One thing that does seem patently wrong in Mr. Domnarski's argument is the suggestion that "ghostwriting" has somehow reduced the intellectual rigor of the judiciary's body of work. A comparison of today's opinions with those of 100 years ago, when "ghostwriting" was less common, refutes that notion. Back in Chief Justice Marshall's day, the federal caseload was low enough that judges could spend the time necessary to produce comprehensive, reasoned opinions on their own - though one can question whether this actually improved the "honesty and transparency" of the conclusions of such strategically obscurantist opinions as Marbury v. Madison or Johnson v. McIntosh (talk about pounding the table).
But by the turn of the last century, the federal court caseload had reached a level where, notwithstanding the literary merit of opinions drafted by the likes of such luminaries as Holmes, Brandeis, Cardozo, Friendly, or Hand, certainly the bulk of what came out was thinly researched and argued and often simply ideological. Or compare the law office history of Justice Black's First Amendment opinions with the dueling historical analyses of Justices Souter and Rehnquist in Seminole Tribe v. Florida. To suggest that a return to the practices of the 30s or earlier would increase rigor is simply fantasy.