Why do we need to prove that functions with "specific properties" exist?

109 Views Asked by At

There are many exercises I come across where the objective is to prove the existence of a function. Consider the following three examples from various posts on this website:

1) Conceptual question about assuming the existence of a function in order to prove the existence of another function

Let $f:\mathbb N \times \mathbb N \to \mathbb N$ be a function, and let $c$ be a natural number. Show that there exists a function $a : \mathbb N \to \mathbb N$ such that $$a(0) =c $$ and $$a(n++)=f(n,a(n)) \text{ for all } n \in \mathbb N$$

2) Show that there exists a unique function $h:X\cup Y\to Z$ such that $h\circ\iota_{X\to X\cup Y}=f$ and $h\circ\iota_{Y\to X\cup Y}=g$.

Show that if $X$ and $Y$ are disjoint sets and $f:X\to Y$ and $g:Y\to Z$ are functions, then there exists a unique function $h:X\cup Y\to Z$ such that $h\circ\iota_{X\to X\cup Y}=f$ and $h\circ\iota_{Y\to X\cup Y}=g$.

3) Exercise 3.5.7 of Tao's Analysis I: A question about direct sums.

Let $X,Y$ be sets, and let $\pi_{X\times Y\to X}: X \times Y \to X$ and $\pi_{X\times Y\to Y}: X \times Y \to Y$ be the maps $\pi_{X\times Y\to X}(x,y):=x$ and $\pi_{X\times Y\to Y}(x,y):=y$; these maps are known as the co-ordinate functions on $X \times Y$. Show that for any functions $f: Z \to X$ and $g: Z \to Y$, there exists a unique function $h: Z \to X \times Y$ such that $\pi_{X\times Y\to X} \circ h =f$ and $\pi_{X\times Y\to Y} \circ g =f$.

I recently saw how Cartesian Products, the Power Set Axiom, and the Axiom of Specification could be used to construct a set, denoted $Y^X$, that consists of all functions from set $X$ to set $Y$...where $f \in Y^X \iff \text {$f$ is a function with domain $X$ and range $Y$}$. The proof can be found here: Hints for proving that the collection of all set functions whose domain is the set $X$ and whose range is a subset or equal to the set $Y$ is a set..


Now, referring back to the three aforementioned examples, in all three cases, we already know the domain and codomain.

For the first example, $a \in \mathbb N^{\mathbb N}$. In the second example, $h \in (X \cup Y) ^{Z}$. In the third example, $h \in Z^{X \times Y}$.

Because we know that the set $codomain^{domain}$ exists, and such a set contains all functions that have these corresponding domains and codomains, I fail to see why any proof needs to occur at all.

Regardless of the properties that functions $a$ and $h$ have, they must already exist because if the set that contains them exists, then clearly so do the elements of said set!

Any clarity would be greatly appreciated. Thanks!

1

There are 1 best solutions below

7
On BEST ANSWER

If you don't see why a problem is nontrivial, it often helps to consider silly variations of it which are outright wrong.


To start with, consider:

Show that there exists a function $f:\mathbb{N}\rightarrow \mathbb{N}$ such that $f(0)=0$, $f(x+1)=f(x)+1$, and $f(3)=17$.

When we write "Show that there exists a function such that [stuff]," what we're proving is that [stuff] is actually a satisfiable constraint. Knowing that the set of all functions with the given domain/codomain exists is irrelevant: we know that if a function with the properties we want exists then it's in that set, but does it actually exist in the first place?

Uniqueness is similarly nontrivial. Consider:

Show that there is a unique function $f:\mathbb{N}\rightarrow\mathbb{N}$ such that $f(0)=0$ and $f(x+2)=f(x)+2$.

Again, we have a certain property $P$ of functions that we want to understand. The set $\mathbb{N}^\mathbb{N}$ of all functions with the appropriate domain/codomain exists, but that doesn't help answer the question: how many functions $\mathbb{N}\rightarrow\mathbb{N}$ satisfy $P$? (In particular, for existence we need "$\ge 1$" and for uniqueness we need "exactly $1$").