The theme of a recent music theory conference was: "in view of a certain element of subjectivity and irrationality inherent in traditional Schenkerian analysis, is it possible to develop its deep insights in the context of a logically non-contradictory, scientifically, and historically well-founded music theory?" Similar criticisms have been leveled at Schenkerian analysis and theory throughout its history. In this paper, I defend the role of subjectivity in Schenkerian analysis (and, peripherally, in music analysis in general), pointing out the following: (1) that subjectivity is essential in music analysis, because analysis, like performance, is interpretation; (2) that other analytical methods with far greater scientific pretensions are equally, if not more, subjective in practice; (3) that Schenker explicitly rejected the characterization of his work as science, preferring to view it as art; (4) that analysis can itself be viewed as a performative act; and (5) that Schenkerian analysis is not only a theory and a system, but also a practice and a craft. In particular, the scientific principle of independent duplication of experimental results is of only limited relevence to Schenkerian analysis. To illustrate the role of multiple readings in Schekerian analysis, I compare three possible readings of the development section of Clementi's G major Sonatina, Op. 36, No. 2, showing how their kaleidoscopic patterns shift into different alignments, affinities, and allegiances; expressing my view of their strong and weak points; and discussing the role of multiple phrase expansions of different types on this very short, yet effective development section.