'Altered recapitulations,' commonly regarded as a distinguishing feature of Joseph Haydn's sonata form movements, are usually explained in terms of the 'monothematic' design of the exposition. According to the logic used in such analytical studies, recomposing the recapitulation would have been aimed at restoring the proportional balance between exposition and recapitulation, a need that resulted from the omission of the seemingly redundant, retransposed secondary theme along with the preceding transition. Though such an explanation has long been considered indisputable, this article casts doubt on the validity of the redundancy principle by showing that Haydn often did retain the monothematic section in the recapitulation. Rather, the recomposition of the recapitulation results from two important structural aspects thus far largely neglected in the literature: (1) the repetitive formal structure of the main theme, which is often considerably reworked in the recapitulation; and (2) the insertion of a separate newly composed dominant zone in the recapitulation that serves to compensate for the lack of a structural dominant at the end of the development section. Finally, it is argued here that Haydn, who was deeply rooted in the late Baroque tradition, by no means regarded multiple 'double returns' as either problematic or redundant, for he may have been thinking more in terms of an overriding ritornello structure.