Zariski topology and product topology on $mathbb A_k^m times mathbb A_k^n$

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











up vote
6
down vote

favorite












Let $k$ be an algebraically closed field and $mathbb A_k^n$ be the affine $n$-space. We can give the Zariski topology on it. Then consider $X=mathbb A_k^m times mathbb A_k^n$ under product topology, where each of $mathbb A_k^m$ and $mathbb A_k^n$ are given the Zariski topology. And also consider $X=mathbb A_k^m+n$ under the Zariski topology. My question is : For which $m,nge 1$, are these two topological spaces homeomorphic ?



Note that I am not asking when two topologies are same ... I'm asking when they are not homeomorphic ...










share|cite|improve this question













migrated from mathoverflow.net Sep 1 at 3:19


This question came from our site for professional mathematicians.














  • Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
    – Donu Arapura
    Aug 28 at 1:06










  • @DonuArapura: can you please explain why is that ?
    – user521337
    Aug 28 at 1:31










  • HInt: What is the intersection of open sets?
    – Jan-Christoph Schlage-Puchta
    Aug 28 at 21:01














up vote
6
down vote

favorite












Let $k$ be an algebraically closed field and $mathbb A_k^n$ be the affine $n$-space. We can give the Zariski topology on it. Then consider $X=mathbb A_k^m times mathbb A_k^n$ under product topology, where each of $mathbb A_k^m$ and $mathbb A_k^n$ are given the Zariski topology. And also consider $X=mathbb A_k^m+n$ under the Zariski topology. My question is : For which $m,nge 1$, are these two topological spaces homeomorphic ?



Note that I am not asking when two topologies are same ... I'm asking when they are not homeomorphic ...










share|cite|improve this question













migrated from mathoverflow.net Sep 1 at 3:19


This question came from our site for professional mathematicians.














  • Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
    – Donu Arapura
    Aug 28 at 1:06










  • @DonuArapura: can you please explain why is that ?
    – user521337
    Aug 28 at 1:31










  • HInt: What is the intersection of open sets?
    – Jan-Christoph Schlage-Puchta
    Aug 28 at 21:01












up vote
6
down vote

favorite









up vote
6
down vote

favorite











Let $k$ be an algebraically closed field and $mathbb A_k^n$ be the affine $n$-space. We can give the Zariski topology on it. Then consider $X=mathbb A_k^m times mathbb A_k^n$ under product topology, where each of $mathbb A_k^m$ and $mathbb A_k^n$ are given the Zariski topology. And also consider $X=mathbb A_k^m+n$ under the Zariski topology. My question is : For which $m,nge 1$, are these two topological spaces homeomorphic ?



Note that I am not asking when two topologies are same ... I'm asking when they are not homeomorphic ...










share|cite|improve this question













Let $k$ be an algebraically closed field and $mathbb A_k^n$ be the affine $n$-space. We can give the Zariski topology on it. Then consider $X=mathbb A_k^m times mathbb A_k^n$ under product topology, where each of $mathbb A_k^m$ and $mathbb A_k^n$ are given the Zariski topology. And also consider $X=mathbb A_k^m+n$ under the Zariski topology. My question is : For which $m,nge 1$, are these two topological spaces homeomorphic ?



Note that I am not asking when two topologies are same ... I'm asking when they are not homeomorphic ...







commutative-algebra general-topology affine-geometry zariski-topology






share|cite|improve this question













share|cite|improve this question











share|cite|improve this question




share|cite|improve this question










asked Aug 28 at 0:51









user521337

30811




30811




migrated from mathoverflow.net Sep 1 at 3:19


This question came from our site for professional mathematicians.






migrated from mathoverflow.net Sep 1 at 3:19


This question came from our site for professional mathematicians.













  • Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
    – Donu Arapura
    Aug 28 at 1:06










  • @DonuArapura: can you please explain why is that ?
    – user521337
    Aug 28 at 1:31










  • HInt: What is the intersection of open sets?
    – Jan-Christoph Schlage-Puchta
    Aug 28 at 21:01
















  • Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
    – Donu Arapura
    Aug 28 at 1:06










  • @DonuArapura: can you please explain why is that ?
    – user521337
    Aug 28 at 1:31










  • HInt: What is the intersection of open sets?
    – Jan-Christoph Schlage-Puchta
    Aug 28 at 21:01















Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
– Donu Arapura
Aug 28 at 1:06




Essentially never. More precisely, $X$ won't be homeomorphic to $mathbbA^mtimes mathbbA^n$ unless $m$ or $n$ is zero.
– Donu Arapura
Aug 28 at 1:06












@DonuArapura: can you please explain why is that ?
– user521337
Aug 28 at 1:31




@DonuArapura: can you please explain why is that ?
– user521337
Aug 28 at 1:31












HInt: What is the intersection of open sets?
– Jan-Christoph Schlage-Puchta
Aug 28 at 21:01




HInt: What is the intersection of open sets?
– Jan-Christoph Schlage-Puchta
Aug 28 at 21:01










1 Answer
1






active

oldest

votes

















up vote
12
down vote



accepted










Clearly the natural map $|mathbb A^n+m| to |mathbb A^n| times |mathbb A^m|$ is not a homeomorphism unless $n = 0$ or $m = 0$ (it's not even a bijection, because of points like $(x-y) in operatornameSpec k[x] otimes k[y]$). We show that in fact there is no homeomorphism, using the following ad hoc notion of dimension:



Definition. Let $X$ be a topological space. Then the intersection dimension $operatornameint. dim(X)$ of $X$ is the supremum of all $n$ for which there exist irreducible closed sets $V_0,ldots,V_n subseteq X$ such that



  1. any intersection $V_I = bigcap_iin I V_i$ for $I subseteq 0,ldots,n$ nonempty is either irreducible or empty;

  2. the intersection $V_I$ for any $I subseteq 0,ldots,n$ with $|I| = n$ is exactly one point;

  3. the intersection $V_0,ldots,n$ is empty.

I feel that I have seen some variant of this somewhere before (probably under a different name), but I do not recall where. References are welcome!



Remark. If there exist sets $V_0,ldots,V_n$ with properties 1, 2, and 3 above, then the same is true for any $m leq n$. Indeed, we may take the sets $V_0, ldots, V_m-1, V_m,ldots,n$.



Lemma 1. Let $k$ be a field. Then $operatornameint. dim(mathbb A^n_k) = n$.



Proof. The hyperplanes $V(x_1+ldots+x_n), V(x_1), ldots, V(x_n)$ show that $operatornameint. dim(mathbb A^n_k) geq n$. Conversely, let $V_0, ldots, V_m$ be irreducible closed subsets satisfying 1, 2, and 3 above. Conditions 2 and 3 imply that no $V_i$ is contained in any intersection of the others, so by induction on $|I|$ the dimension of $V_I$ is at most $n-|I|$ (because all intersections are irreducible). Taking $|I| = m$ gives $m leq n$, showing that $operatornameint. dim(mathbb A^n_k) leq n$. $square$



More generally, this argument shows that if $X$ is a topological space of dimension $n$, then $operatornameint. dim(X) leq dim X$. It's not clear even for smooth proper varieties that the two are equal (in the projective case we might be able to use some Bertini argument, using tangency to reduce the intersections to one point).



Lemma 2. Let $X$ and $Y$ be topological spaces with a closed point. Then
$$operatornameint. dim(X times Y) = max(operatornameint. dim(X), operatornameint. dim(Y)).$$



Proof. By the Lemma of this post, a subset $Z subseteq X times Y$ is closed and irreducible if and only if $Z = V times W$, with $V subseteq X$ and $W subseteq Y$ closed and irreducible.



Let $m$ be the maximum of the intersection dimensions of $X$ and $Y$; without loss of generality let's say $m = operatornameint. dim(X)$. Then there exist sets $V_0,ldots,V_m subseteq X$ satisfying properties 1, 2, and 3 of the definition above. Choose a closed point $y in Y$, and set $W_i = y$ for all $i$. Then the sets $Z_i = V_i times W_i$ are closed and irreducible. Moreover, the finite intersections $Z_I = bigcap_i in I Z_i$ equal $V_I times W_I$. Since properties 1, 2, and 3 hold for the $V_i$ and properties 1 and 2 for the $W_i$, this shows properties 1, 2, and 3 for the $Z_i$. Therefore, $operatornameint. dim(X times Y) geq m$.



Conversely, if $Z_0, ldots, Z_n$ satisfy properties 1, 2, and 3, then we must have $Z_i = V_i times W_i$ for some irreducible closed subsets $V_i subseteq X$, $W_i subseteq Y$. Again we have $Z_I = V_I times W_I$, so both $V_i$ and $W_i$ satisfy 1 and 2. Since $Z_0,ldots,n = varnothing$, we conclude that either $V_0,ldots,n = varnothing$ or $W_0,ldots,n = varnothing$. Hence, either the $V_i$ or the $W_i$ satisfy the conditions above, so $operatornameint. dim(X times Y) leq m$. $square$



Corollary. There exists a homeomorphism $|mathbb A^n+m_k| stackrelsimto |mathbb A^n_k| times |mathbb A^m_k|$ if and only if $n = 0$ or $m = 0$.



Proof. By Lemma 1 and Lemma 2, we have
$$operatornameint. dim(|mathbb A^n_k| times |mathbb A^m_k|) = max(n,m),$$
whereas
$$operatornameint. dim(|mathbb A^n+m_k|) = n + m.$$
These numbers are equal if and only if $n = 0$ or $m = 0$. Clearly the stated homeomorphisms exist in this case. $square$



We can actually mix and match fields, i.e. consider things like $|mathbb A^100_mathbb C| times |mathbb A^17_barmathbb F_3|$. The same proof shows that this cannot be homeomorphic to any space of the form $|mathbb A^117_k|$ for a field $k$.






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: false,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: 10,
    bindNavPrevention: true,
    postfix: "",
    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%2f2901329%2fzariski-topology-and-product-topology-on-mathbb-a-km-times-mathbb-a-kn%23new-answer', 'question_page');

    );

    Post as a guest






























    1 Answer
    1






    active

    oldest

    votes








    1 Answer
    1






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes








    up vote
    12
    down vote



    accepted










    Clearly the natural map $|mathbb A^n+m| to |mathbb A^n| times |mathbb A^m|$ is not a homeomorphism unless $n = 0$ or $m = 0$ (it's not even a bijection, because of points like $(x-y) in operatornameSpec k[x] otimes k[y]$). We show that in fact there is no homeomorphism, using the following ad hoc notion of dimension:



    Definition. Let $X$ be a topological space. Then the intersection dimension $operatornameint. dim(X)$ of $X$ is the supremum of all $n$ for which there exist irreducible closed sets $V_0,ldots,V_n subseteq X$ such that



    1. any intersection $V_I = bigcap_iin I V_i$ for $I subseteq 0,ldots,n$ nonempty is either irreducible or empty;

    2. the intersection $V_I$ for any $I subseteq 0,ldots,n$ with $|I| = n$ is exactly one point;

    3. the intersection $V_0,ldots,n$ is empty.

    I feel that I have seen some variant of this somewhere before (probably under a different name), but I do not recall where. References are welcome!



    Remark. If there exist sets $V_0,ldots,V_n$ with properties 1, 2, and 3 above, then the same is true for any $m leq n$. Indeed, we may take the sets $V_0, ldots, V_m-1, V_m,ldots,n$.



    Lemma 1. Let $k$ be a field. Then $operatornameint. dim(mathbb A^n_k) = n$.



    Proof. The hyperplanes $V(x_1+ldots+x_n), V(x_1), ldots, V(x_n)$ show that $operatornameint. dim(mathbb A^n_k) geq n$. Conversely, let $V_0, ldots, V_m$ be irreducible closed subsets satisfying 1, 2, and 3 above. Conditions 2 and 3 imply that no $V_i$ is contained in any intersection of the others, so by induction on $|I|$ the dimension of $V_I$ is at most $n-|I|$ (because all intersections are irreducible). Taking $|I| = m$ gives $m leq n$, showing that $operatornameint. dim(mathbb A^n_k) leq n$. $square$



    More generally, this argument shows that if $X$ is a topological space of dimension $n$, then $operatornameint. dim(X) leq dim X$. It's not clear even for smooth proper varieties that the two are equal (in the projective case we might be able to use some Bertini argument, using tangency to reduce the intersections to one point).



    Lemma 2. Let $X$ and $Y$ be topological spaces with a closed point. Then
    $$operatornameint. dim(X times Y) = max(operatornameint. dim(X), operatornameint. dim(Y)).$$



    Proof. By the Lemma of this post, a subset $Z subseteq X times Y$ is closed and irreducible if and only if $Z = V times W$, with $V subseteq X$ and $W subseteq Y$ closed and irreducible.



    Let $m$ be the maximum of the intersection dimensions of $X$ and $Y$; without loss of generality let's say $m = operatornameint. dim(X)$. Then there exist sets $V_0,ldots,V_m subseteq X$ satisfying properties 1, 2, and 3 of the definition above. Choose a closed point $y in Y$, and set $W_i = y$ for all $i$. Then the sets $Z_i = V_i times W_i$ are closed and irreducible. Moreover, the finite intersections $Z_I = bigcap_i in I Z_i$ equal $V_I times W_I$. Since properties 1, 2, and 3 hold for the $V_i$ and properties 1 and 2 for the $W_i$, this shows properties 1, 2, and 3 for the $Z_i$. Therefore, $operatornameint. dim(X times Y) geq m$.



    Conversely, if $Z_0, ldots, Z_n$ satisfy properties 1, 2, and 3, then we must have $Z_i = V_i times W_i$ for some irreducible closed subsets $V_i subseteq X$, $W_i subseteq Y$. Again we have $Z_I = V_I times W_I$, so both $V_i$ and $W_i$ satisfy 1 and 2. Since $Z_0,ldots,n = varnothing$, we conclude that either $V_0,ldots,n = varnothing$ or $W_0,ldots,n = varnothing$. Hence, either the $V_i$ or the $W_i$ satisfy the conditions above, so $operatornameint. dim(X times Y) leq m$. $square$



    Corollary. There exists a homeomorphism $|mathbb A^n+m_k| stackrelsimto |mathbb A^n_k| times |mathbb A^m_k|$ if and only if $n = 0$ or $m = 0$.



    Proof. By Lemma 1 and Lemma 2, we have
    $$operatornameint. dim(|mathbb A^n_k| times |mathbb A^m_k|) = max(n,m),$$
    whereas
    $$operatornameint. dim(|mathbb A^n+m_k|) = n + m.$$
    These numbers are equal if and only if $n = 0$ or $m = 0$. Clearly the stated homeomorphisms exist in this case. $square$



    We can actually mix and match fields, i.e. consider things like $|mathbb A^100_mathbb C| times |mathbb A^17_barmathbb F_3|$. The same proof shows that this cannot be homeomorphic to any space of the form $|mathbb A^117_k|$ for a field $k$.






    share|cite|improve this answer
























      up vote
      12
      down vote



      accepted










      Clearly the natural map $|mathbb A^n+m| to |mathbb A^n| times |mathbb A^m|$ is not a homeomorphism unless $n = 0$ or $m = 0$ (it's not even a bijection, because of points like $(x-y) in operatornameSpec k[x] otimes k[y]$). We show that in fact there is no homeomorphism, using the following ad hoc notion of dimension:



      Definition. Let $X$ be a topological space. Then the intersection dimension $operatornameint. dim(X)$ of $X$ is the supremum of all $n$ for which there exist irreducible closed sets $V_0,ldots,V_n subseteq X$ such that



      1. any intersection $V_I = bigcap_iin I V_i$ for $I subseteq 0,ldots,n$ nonempty is either irreducible or empty;

      2. the intersection $V_I$ for any $I subseteq 0,ldots,n$ with $|I| = n$ is exactly one point;

      3. the intersection $V_0,ldots,n$ is empty.

      I feel that I have seen some variant of this somewhere before (probably under a different name), but I do not recall where. References are welcome!



      Remark. If there exist sets $V_0,ldots,V_n$ with properties 1, 2, and 3 above, then the same is true for any $m leq n$. Indeed, we may take the sets $V_0, ldots, V_m-1, V_m,ldots,n$.



      Lemma 1. Let $k$ be a field. Then $operatornameint. dim(mathbb A^n_k) = n$.



      Proof. The hyperplanes $V(x_1+ldots+x_n), V(x_1), ldots, V(x_n)$ show that $operatornameint. dim(mathbb A^n_k) geq n$. Conversely, let $V_0, ldots, V_m$ be irreducible closed subsets satisfying 1, 2, and 3 above. Conditions 2 and 3 imply that no $V_i$ is contained in any intersection of the others, so by induction on $|I|$ the dimension of $V_I$ is at most $n-|I|$ (because all intersections are irreducible). Taking $|I| = m$ gives $m leq n$, showing that $operatornameint. dim(mathbb A^n_k) leq n$. $square$



      More generally, this argument shows that if $X$ is a topological space of dimension $n$, then $operatornameint. dim(X) leq dim X$. It's not clear even for smooth proper varieties that the two are equal (in the projective case we might be able to use some Bertini argument, using tangency to reduce the intersections to one point).



      Lemma 2. Let $X$ and $Y$ be topological spaces with a closed point. Then
      $$operatornameint. dim(X times Y) = max(operatornameint. dim(X), operatornameint. dim(Y)).$$



      Proof. By the Lemma of this post, a subset $Z subseteq X times Y$ is closed and irreducible if and only if $Z = V times W$, with $V subseteq X$ and $W subseteq Y$ closed and irreducible.



      Let $m$ be the maximum of the intersection dimensions of $X$ and $Y$; without loss of generality let's say $m = operatornameint. dim(X)$. Then there exist sets $V_0,ldots,V_m subseteq X$ satisfying properties 1, 2, and 3 of the definition above. Choose a closed point $y in Y$, and set $W_i = y$ for all $i$. Then the sets $Z_i = V_i times W_i$ are closed and irreducible. Moreover, the finite intersections $Z_I = bigcap_i in I Z_i$ equal $V_I times W_I$. Since properties 1, 2, and 3 hold for the $V_i$ and properties 1 and 2 for the $W_i$, this shows properties 1, 2, and 3 for the $Z_i$. Therefore, $operatornameint. dim(X times Y) geq m$.



      Conversely, if $Z_0, ldots, Z_n$ satisfy properties 1, 2, and 3, then we must have $Z_i = V_i times W_i$ for some irreducible closed subsets $V_i subseteq X$, $W_i subseteq Y$. Again we have $Z_I = V_I times W_I$, so both $V_i$ and $W_i$ satisfy 1 and 2. Since $Z_0,ldots,n = varnothing$, we conclude that either $V_0,ldots,n = varnothing$ or $W_0,ldots,n = varnothing$. Hence, either the $V_i$ or the $W_i$ satisfy the conditions above, so $operatornameint. dim(X times Y) leq m$. $square$



      Corollary. There exists a homeomorphism $|mathbb A^n+m_k| stackrelsimto |mathbb A^n_k| times |mathbb A^m_k|$ if and only if $n = 0$ or $m = 0$.



      Proof. By Lemma 1 and Lemma 2, we have
      $$operatornameint. dim(|mathbb A^n_k| times |mathbb A^m_k|) = max(n,m),$$
      whereas
      $$operatornameint. dim(|mathbb A^n+m_k|) = n + m.$$
      These numbers are equal if and only if $n = 0$ or $m = 0$. Clearly the stated homeomorphisms exist in this case. $square$



      We can actually mix and match fields, i.e. consider things like $|mathbb A^100_mathbb C| times |mathbb A^17_barmathbb F_3|$. The same proof shows that this cannot be homeomorphic to any space of the form $|mathbb A^117_k|$ for a field $k$.






      share|cite|improve this answer






















        up vote
        12
        down vote



        accepted







        up vote
        12
        down vote



        accepted






        Clearly the natural map $|mathbb A^n+m| to |mathbb A^n| times |mathbb A^m|$ is not a homeomorphism unless $n = 0$ or $m = 0$ (it's not even a bijection, because of points like $(x-y) in operatornameSpec k[x] otimes k[y]$). We show that in fact there is no homeomorphism, using the following ad hoc notion of dimension:



        Definition. Let $X$ be a topological space. Then the intersection dimension $operatornameint. dim(X)$ of $X$ is the supremum of all $n$ for which there exist irreducible closed sets $V_0,ldots,V_n subseteq X$ such that



        1. any intersection $V_I = bigcap_iin I V_i$ for $I subseteq 0,ldots,n$ nonempty is either irreducible or empty;

        2. the intersection $V_I$ for any $I subseteq 0,ldots,n$ with $|I| = n$ is exactly one point;

        3. the intersection $V_0,ldots,n$ is empty.

        I feel that I have seen some variant of this somewhere before (probably under a different name), but I do not recall where. References are welcome!



        Remark. If there exist sets $V_0,ldots,V_n$ with properties 1, 2, and 3 above, then the same is true for any $m leq n$. Indeed, we may take the sets $V_0, ldots, V_m-1, V_m,ldots,n$.



        Lemma 1. Let $k$ be a field. Then $operatornameint. dim(mathbb A^n_k) = n$.



        Proof. The hyperplanes $V(x_1+ldots+x_n), V(x_1), ldots, V(x_n)$ show that $operatornameint. dim(mathbb A^n_k) geq n$. Conversely, let $V_0, ldots, V_m$ be irreducible closed subsets satisfying 1, 2, and 3 above. Conditions 2 and 3 imply that no $V_i$ is contained in any intersection of the others, so by induction on $|I|$ the dimension of $V_I$ is at most $n-|I|$ (because all intersections are irreducible). Taking $|I| = m$ gives $m leq n$, showing that $operatornameint. dim(mathbb A^n_k) leq n$. $square$



        More generally, this argument shows that if $X$ is a topological space of dimension $n$, then $operatornameint. dim(X) leq dim X$. It's not clear even for smooth proper varieties that the two are equal (in the projective case we might be able to use some Bertini argument, using tangency to reduce the intersections to one point).



        Lemma 2. Let $X$ and $Y$ be topological spaces with a closed point. Then
        $$operatornameint. dim(X times Y) = max(operatornameint. dim(X), operatornameint. dim(Y)).$$



        Proof. By the Lemma of this post, a subset $Z subseteq X times Y$ is closed and irreducible if and only if $Z = V times W$, with $V subseteq X$ and $W subseteq Y$ closed and irreducible.



        Let $m$ be the maximum of the intersection dimensions of $X$ and $Y$; without loss of generality let's say $m = operatornameint. dim(X)$. Then there exist sets $V_0,ldots,V_m subseteq X$ satisfying properties 1, 2, and 3 of the definition above. Choose a closed point $y in Y$, and set $W_i = y$ for all $i$. Then the sets $Z_i = V_i times W_i$ are closed and irreducible. Moreover, the finite intersections $Z_I = bigcap_i in I Z_i$ equal $V_I times W_I$. Since properties 1, 2, and 3 hold for the $V_i$ and properties 1 and 2 for the $W_i$, this shows properties 1, 2, and 3 for the $Z_i$. Therefore, $operatornameint. dim(X times Y) geq m$.



        Conversely, if $Z_0, ldots, Z_n$ satisfy properties 1, 2, and 3, then we must have $Z_i = V_i times W_i$ for some irreducible closed subsets $V_i subseteq X$, $W_i subseteq Y$. Again we have $Z_I = V_I times W_I$, so both $V_i$ and $W_i$ satisfy 1 and 2. Since $Z_0,ldots,n = varnothing$, we conclude that either $V_0,ldots,n = varnothing$ or $W_0,ldots,n = varnothing$. Hence, either the $V_i$ or the $W_i$ satisfy the conditions above, so $operatornameint. dim(X times Y) leq m$. $square$



        Corollary. There exists a homeomorphism $|mathbb A^n+m_k| stackrelsimto |mathbb A^n_k| times |mathbb A^m_k|$ if and only if $n = 0$ or $m = 0$.



        Proof. By Lemma 1 and Lemma 2, we have
        $$operatornameint. dim(|mathbb A^n_k| times |mathbb A^m_k|) = max(n,m),$$
        whereas
        $$operatornameint. dim(|mathbb A^n+m_k|) = n + m.$$
        These numbers are equal if and only if $n = 0$ or $m = 0$. Clearly the stated homeomorphisms exist in this case. $square$



        We can actually mix and match fields, i.e. consider things like $|mathbb A^100_mathbb C| times |mathbb A^17_barmathbb F_3|$. The same proof shows that this cannot be homeomorphic to any space of the form $|mathbb A^117_k|$ for a field $k$.






        share|cite|improve this answer












        Clearly the natural map $|mathbb A^n+m| to |mathbb A^n| times |mathbb A^m|$ is not a homeomorphism unless $n = 0$ or $m = 0$ (it's not even a bijection, because of points like $(x-y) in operatornameSpec k[x] otimes k[y]$). We show that in fact there is no homeomorphism, using the following ad hoc notion of dimension:



        Definition. Let $X$ be a topological space. Then the intersection dimension $operatornameint. dim(X)$ of $X$ is the supremum of all $n$ for which there exist irreducible closed sets $V_0,ldots,V_n subseteq X$ such that



        1. any intersection $V_I = bigcap_iin I V_i$ for $I subseteq 0,ldots,n$ nonempty is either irreducible or empty;

        2. the intersection $V_I$ for any $I subseteq 0,ldots,n$ with $|I| = n$ is exactly one point;

        3. the intersection $V_0,ldots,n$ is empty.

        I feel that I have seen some variant of this somewhere before (probably under a different name), but I do not recall where. References are welcome!



        Remark. If there exist sets $V_0,ldots,V_n$ with properties 1, 2, and 3 above, then the same is true for any $m leq n$. Indeed, we may take the sets $V_0, ldots, V_m-1, V_m,ldots,n$.



        Lemma 1. Let $k$ be a field. Then $operatornameint. dim(mathbb A^n_k) = n$.



        Proof. The hyperplanes $V(x_1+ldots+x_n), V(x_1), ldots, V(x_n)$ show that $operatornameint. dim(mathbb A^n_k) geq n$. Conversely, let $V_0, ldots, V_m$ be irreducible closed subsets satisfying 1, 2, and 3 above. Conditions 2 and 3 imply that no $V_i$ is contained in any intersection of the others, so by induction on $|I|$ the dimension of $V_I$ is at most $n-|I|$ (because all intersections are irreducible). Taking $|I| = m$ gives $m leq n$, showing that $operatornameint. dim(mathbb A^n_k) leq n$. $square$



        More generally, this argument shows that if $X$ is a topological space of dimension $n$, then $operatornameint. dim(X) leq dim X$. It's not clear even for smooth proper varieties that the two are equal (in the projective case we might be able to use some Bertini argument, using tangency to reduce the intersections to one point).



        Lemma 2. Let $X$ and $Y$ be topological spaces with a closed point. Then
        $$operatornameint. dim(X times Y) = max(operatornameint. dim(X), operatornameint. dim(Y)).$$



        Proof. By the Lemma of this post, a subset $Z subseteq X times Y$ is closed and irreducible if and only if $Z = V times W$, with $V subseteq X$ and $W subseteq Y$ closed and irreducible.



        Let $m$ be the maximum of the intersection dimensions of $X$ and $Y$; without loss of generality let's say $m = operatornameint. dim(X)$. Then there exist sets $V_0,ldots,V_m subseteq X$ satisfying properties 1, 2, and 3 of the definition above. Choose a closed point $y in Y$, and set $W_i = y$ for all $i$. Then the sets $Z_i = V_i times W_i$ are closed and irreducible. Moreover, the finite intersections $Z_I = bigcap_i in I Z_i$ equal $V_I times W_I$. Since properties 1, 2, and 3 hold for the $V_i$ and properties 1 and 2 for the $W_i$, this shows properties 1, 2, and 3 for the $Z_i$. Therefore, $operatornameint. dim(X times Y) geq m$.



        Conversely, if $Z_0, ldots, Z_n$ satisfy properties 1, 2, and 3, then we must have $Z_i = V_i times W_i$ for some irreducible closed subsets $V_i subseteq X$, $W_i subseteq Y$. Again we have $Z_I = V_I times W_I$, so both $V_i$ and $W_i$ satisfy 1 and 2. Since $Z_0,ldots,n = varnothing$, we conclude that either $V_0,ldots,n = varnothing$ or $W_0,ldots,n = varnothing$. Hence, either the $V_i$ or the $W_i$ satisfy the conditions above, so $operatornameint. dim(X times Y) leq m$. $square$



        Corollary. There exists a homeomorphism $|mathbb A^n+m_k| stackrelsimto |mathbb A^n_k| times |mathbb A^m_k|$ if and only if $n = 0$ or $m = 0$.



        Proof. By Lemma 1 and Lemma 2, we have
        $$operatornameint. dim(|mathbb A^n_k| times |mathbb A^m_k|) = max(n,m),$$
        whereas
        $$operatornameint. dim(|mathbb A^n+m_k|) = n + m.$$
        These numbers are equal if and only if $n = 0$ or $m = 0$. Clearly the stated homeomorphisms exist in this case. $square$



        We can actually mix and match fields, i.e. consider things like $|mathbb A^100_mathbb C| times |mathbb A^17_barmathbb F_3|$. The same proof shows that this cannot be homeomorphic to any space of the form $|mathbb A^117_k|$ for a field $k$.







        share|cite|improve this answer












        share|cite|improve this answer



        share|cite|improve this answer










        answered Aug 28 at 6:45









        Remy

        3,3641616




        3,3641616



























             

            draft saved


            draft discarded















































             


            draft saved


            draft discarded














            StackExchange.ready(
            function ()
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2901329%2fzariski-topology-and-product-topology-on-mathbb-a-km-times-mathbb-a-kn%23new-answer', 'question_page');

            );

            Post as a guest













































































            這個網誌中的熱門文章

            How to combine Bézier curves to a surface?

            Mutual Information Always Non-negative

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