Opinion of the Court
that §2255(e) does not permit recourse to §2241 in these circumstances).
We now hold that the saving clause does not authorize such an end-run around AEDPA. In §2255(h), Congress enumerated two—and only two—conditions in which a second or successive §2255 motion may proceed. Because §2255 is the ordinary vehicle for a collateral attack on a federal sentence, the straightforward negative inference from §2255(h) is that a second or successive collateral attack on a federal sentence is not authorized unless one of those two conditions is satisfied. See Jennings v. Rodriguez, 583 U. S. ___, ___ (2018) (slip op., at 16) (“ ‘The expression of one thing implies the exclusion of others’ ” (quoting A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 107 (2012))). Even more directly, §2255(h)(2)’s authorization of a successive collateral attack based on new rules “of constitutional law” implies that Congress did not authorize successive collateral attacks based on new rules of nonconstitutional law. Had Congress wished to omit the word “constitutional,” it easily could have done so.
The saving clause does not undermine this strong negative inference. Basic principles of statutory interpretation require that we construe the saving clause and §2255(h) in harmony, not set them at cross-purposes. See, e.g., United States v. Fausto, 484 U. S. 439, 453 (1988); Bend v. Hoyt, 13 Pet. 263, 272 (1839) (Story, J.). That task is not difficult given the distinct concerns of the two provisions. Subsection (h) presumes—as part of its background—that federal prisoners’ collateral attacks on their sentences are governed by §2255, and it proceeds to specify when a second or successive collateral attack is permitted. The saving clause has nothing to say about that question. Rather, like subsection (e) generally, it addresses the antecedent question of the relationship between §§2241 and 2255.
After AEDPA, as before it, the saving clause preserves recourse to §2241 in cases where unusual circumstances
