User:Tropylium/Etymology vs. lexicography

Epistemological notes
Etymologies are not facts. They are theories, formed by examination of words in distinct language varieties and constrained by theories of language change.

Facets of etymologies

 * The existence of an etymological connection between a set of words. Known as the Gleichsetzung, this forms the basic hypothesis for all etymological research, to be elaborated by the examination of detailed evidence.
 * The nature of the connection. The two main options are borrowing and common descent. They however also allow for a gray area in-between.
 * The postulation of an origin. For loan etymologies, an attested word will often serve for this. For common descent, one will typically need to posit origin from an unattested proto-language
 * The
 * The postulation of historical processes, leading to the __ of the words being connected. The two most important are semantic change and sound change. In etymologies bridging several centuries (if not millennia), affixation also often becomes of importance, as commonly it may be the case that a more deeper-range etymology only connects the roots of given words, rather than the words in their entirety. Finally, linguistic connections have to be taken into account as boundary conditions: a loanword etymology cannot be proposed without contact between the languages involved, while an etymology by inheritance cannot be proposed without common origin of the languages involved.

Source grades
As (very briefly) covered at Wiktionary is a secondary source, lexicographical work is rather research-like, and requires intimate attention to primary sources. Secondary sources, such as other dictionaries, should be used sparingly at best.

Note that the absolute phrasing of the previous link — Wiktionary must not rely on other secondary sources, such as other dictionaries — is contradicted by core policy over at WT:ATTEST, which allows the inclusion of words from lesser-used languages on the basis of a single mention from materials "deemed appropriate by the community of editors for that language". (It is unclear to me for how many languages lists of such materials are actually maintained.)

Etymology, by contrast, is not merely research-like — it is a proper sub-branch of linguistics. Etymology takes lexicographical data only as its basic starting point, and requires additional research to generate results that are usable and reliable.

As of September 2015, no editor consensus exists on what grade of an etymological source Wiktionary ought to be. The most strident have argued for a tertiary status, asking for any given etymology to be supported by reliable secondary sources; the most liberal have argued for a primary status, asking for etymologies to be allowed based on the original research of informed editors.

Wiktionary is not Wikipedia, but…
Wiktionary is not Wikipedia, as is widely known.

When it comes to etymological rather than lexicographical work, though, this division becomes less clear than it usually is.