I've come across Konig's Theorem in Jech, and I'm having some trouble understanding the proof which can be found here. I don't understand what Jech means by S_i, the projection into the ith coordinate. Can someone give a basic example so I can wrap my head around it? Thanks.


For a very simple example, suppose that $I=\{0,1,2\}$. Let $T_0,T_1$, and $T_2$ be any sets, let $T=\prod_{i\in I}T_i$, and let $Z\subseteq T$. By the definition of product, the elements of $T$ are the functions $f$ with domain $I$ such that $f(i)\in T_i$ for each $i\in I$. For example, if $T_0=\{0,1\}$, $T_1=\{2,3\}$, and $T_2=\{4,5\}$, there are exactly $2^3=8$ functions in $T$:

$$\begin{align*} &0\mapsto 0,1\mapsto 2,2\mapsto 4\\ &0\mapsto 0,1\mapsto 2,2\mapsto 5\\ &0\mapsto 0,1\mapsto 3,2\mapsto 4\\ &0\mapsto 0,1\mapsto 3,2\mapsto 5\\ &0\mapsto 1,1\mapsto 2,2\mapsto 4\\ &0\mapsto 1,1\mapsto 2,2\mapsto 5\\ &0\mapsto 1,1\mapsto 3,2\mapsto 4\\ &0\mapsto 1,1\mapsto 3,2\mapsto 5 \end{align*}$$

(Of course in the proof we’re dealing with much larger sets.)

If $f$ is any member of $T$, and $i\in I$, then by definition the projection of the single function $f$ on the $i$-th coordinate is $f(i)$. In our toy example suppose that $f$ is the function

$$0\mapsto 0,1\mapsto 3,2\mapsto 5\;;$$

the projection of $f$ to coordinate $0$ is $f(0)=0$, the projection of $f$ to coordinate $1$ is $f(1)=3$, and the projection of $f$ to coordinate $2$ is $f(2)=5$.

Now suppose that we have a set of such functions, i.e., a set $Z\subseteq T$; we might use the example $Z=\{f,g,h\}$, where $f$ is

$$0\mapsto 0,1\mapsto 3,2\mapsto 5\;,$$

$g$ is

$$0\mapsto 1,1\mapsto 2,2\mapsto 5\;,$$

and $h$ is

$$0\mapsto 1,1\mapsto 3,2\mapsto 4\;.$$

The projection of $Z$ to the $0$ coordinate is just $\{f(0),g(0),h(0)\}=\{0,1\}$, the set of projections to the $0$ coordinate of the members of $Z$.

Similarly, the projection of $Z$ to the $1$ coordinate is $\{f(1),g(1),h(1)\}=\{2,3\}$, and the projection of $Z$ to the $2$ coordinate is $\{f(2),g(2),h(2)\}=\{4,5\}$.

Projection here can be thought of in geometric terms. If $I=\{x,y\}$, and $T_x=T_y=\Bbb R$, then $T$ is essentially just $\Bbb R^2$, the plane. If $Z\subseteq\Bbb R^2$, we can project $Z$ onto the $x$-axis by forming the set of all $x$-coordinates of points of $Z$, as if we were using the $x$-axis as a projection screen. Similarly, the set of $y$-coordinates of points of $Z$ is the projection of $Z$ to the $y$-axis.

In my example I took the projections of a single subset of $T$ to each of the coordinates. In the proof of König’s theorem you actually have a set $Z_i$ for each coordinate $i\in I$, and you project the set $Z_i$ only to the $i$-th coordinate. In geometric terminology you have one set, $Z_i$ for each coordinate axis (each $i\in I$), and you project each of those sets to a different axis.

  • $\begingroup$ Wow thank you. This is extremely clear and exactly what I was looking for. $\endgroup$ – user2034 Feb 17 '15 at 14:06
  • $\begingroup$ @user2034: You're welcome; glad it helped. $\endgroup$ – Brian M. Scott Feb 17 '15 at 18:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.