In mathematical analysis, a measure on a set is a systematic way to assign a number to each suitable subset of that set, intuitively interpreted as its size. In this sense, a measure is a generalization of the concepts of length, area, and volume. A particularly important example is the Lebesgue measure on a Euclidean space, which assigns the conventional length, area, and volume of Euclidean geometry to suitable subsets of the ndimensional Euclidean space R^{n}. For instance, the Lebesgue measure of the interval [0, 1] in the real numbers is its length in the everyday sense of the word, specifically, 1.
Technically, a measure is a function that assigns a nonnegative real number or +∞ to (certain) subsets of a set X (see Definition below). It must further be countably additive: the measure of a 'large' subset that can be decomposed into a finite (or countably infinite) number of 'smaller' disjoint subsets is equal to the sum of the measures of the "smaller" subsets. In general, if one wants to associate a consistent size to each subset of a given set while satisfying the other axioms of a measure, one only finds trivial examples like the counting measure. This problem was resolved by defining measure only on a subcollection of all subsets; the socalled measurable subsets, which are required to form a σalgebra. This means that countable unions, countable intersections and complements of measurable subsets are measurable. Nonmeasurable sets in a Euclidean space, on which the Lebesgue measure cannot be defined consistently, are necessarily complicated in the sense of being badly mixed up with their complement.^{[1]} Indeed, their existence is a nontrivial consequence of the axiom of choice.
Measure theory was developed in successive stages during the late 19th and early 20th centuries by Émile Borel, Henri Lebesgue, Johann Radon, and Maurice Fréchet, among others. The main applications of measures are in the foundations of the Lebesgue integral, in Andrey Kolmogorov's axiomatisation of probability theory and in ergodic theory. In integration theory, specifying a measure allows one to define integrals on spaces more general than subsets of Euclidean space; moreover, the integral with respect to the Lebesgue measure on Euclidean spaces is more general and has a richer theory than its predecessor, the Riemann integral. Probability theory considers measures that assign to the whole set the size 1, and considers measurable subsets to be events whose probability is given by the measure. Ergodic theory considers measures that are invariant under, or arise naturally from, a dynamical system.
YouTube Encyclopedic

1/5Views:505 44660 4752 126 1291 457 50847 673

✪ Music And Measure Theory

✪ Measure Theory  Motivation

✪ Math Antics  Area

✪ Math Antics  Angle Basics

✪ The World of Mathematical Reality
Transcription
I have two seemingly unrelated challenges for you. The first relates to music, and the second gives a foundational result in measure theory, which is the formal underpinning for how mathematicians define integration and probability. The second challenge, which I’ll get to about halfway through the video, has to do with covering numbers with open sets, and is very counterintuitive. Or at least, when I first saw it I was confused for a while. Foremost, I’d like to explain what’s going on, but I also plan to share a surprising connection it has with music. Here’s the first challenge. I’m going to play a musical note with a given frequency, let’s say 220 hertz, then I’m going to choose some number between 1 and 2, which we’ll call r, and play a second musical note whose frequency is r times the frequency of the first note, 220. For some values of this ratio r, like 1.5, the two notes will sound harmonious together, but for others, like the square root of 2, they sound cacophonous. Your task it to determine whether a given ratio r will give a pleasant sound or an unpleasant one just by analyzing the number and without listening to the notes. One way to answer, especially if your name is Pythagoras, might be that two notes sound good when the ratio is a rational number, and bad when it is irrational. For instance, a ratio of 3/2 gives a musical fifth, 4/3 gives a musical fourth, of 8/5 gives a minor sixth, etc. Here’s my best guess for why this is the case: a musical note is made up of beats played in rapid succession, for instance 220 beats per second. When the ratio of frequencies of two notes is rational, there is a detectable pattern in those beats, which, when we slow it down, we hear as a rhythm instead of as a harmony. Evidently when our brains pick up on this pattern, two notes sound nice together. However, most rational numbers actually sound pretty bad, like 211/198, or 1093/826. The issue, of course, is that these rational number are somehow more “complicated” than the other ones, our ears don’t pick up on the pattern of the beats. One simple way to measure the complexity of a rational number is to consider the size of its denominator when it is written in reduced form. So we might edit our original answer to only admit fractions with low denominators, say less than 10. Even still, this doesn’t quite capture harmoniousness, since plenty of notes sound good together even when the ratio of their frequencies is irrational, so long as it is close to a harmonious rational number. And it’s a good thing, too, because many instruments such as pianos are not tuned in terms of rational intervals, but are tuned such that each halfstep increase corresponds with multiplying the original frequency by the 12th root of 2, which is irrational. If you’re curious about why this is done, Henry at minutephysics recently did a video which gives a very nice explanation. This means that if you take a harmonious interval, like a fifth, the ratio of frequencies when played on a piano will not be a nice rational number like you expect, in this case 3/2, but will instead be some power of the 12th root of 2, in this case 2^{7/12}, which is irrational, but very close to 3/2. Similarly, a musical fourth corresponds to 2^{5/12}, which is very close to 4/3. In fact, the reason it works so well to have 12 notes in the chromatic scale is that powers of the 12th root of 2 have a strange tendency to be within a 1% margin of error of simple rational numbers. So now you might say a ratio r will produce a harmonious pair of notes if it is sufficiently close to a rational number with a sufficiently small denominator. How close depends on how discerning your ear is, and how small a denominator depends on the intricacy of harmonic patterns your ear has been trained to pick up on. After all, maybe someone with a particularly acute musical sense would be able to hear and find pleasure in the pattern resulting from more complicated fractions like 23/21 or 35/43, as well as numbers closely approximating these fractions. This leads to an interesting question: Suppose there is a musical savant, who find pleasure in all pairs of notes whose frequencies have a rational ratio, even super complicated ratios that you and I would find cacophonous. Is it the case that she would find all ratios r between 1 and 2 harmonious, even the irrational ones? After all, for any given real number you can always find rational numbers arbitrarily close it, just as 3/2 is close to 2^{7/12}. Well, this brings us to challenge number 2. Mathematicians like to ask riddles about covering various sets with open intervals, and the answers to these riddles have a strange tendency to become famous lemmas and theorems. By “open interval”, I just mean the continuous stretch of real numbers strictly greater than some number a, but strictly less than some other number b, where b is of course greater than a. My challenge to you involves covering all the rational numbers between 0 and 1 with open intervals. When I say “cover”, all that means is that each particular rational number lies in at least one of your intervals. The most obvious way to do this is to just use the entire interval from 0 to 1 itself and call it done, but the challenge here is that the sum of the lengths of your intervals must be strictly less than 1. To aid you in this seemingly impossible task, you are allowed to use infinitely many intervals. Even still, the task might feel impossible, since the rational numbers are dense in the real numbers, meaning any stretch, no matter how small, contains infinitely many rational numbers. So how could you possibly cover all rational numbers without just covering the entire interval from 0 to 1 itself, which would mean the total length of your open intervals has to be at least the length of the entire interval from 0 to 1. Then again, I wouldn’t be talking about this if there was not a way to do it. First, we enumerate the rational numbers between 0 and 1, meaning we organize them into an infinitely long list. There are many ways to do this, but one natural way I’ll choose is start with ½, followed by ⅓ and ⅔, then ¼ and ¾, we don’t write down 2/4 since it has already appeared as ½, then all reduced fractions with denominator 5, all reduced fractions with denominator 6, continuing on and on in this fashion. Every fraction will appear exactly once in this list, in its reduced form, and it gives us a meaningful way to talk about the “first” rational number, the “second” rational number, the 42nd rational number, things like that. Next, to ensure that each rational is covered, we are going to assign one specific interval to each rational. Once we remove the intervals from the geometry of our setup and just think of them in a list, each one responsible for only one rational number, it seems much clearer that the sum of their lengths can be less than 1, since each particular interval can be as small as you want and still cover its designated rational. In fact, the sum can be any positive number. Just choose an infinite sum with positive terms that converges to 1, like ½+¼+⅛+... on and on with powers of 2, then choose any desired value epsilon>0, like 0.5, and multiply all terms by epsilon so that we have an infinite sum converging to epsilon. Now scale the nth interval to have a length equal to the nth term in the sum. Notice, this means your intervals start getting really small, really fast, so small that you can’t really see most of them in this animation, but it doesn’t matter, since each one is only responsible for covering one rational. I’ve said it already, by I’ll say it again because it’s so amazing: epsilon can be whatever positive number we want, so not only can our sum be less than 1, it can be arbitrarily small! This is one of those results where even after seeing the proof, it still defies intuition. The discord here is that the proof has us thinking analytically, with the rational numbers in a list, but our intuition has us thinking geometrically, with the rationals as a dense set on the interval, where you can’t skip over any continuous stretch of numbers since each stretch contains infinitely many rationals. So let’s get a visual understanding of what’s going on. Brief side note here: I had trouble deciding on how to illustrate small open intervals, since if I scale the parentheses with the interval, you won’t be able to see them at all, but if I just push the parentheses together, they cross over in a way that it potentially confusing. Nevertheless, I decided to go with the ugly chromosomal cross, so keep in mind that the interval they represent is the tiny stretch between the centers of each parenthesis. Okay, back to the visual intuition. Consider when epsilon = 0.3, meaning if I choose a number between 0 and 1 at random, there is a 70% that it is outside all those infinitely many intervals. What does it look like to be outside the intervals? Well, the square root of 2 over 2 is among those 70%, and I’m going to zoom in it. As I do so I’ll draw the first 10 intervals in the list within our scope of vision. As we get closer to the square root of 2 over 2, even though you will always find rationals within your field of view, the intervals placed on top of those rationals get really small really fast. One might say that for any sequence of rational numbers approaching the square root of 2 over 2, the intervals covering the elements of this sequence shrink faster than that sequence converges. Notice, intervals are really small if they show up very late in the list, and rationals show up late in the list when they have large denominators, so the fact that the square root of 2 over 2 is among the 70% not covered by our intervals is in a sense a way to formalize the otherwise vague idea that the only rational numbers “close” to it have large denominators. That is to say, the square root of 2 over 2 is cacophonous. In fact, let’s use a smaller epsilon, say 0.01, and shift our setup to lie on top of the interval from 1 to 2 instead of from 0 to 1. Then which numbers fall among the elite 1% covered by our tiny intervals? Almost all of them are harmonious! For instance, the harmonious irrational number 2^{7/12} is very close to 3/2, which has a relatively fat interval sitting on top of it, and the interval around 4/3 is smaller, but still fat enough to cover 2^{5/12}. Which members of the 1% are cacophonous? Well, the cacophonous rationals, meaning those with high denominators, and irrationals that are very very very close to them. However, think of the savant who finds harmonic patterns in all rational numbers. You could imagine that for her, harmonious numbers are precisely those 1% covered by the intervals, provided that her tolerance for error goes down exponentially for more complicated rationals. In other words, the seemingly paradoxical fact that you can have a collection of intervals densely populate a range while only covering 1% of its values corresponds to the fact that harmonious numbers are rare, even for the savant. I’m not saying this makes it the result more intuitive, in fact, I find it quite surprising that the savant I defined could find 99% of all ratios cacophonous, but the fact that these two ideas are connected was simply too beautiful not to share.
Contents
Definition
Let X be a set and Σ a σalgebra over X. A function μ from Σ to the extended real number line is called a measure if it satisfies the following properties:
 Nonnegativity: For all E in Σ: μ(E) ≥ 0.
 Null empty set: .
 Countable additivity (or σadditivity): For all countable collections of pairwise disjoint sets in Σ:
One may require that at least one set E has finite measure. Then the empty set automatically has measure zero because of countable additivity, because
which implies (since the sum on the right thus converges to a finite value) that .
If only the second and third conditions of the definition of measure above are met, and μ takes on at most one of the values ±∞, then μ is called a signed measure.
The pair (X, Σ) is called a measurable space, the members of Σ are called measurable sets. If and are two measurable spaces, then a function is called measurable if for every Ymeasurable set , the inverse image is Xmeasurable – i.e.: . In this setup, the composition of measurable functions is measurable, making the measurable spaces and measurable functions a category, with the measurable spaces as objects and the set of measurable functions as arrows. See also Measurable function#Term usage variations about another setup.
A triple (X, Σ, μ) is called a measure space. A probability measure is a measure with total measure one – i.e. μ(X) = 1. A probability space is a measure space with a probability measure.
For measure spaces that are also topological spaces various compatibility conditions can be placed for the measure and the topology. Most measures met in practice in analysis (and in many cases also in probability theory) are Radon measures. Radon measures have an alternative definition in terms of linear functionals on the locally convex space of continuous functions with compact support. This approach is taken by Bourbaki (2004) and a number of other sources. For more details, see the article on Radon measures.
Examples
Some important measures are listed here.
 The counting measure is defined by μ(S) = number of elements in S.
 The Lebesgue measure on R is a complete translationinvariant measure on a σalgebra containing the intervals in R such that μ([0, 1]) = 1; and every other measure with these properties extends Lebesgue measure.
 Circular angle measure is invariant under rotation, and hyperbolic angle measure is invariant under squeeze mapping.
 The Haar measure for a locally compact topological group is a generalization of the Lebesgue measure (and also of counting measure and circular angle measure) and has similar uniqueness properties.
 The Hausdorff measure is a generalization of the Lebesgue measure to sets with noninteger dimension, in particular, fractal sets.
 Every probability space gives rise to a measure which takes the value 1 on the whole space (and therefore takes all its values in the unit interval [0, 1]). Such a measure is called a probability measure. See probability axioms.
 The Dirac measure δ_{a} (cf. Dirac delta function) is given by δ_{a}(S) = χ_{S}(a), where χ_{S} is the indicator function of S. The measure of a set is 1 if it contains the point a and 0 otherwise.
Other 'named' measures used in various theories include: Borel measure, Jordan measure, ergodic measure, Euler measure, Gaussian measure, Baire measure, Radon measure, Young measure, and Loeb measure.
In physics an example of a measure is spatial distribution of mass (see e.g., gravity potential), or another nonnegative extensive property, conserved (see conservation law for a list of these) or not. Negative values lead to signed measures, see "generalizations" below.
 Liouville measure, known also as the natural volume form on a symplectic manifold, is useful in classical statistical and Hamiltonian mechanics.
 Gibbs measure is widely used in statistical mechanics, often under the name canonical ensemble.
Basic properties
Let μ be a measure.
Monotonicity
If E_{1} and E_{2} are measurable sets with E_{1} ⊆ E_{2} then
Measure of countable unions and intersections
Subadditivity
For any countable sequence E_{1}, E_{2}, E_{3}, ... of (not necessarily disjoint) measurable sets E_{n} in Σ:
Continuity from below
If E_{1}, E_{2}, E_{3}, ... are measurable sets and E_{n} is a subset of E_{n + 1} for all n, then the union of the sets E_{n} is measurable, and
Continuity from above
If E_{1}, E_{2}, E_{3}, ... are measurable sets and for all n, E_{n + 1} ⊂ E_{n}, then the intersection of the sets E_{n} is measurable; furthermore, if at least one of the E_{n} has finite measure, then
This property is false without the assumption that at least one of the E_{n} has finite measure. For instance, for each n ∈ N, let E_{n} = [n, ∞) ⊂ R, which all have infinite Lebesgue measure, but the intersection is empty.
Sigmafinite measures
A measure space (X, Σ, μ) is called finite if μ(X) is a finite real number (rather than ∞). Nonzero finite measures are analogous to probability measures in the sense that any finite measure μ is proportional to the probability measure . A measure μ is called σfinite if X can be decomposed into a countable union of measurable sets of finite measure. Analogously, a set in a measure space is said to have a σfinite measure if it is a countable union of sets with finite measure.
For example, the real numbers with the standard Lebesgue measure are σfinite but not finite. Consider the closed intervals [k, k+1] for all integers k; there are countably many such intervals, each has measure 1, and their union is the entire real line. Alternatively, consider the real numbers with the counting measure, which assigns to each finite set of reals the number of points in the set. This measure space is not σfinite, because every set with finite measure contains only finitely many points, and it would take uncountably many such sets to cover the entire real line. The σfinite measure spaces have some very convenient properties; σfiniteness can be compared in this respect to the Lindelöf property of topological spaces. They can be also thought of as a vague generalization of the idea that a measure space may have 'uncountable measure'.
sfinite measures
A measure is said to be sfinite if it is a countable sum of bounded measures. Sfinite measures are more general than sigmafinite ones and have applications in the theory of stochastic processes.
Completeness
A measurable set X is called a null set if μ(X) = 0. A subset of a null set is called a negligible set. A negligible set need not be measurable, but every measurable negligible set is automatically a null set. A measure is called complete if every negligible set is measurable.
A measure can be extended to a complete one by considering the σalgebra of subsets Y which differ by a negligible set from a measurable set X, that is, such that the symmetric difference of X and Y is contained in a null set. One defines μ(Y) to equal μ(X).
Additivity
Measures are required to be countably additive. However, the condition can be strengthened as follows. For any set and any set of nonnegative define:
That is, we define the sum of the to be the supremum of all the sums of finitely many of them.
A measure on is additive if for any and any family of disjoint sets the following hold:
Note that the second condition is equivalent to the statement that the ideal of null sets is complete.
Nonmeasurable sets
If the axiom of choice is assumed to be true, it can be proved that not all subsets of Euclidean space are Lebesgue measurable; examples of such sets include the Vitali set, and the nonmeasurable sets postulated by the Hausdorff paradox and the Banach–Tarski paradox.
Generalizations
For certain purposes, it is useful to have a "measure" whose values are not restricted to the nonnegative reals or infinity. For instance, a countably additive set function with values in the (signed) real numbers is called a signed measure, while such a function with values in the complex numbers is called a complex measure. Measures that take values in Banach spaces have been studied extensively.^{[2]} A measure that takes values in the set of selfadjoint projections on a Hilbert space is called a projectionvalued measure; these are used in functional analysis for the spectral theorem. When it is necessary to distinguish the usual measures which take nonnegative values from generalizations, the term positive measure is used. Positive measures are closed under conical combination but not general linear combination, while signed measures are the linear closure of positive measures.
Another generalization is the finitely additive measure, also known as a content. This is the same as a measure except that instead of requiring countable additivity we require only finite additivity. Historically, this definition was used first. It turns out that in general, finitely additive measures are connected with notions such as Banach limits, the dual of L^{∞} and the Stone–Čech compactification. All these are linked in one way or another to the axiom of choice. Contents remain useful in certain technical problems in geometric measure theory; this is the theory of Banach measures.
A charge is a generalization in both directions: it is a finitely additive, signed measure.
See also
 Abelian von Neumann algebra
 Almost everywhere
 Carathéodory's extension theorem
 Content (measure theory)
 Fubini's theorem
 Fatou's lemma
 Fuzzy measure theory
 Geometric measure theory
 Hausdorff measure
 Inner measure
 Lebesgue integration
 Lebesgue measure
 Lorentz space
 Lifting theory
 Measurable cardinal
 Measurable function
 Measure topology
 Minkowski content
 Noncommutative integration
 Outer measure
 Product measure
 Pushforward measure
 Regular measure
 Vector measure
 Valuation (measure theory)
 Volume form
References
 ^ Halmos, Paul (1950), Measure theory, Van Nostrand and Co.
 ^ Rao, M. M. (2012), Random and Vector Measures, Series on Multivariate Analysis, 9, World Scientific, ISBN 9789814350815, MR 2840012.
Bibliography
 Robert G. Bartle (1995) The Elements of Integration and Lebesgue Measure, Wiley Interscience.
 Bauer, H. (2001), Measure and Integration Theory, Berlin: de Gruyter, ISBN 9783110167191
 Bear, H.S. (2001), A Primer of Lebesgue Integration, San Diego: Academic Press, ISBN 9780120839711
 Bogachev, V. I. (2006), Measure theory, Berlin: Springer, ISBN 9783540345138
 Bourbaki, Nicolas (2004), Integration I, Springer Verlag, ISBN 3540411291 Chapter III.
 R. M. Dudley, 2002. Real Analysis and Probability. Cambridge University Press.
 Folland, Gerald B. (1999), Real Analysis: Modern Techniques and Their Applications, John Wiley and Sons, ISBN 0471317160 Second edition.
 D. H. Fremlin, 2000. Measure Theory. Torres Fremlin.
 Jech, Thomas (2003), Set Theory: The Third Millennium Edition, Revised and Expanded, Springer Verlag, ISBN 3540440852
 R. Duncan Luce and Louis Narens (1987). "measurement, theory of," The New Palgrave: A Dictionary of Economics, v. 3, pp. 428–32.
 M. E. Munroe, 1953. Introduction to Measure and Integration. Addison Wesley.
 K. P. S. Bhaskara Rao and M. Bhaskara Rao (1983), Theory of Charges: A Study of Finitely Additive Measures, London: Academic Press, pp. x + 315, ISBN 0120957809
 Shilov, G. E., and Gurevich, B. L., 1978. Integral, Measure, and Derivative: A Unified Approach, Richard A. Silverman, trans. Dover Publications. ISBN 0486635198. Emphasizes the Daniell integral.
 Teschl, Gerald, Topics in Real and Functional Analysis, (lecture notes)
 Tao, Terence (2011). An Introduction to Measure Theory. Providence, R.I.: American Mathematical Society. ISBN 9780821869192.
 Weaver, Nik (2013). Measure Theory and Functional Analysis. World Scientific. ISBN 9789814508568.
External links
Look up measurable in Wiktionary, the free dictionary. 
 Hazewinkel, Michiel, ed. (2001) [1994], "Measure", Encyclopedia of Mathematics, Springer Science+Business Media B.V. / Kluwer Academic Publishers, ISBN 9781556080104
 Tutorial: Measure Theory for Dummies