Length
When the interval has a length of . In this section we investigate the extent to which we can assign a length to other subsets of .
Outer measure
If a set is contained within a union of intervals, it seems entirely reasonable that any notion of length we might come up with would say the following: the length of is not larger than the sum of the lengths of the intervals.
That is, if
then
holds.
Our first major idea is that, if we choose the intervals wisely, then there may be very little discrepancy between the set and the union of the intervals; in that case the sum
could be considered very close to the length of .
The above idea is formalized mathematically as follows. We define the outer measure of a set by finding the most efficient way of covering it by a union of intervals.
Thus we define
for all sets . It is the infimum that picks out the most efficient way of covering by intervals. Notice that, in the definition, we allow not only unions of a finite number of intervals, but unions of an infinite number as well.
The function is defined on . It assigns a value in to every subset of . It has the following properties, which are very reasonable when thinking of as assigning length to subsets of .
Theorem
The map has the following properties.
- .
- whenever .
- for all .
- for all .
Proof:
First we prove (1). We must show that
has an infimum of zero. Since every member of the set is non-negative we can do this by producing a value in the set smaller than any given . Since any sequence of intervals covers the empty set, we have complete freedom to choose the endpoints, and can for example take for all and for all . For this choice we have
and therefore belongs to the above set. Since every member of the set is non-negative and was arbitrary, the infimum must be zero.
For (2) fix and with . To prove that one infimum is smaller than another, it suffices to prove that one set is larger than the other. Thus, if we can show that
then we will immediately have . But this containment is immediate as any sequence of intervals that covers must cover the smaller set .
Next we prove (3). Fix . For each let
be a cover of with
and let be a sequence of intervals that enumerates all the intervals for all . Certainly they cover the union of the and therefore
which gives what we want as was arbitrary.
Lastly, for (4) note that the translate of a cover of is a cover of and vice versa.▮
Countable additivity
The properties we have proved for are reasonable if we think of as assigning a length. Here is a further reasonable property that we should expect from a device that assigns length.
Countably additive Say that is countably additive if one has
whenever is a sequence of subsets of that is pairwise disjoint.
Recall that a sequence of sets is pairwise disjoint if for all . Informally, a mapping from to is countably additive if we can calculate by breaking up into countably many pieces and summing up the values
We are now faced with a central question: is is countably additive?
Example (Vitali set)
Define an equivalence relation on by if and only if . Let be a set that contains exactly one member of every equivalence class. Any such set is called a Vitali set.
We can use Vitali sets to prove that is not countably additive!
Theorem
The map is not countably additive.
Proof:
Fix a Vitali set . First we show that and are disjoint for every . Indeed, suppose that and . Then and . But so we have contradicted the defining property of .
Now suppose that is countably additive. We want to reach a contradiction. First we verify that
holds. For the second containment, since we certainly have for every . For the first containment fix . There is by definition of some with . Thus for some . We therefore have as desired.
Applying to the above gives
and since is assumed countably additive we have
after applying to each side. Finally
because for all . But the above is impossible no matter the purported value of .▮
Next steps
We have seen, from Vitali sets and the above theorems, that there is no mapping
that has all three of the following properties.
- Countably additive
- Assigns lengths to intervals
- Translation invariant
In order to proceed with a theory of length, we have to give something up. We are going to insist on translation invariance and countable additivity by abandoning the requirement that is defined on . Our next step, therefore, is to try to identify a rich collection of subsets of so that
is countably additive and translation invariant.