This is the main database for bringing together various comparanda from languages and proto-languages that, together, are likely - with various degrees of probability - to constitute the "Khoisan" (or "Macro-Khoisan") family. As such, it is NOT an etymological database for "(Macro-)Khoisan" because no system of regular phonetic correspondences for the languages involved has been presented, and it is not even certain that a single, rigid system of correspondences would necessarily work on the basic lexicon of all the languages that are concerned.

The current state of affairs is more or less as follows:
[a] There is significant evidence that Proto-Peripheral Khoisan, consisting of North Khoisan (Zhu), South Khoisan (!Wi-Taa), and Eastern =Hoan, and Proto-Central Khoisan (Proto-Khoe), consisting of Khoekhoe, West Khoe, and East Khoe, are related on a much higher level. This evidence is often hard to distinguish from traces of more recent contacts between Khoe idioms and various parts of Peripheral Khoisan, but various small patterns of correspondences are extremely hard to explain as contact elements. A more proper approach to the relations between the two will, however, only become possible with the successful reconstruction of Proto-Peripheral Khoisan.
[b] There is also some promising evidence, although not as much, that Sandawe is relatable to both PPK and PCK on an even higher level. Scenarios that relate Sandawe to these proto-languages, among other things, necessarily involve the process of contraction of former bisyllabic roots (preserved in Sandawe) into monosyllabic roots with secondary formation of clicks from former non-clicks in PPK and PCK, although it is not clear if this process took place on the "PPK-PCK" level or happened independently in both subgroups. The latter suggestion is indirectly confirmed by the fact that Sandawe seems to have more obvious isoglosses with Khoe than it has with Peripheral Khoisan; however, this is still a preliminary impression.
[c] A related question is the status of Kwadi, an extinct Khoisan language from which very little data is available. It has recently been shown by T. Güldemann and D. Elderkin in a series of publications that Kwadi is distantly related to Khoe, and that Sandawe may be an even more distant relative of both. However, the question of whether we have to suggest a tripartite division for the entire family (Sandawe; Khoe-Kwadi; Peripheral Khoisan), or one of the two bipartite options (Sandawe-Kwadi-Khoe vs. Peripheral Khoisan; or Kwadi-Khoe-Peripheral Khoisan vs. Sandawe) remains open.
[d] Concerning Hadza, the general consensus is that it is not closely related to any of the postulated or demonstrated branches of Khoisan. At best, Hadza is the earliest branch to have split off the Khoisan stem; at worst, Hadza has no special affinity to Khoisan whatsoever. My current view is that it is at least equally as promising to seek this language's affiliation with Afro-Asiatic as it is to seek it with Khoisan. Nevertheless, even in the "worst" possible case, Hadza is still characterized by a strong Khoisan substrate element, which justifies inclusion of Hadza data into the current database.

The database consists of the following fields:
1. Proto-Khoisan: a sketchy "generalized" invariant for all the comparanda presented below, not a true proto-form in the classic comparative sense. Where data from Hadza and Sandawe seem to indicate that this must have formerly been a bisyllabic root, with contraction (and possibly secondary click formation) in Khoe and Peripheral Khoisan, I mark it as a click-less CVCV type root. Where such data are lacking, I mark it with the same type of click it has in Khoe and/or Peripheral Khoisan (even though it could also go back to a click-less root - we just don't have any external parallels to confirm it).
Items marked with [~] represent Swadesh wordlist entries that have been found in various branches of Khoisan, but do not seem to have any obvious parallels outside one and only one branch. These are not true "proto-Khoisan" comparanda; they have been included for completeness purposes, in order to provide a general lexicostatistical perspective.
2. Meaning: the general meaning(s) of the hypothetical stem.
3. Proto-Bushman (= Peripheral Khoisan): the compared item in the PPK family, linked to the Peripheral Khoisan database.
4. Proto-Khoe (= Central Khoisan): the compared item in the Khoe family, linked to the Khoe database.
5. Sandawe: the compared item in Sandawe, linked to the Sandawe database.
6. Hadza: the compared item in Hadza, linked to the Hadza database.
7. Notes: additional comments and considerations.
8. References: bibliographical links (mostly just references to previous suggestions of the same etymologies; references to language data sources should be found in the subordinate databases).

Transcription conventions for "proto-forms" are rather loose, considering the non-etymological character of the database, but in most cases, are not different at all from similar conventions in the Peripheral Khoisan or Khoe databases; see their respective descriptions for details.