Kendell point (1). Szasz's position is not that diseases must be diagnosed by looking for chemical or structural abnormalities. Rather, such abnormalities are a criterion for having the disease. It may not be possible to find those abnormalities until after the person dies.
Point (2): the existence of such abnormalities is required for a person to have a disease, but not all abnormalities are considered diseases. This is as irrelevant as saying that not all ice cream contains water because some water is not ice cream.
Point (3) is also irrelevant. Szasz's claim is that mental illness is behaviour that a person undertakes for a reason, albeit a reason that people don't like. That idea that such behaviour will turn out to be caused by a disease is like looking for a hardware fault in your iPhone because you have installed an app your dislike.
I addressed (4) earlier. For my reply to point (5) see my reply to point (3).
As for Shorter, note that his statement that mental illnesses are brain diseases is incompatible with Kendell's points since Kendell is attacking the idea that lesions are necessary for disease while Short alleges that such lesions exist. So which position does the poster actually hold? I have already addressed points (1) and (3) by explaining that mental illness is not biological, so let's skip to point (2). Szasz's point was that if a person was found to have a brain disease, brain doctors would treat him with his consent. Psychiatrists don't do this and so there is some difference between psychiatrists and doctors who deal with brain diseases.
Clarke. Point (1): if a person claims to be mentally ill then he is claiming to have an illness when he doesn't. It is better to face up to this squarely and discuss why the person in question is acting the way he acts. If a person claims he is not a moral agent and refuses to face up to his moral disagreements with others, you're not doing him any favours by agreeing.
Point (2): people get labelled as mentally ill because they have a moral problem, that is, behaviour that is regarded as undesirable by somebody who may or may not be the patient. It is true that such problems would not vanish if we stopped calling them mental illness but Szasz didn't claim that they would. Rather, he claimed that we should face up to those problems squarely and stop using quasi-medical excuses to coerce the kind of people commonly labelled mentally ill. If you want to coerce somebody you should have to make an explicit argument that his behaviour should be treated as a crime.
As for point (3), abandoning treatments that the public dislikes is entirely compatible with picking some other set of treatments that the patient dislikes and imposing it on him.