[DanBC]

> and that the qualitative distinction is used to justify all of these.

Yes - does it pose a risk of harm to yourself or other people; does it interfere with your day to day life? These quality statements are used as part of the process of assessing whether someone needs or wants a treatment, and they should be common across all forms of mental illness. (EG people with auditory hallucinations often go unmedicated because they can live with their voices.)

> You can be compassionate and understanding towards someone's problems without categorizing those problems as a medical illness.

Compassion does not treat depression, although it's important part of preventing depression. Talking therapies like CBT are pretty structured, and the evidence says they seem to work. We know the counseling generally doesn't, and can be harmful. And also, if a person needs treatment then they might need money to pay for that treatment and protected time off work to get treatment. Calling dysfunctional forms of sadness "depression" is partly a bureaucratic measure we take to fund treatment and protect people from losing a job.

Strongly agree with your last sentence. A few people on HN equate mental illness with violence but mental illness does not predict violent behaviour (drug addiction; or previous violence are much better predictors, and if you have a combination of either / both of those and a mental illness that's a better predictor, but merely mental illness itself isn't predictive).