# Is there any difference between mapping and function?

I wonder if there is any difference between mapping and a function. Somebody told me that the only difference is that mapping can be from any set to any set, but function must be from $\mathbb R$ to $\mathbb R$. But I am not ok with this answer. I need a simple way to explain the differences between mapping and function to a lay man together with some illustration (if possible).

Thanks for any help.

-
I added the terminology tag. – Alex Becker Jan 2 '12 at 7:00

I'm afraid the person who told you that was wrong. There is no difference between a mapping and a function, they are just different terms used for the same mathematical object. Generally, I say "mapping" when I want to emphasize that what I am talking about pairs elements in one set with elements in another, and "function" when I want to emphasize that the thing I am talking about takes input and returns output. But that's just a personal preference, and there is no convention I'm aware of.

-
 Mapping from any set to $\mathbb C$ or $\mathbb R$ is a function. What can you say about it? – Hassan Muhammad Jan 2 '12 at 16:41

Although in most cases the words function and mapping can be used interchangeably, in several parts of mathematics differences in emphasis, especially in analysis and differential geometry. I can think of two.

First, especially in differential geometry, "mapping" is the universal word, and the word "function" is used for mappings that map to $\mathbb{R}$ or $\mathbb{C}$. Thus a mapping which maps to $\mathbb{R}^n$ for instance would not be called a function. This convention is not always adhered to (you might occasionally read about "vector-valued functions"), but this is the usual interpretation.

Second, especially in analysis, it is not uncommon to call members of $L^p$ "functions", even though they are actually equivalence classes of mappings. Again the idea is that functions should assign numbers to some objects (e.g. points in some space) in a suitable sense. Thus functions are thought of being objects studied in analysis, whereas "mapping" is thought of being a term from set theory.

-
That is why we called a map to $\mathbb R$ or $\mathbb C$ FUNCTIONAL in functional analysis. – Hassan Muhammad Jan 2 '12 at 16:38
@Hassan : Not exactly. A functional usually refers to a mapping with a function space as its domain and $\Bbb R$ or $\Bbb C$ as its range. – Patrick Da Silva Dec 27 '12 at 11:44
@PatrickDaSilva: actually, in context of analysis, I've usually seen a functional to mean a linear function into the scalar field, often implied to be continuous. The meaning you refer to, I've found in context of differential equations and more closely related things than general analysis. – tomasz Dec 27 '12 at 11:57

Not that much difference in the long run. When I use the word function I generally mean that a point maps to a single point. So, if a point might map to several points, I am not going to use that word, more likely mapping or transformation. In a recent article I had one of these, each point went to several points, and each point in the image probably had several pre-images, so I emphasized, in a traditional phrase, that the mapping was "many-to-many." Now, both primage and image were equivalence classes under a weaker equivalence, so the mapping did induce a function from "genus" to "genus," but was not well-defined on the level of isometry classes of quadratic forms.

Anyway, if a point goes to only a single point, you are allowed to call it a function.

EDIT: I see, you have finished college and are just asking about preferences. I've got to think about popularity in English... Function is used for $\mathbb C \mapsto \mathbb C,$ also maps from any smooth manifold to the reals. I might use function for almost any map into $\mathbb R^n$ from almost anything, but would be less likely to use function for a mapping between two other manifolds. Various kinds of mappings in algebra are unlikely to be called function.

-
If I understand you, in a mapping one element from the domain may be map to two or more element in the co-domain? – Hassan Muhammad Jan 2 '12 at 7:08
@HassanMuhammad yes, that is absolutely correct, it does in my paper. But then, sometimes in mathematics, we switch to equivalence classes, so what we are talking about also changes. Suppose you have a group $G$ with a normal subgroup $H.$ There is a mapping taking any element $a \in G$ to all the elements $ah, \; \mbox{for} \; h \; \in H.$ On this level it is one-to-many. But then we define the quotient group $G/H,$ and suddenly we have the function $a \mapsto aH.$ – Will Jagy Jan 2 '12 at 7:28
Most people use the two terms interchaningly, in particular the require mappings to map each point to a single points. For "mappings" that point one point to several, one usually speaks of corespondences, relations, or multifunctions. – Michael Greinecker Jan 2 '12 at 13:58

By Nii: To my best understanding, mapping is just a process of matching elements of one set to elements of another set. Mapping is not a function unless some conditions are defined. Thus every mapping is a retation but not necessary a function.

-