K-T-R and other Stringed Instruments

This is a discussion that occurred on a linguistics page on Facebook. I thought musicians might be interested in it. I’ve changed writers’ names – actually I’ve given them fictional initials (except ME, which is, uh, me) – and provided explanations (in italics) for those unfamiliar with linguistics jargon.

ME: Here's one that touches on two of my interests: linguistics and music. Has anyone else noticed the Eurasian preponderance of words for stringed instruments that have three consonants: a K or S (or a voiced derivative), a T, and an R. Guitar, sitar, santur, zither, kithara, cittern, to name a few. This looks like an example of widespread borrowing (the word went with the thing) but the dictionary etymologies usually suggest different origins. There are a few others that are somewhat similar, from outside of the main area where these words are found: krar and kora (both in Africa and missing the T) and koto (in Japan and missing the R). Are these earlier borrowings? Or is there some common root word here? (Nostraticists might suggest an answer, but I think it's too recent an invention for this to be likely.) Or all some of them coincidences? Other stringed instruments have names that are entirely different. ...And, here's a really weird one -- the word "string" in English begins with the same sequence, minus the vowels...

BC: Guitar, zither, kithara, and cittern are documented cognates. The other stuff just sounds like coincidence to me...

(Cognates: words that share a common origin in related languages, for example, “apple” in English and “Apfel” in German.)

CJ: What if the names describe how they sound, when I think of string instruments I think of k's and t's.... it must be something common, like the fact that most languages I've ran into have m for mother, and b, f, or p for father.

LG: Similar to violin, string bass, zither and harp. (The words sound like, or at least describe, the instrument's sound.)

KT: When you pick six common sounds (k, s, g, z, t, r), a concept, and a continent, and look in languages on that continent for words for that concept containing a couple of those sounds, you are very likely to find a number of them even with no borrowing.

As for cognacy, consider that the proto-language whose existence this suggestion entails would have existed far longer ago than the oldest proto-languages that we have firm evidence for (for example, Indo-European, which is probably about 6,000 years old). Now consider some facts about sound change and meaning change.

(Indo-European: a large family of related languages, including many in India, Europe, and places between, such as Iran [Persia].)

(Proto-language: the [extinct] language from which members of a language family descended.)

Sound change: consider the Proto-Indo-European sound *dʰ. One example of a root containing this sound was *dʰeh₁y, meaning 'suckle, nurse'. The first sound in words starting with that root is now /f/ in Italian (e.g., /fiʎo/), /θ/ in Greek (/θili/), was /d/ in English until these words dropped out of the language (/delu/, /deːon/), and has been completely lost in Spanish (/ixo/), to give a few examples. So in about 6,000 years, *dʰ turned into /f/, /θ/, /d/, and nothing—totally different sounds: three different places of articulation, voiced and voiceless, fricative and stop, and nothing whatsoever.

(The “weird symbols” stand for individual speech sounds, called phonemes. These symbols are used in the science of phonetics and not in any actual language. The asterisk before a word or symbol for a sound indicates that it has been reconstructed and is not in the language now. The paired slashes, before and after, indicate a word being written with symbols, not the way it is actually written in the language.)

Meaning change: that same root *dʰeh₁y has descendants in Romance languages meaning 'son/daughter,' in Greek 'nipple,' in Kurdish 'mother,' etc.

(Romance languages: Languages descended from Latin, an Indo-European language.)

If, as you suggest when you suggest that some of these words (other than the set already pointed out) may be cognate, there was some kind of root *{k,g,s,z}-{t,θ}-r in a language that is ancestral to all of the ones you mention—namely Japanese, an uncertain West African language, and Greek or Arabic depending where the guitar word originally came from—this word would have been used many, many years before Proto-Indo-European was around. Since it was around longer, you would expect it to have changed even more in both sound and meaning than that example Indo-European root did, and it would be pretty likely to have dropped out of the language entirely—words do this a lot (I don't think there's any evidence that the speakers of Proto-Indo-European, for example, had string instruments). In summary, if a root in a proto-language that old did survive into a number of geographically and linguistically distant modern languages, it would almost certainly not have recognizable similarities anymore in those languages.

Also, the kora seems to have been invented in the 16th century [according to online sources], so people would have had to hold onto the ancestor of the word "kora" for literally thousands of years, probably more than 10,000 years, before finally applying it to this new instrument. Similar situation for the koto, whose precursor instrument was first introduced to Japan from China in the 7th and 8th century [according to online sources].

DH: It looks like the examples you give fall into two classes: 1. Guitar, zither, kithara, cittern--from Greek kithara. 2. Sitar--loan from Old Persian. What the two have in common is that both contain a shared root meaning 'string,' which is apparently "tar" in Old Persian, and presumably something similar in PIE or other shared ancestor. It's worth noting that "psalter" is also in this group, with the Greek 'psallein' meaning "pluck". "Santur" might be a loan from this Greek word.

(“Loanword” and “borrowing” in linguistics refers to words moving from one language to another, for example, using the word “sushi” in English.)

(PIE = Proto-Indo-European)

EF: Yes as KT explained so well, it's PIE!

ME: Yeah, I thought it was too recent an invention to be entirely PIE or something even older (see my note about Nostratic, if it ever existed). "Krar", by the way, is Amharic, I think, and "santur" turns out to mean "hundred string(s)" in Old Persian.

KT: Most historical linguists reject the idea of a Nostratic family, fyi. I think the similarities among these words are coincidental, except the ones we already know are borrowings.

(Nostratic: A hypothetical language super-family including Indo-European and several other large families.)

ME: I looked up some lists of stringed instruments in the world, and I think I might have figured this out. There are three groups of words here. 1. Those obviously related to "guitar". 2. Others with the root word "tar", meaning "string", as mentioned above. These are all IE, of course. 3. Non-IE outliers: kora, kwitra, and krar in Africa, kantele in Finnland, and koto in Japan. These look like more ancient, and therefore more altered, forms. More likely, they are accidental resemblances though I should note that the African examples are in the traditionally Orthodox and Islamic areas of Africa (which are geographically closer to the homelands of Indo-European), and other words for stringed instruments in Africa and Asia (and Europe) sound completely different. I might propose an earlier borrowing for one or more of these, though Idon't know which ones.

JW: I wonder whether there is any true acoustic similarity between the names of string instruments and the sounds they produce, say, "s", "z", "c" for the continuous sounds, "k", "t" for the pizzicato or pinching of the strings. Maybe just a coincidence?!...

KT: Yeah, I think just coincidence. The problem is it's pretty easy to imagine an onomatopoetic relationship between any randomly selected segments and things (especially things that make noise).