4
$\begingroup$

Freed's notes give the following definition of oriented bordism.

Definition. Let $\Sigma_0$ and $\Sigma_1$ be the two oriented closed manifolds. An $ n $-dimensional bordism from $\Sigma_0$ to $\Sigma_1$ is a triple $(M, i_0, i_1)$ where

  1. $M$ is an $ n $-dimensional oriented compact manifold with boundary, whose boundary $ \partial M $ inherits the boundary orientation and where we defined a partition of the connected components of $ \partial M $ into in-boundaries and out-boundaries;
  2. $ i_0 $ and $ i_1 $ are smooth orientation preserving embeddings $$ i_0 \colon \Sigma_0 \times [0,\varepsilon[ \to M \qquad i_1 \colon \Sigma_1 \times ]1-\varepsilon, 1] \to M $$ that define (orientation preserving) diffeomorphisms $ \Sigma_0 \cong i_0(\Sigma_0,0)\cong \partial_0M $ and $ \Sigma_1\cong i_1(\Sigma_1,1)\cong \partial_1M $ between the "abstract" in- and out-boundaries $ \Sigma_0 $ and $ \Sigma_1 $, and the (disjoint union of) the chosen in- and out-boundaries $ \partial_0M $ and $ \partial_1M $.

Now, take an oriented $ 2 $-dimensional closed manifold $ \Sigma_0 $ and let's try to define a bordism from $ \Sigma_0 \sqcup \overline{\Sigma}_0 $ to the null manifold $ \emptyset^2 $, where $ \bar\Sigma_0 $ is just $ \Sigma_0 $ with its orientation reversed.

Let $ M $ be the oriented manifold $ M = \Sigma_0 \times [0,1] $, where $ [0,1] $ carries its canonical "from left to right" orientation and where the boundaries of $ M $ acquire the induced "Stokes" orientation [1].

To define an oriented bordism the two maps $$ i_0\colon \left(\Sigma_0\sqcup \bar\Sigma_0\right)\times [0,\epsilon[\to M\qquad i_1\colon \emptyset^2\times [1 - \epsilon, 1[\to M $$ must be defined. Here $ (\Sigma_0 \sqcup \bar{\Sigma}_0) \times [0,\varepsilon[ $ is nothing but the disjoint union of two (times an unspecified number of) cylinders. One of them, $ \Sigma_0 \times [0,\varepsilon[ $, has the same orientation as $ M $, hence I can embed it into $ M $ in an orientation preserving way. On the other hand, the remaining one $ \bar\Sigma_0 \times [0,\varepsilon[ $ has the opposite orientation, and of course I cannot.

Is there something I'm missing? I see basically no other way of defining a bordism from $ \Sigma_0\sqcup \bar \Sigma_0 $ to the null manifold (and it's quite clear that one must exist).


[1] We can depict the manifold $ M $ as a cylinder with two inward-pointing arrows stemming from its bases.

$\endgroup$

1 Answer 1

6
$\begingroup$

The issue arises from incorrectly identifying the boundary orientation. For the cylinder $M = \Sigma_0 \times [0,1]$ with the product orientation, the boundary orientation is $\partial M = (-\Sigma_0) \sqcup \Sigma_1$, where $\Sigma_1$ is $\Sigma_0$ with the same orientation.

To get a cobordism from $\Sigma_0 \sqcup \bar{\Sigma}_0$ to $\emptyset$, you need $\partial M = (-\Sigma_0) \sqcup (-\bar{\Sigma}_0) = (-\Sigma_0) \sqcup \Sigma_0$. This is achieved by taking $M = \Sigma_0 \times [0,1]$ but declaring both boundary components as in-boundaries. Then the embeddings $i_0$ map $\Sigma_0$ to one boundary (which appears as $-\Sigma_0$) and $\bar{\Sigma}_0$ to the other (which appears as $-\bar{\Sigma}_0 = \Sigma_0$). Both maps are orientation-preserving because the boundary orientation automatically reverses the orientation of incoming boundaries relative to their object orientation. Your desired cobordism $M : \Sigma_0 \sqcup \bar{\Sigma}_0 \to \emptyset$ exists by taking the cylinder with both ends as inputs.

$\endgroup$
5
  • $\begingroup$ Can you clarify what you mean by $ \Sigma_0 $, $ \bar \Sigma_0 $, $ -\Sigma_0 $ and $ -\bar\Sigma_0 $? I suspect $ -\Sigma_0 = \bar\Sigma_0 $ so that $ -\bar\Sigma_0 = \Sigma_0 $, am I right? $\endgroup$ Commented yesterday
  • $\begingroup$ Yes. So $-\Sigma_0$ denotes the manifold $\Sigma_0$ with the orientation reversed by the boundary convention, so $-\Sigma_0 = \bar{\Sigma}_0$ and thus $-\bar{\Sigma}_0 = \Sigma_0$. And in the cylinder $M = \Sigma_0 \times [0,1]$, the boundary orientation give $\partial M = (-\Sigma_0) \sqcup \Sigma_0 = \bar{\Sigma}_0 \sqcup \Sigma_0$. By declaring both ends as in-boundaries, this matches the required $\Sigma_0 \sqcup \bar{\Sigma}_0$ (up to reordering), making it a valid cobordism to $\emptyset$. $\endgroup$ Commented 15 hours ago
  • $\begingroup$ It's definitely more clear now, thank you. But one thing still nags me. Let's call the orientation on $ M = \Sigma \times [0,1] $ that makes $ \partial M = (-\Sigma_0) \sqcup \Sigma_0 $ the positive one, and let's try to to define $ i_0 \colon (\Sigma_0 \sqcup \bar\Sigma_0)\times [0, \varepsilon[ \to M $. Then $$ (\Sigma_0\sqcup \bar\Sigma_0) = (\Sigma_0\times [0, \varepsilon[) \sqcup (\bar\Sigma_0\times [0, \varepsilon[). $$ so that we can actually define two maps and then stick them together using the universal property of the coproduct. (Continues below) $\endgroup$ Commented 11 hours ago
  • $\begingroup$ Clearly, then, $ \Sigma_0 \times [0, \varepsilon[ $ is "positive-oriented" cylinder with boundary equal to $ -\Sigma_0 $ that can be embedded into the "$ -\Sigma_0 $" boundary component of $ M $ by one of these maps. On the other hand, $ \bar\Sigma_0 \times [0, \varepsilon[ $ is a "negative-oriented" cylinder with boundary equal to $ \Sigma_0 $. Where do I embed this? I.e., how do I define an orientation-preserving ("collar") map $ \bar\Sigma_0 \times [0, \varepsilon[\to M $? $\endgroup$ Commented 11 hours ago
  • $\begingroup$ Well, the resolve is that for an in-boundary, the standard collar map is actually orientation-reversing in the interval direction. The correct embedding for the $\bar{\Sigma}_0$ component is $i(x,t) = (x, 1-t)$ which uses a decreasing parameterization of the interval. This ensures the map is orientation-preserving when the boundary is declared as incomning. The "outward normal first" convention for the boundary orientation then forces the collar parameter to run backwards for in-boundaries. $\endgroup$ Commented 5 hours ago

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.