Hack Foreign Vocabulary with the DNA of Language

We've looked at Spatial Memory and Language Families, and now we can apply similar concepts that we understand from biology to track how languages evolve over time and space.

In our next post, I'm going to tell you about writing systems, what they are, and most importantly, what they most certainly are not.

Etymology

Etymology is the study of the history of words. To many people, it seems to be a useless field of study, even for those who study foreign languages. But with just a little bit of etymology, you can save countless hours on your studies of foreign languages. Especially if they belong to the same language family.

The phonemes of language are very much like the DNA in biology. This is because over time phonemes evolve into different forms. Unlike biology, they don't evolve out of necessity from the environment, but it is not impossible to rule out environmental factors. It is apparent that languages do simplify over time.

A Tale of Two Lects

This is the tale of the two fictional languages Fûlz and Radubuy, whose names descended from the same name.

Let's take a look at a likely scenario. An early language spoken tens of thousands of years ago may have been made up of single syllables. Let's call this language Vol. As people used the language Vol more and more, they started assembling these single syllables into more complex combinations. Over time, some vowels were less emphasized and what resulted were longer words with many consonant clusters. So let's apply something like this to the language name and call it Protovols.

Over time, native speakers in the west started fusing consonants together, which elders may have considered "lazy speech", even though it is a natural course of language change. This resulted in a Western Protovols language. And speakers in the east started to change their consonants so they flowed together easily, for example changing stops to fricatives, giving rise to Eastern Protovols. Over time, the people in the west and east who once spoke the same language no longer were able to understand each other. In fact, by now, they even called their own language by different names: Prtols in the west and Hrodhovol in the east.

As more time passed, the western language may have returned to a completely new set of single-syllable words through many generations of simplification. Since the original stresses on the words would have fused together into single syllables, this language may have applied the stress contours onto the final resulting syllable, giving rise to a tonal language. And the language name may have become something like Fûlz (Pŕtòls > Fθûlz > Fûlz). An example of tones starting in languages can be seen today in languages like Tibetan and Swedish.

The eastern language may have instead simplified its consonant and vowel inventory, keeping its vowels intact, and its name may eventually have become Radubuy. After such a long period of time and language change, the two languages appear nothing like each other, but there may only be left a slight hint of a relationship based on the structure of words. A similar situation can be seen in how the Austronesian languages developed from a greatly disputed ancient language.

In fact, sounds can and do change in predictable ways over time. Although the relationship between Austronesian and Chinese is disputed, some slight traces of relationships can still be seen. For example there are several words in Austronesian that mean dark that almost always have two syllables ending in -m. In Chinese there are two semantically related words, with only one syllable ending in -m: 暗 and 陰 (now pronounced -n in Mandarin). The following words come from a variety of Austronesian languages and vary in meaning from darkness, night, obscure, dark green/blue, black, dusk, cloud(y), dim, shade: lilim, limlim, lindum, linggom, malam, balam, bilem, dilum, urum, kulem, lomlom, dumdum, humhum, simdim, milem, tilem, ulem, veljelem, lulem, andum, jarem, merem, surem, tedem. In some languages, this -m ending may disappear or become -ng instead. If you're learning an Austronesian or Pacific language, a tip like this may help in acquiring vocabulary.

We can use these sound changes to help us recognize and acquire new vocabulary quickly. At first the vocabulary may look completely foreign to us, but by noticing the sound correspondences, we can easily remember new words.

A Sample Austronesian Evolution

One of the best ways to examine sound changes over time and space is to look at numbers. Let's look at the number four as it evolves throughout the Austronesian languages:

Proto-Austronesian: *ʃepat
Formosan languages: supat, ʃpat, səpat, səpats, pat, səptə, spat
Philippine languages: uppat, opat, əpat, apat, upat, ofat
Indonesian languages: papat, pat, opat, əmpat, ampɛ, apat, hat
Madagascar: efatra
Vietnam's Austronesian languages: pāˀ
Oceanic & Polynesian languages: vasi, paw, vā, fā, hā, ehā, whā

From this sample you can see how some languages drop s-, and how others morph the -p- into other realizations such as [h, wh, f, v] and it seems the farther afield you go (far out in the Pacific) the more change has happened. It's worth to note that all these letters are related to each other: they are all pronounced on the lips and referred to as labials, except [h] which has lost the rounding in the lips. So over time [h] is a possible reduction from underlying /p/.

A little bit of nasalization is happening in languages like Indonesian with the -m- inserted. This nasalization was not there originally, but in other language families, there may have originally been nasalization which then disappeared. Sometimes, you'll find one language retaining nasalization where others have dropped it, compare English "whence" and Polish "-kąd" both with nasalization in the same root position, but Czech "-kud" without it.

hʷ ɛ n s = the English word "whence" (h is rarely pronounced now)

k ɔ n t = kąd in the Polish word "skąd" meaning whence

k u - t = kud in the Czech word "odkud" meaning whence

A Sample Indo-European Evolution

Let's take a look at the word "four" among the Indo-European languages. Since English has /f/ in the word just like Samoan "fā" above, wouldn't it be interesting to see if all the other Indo-European languages follow a similar pattern of sound change like the Austronesian languages?

Proto-Indo-European: *kʷetwṓr
Indic: tsor, tʃar, sari, tʃari
Iranian: tsuppar, tsubur, tsalor, tʃəhar, tʃwar, tʃahar, tʃar
Armenian: tʃors

Albanian/Greek: katər, tɛsɛra
Baltic: tʃɛtri, keturi
Slavic: tʃitiri, tʃteri, ʃtiri, tʃetiri
Germanic: fyra, fir, fjor-, four
Italic: petor, patru, kwattor, kwatr-, katʀ
Celtic: pedwar, kʲer, kʲehirʲ

In the above examples, we don't see so much of the same happening as in the Austronesian examples primarily because of sound change coming from a different source. In this case the languages with /k/ are the oldest as /k/ changing to /tʃ/ is quite common (called palatalization). We can see this even in English when {c} is followed by {i, e} in the same root:

CAPtive > aCHIEVe, /kap/ > /tʃiv/

When we see that the number four in Indic, Iranian, Baltic, and Slavic languages all have [tʃ, ts], this is now very easy to remember. And it will apply to a great number of words in all these languages. In Greek, simply split the [ts] with a vowel in between, but drop the /ka-/ prefix.

The Germanic languages underwent a process similar to the Polynesian languages. The combination of /khw/ can lead to [kw] in one direction (Romance languages) and [hw] in another direction (Germanic languages). This [hw] eventually fuses together to [f]. In Māori (New Zealand), they are halfway there: some dialects pronounce whā [hwā], and some dialects pronounce it [fā] like the Samoans.

The /khw/ and /f/ variation can be seen in other language families as well. Compare Chinese 快 with Mandarin /khwaj/ and Cantonese /faj/. The Austronesian /hw/ and /f/ variation also shows up in Chinese 花 in Mandarin as /hwa/ and Cantonese as /fa/.

Palatalization in Chinese Languages

Another example of /k/ becoming /ts/ can be seen among the Chinese languages.

Remember when Beijing used to be spelled Peking? How about when Chongqing was spelled Chungking? Did they change the name, or was that just language undergoing change?

Mandarin underwent a palatalization change at the end of the Qing (Ts'ing) Dynasty, which is in recent history: 1911. But not everybody spoke the same way, so it was still unstable across Mandarin's many dialects. Since contact with the West was already common, the "Peking" spelling had already become standard. Once pinyin was adopted as a standard romanization, it was apparent that Mandarin's palatalization shift was stable and mature, so pinyin embraced it. Since Mandarin had no /b/, to make things easy, they re-wrote the original {p} as {b} instead. Mandarin /p/ sounds like Italian /p/, not English /p/ (there's a difference in the puff of air on that letter).

To summarize the Mandarin sound correspondences:

{k'} /kʰ/ /ʦʰ/ all become /ʦʰ/, written {q}, pronounced [ʨʰ]

{k} /k/ /ʦ/ all become /ʦ/, written {j}, pronounced [ʨ]

{h} /h/ /s/ all become /s/, written {x}, pronounced [ɕ]

The pronunciation of {g} in the English word "gigantic" actually follows the same palatalization rule and is quite universal among the Romance languages: Portuguese, Galician, Spanish, Catalan, French, Italian, Romanian.

More Sound Groups

Learn to recognize vocabulary by their roots and the various sound correspondences that exists therein. English already has a lot of data to learn from.

The Royal Lawyer Rule: r / l

This isn't an official linguistics rule, it's just a fun name to help you remember.

R and L are both "liquids", and they frequently interchange to create variations on meaning. Everything related to the law, and what is governed within a country, including reading (what only ancient scribes were able to do), all came from a single root: rex/lex. The following is a sampling of English words derived from this root (with other languages in parentheses).

R - X: rex

R - K: reckon, rich, right, correct, (rechnen, Recht, rector, rectus)

R - G: regency, regimen, region, reign, regulate, regular, rule, regal, (regere, regnum, regula, regina, regius)

R - Y: royal, real, royalty

L - X: lex, lexeme, lexis, lexical, lexicon, (lexikos, legein)

L - K: lect, collect, elect, lesson, lection, lector, lectern (eligere, lectio, lectus)

L - G: logic, (logikos), college, legacy, legate, legend, legible, legion, legal (legare)

L - Y: loyal, loyalty

L - W: law, lawyer

The Guard Yard Rule: k / g / h / y / w

As can be seen from the examples above in the Royal Lawyer Rule, there's a decline of /k/ all the way down to /g/ and /y/. There is an example where English "day" is spelled "dag" in Danish, but pronounced the same as in English -- because Danish has been evolving speech faster than their writing can keep up, which is also why Norwegians and Swedes have a hard time understanding them.

This root extends widespread throughout languages. The root is related semantically to the place that provides protection and derivations extend the meaning in various ways: fence, field, enclosure, town, citadel, fortification, house.

K - RT/D: garden, gird, girdle, girth, yard (garðr, gård, gaard, gjǫrð, Gürtel, garth, gort, jardin, gar̃das, город, град, grȃd, gród, गृह gṛhá, գերդաստան)

H - RT: horticulture, cohort > court, (hortus, χόρτος, hrad, jardin)

GW - RD/N: warn, ward, guard, garnish, aware, wary, war, wear, guarantee, warranty, ware, garage, hlāford > lord, regard, reward, guardian, (waar, var, wehren, verja, weer, vǫrn, garer, wahren, vara, warten, vǫrðr)

Interesting to note is that the Central Asian words "yurt" and "ger" may not have descended from a common root, but they have the same sounds meaning "house".

The Wagon Weight Rule: b / v / w - g / z / h

From the Indo-European root *weǵʰ- we get a large number of words that are easy to remember.

W - K: in Greek/Latin: ἔχος ‎ékhos, vectum, (transport)

W - G/Y: in Germanic: weigh, wagon, wicht, way, vagn, wain, (weight, transport)

V - H: in Latin/Indic: वहति vahati, vehō, vehere, vehicle (transport)

V - Z/S: in Balto-Slavic/Iranian: vežti, вести, wieść, vésti, وز vaz, վազ vaz, bazdan (lead, convey), Georgian გზა (gza) may have come from Armenian.

V - D: in Balto-Slavic: водить, wodzić, voditi (conduct)

V - D: in Germanic: wed, wage, wetten, veðja (wager, pledge, join)

It is also worth noting that Romance "peso" begins with a bilabial, and for those who know Slavic languages will be familiar with the relationship between h and s making this weighty word easy to remember.

Practice

Below is a list of vocabulary that means "five" from various languages. Try to group similar vocabulary together into families. As a hint, there are 9 language families represented (two language families are very similar) and each family has at least 2 languages represented.

  1. Worry about vowels last. Beware, in at least one language family here vowels can switch positions between consonants.
  2. Check how many syllables are in the words. Does the word have an extra syllable that matches the structure of words in other languages?
  3. Check the point of articulation in the mouth. Do they match another word? (Lips = p b f v m, Teeth = t d s r n l ɬ θ ð, Roof of Mouth = (ʃ ʒ) c ɟ j/i, Back of Mouth = k ɡ ŋ h x ɣ w/u). Normally to simplify things, I use a strategy of applying P-T-C-K for those four categories. I use W J for glides, and V for vowels. Add more detail (like nasals) if needed.
ammist
beʃ
biæs
sáhta
dimy
fondz
gataanu
hima
hing
ɬáno
mítánu
ntɬʼanu
ŋ
ŋa
øt
pãːtʃ
pet
pump
rima
rŋa
tfɨ
txʷə
u
vittâ
wet
xamʃa
zaptã

Evolution

Most of today's languages descended from just a few languages spoken thousands of years ago. For example there was one proto-language for all the Indo-European (Indo-Hittite) languages. There was one proto-language for the Uralic languages.

Click here to see the relationship between all the sounds found in Indo-European languages.

Does that mean there were fewer languages then than there are now?

No, it simply means that the other languages didn't survive. Languages die all the time. And with them disappear a long human history encoded in them.

The sounds in languages are like biological DNA because we can track how they mutate and change over time. From all the existing languages today, we can reconstruct what languages of the past sounded like. We can even predict how our own languages will sound in the future.

But we cannot use the letters of the alphabet to analyze sounds. Orthographic spelling can easily mislead us because they stand for different sounds in every language. Letters are too arbitrary. Instead we use phonetic symbols that actually look like letters, but they behave very differently from the letters we know in written language. A phonetic symbol tells us where our tongue touches, whether to add voice from the throat, and how much pressure we use to release the sound. This collection of phonetic symbols is used by linguists everywhere in the world and is called the International Phonetic Alphabet (IPA).

With the IPA, it's easy to track sounds across languages, and see how words change. Each language writes these sounds in their own ways and spellings, and that can be quite misleading. Spellings are just conventions used for native speakers to record what they say in an easy, un-scientific way. Spelling is supposed to be a rough approximation to sounds and ideas and easy enough for a child to learn.

In an upcoming post we'll discuss why spelling and using an alphabet creates handicaps to becoming fluent in a foreign language.

Answers

  • Siouan: zaptã (Lakota), sáhta (Osage), structure = TVPTV, TVKTV (Osage looks newer, Lakota looks conservative)
  • Indo-European: hing (Armenian), fondz (Ossetic), pump (Welsh), pet (Slovene), pãːtʃ (Nepali), structure = KVKK, PVTT, PVPP, PVT, PVTT
  • Austronesian: dimy (Malagasy), hima (Bunun), rima (Maori), structure = TVPV, KVPV, TVPV
  • Caucasian: tfɨ (Adyghe), txʷə (Kabardian), structure = TKV (both are TKV because /xʷ/ has fused into /f/ in Adyghe as described at the beginning of this article)
  • Uralic: vittâ (Inari Sami), wet (Khanty), øt (Hungarian) structure = PVTTV, PVT, -VT
  • Altaic: beʃ (Kyrgyz), biæs (Yakut), structure = PVT, PVT
  • Semitic: xamʃa (Neo-Aramaic Assyrian), ammist (Amharic), structure = KVPTV, VPVTT
  • Bantu: gataanu (Rwanda), ntɬʼanu (Xhosa), mítánu (Lingala), ɬáno (Northern Sotho), structure = KTVTV, TTVTV, PTVTV, -TVTV
  • Sino-Tibetan: ŋa (Burmese), rŋa (Tibetan), ŋ (Hakka), u (Mandarin), structure = -KV, TKV, K-, -V