Help with a proof regarding convergent sequence (proof by contradiction)

malawi_glenn
Science Advisor
Messages
6,735
Reaction score
2,433
Ok I am trying to brush up my real analysis skills so that I can study some topology and measure theory at some point.

I found this theorem in my notes, that is proven by using proof by contradiction. However, I have a hard time understanding what the contradiction really is...

Here is the theorem and the proof.
- - - - - - - - - - - - - - - - - - - - - - - -
Let ##(x_n)## be a convergent sequence. Assume that there exists a natural number ##N## such that ##x_n \leq A## (where ##A## is some real number) holds for each ##n \geq N##. Then the following holds: ##\lim_{n \to \infty} x_n \leq A##.

Proof: Assume the contrary, that ##\lim_{n \to \infty} x_n = L > A##. Let ##\epsilon = L - A##. Since ##(x_n)## is convergent, all except a finite number of elements in the sequence belongs on the interval ##(L - \epsilon , L + \epsilon)##. But all numbers on this interval is strictly greater than ##A = L - \epsilon##, which is a contradiction.
- - - - - - - - - - - - - - - - - - - - - - - -

I have no idea what is contradicted and why this is a contradiction. Thanks in advance for any kind of illumination on this theorem and proof.
 
Physics news on Phys.org
The contradiction is that your initial statement is ##x_n \leq A## (except maybe a finite number of cases). But if you assume ##\lim x_n > A##, then you need an infinite number of cases where ##x_n > A##.
This is a contradiction and therefore you must have ##\lim x_n \leq A##
 
  • Like
Likes malawi_glenn
drmalawi said:
Ok I am trying to brush up my real analysis skills so that I can study some topology and measure theory at some point.

I found this theorem in my notes, that is proven by using proof by contradiction. However, I have a hard time understanding what the contradiction really is...

Here is the theorem and the proof.
- - - - - - - - - - - - - - - - - - - - - - - -
Let ##(x_n)## be a convergent sequence. Assume that there exists a natural number ##N## such that ##x_n \leq A## (where ##A## is some real number) holds for each ##n \geq N##. Then the following holds: ##\lim_{n \to \infty} x_n \leq A##.

Proof: Assume the contrary, that ##\lim_{n \to \infty} x_n = L > A##. Let ##\epsilon = L - A##. Since ##(x_n)## is convergent, all except a finite number of elements in the sequence belongs on the interval ##(L - \epsilon , L + \epsilon)##. But all numbers on this interval is strictly greater than ##A = L - \epsilon##, which is a contradiction.
- - - - - - - - - - - - - - - - - - - - - - - -

I have no idea what is contradicted and why this is a contradiction. Thanks in advance for any kind of illumination on this theorem and proof.
It's quite neat to do it that way. Can you do a different proof? Along similar lines.
 
I think that you should fill in some details in a more step-by-step manner.
Let ##N_1## denote the integer such that ##|x_n -L| \lt \epsilon## ## \forall n\gt N_1##. Let ##N' = max\{N_1, N\}##. If ##n \gt N'##, what can you say about ##x_n##?
Although this amount of detail might seem unnecessary, it is good practice for problems that are more complicated.
 
  • Like
Likes malawi_glenn
There's another useful technique, as follows:

Let ##\epsilon > 0##. Show that ##L \le A + \epsilon##. Then use the result (you could prove this separately) that: $$\text{If} \ \forall \ \epsilon > 0 \ \text{we have} \ a \le b + \epsilon, \ \text{then} \ a \le b$$
 
  • Like
Likes malawi_glenn

Similar threads

Replies
1
Views
3K
Replies
5
Views
2K
Replies
17
Views
1K
Replies
6
Views
1K
Replies
8
Views
2K
Replies
77
Views
12K
Back
Top