Chapter 2. Lebesgue Measure: Section 2.1. Introduction
Chapter 2. Lebesgue Measure: Section 2.1. Introduction
Note. We “weigh” an interval by its length when setting up the Riemann integral.
So to generalize the Riemann integral, we desire a way to weigh sets other than
intervals. This weight should be a generalization of the length of an interval.
Note. Since we know an open set is a countable union of disjoint open intervals,
we would define its “weight” (or “measure”) to be the sum of the lengths of the
open intervals which compose it.
Note. We want a function m which maps the collection of all subsets of R, that
is the power set of the reals P(R), into R+ ∪ {0, ∞} = [0, ∞]. We would like m to
satisfy:
1. For any interval I, m(I) = `(I) (where `(I) is the length of I).
2. For all E on which m is defined and for all y ∈ R, m(E + y) = m(E). That
is, m is translation invariant.
3. If {Ek }∞
k=1 is a sequence of disjoint sets (on each of which, m is defined), then
P
m(∪·Ek ) = m(Ek ). That is, m is countably additive.
4. m is defined on P(R).
Note. We will see in Section 2.6 that there is not a function satisfying all four
properties. In fact, there is not even a set function satisfying (1), (2), and (4)
Xn
n
for which m (∪·k=1 Ek ) = m(Ek ) for disjoint Ek (this property is called finite
k=1
additivity). See Theorem 2.18 for details.
Note. It is “unknown” whether m exists satisfying properties (1), (3), and (4) (if
we assume the Continuum Hypothesis, then there is not such a function).
Note. We will weaken Property (4) and try to find a function defined on as large
a set as possible. We will require (by (3)) that our collection of sets, M, on which
m is defined, be countably additive and therefore M will be a σ-algebra.
Revised: 8/28/2020