## Group Theory III.1 (Subgroups: Definition)

The astute reader might have noticed in our list of examples of groups that some groups are contained in others. E.g. $A_n \subset S_n$, $SL_n(\mathbf{R}) \subset GL_n(\mathbf{R})$, the group of symmetries of the icosahedron is contained in S12 etc. Furthermore, this subset inherits the group operation from the big group.

Warning : be careful with the cyclic groups Cn. For example, even though we have $C_5 \subset C_{10}$ set-theoretically (recall that the former is {0, …, 4} while the latter is {0, …, 9}), the group operation is inconsistent. E.g. 3 + 4 = 2 in C5 but 3 + 4 = 7 in C10. Nevertheless, it is possible to “embed” C5 into C10, as we shall see later.

Just as sets have subsets, it is only natural to define the following: a subgroup of G is a subset H which forms a group under the * operation inherited from G. To ensure that (H, *) forms a group as well, we need to go back to check those 3 axioms of a group. Associativity clearly comes for free, and the only remaining conditions to check are:

• $e \in H$;
• if $h, h' \in H$, then the product $hh' \in H$;
• if $h \in H$, then the inverse $h^{-1} \in H$.

In fact, it suffices to reduce to the following conditions.

A subset H of G is a subgroup if and only if H is non-empty, and for any $h, h' \in H$, we have $h h'^{-1} \in H$.

Clearly the boxed condition is necessary for H to be a subgroup. The proof for sufficiency is easy and left as an exercise. Note that non-emptiness of H is important, since without it, the empty set would also become a subgroup of G, whereas our earlier definition forces H to contain e at least. This reason may appear cosmetic, but disallowing empty sets as subgroups turns out to be wise in hindsight when we encounter Lagrange’s theorem in future.

This entry was posted in Group Theory Notes and tagged , , , , , . Bookmark the permalink.