Littlewood's Three Principles

Spring 2021 ETH Zürich
Measure and Integration Maran Mohanarangan
Exercise Class 8
Littlewood’s Three Principles1
What does a measurable set or function look like? This is a somewhat futile question.
After all, the Borel hierarchy is huge. Carathéodory’s criterion of measurability gives
us an intuition what a measurable set is capable of doing, namely splitting up arbitrary
sets without changing its measure, however, it does not really tell us what a measurable
set looks like.
The notions of measurable sets and measurable functions represent new tools, how-
ever, they are still related to older concepts. Littlewood thus introduced three principles
of real analysis as heuristics to help understand the essentials of measure theory. This
exercise class is devoted to study and exposit them.
“The extent of knowledge [of real analysis] required is nothing like as great
as is sometimes supposed. There are three principles, roughly expressible in
the following terms:
1. Every set is nearly a finite sum of intervals.
2. Every function is nearly continuous.
3. Every convergent sequence is nearly uniformly convergent.”
- John Littlewood.
Naturally, the sets and functions referred to above are assumed to be measurable. The
catch is in the word “nearly”, which has to be understood appropriately in each context.
In this class, we make these statements precise.
For the remainder of this class, let µ denote the Lebesgue measure on R and Σ the
σ-algebra of µ-measurable sets.
Littlewood’s First Principle

Theorem. Every finite measurable set is nearly a finite sum of intervals. That is, if
E ∈ Σ and µ(E) < ∞, then for every ε > 0, there is a set F that is a finite union of
open intervals such that µ(E4F ) < ε.
Proof. Let E ∈ Σ and ε > 0. By Theorem 1.3.8. in the Lecture Notes, we know that
there exists an open set G ⊂ R such that E ⊂ G and
ε
µ(G \ E) < .
2
This shows in particular that µ(G) < ∞. Thus G can be written as a countable union
of disjoint open intervals {(an , bn ) | n ∈ N}. Furthermore, by countable additivity of µ,
we have
[∞ X∞
µ(G) = µ( (an , bn )) = µ((an , bn ))
n=1 n=1
or equivalently
k
X
µ(G) = lim µ((an , bn )).
k→∞
n=1
1
The notes this week closely follow [SS09].
Questions and Remarks to: [email protected] 1

The sum on the r.h.s. converges, so we may find N sufficiently large such that
N
X ε
µ(G) − µ((an , bn )) < .
2
n=1
Now, let F = N
S
n=1 (an , bn ). Then F ⊂ G and F is open as the finite union of open
intervals. We obtain that
ε
µ(F \ E) ≤ µ(G \ E) < .
2
Similarly, since E ⊂ G, we know that E \ F ⊂ G \ F and so it follows (by using
measurability of F ) that
N
X ε
µ(E \ F ) ≤ µ(G \ F ) = µ(G) − µ(F ) = µ(G) − µ((an , bn )) < .
2
n=1
Since E \ F and F \ E are disjoint, it follows from the disjoint additivity of µ that
ε ε
µ(E4F ) = µ (E \ F ) ∪ (F \ E) = µ(E \ F ) + µ(F \ E) < + = ε.
2 2
The proof can easily be generalized for Rd . One simply has to write open sets in
Rd as the union of disjoint (dyadic) cubes and then the proof can essentially be copied
from above.
The latter principles are more striking but Littlewood’s first principle is of help when
thinking about measurable sets. Additionally, it provides a straightforward proof of the
Riemann–Lebesgue lemma which states that the Fourier transform of an integrable
function vainshes at infinity.
Lemma (Riemann–Lebesgue). If f : R → R is Lebesgue integrable, then

Z
lim f (x)eikx dx = 0.
s→±∞ R
Proof. First, we prove the result for f = χI , where I is a bounded interval, say I = (a, b).
This is trivial, since
Z b
eika − eikb
eikx dx = → 0 for s → ±∞.
a ik
Just as trivially, the result extends to any finite union of bounded intervals.
Next, consider f = χE , where E is any Lebesgue measurable set with finite measure.
Let ε > 0, then by Littlewood’s first principle, there exists a finite union F of bounded
intervals such that µ(E4F ) < ε. By the first part of the proof, we have that
Z
eikx dx < ε

F
for |k| large enough, in which case it follows that

Z Z Z

ikx ikx ikx

e dx ≤ e dx + e dx < 2ε.

E F E4F

The result thus naturally extends to any simple function.

Finally, if f is integrable, then given ε > 0, there exists a simple integrable function
g so that Z

f (x) − g(x) dx < ε.
R
By what we have proved before, it holds that
Z
ikx

g(x)e dx < ε

R
when |k| is sufficiently large. This allows us to conclude as

Z Z Z
f (x)eikx dx ≤ (f (x) − g(x))eikx dx + g(x)eikx dx

R
ZR R

< f (x) − g(x) dx + ε < 2ε.
R
Littlewood’s Third Principle

Littlewood’s Third Principle states that every convergent sequence is nearly uniformly
convergent. This is better known as Egorov’s theorem.
Theorem (Egorov). Suppose (fk )k is a sequence of measurable functions defined on a
measurable set Ω with µ(Ω) < ∞, and assume that fk → f µ-a.e. on Ω. Given ε > 0,
we can find a compact set K ⊂ Ω such that µ(Ω \ K) < ε and fk → f uniformly on K.
Proof. See Theorem 2.3.1 in the Lecture Notes
The reason for introducing the Third Principle before the Second one will become
apparent in the next section.
Littlewood’s Second Principle

The result that “every function is nearly continuous” is better known as Lusin’s theorem.
Theorem (Lusin). Suppose f is measurable and finite valued on Ω with µ(Ω) < ∞.
Then for every ε > 0 there exists a closed set F with
F ⊂Ω and µ(Ω \ F ) < ε
and such that f |F is continuous.

By f |F we mean the restriction of f to the set F . The theorem states that if f is
viewed as a function defined only on F , then it is continuous. However, the theorem
does not make the stronger assertion that the function f defined on Ω is continuous at
the points of F .
The idea is to approximate f pointwise almost everywhere with a sequence of step
functions. Outside a set of small measure, the set functions are continuous. So then by
Egorov’s theorem, outside a set of small measure, f is the uniform limit of continuous
functions and so f is continuous.

Proof. Let fn be a sequence of simple functions so that fn → f µ-a.e. Then we can find
sets En so that µ(En ) < 2−n and fn is continuous outside of En . By Egorov’s theorem,
we can find a set K on which fn → f uniformly and µ(Ω \ K) < ε/3. Then we consider
[
F0 = K \ En
n≥N
for N so large that n≥N 2−n < ε/3. Now for every n ≥ N the function fn is continuous
P
on F 0 . Thus f being the uniform limit of (fn )n is also continuous on F 0 . To finish the
proof, we merely need to approximate F 0 by a closed set F ⊂ F such that µ(F 0 \ F ) <
ε/3, which can be easily done with Littlewood’s First Principle. It then follows that
µ(Ω \ F ) < ε and f |F is continuous.
Remark. Observe that we stated Egorov’s and Lusin’s theorem in less generality than
in the lecture. For simplicity, we assumed that f is finite everywhere.
Sources
[SS09] Elias M. Stein and Rami Shakarchi. Real Analysis. Princeton University Press,
2009. doi: doi:10.1515/9781400835560.

Littlewood's Three Principles

Uploaded by

Copyright:

Available Formats

Littlewood's Three Principles

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Littlewood's Three Principles

Uploaded by

Copyright:

Available Formats

Spring 2021 ETH Zürich

Measure and Integration Maran Mohanarangan

Littlewood’s First Principle

Questions and Remarks to: [email protected] 1

Lemma (Riemann–Lebesgue). If f : R → R is Lebesgue integrable, then

for |k| large enough, in which case it follows that

Questions and Remarks to: [email protected] 2

The result thus naturally extends to any simple function.

when |k| is sufficiently large. This allows us to conclude as

Littlewood’s Third Principle

Littlewood’s Second Principle

F ⊂Ω and µ(Ω \ F ) < ε

and such that f |F is continuous.

Questions and Remarks to: [email protected] 3

Questions and Remarks to: [email protected] 4

You might also like