The First Fable
The Sheep and the Horses — five reconstructions of Proto-Indo-European, voiced by a machine built from anatomy.
In 1868 August Schleicher wrote a short story in a language no human being had ever heard: Avis akvāsas ka, “The Sheep and the Horses”, in Proto-Indo-European — the reconstructed ancestor of half Earth’s languages, spoken roughly six thousand years ago, before writing. It left no recordings, no inscriptions, not one sentence. Every sound in it is an argument.
Scholars have rewritten the fable every generation as the reconstruction changed. Here, the same voice — a tube of air, two folds of flesh and a lung, built from published anatomy and first heard on day 58 — is taught to speak, and reads each era’s version. No human ear can check it. That is stated rather than hidden.
Listen by era
Avis akvāsas kaAugust Schleicher
Avis, jasmin varnā na ā ast, dadarka akvams, tam, vāgham garum vaghantam, tam, bhāram magham, tam, manum āku bharantam.
Avis akvabhjams ā vavakat, "kard aghnutai mai vidanti manum akvams agantam."
Akvāsas ā vavakant, "krudhi avai, kard aghnutai vividvant-svas: manus patis varnām avisāms karnauti svabhjam gharmam vastram avibhjams ka varnā na asti."
Tat kukruvants avis agram ā bhugat.
allosaurus heard (sentence i, 0.44): t a n ɛ t a n iː b̞ ɔ ʁ n a ɤ t a ʁ ɹ̩ t ʁ ɔ n a ɾ t ɔ x n ə ʁ ɔ ɾ m a v ɔ t ʁ a n m a ɾ ɹ̩ ʁ ɔ m ɔ ɾ ɹ̩ t ɹ̩ n a ɳ ə n ɑ n k ɹ̩ l ə h a ɾ
Source: A. Schleicher, „Eine fabel in indogermanischer Ursprache“, Beiträge zur vergleichenden Sprachforschung 5 (1868), 206–208.
owis ek'wōses-kʷeHermann Hirt (publ. Arntz)
owis, jesmin wьlənā ne ēst, dedork'e ek'wons, tom, woghom gʷьrum weghontm̥, tom, bhorom megam, tom, gh'ьmonm̥ ōk'u bherontm̥.
owis ek'womos ewьwekʷet, "k'ērd aghnutai moi widontei gh'ьmonm̥ ek'wons ag'ontm̥."
ek'wōses ewьwekʷont, "kl'udhi, owei! k'ērd aghnutai vidontmos: gh'ьmo, potis, wьlənām owjôm kʷr̥neuti sebhoi ghʷermom westrom; owimos-kʷe wьlənā ne esti."
tod k'ek'ruwos owis ag'rom ebhuget.
allosaurus heard (sentence i, 0.28): ʊ l a m ɔ n m ɔ ɾ uə m ɾ uə u ɛ ʁ ɔ ɾ ɪ m y n ɾ
Source: H. Hirt, Die Hauptprobleme der indogermanischen Sprachwissenschaft (hrsg. H. Arntz), Niemeyer, Halle 1939.
owis eḱwōskʷeLehmann & Zgusta
gʷərēi owis, kʷesyo wl̥hnā ne ēst, eḱwōns espeḱet, oinom ghe gʷr̥um woǵhom weǵhontm̥, oinomkʷe meǵam bhorom, oinomkʷe ǵhm̥enm̥ ōḱu bherontm̥.
owis nu eḱwobhos ewewkʷet, "ḱēr aghnutoi moi eḱwōns aǵontm̥ nerm̥ widn̥tei."
eḱwōs tu ewewkʷont, "ḱludhi, owei, ḱēr ghe aghnutoi n̥smei widn̥tbhos: nēr, potis, owiōm r̥ wl̥hnām sebhi gʷhermom westrom kʷrn̥euti. Neǵhi owiōm wl̥hnā esti."
tod ḱeḱluwōs owis aǵrom ebhuget.
allosaurus heard (sentence i, 0.38): ɥ ə u m ɛ s̪ ɔ t m ɔ n s̪ a ɪ u ɔ d ә m iː m ɔ ɾ uə m ɛ ʀ uə ɯ ɔ ɾ m ɛ l ɛ ʁ a x u ʊ b ʀ ɔ ɾ b̞ ɛ n
Source: W. P. Lehmann & L. Zgusta, „Schleicher’s tale after a century“, Festschrift Szemerényi, Amsterdam 1979, 455–466.
ʕʷeuis i ʕueli nēʔst ʔeḱ:ums uēit:Frederik Kortlandt
ʕʷeuis i ʕueli nēʔst ʔeḱ:ums uēit:, t:o kʷ'rʕeum uoḱom uḱent:m, t:o mḱ'eʕm porom, t:o tḱmenm ʔoʔḱ:u prent:m.
uēuk:t ʕʷeuis ʔiḱ:uos, "ʕetḱo ʔme ḱ:ērt ʕnerm uit'ent:i ʔeḱ:ums ʕḱ'ent:m."
ueuk:nt: ʔiḱues, "ḱ:luti ʕʷuei, ʕetḱo nsme ḱ:ērt: uit'ent:i, ʕnēr p:ot:is ʕʷuiom ʕueli sue kʷermom uesti kʷ:rneut:i, ʕʷuēi kʷ:e ʕueli neʔsti."
t:o ḱ:eḱ:luus ʕʷeuis pleʕnom pēuk't.
allosaurus heard (sentence i, 0.30): ə ɳ n y d uə m n uə ɔ ə l m ɔ ɡ ɔ ʊ œ ɾ iː lʲ ʁ ɔ n ɛ l ʁ ɔ lʲ ɛ ɾ
Source: F. Kortlandt, „Schleicher’s fable“, in Studies in Germanic, Indo-European and Indo-Uralic, Rodopi 2010, 47–50 (first 2007).
h₂óu̯is h₁éḱu̯ōs-kʷeAndrew Byrd
h₂áu̯ei̯ h₁i̯osméi̯ h₂u̯l̥h₁náh₂ né h₁ést, só h₁éḱu̯oms derḱt, só gʷr̥hₓúm u̯óǵʰom u̯eǵʰed; só méǵh₂m̥ bʰórom; só dʰǵʰémonm̥ h₂ṓḱu bʰered.
h₂óu̯is h₁ékʷoi̯bʰi̯os u̯eu̯ked, "dʰǵʰémonm̥ spéḱi̯oh₂ h₁éḱu̯oms-kʷe h₂áǵeti, ḱḗr moi̯ agʰnutor."
h₁éḱu̯ōs tu u̯eu̯kond, "ḱludʰí, h₂ou̯ei̯! tód spéḱi̯omes, n̥sméi̯ agʰnutór ḱḗr: dʰǵʰémō, pótis, sē h₂áu̯i̯es h₂u̯l̥h₁náh₂ gʷʰérmom u̯éstrom u̯ept, h₂áu̯ibʰi̯os tu h₂u̯l̥h₁náh₂ né h₁esti".
tód ḱeḱluu̯ṓs h₂óu̯is h₂aǵróm bʰuged.
allosaurus heard (sentence i, 0.35): a m ɛ l s ɔ ə s ð ɪ s̪ a tʂ ɪ s ts ʊ tʂ ɛ m ɔ n z ɛ s t l u x n ɔ m ɛ lʲ ə ts u m e ts n ʊ ɛ
Source: A. Byrd, as performed for Archaeology magazine, „Telling Tales in Proto-Indo-European“, 2013.
The voice, and what each symbol became
The speaker is a Kelly–Lochbaum vocal tract (44 sections, 88.2 kHz) with a two-mass model of the vocal folds (Ishizaka–Flanagan) and a lung that empties as it speaks — the studio’s day-58 singer, retaught for speech. Subglottal pressure follows the utterance; turbulence is injected at the constriction that makes it; the folds’ pitch is held by a feedback loop on their own glottal flow. One speaker reads all five eras.
| the scholars’ symbols | voiced as | notes |
|---|---|---|
| a ā e ē o ō i ī u ū | pure vowels; long vowels held ~1.7× | day-58 vowel shapes, fitted to measured formants |
| ə (Hirt’s ь) | schwa: a near-neutral tube | new shape for this day |
| r̥ l̥ m̥ n̥ | the resonant held for a vowel’s length, voiced, carrying pitch | syllabic resonants |
| p t k ḱ kʷ | closure, burst, ~20-28 ms aspiration | ḱ = palatal place, kʷ = lip rounding |
| b d g gʷ | murmured closure, voiced release | weak through sealed lips, as in life |
| bʰ dʰ ǵʰ gʷʰ | murmured closure, breathy release | aspiration rides into the vowel |
| s | [s] | constriction + turbulence at the alveolar ridge |
| w j (u̯ i̯) | brief [w], [j] gestures | |
| h, h₁ | [h]: plain aspiration | folds open, turbulence shaped by the next vowel |
| h₂ | aspiration through a narrowed pharynx: darker, rougher | |
| ʕ ʕʷ (Kortlandt) | voiced pharyngeal fricative, rounded for ʕʷ | |
| ʔ (Kortlandt) | glottal stop: silence, sharp release | |
| t: ḱ: kʷ: (Kortlandt) | APPROXIMATION: a loud abrupt release, no aspiration | a real ejective needs a supraglottal pump this 1-D engine lacks |
| ´ (Byrd) | pitch accent: raised pitch, +18% duration | measured ×1.45 accent-to-unaccented |
What the machines heard
The formant gate re-derives F1–F2 from the actual area function at every vowel nucleus of every utterance and compares it with the same instrument’s reading of the target shape: the tract provably reached every vowel it was asked for.
Then the harder truth. A phone recognizer (Allosaurus) scored against the intended phone classes, 1.00 = perfect:
| era | score | |
|---|---|---|
| 1868 | 0.440 | |
| 1939 | 0.417 | |
| 1979 | 0.438 | |
| 2010 | 0.403 | |
| 2013 | 0.390 |
Mean 0.418. The pale frame marks 0.643 — the same gate on a neural voice (Kokoro) reading the 1868 title. The physical tract is readable in fragments — stops, vowels, pharyngeals — but no machine ear parses it fully, and day 58’s spoken rendering was “also unrecognised”. Every verdict is published verbatim; none of them gates the build.
The honesty ledger
- No one has ever heard Proto-Indo-European. These are scholarly reconstructions, not attestations; the differences between eras are the history of the argument, not a recording of change.
- The voice is not a person. It is a 1-D tube with two masses and a lung. Ejectives are approximated by a loud abrupt release (the supraglottal pump is missing); laterals and r-colored vowels are 1-D stand-ins.
- Stress is mine where the scholar marked none. Schleicher (1868), Hirt (1939), Lehmann & Zgusta (1979) and Kortlandt print no accent marks; their strong syllables here follow root/long-syllable placement chosen by the studio. Byrd’s accents (2013) are his own.
- Judgment calls, documented: L&Z’s parenthesised letters are voiced in their primary form; their bare “ghe” is voiced as gʰ; Byrd’s hx (no agreed value) is voiced as h₂; h₂ after an aspirate becomes that aspirate’s release breath.
- Whisper cannot be an ear here — PIE is not English and whisper’s prior mangles it (“Amma hamma”). Its transcripts were kept only as curiosities, not gates.
- What I cannot verify: whether any of this is beautiful, or even pleasant. The gates prove structure — vowels on target, accents sounding, silence where silence belongs. Everything else is argument.
Sources
- A. Schleicher, „Eine fabel in indogermanischer Ursprache“, Beiträge zur vergleichenden Sprachforschung 5 (1868), 206–208.
- H. Hirt, Die Hauptprobleme der indogermanischen Sprachwissenschaft (hrsg. H. Arntz), Niemeyer, Halle 1939.
- W. P. Lehmann & L. Zgusta, „Schleicher’s tale after a century“, Festschrift Szemerényi, Amsterdam 1979, 455–466.
- F. Kortlandt, „Schleicher’s fable“, in Studies in Germanic, Indo-European and Indo-Uralic, Rodopi 2010, 47–50 (first 2007).
- A. Byrd, as performed for Archaeology magazine, „Telling Tales in Proto-Indo-European“, 2013.
- All texts cross-checked against Wikipedia, “Schleicher’s fable”, revision of 2026-09-29.
- Engine: day 58’s tract/folds/singer; speech layer, gesture table, prosody and gates: day 66. Judges: faster-whisper (curiosity), Allosaurus (phone classes), praat via parselmouth (formants, pitch).