Lewin is graceful, and my initial reaction is that I like that a lot. He begins by refuting Lawrence Kramer, who reads, in Wagner's Tristan and Isolde, the dissolution of gender boundaries. In the intensity of their desire for one another, Tristan and Isolde somehow become genderless, mirror images of their fundamental humanity. (As per my MO, I would read this against Irigaray, specifically her essay on masculine subjectivity and the metaphysical longing for the mirror instead of different. Kramer himself anticipates this argument by centering the mirror in his metaphor.) Lewin avoids philosophical distinctions, and instead simply points out that Isolde is always a woman's voice, and can only be sung by a woman. (Where do trouser roles fit into this?)
He slides from here into McClary's chapter on madness, women, and opera, in which she traces a genealogy of mad women from Monteverdi to Schoenberg. Lewin agrees with McClary on a few points, but is unwilling to apply the "madwoman" label to everyone on McClary's list. Isolde, for one, is not mad, says Lewin, but rather transendentally enamored. McClary sees madness differently than does Lewin, to be sure. For Lewin, madness is or is not. McClary, on the other hand, relies on Foucault for her understanding of madness, and therefore is more concerned with how the ostensibly mad person is received and constrained by society: Isolde may well be mad by this definition, depending on the audience's expectations. However, the fact that Lewin takes issue at all suggests that, at least in contemporary circles, McClary may be a little off her mark here.
Rather than dwell on the point, Lewin looks toward a slightly later piece from McClary's list: Schoenberg's Erwartung (listen to this piece, and follow the libretto... goooood stuff). But again, so gracefully, Lewin slides from a work McClary has already read to a slightly earlier Schoenberg piece: the second string quartet, which uses a female voice in two of the movements. [pardon my parallel syntax] Again, Lewin argues that the gender of the singer is essential (though he also problematizes vocal gender, suggesting that register and timbre may well be entirely constructed, a position with which McClary would probably agree). Due to structural features of the music (in all three cases, not simply in the second quartet), a man's voice simply would not fill the shoes.
What ties these roles together, aside from gender, is control. Lewin argues, and supports his arguments with brief analyses from the second quartet, that the female voice is what governs the flow of the music, permitting specific harmonic paths and contradicting and curtailing others. These paths are, to be sure, licentious, but rather than marking this as a negative or transgressive trait, Lewin tacitly remembers that this is the spirit of modernism.
Jean-Phillipe Rameau's Traité de l'harmonie is resurrected briefly to serve as a counter-model. In 1722 (roughly 200 years before Schoenberg's piece), Rameau claimed that the bass voice was not only the exclusive provence of men (not a shocking claim) but that it was also the sole generator of the higher voices. The bass governed harmony (it was fundamental), and harmony governed melody. Women's voices were twice subordinate. The role Rameau's treatise plays in the history of tonal music suggests that the casting aside of women was a common-practice practice, and that modernism--Wagner included, here--resurrects the female voice in the service of progress and change and life. (Irigaray's mirror, as well as much feminist work on appropriation and speaking lends Lewins vision a darker cast, but I shall leave that aside for now.)
Lewin closes by encouraging his readers to take this and run with it (which perhaps I shall), but cannot resist a few closing remarks of his own. What does this mean for academic music? Well, pragmatically, it means we ought to be better aware of the differences in voice. Men often grow up singing much higher than they do upon maturity, and consequentially have experience singing in all four traditional registers. Women frequently do not share this advantage. Indeed, in my own teaching experience I've found that women are more likely to not be able to transcribe the bass, or to confuse it with an upper voice (and men, though less often, will have a hard time grabbing the soprano). The ideal solution, per Lewin (and I agree) is early piano instruction, before theory and ear training. Lacking this, learning an instrument that does not share your vocal register would be advantageous. This clearly does not happen in most pre-university bands, where the women are too often encouraged to play "feminine" instruments such as the flute or clarinet. Here again, the gents have the edge, since the trumpet has both the high pitch their voices will eventually lack, and the aggression they'd better find for themselves if they're to keep from being ridiculed.