Half the people alive today speak a language descended from a single prehistoric ancestor: Proto-Indo-European (PIE). English, Spanish, Hindi, Russian, Persian, Bengali, German, French, Greek, Italian, Polish, Punjabi, Marathi — all of them, plus dozens of extinct languages like Latin, Sanskrit, Hittite, Tocharian, and Gothic — descend from a language spoken roughly 6,000 years ago on the Pontic-Caspian steppe (according to the dominant Kurgan hypothesis of Marija Gimbutas and David Anthony) or, in a competing theory, in Anatolia.
No one ever wrote PIE down. Yet linguists confidently quote PIE words: *ph₂tḗr "father", *méh₂tēr "mother", *wódr̥ "water", *ǵneh₃- "to know", *kʷékʷlos "wheel". How can they?
The breakthrough came in 1786, when the British judge Sir William Jones announced before the Asiatic Society of Bengal that Sanskrit, Greek, and Latin showed a relationship "so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists." Over the next century, Franz Bopp, Rasmus Rask, Jacob Grimm, and August Schleicher developed the comparative method: align cognate words across daughter languages, identify regular sound correspondences, and project them back to a reconstructed parent.
The most famous example is Grimm's Law, formulated in 1822, which describes the systematic shift of stop consonants in the ancestor of Germanic. PIE *p became Germanic *f: hence Latin pater ↔ English father, Latin piscis ↔ English fish, Latin pēs ↔ English foot. PIE *t became *þ (the th sound): Latin trēs ↔ English three. The correspondences are exception-free once secondary rules (Verner's Law, 1875) are added.
Then comes the moment that, for many linguists, settled the question of whether reconstruction is real science. In 1879, the young Swiss linguist Ferdinand de Saussure — twenty-one years old and not yet famous as the founder of modern semiotics — published Mémoire sur le système primitif des voyelles dans les langues indo-européennes. To make PIE vowel alternations work out cleanly, he proposed that PIE had contained mysterious additional coefficients sonantiques that had since disappeared in every known daughter language — sounds we now call laryngeals (*h₁, *h₂, *h₃).
"There must once have existed in Indo-European a phoneme A which has left no direct trace in the historical languages..." — Saussure, 1879
Saussure's proposal initially rested on internal reconstruction rather than direct attestation. After Bedřich Hrozný deciphered Hittite in 1915, the Polish linguist Jerzy Kuryłowicz argued in 1927 that the Hittite consonant written ḫ occurred in positions corresponding to some of the coefficients Saussure had proposed. This supplied direct comparative evidence for the laryngeal theory.
Today PIE reconstruction includes hundreds of well-established roots, a grammar conventionally reconstructed with eight cases and three genders, ablaut alternations, and inherited poetic formulas — phrases like *ḱléwos n̥dʰgʷʰitóm "imperishable fame", reflected in both Homeric Greek (kléos áphthiton) and Vedic Sanskrit (śrávas ákṣitam). The reconstruction remains a model rather than a recorded text, and individual forms continue to be revised as the evidence is reassessed.