‘With the least possible editorial intervention’

Much about the order of composition of Remarks on Colour is debatable but not that Part I was composed after Part III. While a higher manuscript number does not guarantee a manuscript is a later addition, it is practically certain that MS 176, the source of Part I, follows MS 173, the source of Part III. Only at the expense of seriously distorting what Wittgenstein says and how he proceeds can Part III be taken to expand Part I, rather than Part I taken to compress Part III. The majority of the 88 remarks of Part I (MS 176, pp. 1r–22r) clearly derive from the 350 remarks of Part III, some from III.1–130 but mostly from III.131–295. Much is copied verbatim, or almost so, but many remarks are revised in ways that strongly suggest Wittgenstein was working with MS 173 in hand. More tellingly still, when the two sets of remarks are compared, there can be little doubt that Wittgenstein was paring down and reorganizing the remarks of Part III. The remarks of Part I are mostly better worked up and more deftly fashioned, and there are many fewer erasures and inserted words. Moreover, counterparts of Part I in Part III are struck through, a fairly conclusive indication that Part III is the source and its remarks were being selected and rewritten (compare Rothhaupt, Farbthemen, pp. 602–719).

In her ‘Editor's Preface’, Anscombe reports that Wittgenstein's literary executors, Rhees, von Wright and herself, elected to publish the remarks on colour in MSS 172, 173 and 176 since they give ‘a clear sample of first-draft writing and subsequent selection’. This decision is hard to quarrel with, a few quibbles aside. Wittgenstein may have been relying on notes when he wrote Part III (and, conceivably, Part II as well), and a stickler for the truth might cavil at Anscombe's observation that Part I is ‘a selection and revision of the earlier material, with few additions’. There is nothing in Part I from Part II, only a few echoes of what is said there, and almost a quarter of the remarks in Part I are new, some 20 or so of its 88 remarks having no clear antecedents in Part III.