via: Georg Henrik von Wright, “The Wittgenstein Papers,” The Philosophical Review 78:4 (1969), 483–563, here: 487.

von Wright seems to indicate that Wittgenstein created typescripts which he cut up into zettel and then was able to rearrange them into final forms.

Llewelyn, J. E. “Zettel. By Ludwig Wittgenstein. Edited by G. E. M. Anscombe and G. H. von Wright. Translated by G. E. M. Anscombe. (Oxford: Basil Blackwell. 1967. Pp. v + ve + 124 + 124e. Price 37s 6d).” The Philosophical Quarterly 18, no. 71 (April 1968): 176. https://doi.org/10.2307/2217524.

Perhaps apropos from Wittgenstein's own zettelkasten? 🐎

1. ''Putting the cart before the horse" may be said of an explanation like the following: we tend someone else because by analogy with our own case we believe that he is experiencing pain too.—Instead of saying: Get to know a new aspect from this special chapter of human behaviour—from this use of language. (p96)

Wittgenstein, Ludwig. Zettel. Edited by Gertrude Elizabeth Margaret Anscombe and Georg Henrik von Wright,. Translated by G. E. M. Anscombe. Second California Paperback Printing. 1967. Reprint, Berkeley and Los Angeles, California: University of California Press, 2007.

This section makes me question whether or not the editors of this work were aware of the zettelkasten tradition?!?

This brings up the question of how Ludwig Wittgenstein arranged his own zettelkasten...

Peter Geach made an arrangement of Wittgenstein's zettels which was broadly kept in the edited and published version Zettel (1967). Apparently fragments on the same topic were clipped together indicating that Wittgenstein's method was most likely by topical headings. However there were also a large number of slips "lying loose in the box." Perhaps these were notes which he had yet to file or which some intervening archivist may have re-arranged?

In any case, Geach otherwise arranged all the materials as best as he could according to subject matter. As a result the printed book version isn't necessarily the arrangement that Wittgenstein would have made, but the editors of the book felt that at least Geach's arrangement made it an "instructive and readable compilation".

This source doesn't indicate the use of alphabetical dividers or other tabbed divisions.

Based on the dating provided by Anscombe and von Wright, Wittgenstein's zettelkasten slips dated from 1929 to 1948.

for reference LW's dates were 1889-1951

Supposing that the notes preceded the typescripts and not the other way around as Anscombe and von Wright indicate, the majority of the notes were turned into written work (typescripts) which were dictated from 1945-1948.

What was LW's process? Note taking, arranging/outlining, and then dictation followed by editing? Dictating would have been easier/faster certainly if he'd already written down his cards and could simply read from them to a secretary.

Some of the manuscript notes in Wittgenstein's zettelkasten were "apparently written to add to the remarks on a particular matter preserved in the box".ᔥ So much like Niklas Luhmann's wooden conversation partner, Wittgenstein was not only having conversations with the texts he was reading, he was creating a conversation between himself and his pre-existing notes thus extending his lines of thought within his zettelkasten.

In Zettel, the editors indicate that many of Wittgenstein's zettels "were for the most part cut from extensive typescripts of his, other copies of which still exist." Perhaps not knowing of the commonplace book or zettelkasten traditions, they may have mistook the notes in his zettelkasten as having originated in his typescripts rather than them having originated as notes which then later made it into his typescripts!

What in particular about the originals may have made them think it was typescript to zettel?

In 1967, G. E. M. Anscombe and G. H. von Wright published a collection of notes from Ludwig Wittgenstein's zettelkasten which they aptly titled Zettel.