Cardinality of continuous functions f:R->R.

AI Thread Summary
The discussion focuses on determining the cardinality of the set of continuous functions from R to R. It is established that this cardinality is less than or equal to 2^c, where c is the continuum cardinal. Participants explore the challenge of showing that the cardinality is also greater than or equal to 2^c, with suggestions to construct subsets of continuous functions to demonstrate this. The conversation highlights that any continuous function is determined by its values on the rationals, leading to an upper bound for the cardinality. Ultimately, it is concluded that the cardinality of continuous functions is indeed 2^c, with a reference to counting constant functions as part of the proof.
MathematicalPhysicist
Science Advisor
Gold Member
Messages
4,662
Reaction score
372
i need to find the cardinality of set of continuous functions f:R->R.
well i know that this cardinality is samaller or equal than 2^c, where c is the continuum cardinal.
but to show that it's bigger or equals i find a bit nontrivial.
i mean if R^R is the set of all functions f:R->R, i need to find a 1-1 function from it to the set of continuous functions (lets call it A).
i tried this function g:R^R->A:
g(f)=f if f is continuous function.
g(f)=h if f isn't continuous on R.
where h is a continuous function on Q.
i feel that again i did something wrong my problem is how to correct it, obviously i didnt defined h as i should be, and what's the connection of it to f, but i don't see a way to pass it.
any help?
 
Physics news on Phys.org
You do not have to find a injection from R^R, the set of all functions. A map from any set with that cardinailty will do.

About the best way of doing this (again), is to construct some subset of all continuous functions with cardinality 2^c. It might take some ingenuity, but it isn't that hard (just think of Taylor series).
 
Last edited:
any suggestions?
i mean possible candidates are:
{0,1}^[0,1] P([0,1]) P(R)
i think the third set is the easiest, but gain finding the function is hard task here.
i need to find a function from P(R)->A
perhaps we can mapp from a subset of P(R), say B to a function in A, perhaps f|B (which is a continuous function f:B->R), will that suffice or i need something else?
 
Reread the edited post above.
 
i don't understand, we need to show that the cardinality of A is 2^c, iv'e showed that it's smaller or equals 2^c, now i need to show that it's bigger or equals 2^c.
and you suggest me to:"About the best way of doing this (again), is to construct some subset of all continuous functions with cardinality 2^c"
the only subset i can think of is the set of all functions f:R\Q->R\Q, it's cardinality is 2^c, but is it a subset, i mean every function in this subset is a continuous function cause f(R\Q) is a subset of R\Q.

p.s
i feel I am writing jibberish. )-:
 
well i don't see the connection to here.
i mean f(x)=f(x0)+(x-x0)f'(x0)+(x-x0)^2f''(x0)/2+...
you mean the set of functions which can be written as a taylor expansion as a subset to the set of all continuous functions from R to R.
or perhaps the set of functions of different estimations to f, such as:
f(x0)+(x-x0)f'(x0),f(x0)+(x-x0)f'(x0)+(x-x0)^2f''(x0)/2,...
 
I don't see how Taylor series will help. The most obvious interpretation of that hint shows that there are c|N| analytic functions, and although analytic functions are continuous, c|N| < 2c.
 
Yep. That's why I pulled out to think about it some more.
 
so do you have any advice on this question, perhaps the cardinality isn't 2^c?
 
  • #10
loop quantum gravity said:
so do you have any advice on this question, perhaps the cardinality isn't 2^c?

I didn't think it was -- I thought the restriction of continuity made it c -- but I don't have a proof of this offhand. Let me think about it.
 
  • #12
Oh, it turns out the proof's easy -- any continuous function is completely determined by its values on Q. So,

|C(R, R)| <= |RQ| = |R||Q| = (2|N|)|Q| = 2|N||Q| = 2|NxQ| = 2|N| = |R|
 
Last edited:
  • #13
Duh... And there was me lamenting that analysis isn't much help 'cos I couldn't think of a way of making (different) density arguments work.
 
  • #14
hurkyls argument gives only an upper bound for the cardinality. you have to show it achieves this bound too. but maybe i came in late and missed this part of the discussion.
 
  • #15
mathwonk said:
hurkyls argument gives only an upper bound for the cardinality. you have to show it achieves this bound too. but maybe i came in late and missed this part of the discussion.
Well this is entirely obvious, just count the constant functions.

Nice proof Hurkyl.
 

Similar threads

Back
Top