A Fable in a Language with No Surviving Texts

In 1868, August Schleicher published a story in a language for which no ancient text survives.

It begins with a sheep that has no wool. The sheep sees horses doing different kinds of work for a man and feels sorry for them. The horses answer that the sheep has more reason to grieve: the man turns a sheep's wool into a warm garment, while the sheep is left bare. The sheep hears this and runs away.

Schleicher called it Avis akvāsas ka, "The Sheep and the Horses." He had not discovered the story on a tablet. He wrote it as an experiment in Proto-Indo-European, the reconstructed ancestor of languages that now stretch from English and Irish to Greek, Persian, and Hindi.

The fable has been rewritten many times since. That is not a failure. Each version is a dated map of what linguists thought they could recover.

The asterisk before the sheep

A modern reconstruction of the word "sheep" is often written *h₂ówis.

Every mark in that compact form carries a warning or a claim:

• The asterisk means the word is reconstructed, not directly attested.
• The h₂ represents one of the lost consonants conventionally called laryngeals. Its exact pronunciation is uncertain.
• The remaining shape connects with forms such as Latin ovis, Sanskrit ávi-, and English ewe.

So *h₂ówis does not mean "this is exactly what one prehistoric speaker said." It means "this form best explains a network of regular correspondences in the surviving languages."

The fable's horses offer another example: *h₁éḱwos, "horse," stands behind forms including Latin equus and Sanskrit aśva-. The match is not based on a vague resemblance between two words. Linguists compare whole systems of sound changes, grammatical endings, and vocabulary across related languages. A proposed ancestor has to explain many descendants at once.

That is the comparative method. It works backward, but it does not work by free association.

Why the fable keeps changing

Schleicher's opening was Avis, jasmin varnā na ā ast, dadarka akvams: "A sheep that had no wool saw horses." The sentence looks relatively friendly to anyone who knows Latin or Sanskrit because Schleicher's reconstruction leaned heavily on those languages.

Later linguists changed the vowels, consonants, endings, and even the sounds they believed had disappeared before most Indo-European languages were written down. The modern sheep acquired h₂. The horse acquired h₁ and a palatal consonant written ḱ. The tidy nineteenth-century sentence became stranger as the model grew more exact.

That oddity is useful. A reconstruction should not be rewarded for looking familiar to a modern reader. It should account for the evidence.

The fable also exposes the difference between reconstructing words and reconstructing a world. Shared vocabulary can tell us that speakers inherited terms for sheep, horses, kinship, and many other things. It does not prove that Schleicher's little scene happened, or that every object and social arrangement in the story belonged to one precise period of Proto-Indo-European. The tale is a linguistic demonstration, not an eyewitness report from prehistory.

The consonants predicted before they were found

The most dramatic revision came from sounds that were first proposed as algebra.

In 1878, the young Ferdinand de Saussure argued that lost elements had once shaped Indo-European vowel patterns. He did not have an inscription containing those sounds. He inferred them from gaps and alternations in the daughter languages. Other scholars later developed the proposal into what became known as laryngeal theory.

Then Hittite entered the comparison. The language, preserved on cuneiform tablets from Anatolia, contained a consonant written with the sign scholars transliterate as ḫ in positions that matched some of the sounds the theory had predicted. In 1927, Jerzy Kuryłowicz made the correspondence explicit.

Hittite did not settle the exact pronunciation of every reconstructed laryngeal, and specialists still debate their number and phonetic values. It did something more important than making the theory neat. It supplied new evidence from outside the set of languages on which the original proposal had been built.

A structure inferred from vowel patterns had left a consonantal trace on clay.

What a reconstructed language sounds like

Linguists can pronounce Proto-Indo-European reconstructions, and Andrew Byrd of the University of Kentucky has recorded reconstructed fables. The result is not a recording recovered from the past. It is a model made audible.

To read a form such as *h₂ówis aloud, a speaker must choose values for symbols whose precise sounds remain uncertain. The rhythm, clusters, and syllable structure also depend on a particular analysis. A recording can reveal what that analysis implies. It cannot make the analysis uniquely true.

That distinction is the pleasure of the exercise. The fable lets us hear scholarship changing. Schleicher's avis and a modern *h₂ówis are not rival magic spells. They are stages in an argument, separated by Hittite tablets, better sound laws, and more than a century of criticism.

No ancient scribe left us a Proto-Indo-European paragraph. What survives is distributed across its descendants. Latin ovis, Sanskrit ávi-, English ewe, and many other forms each preserve part of the path. Grammar and sound correspondences bring those pieces into focus, while the asterisk keeps the uncertainty visible.

PRAVIEL is built around that same habit of reading closely: not treating an ancient form as a decorative relic, but asking what every sound, ending, and scholarly mark allows us to know.

Sources

• Keith Johnson, The History of Early English, Schleicher's 1868 fable and translation (Routledge)
• University of Kentucky, "Fables of the Reconstruction: Andrew Byrd"
• Linguistics Research Center, University of Texas at Austin, "The Laryngeal Theory"
• Linguistics Research Center, University of Texas at Austin, Indo-European reflexes of "ewe, sheep"
• H. Craig Melchert, "Practice Reconstruction" (UCLA Linguistics)