If is an invertible matrix, then
If instead is an invertible matrix, then
Thus in either case and are similar matrices. Explicitly, whenever and with , one has and . Applying the same argument with in place of proves equality of the sets of eigenvalues, including zero.
Yes. The sums of the columns of are the sums of the rows of its transpose . By the previous argument, , so . The determinant is unchanged by transpose, giving
Therefore . The vector need not itself be an eigenvector of in the column-sum case.
Take the nonzero vector . Its th image entry under matrix multiplication is the sum of row :
Thus , exhibiting as an eigenvector with eigenvalue .
We have . Define a function of the columns of by
It is an alternating multilinear map, because both summands are, and . Hence is a normalized determinantal function. By the asserted uniqueness of the determinant, for every . Rearranging yields
This argument also covers without dividing by it.
Write . Since and is a linear map, the multilinearity of the determinant gives, for any column position ,
If for , the corresponding columns also agree, so . Thus is an alternating multilinear map of the columns of , even when is not invertible.
A scalar is an eigenvalue of if and only if there is a nonzero vector with . By the preceding kernel criterion,
The last condition is , exactly the definition of a polynomial root. For a real matrix, this proves the assertion for real eigenvalues; the same proof over the complex numbers gives the usual complex version.
We use the monic polynomial convention for the characteristic polynomial of an matrix:
It has degree , since the only contribution to the coefficient of in the determinant is the product of the diagonal terms . The alternative convention has the same polynomial roots.
For the linear map on the finite-dimensional vector space , the rank-nullity theorem implies that injectivity and surjectivity are equivalent. Consequently is invertible if and only if .
Using the permitted equivalence between invertibility and a nonzero determinant, we obtain
If with , then if and only if , so their kernels agree. Conversely, suppose both nonzero linear functionals have kernel . Choose with , by scaling a vector on which is nonzero. For each ,
Therefore . Moreover , since . Thus , with a nonzero scalar.
Every nonzero linear functional is surjective, so its kernel has dimension by the rank-nullity theorem. Conversely, every hyperplane is such a kernel: extend a basis of by one vector , and assign value zero on that basis and value one on . Hence the hyperplanes containing are exactly the kernels of the nonzero linear functionals counted above. Each kernel occurs for exactly of them, its distinct nonzero scalar multiples. The required number is
Put . Every two-dimensional vector subspace containing is for some . There are choices for in . Within any one such , exactly choices of yield that same : each forms a linearly independent pair with , and that pair is a basis of .
These sets of choices partition , because determines uniquely. Thus the number is
Equivalently these vector subspaces correspond to the one-dimensional vector subspaces of the quotient vector space . For there are none and the formula gives zero.
Use basis extension to choose a basis of the vector space over the finite field . The condition fixes the first value of a linear functional; its remaining values are arbitrary. There are such linear maps, exactly one of which is zero. Hence the required number is
For this is zero, as it should be: vanishing at a basis vector then forces the entire linear functional to vanish.
Every has a unique basis expansion . Define the linear functional
Uniqueness of the coordinates makes this definition well defined. The coordinates of a sum or a scalar multiple are the corresponding sums or scalar multiples of the coordinates, so this formula defines a linear map. It has . Conversely, any linear map with those values must satisfy the same formula by linearity, proving uniqueness.
If the finite field has elements, each of the values can be chosen independently in ways. The number of linear maps is therefore , including the zero linear map.
A map between vector spaces over the same field is a linear map precisely when it preserves vector addition and scalar multiplication:
Equivalently, for all and .
A vector space over a field is a set with vector addition and scalar multiplication satisfying the following vector space axioms. Under vector addition, is an abelian group: addition is associative and commutative, there is a zero vector , and each has an additive inverse . For all and ,
The operations are defined for every indicated pair and take their values in .
Write , where
The binomial theorem for commuting matrices applies because . All terms of degree at least three vanish, so for every positive integer ,
The same expression directly gives when , so no small exponent is excluded.
If is a strictly upper triangular matrix, then unless . We prove by mathematical induction that
The assertion for is the definition. In the matrix multiplication formula
a nonzero summand would require and , hence . This proves the induction step. Since for all entries, . In particular, every strictly upper triangular matrix is a nilpotent matrix.
Perform the elementary row operations and , then , , and finally , :
Consequently the reduced row echelon form is . Applying the same operations to the augmented column gives
Thus every admits exactly one solution. This follows because the elementary row operations preserve the solution set of the augmented linear system, and the reduced coefficient matrix is the identity matrix.
For an matrix, the corresponding elementary matrices are , because the operation is implemented by left matrix multiplication. Apply the desired elementary row operation to the identity matrix . Writing for the matrix unit with its only nonzero entry, , in position gives:
Indeed, these matrix products interchange rows , scale row , and add times row to row , respectively. The displayed inverses undo precisely those operations. Algebraically, when , so ; the other two inverse identities follow immediately from their row actions. Thus every elementary matrix is an invertible matrix.
The three elementary row operations on an real matrix are:
  • Interchange two distinct rows: .
  • Multiply one row by a nonzero real number: , with .
  • Add a scalar multiple of one row to a different row: , with .
Each operation changes whole rows, so it acts on every column by the same rule. Each can be undone by an elementary row operation.
Eliminate the two pure second derivatives by taking to be the roots of . They are distinct real numbers exactly when
Thus the required function is . The mixed coefficient becomes , so the transformed equation is . Integrating once in each coordinate gives
with arbitrary sufficiently smooth functions . This is the characteristic form of a strictly hyperbolic partial differential equation.

There are unlisted articles, also show them or only show them.