cripts die in silence. No funeral rites mark their passing; no mourners gather when the last person who can read a particular arrangement of lines and dots expires, taking an entire system of meaning with them. Of the approximately 400 writing systems devised by humanity, fewer than 30 remain in daily use. The rest exist only as marks on stone, scratches on clay, stains on papyrus—artifacts awaiting decipherment or, more often, cataloguing without comprehension.

Consider the Indus Valley script: over four thousand inscriptions survive on seals, tablets, and pottery from the Harappan civilization, dating to c. 2600–1900 BCE. Despite a century of computational analysis, pattern matching, and scholarly debate, these symbols have yielded no consensus translation. Each glyph is a locked room. Each seal is a message with no addressee.

The archive of archaic.works approaches these lost scripts not as puzzles to be solved but as objects to be contemplated. A writing system is a technology—an invention as significant as the wheel or the lever—and its death is a form of technological extinction. We gather the remnants here: the Phaistos Disc with its spiraling stamped symbols, the Cascajal Block with its Olmec glyphs, the Vinča signs from Neolithic Serbia that may or may not constitute true writing.

¶ Proto-Elamite, used in ancient Iran c. 3100–2900 BCE, is the oldest undeciphered writing system with a significant corpus.

Each entry in this section presents the script as it appeared in its native medium: carved, painted, stamped, or incised. The material matters. Cuneiform was shaped by the reed stylus pressed into soft clay; its wedge-shaped marks are inseparable from the gesture that produced them. Hieratic script evolved as a cursive shorthand of hieroglyphics, its flowing forms dictated by the brush moving across papyrus. To study a script without its substrate is to study a footprint without the ground.

rtifacts are the vocabulary of silence. Where written records fail, objects speak—or rather, they gesture. A corroded bronze fibula tells us that someone fastened a garment; a shard of Samian ware tells us that someone ate. The archaic artifact is simultaneously mute and eloquent, a paradox of material culture that resists the tidy narratives we impose upon it.

Below, we present a collection of objects recovered from the margins of history: tools whose purposes we can only guess, ornaments whose symbolism is lost, instruments whose music has fallen silent. Each artifact is documented with its provenance, material composition, and approximate date. What we cannot document is its meaning.

rafts perish differently from scripts. A writing system can survive in stone for millennia; a craft technique lives only in the hands that practice it. When the last master damascener dies without an apprentice, the art of inlaying gold wire into steel vanishes not from the historical record but from the realm of the possible. The knowledge becomes theoretical—we know that it was done but no longer how.

The distinction matters. archaic.works catalogs crafts along a spectrum of endangerment borrowed from conservation biology: Vulnerable (fewer than 100 active practitioners), Endangered (fewer than 20), Critical (fewer than 5), and Extinct (no living practitioners). The taxonomy is imperfect—how does one count practitioners of a craft that was never formally named?—but it provides a framework for urgency.

¶ The UNESCO Intangible Cultural Heritage list now recognizes over 700 craft traditions, but many thousands more remain undocumented and unprotected.

Among the entries you will find here: niello, the art of filling engraved metal with a black alloy of silver, copper, and lead, practiced from ancient Egypt through the Byzantine empire and now reduced to a handful of workshops in Tula, Russia. Cloisonné enameling, in which thin metal wires partition a surface into cells filled with powdered glass and fired—a technique perfected in the 6th century and now largely mechanized beyond recognition. Parchment making itself, the preparation of animal skins for writing, which requires knowledge of liming, scraping, stretching, and finishing that took medieval craftsmen years to acquire.

Each craft entry includes: a description of the technique, its historical range and peak period, the materials and tools required, the estimated number of current practitioners, and—where possible—a first-person account from someone who still practices the art. These testimonies are the most valuable content on this site. They are the voice of living memory before it becomes archaeological silence.

anguages carry worlds within them. Each tongue that falls silent takes with it not merely vocabulary and grammar but an entire architecture of thought—categories of color, concepts of time, kinship structures, botanical taxonomies, philosophical distinctions that exist in no other language. When Eyak died with its last speaker, Marie Smith Jones, in 2008, the world lost not just a medium of communication but a unique way of parsing reality.

The languages cataloged here are of two kinds: those that are dead (no living speakers, known only from written records or recordings) and those that are dormant (no native speakers, but undergoing revival efforts). The distinction between dead and dormant is a matter of political will as much as linguistic fact. Hebrew was dormant for nearly two millennia before its revival; Cornish was declared dead in 1777 and is now spoken by several hundred people. Death, in linguistics, is not always permanent.

¶ Sumerian, the oldest known written language (c. 3100 BCE), was already dead as a spoken tongue by 2000 BCE but continued as a scholarly language for another two millennia—a precedent for Latin's later trajectory.

We present each language with: its linguistic classification, geographic range, approximate dates of use, number of known texts or recordings, and a sample passage (with interlinear gloss where possible). For extinct languages without written records, we note only what can be inferred from loanwords, place names, and substrate effects in successor languages—the ghostly traces of a vanished tongue in the pronunciation and idiom of those that replaced it.

This is the palimpsest made literal: beneath every living language lie the erased traces of its predecessors, visible only to those who know how to look.