# What's the use of a complete measure?

A complete measure space is one in which any subset of a measure-zero set is measurable.

For what reasons would I want a complete measure space? The only reason I can think of is in the context of probability theory: using complete probability spaces forces almost-everywhere equal random variables to generate the same sigma-sub-algebra.

Am I missing some other important technical reasons?

• It certainly seems useful to have any subset of a set of measure zero be measurable: this is a great way to show that a set has measure zero. If you had to stop every time and check that such sets were actually measurable -- or worse, deal with the possibility that they are not -- it would be a real pain. Jan 12, 2010 at 17:00

Since the existence of non-measurable sets is often seen as undesirable, we naturally want to have as many measurable sets as possible. With Lebesgue measure on the reals, for example, if we were to stop with the collection of Borel sets, we would only have continuum c many measurable sets. But when completing the measure, we gain 2c many more measurable sets, incomparably more. The newly measurable sets are not just measure zero sets, of course, but all those sets that differ from a previously measurable set by (a subset of) a measure zero set.

But it isn't just about the number of measurable sets. Rather, completing the measure allowed us to increase (or even maximize in a sense) our collection of measurable sets in a way that seems to accord completely with how we wanted to measure sets in the first place. It's a basic part of what we were trying to do with measure to be able to say that something that is less than negligible is also negligible.

• This is a good formulation of what I wanted to say. Jan 16, 2010 at 3:22

Wikipedia gives one instance of a situation in which complete measures are needed, for the purpose of defining measures on product spaces.

I suggest that you look into Rudin's "Real and Complex Analysis". There he makes an argument that a completion of an ordinary measure space into a complete measure space is just as fundamental to real analysis, as the completion of the rationals to the reals is.

Many theorems in measure theory, for instance Fubini or Radon-Nikodym, needs completeness to make full sense. Fubini is explained in the wikipedia example. To make the other aspect clear -- quite a few statements in measure theory uses the notion of "almost everywhere" -- for instance the definition of $L^p$ spaces, or Radon-Nikodym.

But this notion of "almost everywhere"(rather, "almost nowhere") becomes better if the measure space is complete. It would look really odd if you declare that some property holds true almost nowhere because it holds only on some set with measure zero, and you so arrange things that some other property holds on a smaller set, and then you are no longer able to make the assertion! The product measure example above is a specific illustration in which the concerned property is simply "being measurable", and the consequences are particularly notable.

Added(Jan 16): There are problems into applications into Ergodic theory, for instance. This definition of ergodic transformation and an ergodic theory built on it will run into all sorts of problems if the underlying measure space is not complete. This is again because you need a proper notion of "almost everywhere" and "almost nowhere".

• Anweshi -- I think this is the right answer (especially Fubini) so I upvoted it. Note though that what you say at the end about rationals versus irrationals is too simplistic: no problems arise in adding or subtracting countable sets! Jan 12, 2010 at 19:02
• @Pete. Thanks for the observation. I have made an edit accordingly, removing the specific example and replacing with a general situation. Jan 12, 2010 at 19:12
• I have a few complaints about this answer: 1) the wikipedia article does not give an example where completeness is needed. it merely states that the product of complete measures is not necessarily a complete measure over the product space 2) fubini-tonelli does not require completeness, however it can make use of it to integrate messier things. 3) radon-nikodym does not depend on completeness. Jan 13, 2010 at 2:27
• I didn't mention that argument since it was meta-mathematical; I think it is equally weird that subsets of any finite-measure set may be non-measurable, but this is something one must accept in order for measure theory to work at all. When working with a measure space, the $\sigma$-algebra dictates which sets exists at all; it's irrelevant what the measure of a non-existent set is. I think it is a sign that something is funny when the completion of Borel measure takes its size from $\mathbb{N}$ to $\mathbb{R}$ .. Jan 13, 2010 at 9:51
• Oh somehow I got an old version of your comment? About fubini-tonelli, the proof goes through fine without completeness, and actually needs extra band-aid work for completeness. As far as Lebesgue measure is concerned, I personally know of no example where the completness does any good! And in fact, below you'll see I gave an answer where completeness caused things to go wrong. I guess what I'm trying to say is, if we built up measure theory without complete measures, would theorems would we lose? I know of none, however Bridge below gives an example.. sorry if i sounded accusatory Jan 13, 2010 at 9:57

From the categorical viewpoint there is no difference, because the category of measurable spaces is equivalent to the category of complete measurable spaces with the equivalence given by the completion functor. Moreover, we are forced to identify objects that are different only on a set of measure 0 or a subset of such a set (otherwise some theorems simply do not make any sense), hence we cannot even see the difference. However, working with complete measurable spaces is technically easier. More precisely, objects of the category of measurable spaces are triples (X,A,N), where X is a set, A is a σ-algebra of measurable subsets of X, N is a σ-ideal of null sets in A. A morphism from (X,A,N) to (Y,B,O) is an equivalence class of maps of sets f: X→Y such that the preimage of every element of B is a union of an element of A and a subset of an element of N and the preimage of every element of O is a subset of an element of N. Two maps are equivalent if they differ on a subset of an element of N. If we restrict our attention to complete measurable spaces, then the definition of morphism becomes significantly simpler: we have to require that the preimage of every element of B is an element of A and likewise for O and N and two maps are equivalent if they differ on an element of N.

This definition is too general to be useful for measure theory. Once we restrict ourselves to the subcategory of localizable measurable spaces (all major theorems of measure theorem such as Riesz representation theorem and Radon-Nikodym theorem imply the property of localizability) the resulting category becomes contravariantly equivalent to the category of commutative von Neumann algebras, also known as W*-algebras. In my opinion this constitutes the best possible definition for the main category of measure theory, both in terms of conceptuality and effectiveness, just as the best way to define the category of affine schemes is to make it equal to the opposite category of the category of commutative rings. Such a viewpoint is unfortunately highly unlikely to be adopted by analysts (especially hard analysts) considering their unwillingness to study even the most elementary notions of category theory.

• I think the viewpoint taken by you, ie represent a compact Hausdorff space by its $C^*$-algebra, and declare a measure to be a functional on it(like in Riesz Representation theorem) is what is adopted by noncommutative geometers, for their noncommutative analogies. So it is not true that analysts are not willing to take up category theory. Still, the base of the subject is in good old Lebesgue theory, and that must be first carried out in the traditional fashion. That is what analysis is all about. The essence of analysis is contained in that type of stuff, not in category theory. Jan 14, 2010 at 13:58
• @Anweshi: I doubt that measure theory should be considered a part of analysis at all. For example, smooth manifolds were once considered part of analysis (think of multivariable calculus) but now they are not. The subject became much more clear and conceptual when it was detached from analysis. (Of course we still sometimes use analysis to prove theorems like Calabi-Yau theorem, but such proofs are not considered final and in the end of the day they will be replaced by more conceptual and geometric proofs.) Measure theory will undergo the same transition. Jan 14, 2010 at 15:01
• “…is what is adopted by noncommutative geometers, for their noncommutative analogies. So it is not true that analysts are not willing to take up category theory.” Noncommutative geometry is a part of geometry, not analysis. From Connes' book it is clear that the driving force behind NCG is topology and geometry, not analysis, e.g., we have noncommutative Chern character, noncommutative smooth manifolds and differential forms, noncommutative K-theory etc. All these notions come from geometry and topology, not analysis. Jan 14, 2010 at 15:08
• You seem to be saying that since category theory cannot see the difference between a measure and its completion, then there is no mathematically substantive difference. But surely huge structural differences between the Borel sets and the Lebesgue measurable sets of reals, say, are revealed by descriptive set theory. Jan 14, 2010 at 22:36
• Perhaps that is an unneccessarily narrow view of what measure theory is? After all, perhaps some day a measure theorist will want to take a continuous image of a Borel set...and then find that she is actually a set theorist! :-) Jan 16, 2010 at 4:05

In light of the comments here, I'm going to show why completeness can be a pain. In exercise 9 of section 2.1 of Folland, he develops a function $g: [0,1] \to [0,2]$ by $g(x) = f(x) + x$ where $f : [0,1] \to [0,1]$ is the Cantor function. In that exercise it is established that $g$ is a (monotonic increasing) bijection, and that its inverse $h = g^{-1}$ is continuous from $[0,2]$ to $[0,1]$.

Since $h$ is continuous, it is Borel measurable. On the other hand, $h$ is not $(\mathcal{L}, \mathcal{L})$-measurable!! In particular, let $C$ be the Cantor set; $m(g(C)) = 1$, but this means there is a subset $A \subseteq g(C)$ which is not Lebesgue measurable. On the other hand $B := g^{-1}(A) \subseteq C$ whereas $m(C) = 0$; thus this preimage $B$ is Lebesgue measurable (with measure zero). But therefore $h^{-1}(B) = A$ is not Lebesgue measurable, meaning $h$ is not $(\mathcal{L}, \mathcal{L})$-measurable.

On one hand, this function is contrived. On the other hand, it shows that completing measures can mess things up. The typical definition of "measurable function" is a Borel measurable function, and I suppose reasons like the above led to this convention. I do not know the material Bridge references above, and so can't say what breaks when completeness is dropped. Although it seems mathematically convenient to throw in completeness, I don't know any examples in basic probability theory where it helps. For instance, Fubini-Tonelli can be formulated just fine without completeness. Your statement of the theorem only need mention completeness if your measures happen to be complete!

EDIT I corrected the nonsense in the second paragraph; also I meant to talk about $(\mathcal L, \mathcal L)$-measurable functions, which I accidentally refered to as Lebesgue measurable (which means $(\mathcal L, \mathcal B)$-measurable). My whole point is that if you take completion in $\sigma$-algebra of the range space, the extra sets you added could map back to basically anything. IE it is somewhat nonsensical to add in all sorts of null sets, but not all sorts of finite measure sets. Sometimes completion gives you something you want, but sometimes it does not, as I showed here--the function is better behaved wrt the non-completed measure.

• I think you mean g^{-1}(g(C))=C, and not g^{-1}([0,2])=C, since clearly g^{-1}([0,2])=[0,1]. But the former supports the rest of your argument. Jan 13, 2010 at 14:13
• “On the other hand, it shows that completing measures can mess things up.” I disagree. The function g is a perfectly good morphism between two complete measurable spaces. The codomain of g is the interval [0,2] equipped with the standard σ-algebra of Lebesgue measurable sets and the standard σ-ideal of Lebesgue sets of measure 0, whereas the domain of g is the completion of the following measurable space: The interval [0,1] is eqipped with the standard σ-algebra of Borel sets and for the σ-ideal of sets of measure 0 we take all Borel sets whose image under the map g has measure 0. Jan 13, 2010 at 16:56
• @dmitri i agree with what you say, in fact what i'm trying to say is that the "naive" completion of just throwing in all subsets of null sets is a bad idea (this is the standard definition). on the other hand, i've never heard of a recipe for the kind of completion you describe. can you send me some pointers? Jan 14, 2010 at 15:37
• @Matus: So the point is that one should be careful about what one means by “null sets”. Concerning references, I have been looking for a good reference myself for quite a while. Meanwhile you can take a look at Fremlin's Measure Theory (Volume 3) or at Takesaki's Theory of Operator Algebras (Volume 1, Chapter III.1) for an exposition of the ideas that I mentioned in my comments. Jan 14, 2010 at 15:57
• @Dmitri: I will look into this! Thanks very much. I'll close by saying that, in my experience, Borel $\sigma$-algebras are adequate/nice on the reals, but where completeness is convenient is in stochastic processes. Perhaps some of the notions you describe can be used to give a clean construction of them? (completeness feels like a gross bandaid to me.) Jan 14, 2010 at 16:10

Hi Tom E,

For Stochastic Processes, completion of the initial sigma-field of the natural filtration of the process with the negligeable sets of the Probability measure of the limit sigma-field of the filtration (coupled with right continuity of the underlying filtration) is really uselfull.

It allows to find version of processes that are càdlàg (right continuous with left limit) under very general conditions. Càdlàg processes are the main object of study in Stochastic Processes analysis (only my point of view).

As a matter of fact, this is so usefull, that those two conditions are called the "usual conditions" for the probability space and filtration on which process lives (the 3-tuple $(\Omega, (\mathcal{F}_t),P)$ is named a stochastic basis).

If interested, you can have a look at Karatzas and Shreve's book on Brownian motion and Stochastic Calculus, but even if I realise that the matter might be quite far from your day-to-day mathematical activities this is certainly an example which shows how usefull completion of sigma field might be.

Regards

Hi Tom E,

Here is another reason: Let $E$ be a Borel set in Euclidean space. Then its image under a continuous map is always Lebesgue-measurable but in general not Borel measurable. Results like this make the completion useful; The theory behind this is the theory of analytic sets or the Souslin operation.

What's missing here is the Caratheodory extension process creates a complete measure space. Hence, to have a product measure that is not complete requires one use a different method to create it. If this is the case, then completeness is not required in Fubini or Tonelli, e.g. (R, A, mu), (R, B, nu) Borel measure spaces and product measure defined as mu X nu for sigma-algebra A X B. However, this is not the traditional way we construct a measure and in particular would not construct the Lebesgue measure on R2. However, if we take two measure spaces which are both not complete and create their product measure using the Caratheodory extension of mu X nu (a complete measure), the equality in the conclusion of Fubini and Tonelli would not necessarily hold.

Hence, completeness of one or both measure spaces in the hypotheses is only important (required) if we want to construct our product measures using the Caratheodory extension process or if our product measure is complete.