How we know languages like proto-Indo-European existed [video]

Here is an alternative Piped link(s): https://piped.video/watch?v=jIgoBRbfkUA

Piped is a privacy-respecting open-source alternative frontend to YouTube.

I'm open-source, check me out at GitHub.

Around 16min or so, he mentions the correspondence of Germanic /f/ with Latin /p/. There's a related sound correspondence between Proto-Germanic (PG) /h/ and Latin /k/, that illustrates well how sound changes pile up over time (something that he mentioned earlier). Like this:

• PG *hundą "hundred" /h/ vs. Latin ⟨centum⟩ "hundred" /k/
• PG *haubudą~*hafud "head" /h/ vs. Latin ⟨caput⟩ "head" /k/
• PG *ahwō "river, waters" /h/ vs. Latin ⟨aqua⟩ "water" /k/¹

It's a rather simple relationship, right? Well, now look at the descendants of those words in English and in Spanish:

• English ⟨hundred⟩ /h/ vs. Spanish ⟨ciento⟩ "hundred" /θ/~/s/
• English ⟨head⟩ /h/ vs. Spanish ⟨cabo⟩ "end, edge, cable, cape" and ⟨cabeza⟩ "head" /k/
• English ⟨ea⟩ Ø vs. Spanish ⟨agua⟩ "water" /ɣ/

The correspondence went from /h/ vs. /k/ to /h Ø/ vs. /θ~s k ɣ/. You'll also see something similar with PG *θ vs. Latin /t/, specially if you use German instead of English (as German got rid of /θ/, English still kept it as /θ/ or /ð/).

EDIT: I'm a dumbarse and I should've watched the whole video before commenting. He just mentioned /k/ vs. /h/.

Also, note something: how much semantic drift also takes it toll. He mentioned English ⟨hound⟩ vs. Latin ⟨canis⟩, but if you were to simply look at the words for "dog", you wouldn't see any correspondence between ⟨dog⟩ and ⟨canis⟩.

---

25:40, on the words for "father": it's worth noting that the cognate of "atta" did survive in English, although reduplied. It's "dad". That "atta" is likely an old word that eventually got replaced with *ph₂tér in Late PIE.

Note that the spelling for later languages also provides some clues. Here it's shown with Irish, the fact that /h/ is spelled as ⟨th⟩ is meaningful here - it's from a lenited /t/. For another example of that, check my Spanish example above, why that /s/~/θ/ is spelled with the same ⟨c⟩ as /k/.

29:00 - the "typical" sound change sequence for [p] and [f] is usually something like [p]→[ɸ]→[f]→[h]→Ø.

30:00 - if I had to take an educated guess, PIE *r was likely *r~ɾ. Perhaps slightly retracted. This sort of alternation is extremely common, and Latin itself likely had [r ɾ] for /r: r/ (Italian keeps it; and even Spanish and Portuguese two rhotics can be still analysed as gemination.)

32:00 - even if it contradicts Ockham's Razor, I still believe that *h₂ and *h₃ were at least two consonants each, a voiceless-voiced pair. That explains the "odd" voicing in the root peh₃- "to drink" when reduplicated as *pibh₃- for the perfect aspect, without weird stuff like "fricatives were mostly voiceless, except that one way on the back of the mouth that is voiced".