A verb appearing as ᴱQ. sar- “write” in Early Qenya Word-lists from the 1920s. Hints at its later validity may be seen in Q. sarat, the names of the letters in the Sarati alphabet, and the root √SAR “score, incise; write”, both mentioned in the Quendi and Eldar essay of 1959-60.
Neo-Quenya: Most Neo-Quenya writers use [ᴹQ.] tec- as the ordinary word for write, given its obvious association to Q. tehta and Q. tengwa. One of the challenges with sar- for “write” is that in Tolkien’s later writings, √SAR is also associated with “stone”, as in S. sarn/Q. sar “(small) stone”. In a discussion on the Vinyë Lambengolmor Discord Server (VLDS) in April 2022, Röandil suggested restoring ᴺQ. sar- with the sense “to mark (i.e. by incision), notch, score; (by later extension) to write”. Based on this I invented the following (fan-based) etymology:
Given its derivatives, it seems likely that √SAR is associated with stone but also with marks, probably originally marks in stone but generalized from there to marks on other substances. From this sar- was used as the basis for the earliest system of writing, both as a verb and in the name sarati. However, with the introduction of Feanor’s tengwar alphabet, this word no longer seemed suitable for as a general writing word, and was displaced by tec-, originally also meaning “make a mark”, but not specifically associated with stone.
Since tec- now usually meant “write”, the archaic sense “mark, score” was restored to sar-. This was aided by the fact that sar- was primarily a transitive verb, and used some awkward constructions. In particular, with the verb sar-, the direct object was always the thing written on, rather than the thing written. Thus with sar-, the proper way to say “write a sentence on paper” was actually sare hyalin quettalénen “write [on] paper with a sentence [instrumental]”, a construction that still survives, though now somewhat archaic. Meanwhile, the direct object of tec- is the thing written, so one would say tece quettale hyalinesse “write a sentence on paper”.
Note that this last construction [sare hyalin quettalénen] is inspired by Tolkien’s usage of the ᴱQ. verb teke- from the 1910s (QL/90), and essentially reverses of the Early Qenya paradigm of the 1910s and 20s, switching the roles of tek- and sar-. There is no evidence that the above Neo-Quenya discussion matches Tolkien’s own thoughts on the subject, so it should be considered “etymological fan-fiction”.
A verb appearing as ᴱQ. sara- “saw” in the Qenya Lexicon of the 1910s derived from the early root {ᴱ√SAŘA >>} ᴱ√SARA, apparently of the same meaning (QL/82). Its Gnomish cognate G. thar- “saw (up)” indicates an actual root form of ✱ᴱ√ÞARA.
Neo-Quenya: The root √THAR means “across” in Tolkien’s later writings, but I think ᴺQ. sara- “to saw [wood]” is still salvageable from an original sense of “✱(cut) across”.