Artin's induction theorem: Difference between revisions
| Line 49: | Line 49: | ||
| 3 || To show what we need to prove, it suffices to show that every class function is a <math>\mathbb{C}</math>-linear combination of characters induced from members of <math>X</math>. || || || (1), (2) || <toggledisplay>By step (1), all characters induced from characters of <math>X</math> are characters, hence they are in the <math>\mathbb{Q}</math>-span of the space of all characters. By step (2), if we show that every class function is in the <math>\mathbb{C}</math>-span of these, we would show that every character is in the <math>\mathbb{Q}</math>-span of these. This would show what we need to prove.</toggledisplay> | | 3 || To show what we need to prove, it suffices to show that every class function is a <math>\mathbb{C}</math>-linear combination of characters induced from members of <math>X</math>. || || || (1), (2) || <toggledisplay>By step (1), all characters induced from characters of <math>X</math> are characters, hence they are in the <math>\mathbb{Q}</math>-span of the space of all characters. By step (2), if we show that every class function is in the <math>\mathbb{C}</math>-span of these, we would show that every character is in the <math>\mathbb{Q}</math>-span of these. This would show what we need to prove.</toggledisplay> | ||
|- | |- | ||
| 4 || To show what we need to prove, it suffices to show that every class function is a <math>\mathbb{C}</math>-linear combination of class functions induced from members of <math>X</math>. || || || (3) || <toggledisplay>This is the same as (3), except that we've replaced ''characters induced'' by ''class functions induced''. This replacement is valid because class functions induced are in turn linear combinations of characters induced.</toggledisplay> | | 4 || To show what we need to prove, it suffices to show that every class function is a <math>\mathbb{C}</math>-linear combination of class functions induced from members of <math>X</math>. || || || (3) || <toggledisplay>This is the same as Step (3), except that we've replaced ''characters induced'' by ''class functions induced''. This replacement is valid because class functions induced are in turn linear combinations of characters induced.</toggledisplay> | ||
|- | |- | ||
| 5 || Let <math>V</math> be the <math>\mathbb{C}</math>-span of all the class functions of <math>G</math>. <math>V</math> is also the space of class functions of <math>G</math>. Let <math>W</math> be the span of all class functions induced from characters of members of <math>X</math>. In other words: <math>\! W = \langle \operatorname{Ind}_H^G \psi \mid H \in X \rangle</math>. || || || || | | 5 || Let <math>V</math> be the <math>\mathbb{C}</math>-span of all the class functions of <math>G</math>. <math>V</math> is also the space of class functions of <math>G</math>. Let <math>W</math> be the span of all class functions induced from characters of members of <math>X</math>. In other words: <math>\! W = \langle \operatorname{Ind}_H^G \psi \mid H \in X \rangle</math>. || || || || | ||
Revision as of 00:34, 24 February 2011
This article states an induction theorem: a result relating the linear characters and linear representations of a group with the characters/representations induced from the linear characters/representations of subgroups
View a complete list of induction theorems
This fact is related to: linear representation theory
View other facts related to linear representation theory | View terms related to linear representation theory
Statement
Let be a finite group and a family of subgroups of . Then the following are equivalent:
- The union of conjugates of elements of cover the whole of
- Every character of over is a rational linear combination of characters induced from characters of members of
Further, these equivalent conditions hold if is the collection of all cyclic subgroups of .
Related facts
Facts used
Proof
This proof uses a tabular format for presentation. Provide feedback on tabular proof formats in a survey (opens in new window/tab) | Learn more about tabular proof formats|View all pages on facts with proofs in tabular format
Proof idea
With a little linear algebra, we can show that if a character of is a complex linear combination of characters induced from members of , all the coefficients are in fact rational. Thus, the problem reduces to showing that the class functions induced from members of span the space of all class functions on .
This proof follows by using Frobenius reciprocity, and the fact that the only class function on which restricts to the zero function on every member of , is the zero function on the whole of .
Proof details: (1) implies (2)
Given: A finite group with a family of subgroups such that the union of conjugates of members of is .
To prove: Every character of is a rational linear combination of characters induced from characters of members of .
Proof:
| Step no. | Assertion/construction | Given data used | Facts used | Previous steps used | Explanation |
|---|---|---|---|---|---|
| 1 | The (rational) vector space span of the irreducible characters contain all the characters. | [SHOW MORE] | |||
| 2 | A subset of the -span of characters with the property that its -span contains all class functions also has the property that its -span contains all characters. | Something about linear algebra; fields are linearly closed | (1) | ||
| 3 | To show what we need to prove, it suffices to show that every class function is a -linear combination of characters induced from members of . | (1), (2) | [SHOW MORE] | ||
| 4 | To show what we need to prove, it suffices to show that every class function is a -linear combination of class functions induced from members of . | (3) | [SHOW MORE] | ||
| 5 | Let be the -span of all the class functions of . is also the space of class functions of . Let be the span of all class functions induced from characters of members of . In other words: . | ||||
| 6 | Let be the orthogonal complement to in with respect to the inner product of class functions: | (5) | |||
| 7 | Suppose . Then, for any and any class function of , we have: . | Fact (1) (Frobenius reciprocity) | |||
| 8 | (5), (6) | By definition, , so its inner product with anything in is zero. | |||
| 9 | (7), (8) | By (8), the left side of the expression in (7) is 0. Hence, so is the right side. | |||
| 10 | is orthogonal to every class function of . | (9) | [SHOW MORE] | ||
| 11 | for every | The inner product is an inner product. | (10) | By Step (10), and the fact that we are dealing with an inner product, is the zero function. Further, since was an arbitrary member of , this result holds for all .</toggledisplay> | |
| 12 | for every element of in the union of conjugates of members of . | (6), (7): is a class function; (11) | [SHOW MORE] | ||
| 13 | is identically the zero function | is the union of all conjugates of members of | (12) | [SHOW MORE] | |
| 14 | (13) and (7) | [SHOW MORE] | |||
| 15 | , i.e., every class function on is in the -span of those induced from class functions on members of . | (14), (5), and (6) | [SHOW MORE] | ||
| 16 | We are done. | (4) and (15) | Follows directly. |
Proof details: (2) implies (1)
The proof here is essentially the same; it uses Frobenius reciprocity to reason in the opposite direction.
Given: A finite group with a family of subgroups such that every character of is a rational linear combination of characters induced from .
To prove: is the union of conjugates of members of .
Proof: Since every character of is a rational linear combination of the characters induced from , it is in particular true that the -span of class functions induced from class functions of , is the whole space of class functions on .
Taking the usual inner product of class functions:
.
Now, suppose is a class function of that takes the value on the union of conjugates of and is outside. Then we have that for every and every class function of :
.
By Frobenius reciprocity, we get:
.
In other words, is orthogonal to all the class functions induced from members of . By assumption, is thus orthogonal to every class functino of , forcing . By te way we defined , we obtain that the union of conjugates of must be the whole group .
Proof details for the additional observation
The additional observation follows from fact (2): every group is a union of cyclic subgroups.