Topology - The Classification of Surfaces Oxford Mathematics 2nd Year Student Lecture

The Opinion Lighthouse
32 followers
7 Video Views·Nov 7, 2022
Topology - The Classification of Surfaces Oxford Mathematics 2nd Year Student Lecture
Okay, now we're ready to finish prove the main theorem,
classification of closed surfaces.
And here are the four lemmas which I will be using,
which are put here for easy reference.
So, we start with polygon with a complete set of edge identifications,
and then we want to show that the corresponding surface is one of these.
So we start with a word, and case one is that there exists a letter
which appears twice in the word, both as x, not as x and x inverse,
or once as x inverse, which of course would be a bit stupid,
because we can redefine it to be x,
and then this case visibly reduces to that case.
So, this case is the case of non-orientable surfaces.
Okay, so our word is of the form, so up to cyclic permutation,
we assume that it starts with an x,
so x and then some word and the x and some other words.
So it is exactly of that form, x A, x B.
And by this lemma, so let's name these lemmas,
one, two, three, four, by lemma one,
by lemma one we have that the surface, which is the one we care about,
is, okay, so we rewrite it as x, x and then this thing,
and then that's the same thing as a connected sum of rp2 with some other surface.
And now we may invoke induction on the length of the word.
So by induction on the length of the word, this is a shorter word,
we may assume known the result of the theorem for this.
So by induction of length of the word, this is, well,
either S2 or a bunch of B2s or a bunch of rp2s.
So we have three subcases, and we get that our surface is homeomorphic to either S2,
so rp2 connect some S2, but connect some S2 does nothing,
because if S2, I remove a disk, then that's just,
I just have a disk and then you put, you're doing nothing.
So that's just rp2.
Or rp2 connect some n copies of T2,
which by lemma four, applied n times,
is the same thing as a connected sum of 2n plus 1 copies of rp2.
Or this is rp2 and n copies of rp2,
in which case that's just n plus 1 copies of rp2.
So in all cases, we're good.
And we have shown using the induction hypothesis that our surface is one of these.
So let's do the case two.
So case two, now we're assuming that no letter appears twice the same,
like every letter appears as x and then x inverse.
So this is the case of orientable surfaces.
So now I'm going to assume that there exists two letters, x and y,
which are interlocking.
And what do I mean by interlocking?
I mean that they appear, if you draw your word cyclically,
they appear not like that, but like that.
So something x and then y and then x inverse and then y inverse.
That's what it means that they're interlocking.
Or if you draw them on the surface of a circle,
then you have the x and the x inverse pair somewhere,
then the y and the y inverse pair somewhere.
And if you draw a line that connects them, these two lines cross.
That's what it means to be interlocking.
Okay, so we have our word like that.
And we may as well, by cyclic permutation, assume that it starts with x.
So here it is.
So it is x and then a and then y and then some other part
and then x inverse and then something and then y inverse and something.
And now we're going to apply lemma two a number of times.
So first we're going to apply lemma two.
So lemma two says that between an x and an x inverse,
I can perform a cyclic permutation.
So here, between x and x inverse, I can perform a cyclic permutation
to bring y close to x inverse.
So I have d, a, y, x inverse.
And here, that's also between an x and an x inverse.
It's between x inverse and this other one, which is x.
You can imagine it here.
And so now again, I'm going to put this y inverse close to this.
y inverse dc.
And then finally, I'm going to apply once more the same lemma two,
but now to this chunk here, which is between this y and this y inverse.
And by doing that, I'm going to do a cyclic permutation so that x is close.
So I have x, y, x inverse, y inverse.
And then the rest is x, well, it's d, a, dc, which I can, for example, put here.
So the manifold associated to this word
is homeomorphic to the manifold associated to this word,
is homeomorphic to the manifold associated to this word.
And these are all instances of lemma two.
And what is that?
Well, I have here the word for a thorus and here's some other word.
So this is the same thing as the manifold associated to the word.
This is the same thing as the torus.
We too connect some with the manifold associated to d, a, dc.
But once again, we may assume by induction on the length of the word.
And also we know that this is, this is not creating any x x's or x inverse x inverses.
So we're still in the orientable case.
This is either s2 or a bunch of t2's.
And so, well, if I connect some t2 with s2, that's just t2.
And if I connect some t2 with a bunch of t2's, that's a bunch of t2's.
So this is either t2 or a bunch of t2's.
And so great, we're again, well, so here in this case one,
we found that we are one of these.
In case two, we found that we're one of these.
And now case three is the case when there are no interlocking letters.
So, and if there are no interlocking letters,
if I draw the letters on the boundary of the polygon, just like I did here,
and I draw, whenever I see an x and an x inverse, I draw a line connecting them,
these lines never cross.
And so we have a pattern that looks a bit like that.
And in particular, we have the closest arcs, the arcs that are sort of closest,
the smallest arcs that correspond to edges that are in cancelling position.
We have y, y inverse.
So there exist letters to which one can apply lemma three.
And this brings us to a smaller word, which again has no interlocking letters,
so you cancel, cancel, cancel, cancel, cancel, until you're left with nothing.
Well, with the smallest possible word, which is just x, x inverse,
if you admit bygones, or if you don't admit bygones because you want edges to remain straight,
it's going to be like x, x inverse, y, y inverse, which is s2.
So, so this is case two.
And that's it.
That is, that is the proof of this theorem using these lemmas.
And that is also the end of this sequence of data points.
Thank you for watching.
