There are a handful of results establishing conditions on the measurable real-valued function $V(x)$ under which the operator:

$$-\Delta + V(x)$$

Is (essentially) self-adjoint on $L^2(\mathbb{R}^n)$. My question is: Is the operator $V(x)$ by itself always self-adjoint? My instinct is no: Something sufficiently ugly like:

$$V(x) = \frac{1}{\sin(1/x)}$$

With infinitely many singularities near the origin ought to do it. And, if it is not true that multiplication by all measurable functions is self-adjoint, is there a relatively simple necessary and sufficient condition? Being integrable on every compact set seems sufficient but perhaps too strong.


  • $\begingroup$ I'm fairly sure that multiplication by $V$ will be self-adjoint if $V\in L^2_\text{loc}$. EDIT: This source seems to corroborate: books.google.com/… $\endgroup$ – Alex Youcis Feb 21 '13 at 8:51
  • $\begingroup$ It says "In conclusion, we note that the condition of local square-integrability can be weakened to the condition that V(x) be finite almost everywhere and measurable".. that seems like basically not a restriction at all. $\endgroup$ – Alex Zorn Feb 21 '13 at 17:19

Weidmann: Linear Operators in Hilbert spaces

Example 1. (page 51) is where the multiplication operator is defined. Example 2 on page 91 shows that for an arbitrary measurable $V$ you always get a (densely defined) selfadjoint (but not necessarily bounded) operator. If you want it a bounded operator, you need it essentially bounded.

Summarizing, there is no restriction at all.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.