As an American I've been confused by this argument against a second referendum. It has for a long time seemed like the best option. If people still want Brexit, they'd vote for it (note that a second vote would only strengthen confidence that Brexit is really what people want). If not, then the will of the people has changed.
It seems like a second referendum would optimally be framed differently than the first, but it seems like the best option as an outsider. The current scenario doesn't seem to be working.
I read it differently, but I don't think the author was saying art has no quality in the way you're interpreting it at the end of your sentence, in the sense of value or anything to offer. I interpreted them to mean that there's no way to objectively determinine that value, so practically speaking there's no objective quality in art. I think there's even a sentence or two where they try to make clear that they're not questioning art's value, just our ability to determine that value in real terms.
I agree that the author's choice of words was probably kind of poor. But I interpreted the article to mean that as it becomes more difficult to determine inherent worth in an area, success becomes more and more dependent on social networking effects per se. If you need a hammer, it has a specific function, and you can quantify that function in various ways. But if you're talking about something where the value is unknown, either because no one knows the ultimate truth, or because it's based on subjective experience or something, you can't quantify it, so you go back to networks.
This author has been making the rounds a bit in the last few years, and I think they've done research on these topics. Their argument is that network effects are very strong, and account for a lot of variability in outcomes that are popularly attributed to individuals (broadly defined).
Something like the SCID is useful as a structured way of measuring or assessing behavior. A binary output is probably not helpful though.
What's interesting is that the people who developed the SCID (and the DSM and its predecessor, the RDC) actually developed a very similar tool in the 60s, but instead of it producing binary diagnoses, it produced scale scores more like test scores. It actually had better statistical properties and provided fairly detailed information. It's always been a mystery why they moved away from that, except that the DSM as a whole moved away from that.
I do research on the topic of the article. I actually initially clicked on it because the title seems so uncannily similar in content to an article I wrote that I thought they might mention something about it. They did not but it's still really salient to me.
The article is sort of strange in that it brings up these issues and then fails to discuss the huge range of research in this area. The p-factor stuff is sort of hot lately but is just one subtopic in a vast area, somewhat controversial, and more of theoretical interest than clinical utility.
Part of the problem is that there are different purposes for classification systems. I think you're right in that categorical diagnoses have all the appeals you mention at least among some subgroups, but those are a little detached from (1) how many clinicians actually think about a lot of problems, and (2) how many researchers think about these things. Putting aside the "syndrome"-favoring clinicians, a lot of clinicians focus more on behavioral patterns, which tend to be very specific, more targetable, and far removed from broad categorical labels. Many researchers are moving away from these categorical labels also, for many reasons, but mostly generally because the patterns you see in behavior don't map onto the categories in the DSM in reality. Think blood pressure or height as analogies for depression or disorganized thinking, rather than Huntington's disease or malaria.
RDoC is important to mention but it's ending up to be about as controversial as the DSM for various reasons. First, it suffers from the same "names by committee fiat" as the DSM, which is problematic when you are dealing with something like neurobehavioral pathology, which can be examined at many levels and from many directions, which is fuzzy and complex, and about which we know very little. Second, despite its noble intents, was kind of developed by neuroscientists with little connection to actual clinical human behavior. I don't mean any offense to them, but when you get together a bunch of researchers, a substantial number of whom study rodent neurobiology, you're going to start losing touch with what clinicians are actually wrestling with in clinical settings. Look for psychosis or subdimensions of psychosis, for example, and it's missing from RDoC. I like to think of RDoC as "revenge of the neuroscientists," people who got tired of trying to connect their research to human public health, and decided to just redefine problems so it's closer to their own research. A similar initiative is HiTOP, which is more phenotypically focused, and uses statistical/quantitative models for classification rather than committee decisions (I'm not saying HiTOP is better, it just is a similar but different approach from RDoC).
I think the issue is not whether or not mental disorders should have names, it's whether or not there are mental disorders per se, or how to best describe mental-neural-behavioral patterns. You need to be able to quantify to study something scientifically, and quantifying implies labeling; the question is how.
It seems like a second referendum would optimally be framed differently than the first, but it seems like the best option as an outsider. The current scenario doesn't seem to be working.
Strange times we all live in.