Various letters have been used to write the click consonants of southern Africa. The precursors of the current IPA letters were created by J. G. Krönlein, popularized by Karl Richard Lepsius, and continued by Wilhelm Bleek.
Individual languages have had various orthographies, usually based on either the Lepsius alphabet or on the Latin alphabet. They may change over time or between countries. Latin letters, such as c q x ç, have case forms; the pipe letters, ? ? ? ?, do not.
By the early 19th century, the otherwise unneeded letters c q x were used as the basis for writing clicks in Zulu by British and German missions. However, for general linguistics this was confusing, as each of these letters had other uses. There were various ad hoc attempts to create letters--often iconic symbols--for click consonants, with the most successful being that of Krönlein popularized by Lepsius. Doke later created a different system, based graphically on the IPA letters of 1921 and theoretically on an empirically informed conception of the nature of click consonants.
|Lingvarium (ca. 2005)|
|typewriter substitutions||@||/||//||!||§ or =|
Besides the difference in letter shape (variations on a pipe for Lepsius, modifications of Latin letters for Doke and Beach), there was a conceptual difference: Lepsius used one letter as the base for all click consonants of the same place of articulation (called the 'influx'), and added a second letter or diacritic for the manner of articulation (called the 'efflux'), treating them as two distinct sounds (the click proper and its accompaniment), whereas Doke used a separate letter for each tenuis, voiced, and nasal click, treating each as a distinct consonant, and thus following the example of the Latin alphabet, where the voiced and nasal occlusives also treated as distinct consonants (p b m, t d n, c j ñ, k g ?).
Doke's nasal-click letters were based on the letter ⟨n⟩, continuing the pattern of the pulmonic nasal consonants ⟨m ? n ? ? ? ?⟩. For example, the letter for the dental nasal click is ⟨?⟩; the alveolar is similar but with the curl on the left leg, the lateral has a curl on both legs, and the retroflex and palatal are ?, ? with a curl on their free leg: . The voiced-click letters are more individuated, a couple were simply inverted versions of the tenuis-click letters. The tenuis-voiced pairs were ⟨? ?⟩ (the letter ⟨?⟩ had not yet been added to the IPA for the voiced velar fricative), ⟨? Q⟩, ⟨? ?⟩, similarly ⟨?⟩ and its inverse, and lateral ⟨?⟩ paired with a double loop (an inverted ?): . A proposal to add Doke's letters to Unicode is not yet decided finally (as of December 2017, Unicode version 10.0).
Beach wrote on Khoekhoe and so had no need for letters for the voiced clicks; he created letters for nasal clicks by adding a curl to the bottom of the tenuis-click letters: double-barred ⟨?⟩ for nasal ⟨?⟩, stretched ⟨?⟩ for nasal ⟨?⟩, turned ⟨?⟩ for nasal ⟨?⟩ (though with the curl on the bottom), and something like a topless ⟨?⟩ for nasal ⟨?⟩: .
Doke had run "admirable" experiments establishing the nature of click consonants. Nonetheless, Bleek in his highly influential work on Bushman languages rejected Doke's orthography on theoretical grounds, arguing that Doke's letters stood for two sounds each, "a combination of the implosive sound with the sound made by the expulsion of the breath" (that is, influx plus efflux), and that using Doke's orthography it was impossible to write the clicks themselves, as "we cannot call [them] either unvoiced, voiced, or nasal." Bleek therefore used digraphs based on the Lepsius letters, as Lepsius himself had done for the same reason. Ironically, linguists have since taken the co-articulation to be inherent in the Lepsius (pipe) letters, since the 'influx' can never occur alone, and therefore use the simple letters for the tenuis clicks rather than for some abstract 'clickness' as Bleek had. However, since the Lepsius letters have become standard (and even when the Doke letters were official in the IPA, only the letters for the tenuis clicks had been adopted, being treated as conceptually equivalent to the Lepsius alphabet), today if linguists wish to reflect the dominant view, and to use the IPA, they must resort to diacritics that would not be used for non-click consonants.
Summarized below is the evolution of formal click transcription, from Bleek's digraphs reflecting co-articulated consonants, to ligatures intended to function as single letters, to full IPA with diacritics, along with an equivalent treatment of the tenuis, voiced, and nasal non-click occlusives [t d n] (for illustrative purposes).
Written languages with clicks generally use an alphabet either based on the Lepsius alphabet, with multigraphs based on the pipe letters for clicks, or on the Zulu alphabet, with multigraphs based on c q x for clicks. In the latter case, there have been several conventions for the palatal clicks. Some languages have had more than one orthography over the years. For example, Khoekhoe has had at least the following, using palatal clicks as an example:
Historical roman orthographies have been based on the following sets of letters:
There are two principal conventions for writing the manners of articulation (the 'effluxes'), which are used with both the Lepsius and Zulu orthographies. One uses g for voicing and x for affricate clicks; the other uses d for voicing and g for affricate clicks. Both use n for nasal clicks, but these letters may come either before or after the base letter. For simplicity, these will be illustrated across various orthographies using the lateral clicks only.
|Zulu||> ca. 1850||x||xg||xn||xh|
|Ju?'hoansi||modern||?||g?||n?||?'||?h||?x, g?x||?k, g?k||?'h||g?h||n?h|
|1987||x||dx||nx||x'||xh||xg, dxg||xg', dxg'||x'h||dxh||nxh|
Essential to the [clicks] is the peculiarity of stopping in part, and even drawing back the breath, which appears to be most easily expressed by a simple bar ?. If we connect with this our common marks for the cerebral [i.e. retroflex: the sub-dot] or the palatal [the acute], a peculiar notation is wanted only for the lateral, which is the strongest sound. We propose to express it by two bars . As the gutturals [i.e. posterior articulations] evidently do not unite with the clicks into one sound, but form a compound sound, we may make them simply to follow, as with the diphthongs. (Note: Lepsius used short bars which are not available with Unicode 6.3 and are approximately represented here by a dotless ?, but in fact are bars without serifs.)-- Lepsius (1863:80-81)