The question
Novenson opens with Voltaire's observation that no people ever took for themselves the name of idolater: idolatry exists only in the mouths of its accusers. Some large ideas are like that, an idea in search of a proponent. The chapter asks whether justification from works of the law is another, and it asks it as a question of attribution. Who, in antiquity, actually held the position Paul attacks?
The tradition has never been short of an answer. Luther named the holders freely: "as many as are in the world that hold not this doctrine are either Turks, Jews, papists, or heretics." Modern criticism narrowed the list to Jews, as when Dunn wrote that it was Paul's fellow Jews in particular who needed to hear that no one is justified by works of the law. The Paul-within-Judaism school, which disagrees with both about nearly everything else, agrees with them that ancient Jews held the view.
The evidence
The three terms do not cluster. Righteousness, works and law never once appear together in the Jewish scriptures; nor in Philo, nor Josephus, nor in the sayings attributed to Yohanan ben Zakkai or Aqiba, nor in the Mishnah, the Tosefta, either Talmud, or any midrash. The first person in antiquity expressly to attest the motif is Paul.
Novenson can point at the workshop. Paul adds "from works of the law" to a psalm that mentions neither works nor law, and, as Francis Watson argues, coins the phrase as the negative image of Habakkuk's "the righteous from trust shall live". If Paul reasons from solution to plight, as Sanders taught the field, he also reasons from solution to counter-solution. The phrase is his own polemical invention, and the only people who hold the position are the people he says hold it.
The obvious objection is 4QMMT, the halakhic letter published in 1994, whose miqsat ma'aseh ha-Torah, "some works of the law", arrives with Genesis 15:6 attached. Interpreters treated it as decisive. Novenson quotes Wright's assessment that it showed justification by works of Torah to have characterised sectarian Judaisms of various sorts, and perhaps mainstream Judaism as well, and argues that the text does not bear the weight. In 4QMMT the works of the law are particular halakhot recommended over against competing halakhot, a 364-day calendar against a 365; and "reckoned to you for righteousness" means that those who adopt them will be proved right in the end. That is vindication in a disputed opinion. Paul means by justification the obliteration of sin and transfer into undying life, which is his own technical usage rather than a current general term for good religion. The same argument is run at equal length against Philo, whose ethical circumcision is the nearest conceptual relative and still not the thing.
The objections
Wright would accept the correction on the fragment and refuse the conclusion. The generalisation from 4QMMT to mainstream Judaism outran the evidence and he would withdraw it. What he would not withdraw is the historical situation: somebody at Antioch made a demand and somebody in Galatia preached a message, and the conflict over the boundary of the covenant was real whoever coined its slogan. Relocating the mint does not shrink the war.
Luther would grant the archival finding and deny that it touches him. No ancient Jew wrote the doctrine in a book; he would say he never met it in a book either. He met it in himself, as the opinio legis, the inborn assumption that doing makes righteous, which needs no publisher and leaves no treatise. He would then press the chapter's own concession: Novenson allows that Paul aims the phrase at Jews twice, at himself before his call and at kinsmen who keep Torah and refuse the crucified Messiah. If so, the coin circulated beyond the mint. This is reconstruction, though the 1535 Galatians lectures make the move directly.
Sanders would answer Luther from the text with a distinction he did not draw himself. Paul holds that no one is justified from works of the law, and holds that Jews know this, because they have the law and the prophets; gentiles do not know it, which is why they keep getting circumcised in the mistaken belief that it will effect their justification. Galatians 2:15 to 16 says so in order: "we who are Jews by nature know that a person is not justified from works of the law." That is why Antioch is a charge of hypocrisy rather than ignorance. Paul does not present justification by works as a characteristically Jewish error; he presents it as a gentile one.
Fredriksen would collect a partial endorsement and enlarge it. Her proposal that Paul's righteousness is the Second Table, justice towards the neighbour, the book judges basically right and then extends: Paul's δικαιοσύνη subsumes all the duties, performed maximally and permanently, what Christine Hayes calls robo-righteousness. She would accept the amendment and press what follows from it, which is that Paul's righteousness-talk is eschatological and impracticable, and describes nothing that either Judaism or Christianity offers as a lived religion.
Paul's own account
The letters show the coinage being made. Paul quotes the psalm that no flesh shall be justified and supplies the qualifier the psalm lacks; he sets against it Habakkuk's righteous who live from trust; and he uses the resulting phrase as a polemical instrument rather than as a description of anyone's programme.
The letters also show where he aims it. In Galatians the target is the counting of foreskins as credit; in Philippians it is his own former standing, "my own righteousness, from the law", which he sets aside not because it was false but because it was mortal. He states in the same breath that he was blameless in that righteousness. Both things are said, and the chapter's reading depends on holding them together: everyday righteousness in the law is real and is not what he means by justification.
On his kinsmen the letters are less tidy than either side would like. Paul writes that Israel, pursuing a law of righteousness, did not attain it, and that they pursued it not from trust but as if from works. The phrase does appear in his mouth about Jews. What the chapter denies is that anyone else in antiquity put it in their own.
The remaining disagreement
The attribution is settled and the significance is not. Novenson has shown that the position Paul attacks is not documented outside his attack on it, and nobody at the table disputes the search. Wright holds that a real conflict existed whoever named it; Luther holds that the position needs no documentation because it is a disposition rather than a doctrine; Sanders holds that Paul locates the error among gentiles and that this settles the matter; Fredriksen holds that the whole vocabulary describes an eschatological condition and so belongs to neither lived religion.
The chapter does not need to choose, because its claim is narrower than the quarrel. It establishes who first used the phrase, and leaves open what work the phrase was doing.
The same chapter as a dialogue sets these positions in the participants' own voices, which is how this was first written.