with whom Justice Stevens joins, dissenting.
I join Justice O’Connor’s dissent. As the dissent says, and as Justice Souter agrees, the Montana Supreme Court did not understand Montana’s statute to have redefined the mental element of deliberate homicide. In my view, however, this circumstance is not simply happenstance or a technical matter that deprives us of the power to uphold that statute. To have read the statute differently — to treat it as if it had redefined the mental element — would produce anomalous results. A statute that makes voluntary intoxication the legal equivalent of purpose or knowledge but only where external circumstances would establish purpose or knowledge in the absence of intoxication, see ante, at 58 (Ginsburg, J., concurring), is a statute that turns guilt or innocence not upon state of mind, but upon irrelevant external circumstances. An intoxicated driver stopped at an intersection who unknowingly accelerated into a pedestrian would likely be found guilty, for a jury unaware of intoxication would likely infer knowledge or purpose. An identically intoxicated driver racing along a highway who unknowingly sideswiped another car would likely be found innocent, for a jury unaware of intoxication would likely infer negligence. Why would a legislature want to write a statute that *80draws such a distinction, upon which a sentence of life imprisonment, or death, may turn? If the legislature wanted to equate voluntary intoxication, knowledge, and purpose, why would it not write a statute that plainly says so, instead of doing so in a roundabout manner that would affect, in dramatically different ways, those whose minds, deeds, and consequences seem identical? I would reserve the question of whether or not such a hypothetical statute might exceed constitutional limits. Cf. McMillan v. Pennsylvania, 477 U. S. 79, 85-86 (1986); Patterson v. New York, 432 U. S. 197, 210 (1977); Mullaney v. Wilbur, 421 U. S. 684, 698-699 (1975).