Okay, half of you just went away.
My brother is a theatrical lighting designer, and has witnessed enormous technology changes in the field over his career. He mentioned once that a senior designer in an academic setting – which means, even older than us – was distressed about what was not being taught to current students. She brought out a carefully constructed lighting plot. They look a bit like this.
“This is also art.” she said emphatically. My brother smiled uncomfortably and moved on to other subjects. It is easy to see how it comes about. Not only are lighting plots interestingly patterned, but they take craft and knowledge to create, and to a professional, much that is additional – how this light operates on the stage – is implied as well. The skill is not useless, but it has been transformed to such a degree as to become progressively less useful. Lighting design has not become useless, but a particular technique, even a central technique, has apparently become so. (I will check with my brother whether it has entirely lost usefulness or only diminished in usefulness.)
The drawn lighting plot came about to meet a certain need. It does not have intrinsic value.
I have gotten involved (again) in the controversy about language superfamilies, which most historical linguists believe are still speculative, and likely unprovable. A minority believe that there remain enough traces to show common ancestry of the Uralic, Altaic, and Indo-European languages in an older Nostratic superfamily, and some believe there is evidence of a single origin of language and ultimate relatedness of all tongues.
The doubters, the majority, believe that such things might be so, but there is not sufficient evidence to sign off on it, nor is there ever likely to be, because the time-depth is beyond what side-by-side language comparisons can show. Sounds, structure and vocabulary change too quickly – demonstrably so among studied languages in historical timeframes – to give much credence to reconstructions beyond 5,000 years ago. Oops, I mean 10,000 years ago. That has changed in my lifetime.
The insistence is “we do comparisons in a particular way, which we have developed over time and provide protection against leaping to faddish conclusions.” The difficulty is that the supposedly speculative theories point to a relatedness among peoples for which genetic evidence is accumulating. Greenberg proposed three main language groups in the New World (perhaps chief among those which annoy his opponents) correlates with DNA pretty well. Not perfectly, but enough to lend considerable weight to Greenberg’s classification, even though he made many mistakes in understanding and reporting his data.
I know the argument that one can be led astray by attractive but false theories if one does not adhere to the side-by-side comparison rules of determining linguistic relatedness. Yes, errors might occur. But historical linguistics came into being because it looked like it might help answer certain questions about the journey of mankind, and give clues to a history that was not written down.The techniques are not art forms of intrinsic value, even though they might have some use, require skill, and be beloved by practitioners.