Why the ‘Coffee’ Words Are Not Cognates

Dec 12, 2014 by

coffee_map.png.CROP.promo-mediumlargeA former student of mine drew my attention to a recent article in Slate written by Alyssa Pelish and titled “The Stimulating History of Coffee: Why You Hear This Word Around the World” (the image on the left is reproduced from the article). Pelish starts with a little thought-experiment about how one would order a coffee while travelling around the world: Kaffee in Berlin, caffè in Rome, kofi in Lagos, Nigeria, kŏfī in Delhi, India, and кофе (pronounced /’kofè/) in St. Petersburg, Russia. She correctly points out that these words sound alike in many languages, describing these words very poetically as “the two reliable syllables, the seesaw of vowel sounds punctuated by velar stops and fricatives”. I am not sure about the reliability of syllables or how one would go about measuring it, or whether the alternating consonant-vowel-consonant-vowel (CVCV) pattern can be called a “seesaw”. But the explanation Pelish provides for why these ‘coffee’ words are so similar the world over is entirely wrong and ignorant.

She claims that the recognizable nature of the different words for ‘coffee’ that one might hear around the world is due to the fact that they are cognates. She writes: “These are words that share the same root, and, often enough, they’re a good indication that two languages developed from a common ancestor”. In the next paragraph, she adds:

“But one of the tricky things about cognates—true cognates—is that they aren’t always an indication that two languages share a common ancestor”.

I had to re-read this sentence several times to make sure the negation is really there because the sentence would be perfect if the little “n’t” weren’t there. The one tricky thing about cognates—which all too many researchers and journalists mess up—is that they always are an indication that two languages share a common ancestor. In fact, that is precisely the definition of cognates, which one can easily find in any introductory textbook if one desires to provide one’s readers with correct and well-informed stories: cognates are words from different languages that share a common ancestor.

The word that Pelish should have used is “look-alikes”. Or perhaps more precisely “sound-alikes”. As she later admits, “coffee is a loan word. That is, it was borrowed, fully formed, from another language”. But that’s just it: loanwords are not cognates! Sometimes it is hard to recognize loanwords as distinct from cognates, especially if the loanword comes from a closely related language. But that does not make all words that “share the same root” cognates. Nor is the use of the term “cognates” in connection to ‘coffee’ words in the above-cited passage an isolated instance in this article; further down Pelish writes:

“To trace the history of the word, which encompasses the development of its many cognates, is to trace the routes of travelers, traders, and colonists in the early modern world—and beyond.”

Her other example of cognates is the English milk and its ilk: German milch, Dutch and Norwegian melk, Czech mléko. These words in Indo-European languages all descend from a common ancestor. As Pelish so cutely notesProto-Indo-European Air would serve you melǵ”. Indeed, the ‘coffee’ and ‘milk’ words are a perfect opportunity—so perfectly missed by Pelish—to explain the difference between loanwords and cognates. Since cognates develop independently from the same source, they show the sound changes that each language underwent. For instance, in the case of ‘milk’, the Russian form moloko exhibits the so-called pleophony, a -olo- “sound signature” characteristic of East Slavic languages. The corresponding word in Czech, a West Slavic language, mléko, is missing the vowel corresponding to the first -o- in the East Slavic form. In Germanic languages, another branch of Indo-European, the vowel corresponding to the first -o- of the East Slavic -olo- is present, but the second one is missing: hence, milk. Because sound changes are regular, we find the exact same pattern in other sets of cognates. For example, the East Slavic pattern is evident in the Russian words gorod ‘town, city’ (as in Novgorod), ogorod ‘vegetable garden’, and izgorod’ ‘fence’. (Never mind the meaning differences between the words: they all come from the same root that means ‘enclosure’.) The West/South Slavic pattern is illustrated by the Old Church Slavonic form grad (which Russian borrowed, as in Leningrad), and the Germanic pattern is observable in the English garden.

Loanwords, in contrast, do not exhibit the sound changes that the target language underwent in its history; in the words of Pelish, loanwords are taken over “fully formed”. For example, the Polish (West Slavic) word koromisło ‘shoulder yoke’ was borrowed from Ukrainian (East Slavic), and therefore carries with it the “sound signature” of East Slavic pleophony, the full-sounding -olo-. In the case of kofe, it too bears a clear mark of its foreignness: the sound /f/. As it so happens, all words with /f/ in Russian—grafika, forma, fil’m, sofa, pufik, and many more—are loanwords.

Loanwords may wear their foreignness on their sleeves, so to speak, in other ways besides their peculiar sound patterns. Let’s take the Russian ‘coffee’ word again. Russian has a sex-based gender system, classifying all its nouns, even those for inanimate things, as either masculine, feminine, or neuter. Adjectives, (past tense) verbs, and certain other elements in the sentence must agree with the noun in gender. Thus, ‘In the corner stood a little chair’ is V uglu stojal malen’kij stul, whereas ‘In the corner stood a little chair stool’ is V uglu stojala malen’kaja taburetka. Yet, ‘coffee’ is problematic from this point of view: some speakers treat it as masculine and others as neutral. The problem is that kofe ends in –e and therefore should be neuter. Since it does not change for case (i.e. is indeclinable) and refers to an inanimate thing, it should also be neuter, like boa the scarf (which is neuter) as opposed to boa the snake (which is masculine). Yet, prescriptive grammars have for decades required kofe to be treated as masculine, and now different prescriptive sources vacillate between the masculine and neuter options. A declinable, “native Russian” word (or even a word with a longer history within the language) would not cause this type of problem: more ‘sea’, solnce ‘sun’, and serdce ‘heart’ are all indisputably neuter. Curiously, kakao ‘cocoa, hot chocolate’ causes exactly the same problem as kofe, and for the same reason: it ends in –o, which is the other typical neuter ending, but it is an indeclinable “import”, and for a long time has been treated by prescriptive grammars as masculine. Thus, deciding if your ‘coffee’ or your ‘cocoa’ in Russian is vkusnyj ‘tasty’ (masculine) or vkusnoje (neuter) is quite a dilemma. Luckily, the same problem does not arise with ‘tea’, čaj, which is both declinable and uncontrovertibly masculine. Still, like ‘coffee’ and ‘cocoa’, ‘tea’ is a relatively recent loanword; it simply lucked out by having a form that is more easily absorbed into the Russian grammar.

The Take-Home Message: loanwords are not cognates!





Subscribe For Updates

We would love to have you back on Languages Of The World in the future. If you would like to receive updates of our newest posts, feel free to do so using any of your favorite methods below: