The thought experiment
The chapter opens by asking how the field would classify an anonymous ancient Jewish writer among whose sayings we made out the following: "do you not heed the law?"; "as many as sinned in the law, through the law they shall be judged"; "everyone who undergoes circumcision is obligated to do the whole law"; "it is not hearers of the law who are right with God, but doers of the law shall be proved right"; "in respect of the law, I am a Pharisee; in respect of righteousness in the law, blameless."
We would class the writer as a halakhic rigorist, or in the older term, a legalist. Every one of those sayings is Paul's. Novenson's explanation for why we cannot see it is that the field reads Paul by what he calls Nixon's rule for legality: when the apostle says it, it is not legalistic, though the same sentence in another mouth patently is.
The word
The history of the term matters to the argument. "Legalism" is a seventeenth-century Puritan coinage, and served as a term of intra-Christian contempt for two hundred years before it was applied to Jews; Bernard Jackson finds the stereotype of Jewish legalism no earlier than 1861. It has always been a term of abuse rather than of analysis, and to invoke it is to pick a fight.
Of Simon Gathercole's five senses of the word, Novenson holds that Paul plainly fits two: the detailed regulation of practice, covering food, sex, veils, idol-meat and lawsuits; and the belief that God will repay each person according to his works, which Paul states without embarrassment at Romans 2:6.
Stripped of the abuse, what remains is what James Kugel called the genius of a religion of laws: that in the small encounters of daily life the Torah sets out the precise form of behaviour God prescribes, so that living becomes a constant act of reaching towards God.
The mechanism
The chapter's engine is a neglected argument of Schweitzer's. For Paul there are precisely three deadly sins which, of themselves, sever a baptised person's union with Christ: πορνεία, communion with demons, and proselyte circumcision. The theory has been largely ignored, and Novenson judges it basically right.
The reason it explains the legalism is that these are exactly the three matters on which Paul issues lengthy and detailed halakhah, and all three are modifications of the body: penetration, ingestion, laceration. The offender is severed, in Collman's image, like a gangrenous limb.
On circumcision Novenson follows Gaventa against the tradition that treats it as adiaphora: Paul does not say his gentiles need not be circumcised, he says they must not. Matthew Thiessen supplies the halakhic rationale. Genesis 17 in Greek stipulates circumcision on the eighth day or not at all, so proselyte circumcision is not merely unnecessary but forbidden. That reading makes sense of an otherwise awkward verse: everyone who undergoes circumcision is obligated to do the whole law of circumcision, which a proselyte, being past his eighth day, cannot keep.
Fredriksen's older point follows. The "law-free gentile mission" is a misnomer. Paul forbids gentile cult, builds a high fence around adultery, and issues controversial halakhah about post-eighth-day circumcision. Whatever this is, it is not law-free.
The objections
Sanders is the figure the chapter argues with most directly, and would begin by conceding the history. He coined "covenantal nomism" because "legalism" had spent four centuries as an insult; nomism and legalism are etymological twins, and only one of them triggered his readers. Novenson's charge is that by ceding the word he missed the deeper criticism, the one Philip Alexander put in 1986: what exactly is so bad about legalism, and is it self-evident that a religion of grace is superior to a religion of law? Neusner put it harder, and Sanders would quote it against himself: with friends like Sanders, Judaism hardly needs enemies. His defence would be chronological. In 1977 the field called Judaism a religion of works-righteousness and meant it as a diagnosis; one fights the battle of one's own day with the weapons that exist, and a later scholar working in a cleared field can afford the word.
Luther would refuse the term while conceding most of the content. He never denied the law its use among the justified; a father instructs his children without making them earn their sonship. The question is not whether Paul gives rules but what the rules are for, and Galatians 5 gives the order: first the severance and the falling from grace, then the fruit, love and joy and peace and patience. First the tree is made good and then it bears; no one made a tree good by hanging fruit on it. He would accept "legislator" and fight "legalist", because the second word says that the ground of standing before God is performance.
Hart would observe that Luther's reply concedes the chapter. He commands; the commands are not the ground of standing: that is Novenson's position without the frightening word. And the word is doing work, because it forces the admission that judgment according to deeds is a structural beam in Paul rather than a stray thread. Hart translated those verses without softening and was accused of making Paul semi-Pelagian.
Wright would offer the resolution he has urged for decades: justification by faith in the present, judgment according to the whole life led in the Spirit at the end, two verdicts in one story, joined by the Spirit's work. On that reading the chapter is useful exactly insofar as it makes the second verdict undeniable, and misleading exactly insofar as "legalist" suggests that the doing generates the belonging. It displays it.
Paul's own account
The letters contain the rigorism the chapter describes, in Paul's own hand and without apology. He writes that God will repay each person according to his works; that doers of the law shall be proved right; that he was, in respect of righteousness in the law, blameless. He legislates in detail about food offered to idols, about sexual conduct, about head coverings, about lawsuits between members.
He also writes the sentences the tradition has built on, and does not treat them as in tension with the others. He does not offer a formula reconciling repayment according to works with justification apart from works of the law, because he does not appear to have experienced the two as a problem requiring a formula.
What the letters do supply is the ordering that Luther points to. In Galatians the severance and the fruit are stated in that sequence. Whether the sequence bears the weight of a doctrine of the ground of standing is the argument between Luther and Novenson, and the text does not settle it.
The remaining disagreement
By the end of the chapter every party agrees that Paul issues law and expects performance. The quarrel is over the relation between the performance and the standing, and Novenson's own view is that this is where the interesting disagreement has always been hiding, obscured by a word nobody would accept about themselves.
Whether "legalist" clarifies that relation or merely restarts the fight is the open question, and Novenson chose the word knowing it would do both.
The same chapter as a dialogue sets these positions in the participants' own voices, which is how this was first written.