Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am self studying from A Book of Abstract Algebra by Charles C. Pinter. In chapter 3 problem set B problem 4 it asks to show whether $(a,b)\star(c,d)=(ac-bd,ad+bc)$ on the set $ \{(x,y) \in \mathbb R^2 | (x,y) \not= (0,0)\} $ is a group. Currently I am in the middle of showing that there is an inverse. This is the equation I am trying to solve to show that there is an inverse.

$(a,b) \star (a',b')=(1,0)=(aa' - bb',ab' + ba')$

I am pretty sure I have the answer for the first entry of the inverse element but I pass through steps with division. In particular I am wondering whether $[1]$ to $[2]$ is a valid step and if it is why. In other words if I exclude $0$ at one step do I have to keep $0$ excluded for the rest of the proof? If its not legal would I just have to split the problem into a couple of cases?

$[1]$ $\frac{aa'-1}{b} =\frac{-ba'}{a}$ where $a\not=0$ and $b\not=0\\$

$[2]$ $a^2 a' - a = -b^2 a'$ (note I did not exclude $0$)

share|cite|improve this question
Are you allowed to notice that this is just the multiplicative group of nonzero complex numbers? In the right solution the only divisor is $a^2+b^2$ which is known to be nonzero because $a$ and $b$ are not both $0$. – Henning Makholm Jun 28 '12 at 20:15
I would say no because complex numbers have not been introduced in the text and its an exercise to learn how to prove something is a group. – Paul Plummer Jun 28 '12 at 20:45
Note: $(a,b)\neq (0,0)$ does not mean $a\neq 0$ and $b\neq 0$. It means either $a\neq 0$ or $b\neq 0$. So I don't think that [1] is valid. – Arturo Magidin Jun 28 '12 at 20:49
@Paul: If what you're supposed to learn is just how to prove it is a group, then it is allowed simply to pull the correct expression for the inverse out of a hat and then prove that it happens to satisfy the required condition. You don't have to derive it from scratch. The hat might involve lucky guesses and/or previous knowledge of complex numbers; that doesn't affect the validity of the resulting proof. – Henning Makholm Jun 28 '12 at 20:53
@Paul: That depends on what the exercise is. If the exercise is "prove such-and-such", then where the ideas behind the proof comes from is not part of the answer -- just conducting the proof itself is. Indeed one (minor) point of such exercises might be to develop an understanding of the difference between the a derivation and a proof. – Henning Makholm Jun 28 '12 at 23:46
up vote 1 down vote accepted

You want $a'$ and $b'$ such that $aa'-bb'=1$ and $ab'+ba'=0$. In effect you want to show that the system

$$\left\{\begin{align*} &ax-by=1\\ &bx+ay=0 \end{align*}\right.\tag{1}$$

always has a solution with $\langle x,y\rangle\ne\langle 0,0\rangle$ provided that $\langle a,b\rangle\ne\langle 0,0\rangle$. You can do this in several ways. If you already know the relevant linear algebra, you can note that $$\det\pmatrix{a&-b\\b&a}=a^2+b^2\ge 0\;,$$

with equality iff $a=b=0$, so $(1)$ always has a unique solution, which clearly (by virtue of the first equation of $(1)$) cannot be $x=y=0$.

Alternatively, you can work out the solution by whatever technique appeals to you and then show directly that it exists and is a non-$\langle 0,0\rangle$ solution whenever $a\ne 0\ne b$.

Your approach, however, does require you to consider separate cases, since it’s entirely possible that one side of $[1]$ is undefined.

share|cite|improve this answer

Hint: Let $a=r\cos\theta$, $b=r\sin\theta$, and let $a'=r'\cos\theta'$, $b'=r'\sin\theta'$. (As usual choose $r$, $r'$ positive.) It is easy to see that $r'=1/r$. Now you can identify $\theta'$ by recalling some trigonometric identities.

Or else if you are familiar with complex numbers, there is a much more elegant version of the same thing. We have $(a+bi)(c+di)=ac-bd++(ad+bc)i$. So $$\frac{1}{a+bi}=\frac{a-bi}{(a+bi)(a-bi)}=\frac{a}{a^2+b^2}-\frac{b}{a^2+b^2}i.$$

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.