Uncountably many 8s

Geometry Level 5

On the Cartesian plane, is it possible to place an uncountable number of figure 8's (two circles that are externally tangential) that do not intersect each other?


Note: An infinite set is uncountable if there exists no bijection to the integers.

Yes No

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Jul 15, 2017

There are many solutions to be found for this one. The easiest I know of is this...

Each loop of a figure- 8 8 is open, and so must contain a point with rational coordinates.

If we have a collection S \mathcal{S} of disjoint figure- 8 8 s, we can define a map f : S Q 2 × Q 2 f \,:\, \mathcal{S} \to \mathbb{Q}^2 \times \mathbb{Q}^2 , where f ( E ) = ( q 1 , q 2 ) f(E) = (\mathbf{q}_1,\mathbf{q}_2) , where q 1 \mathbf{q}_1 is a point with rational coordinates inside one loop, and q 2 \mathbf{q}_2 is a point with rational coordinates inside the other loop of the figure- 8 8 E E .

Suppose that E 1 , E 2 S E_1,E_2 \in \mathcal{S} are distinct and such that f ( E 1 ) = f ( E 2 ) = ( q 1 , q 2 ) f(E_1) = f(E_2) = (\mathbf{q}_1,\mathbf{q}_2) . Since E 1 E_1 and E 2 E_2 are disjoint, the loop of E 2 E_2 that contains q 1 \mathbf{q}_1 is either totally inside the loop of E 1 E_1 that contains q 1 \mathbf{q}_1 , or else contains the whole of E 1 E_1 . Similarly, the loop of E 2 E_2 that contains q 2 \mathbf{q}_2 is either totally inside the loop of E 2 E_2 that contains q 2 \mathbf{q}_2 , or else contains the whole of E 1 E_1 . The only way in which these two loops can be mutually tangent is if they both contain the whole of E 1 E_1 . But, in this case, both q 1 \mathbf{q}_1 and q 2 \mathbf{q}_2 are inside both loops, which means that E 2 E_2 is not a figure- 8 8 . This is impossible.

Thus we deduce that f f is injective, and hence that S \mathcal{S} is countable.

@Sharky Kesa What's with the figure 8? Daniel Liu asked the same question.

Agnishom Chattopadhyay - 3 years, 11 months ago

Log in to reply

Well, figure- 8 8 s are more interesting than circles (we can have an uncountably infinite disjoint number of these); they have measure zero, but are complex enough for it only to be possible to have a countable number of disjoint ones...

Mark Hennings - 3 years, 11 months ago

Log in to reply

we can have an uncountably infinite disjoint number of these

Really, that's interesting. How so?

Agnishom Chattopadhyay - 3 years, 11 months ago

Log in to reply

@Agnishom Chattopadhyay Centre ( 0 , 0 ) (0,0) , radius r r for all r > 0 r > 0 .

Mark Hennings - 3 years, 11 months ago

Log in to reply

@Mark Hennings Okay, I did not think this through

Agnishom Chattopadhyay - 3 years, 11 months ago

Is there a subset of R 2 \mathbb{R}^2 which has positive Lebesgue measure, and yet, uncountably many isomorphic, but non-intersecting copies of it exist?

Agnishom Chattopadhyay - 3 years, 11 months ago

Log in to reply

No. If the shape has positive measure, then any bounded subset, such as [ N , N ] × [ N , N ] [-N,N] \times [-N,N] , of R \mathbb{R} can only contain finitely many (if they all have the same measure) or countably many (if they can be scaled in size). To see this last, there can only be finitely many with measure greater than n 1 n^{-1} for any integer n n . Now take the union of these over n n .

Thus there can only be countably many in the whole of R 2 \mathbb{R}^2 .

Where it gets interesting is with shapes of measure zero, like the figure- 8 8 .

Mark Hennings - 3 years, 11 months ago

Simply out of curiosity, is there any way to do this without implicitly invoking the axiom of choice. Any finitely-large circle is going to have an infinite number of choices for q 1 q_1 , and I cannot deduce any simple rule that could easily select a rational number given an arbitrary circle. Would it be possible to set up a version of this proof that holds without invoking the axiom of choice?

Kyle Coughlin - 3 years, 10 months ago

Log in to reply

I shouldn't think so. The standard proof that you cannot have uncountably many disjoint open disks works by the same method (of choosing a point with rational coordinates inside each disk), and therefore also requires the axiom of choice.

Mark Hennings - 3 years, 10 months ago

Is Q^2 x Q^2 countable? I guess so since your answer doesn't make sense if it isn't, but I realized while considering this that I didn't understand the diagonal slash properly.

Matt McNabb - 3 years, 10 months ago

Log in to reply

Any Cartesian product of countable sets is countable. If f : A N f:A\to\mathbf{N} and g : B N g:B\to\mathbf{N} are injective, so is h : A × B N h:A\times B\to\mathbf{N} given by h ( a , b ) = 2 f ( a ) 3 g ( b ) h(a,b)=2^{f(a)}3^{g(b)}

Mark Hennings - 3 years, 10 months ago

Just a doubt....since the cartesian plane extends infinitely cant we place infinite figures of 8 that do not intersect....?

Ben John Toms - 3 years, 6 months ago

Log in to reply

Infinite, yes. Uncountably infinite, no.

Mark Hennings - 3 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...