Consider the title Johannes Brahms gave to his Sonate für Pianoforte und Violine (op. 78), in which he named the piano before the violin--contrary to standard practice--to emphasize the equality between the two instruments. Now consider where this work should be classed, using notes in their state prior to a recent set of updates in the 780s. Our starting point might be 785 Ensembles with only one instrument per part, with chamber music in its class-here note. A class-elsewhere note there instructed, "Class works for solo melody instrument with keyboard or other accompaniment in 786–788." (That is, chamber music for two instruments, one of which is a keyboard instrument, generally is classed with the non-keyboard instrument.) The class-elsewhere note at 785 was reinforced by the corresponding class-here note at 786–788, "Class here music for solo instrument, music for solo instrument accompanied by one other instrument when the accompanying instrument clearly has a subsidiary role . . ."But wait! What if the accompanying instrument doesn't clearly have a subsidiary role (or clearly doesn't have a subsidiary role)? Should such a work be classed instead as an ensemble with two instruments, for example, in 785.28 Ensembles of strings and keyboard? But a scope note there limits its applicability to "three or more instruments."
The problem faced by classing such works (which are typical of chamber music involving the piano) lies in the assumption that keyboard accompaniment always plays a "subsidiary role," an assumption that is also, to some degree, built into the word "accompaniment." We have addressed this by supplementing our musical accompaniment verbiage with musical collaboration language. For example, the note at 786–788 now reads: "Class here music for solo instrument, music for solo melody instrument with keyboard or other accompanying or collaborative instrument . . ." Note that these changes simply make clear what was intended all along. Specifically, a work like Brahms' sonatas for piano and violin is still considered part of the violin repertoire and should be classed in 787.2183 Violin sonatas (built from 787.2 Violins, plus 183 from 784.183 Sonatas, following the instructions at 784–788).
Does music exist in which a keyboard or other instrument is an accompanying instrument, but not a collaborative instrument? Yes. Consider, for example, J. S. Bach's Sonate für Flöte und Basso continuo e-moll (BWV 1034). Here we have a sonata written with two parts. One part is for the solo melody instrument, in this case a flute. The second part is a bass part, supplied with numerical figures indicating the harmonies to be filled in, known variously as a basso continuo or continuo or figured bass line. While this line might be played by multiple instruments, at least one continuo instrument would be capable of filling in the harmonies through a realization of the figured bass. Often this would be a keyboard instrument, but could also be a lute, a guitar, a harp. Works like Bach's flute sonatas should be classed in 788.32183 Flute sonatas (built from 788.32 Flutes, plus 183 from 784.183 Sonatas, following the instructions at 784–788). Note that the classification of such works does not depend on determining whether the keyboard instrument plays an accompanying role or a collaborative role, since the end result is the same.