Recipe
Build the quotient group of an index-2 subgroup
Use when: you're given a subgroup H with [G:H]=2, exactly two cosets, and need both a normality proof and a description of G/H
Goal: a proof that H is normal, plus the complete operation table of the two-element group G/H
Taught in Quotient groups of permutation groups
Quick recipe
-
1
Compute [G:H]=|G|/|H|, or count directly. Confirm it equals 2 before doing anything else. This shortcut only works at index 2, never any larger index.
Example
|S_3|=6 and |A_3|=3, so [S_3:A_3]=6/3=2.
Why: the whole reasoning below only holds because a partition into exactly two blocks leaves no room for gH and Hg to disagree.
-
2
An index-2 subgroup is automatically normal. Its only two left cosets, H and G\setminus H, are also its only two right cosets. That forces gH=Hg for every g.
Example
A_3\trianglelefteq S_3 follows immediately from Step 1's division, with no conjugation computation needed.
Why: coset agreement is equivalent to being normal, and index 2 forces that agreement automatically, with no exceptions.
-
3
G/H has exactly two elements, H itself and O=G\setminus H. H acts as the identity automatically. So the only genuinely new computation is O\cdot O.
Example
In S_3/A_3, representative (1\,2)\in O gives (1\,2)(1\,2)=e\in A_3, so O\cdot O=A_3 directly from one representative.
- G is some S_n and H=A_n — skip the representative computation entirely, using the shortcut below.
- G is some other group with an index-2 subgroup — pick any representative from O, multiply it by itself in G, and see which coset catches the result.
⚡ Shortcut
For G=S_n and H=A_n specifically, two odd permutations always compose to an even one, by chapter 4's parity rule. So O\cdot O=A_n is guaranteed without picking any representative at all. The same shortcut works unchanged from S_3/A_3 up to S_n/A_n for any n.
Why: a two-element set with H as the identity has only one product left undetermined by the group axioms. Finding it completes the whole table.
Common mistakes
- Trying to apply this shortcut to a subgroup whose index isn't exactly 2. The reasoning breaks down completely; it doesn't just get harder.
- Building a full multiplication table by hand anyway. This misses that three of the four entries are forced for free once H is confirmed to be the identity.
- Forgetting to check the index equals 2 first, and assuming any roughly-half-sized split works the same way.
- Confusing H\trianglelefteq G, a subgroup property, with G/H having the same two-row multiplication table as \mathbb{Z}_2. That second fact is about the resulting quotient, not about H itself.