Contrapositive proof: If $H$ and $K$ are nontrivial subgroups of $\Bbb Q$, then $H\cap K$ is also nontrivial.

1.1k Views Asked by At

I'm reading "Contemporary Abstract Algebra," by Gallian.

This is (part of) Exercise 26 of the supplementary exercises for chapters 1-4 ibid., although I am requesting a proof of the contrapositive just out of interest, preferably using prior material available from the textbook.

I intend to use abstract algebra, keeping in the spirit of the textbook; thus proofs that rely on, say, real analysis (too much), will not be accepted.

$\mathcal{O}$: Suppose that $H$ and $K$ are nontrivial subgroups of $\Bbb Q$ under addition. Show that $H \cap K$ is a nontrivial subgroup of $\Bbb Q$.

The contrapositive problem can be stated as follows:

$\mathcal{C}$: Let $H$ and $K$ be subgroups of $\Bbb Q$. Suppose $H\cap K=\{0\}$ is trivial. Show that either $H$ or $K$ is trivial.


Thoughts:

To me, it seems $\mathcal{C}$ is one of those things that are so tauntingly "obvious", it's hard to know where to begin.

My problem is also that my only thought I have worth sharing for this is to try assuming - no joke! - that neither $H$ nor $K$ is trivial, then deducing $H\cap K$ is nontrivial, contradicting the hypothesis that $H\cap K=\{0\}$. That's just silly.


A lovely proof of the original statement $\mathcal{O}$ is given here, with variations on its theme here and here.

Please help :)

2

There are 2 best solutions below

1
On BEST ANSWER

Let $P$ denote the property that triviality of an intersection of two subgroups implies triviality of either.

1: If all $G_i$ have the property $P$, then so does

$$\lim_{\longrightarrow}G_i .$$

2: Define $\mathbb{Z}^{(n)}=\mathbb{Z}$, and for $n|m$,

\begin{align} f_{n,m}:\mathbb{Z}^{(n)} &\longrightarrow\mathbb{Z}^{(m)}, \\ k &\mapsto\frac{m}{n}\cdot k. \end{align}

Then $$\mathbb{Q}=\lim_{\longrightarrow}\mathbb{Z}^{(n)}.$$

3: $\mathbb{Z}$ has the property $P$. In fact, $H\cap K=0$ implies the injectivity of $$\mathbb{Z}\longrightarrow \mathbb{Z}/H \times \mathbb{Z}/K,$$ and thus $H=0$ or $K=0$.

7
On

This follows from the fact that any two rational non-zero numbers $a/b$ and $c/d$ have a common multiple, say $$ cb \cdot \frac{a}{b} = ad \cdot \frac{c}{d}. $$

It will be unnatural to give a proof using abstract algebra only, as this is a simple arithmetic fact depending on specific properties of rational numbers.