Let’s Build a PIE Conlang 3: Hello Hubris My Old Friend

Part 1: Introduction and *h₁n̥gʷnís
Part 2: *wódr̥ and *péh₂wr̥

This is the extremely predictable post where I eat a little crow. The project as I had been building it up in the last two installments (and many, many pages of notes in between) became more stressful than fun and its time for a refresh.

In my infinite eyes-bigger-than-stomach wisdom, I got it into my head that I wanted to do this on doubleplus hard mode and try recreating an earlier stage of PIE that contained all the theories that I thought were interesting: it shouldn’t take a certified genius (which I am not) to point out that it’s really fucking difficult to reverse-engineer a coherent predecessor to a barely-functional theoretical model, and twice as hard to do that out of random parts that don't really fit together in the first place. With Frankenstein-like enthusiasm I created a shambling monstrosity that sparks no joy; time to put it out to pasture and chalk it up as a lesson learned.

So this post is going to be a step back and re-orientation, hopefully. Bear with me as I talk myself through the steps. Maybe writing it down for other people to read will make all the pieces fit together 

What Went Wrong

Scope creep! One slightly idiosyncratic choice (keeping consonantal laryngeals) invited friends (every other theory that made enough sense for me to go "yeah, that makes sense, I like it") and then they invited even more friends and suddenly this house party is out of control. I forced myself into greater and greater leaps of logic and wild justifications to keep all the plates spinning.

Rabbit-holing and what-if-ism took their cut as well; the issue with infinite fractal complexity is that it is, indeed, infinite - you can toggle and adjust forever and accomplish nothing, eternally trying to get it just right.

What to Do Now

First step is backburner the project (done) and re-assess (what I’m doing right here). I've got a decent shot at actually finishing something for Cursed Conlang Circus 4 this year, so whatever I choose to do with this particular project will be later on down the road.

Option A) Toss out everything and go back to basics with a nice, bog-standard post-Anatolian IE language with minimal hypotheticals. No going backwards in time, no major divergences from the standard reconstruction, play it by ear and see how is shakes out.

Option B) Take all the wacky theories I liked and apply them to an original conlang where they can co-exist in harmony, instead of forcing them to fit into the reconstruction.

Option C) Scuttle the entire thing and chalk it up as a wash. If I return to it, I return to it a good ways into the future.

All of these are viable. For Option A, it's basically a reset to where I was at the beginning of the year, tooling around with sound changes and going more on vibes than a concrete plan. For Option B I could easily just generate new roots for the wordlist I am already using and work from there. Option C is easiest, and the more I consider it probably the wisest. I've considered picking another protolanguage to evolve from, but I have a sinking suspicion that I will end up in a similar spot if I try that. It might just be that this particular methodology doesn't work with my brain, it might be that I am just under so much god-damn stress both external and self-inflicted that it just won't compute. 

 

What was I Trying to Do

Same as it was in the beginning: evolve a unique Indo-European branch with a signature language. The why is because it's a fascinating subject and I'm terrible at actually sitting down and actually churning out words, so having an already-extant word list & grammar as a starting point helps (this is called an a priori conlang, in the biz). It does lead to analysis paralysis via sound changes (boy does it ever), but it does dangle a nice big carrot labeled "you can transform Thing A into Thing B and it'll probably feel pretty satisfying when you do".

Exactly what this would look like has changed over time, which may very well be part of the problem; "do a thing" is not a particularly concrete goal. In its most recent form, I wanted to evoke the aesthetics and sounds of Ithkuil, Klingon, and the Caucasian languages - uvular consonants, lots of secondary articulations, few vowels, etc - for vibes reasons, because I find PIE as reconstructed to have a vibe-kinship with those. But I ended up kinda falling away from that - ditching the Klingon retroflexes and Ithkuil's ludicrous consonant clusters - so I don't know how useful it is as a goal anymore. I have a vague target phonology (see below) that I can achieve without keeping the laryngeals around (and, for the sake of my own sanity, should.

Damn it I forgot to put θ on there.

Vowels are 90% likely to just be a / i / u / ə. 

Theories I Want to Keep

• *e and *o were more like *ə~ɛ~æ and *ɑ
• The laryngeals were originally **ʔ/x~h, **χ, **ʁ, and later became *h, *x/ħ, *χ
• The plain velar series were back enough to turn into the uvulars *q, *ɢ and *ɢʰ

 

Dealing with the Laryngeals

Fuck 'em, too much trouble. The traces left behind are:

• Vowel-coloration and lengthening, as standard. Sometimes they will get nuked without lengthening, per a few varied sound laws.
• Proto-Balto-Slavic style "acute" glottalization of preceding vowels (which will later migrate to consonants to form the ejectives.)
• Will be deleted between consonants if it doesn't cause illegal clusters; will vocalize into *e *a *o if the result would otherwise be illegal.
• Will vocalize into *e *a *o when at beginning of words preceding a consonant, per Greek Triple Reflex.
• Cases where that's the unknown laryngeal *H will just be dropped, cause I can't be assed.   

 

Final Notes 

The advice is always step back and take a break when dealing with burnout, and that is something I have never, ever been good at. This shit doesn't turn off; it's why I still have over 300 pages of notes and drafts on gdocs alone. It just accumulates like pressure behind a damn. It certainly doesn't help that it's September 2025 and it's been a dogshit year even when you don't factor in The Everything.

Ah well. This project, at least, is going to get backburnered for now. Get it out of my drafts.