3
$\begingroup$

This is a question about terminology. I hope it is not too silly but I haven't been able to find a clear answer. Basically, is there a more or less standard name for a function from an n-dimensional space into another n-dimensional space that maps each coordinate independently?

In formal terms, let's suppose we have two sets $A$ and $B$ and we can define functions mapping one set to the other $f \colon A \to B$. If, for an integer $n > 1$ we take the Cartesian products $A^n$ and $B^n$, then we can define functions $g \colon A^n \to B^n$. One simple form of such a function would be $$g(a_1, \dots, a_n) = (f_1(a_1), \dots, f_n(a_n)),$$ where each $f_i$ is a function from $A$ to $B$. But of course there will also be more general $g$ functions that don't map each coordinate independently. I'm looking for the English words that express this distinction as in $g$ is a 'whatever' function vs $g$ is a 'non-whatever' function.

I am aware of certain particular cases, like the diagonal vs non-diagonal linear maps in a vector space. But I'm wondering if there are any standard names for this distinction in the general case where we simply have sets, Cartesian products, and functions.

$\endgroup$
1
$\begingroup$

Let $n=2$ for simplicity. Assume you are given two maps $f_i : A_i \to B_i$. Then the function you are referring to is the cartesian product $f_1 \times f_2$.

Here we are thinking of $f_i$ as a subset of $A_i \times B_i$:

Definition A function $f : A\to B$ is subset (also called $f$) of $A\times B$, so that (a) for all $a\in A$, $\{ b\in B : (a, b)\in f\}$ is nonempty, and (b) if $(a, b), (a, b')\in f$, then $b= b'$.

Of course the above implies that it is unambiguous to write, for all $a\in A$, $f(a)$ so that $f(a) \in \{ b\in B : (a, b)\in f\}$. So $(a, f(a) ) \in f$.

From the above definition of functions as subsets, we see that $f_1\times f_2$ is a subset of $(A_1\times B_1) \times (A_2 \times B_2)$ and this latter set is identified as $(A_1 \times A_2) \times (B_1 \times B_2)$.

Claim: $f_1 \times f_2$, treated as a subset of $(A_1 \times A_2) \times (B_1 \times B_2)$, is a function $A_1\times A_2 \to B_1\times B_2$ and $$\tag{1}(f_1\times f_2)(a_1, a_2) = (f_1(a_1), f_2(a_2))$$ for all $a_i \in A_i$.

To prove the claim that $f_1\times f_2$ is a function $A_1\times A_2 \to B_1\times B_2$, let $(a_1, a_2) \in A_1\times A_2$. Then first of all, $$(a_i, f_i(a_i)) \in f_i \subset A_i \times B_i$$ for $i=1,2$. So $$\tag{2}(a_1, a_2, f_1(a_1), f_2(a_2)) \in f_1\times f_2.$$

Secondly, if $$(a_1, a_2, b_1, b_2), (a_1, a_2, b'_1, b'_2) \in f_1\times f_2,$$ then $$(a_i, b_i)\in f_i,\ \ \ (a_i, b'_i)\in f_i.$$ This implies $b_i = b'_i$ since $f_i$ is a function. Thus $(b_1, b_2) = (b'_1, b'_2)$ and so $f_1\times f_2$ defines a function. Now (1) follows from (2).

$\endgroup$
  • $\begingroup$ This does not really answer the question, which is to find a generic name for functions that can be written as a cartesian product of functions. $\endgroup$ – J.-E. Pin Feb 9 '18 at 5:01
  • $\begingroup$ @J.-E.Pin : From the question it seems to me that the questioner did not realize that those functions are just the cartesian product (If they realized I guess they would not ask this question). $\endgroup$ – user99914 Feb 9 '18 at 5:15
  • $\begingroup$ @John, thanks for your answer. I have decided to accept it even if, as J.-E. Pin pointed out above, I was looking for a couple of opposing terms rather than an explanation that the "simple form" I mentioned is also a Cartesian product. But I really appreciate the explanation and it is true that this means that I can make the distinction I was looking for by referring to "n-dimensional functions that are a Cartesian product" vs. those "that are not a Cartesian product". While this is wordier than I wanted, I suppose there is no common simpler way of putting it. $\endgroup$ – Ángel José Riesgo Feb 14 '18 at 9:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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