For that matter, why is it that these non-linguists’ linguistics papers get more general attention than those by linguists?
This phenomenon isn’t limited to linguistics. Bad papers often get good press across a variety of disciplines. I know of two reasons:
1) Although this reason doesn’t apply in this case, quacks and pseudoscientists often use the press to get their ideas across. They can’t stand up to the rigor imposed by the academy, so they take the debate to a more forgiving forum, journalism.
2) University public relations offices often promote articles that have a good hook. Work that makes sweeping or big claims is more attractive to the press. (If it seems that shaky linguistic articles are more common than in other disciplines, it may be because of this. Stories about language are perceived to have wide appeal.) And often what is actually a decent paper will be blown out of proportion and misreported by the press.
Also contributing to the perception that this happens more often in linguistics is selection bias. Those who read this site are also more likely to read more linguistics articles. Since we see more of them overall, we’ll also see more bad ones. If this were a astrophysics site, the readers would probably think astrophysics had a disproportionate number of bad articles receiving press attention.
In this particular case, contributing to the press interest is that the article is in PNAS, a rather prestigious journal. That adds to the fuel for both the university PR office and the journalists reporting on it. The fact that PNAS doesn’t (as far as I know) often publish linguistics articles probably contributed to the problem of this being a bad paper (They don’t have editors versed in the subject and may not know the proper peer reviewers to contact.)
Note also that this article did not follow PNAS’s standard submission procedure. It had a “pre-arranged editor.” A PE is used by PNAS and some other journals when, “an article falls into an area without broad representation in the Academy, or for research that may be considered counter to a prevailing view or too far ahead of its time to receive a fair hearing.” Basically, the authors get to pick their editor. It doesn’t bypass peer review, but it can grease the skids. Depending on the authors’ influence over the editor, they may be able to pick their reviewers and a shaky paper that might be rejected by a more neutral editor is more likely to be accepted by a PE. Anytime you see an article with a PE, that’s a flag that the work is on the fringe—not necessarily bad, but a warning sign that it is more likely to be so. Journalists, of course, don’t pick up on this nuance.