Suppose $X$ is infinite and $A$ is a finite subset of $X$. Then $X$ and $X setminus A$ are equinumerous

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP











up vote
2
down vote

favorite













Suppose that $X$ is infinite and that $A$ is a finite subset of $X$. Then $X$ and $X setminus A$ are equinumerous.





My attempt:



Let $|A|=n$. We will prove by induction on n. It's clear that the the theorem is trivially true for $n=0$. Assume the theorem is true for all $n=k$. For $n=k+1$, then $|A setminus a|=k$ for some $a in A$. Thus $X setminus (A setminus a) sim X$ by inductive hypothesis, or $(X cap a) cup (X setminus A) sim X$, or $a cup (X setminus A) sim X$. We have $a cup (X setminus A) sim X setminus A$ since the theorem is true for $n=1$. Hence $X setminus A sim a cup (X setminus A) sim X$. Thus $X setminus A sim X$. This completes the proof.





Does this proof look fine or contain gaps? Do you have suggestions? Many thanks for your dedicated help!





Update: Here I prove that the theorem is true for $n=1$.



Assume that $A = a$ and consequently $X setminus A= X setminusa$. It's clear that $|X setminus A| le |X|$. Next we prove that $|X| le |X setminus A|$. Since $X$ is infinite, there exists $B subsetneq X$ such that $B sim X$ (Here we assume Axiom of Countable Choice). Thus $|X|=|B|$. There are only two possible cases.



  1. $a in X setminus B$

Then $B subseteq X setminus a=X setminus A$ and consequently $|X|=|B| le |X setminus A|$. Thus $|X| le |X setminus A|$ and $|X setminus A| le |X|$. By Schröder–Bernstein theorem, we have $|X setminus A| = |X|$. It follows that $X setminus A sim X$.



  1. $a in B$.

Let $b in X setminus B$. We define a bijection $f:X setminus a to X setminus b$ by $f(x)= x$ for all $x in X setminus a,b$ and $f(b)=a$. Thus $X setminus a sim X setminus b$. Since $b in X setminus B$, it follows from Case 1 that $X setminus b sim X$. Hence $X setminus a sim X setminus b sim X$. Thus $X setminus a = X setminus A sim X$.



To sum up, $X setminus A sim X$ for all $|A|=1$.










share|cite|improve this question























  • What's your definition of $X$ being infinite?
    – Paul K
    Sep 10 at 15:51










  • @PaulK $X$ is infinite if and only if $X$ is not finite.
    – Le Anh Dung
    Sep 10 at 15:55










  • @PaulK of infinite cardinality
    – Michał Zapała
    Sep 10 at 15:57










  • The proof is somewhat lacking in elegance, but formally correct
    – Michał Zapała
    Sep 10 at 15:58










  • @MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
    – Le Anh Dung
    Sep 10 at 16:08














up vote
2
down vote

favorite













Suppose that $X$ is infinite and that $A$ is a finite subset of $X$. Then $X$ and $X setminus A$ are equinumerous.





My attempt:



Let $|A|=n$. We will prove by induction on n. It's clear that the the theorem is trivially true for $n=0$. Assume the theorem is true for all $n=k$. For $n=k+1$, then $|A setminus a|=k$ for some $a in A$. Thus $X setminus (A setminus a) sim X$ by inductive hypothesis, or $(X cap a) cup (X setminus A) sim X$, or $a cup (X setminus A) sim X$. We have $a cup (X setminus A) sim X setminus A$ since the theorem is true for $n=1$. Hence $X setminus A sim a cup (X setminus A) sim X$. Thus $X setminus A sim X$. This completes the proof.





Does this proof look fine or contain gaps? Do you have suggestions? Many thanks for your dedicated help!





Update: Here I prove that the theorem is true for $n=1$.



Assume that $A = a$ and consequently $X setminus A= X setminusa$. It's clear that $|X setminus A| le |X|$. Next we prove that $|X| le |X setminus A|$. Since $X$ is infinite, there exists $B subsetneq X$ such that $B sim X$ (Here we assume Axiom of Countable Choice). Thus $|X|=|B|$. There are only two possible cases.



  1. $a in X setminus B$

Then $B subseteq X setminus a=X setminus A$ and consequently $|X|=|B| le |X setminus A|$. Thus $|X| le |X setminus A|$ and $|X setminus A| le |X|$. By Schröder–Bernstein theorem, we have $|X setminus A| = |X|$. It follows that $X setminus A sim X$.



  1. $a in B$.

Let $b in X setminus B$. We define a bijection $f:X setminus a to X setminus b$ by $f(x)= x$ for all $x in X setminus a,b$ and $f(b)=a$. Thus $X setminus a sim X setminus b$. Since $b in X setminus B$, it follows from Case 1 that $X setminus b sim X$. Hence $X setminus a sim X setminus b sim X$. Thus $X setminus a = X setminus A sim X$.



To sum up, $X setminus A sim X$ for all $|A|=1$.










share|cite|improve this question























  • What's your definition of $X$ being infinite?
    – Paul K
    Sep 10 at 15:51










  • @PaulK $X$ is infinite if and only if $X$ is not finite.
    – Le Anh Dung
    Sep 10 at 15:55










  • @PaulK of infinite cardinality
    – Michał Zapała
    Sep 10 at 15:57










  • The proof is somewhat lacking in elegance, but formally correct
    – Michał Zapała
    Sep 10 at 15:58










  • @MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
    – Le Anh Dung
    Sep 10 at 16:08












up vote
2
down vote

favorite









up vote
2
down vote

favorite












Suppose that $X$ is infinite and that $A$ is a finite subset of $X$. Then $X$ and $X setminus A$ are equinumerous.





My attempt:



Let $|A|=n$. We will prove by induction on n. It's clear that the the theorem is trivially true for $n=0$. Assume the theorem is true for all $n=k$. For $n=k+1$, then $|A setminus a|=k$ for some $a in A$. Thus $X setminus (A setminus a) sim X$ by inductive hypothesis, or $(X cap a) cup (X setminus A) sim X$, or $a cup (X setminus A) sim X$. We have $a cup (X setminus A) sim X setminus A$ since the theorem is true for $n=1$. Hence $X setminus A sim a cup (X setminus A) sim X$. Thus $X setminus A sim X$. This completes the proof.





Does this proof look fine or contain gaps? Do you have suggestions? Many thanks for your dedicated help!





Update: Here I prove that the theorem is true for $n=1$.



Assume that $A = a$ and consequently $X setminus A= X setminusa$. It's clear that $|X setminus A| le |X|$. Next we prove that $|X| le |X setminus A|$. Since $X$ is infinite, there exists $B subsetneq X$ such that $B sim X$ (Here we assume Axiom of Countable Choice). Thus $|X|=|B|$. There are only two possible cases.



  1. $a in X setminus B$

Then $B subseteq X setminus a=X setminus A$ and consequently $|X|=|B| le |X setminus A|$. Thus $|X| le |X setminus A|$ and $|X setminus A| le |X|$. By Schröder–Bernstein theorem, we have $|X setminus A| = |X|$. It follows that $X setminus A sim X$.



  1. $a in B$.

Let $b in X setminus B$. We define a bijection $f:X setminus a to X setminus b$ by $f(x)= x$ for all $x in X setminus a,b$ and $f(b)=a$. Thus $X setminus a sim X setminus b$. Since $b in X setminus B$, it follows from Case 1 that $X setminus b sim X$. Hence $X setminus a sim X setminus b sim X$. Thus $X setminus a = X setminus A sim X$.



To sum up, $X setminus A sim X$ for all $|A|=1$.










share|cite|improve this question
















Suppose that $X$ is infinite and that $A$ is a finite subset of $X$. Then $X$ and $X setminus A$ are equinumerous.





My attempt:



Let $|A|=n$. We will prove by induction on n. It's clear that the the theorem is trivially true for $n=0$. Assume the theorem is true for all $n=k$. For $n=k+1$, then $|A setminus a|=k$ for some $a in A$. Thus $X setminus (A setminus a) sim X$ by inductive hypothesis, or $(X cap a) cup (X setminus A) sim X$, or $a cup (X setminus A) sim X$. We have $a cup (X setminus A) sim X setminus A$ since the theorem is true for $n=1$. Hence $X setminus A sim a cup (X setminus A) sim X$. Thus $X setminus A sim X$. This completes the proof.





Does this proof look fine or contain gaps? Do you have suggestions? Many thanks for your dedicated help!





Update: Here I prove that the theorem is true for $n=1$.



Assume that $A = a$ and consequently $X setminus A= X setminusa$. It's clear that $|X setminus A| le |X|$. Next we prove that $|X| le |X setminus A|$. Since $X$ is infinite, there exists $B subsetneq X$ such that $B sim X$ (Here we assume Axiom of Countable Choice). Thus $|X|=|B|$. There are only two possible cases.



  1. $a in X setminus B$

Then $B subseteq X setminus a=X setminus A$ and consequently $|X|=|B| le |X setminus A|$. Thus $|X| le |X setminus A|$ and $|X setminus A| le |X|$. By Schröder–Bernstein theorem, we have $|X setminus A| = |X|$. It follows that $X setminus A sim X$.



  1. $a in B$.

Let $b in X setminus B$. We define a bijection $f:X setminus a to X setminus b$ by $f(x)= x$ for all $x in X setminus a,b$ and $f(b)=a$. Thus $X setminus a sim X setminus b$. Since $b in X setminus B$, it follows from Case 1 that $X setminus b sim X$. Hence $X setminus a sim X setminus b sim X$. Thus $X setminus a = X setminus A sim X$.



To sum up, $X setminus A sim X$ for all $|A|=1$.







elementary-set-theory proof-verification infinity






share|cite|improve this question















share|cite|improve this question













share|cite|improve this question




share|cite|improve this question








edited Sep 12 at 7:21

























asked Sep 10 at 15:48









Le Anh Dung

1,2041421




1,2041421











  • What's your definition of $X$ being infinite?
    – Paul K
    Sep 10 at 15:51










  • @PaulK $X$ is infinite if and only if $X$ is not finite.
    – Le Anh Dung
    Sep 10 at 15:55










  • @PaulK of infinite cardinality
    – Michał Zapała
    Sep 10 at 15:57










  • The proof is somewhat lacking in elegance, but formally correct
    – Michał Zapała
    Sep 10 at 15:58










  • @MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
    – Le Anh Dung
    Sep 10 at 16:08
















  • What's your definition of $X$ being infinite?
    – Paul K
    Sep 10 at 15:51










  • @PaulK $X$ is infinite if and only if $X$ is not finite.
    – Le Anh Dung
    Sep 10 at 15:55










  • @PaulK of infinite cardinality
    – Michał Zapała
    Sep 10 at 15:57










  • The proof is somewhat lacking in elegance, but formally correct
    – Michał Zapała
    Sep 10 at 15:58










  • @MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
    – Le Anh Dung
    Sep 10 at 16:08















What's your definition of $X$ being infinite?
– Paul K
Sep 10 at 15:51




What's your definition of $X$ being infinite?
– Paul K
Sep 10 at 15:51












@PaulK $X$ is infinite if and only if $X$ is not finite.
– Le Anh Dung
Sep 10 at 15:55




@PaulK $X$ is infinite if and only if $X$ is not finite.
– Le Anh Dung
Sep 10 at 15:55












@PaulK of infinite cardinality
– Michał Zapała
Sep 10 at 15:57




@PaulK of infinite cardinality
– Michał Zapała
Sep 10 at 15:57












The proof is somewhat lacking in elegance, but formally correct
– Michał Zapała
Sep 10 at 15:58




The proof is somewhat lacking in elegance, but formally correct
– Michał Zapała
Sep 10 at 15:58












@MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
– Le Anh Dung
Sep 10 at 16:08




@MichałZapała I'm very curious about your elegant approach. Could you please elaborate on it?
– Le Anh Dung
Sep 10 at 16:08










3 Answers
3






active

oldest

votes

















up vote
0
down vote



accepted










The proof (with the update) seems correct.



Assuming choice (or at least countable choice), we can do it perhaps more easily.



Since $A$ is finite, there is a bijection $gcolon0,1,dots,n-1to A$, for some $ninmathbbN$.



Fix an injection $fcolonmathbbNto Xsetminus A$ (which exists because $Xsetminus A$ is infinite, assuming countable choice) and define $psicolon Xsetminus Ato X$ by
$$
psi(x)=begincases
x & xnotin f(mathbbN) \[4px]
g(m) & x=f(m),quad 0 le m < n \[4px]
f(m-n) & x=f(m),quad m ge n
endcases
$$
Prove $psi$ is a bijection.






share|cite|improve this answer






















  • Thank you so much!
    – Le Anh Dung
    Sep 12 at 8:55










  • After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
    – Le Anh Dung
    Sep 12 at 11:13










  • Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
    – Dan Velleman
    Sep 13 at 21:32










  • @DanVelleman Thanks for noting!
    – egreg
    Sep 13 at 21:34

















up vote
3
down vote













Your proof is correct except for the step where you say that $a cup (X setminus A) sim X setminus A$ by inductive hypothesis. I assume you are applying the inductive hypothesis (to the set $a cup (X setminus A)$) in the case $n=1$, which is fine as long as $k ge 1$. But your proof does not work in the case $k=0$. In other words, your proof correctly shows that if the theorem holds for $n=1$, then it holds for all larger values of $n$. But it does not prove that it holds for $n=1$.



In fact, the proof for $n=1$ is rather tricky. Here's a nice exercise: prove that the $n=1$ case for an infinite set $X$ is equivalent to the statement that $X$ contains a subset that is equinumerous to the set of positive integers. Now, the statement that every infinite set contains a subset equinumerous with the positive integers cannot be proven without some form of the axiom of choice. Therefore the proof of the $n=1$ case will also require the axiom of choice.






share|cite|improve this answer



























    up vote
    0
    down vote













    I found that @egreg's solution is very elegant, so I want to re-formalize it into the below proof. All credits go to @egreg.




    Lemma 1: If $A$ is finite and $B$ is countably infinite, then $Acup B$ is countably infinite.



    Lemma 2: If $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite.



    Lemma 3: If $Y$ is infinite, then there exists $Bsubsetneq Y$ such that $B$ is countably infinite. (Here we assume Axiom of Countable Choice)




    Since $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite by Lemma 2.



    Since $Xsetminus A$ is infinite, there exists $Bsubsetneq Xsetminus A$ such that $B sim Bbb N$ by Lemma 3.



    Since $A$ is finite and $B$ is countably infinite, then $Acup B sim Bbb N$ by Lemma 1.



    Since $B sim Bbb N$ and $Acup B sim Bbb N$, $B sim Acup B$ and thus there exists an bijection $f_1:B to Acup B$.



    Let $f_2:Xsetminus Asetminus B to Xsetminus Asetminus B$ be the identity map on $Xsetminus Asetminus B$. Then $f_2$ is a bijection.



    We define $f:Xsetminus A to X$ by $f(x)=f_2(x)$ for all $x in Xsetminus Asetminus B$ and $f(x)=f_1(x)$ for all $x in B$. Thus $f$ is a bijection.



    Hence $Xsetminus A sim X$.






    share|cite|improve this answer




















      Your Answer





      StackExchange.ifUsing("editor", function ()
      return StackExchange.using("mathjaxEditing", function ()
      StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
      StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
      );
      );
      , "mathjax-editing");

      StackExchange.ready(function()
      var channelOptions =
      tags: "".split(" "),
      id: "69"
      ;
      initTagRenderer("".split(" "), "".split(" "), channelOptions);

      StackExchange.using("externalEditor", function()
      // Have to fire editor after snippets, if snippets enabled
      if (StackExchange.settings.snippets.snippetsEnabled)
      StackExchange.using("snippets", function()
      createEditor();
      );

      else
      createEditor();

      );

      function createEditor()
      StackExchange.prepareEditor(
      heartbeatType: 'answer',
      convertImagesToLinks: true,
      noModals: true,
      showLowRepImageUploadWarning: true,
      reputationToPostImages: 10,
      bindNavPrevention: true,
      postfix: "",
      imageUploader:
      brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
      contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
      allowUrls: true
      ,
      noCode: true, onDemand: true,
      discardSelector: ".discard-answer"
      ,immediatelyShowMarkdownHelp:true
      );



      );













       

      draft saved


      draft discarded


















      StackExchange.ready(
      function ()
      StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2912030%2fsuppose-x-is-infinite-and-a-is-a-finite-subset-of-x-then-x-and-x-setm%23new-answer', 'question_page');

      );

      Post as a guest






























      3 Answers
      3






      active

      oldest

      votes








      3 Answers
      3






      active

      oldest

      votes









      active

      oldest

      votes






      active

      oldest

      votes








      up vote
      0
      down vote



      accepted










      The proof (with the update) seems correct.



      Assuming choice (or at least countable choice), we can do it perhaps more easily.



      Since $A$ is finite, there is a bijection $gcolon0,1,dots,n-1to A$, for some $ninmathbbN$.



      Fix an injection $fcolonmathbbNto Xsetminus A$ (which exists because $Xsetminus A$ is infinite, assuming countable choice) and define $psicolon Xsetminus Ato X$ by
      $$
      psi(x)=begincases
      x & xnotin f(mathbbN) \[4px]
      g(m) & x=f(m),quad 0 le m < n \[4px]
      f(m-n) & x=f(m),quad m ge n
      endcases
      $$
      Prove $psi$ is a bijection.






      share|cite|improve this answer






















      • Thank you so much!
        – Le Anh Dung
        Sep 12 at 8:55










      • After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
        – Le Anh Dung
        Sep 12 at 11:13










      • Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
        – Dan Velleman
        Sep 13 at 21:32










      • @DanVelleman Thanks for noting!
        – egreg
        Sep 13 at 21:34














      up vote
      0
      down vote



      accepted










      The proof (with the update) seems correct.



      Assuming choice (or at least countable choice), we can do it perhaps more easily.



      Since $A$ is finite, there is a bijection $gcolon0,1,dots,n-1to A$, for some $ninmathbbN$.



      Fix an injection $fcolonmathbbNto Xsetminus A$ (which exists because $Xsetminus A$ is infinite, assuming countable choice) and define $psicolon Xsetminus Ato X$ by
      $$
      psi(x)=begincases
      x & xnotin f(mathbbN) \[4px]
      g(m) & x=f(m),quad 0 le m < n \[4px]
      f(m-n) & x=f(m),quad m ge n
      endcases
      $$
      Prove $psi$ is a bijection.






      share|cite|improve this answer






















      • Thank you so much!
        – Le Anh Dung
        Sep 12 at 8:55










      • After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
        – Le Anh Dung
        Sep 12 at 11:13










      • Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
        – Dan Velleman
        Sep 13 at 21:32










      • @DanVelleman Thanks for noting!
        – egreg
        Sep 13 at 21:34












      up vote
      0
      down vote



      accepted







      up vote
      0
      down vote



      accepted






      The proof (with the update) seems correct.



      Assuming choice (or at least countable choice), we can do it perhaps more easily.



      Since $A$ is finite, there is a bijection $gcolon0,1,dots,n-1to A$, for some $ninmathbbN$.



      Fix an injection $fcolonmathbbNto Xsetminus A$ (which exists because $Xsetminus A$ is infinite, assuming countable choice) and define $psicolon Xsetminus Ato X$ by
      $$
      psi(x)=begincases
      x & xnotin f(mathbbN) \[4px]
      g(m) & x=f(m),quad 0 le m < n \[4px]
      f(m-n) & x=f(m),quad m ge n
      endcases
      $$
      Prove $psi$ is a bijection.






      share|cite|improve this answer














      The proof (with the update) seems correct.



      Assuming choice (or at least countable choice), we can do it perhaps more easily.



      Since $A$ is finite, there is a bijection $gcolon0,1,dots,n-1to A$, for some $ninmathbbN$.



      Fix an injection $fcolonmathbbNto Xsetminus A$ (which exists because $Xsetminus A$ is infinite, assuming countable choice) and define $psicolon Xsetminus Ato X$ by
      $$
      psi(x)=begincases
      x & xnotin f(mathbbN) \[4px]
      g(m) & x=f(m),quad 0 le m < n \[4px]
      f(m-n) & x=f(m),quad m ge n
      endcases
      $$
      Prove $psi$ is a bijection.







      share|cite|improve this answer














      share|cite|improve this answer



      share|cite|improve this answer








      edited Sep 13 at 21:34

























      answered Sep 12 at 8:53









      egreg

      172k1283194




      172k1283194











      • Thank you so much!
        – Le Anh Dung
        Sep 12 at 8:55










      • After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
        – Le Anh Dung
        Sep 12 at 11:13










      • Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
        – Dan Velleman
        Sep 13 at 21:32










      • @DanVelleman Thanks for noting!
        – egreg
        Sep 13 at 21:34
















      • Thank you so much!
        – Le Anh Dung
        Sep 12 at 8:55










      • After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
        – Le Anh Dung
        Sep 12 at 11:13










      • Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
        – Dan Velleman
        Sep 13 at 21:32










      • @DanVelleman Thanks for noting!
        – egreg
        Sep 13 at 21:34















      Thank you so much!
      – Le Anh Dung
      Sep 12 at 8:55




      Thank you so much!
      – Le Anh Dung
      Sep 12 at 8:55












      After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
      – Le Anh Dung
      Sep 12 at 11:13




      After I read your answer carefully, I found that your approach is much more elegant and simpler than mine. I decided to re-formalize it into a proof in my own words and posted as an answer at the end of this thread. If you don't mind, please have a look at it! Many thanks for you!
      – Le Anh Dung
      Sep 12 at 11:13












      Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
      – Dan Velleman
      Sep 13 at 21:32




      Typo: I think you mean $psi(x) = f(m-n)$ when $x = f(m)$, $m ge n$.
      – Dan Velleman
      Sep 13 at 21:32












      @DanVelleman Thanks for noting!
      – egreg
      Sep 13 at 21:34




      @DanVelleman Thanks for noting!
      – egreg
      Sep 13 at 21:34










      up vote
      3
      down vote













      Your proof is correct except for the step where you say that $a cup (X setminus A) sim X setminus A$ by inductive hypothesis. I assume you are applying the inductive hypothesis (to the set $a cup (X setminus A)$) in the case $n=1$, which is fine as long as $k ge 1$. But your proof does not work in the case $k=0$. In other words, your proof correctly shows that if the theorem holds for $n=1$, then it holds for all larger values of $n$. But it does not prove that it holds for $n=1$.



      In fact, the proof for $n=1$ is rather tricky. Here's a nice exercise: prove that the $n=1$ case for an infinite set $X$ is equivalent to the statement that $X$ contains a subset that is equinumerous to the set of positive integers. Now, the statement that every infinite set contains a subset equinumerous with the positive integers cannot be proven without some form of the axiom of choice. Therefore the proof of the $n=1$ case will also require the axiom of choice.






      share|cite|improve this answer
























        up vote
        3
        down vote













        Your proof is correct except for the step where you say that $a cup (X setminus A) sim X setminus A$ by inductive hypothesis. I assume you are applying the inductive hypothesis (to the set $a cup (X setminus A)$) in the case $n=1$, which is fine as long as $k ge 1$. But your proof does not work in the case $k=0$. In other words, your proof correctly shows that if the theorem holds for $n=1$, then it holds for all larger values of $n$. But it does not prove that it holds for $n=1$.



        In fact, the proof for $n=1$ is rather tricky. Here's a nice exercise: prove that the $n=1$ case for an infinite set $X$ is equivalent to the statement that $X$ contains a subset that is equinumerous to the set of positive integers. Now, the statement that every infinite set contains a subset equinumerous with the positive integers cannot be proven without some form of the axiom of choice. Therefore the proof of the $n=1$ case will also require the axiom of choice.






        share|cite|improve this answer






















          up vote
          3
          down vote










          up vote
          3
          down vote









          Your proof is correct except for the step where you say that $a cup (X setminus A) sim X setminus A$ by inductive hypothesis. I assume you are applying the inductive hypothesis (to the set $a cup (X setminus A)$) in the case $n=1$, which is fine as long as $k ge 1$. But your proof does not work in the case $k=0$. In other words, your proof correctly shows that if the theorem holds for $n=1$, then it holds for all larger values of $n$. But it does not prove that it holds for $n=1$.



          In fact, the proof for $n=1$ is rather tricky. Here's a nice exercise: prove that the $n=1$ case for an infinite set $X$ is equivalent to the statement that $X$ contains a subset that is equinumerous to the set of positive integers. Now, the statement that every infinite set contains a subset equinumerous with the positive integers cannot be proven without some form of the axiom of choice. Therefore the proof of the $n=1$ case will also require the axiom of choice.






          share|cite|improve this answer












          Your proof is correct except for the step where you say that $a cup (X setminus A) sim X setminus A$ by inductive hypothesis. I assume you are applying the inductive hypothesis (to the set $a cup (X setminus A)$) in the case $n=1$, which is fine as long as $k ge 1$. But your proof does not work in the case $k=0$. In other words, your proof correctly shows that if the theorem holds for $n=1$, then it holds for all larger values of $n$. But it does not prove that it holds for $n=1$.



          In fact, the proof for $n=1$ is rather tricky. Here's a nice exercise: prove that the $n=1$ case for an infinite set $X$ is equivalent to the statement that $X$ contains a subset that is equinumerous to the set of positive integers. Now, the statement that every infinite set contains a subset equinumerous with the positive integers cannot be proven without some form of the axiom of choice. Therefore the proof of the $n=1$ case will also require the axiom of choice.







          share|cite|improve this answer












          share|cite|improve this answer



          share|cite|improve this answer










          answered Sep 10 at 16:40









          Dan Velleman

          1663




          1663




















              up vote
              0
              down vote













              I found that @egreg's solution is very elegant, so I want to re-formalize it into the below proof. All credits go to @egreg.




              Lemma 1: If $A$ is finite and $B$ is countably infinite, then $Acup B$ is countably infinite.



              Lemma 2: If $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite.



              Lemma 3: If $Y$ is infinite, then there exists $Bsubsetneq Y$ such that $B$ is countably infinite. (Here we assume Axiom of Countable Choice)




              Since $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite by Lemma 2.



              Since $Xsetminus A$ is infinite, there exists $Bsubsetneq Xsetminus A$ such that $B sim Bbb N$ by Lemma 3.



              Since $A$ is finite and $B$ is countably infinite, then $Acup B sim Bbb N$ by Lemma 1.



              Since $B sim Bbb N$ and $Acup B sim Bbb N$, $B sim Acup B$ and thus there exists an bijection $f_1:B to Acup B$.



              Let $f_2:Xsetminus Asetminus B to Xsetminus Asetminus B$ be the identity map on $Xsetminus Asetminus B$. Then $f_2$ is a bijection.



              We define $f:Xsetminus A to X$ by $f(x)=f_2(x)$ for all $x in Xsetminus Asetminus B$ and $f(x)=f_1(x)$ for all $x in B$. Thus $f$ is a bijection.



              Hence $Xsetminus A sim X$.






              share|cite|improve this answer
























                up vote
                0
                down vote













                I found that @egreg's solution is very elegant, so I want to re-formalize it into the below proof. All credits go to @egreg.




                Lemma 1: If $A$ is finite and $B$ is countably infinite, then $Acup B$ is countably infinite.



                Lemma 2: If $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite.



                Lemma 3: If $Y$ is infinite, then there exists $Bsubsetneq Y$ such that $B$ is countably infinite. (Here we assume Axiom of Countable Choice)




                Since $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite by Lemma 2.



                Since $Xsetminus A$ is infinite, there exists $Bsubsetneq Xsetminus A$ such that $B sim Bbb N$ by Lemma 3.



                Since $A$ is finite and $B$ is countably infinite, then $Acup B sim Bbb N$ by Lemma 1.



                Since $B sim Bbb N$ and $Acup B sim Bbb N$, $B sim Acup B$ and thus there exists an bijection $f_1:B to Acup B$.



                Let $f_2:Xsetminus Asetminus B to Xsetminus Asetminus B$ be the identity map on $Xsetminus Asetminus B$. Then $f_2$ is a bijection.



                We define $f:Xsetminus A to X$ by $f(x)=f_2(x)$ for all $x in Xsetminus Asetminus B$ and $f(x)=f_1(x)$ for all $x in B$. Thus $f$ is a bijection.



                Hence $Xsetminus A sim X$.






                share|cite|improve this answer






















                  up vote
                  0
                  down vote










                  up vote
                  0
                  down vote









                  I found that @egreg's solution is very elegant, so I want to re-formalize it into the below proof. All credits go to @egreg.




                  Lemma 1: If $A$ is finite and $B$ is countably infinite, then $Acup B$ is countably infinite.



                  Lemma 2: If $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite.



                  Lemma 3: If $Y$ is infinite, then there exists $Bsubsetneq Y$ such that $B$ is countably infinite. (Here we assume Axiom of Countable Choice)




                  Since $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite by Lemma 2.



                  Since $Xsetminus A$ is infinite, there exists $Bsubsetneq Xsetminus A$ such that $B sim Bbb N$ by Lemma 3.



                  Since $A$ is finite and $B$ is countably infinite, then $Acup B sim Bbb N$ by Lemma 1.



                  Since $B sim Bbb N$ and $Acup B sim Bbb N$, $B sim Acup B$ and thus there exists an bijection $f_1:B to Acup B$.



                  Let $f_2:Xsetminus Asetminus B to Xsetminus Asetminus B$ be the identity map on $Xsetminus Asetminus B$. Then $f_2$ is a bijection.



                  We define $f:Xsetminus A to X$ by $f(x)=f_2(x)$ for all $x in Xsetminus Asetminus B$ and $f(x)=f_1(x)$ for all $x in B$. Thus $f$ is a bijection.



                  Hence $Xsetminus A sim X$.






                  share|cite|improve this answer












                  I found that @egreg's solution is very elegant, so I want to re-formalize it into the below proof. All credits go to @egreg.




                  Lemma 1: If $A$ is finite and $B$ is countably infinite, then $Acup B$ is countably infinite.



                  Lemma 2: If $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite.



                  Lemma 3: If $Y$ is infinite, then there exists $Bsubsetneq Y$ such that $B$ is countably infinite. (Here we assume Axiom of Countable Choice)




                  Since $X$ is infinite and $A$ is finite, then $Xsetminus A$ is infinite by Lemma 2.



                  Since $Xsetminus A$ is infinite, there exists $Bsubsetneq Xsetminus A$ such that $B sim Bbb N$ by Lemma 3.



                  Since $A$ is finite and $B$ is countably infinite, then $Acup B sim Bbb N$ by Lemma 1.



                  Since $B sim Bbb N$ and $Acup B sim Bbb N$, $B sim Acup B$ and thus there exists an bijection $f_1:B to Acup B$.



                  Let $f_2:Xsetminus Asetminus B to Xsetminus Asetminus B$ be the identity map on $Xsetminus Asetminus B$. Then $f_2$ is a bijection.



                  We define $f:Xsetminus A to X$ by $f(x)=f_2(x)$ for all $x in Xsetminus Asetminus B$ and $f(x)=f_1(x)$ for all $x in B$. Thus $f$ is a bijection.



                  Hence $Xsetminus A sim X$.







                  share|cite|improve this answer












                  share|cite|improve this answer



                  share|cite|improve this answer










                  answered Sep 12 at 11:11









                  Le Anh Dung

                  1,2041421




                  1,2041421



























                       

                      draft saved


                      draft discarded















































                       


                      draft saved


                      draft discarded














                      StackExchange.ready(
                      function ()
                      StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2912030%2fsuppose-x-is-infinite-and-a-is-a-finite-subset-of-x-then-x-and-x-setm%23new-answer', 'question_page');

                      );

                      Post as a guest













































































                      這個網誌中的熱門文章

                      Is there any way to eliminate the singular point to solve this integral by hand or by approximations?

                      Why am i infinitely getting the same tweet with the Twitter Search API?

                      Solve: $(3xy-2ay^2)dx+(x^2-2axy)dy=0$