This question — “Do you agree that the Supreme Court’s decision in Baker v. Nelson . . . is  settled law?” — is an easy one to answer “yes,” and make the whole issue go away. Why push back? Because she thinks the case was wrongly decided, or otherwise not entitled to precedential weight. But which one is it?
By way of background, Baker v. Nelson, 409 U.S. 810 (1972) constitutes the Supreme Court’s first (and only) opinion on the direct issue of whether a federal right exists to same-sex marriage. The Court summarily dismissed the case, which came to it on certiorari from the Minnesota Supreme Court. The entire text of the opinion follows:
The appeal is dismissed for want of a substantial federal question.
This represents a jurisdictional dismissal. Jurisdictional dismissals are normally not entitled to precedential value, because the whole point of such rulings is that the central question of the case never comes before the Court. But Baker was decided under the Supreme Court’s mandatory appellate jurisdiction (it became discretionary in 1988, inaugurating the modern certiorari process), and the decision therefore is, in fact, treated as one “on the merits,” making Baker binding precedent for the proposition that there is no federal right to same-sex marriage.
But, the decision is clearly wrong. No matter what you think about whether a federal right to same-sex marriage does exist, whether such right emanates from an amendment to the federal Constitution is a “substantial” federal question. “Substantial” here does not mean “meritorious”: it means “capable of adjudication,” which this question clearly was.
While review of a federal question is barred if the decision of a state high court is supported by “independent and adequate state-law grounds” — in other words, if, regardless of the resolution of the federal question, the state court’s construction of state law resolves the matter — the judgment cannot be reviewed. But the opinion below, 291 Minn. 310 (1971), rests purely on federal decision, and finds no support in state law. The matter was therefore reviewable, and the Supreme Court should have taken the case on the merits. Baker is therefore wrong: technically. But wrong in a way that says nothing about the merits of the underlying case.
I would not expect Grassley to understand the finer points of federal courts jurisprudence. Surely Kagan feels the same way. Her answer, then, may have been a safe way to acknowledge a wonkish point of federal law and signal her disagreement with Grassley’s oversimplification, hopefully without raising his hackles. If it was that, we can conclude that Kagan is an abnormally candid nominee, risking political opposition to correct a technical error without actually signaling anything about her opinion of the case itself. Or she thought it was important to record her disagreement with the case’s principles. That’s something very different… and very exciting.