Indo-European

Question: What methods do linguists use to reconstruct proto Indo European words from languages like Old English and Sanskrit? How do they translate these words into their original form?
https://www.quora.com/What-methods-do-linguists-use-to-reconstruct-proto-Indo-European-words-from-languages-like-Old-English-and-Sanskrit-How-do-they-translate-these-words-into-their-original-form/answer/Hans-Georg-Lundahl-2

Hans-Georg Lundahl

amateur linguist

4.X.2024

St. Francis of Assisi

It’s not a question of translating.

It’s a question of showing the sound changes.

If I wanted to reconstruct oculum from Italian occhio, Spanish ojo, French oeil, I would probably try out sound changes such that from oculum it leads to each of these. If I did that backwards, I’d not be able to construct “-um” but more like “-o” but I would probably get sth like *oklo.

In order to do that backwards, I’d like to start with a word list.

occhio, orecchio, naso, bocca
ojo, oreja, nariz, boca
oeil, oreille, nez, bouche

I see that -cchi- between vowels in Italian corresponds to Spanish -j- and French yod written with an -il group (word final or not). I see that in French this but not a following simple -r- will affect the quality of the o, and also note that French has ou for o in bouche, I’ll presume that o became ou because the following consonant was not simple, I note that French final silent -e corresponds to Spanish and Italian -a in bocca, boca, bouche, and to Spanish, but not Italian -a in orecchio, oreja, oreille. I’ll presume that in Italian orecchio was originally orecchia (and I would be wrong, oreja and oreille come from an original plural form, while the singular was auriculum).

I’ll wager that -cchi- / -j- / -il(l)- go back to a group like “kl” … even if it’s a bit risky (and I would be right, but that’s why I’m not sure how I could have figured it out without knowing the Latin).

I’ll also wager that French oreille with feminine ending -e corresponds to what in Italian would be orecchia, this means every “majority -o” word has lost the -o in French.

Back-track this into the words:

oklo, oreklo, naso, bocca
oklo, orekla, narizo, boca
oklo, orekla, nezo, bocha

Wait, Spanish wasn’t losing final -o, so what is “narizo” doing with one?

I’ll wager that “narizo” is actually another word and doesn’t quite belong. I’ll also wager that, as French has a different vowel from o- in bouche, but the same in oreille, bouche originally had a group making the first syllable closed, like a geminated kk sound.

oklo, oreklo, naso, bokka
oklo, orekla, — , bokka
oklo, orekla, nezo, bokka

Finally, I’ll imagine that -s- between vowels became -z- in French before French lost the final -o, and that a- becomes e- in French, perhaps under conditions like for bouche.

oklo, orekla, nasso, bokka
oklo, orekla, — , bokka
oklo, orekla, nasso, bokka

= oklo, orekla, nasso, bokka.

Let’s test nasso becomes nezo, becomes nez … If the -z- referred to a sound like zebra in English, improbable. But if it was more like -ts- it could stand.

oklo, orekla, nasso, bokka to French:
1) get -kl- to -gl- and geminates to affricates:
oglo, oregla, natso, bokkha
2) turn -gl- to -ly- and change vowel before as affricates:
olyo, orelya, netzo, boukkha
3) turn off final -o and change kkh to achlaut
oly, orelya, netz, boukha
4) turn oly to öly, achlaut to sh, -a to -e
öly, orelye, netz, boushe
5) turn ly to yy and make -e silent
öyy, oreyy, netz, boush
6) turn netz to néé by way of -s, -z (like English zebra), -zero or by way of -s, -hh, -zero
öyy, oreyy, né, bouch
7) fix the spelling by presuming some got written down further back:
oeil, oreille, nez, bouche

Now, let’s compare with another word.

occhio, orecchio, naso, bocca, passo
ojo, oreja, nariz, boca, paso
oeil, oreille, nez, bouche, pas

Oops, my reconstruction of “nez” as “nasso” was wrong, very wrong. “Nasso” would have become “nas” and not “nez” … let’s start all over. This time, the -as- is different by having a long a- instead of a long -s- for nose.

oklo, orekla, naaso, bokka, passo

Apart from there being no general geminate to affricate rule, apart from -kka becoming -che because of the -a sound, like caballum became cheval, apart from all long aa becoming æ or éé, and apart from -s- in naso instead going the path of -z- (as in English zebra) … which is pretty many exceptions, the first reconstruction was basically right.

Within a presumed “branch of” Indo-European, this may be pretty straightforward. Latin to Italian or to French, that’s lots of words in common, most constructions still extant (except those that depended on presence of six cases or absence of the Romance future and conditional, and sometimes a construction with conditional can be replaced by an older one.

Between the branches, it is far less straightforward. Remember that irregularity between Italian orecchio and the rest? Remember deciding to skip nariz in the comparison? There is far more of that. There are some twelve known “branches of Indo-European”, including extinct ones, and many words considered Indo-European come only in three of them. After going through 36 Indo-European reconstructed words, I have 5 extant in two extant branches, 11 in three branches, in some cases plus extinct ones, six in four groups sometimes plus extinct, and plus Albanian once, 7 in five branches, also once including Albanian, 4 in six groups, 1 in seven groups, 1 in eight groups, 1 in nine groups, 1 in ten of them. I treated Albanian as not an extant group, since it could be influenced by Greek or Slavic or Romanian, which have been neighbouring languages. If you find a word in Albanian, it could be from there.

So, this situation could also occur when different languages at different times were neighbouring each other, and that might in some cases simplify the sound changes. For instance why Indo-European yod becomes initial h- and initial dz- in Greek (ho, dzugon supposedly from yo, yugom). My idea behind the check up was to see if the words would be evenly distributed between the groups or cluster very much, though the latter could also mean intermediate larger language families, it could also mean a series of consecutive language unions, Sprachbünder. The one created by Yamnaya invasion would just be one of them and not the first. I am not the first in linguistics, and not limited to amateurs like myself, who has preferred the Balkan model (confer what I said about Albanian) over the Romance model for what “Indo-European” really is.

By the way, the “nasso” thing was a deliberate plausible error in the guesswork to show the process is not straightforward even within a group like Romance.

Hans-Georg Lundahl

PIE vs Sprachbund Project : Intentions and Theorems, Continued
https://pokorny-vs-trubetskoy.blogspot.com/2022/06/intentions-and-theorems-continued.html

If you are interested, about the same question was posed me about a year ago, and it's on quora as well as on this blog, Q I of the message: Language Related, Again