Click consonants?

[ProfSaber]

Hello all,

I'm working on a naming language at the moment and would like it to have clicks in it. Does anyone happen to know how clicks develop from non-click consonants, or know of/have any conlangs that have them?

[Pabappa]

I emailed a professor in S Africa once and he said even the scholars don't know. Your guess is as good as mine, but there are enough of us here that we can probably help you find an idea you can work with.

[ProfSaber]

Thank you.

[malloc]

I have read suggestions that they developed out of labiovelarized consonants in which the velar articulation fortified to become coarticulated. In other words, something like /tw/ > tʷ > tˠ > t͡k > /k!/ for the tenuis alveolar click with analogous changes for other clicks. It seems that some languages have undergone the change /mw/ > /ŋʘw/ through the aforementioned velarization mechanism. The wikipedia article on Chitonga, a Bantu language spoken in Zambia and Zimbabwe, notes that:

At least some speakers have a bilabial nasal click where neighboring dialects have /mw/, as in mwana 'child' and kumwa 'to drink'.[

[Ketsuban]

There's also the suggestion that some clicks arose from consonant clusters - quoth Wikipedia: "For example, the Sandawe word for 'horn', /tɬana/, with a lateral affricate, may be a cognate with the root /ᵑǁaː/ found throughout the Khoe family, which has a lateral click. This and other words suggests that at least some Khoe clicks may have formed from consonant clusters when the first vowel of a word was lost; in this instance *[tɬana] > *[tɬna] > [ǁŋa] ~ [ᵑǁa]."

[WeepingElf]

Yes. My personal theory is that clicks emerge from doubly-articulated consonants involving a velar closure, e.g. *tk > !. Such consonants are frequent in equatorial Africa, and I consider it likely that before the Bantu expansion, most languages south of the Equator had clicks.

[KathTheDragon]

Some speakers of German reportedly have allophonic clicks in such clusters.

[Salmoneus]

There's also the suggestion that some clicks arose from consonant clusters - quoth Wikipedia: "For example, the Sandawe word for 'horn', /tɬana/, with a lateral affricate, may be a cognate with the root /ᵑǁaː/ found throughout the Khoe family, which has a lateral click. This and other words suggests that at least some Khoe clicks may have formed from consonant clusters when the first vowel of a word was lost; in this instance *[tɬana] > *[tɬna] > [ǁŋa] ~ [ᵑǁa]."

I think one point worth making is that the question of how clicks arise in a language may not be the same as the question of how a particular click arose in a language.

Specifically, AIUI, once a language has clicks, it's much easier for more clicks to be created (some languages, for instance, have simply swapped non-clicks for clicks in some words, probably originally having an emphatic, onomatopoeiaic, deformative or playful function. But this doesn't happen until you already have clicks to act as models for these transformations. Likewise, I suspect that changes like the one you mention there could easily create clicks - once 'clicks' are a thing in the language. They may not be as likely to generate the first clicks in the language.

[Travis B.]

There is also evidence of clicks being lost by turning into affricate ejectives, as in the case of some South Cushitic languages that were likely once in contact with now-extinct languages with clicks. I wonder if possibly this could work in reverse as well - affricate ejectives turning into clicks?

[Vardelm]

There is also evidence of clicks being lost by turning into affricate ejectives, as in the case of some South Cushitic languages that were likely once in contact with now-extinct languages with clicks. I wonder if possibly this could work in reverse as well - affricate ejectives turning into clicks?

I always felt that ejective stops were a potential source for clicks: velar & uvular especially. That is based on nothing more than the "logic" of amateurish guesswork, so take it for what you will.

Does anyone happen to ... know of/have any conlangs that have them?

My Jin conlang has them (though it's mostly a sketch still) and I'm sure there are some others.

If you're looking to learn more about clicks, I found audio samples of clicks in the !Xóõ / Taa language. It was super helpful to me.


EDIT: You might be able to find some via this thread about African conlangs on the old board.

[ProfSaber]

Thank you all, this is all very interesting. I like the consonant cluster idea in particular, and losing the first vowel.

I took a glance at Jin @Vardelm and the phonology is quite complex--though most click languages are, I know. I've been planning on this language having "simple" clicks, i.e. plain /! | ǁ / and so forth. I don't know if this is realistic, though--though bear in mind this language is spoken by non-humans.

[akam chinjir]

A bit tangential, but related to Sal's point: am I right to think that both the following are true?

• Languages very rarely acquire clicks as part of their core phonology, except under the influence of another language (though paralinguistic uses of clicks, like English "tsk tsk," are common).
• A language with clicks has no special tendency to lose them, and in particular, children have no special difficulty acquiring them. (I'm also curious how difficult they are for adult learners, compared to, say, contrastive voicing, or lexical tone.)

[Moose-tache]

A bit tangential, but related to Sal's point: am I right to think that both the following are true?

• Languages very rarely acquire clicks as part of their core phonology, except under the influence of another language (though paralinguistic uses of clicks, like English "tsk tsk," are common).
• A language with clicks has no special tendency to lose them, and in particular, children have no special difficulty acquiring them. (I'm also curious how difficult they are for adult learners, compared to, say, contrastive voicing, or lexical tone.)

On the first point, how would we know how "rare" it is? There are three Khoi-San families with clicks, and it's pure speculation how they got them. Same story with Hadza. Outside of that area any use of clicks (e.g. in Damin) appears isolated, and therefore not an influence from another language. It could be that clicks arose once in southern Africa, or multiple times, and could have been borrowed into various Khoi-San languages, or not.
On the second point, apparently Kirk Miller thinks they're easy for children to master, and I suspect adults don't struggle much with the primary articulation. But distinguishing eight kinds of alveolar click? There's no way that's easier than other notorious phonemes. Most Bantu languages with clicks don't bother with secondary articulation.

[Vardelm]

I've been planning on this language having "simple" clicks, i.e. plain /! | ǁ / and so forth. I don't know if this is realistic, though--though bear in mind this language is spoken by non-humans.

If you are planning on this language developing from one that didn't have clicks (which it seems you are), then this sounds perfectly plausible. To an extent, that might depend on the original language. If the parent lang had a voiced/unvoiced distinction, or ejectives, or aspiration, labialization, or other secondary articulations, then I would guess those would carry over to the clicks in some way. It's no different than stops, I would think.

[Salmoneus]

A bit tangential, but related to Sal's point: am I right to think that both the following are true?

• Languages very rarely acquire clicks as part of their core phonology, except under the influence of another language (though paralinguistic uses of clicks, like English "tsk tsk," are common).
• A language with clicks has no special tendency to lose them, and in particular, children have no special difficulty acquiring them. (I'm also curious how difficult they are for adult learners, compared to, say, contrastive voicing, or lexical tone.)

On the first point, how would we know how "rare" it is? There are three Khoi-San families with clicks, and it's pure speculation how they got them. Same story with Hadza. Outside of that area any use of clicks (e.g. in Damin) appears isolated, and therefore not an influence from another language. It could be that clicks arose once in southern Africa, or multiple times, and could have been borrowed into various Khoi-San languages, or not.
On the second point, apparently Kirk Miller thinks they're easy for children to master, and I suspect adults don't struggle much with the primary articulation. But distinguishing eight kinds of alveolar click? There's no way that's easier than other notorious phonemes. Most Bantu languages with clicks don't bother with secondary articulation.

On the first point: out of the whole panoply of human languages, we know of only six families that have ever developed clicks. Three of those families are in close geographic contact, two (of one family each) are near each other, one of the first three is believed to have come from the vicinity of the latter pair and may be related to one of them, and the sixth family is known to have gained clicks from contact with the first three families, relatively recently.

That strongly suggests that clicks do not arise often. Even if we believed that the whole of subsaharan africa was clicky before the bantu came, that would still be a strongly areal pattern. If clicks were something that just happened relatively easily, you wouldn't expect their existence to rely on geography - either due to contact or due to kinship. Instead, they'd be randomly distributed throughout the world. But they're not. All those weird languages in Papua New Guinea, the Caucasus, the Pacific Northwest, the Amazon - dozens and dozens of language families perfectly willing to explore complicated sounds, and repeatedly innovating ejectives, uvulars, implosives and so on, but never, ever, so far as we can see, ever innovating clicks? And yet a whole bunch of language families in contact with one another in just one part of the world do have clicks?

That's overwhelming evidence for the hypothesis that clicks are rarely innovated, but not that difficult to borrow.

On the second point: languages that have fully incorporated clicks into their phonologies seem to have a tendency toward runaway phonological inventories - outside Bantu, pretty much every click language has at least five different articulations for them, and some have up to twenty or more.

Of course, since we know areal effects are in play, it's hard to know what's what. However, I think the most sensible approach is to assume two areas - Tuu-Kx'a and Macro-Khoe-Hadza (MK being Khoe plus Sandawe). The former had immense phonologies, while the latter didn't, but some Khoe languages have expanded their inventories greatly since the family's arrival into the T-K sphere. And yet, even the relatively sparing northeastern (MK-H) area still has 15-20 clicks as the norm. With a sample size of two, it's hard to be certain of anything, but this at least strongly suggests that clicks do tend to multiply (as does the ease with which some individual languages have rapidly multiplied click articulations, while few seem to have significantly reduced them).

[Moose-tache]

I guess from your response that you interpreted my post as saying "we have no way of knowing whether or not click consonants occur very often on Earth." If my posts are generally so pants-on-head stupid that you seriously thought such an idiotic claim was a likely thing for me to say, then I can only apologise and promise to display at least the same intelligence as a chimpanzee or a gifted horse in the future. ;p

Since it apparently wasn't obvious, I was responding to Akam Chinjir's question of how rare it is to develop clicks independently vs borrowing them. Since we have no way of knowing how many times clicks developed (i.e. we cannot know that all three Khoi-San families developed clicks independently), we cannot know how rare they are as a-priori developments. It could be that clicks, outside of cases like Damin, emerged only once in human history. Then again, they may have developed several times in Sub-Saharan Africa, which would make them an areal feature, but no more absurd than the curious features that only appear in one part of the world despite being perfectly reasonable, like a trigger alignment system or grammatical evidentiality. That feels like two different kinds of "rare" to me, but I admit "rare" isn't a terribly useful term for exploring this distinction.

Also, when it comes to secondary articulations, I'd have to check but I'm fairly certain that the Bantu languages that have borrowed clicks either have few to no secondary articulations, or show reduction of secondary articulations over time. I was unaware of Khoi languages that have increased their inventory in recent years, but then those are languages that already had numerous clicks, with secondary articulations, as a major part of their phonologies. Finnish might add another noun case, but if German did the same we'd all be pretty surprised. Neither fact tells us that noun cases have a "runaway" tendency to proliferate over time. In other words, I don't think it tells us much about the nature of clicks to say that languages which already have lots of them sometimes gain more, while languages that recently borrowed a few of them tend to pare them down or not bother with the secondary articulations at all. We could probably say something similar about ejectives or nasal vowels.

Basically, the sort of questions you would want to ask as a conlanger require data that's just not there. We can describe phenomena, but we can't build a predictive model yet.

[Tropylium]

There's also the suggestion that some clicks arose from consonant clusters - quoth Wikipedia: "For example, the Sandawe word for 'horn', /tɬana/, with a lateral affricate, may be a cognate with the root /ᵑǁaː/ found throughout the Khoe family, which has a lateral click. This and other words suggests that at least some Khoe clicks may have formed from consonant clusters when the first vowel of a word was lost; in this instance *[tɬana] > *[tɬna] > [ǁŋa] ~ [ᵑǁa]."

I've learned not too long ago that the Sandawe word in question can however also be a loanword: Burunge has /tɬʼaana/ 'upper leg, thigh' — not the most convincing semantics, but this is as a part of a larger argument that the lateral obstruents in Sandawe would have in general come about as loanword phonemes from Southern Cushitic (Kießling, Mous & Nurse 2008).

But this does not have to torpedo the entire idea — perhaps the ancestors of Khoe arrived from Eastern Africa (as has been long suspected to be the case due to the animal husbandry) and had also picked up some Cushitic loanwords at the time (cf. Blench 2009).

[jal]

I'm working on a naming language at the moment and would like it to have clicks in it.

What hasn't been observed, as far as my quick scan of the replies revealed, is that it would be pretty unrealistic if people speaking languages without any clicks, would use names that have them. And if you need a naming language for a story or something, your readers won't probably be even able to imagine what clicks sound like...


JAL

[Whimemsz]

I'm working on a naming language at the moment and would like it to have clicks in it.

What hasn't been observed, as far as my quick scan of the replies revealed, is that it would be pretty unrealistic if people speaking languages without any clicks, would use names that have them. And if you need a naming language for a story or something, your readers won't probably be even able to imagine what clicks sound like...


JAL

"Naming language" means a language that you create just enough of to use to form names in your novel or other creative outlet or the blank part of your maps whose cultures aren't your main focus or etc. etc., as you seem to acknowledge in your second sentence. But ... it doesn't have anything to do with "names" being different from other words, you're just using the phonology and vocabulary and grammar you devise to create names and nothing else, which is different from saying the phonology (and vocabulary and grammar) you devise are uniquely found in names in that culture. So I'm not sure what the first part of your argument is here. (The second part is a good point to bring up, though, depending on what the OP's intended use of the naming language is.)

[Zaarin]

(The second part is a good point to bring up, though, depending on what the OP's intended use of the naming language is.)

Though Tolkien would like to remind you that you can create a perfectly accessible phonology and still have people butcher your names--like /s/ in Celeborn or /ɒ/ in Sauron. The former could have been helped by a less Latin-esque orthography such as Keleborn, but the latter is about as clear as one could hope to be for Anglophones. I don't think the mispronunciations of the audience need be considered excessively. Alternately, one can imagine how the name might have been passed down to the audience mediated by a more accessible language like Greek or Latin. For example, the Hebrew names we have in English Bibles are utterly butchered, but they also fit Anglophone mouths better than the original Hebrew would in most cases. One can also look at how English-speaking settlers rendered Native American names.