What is an Homomorphism/Isomorphism "Saying"?

Outside of the technical definitions, what exactly is a homormorphism or an isomorphism "saying"?

For instance, let's we have a group or ring homomorphism $f$, from $A$ to $B$. Does a homomorphism mean that $f$ can send some $a_i$ in $A$ to $b_j$ in $B$, but has no way to "get it back"?

Similarly, if we have a group or ring isomorphism $g$ from $A$ to $B$, does it mean that $g$ can both send and "take back" some $a_i$ in $A$ to/from $b_j$ in $B$?

I'm sorry if this question sounds stupid, but I'm just trying to understand the meaning behind homomorphisms and isomorphisms outside of the technical definitions. I think it will help me tremendously to be able to put them into "dumbed down" definitions. Thank you for your help!

One analogy I use in class says (loosely) that various kinds of morphisms (iso-, homo-) can be thought of as translations from one language to another.

An isomorphism provides a perfect translation in both directions. Words correspond one to one. Anything you can say in one language you can say equally well in the other.

A homomorphism can map many words in one language to the same word in another, effectively creating synonyms. There's an old saying that Eskimo languages have many different words for "snow". The Eskimo-to-English dictionary (homomorphism) would show that (it's not one-to-one, injective), and there would be words in English that didn't come from an Eskimo word (it's not onto, surjective).

Footnote: See THE GREAT ESKIMO VOCABULARY HOAX .

• Interesting analogy maybe you could form a quotient group with the Eskimo Language by identifying those names for the same word as individual equivalence classes then our "quotient language" would be isomorphic to English hahaha Mar 6 '18 at 21:34

I'll consider groups. A homomorphism of groups (or an isomorphism) is simply a map between groups that preserves the structure of a group. What does this mean? Well, the structure we have on a group is an operation $*$. Then, to say that a map $\varphi$ preserves this structure is to say that multiplication in the domain translates to multiplication in the codomain. Formally, if

$$\varphi \colon G \to H$$

is a homomorphism of groups, then $\varphi (xy) = \varphi (x) * \varphi (y)$ for $x, y \in G$. If $\varphi$ is an isomorphism, then the groups (if they're finite) have identical multiplication tables (because $\varphi$ is a bijection and hence every product $h_1 h_2$ where $h_1, h_2 \in H$ corresponds to a product $g_1 g_2$ in our first group), which is what we would want if we say that groups are "equivalent".

One would perhaps want a homomorphism to preserve other structure between groups such as identity or inverses. In other words, we could wish for $varphi$ to map the identity in $G$ to the identity in $H$ (i.e. $\varphi (e_G) = \varphi (e_H)$) or we could want $\varphi$ to map inverses to inverses, meaning $\varphi (x^{-1}) = \varphi (x)^{-1}$. It happens that these properties are actually encapsulated in the definition of a homomorphism.

Isomorphims means that you can identify $A$ with $B$ (whatever you do in $A$) has a correspondent in $B$ and viceversa.

Homomorphisms go only one way, however, if you factor $A$ into equivalence classes (two elements $a,b\in A$ are equivalent iff $f(a)=f(b)$), then there is an isomorphism between the the quotient class and the image $f(A)$. This is called the Fundamental theorem on homomorphisms

• If someone is a complete newbie in group theory,what should be the motivation of studying homomorphism which maps from the same set wrt same operation ? (If there are different sets,one can understand why) Oct 20 '19 at 19:14