What are subsets of $A\times B$ called in mathematics?
In logic (algebra and analysis), a binary relation on/over $A$ is a subset of $A^2$. Abstractions of $=$, $\leq$, etc. While using subsets of $A\times B$ extensively in the role of functor, logicians have no term for these objects. I have always called a subset of $A\times B$ a binary relationship, but always explain that this is just my term.
In the computer science ER Model, a binary relation is a subset of $A\times B$, with a subset of $A^2$ being called a recursive binary relation.
Wikipedia currently makes a complete mess of the situation, by defining a binary relation as
In mathematics, a binary relation over two sets $X$ and $Y$ is a set of ordered pairs $(x, y)$ consisting of elements $x\in X$ and $y\in Y$. That is, it is a subset of the Cartesian product $X × Y$.
yet defining an equivalence relation, via reference to the above definition, as
In mathematics, an equivalence relation is a binary relation that is reflexive, symmetric and transitive.
The latter definition is complete nonsense for subsets of $A\times B$, only making sense for subsets of $A^2$. Reflexivity cannot be defined for subsets of $A\times B$.
- What are subsets of $×$ called in mathematics?
- I have never encountered these being called binary relations in mathematics, except for Wikipedia. I thought this might be my algebra-logic bias, but have checked Analysis books, and yes, (binary) relations are as we have known them from primary school, subsets of $A^2$. Abstraction of $=$, $\leq$, etc.
- Unary operations ($o:A\rightarrow A$) are to binary relations ($r\subseteq A^2$) as functions ($f:A\rightarrow B$) are to ??? ($R\subseteq A\times B$)?
I am not looking for a proposed term, nor asking for reference back to Wikipedia, rather, I am asking if there is some term already in widespread use in the literature, and ideally with reference.
Based on answers received so far, an additional question would then be,
- in the context that binary relations are subsets of $A\times B$, what term distinguishes the $A^2$ case? Endomorphic? Recursive?
Some have argued that the Wikipedia definition of equivalence relation is correct as stands. It is not. The definition of an equivalence relation cannot be phrased for subsets of $A\times B$. Reflexivity is impossible to define for subsets of $A\times B$. Further, no one would define an equivalence relation as a subset of $A\times B$ such that..., and then expect the reader to deduce that $A$ must equal $B$. No. That is not how equivalence relations are defined. Currently Wikipedia is wrong.
@TheEmptyFunction has drawn my attention to the fact that Wikipedia currently defines an $n$-ary operation as a function $A_1\times...\times A_n\rightarrow B$, rather than the mathematicians $A^n\rightarrow A$. This definition strips the notion of operation of all its logical/algebraic/analytic meaning. Homomorphisms no longer make sense. The same is true for calling a subset of $A\times B$ a binary relation.
As with all math, you have to be sensitive to the context - different fields are going to use slightly different terms depending on their needs, and will often use terminology from other fields imprecisely via metaphor. You shouldn't assume that every mathematician means the exact same formal definition as every other mathematician by such terms, especially when it comes down to small details like domain/codomains, and distinctions like those between objects like "a boolean valued function on a set" and "a subset of a set" where mathematicians often differ in the precise details (and may even just treat them as interchangeable).
This said, I think all the evidence you need is in your question: subsets of $A\times B$ are most commonly known as relations from $A$ to $B$ (with this "from _ to _" phrasing mostly coming from the category Rel - in some contexts where order doesn't matter you'll hear "between _ and _" instead). There are some other terms that sometimes get used (e.g. "binary predicate" which is subtly different, but interchangeable in most classical contexts - note that predicates commonly take any number of arguments, so "binary" is important here) as well as some qualifiers (e.g. "binary relation" to clarify that two sets are involved in the product). In my experience, this term is very widespread (and Wikipedia would not invent a name for a concept like this!), but not so technical as to have everyone in agreement about the specifics - sort of like how $\mathbb N$ might include $0$ or might not, but it's should never be hard to figure out which convention the author meant or if it matters.
The definition of an equivalence relation does not contradict this definition of a relation. An equivalence relation is a (binary) relation, but the properties of symmetry, transitivity, and reflexivity all implicitly assume that it is a relation from a set to itself. This is in the same way that an endomorphism is a kind of function where the domain and codomain are equal.