Practice: Diagonalization

Recognition · Comparison

Both S = ( 3 0 0 3 ) and Q = ( 3 1 0 3 ) have characteristic polynomial ( λ − 3 ) 2 . Which are diagonalizable?

2 hints available, least help first.

Hint 1: Retrieval cue

Form A − 3 I for each matrix and find the dimension of its kernel.

Hint 2: Concept cue

Diagonalizability needs the eigenspace dimensions to sum to n . Compare each against 2.

Direct application

A matrix A is diagonalized as A = P D P − 1 with

P = ( 1 1 1 − 2 ) , A ( 1 − 2 ) = ( 2 − 4 ) .

What is the entry D 22 , in the second row and second column of D ?

Enter the value. It is checked against the answer and the precision this task asks for.

2 hints available, least help first.

Hint 1: Retrieval cue

Which column of P is ( 1 , − 2 ) T , and which diagonal entry of D shares its index?

Hint 2: Concept cue

( 2 , − 4 ) = λ ( 1 , − 2 ) for what λ ?

Direct application

A matrix satisfies A = P D P − 1 with D = ( 5 0 0 2 ) .

In computing A 5 = P D 5 P − 1 , what is the entry ( D 5 ) 11 ?

Enter the value. It is checked against the answer and the precision this task asks for.

2 hints available, least help first.

Hint 1: Retrieval cue

What is diag ⁡ ( a , b ) k ?

Hint 2: Concept cue

Raise the first diagonal entry to the fifth power.

Direct application

A candidate diagonalization of A = ( 1 2 2 4 ) uses

P = ( 2 1 − 1 2 ) , D = ( 0 0 0 5 ) .

The check is whether A P = P D . Compute the entry in row 2, column 2 of A P .

Enter the value. It is checked against the answer and the precision this task asks for.

2 hints available, least help first.

Hint 1: Retrieval cue

The entry in row i , column j of a product is row i of the first matrix dotted with column j of the second.

Hint 2: Concept cue

Row 2 of A is ( 2 , 4 ) ; column 2 of P is ( 1 , 2 ) .

Construction · Classification · Explanation

(a) For A = ( 1 2 2 4 ) , decide whether A is diagonalizable over R . If it is, give P and D and verify by computing A P and P D .

(b) Using your P and D , compute A 3 via P D 3 P − 1 . Check your answer against A 3 computed directly.

(c) For R = ( 0 − 1 1 0 ) , decide diagonalizability over R and then over C , giving P and D where one exists.

(d) For Q = ( 3 1 0 3 ) , decide diagonalizability over C , and explain why no field makes it diagonalizable.

(e) Parts (c) and (d) both fail over R , for different reasons. State the difference precisely, and say what it means for what can be done about each.

Write your answer, then compare it with the worked solution.

3 hints available, least help first.

Hint 1: Retrieval cue

Check the trace and determinant first; for a 2 × 2 they give the characteristic polynomial immediately.

Hint 2: Concept cue

Verify with A P = P D rather than forming P − 1 . It is cheaper and catches a pairing error at once.

Hint 3: Strategy cue

For part (e), ask of each matrix: does the polynomial have its roots here, and once it does, are there enough independent eigenvectors?

Compare with the worked solution

Comparing does not record a result. Judging your own written answer cannot show that you can do this without help.

(a) A is diagonalizable over R . tr ⁡ A = 5 , det A = 4 − 4 = 0 , so p ( λ ) = λ 2 − 5 λ = λ ( λ − 5 ) : eigenvalues 0 and 5 , distinct, each of algebraic multiplicity 1. Two distinct real eigenvalues in a 2 × 2 , so by the corollary A is diagonalizable. E 0 = ker ⁡ A : from v 1 + 2 v 2 = 0 , take ( 2 , − 1 ) . E 5 : A − 5 I = ( − 4 2 2 − 1 ) , giving 2 v 1 − v 2 = 0 and the vector ( 1 , 2 ) . Taking the eigenvectors in the order ( 2 , − 1 ) then ( 1 , 2 ) :

P = ( 2 1 − 1 2 ) , D = ( 0 0 0 5 ) .

Verification.

A P = ( 1 2 2 4 ) ( 2 1 − 1 2 ) = ( 0 5 0 10 ) ,
P D = ( 2 1 − 1 2 ) ( 0 0 0 5 ) = ( 0 5 0 10 )   ✓

Column 1 says A ( 2 , − 1 ) T = 0 , and column 2 says A ( 1 , 2 ) T = 5 ( 1 , 2 ) T . Invariant check. tr ⁡ D = 0 + 5 = 5 = tr ⁡ A ✓, and the diagonal product 0 × 5 = 0 = det A ✓. (b) A 3 by the factorisation. det P = 2 ( 2 ) − 1 ( − 1 ) = 5 , so P − 1 = 1 5 ( 2 − 1 1 2 ) . D 3 = diag ⁡ ( 0 3 , 5 3 ) = diag ⁡ ( 0 , 125 ) .

P D 3 = ( 2 1 − 1 2 ) ( 0 0 0 125 ) = ( 0 125 0 250 ) ,
A 3 = P D 3 P − 1 = ( 0 125 0 250 ) ⋅ 1 5 ( 2 − 1 1 2 ) = 1 5 ( 125 250 250 500 ) = ( 25 50 50 100 ) .

Direct check. A 2 = ( 5 10 10 20 ) = 5 A , so A 3 = 5 A 2 = 25 A = ( 25 50 50 100 ) ✓. (That A 2 = 5 A is the eigenvalue structure showing through: A acts as 0 on one direction and 5 on the other, so A k = 5 k − 1 A .) (c) R : not over R , yes over C . tr ⁡ R = 0 , det R = 1 , so p ( λ ) = λ 2 + 1 . Over R this has no roots, so R has no real eigenvalues and no real basis of eigenvectors: not diagonalizable over R . Over C the roots are i and − i , distinct, so R is diagonalizable. For λ = i , ( R − i I ) v = 0 gives − i v 1 − v 2 = 0 , so v = ( 1 , − i ) ; conjugating, λ = − i has v = ( 1 , i ) . Thus

P = ( 1 1 − i i ) , D = ( i 0 0 − i ) .

Verification of column 1. R ( 1 , − i ) T = ( 0 ( 1 ) + ( − 1 ) ( − i ) , 1 ( 1 ) + 0 ( − i ) ) T = ( i , 1 ) T , and i ( 1 , − i ) T = ( i , − i 2 ) T = ( i , 1 ) T ✓. (d) Q : not diagonalizable over any field. p ( λ ) = ( λ − 3 ) 2 , which already splits over R , and over C , and over every field containing 3. So there is no field to move to. Q − 3 I = ( 0 1 0 0 ) has rank 1, so dim ⁡ E 3 = 2 − 1 = 1 . The geometric multiplicities sum to 1, short of 2, so no basis of eigenvectors exists and Q is not diagonalizable over C either. The rank of Q − 3 I is computed by row reduction over whatever field contains the entries, and extending the field does not change it: a matrix of rank 1 over R has rank 1 over C . So the shortfall is permanent. (e) The two obstructions. For R the obstruction is that the characteristic polynomial does not split over the field in question. That is a property of the field, not of the matrix: λ 2 + 1 has roots somewhere, and moving to a field containing them produces distinct eigenvalues and a full basis of eigenvectors. The obstruction is removable, and what removes it is enlarging the field. For Q the polynomial splits and the obstruction is that the geometric multiplicity falls short of the algebraic. That is a property of the matrix, expressed as the rank of Q − 3 I , and rank is invariant under field extension. There is nowhere to move to, and the obstruction is permanent. What can be done about each: for R , work over C and the factorisation exists. For Q , accept that no diagonal representative exists in its similarity class and settle for the nearest normal form. The Jordan form, which for Q is Q itself, a single 2 × 2 block with 3 on the diagonal and 1 above it. Reporting "not diagonalizable" without naming which obstruction applies leaves out exactly the part that decides whether anything can be done.

A complete answer does each of these:

  • applies the criterion
  • pairs columns with eigenvalues
  • verifies the factorisation
  • computes a power
  • distinguishes the obstructions
Practice data

Your practice record is stored in this browser only. Clearing it removes every answer and every scheduled review, and cannot be undone.

Results update as you type. Use the up and down arrow keys to move between results, Enter to open one, and Escape to close.

Type to search.

Settings

Appearance

Interface density

Your record

Your progress is stored in this browser and nowhere else: an identifier, the answers you have given, the mastery states and review schedule derived from them, and the lesson you last opened. Clearing it makes you a new learner on this device. It cannot be undone, and it will not affect your appearance or density settings.

Focus timer

Focus--minutes remaining

Phase

Kept in this browser only, and used to label the session in your own history.

Today

Nothing recorded yet. Finish a focus session and it will appear here.

Settings

Focus sessions between long breaks.

Sessions you are aiming for in a day.

Notifications

Your history

Sessions are stored in this browser and nowhere else. They are not evidence and never reach your mastery record.