximations once were or remain.
-Probability theory is like the flat earth theory. It’s useful perhaps (for gardening) - but if you are thinking about what the laws of nature are then false theories are an impediment to understanding reality.-
Probability has persisted in quantum theory because of the so-called “Born Rule” which is about how prior to a measurement there exists a “wave function” (and hence a distribution of values for all the physical properties of a (or even a collection of) physical object(s)) - but at the moment of measurement only one value - not a distribution - is observed. This leads many to say: well the one observed is a function of the “probability” is had given the wave function. But this is all false because it rests on the wave function “collapsing” (disappearing except at the point of observation onto the value(s) observed at observation. So called “unitary quantum theory” (the multiverse) does not add in the extra assumption about collapse because there is no good logical, scientific or philosophical reason to add this additional assumption (to my mind it would be rather like astronomers adding to planetary science an assumption that “our theories of planetary formation are valid but not for other galaxies where, presently, we cannot observe other planets and so we assert those planets do not exist because they cannot be directly observed.” Or something like that.
But what does it matter if we just go on acting as if probability works reasonably well much of the time and so therefore treat it as an explanation? David says “fundamental falsehoods don’t always rear up and bite you. You can believe in a flat earth and it might never affect your life and thinking. But belief in a flat earth theory could eventually conflict with a theory for avoiding asteroid strikes.” Then you’ve got a problem to put it mildly. Again, David is interested in making progress. He wants to understand literal reality. Not merely the approximations to it (especially ones already refuted).
-Similarly you might very well continue to think probability theory is true or randomness is real and so on. You might even be able to invent quantum computers using this whole misconception. But because the Born rule contains misconceptions then you might not end up developing successors to quantum theory. Constructor theory for example is incompatible with probability. In other words, taking literal falsehoods too seriously impedes progress. And that, ultimately, is why this talk is so important. We want to solve our problems and make progress. We don’t want to rest of our laurels and hope that relying on existing knowledge and all of its concomitant misconceptions will be sufficient for the unknown future. At the end of the summary over the page, and in the original podcast, I spend time writing (and speaking) about risk management. I commend that section separately to the reader rather than attempt to summarise it there. It is titled “Risk Management”.
Over the page the rather lengthy transcript to my podcast begins.
It is useful to keep in mind that as with all of David’s work, and what sets him apart from many other speakers and writers coming from a scientific background and who author books and deliver public lectures - he is often not only summarising our best explanations known at any given time but rather he is taking them seriously as statements about reality and taking them further. This talk takes physics seriously (and by implication “Constructor Theory” although CT is not needed for the argument he makes herein) and in doing so shows what logically follows about probability. David states, and I emphasise, the probability calculus - rather like Newtonian physics - is a useful fiction. Insofar as it is used in a sort of “instrumentalist” fashion (as a tool) - then it can “work” in much the same way as “Newtonian Physics” can work if one is just flying rockets to the moon (for example). But there is no actual force of gravity that permeates all of space instantly as required by that explanation. So too with probability. A “stochastic process” is a random process which would obey the probability calculus precisely. The assertion herein is that no such physical process exists. They only “seem to exist approximately” - which is not the same thing and again puts us in a class of phenomena like “Newton’s force of gravity”.
Probability theory originally came from Cardano and others in the 15th century who were trying to understand how to win at games of chance. They never imagined it actually would apply to the real world (let alone be applied everywhere from philosophy to high finance).
There are many interpretations of probability. One entails interpreting probabilities as “frequencies”. Frequencies are how often events occur. In my summary I use coin tosses. The fact is we can only toss a coin a finite number of times and in reality the frequency almost never matches the exact probability. This means that theory is, strictly, falsified (as compared to actual quantum theory which does govern the universe to the best of our knowledge). The “frequentist” can say at this point that the frequency will match the probability for an infinitely long sequence. Which, again, places probability well outside of physics because in physics - the real world - we do not have infinitely long sequences of things like coin tosses, dice rolls or anything else. One might argue that we should be interpreting probabilities are what they are “a priori” - i.e: without referencing the real world because they must be what they are by definition. And, again, we have left the real world and entered into a domain of pure mathematics. If the coin has a probability of “heads” that is defined to be “1/2” then we assume it is an ideal mathematical object no longer governed by literal physics that can cause it to deviate from that precise 1/2. Sure, it might be “approximately 1/2” but our business here is to consider what reality literally consists of. Not what rules it approximately conforms to. Finally it is sometimes said we should be interpreting probability as being a purely subjective thing - it is about what we should believe about the world given the limited information we have. This then turns probability from a theory of physics into a theory of psychology about “objects in the mind” rather than objects our there in objective reality. We should believe (set our credences to) what the probability calculus says about things even if it does not apply strictly. We are not told why exactly - and this leads David to compare that kind of denial about reality and probability to a kind of mysticism. A thing is “likely” on that account in the same way as it has “favourable omens”.
David does not need constructor theory to make his case because even absent constructor theory in the universe there is a simple heuristic I continue to come back to:
Things happen. They do not probably happen.
That’s simply physical realism as we know it. But constructor theory does build a mode of explanation out of what can possibly happen and not possibly happen. There are many modes of explanation (David goes through in his talk) and probability is just one - and it is false. Our purpose here, again, is to consider reality. Not merely approximations to reality that have been superseded - no matter how useful those approximations once were or remain.
-Probability theory is like the flat earth theory. It’s useful perhaps (for gardening) - but if you are thinking about what the laws of nature are then false theories are an impediment to understanding reality.-
Probability has persisted in quantum theory because of the so-called “Born Rule” which is about how prior to a measurement there exists a “wave function” (and hence a distribution of values for all the physical properties of a (or even a collection of) physical object(s)) - but at the moment of measurement only one value - not a distribution - is observed. This leads many to say: well the one observed is a function of the “probability” is had given the wave function. But this is all false because it rests on the wave function “collapsing” (disappearing except at the point of observation onto the value(s) observed at observation. So called “unitary quantum theory” (the multiverse) does not add in the extra assumption about collapse because there is no good logical, scientific or philosophical reason to add this additional assumption (to my mind it would be rather like astronomers adding to planetary science an assumption that “our theories of planetary formation are valid but not for other galaxies where, presently, we cannot observe other planets and so we assert those planets do not exist because they cannot be directly observed.” Or something like that.
But what does it matter if we just go on acting as if probability works reasonably well much of the time and so therefore treat it as an explanation? David says “fundamental falsehoods don’t always rear up and bite you. You can believe in a flat earth and it might never affect your life and thinking. But belief in a flat earth theory could eventually conflict with a theory for avoiding asteroid strikes.” Then you’ve got a problem to put it mildly. Again, David is interested in making progress. He wants to understand literal reality. Not merely the approximations to it (especially ones already refuted).
-Similarly you might very well continue to think probability theory is true or randomness is real and so on. You might even be able to invent quantum computers using this whole misconception. But because the Born rule contains misconceptions then you might not end up developing successors to quantum theory. Constructor theory for example is incompatible with probability. In other words, taking literal falsehoods too seriously impedes progress. And that, ultimately, is why this talk is so important. We want to solve our problems and make progress. We don’t want to rest of our laurels and hope that relying on existing knowledge and all of its concomitant misconceptions will be sufficient for the unknown future. At the end of the summary over the page, and in the original podcast, I spend time writing (and speaking) about risk management. I commend that section separately to the reader rather than attempt to summarise it there. It is titled “Risk Management”.
Over the page the rather lengthy transcript to my podcast begins.