Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Why is this way of computing imaginary arctangents wrong?

  1. Dec 10, 2014 #1
    I was doing some things in my head the other day (these moments usually don't come out so well :rolleyes:). And I "came up" with the following way to compute the arctangent with imaginary arguments.

    Consider the identity x2 = - 2, where x = ± i√2.

    Now rearrange this identity to x2 + 1 = -1

    Invert the equation to get the reciprocals 1/(x2 + 1) = - 1

    Multiply both sides by arctan(x)

    arctan(x)(1/(x2 + 1)) = - arctan(x)

    Let y = arctan(x)

    Then we have the differential equation yy' = - y (Because 1/(x^2 + 1) is the derivative of arctan(x))

    Divide both sides by y to get y' = - 1

    Now integrate both sides ∫dy = - ∫dx

    The result is y = - x

    But since y = arctan(x), it follows that arctan(x) = - x

    And since we know that x = ± i√2, it follows that arctan(i√2) = - i√2 and arctan(-i√2) = i√2

    I know I must have done something wrong here. Could someone please point out the errors (which I'm sure are many o0)) in the above?
  2. jcsd
  3. Dec 10, 2014 #2


    Staff: Mentor

    The above is not an identity. It is a conditional equation that is true only for two values of x. Here are two examples of identities:
    1. x2 + 2x + 1 = (x + 1)2
    2. sin(2x) = 2sin(x)cos(x)
    arctan(x) is a constant, so its derivative is 0.
  4. Dec 10, 2014 #3
    Thanks, Mark44. I actually just realized the error about 17 seconds after I clicked on "Create Thread".

    I definitely should not quit my day welfare.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook