Okay, as this is a fairly expansive topic, I'll try to keep this somewhat succinct. What I'm proposing here is that schizophrenia (and, by extension, a significant proportion of other mental conditions currently treated by psychotherapy and pharmacotherapy) is a genuine medical condition according to any reasonable definition of the term, featuring underlying physical causes. I'm certainly not trying to argue that every historical treatment for mental illness has been a good idea, or even that all current treatments are effective/ethical/economic, although I am willing to argue that some treatments do provide at least a modicum of relief for patients, and that broadly speaking the current range of psychiatric interventions are probably the best options available. Please note that the evidence I'm providing represents a quick-and-dirty search on google scholar and pubmed; there are plenty of areas I haven't even touched on, and for all of my citations there are most likely a number of additional papers (some of which may well be better than the ones I've chosen) which examined the same issue.
The underlying physical changes in the brains of schizophrenics are gradually becoming better understood. Patients exhibiting the cluster of symptoms associated with schizophrenia can be identified and put through a wide variety of tests which, when analysed, demonstrate significant structural differences between the brains of schizophrenics and the brains of the "normal" population. This ranges from relatively obvious large-scale changes such as shrinkage of the ventricles to more subtle changes in white matter connectivity and hippocampus volume. In the latter instance, they specifically compared structural abnormalities in first-episode patients with those who have suffered from the condition for a long time and found similar volume reductions, suggesting that the changes do not stem from from extended periods of pharmaceutical treatment, social withdrawal, or any similar factors. In addition to structural differences, there have been differences observed in receptor function, with post-mortem brains of schizophrenics exhibiting lower concentrations of glutamate receptors, among other changes.
Further supporting an underlying physical cause, there is well-established heritability in the condition. Having a close relative with schizophrenia significantly increases the risk of the condition, getting up to 40% or so if one has an identical twin with it. This link is maintained even across people who have been adopted, indicating that it’s not an environmental factor. With more recent advances in genotyping and bioinformatics, it has become possible to identify specific regions of the genome strongly associated with schizophrenia. While there is plenty of support for genetic factors playing a major role in the development of the condition, they aren’t the whole story, with a range of environmental factors also playing a role.
Like many other physical disorders, schizophrenia responds to treatment in a somewhat predictable manner. Obviously there are a range of different sub-types of the condition and a huge range of variables specific to each patient which could confound treatment, but nevertheless there is a variety of somewhat useful treatment options. Pharmacotherapy can be effective at improving quality of life for patients, with modern atypical antipsychotics providing significant improvement in the mental state of patients (measured by the standardised PANSS test) when compared with placebo in randomised trials. If it weren’t a physical condition, one would expect the placebo to perform as well as the active compound. With regards to psychotherapy, there is evidence that cognitive behavioural therapy significantly benefits schizophrenics, beyond the benefit provided by talk therapy in general.
Taken together, this brief roundup of evidence strongly suggests that the condition (or range of conditions) characterised as schizophrenia has an underlying physical cause, and that it is amenable to treatment through conventional medical means. The condition is relatively stable in its incidence across countries and cultures, follows a well-defined set of symptoms, and is not under the control of the patient suffering from it, all of which says to me that it can reasonably considered an illness. More generally, I don’t really see how anyone can seriously argue that mental illness doesn’t exist at all. There are certainly reasonable arguments to be made about the degree of medicalisation of physiological/psychological phenomena, the extent of use of pharmacotherapy, the application of force when dealing with people suspected of posing a danger to themselves or others, etc., but none of this changes the fact that mental illness exists. Indeed, it would be surprising if it didn’t exist - the brain is a physical system like any other organ and, as such, is prone to defects and pathological functions (perhaps even more so than most other organs, given its exquisite complexity). Given that the brain controls behaviour, among other things, it would be downright amazing if people never developed physical malfunctions which manifested as behavioural problems. That’s not to say that all behavioural problems have a physiological underpinning (or even that they’re all necessarily “problems”), and I don’t think that having a mental illness necessarily excuses one from behaving in a reasonable manner, even if doing so is more difficult than for a “normal” person.
Hope that was of some interest, do feel free to ask questions, demand clarification, or whatnot, and I’ll get back to you in due course. Also, if you’re having trouble getting through the paywall on any of the journal sites, I should be able to find and send you PDF copies of any papers you’re particularly interested in reading.