Let’s take a moment to consider Mueller himself. The cut of his jib is likable, and the trad Brooks Brothers vibe of his wardrobe is a perfect complement to his job title. But it’s hard to avoid the suspicion that he’s playing a political game at this point. To be fair, I’m vulnerable to confirmation bias of my own in this assessment, since about a year ago I suggested that Mueller was going to drag out his investigation until 2019, when Democrats were likely to be back in charge of the House, and seeing a prediction play out can lead to unwarranted certitude. But the reports we’re starting to see suggest a man who’s fallen prey to the same state of mind that warped Ken Starr—namely disgust over the people you’re investigating and a desire to justify the sunk capital.
Our justice system gives prosecutors a frightening amount of power as it is, and nothing tempts misuse of it quite like the belief in a narrative in the face of a disappointing witness. George Papadopoulos has told people he pleaded guilty to perjury because Mueller was threatening to prosecute him as an unregistered agent of Israel. Jerome Corsi insists that Mueller was (and is) threatening him with a raft of indictments unless he signed on to an untrue story of how he came to believe (or know) that WikiLeaks had hacked the e-mails of John Podesta.
We don’t know why Mueller feels Manafort is lying to prosecutors, but we do know that Mueller is either asking him about things that have little to do with Manfort’s guilty plea, i.e. acting as an unregistered agent of Ukraine, or else asking him things that have little to do with the original purpose of Mueller’s investigation, i.e. Russian conspiracy. The former would mean Mueller was tempting Manafort, deliberately or not, to make up a story to please federal prosecutors (“not just sing,” but “also compose,” as a judge on the case warned last May). The latter would mean Mueller was getting out on tangents and allowing his investigation, Starr-style, to lapse into a shape-shifting creature with few self-imposed limits.