WEBVTT NOTE A Challenge to Mises’s Theory of Probability 1 00:00:00.000 --> 00:00:08.000 This is kind of a boring topic, definition of probability, but hopefully this will interest some people here. 2 00:00:10.000 --> 00:00:17.000 The paper that I'm presenting builds upon the paper that I published last year in Libertarian Papers, 3 00:00:17.000 --> 00:00:27.000 where I defended the subjective definition of probability against, more in that paper 4 00:00:27.000 --> 00:00:31.120 against Richard von Mises. 5 00:00:31.120 --> 00:00:38.200 So in this paper I turn more to examine Ludwig von Mises' specific definition of theory of 6 00:00:38.200 --> 00:00:51.880 Probability, let's see, so the way that I'd like to proceed is not exactly in the order 7 00:00:51.880 --> 00:01:00.460 that I wrote the paper, just because I did not present that paper last year, so I'd like 8 00:01:00.460 --> 00:01:07.640 to lay out why I think probability should be defined subjectively and then I can examine 9 00:01:07.640 --> 00:01:14.640 and Ludwig von Mises' theory of probability in the light of the subjective definition. 10 00:01:19.760 --> 00:01:25.820 So there's two general ways we can go about defining probability, speaking very generally. 11 00:01:25.820 --> 00:01:32.820 We can say that probability is a real physical feature of the world, like the weight of a 12 00:01:32.820 --> 00:01:42.820 Like the weight of a chair or the height of something, we can think of it as an actual physical property of the world. 13 00:01:42.820 --> 00:01:49.820 And this is the way that Richard von Mises, Ludwig's brother, thought of probability. 14 00:01:49.820 --> 00:02:05.320 He speaks of probability in exactly these terms. He says the probability of a six, this is his quote, 15 00:02:05.320 --> 00:02:15.320 the probability of a six is a physical property of a given die and is a property analogous to its mass, specific heat or electrical resistance. 16 00:02:15.320 --> 00:02:34.320 Similarly, for a given pair of dice, including of course the total setup, the probability of a 6 is a characteristic property, a physical constant belonging to the experiment, as a whole, and comparable with its other physical properties. 17 00:02:34.320 --> 00:02:46.320 So that's the objective, if we were to conceive of probability as an objective physical feature of the world, that's basically the conclusion we would have to draw about probability. 18 00:02:46.320 --> 00:02:56.320 But there's another way that we can conceive of probability, and this is the subjective conception of probability. 19 00:02:56.320 --> 00:03:21.320 And under this conception, probability is not a physical feature of the world, it's a... it's hard to speak of this, the terms here get confusing, but it's not a physical feature of the world, it's a feature of man, it's something inside man. 20 00:03:21.320 --> 00:03:32.320 It's a consequence of the fact that man is not omniscient. 21 00:03:32.320 --> 00:03:43.320 The famous quote from one of the most famous subjectivists is from Bruno de' Finetti, who said that probability does not exist. 22 00:03:43.320 --> 00:03:49.840 And he was a probabilist, so he believed in using probability, but he did not conceive 23 00:03:49.840 --> 00:03:55.120 a probability as something out there in the world to be measured like Richard von Mises 24 00:03:55.120 --> 00:04:04.560 thought probability was. 25 00:04:04.560 --> 00:04:09.900 So the question then is, which of these conceptions of probability is correct? 26 00:04:09.900 --> 00:04:18.140 An Objective Physical Feature of the World or a Subjective Quality Inside Man? 27 00:04:18.140 --> 00:04:27.260 The answer, I think, comes from the probabilist I.J. Good, who pointed out that the question 28 00:04:27.260 --> 00:04:30.780 really depends upon the nature of the world. 29 00:04:30.780 --> 00:04:36.100 If the world is deterministic, then we have to say that probability is a measure of something 30 00:04:36.100 --> 00:04:48.680 and Man. If, on the other hand, the world is considered indeterministic or random in 31 00:04:48.680 --> 00:04:54.960 some sense, then we can define it the way that Richard von Mises defines it. And Richard 32 00:04:54.960 --> 00:05:03.760 Ludwig von Mises specifically defines probability as the limit, let's see, he defines it as 33 00:05:03.760 --> 00:05:19.260 the limit of the relative frequency of occurrence, he defines it as the limit of the relative 34 00:05:19.260 --> 00:05:29.540 Frequency of Occurrence and an indefinite sequence of observations in a collective. 35 00:05:29.540 --> 00:05:35.860 So the question is, is the world deterministic? Does everything that happens in the world 36 00:05:35.860 --> 00:05:46.860 have a cause? If I flip a coin, is there something random about the outcome? Or is it just a 37 00:05:46.860 --> 00:05:56.860 is a matter of, is it just a collection of various forces acting on the coin that if we knew what all those factors were, we would know the outcome. 38 00:06:00.860 --> 00:06:14.860 And I, of course, take the position that this comes from the paper I published last year, but I take the position that everything that happens in the world has a cause. 39 00:06:14.860 --> 00:06:22.300 It's not even really conceivable that something would have, something would occur completely 40 00:06:22.300 --> 00:06:29.660 randomly or without a cause. 41 00:06:29.660 --> 00:06:38.260 So let me just add as well that even if we did conceive of the world as having a certain 42 00:06:38.260 --> 00:06:44.460 amount of randomness in the world, as Richard von Mises did, and Richard von Mises drew 43 00:06:44.460 --> 00:07:00.460 Heisenberg established that at certain levels of smallness, things just happen. 44 00:07:00.460 --> 00:07:07.800 I think this is a misreading of Heisenberg, but it was interpreted largely to, especially 45 00:07:07.800 --> 00:07:14.060 by Richard von Mises as saying that there's a certain amount of randomness in the world. 46 00:07:14.060 --> 00:07:24.860 Even if that's the case, there's still uncertainty in man, in the sense that, even if there's 47 00:07:24.860 --> 00:07:31.340 uncertainty about whether a coin will come up heads or tails, we don't know which way 48 00:07:31.340 --> 00:07:45.340 So, even in the case that, let me just rephrase this, if the world is deterministic, everything 49 00:07:45.340 --> 00:07:53.020 has a cause, then we have to be subjectivist, we have to define probability subjectively. 50 00:07:53.020 --> 00:08:00.000 If the world has an element of uncaused randomness in it, then we can define it as a physical 51 00:08:00.000 --> 00:08:08.000 is a fundamental feature of the world, but there's still uncertainty in man, even in that case. 52 00:08:08.000 --> 00:08:24.000 So, the central philosophical question then is whether everything in the world has a cause. 53 00:08:24.000 --> 00:08:28.000 And for Austrians, this is rather straightforward because, 54 00:08:28.000 --> 00:08:34.880 for Austrians who follow Mises because Mises was a determinist and he was very unequivocal 55 00:08:34.880 --> 00:08:38.960 that everything has a cause. 56 00:08:38.960 --> 00:08:45.600 And Hans-Hermann Hoppe is the one who really established this as an axiom, but this is 57 00:08:45.600 --> 00:08:50.600 This is what Mises himself had to say about this. 58 00:09:01.600 --> 00:09:04.600 This is Mises, Ludwig von Mises. 59 00:09:04.600 --> 00:09:08.600 In a world without causality and regularity of phenomena, 60 00:09:08.600 --> 00:09:12.600 there would be no field for human reasoning and human action. 61 00:09:12.600 --> 00:09:18.600 The world would be a chaos in which man would be at a loss to find any orientation and guidance. 62 00:09:18.600 --> 00:09:23.600 Man is not even capable of imagining the conditions of such a chaotic universe. 63 00:09:23.600 --> 00:09:31.600 And this is what makes it somewhat difficult to talk about Mises' theory of probability 64 00:09:31.600 --> 00:09:43.180 because he discusses probability very much in line with his brother in certain respects. 65 00:09:43.180 --> 00:09:47.840 He talks about class probability which is almost identical with his brother's idea 66 00:09:47.840 --> 00:09:54.440 of collectives and his brother was a dogmatic frequentist. But on the other hand, Mises 67 00:09:54.440 --> 00:10:01.280 This puts his discussion of probability in the section on uncertainty, and as a matter 68 00:10:01.280 --> 00:10:06.760 of fact his discussion of probability takes up almost the entire chapter on uncertainty. 69 00:10:06.760 --> 00:10:14.040 So he clearly understood that probability was a problem of human knowledge and not a 70 00:10:14.040 --> 00:10:17.000 problem of the natural sciences. 71 00:10:17.000 --> 00:10:23.960 And he says specifically that probability is a concern of praxeology and not a concern 72 00:10:23.960 --> 00:10:29.920 of the Physical Sciences, whereas Richard thought that it was exclusively the concern 73 00:10:29.920 --> 00:10:36.080 of the physical sciences. 74 00:10:36.080 --> 00:10:43.860 So the other way that we know that the world is causally deterministic in the sense that 75 00:10:43.860 --> 00:10:50.440 everything that happens has a cause, we know this because the relative frequency method 76 00:10:50.440 --> 00:10:59.260 for Generating Probabilities relies upon this principle. If it was not the case, if the 77 00:10:59.260 --> 00:11:11.880 world was assumed to not be governed by causally determined, the term Hoppe uses is time invariant 78 00:11:11.880 --> 00:11:19.560 causal forces. If it was the case that the world operated in that way, that these sorts 79 00:11:19.560 --> 00:11:25.600 of Causes did not exist. We could not construct collectives or classes from which we could 80 00:11:25.600 --> 00:11:32.600 generate relative frequencies of occurrence. So Richard von Mises' favorite method for 81 00:11:32.720 --> 00:11:39.720 – not favorite, he claims it's the exclusive method for generating numerical probabilities 82 00:11:40.840 --> 00:11:47.840 itself relies upon the principle of causality. Governing the phenomena that you're studying, 83 00:11:49.560 --> 00:12:03.660 So, given what I've just said, this leads us inexorably to the, I think, to the position 84 00:12:03.660 --> 00:12:10.760 that probability must be defined subjectively in the following way. 85 00:12:10.760 --> 00:12:16.600 It should be defined as a measure of human uncertainty about the likelihood of occurrence 86 00:12:16.600 --> 00:12:27.280 of Events in the World, and this is radically different from the definition that Richard 87 00:12:27.280 --> 00:12:35.880 von Mises gave us, and I'd like to say that this is radically different from what Ludwig 88 00:12:35.880 --> 00:12:44.480 von Mises gave us in terms of a definition for probability. 89 00:12:44.480 --> 00:12:52.400 And in order to see this, let me just step back and say a few things about how Ludwig 90 00:12:52.400 --> 00:13:00.600 von Mises' theory of probability differed from his brother's theory. 91 00:13:00.600 --> 00:13:09.800 First on the similarity side, they both were very clear that they thought that numerical 92 00:13:09.800 --> 00:13:13.480 Numerical probability only applied to classes of events. 93 00:13:13.480 --> 00:13:21.280 Ludwig von Mises uses the term classes, and Richard von Mises uses the term collectives. 94 00:13:21.280 --> 00:13:27.480 But essentially what they mean, what they're talking about is the same thing, that unless 95 00:13:27.480 --> 00:13:34.360 you can construct a collective or a class of events that are virtually identical in 96 00:13:34.360 --> 00:13:43.480 In every way, there can be no numerical probability and this, for both of them, this rules out 97 00:13:43.480 --> 00:13:49.800 by definition the possibility of calculating numerical probabilities for singular events 98 00:13:49.800 --> 00:14:02.520 like the Super Bowl or events like this where you can't put it into a class, a class of 99 00:14:02.520 --> 00:14:10.660 of Similar Events and Generated Relative Frequency of Occurrence, which they called the probability 100 00:14:10.660 --> 00:14:15.620 that was generated that way, the numerical probability. 101 00:14:15.620 --> 00:14:20.720 But Ludwig von Mises' theory of probability is different from Richard von Mises' theory 102 00:14:20.720 --> 00:14:24.020 in very important ways, I think. 103 00:14:24.020 --> 00:14:28.560 In the first place, Richard von Mises was an indeterminist. 104 00:14:28.560 --> 00:14:35.960 He was a proponent or he was a disciple of Heisenberg who thought that things could happen 105 00:14:35.960 --> 00:14:45.040 in the world that were indeterminate, random, without cause. 106 00:14:45.040 --> 00:14:53.120 He felt this not just at the micro level, which is how most people interpret Heisenberg 107 00:14:53.120 --> 00:15:01.120 has meaning that at the micro level of smallness, the ultimate level of smallness, things move randomly. 108 00:15:01.120 --> 00:15:06.120 But Richard von Mises went further than that and said that even at the macro level, 109 00:15:06.120 --> 00:15:11.120 certain things are indeterminate like length. 110 00:15:11.120 --> 00:15:15.120 This is completely different from Ludwig von Mises' position here. 111 00:15:15.120 --> 00:15:21.120 This is a determinist, as the quote that I just read indicates. 112 00:15:27.120 --> 00:15:35.120 Another difference between the two is that Ludwig von Mises emphasizes a great deal that probability deals with uncertainty. 113 00:15:36.120 --> 00:15:42.120 And this is almost common sense for anybody who hasn't read Richard von Mises anyway. 114 00:15:42.120 --> 00:15:48.520 Anyway. But if you read Richard von Mises, you could walk away from his book without 115 00:15:48.520 --> 00:15:55.600 even the slightest indication that probability and uncertainty are related to one another. 116 00:15:55.600 --> 00:15:59.400 But most importantly, and I think this is the thing to take away from my paper anyway, 117 00:15:59.400 --> 00:16:06.080 is that I think that Ludwig von Mises never gave a definition of probability in anything 118 00:16:06.080 --> 00:16:19.080 In Human Action, he gives a general definition of the word probable, but he does not give us a definition of the concept of probability. 119 00:16:19.080 --> 00:16:31.080 So, defining the word probable, and really it's just a rehash of a standard dictionary definition of the word probable, that doesn't really clarify anything for us. 120 00:16:31.080 --> 00:16:39.200 for us, this is a philosophical problem that you can't really just dispense with by giving 121 00:16:39.200 --> 00:16:45.440 the definition of the word probable. 122 00:16:45.440 --> 00:16:57.400 So since he lacks an explicit definition of probability, my position is that he ought 123 00:16:57.400 --> 00:17:00.360 to have adopted the subjective definition. 124 00:17:00.360 --> 00:17:06.440 fits in with the entirety of his philosophical, praxeological, methodological system. 125 00:17:06.440 --> 00:17:15.440 He was a determinist, and I think that he adopted this, his brother's position here 126 00:17:15.440 --> 00:17:25.120 because, initially, because he was able to neatly, conceptually integrate class probability 127 00:17:25.120 --> 00:17:29.160 and Case Probability into what he thought was a, 128 00:17:31.800 --> 00:17:36.040 well, he thought that by integrating class probability 129 00:17:36.040 --> 00:17:39.240 into the natural sciences and case probability 130 00:17:39.240 --> 00:17:42.720 into the social sciences, that this was a neat separation 131 00:17:42.720 --> 00:17:45.600 and an important division. 132 00:17:45.600 --> 00:17:49.320 But if we adopt a subjective definition of probability, 133 00:17:49.320 --> 00:17:51.880 this is totally arbitrary. 134 00:17:55.120 --> 00:18:16.120 And in addition to this, if we define probability subjectively, this also undercuts entirely the idea that numerical probabilities cannot be applied to singular cases. 135 00:18:16.120 --> 00:18:24.520 There's no non-arbitrary reason for us to say that, for example, putting a number of 136 00:18:24.520 --> 00:18:35.440 .6 on the NHL finals this year is not a probability just like any other. 137 00:18:35.440 --> 00:18:40.760 That makes sense if we adopt Richard von Mises' definition of probability, but if we have a 138 00:18:40.760 --> 00:18:43.860 have a subjective definition that's just not defensible. 139 00:18:43.860 --> 00:18:45.100 It's totally arbitrary. 140 00:18:47.700 --> 00:18:52.700 And there's no non-arbitrary reason to say that, for example, 141 00:18:54.880 --> 00:18:57.020 the terror alert system in the United States 142 00:18:57.020 --> 00:19:00.900 is not also a probability. 143 00:19:00.900 --> 00:19:05.900 It tells us, it gives us a likelihood of occurrence 144 00:19:05.900 --> 00:19:27.900 And that doesn't mean we have to accept it, because in a sense it's an opinion. Probabilities are opinions, but nevertheless it is a probability, if we adopt the subjective definition of probability. 145 00:19:27.900 --> 00:19:44.900 And finally, let me just close by talking about these two famous subdivisions that Ludwig von Mises came up with, case probability on the one hand and class probability. 146 00:19:44.900 --> 00:19:48.180 If we adopt a subjective definition of probability, 147 00:19:48.180 --> 00:19:53.020 these are totally arbitrary and they are essentially 148 00:19:53.020 --> 00:19:58.020 nothing more than methodological subcategories. 149 00:19:58.180 --> 00:20:01.100 It's a way of saying some probabilities are generated 150 00:20:01.100 --> 00:20:03.740 with the relative frequency approach and some are not, 151 00:20:03.740 --> 00:20:05.480 and nothing more than that. 152 00:20:05.480 --> 00:20:10.180 So, in conclusion, Mises should have been 153 00:20:10.180 --> 00:20:11.880 a subjectivist, I suppose. 154 00:20:14.900 --> 00:20:16.900 Thank you very much.