The fact that your question already begins by calling the ideas "dogma," makes me doubt that you'll be very amenable to discussion, but in case other people reading this are curious:
1) In comments you claim to have learned this dogma from "the texts." "The texts" say no such thing; in fact, they say no single thing at all. There is a vast collection of theory texts, with a vast range of focus and pedagogical intent. Some aren't very good, some are great; some are old and benefit from immediacy but lack perspective, some are new and benefit from hindsight but occasionally devolve into less-useful abstractions. I've taught from the Kostka/Payne text, the Aldwell/Schachter, the Salzer/Schachter and the new Clendinning/Marvin text. These are some of the most widely-used and respected textbooks in tonal harmony, and absolutely none of them claim that composers conceived of their music as an elaborated chorale structure or that they consciously wrote music that adhered to the specific voice-leading conventions generally taught to undergraduates.
2) "Piano music" isn't a genre. Depending on the century, the country, the genre and the general time in a specific composer's life, you will find a huge assortment of harmonic, melodic, rhythmic and timbral techniques, obsessions and dogmas. Some will indeed essentially be elaborated chorale-like structures, some will have clearly contrapuntal origins that are obscured by ornamentation, and for some a voice-leading graph will be less explanatory. There are some general aspects of piano writing that make things like, for example, parallel fifths, a somewhat more common occurrence. However, it is true that if the piano music is being written in a common-practice tonal idiom during Baroque, Classical and Romantic eras in European music, most general principles of voice-leading will be adhered to more often than they aren't. This isn't because the composers were consciously following rules, it's because the general principles of counterpoint were baked-in; they are part of the DNA of common-practice tonality.
3) Principles of voice-leading are generally taught through 4-voice chorales, but that does not equate to a claim that all common-practice music is fundamentally based on four voices. Four voices is a useful number for pedagogical purposes—not so many voices that it's overwhelming, not so few that you have to constantly use incomplete triads and sevenths. It's also helpful in that it guarantees that many complex issues of contrapuntal doubling will come up organically and can be practiced. Furthermore, it's helpful that there's a lot of music written for four-voices that can be looked at for examples that don't stray as far from the exercises. None of this is the same thing as claiming that common-practice composers always start with a four-voice framework when composing, consciously or otherwise. It's a starting place for discussion of complex principles that isn't burdened by too many variables.
4) Musical analysis is not about reconstructing the compositional process for a particular work. It is about discovering potentially interesting or revelatory aspects of a piece that lie beneath the surface. My old theory teacher says it's about telling an interesting story about a composition, never about "explaining" it. Sometimes an earlier step in such analysis does involve boiling away surface ornamentation in order to get a better sense of the underlying structure—not because such details aren't important, just because it can help an analyst notice things that were previously more obscure. Most certainly, some analysts go too far and make assumptions that aren't borne out by the actual piece. Theorists argue constantly about what is too abstract and what isn't. But that doesn't mean the entire concept of fundamental voice-leading structure in common-practice tonality is wrong.
Beethoven, Mozart et al, most certainly weren't consciously writing in terms of voice-leading principles as taught in a theory class. They didn't have to. It was in their ears, their fingers and their muscle memory. Composition instruction in those days in Europe began with intensive contrapuntal instruction—generally starting with two- then three-voice Species Counterpoint and then moving on to four-voice chorales and beyond. Some of the rules they learned and taught would be immediately familiar to a modern-day student, some wouldn't. Mozart thought that it was important to introduce double-neighbor figures in Third Species, Fux thought it was irrelevant to proper counterpoint. The point was never any one specific rule, but rather a keener understanding of how two or more independent voices interact. By the time they were writing their own pieces, they no more had to consciously avoid doubled leading tones and parallel fifths than they had to consciously keep breathing. It was part and parcel of the music. When they had an idea that ran counter to the principles, they used it; exceptions abound and are often the most interesting parts of a composition. Nevertheless, a firmer grasp of how their music functions structurally by default is enormously helpful to understand both the basic and the exceptional segments of their music.
