Left inverse in $F_{A}$ iff injective proof.

Solution 1:

Suppose $f$ has a left inverse $g$.
Suppose $f(x)=f(y)$ for some $x,y\in A$.
Then $gf(x)=gf(y)$.
Since $gf=i_A$, we have $x=y$ and hence $f$ is injective.

Suppose $f$ is injective. Let $a_0$ be a fixed element in $A$.
Define $g:A\rightarrow A$ by the rules:
For $x\in A$
(i) If there is an element $y\in A$ such that $f(y)=x$, then $g(x)=y$.
(ii) If no such element $y$ exists in $A$, then $g(x)=a_0$.

For case (i), the element $y$ is unique as if $g(x)=y_1$ and $g(x)=y_2$, then $f(y_1)=x=f(y_2)$ which implies $y_1=y_2$.
For case (ii), $g(x)=a_0$ is also unique by the way we define it.
So $g$ is a function from $A$ to $A$.

For $a\in A$, $gf(a)=a$ as $f(a)=x\in A$ implies $g(x)=a$ by (i).
Hence $g$ is a left inverse of $f$.

Solution 2:

Each of the two answers in this article have flaws in their proofs that every injective function has a left inverse, so here we set the record straight.

The accepted answer doesn't account for the fact that $A$ might be empty (this is really hair splitting). The answer by Reuben Stern only constructs a left inverse when the function $f$ is a bijection.

If $f: A \to A$ is injective, the converse relation $f^{-1}$ will define a bijective function

$$ f^{-1}: f(A) \to A$$

satisfying $f^{-1} \circ f = id_A$.

If $f(A) = A$ then $f^{-1} \in F_A$ and we are done.

If $f$ is not surjective let $a_0 \in A \setminus f (A)$ and define the function $g: A \to A$ by

$g(a) = \left\{\begin{array}{lr} f^{-1}(a)\, \;\;\;\text{ |} & \text{for } a \in f(A)\\ a_0 \quad \quad \,\,\;\; \text{ |} & \text{for } a \in A \setminus f (A) \end{array}\right\}$

It is easily checked that $g \circ f = id_A$.

Solution 3:

First of all, a one-to-one set function $f: A \to A$ certainly need not be a bijection; consider the multiplication by two map $\times 2: \mathbb{Z} \to \mathbb{Z}$. This is injective, but no odd numbers are in the image.

Now for your main question, assume $f$ is injective. Then define a left inverse $g$ by setting $g(x) = f^{-1}(x)$ wherever $f^{-1}(x)$ is defined. Because $f$ is injective, there is only one possible value for $f^{-1}(x)$, whenever it exists. Thus, $g$ is a left inverse of $f$ (because $g(f(x)) = x$ by definition). In the other direction, suppose that $g$ is a left inverse to $f$, i.e. for all $x \in A$, $g(f(x)) = x$. Now suppose that $f(x) = f(y)$. Applying $g$ to both sides, we see that $g(f(x)) = g(f(y))$, and since $g$ is a left inverse, this gives $x = y$. Thus $f$ is injective.