Intuition behind neglecting higher order differentials in visual proofs of the Product Rule

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











up vote
3
down vote

favorite
1












In most visual explanations of the product rule (like this video), the justification for the remaining square ($dfcdot dg$) being neglected is that it is "very small". But the whole point of the intuition is that the $df cdot g$ and $dg cdot f$ are very small too.



As such, I'm interested in either an intuitive explanation of this, or an other intuitive explanation for the product rule that takes this into account.



Thanks in advance







share|cite|improve this question

























    up vote
    3
    down vote

    favorite
    1












    In most visual explanations of the product rule (like this video), the justification for the remaining square ($dfcdot dg$) being neglected is that it is "very small". But the whole point of the intuition is that the $df cdot g$ and $dg cdot f$ are very small too.



    As such, I'm interested in either an intuitive explanation of this, or an other intuitive explanation for the product rule that takes this into account.



    Thanks in advance







    share|cite|improve this question























      up vote
      3
      down vote

      favorite
      1









      up vote
      3
      down vote

      favorite
      1






      1





      In most visual explanations of the product rule (like this video), the justification for the remaining square ($dfcdot dg$) being neglected is that it is "very small". But the whole point of the intuition is that the $df cdot g$ and $dg cdot f$ are very small too.



      As such, I'm interested in either an intuitive explanation of this, or an other intuitive explanation for the product rule that takes this into account.



      Thanks in advance







      share|cite|improve this question













      In most visual explanations of the product rule (like this video), the justification for the remaining square ($dfcdot dg$) being neglected is that it is "very small". But the whole point of the intuition is that the $df cdot g$ and $dg cdot f$ are very small too.



      As such, I'm interested in either an intuitive explanation of this, or an other intuitive explanation for the product rule that takes this into account.



      Thanks in advance









      share|cite|improve this question












      share|cite|improve this question




      share|cite|improve this question








      edited Jul 27 at 19:13









      Blue

      43.6k868141




      43.6k868141









      asked Jul 27 at 19:03









      Gabriel Alfour

      243




      243




















          3 Answers
          3






          active

          oldest

          votes

















          up vote
          5
          down vote



          accepted










          It might help to compare the product rule against this (exact) equation for finite differences, where $Delta f(x)$ means $f(x + delta) - f(x)$:



          $$ Delta( fg(x) ) = g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)$$



          If you find this equation helps you adequately intuit the product rule, then to answer the posted question observe the approximation



          $$ g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)
          approx g(x) cdot Delta f(x) + f(x) cdot Delta g(x) $$



          ought to have negligible error; the difference between the two formulas is precisely $Delta f(x) Delta g(x)$.




          At some point you may be better off simply working through the algebra behind the proof of the product rule and basing your intuition off of the algebra directly, rather than looking for heuristic rationalizations.






          share|cite|improve this answer





















          • I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
            – Gabriel Alfour
            Jul 27 at 20:33

















          up vote
          8
          down vote













          Indeed $dfcdot g$ and $gcdot df$ are very small but $dfcdot dg$ is even smaller.



          For the intuition, let assume for examaple



          • $Delta fapprox Delta gapprox10^-3$

          • $fapprox gapprox10$

          then



          $$Delta f cdot g approx10^-2ggDelta f cdot Delta g approx10^-6$$






          share|cite|improve this answer



















          • 1




            @AOrtiz Thanks for the edit and correction for the typo!
            – gimusi
            Jul 27 at 19:11

















          up vote
          8
          down vote













          The derivative is about the first order term in the expanded functions, higher order terms vanish in the limit:
          beginalign
          (f(x) , g(x))'
          &= lim_hto 0 fracf(x+h), g(x+h) - f(x), g(x)h \
          &= lim_hto 0 frac(f(x) + f'(x) h + O(h^2)), (g(x) + g'(x) h + O(h^2))) - f(x), g(x)h \
          &= lim_hto 0 fracf(x) , g(x) + f'(x) , g(x) h + f(x) , g'(x) h + O(h^2) - f(x), g(x)h \
          &= lim_hto 0 f'(x) , g(x) + f(x), g'(x) + O(h) \
          &= f'(x) , g(x) + f(x) , g'(x)
          endalign
          So too small in this context means being a term of order $k$ in $h$ with $k ge 2$.






          share|cite|improve this answer

















          • 1




            Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
            – Hurkyl
            Jul 27 at 19:35











          • In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
            – Gabriel Alfour
            Jul 27 at 20:38










          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%2f2864692%2fintuition-behind-neglecting-higher-order-differentials-in-visual-proofs-of-the-p%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
          5
          down vote



          accepted










          It might help to compare the product rule against this (exact) equation for finite differences, where $Delta f(x)$ means $f(x + delta) - f(x)$:



          $$ Delta( fg(x) ) = g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)$$



          If you find this equation helps you adequately intuit the product rule, then to answer the posted question observe the approximation



          $$ g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)
          approx g(x) cdot Delta f(x) + f(x) cdot Delta g(x) $$



          ought to have negligible error; the difference between the two formulas is precisely $Delta f(x) Delta g(x)$.




          At some point you may be better off simply working through the algebra behind the proof of the product rule and basing your intuition off of the algebra directly, rather than looking for heuristic rationalizations.






          share|cite|improve this answer





















          • I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
            – Gabriel Alfour
            Jul 27 at 20:33














          up vote
          5
          down vote



          accepted










          It might help to compare the product rule against this (exact) equation for finite differences, where $Delta f(x)$ means $f(x + delta) - f(x)$:



          $$ Delta( fg(x) ) = g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)$$



          If you find this equation helps you adequately intuit the product rule, then to answer the posted question observe the approximation



          $$ g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)
          approx g(x) cdot Delta f(x) + f(x) cdot Delta g(x) $$



          ought to have negligible error; the difference between the two formulas is precisely $Delta f(x) Delta g(x)$.




          At some point you may be better off simply working through the algebra behind the proof of the product rule and basing your intuition off of the algebra directly, rather than looking for heuristic rationalizations.






          share|cite|improve this answer





















          • I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
            – Gabriel Alfour
            Jul 27 at 20:33












          up vote
          5
          down vote



          accepted







          up vote
          5
          down vote



          accepted






          It might help to compare the product rule against this (exact) equation for finite differences, where $Delta f(x)$ means $f(x + delta) - f(x)$:



          $$ Delta( fg(x) ) = g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)$$



          If you find this equation helps you adequately intuit the product rule, then to answer the posted question observe the approximation



          $$ g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)
          approx g(x) cdot Delta f(x) + f(x) cdot Delta g(x) $$



          ought to have negligible error; the difference between the two formulas is precisely $Delta f(x) Delta g(x)$.




          At some point you may be better off simply working through the algebra behind the proof of the product rule and basing your intuition off of the algebra directly, rather than looking for heuristic rationalizations.






          share|cite|improve this answer













          It might help to compare the product rule against this (exact) equation for finite differences, where $Delta f(x)$ means $f(x + delta) - f(x)$:



          $$ Delta( fg(x) ) = g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)$$



          If you find this equation helps you adequately intuit the product rule, then to answer the posted question observe the approximation



          $$ g(x) cdot Delta f(x) + f(x+delta) cdot Delta g(x)
          approx g(x) cdot Delta f(x) + f(x) cdot Delta g(x) $$



          ought to have negligible error; the difference between the two formulas is precisely $Delta f(x) Delta g(x)$.




          At some point you may be better off simply working through the algebra behind the proof of the product rule and basing your intuition off of the algebra directly, rather than looking for heuristic rationalizations.







          share|cite|improve this answer













          share|cite|improve this answer



          share|cite|improve this answer











          answered Jul 27 at 19:18









          Hurkyl

          107k9112253




          107k9112253











          • I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
            – Gabriel Alfour
            Jul 27 at 20:33
















          • I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
            – Gabriel Alfour
            Jul 27 at 20:33















          I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
          – Gabriel Alfour
          Jul 27 at 20:33




          I already worked through the algebra. I'm looking for intuitions, so that I can manipulate the equations as mental concepts rather than going through computations everytime. But I think you are pointing at the right thing. These $df$/$dg$ are entailed by the $dx$ representation. I think I should be looking at something like the slope of a graph at a given point, but for function of $R^2$ to $R$. After checking Google, it's actually called the total derivative, and makes much more sense framed this way.
          – Gabriel Alfour
          Jul 27 at 20:33










          up vote
          8
          down vote













          Indeed $dfcdot g$ and $gcdot df$ are very small but $dfcdot dg$ is even smaller.



          For the intuition, let assume for examaple



          • $Delta fapprox Delta gapprox10^-3$

          • $fapprox gapprox10$

          then



          $$Delta f cdot g approx10^-2ggDelta f cdot Delta g approx10^-6$$






          share|cite|improve this answer



















          • 1




            @AOrtiz Thanks for the edit and correction for the typo!
            – gimusi
            Jul 27 at 19:11














          up vote
          8
          down vote













          Indeed $dfcdot g$ and $gcdot df$ are very small but $dfcdot dg$ is even smaller.



          For the intuition, let assume for examaple



          • $Delta fapprox Delta gapprox10^-3$

          • $fapprox gapprox10$

          then



          $$Delta f cdot g approx10^-2ggDelta f cdot Delta g approx10^-6$$






          share|cite|improve this answer



















          • 1




            @AOrtiz Thanks for the edit and correction for the typo!
            – gimusi
            Jul 27 at 19:11












          up vote
          8
          down vote










          up vote
          8
          down vote









          Indeed $dfcdot g$ and $gcdot df$ are very small but $dfcdot dg$ is even smaller.



          For the intuition, let assume for examaple



          • $Delta fapprox Delta gapprox10^-3$

          • $fapprox gapprox10$

          then



          $$Delta f cdot g approx10^-2ggDelta f cdot Delta g approx10^-6$$






          share|cite|improve this answer















          Indeed $dfcdot g$ and $gcdot df$ are very small but $dfcdot dg$ is even smaller.



          For the intuition, let assume for examaple



          • $Delta fapprox Delta gapprox10^-3$

          • $fapprox gapprox10$

          then



          $$Delta f cdot g approx10^-2ggDelta f cdot Delta g approx10^-6$$







          share|cite|improve this answer















          share|cite|improve this answer



          share|cite|improve this answer








          edited Jul 27 at 19:10









          AOrtiz

          8,84621238




          8,84621238











          answered Jul 27 at 19:09









          gimusi

          64.9k73483




          64.9k73483







          • 1




            @AOrtiz Thanks for the edit and correction for the typo!
            – gimusi
            Jul 27 at 19:11












          • 1




            @AOrtiz Thanks for the edit and correction for the typo!
            – gimusi
            Jul 27 at 19:11







          1




          1




          @AOrtiz Thanks for the edit and correction for the typo!
          – gimusi
          Jul 27 at 19:11




          @AOrtiz Thanks for the edit and correction for the typo!
          – gimusi
          Jul 27 at 19:11










          up vote
          8
          down vote













          The derivative is about the first order term in the expanded functions, higher order terms vanish in the limit:
          beginalign
          (f(x) , g(x))'
          &= lim_hto 0 fracf(x+h), g(x+h) - f(x), g(x)h \
          &= lim_hto 0 frac(f(x) + f'(x) h + O(h^2)), (g(x) + g'(x) h + O(h^2))) - f(x), g(x)h \
          &= lim_hto 0 fracf(x) , g(x) + f'(x) , g(x) h + f(x) , g'(x) h + O(h^2) - f(x), g(x)h \
          &= lim_hto 0 f'(x) , g(x) + f(x), g'(x) + O(h) \
          &= f'(x) , g(x) + f(x) , g'(x)
          endalign
          So too small in this context means being a term of order $k$ in $h$ with $k ge 2$.






          share|cite|improve this answer

















          • 1




            Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
            – Hurkyl
            Jul 27 at 19:35











          • In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
            – Gabriel Alfour
            Jul 27 at 20:38














          up vote
          8
          down vote













          The derivative is about the first order term in the expanded functions, higher order terms vanish in the limit:
          beginalign
          (f(x) , g(x))'
          &= lim_hto 0 fracf(x+h), g(x+h) - f(x), g(x)h \
          &= lim_hto 0 frac(f(x) + f'(x) h + O(h^2)), (g(x) + g'(x) h + O(h^2))) - f(x), g(x)h \
          &= lim_hto 0 fracf(x) , g(x) + f'(x) , g(x) h + f(x) , g'(x) h + O(h^2) - f(x), g(x)h \
          &= lim_hto 0 f'(x) , g(x) + f(x), g'(x) + O(h) \
          &= f'(x) , g(x) + f(x) , g'(x)
          endalign
          So too small in this context means being a term of order $k$ in $h$ with $k ge 2$.






          share|cite|improve this answer

















          • 1




            Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
            – Hurkyl
            Jul 27 at 19:35











          • In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
            – Gabriel Alfour
            Jul 27 at 20:38












          up vote
          8
          down vote










          up vote
          8
          down vote









          The derivative is about the first order term in the expanded functions, higher order terms vanish in the limit:
          beginalign
          (f(x) , g(x))'
          &= lim_hto 0 fracf(x+h), g(x+h) - f(x), g(x)h \
          &= lim_hto 0 frac(f(x) + f'(x) h + O(h^2)), (g(x) + g'(x) h + O(h^2))) - f(x), g(x)h \
          &= lim_hto 0 fracf(x) , g(x) + f'(x) , g(x) h + f(x) , g'(x) h + O(h^2) - f(x), g(x)h \
          &= lim_hto 0 f'(x) , g(x) + f(x), g'(x) + O(h) \
          &= f'(x) , g(x) + f(x) , g'(x)
          endalign
          So too small in this context means being a term of order $k$ in $h$ with $k ge 2$.






          share|cite|improve this answer













          The derivative is about the first order term in the expanded functions, higher order terms vanish in the limit:
          beginalign
          (f(x) , g(x))'
          &= lim_hto 0 fracf(x+h), g(x+h) - f(x), g(x)h \
          &= lim_hto 0 frac(f(x) + f'(x) h + O(h^2)), (g(x) + g'(x) h + O(h^2))) - f(x), g(x)h \
          &= lim_hto 0 fracf(x) , g(x) + f'(x) , g(x) h + f(x) , g'(x) h + O(h^2) - f(x), g(x)h \
          &= lim_hto 0 f'(x) , g(x) + f(x), g'(x) + O(h) \
          &= f'(x) , g(x) + f(x) , g'(x)
          endalign
          So too small in this context means being a term of order $k$ in $h$ with $k ge 2$.







          share|cite|improve this answer













          share|cite|improve this answer



          share|cite|improve this answer











          answered Jul 27 at 19:26









          mvw

          30.2k22250




          30.2k22250







          • 1




            Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
            – Hurkyl
            Jul 27 at 19:35











          • In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
            – Gabriel Alfour
            Jul 27 at 20:38












          • 1




            Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
            – Hurkyl
            Jul 27 at 19:35











          • In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
            – Gabriel Alfour
            Jul 27 at 20:38







          1




          1




          Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
          – Hurkyl
          Jul 27 at 19:35





          Aside; this calculation assumes $f$ and $g$ are at least twice differentiable (which is not necessarily a defect). It can be made more general by using the little-oh bound $f(x+h) = f(x) + f'(x) h + o(h)$.
          – Hurkyl
          Jul 27 at 19:35













          In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
          – Gabriel Alfour
          Jul 27 at 20:38




          In the algebra I see that the higher order term vanish. The remaining term is a $f(x +h) cdot g(x+h) / h$, so one of them is a derivative, while the other is equal to zero. However, I don't see how it can be represented on those rectangles.
          – Gabriel Alfour
          Jul 27 at 20:38












           

          draft saved


          draft discarded


























           


          draft saved


          draft discarded














          StackExchange.ready(
          function ()
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2864692%2fintuition-behind-neglecting-higher-order-differentials-in-visual-proofs-of-the-p%23new-answer', 'question_page');

          );

          Post as a guest













































































          Comments

          Popular posts from this blog

          What is the equation of a 3D cone with generalised tilt?

          Color the edges and diagonals of a regular polygon

          Relationship between determinant of matrix and determinant of adjoint?