>Such harms can be reliably detected, with stringent enough criteria.
Mental harms, in many cases, can also be detected by competent professionals; besides that, it is entirely possible for physical harms to heal and for supporting evidence of their infliction be used to convict. Further, many physical harms depend at least partially on the victim's characteristics or situation; a concert pianist is arguably harmed more by someone cutting off his finger than a schoolteacher would be, for instance. Many physical harms that are rightfully legislated against often require the testimony of the victim for the case to succeed. For a wide class of 'mental harms' it is accurate to say that they are indeed physiological responses - from PTSD to lethargy and insomnia. This is in contrast to the caricature that mental harms are necessarily merely 'hurt feelings'.
I also have concerns that the difficulty or the fact of sometimes being nebulous features of mental harms should necessarily rule out such lawmaking. At best, the minimum for proving such harm should at least be set out by the legislators or judiciary, if the standard of evidence is the roadblock to legislation.
It's also worth remembering that we're talking about harms here, not mere hurts. Harms are much harder to fabricate than hurts are.