Is there any evidence against the following hypothesis:
What we call mental illnesses are (with possible exception of Schizophrenia and Bipolar disorder), not qualitatively different from ordinary experience. But in order to provide treatment (e.g. various kinds of therapy) it is useful to assign labels to people. But this usefulness doesn't make mental illness as "real as cancer", in the sense that it is a discrete category that we can definitively assign people too.
One reason people prefer a diagnosis is that they prefer a label (even if it does come with some social stigma) to the greater social stigma associated with the same symptoms absent a "mental illness". Assigning such a label allows people (both the individual with the mental illness, and people around them) to view their symptoms in a different and more compassionate light. Most importantly, it allows people to think rationally about the best course of action for the person.
My claim is that if society could be more compassionate, without needing to first assign a label in order to remove "blame" for people's symptoms, we wouldn't need the category of mental illness as much. E.g. instead of saying "I thought I was just a loser, but then I found out I suffered from depression", a person might say "I thought I was a loser, but then I realized that what society calls a loser is a person with certain characteristics, thought patterns, and patterns of behavior, and I decided that instead of judging myself for having these things, I would find the best approach to changing them".
There is one more corollary to this, which is that "professional" help, in the form of drugs or therapy, may not be the only or best approach for such people.