The shift could be explained historically, but I am inclined to suspect of our tendency to organize differences along a chronological grid. I would prefer to adopt along with it also a genre-distinction. It is possible that the genres evolved one after the other, so that their beginning is chronologically identifiable, although texts belonging to all genres have then been composed at the same time.
Can one still detect the transition from one genre to the other? Consider the following quote from Jan Gonda, The ritual sūtras, 1977: 515-6:
Especially the last chapters (praśna) of the [Baudhāyana] Śrautasūtra […] are something between a brāhmaṇa and a sūtra because the motivation of the prescripts is so often added. Among the explanations are many myths, part of them unknown from other sources. […] It is in perfect harmony with the character of this work that its author has not aimed at the well-known brevity and conciseness of the sūtra style. On the contrary, cases of repetition (also of longer passages), prolixity and diffuseness are far from rare. The several rites are dealt with independently. Verbs are often repeated, implications avoided; the phrases are as a rule long and the syntax is free.I think that this specific place for all controversial topics may derive from the Brāhmaṇa structure, where controversies where altogether absent, influenced by the dialectical style of the Kalpa Sūtras, which demands that one deals with controversial points. In fact, Gonda himself follows (p.516):
Another remarkable feature is the structure of the work […]. Whereas the different opinions of other authorities are in the other sūtras subjoined to the several views of the author and consequently dispersed over the text the compiler of the Baudhāyana-Sūtra has brought all this controversial matter together in the four chapters XX-XXIII, the so-called dvaidha-sūtra ("on variant or different opinions").
After this part of the book there follows, closely connected with the dvaidhasūtra, the so-called karmāntasūtra (XXIV-XXVI) which countains pariśiṣṭas (i.e. paraliomena, topics that were not sufficiently explained in the preceding chapters). Caland had good reasons for supposing that the dvaidha- as well as the karmāntasūtra are additions of material that, through dating back to the same period as the main text, was drawn up somewhat later, perhaps by a pupil.