Definition
Positive and negative parts of a self-adjoint element
The canonical orthogonal positive elements whose difference is a given self-adjoint element.
Definition
Let be a self-adjoint element of a -algebra , and put . The positive part and negative part of are
Both belong to the positive cone of , and they satisfy
They are the unique positive elements with and . This decomposition is also called the Jordan decomposition of .
Functional-calculus description
Under the continuous functional calculus for ,
where and . Hence is the positive part of , and both parts commute with every element that commutes with . Their orthogonality follows pointwise from Murphy, §2.2.
Order and norm consequences
The formulas give , , and
Moreover, is positive exactly when , and is negative exactly when . Applying a -homomorphism commutes with taking positive and negative parts because -homomorphisms commute with continuous functional calculus.
Examples and cautions
For a Hermitian matrix, keeps the positive eigenvalues and replaces the negative ones by zero; replaces each negative eigenvalue by . For a real-valued function, the construction is pointwise. The notation denotes a positive element, not the nonpositive function . The decomposition also should not be confused with the Jordan decomposition of a functional or measure, which requires a separate uniqueness theorem.
References
- Gerard J. Murphy, C-Algebras and Operator Theory*, Academic Press, 1990. Elsevier DOI record. Relevant: §2.2 on positive elements and continuous functional calculus.
- Gert K. Pedersen, C-Algebras and Their Automorphism Groups*, 2nd ed., Academic Press, 2018. Elsevier DOI record. Relevant: §§1.4–1.5 on positivity, order, and functional calculus.