Source Count: 16 | Weighted Score: 28 | Source Confidence: [3/5] | Primary Tier: 2 | Last Updated: Apr 12, 2026
Keywords: decipherment, rosetta-stone, champollion, linear-b, michael-ventris, hieroglyphs, cuneiform, undeciphered-scripts, maya-glyphs, rawlinson
Category Tags: decipherment, writing-systems, epigraphy, linguistic-history
Cross-References: ZG_1_18 — Sound Symbolism · ZG_3_17 — Historical Linguistics · A_1_01 — Foundations Overview

QUICK SUMMARY

The history of decipherment — the recovery of lost writing systems and languages — represents some of the greatest intellectual achievements in the humanities, revealing entire civilizations whose written records had been unreadable for centuries or millennia. KEY FINDING The pivotal breakthrough was Jean-François Champollion's decipherment of Egyptian hieroglyphs in 1822, using the Rosetta Stone (a trilingual decree from 196 BCE in hieroglyphic, Demotic, and Greek, discovered by French soldiers near Rashid, Egypt in 1799). Champollion's key insight — announced in his famous Lettre à M. Dacier (September 27, 1822) — was that hieroglyphs were not purely symbolic or pictographic (as Thomas Young and others had assumed) but represented a mixed system: some signs were logograms (representing words), while others were phonetic (representing sounds, including consonants and syllabic groups). By matching the cartouches of royal names (Ptolemy, Cleopatra, Ramesses) between the Greek text and hieroglyphic text, Champollion established phonetic values, then extended the system to read the entire script. The second great decipherment was Henry Rawlinson's reading of Old Persian cuneiform (1837–1846, using the trilingual Behistun inscription of Darius I, carved on a cliff face in Iran ~520 BCE — Old Persian, Elamite, and Akkadian/Babylonian). Rawlinson's work opened the door to the decipherment of Mesopotamian Akkadian by Edward Hincks and others by the 1850s. The third great triumph was Michael Ventris's decipherment of Linear B (1952) — a syllabic script used for palatial record-keeping in Mycenaean Greece (~1450–1200 BCE). Ventris, an architect by profession, demonstrated through systematic cryptographic analysis that Linear B encoded an early form of Greek (Mycenaean Greek), 500 years older than Homer — a finding that stunned scholars who had assumed the script was non-Greek. Yuri Knorozov's decipherment of Maya glyphs (1952: proposed that Maya writing was logosyllabic, with signs representing syllables and logograms — initially rejected by Western Mayanists, particularly J. Eric S. Thompson, but vindicated by the 1970s–1980s through the work of Linda Schele, David Stuart, and others) opened an entire civilization's historical record. Major scripts that remain undeciphered include: the Indus Valley script (~2600–1900 BCE, ~400 signs, ~4,000 inscriptions — language unknown), Proto-Elamite (~3100–2900 BCE), Linear A (Minoan Crete, ~1800–1450 BCE), and the Rongorongo script of Easter Island.

1. VERIFIED CLAIMS (Tier 1 — Peer-Reviewed / Established)

• KEY FINDING Champollion and hieroglyphs: the Rosetta Stone (now in the British Museum) provided the bilingual key. Thomas Young (1814–1819) identified several phonetic hieroglyphs in Ptolemy's cartouche but believed most signs were ideographic. Champollion (1822) proved that the phonetic principle extended throughout the script — hieroglyphs could represent one, two, or three consonants, with logograms (determinatives) aiding word identification. His Précis du système hiéroglyphique (1824) systematized the grammar and sign list. This unlocked 3,000 years of Egyptian textual records.
• Rawlinson and cuneiform: Henry Rawlinson independently deciphered Old Persian cuneiform between 1837–1846, using the Behistun inscription. He confirmed Georg Friedrich Grotefend's earlier partial decipherment (1802, identifying the names of Darius and Xerxes). The trilingual Behistun text — analogous to the Rosetta Stone — then enabled the decipherment of Elamite and Babylonian/Akkadian cuneiform. In 1857, the Royal Asiatic Society conducted a famous test: Rawlinson, Hincks, Jules Oppert, and William Fox Talbot independently translated the same previously unseen Akkadian inscription and produced substantially similar results — confirming the decipherment's validity.
• Ventris and Linear B: Michael Ventris (1952) announced his decipherment in a BBC radio broadcast (July 1, 1952) and published the landmark paper with John Chadwick in the Journal of Hellenic Studies (1953). Ventris used a grid-based method to identify patterns in sign frequency, co-occurrence, and spelling alternation, hypothesizing that Linear B tablets from Knossos and Pylos recorded an early form of Greek. The decipherment was confirmed by the reading of new tablets unearthed at Pylos, where predicted words appeared correctly.
• Maya decipherment: Yuri Knorozov (1952, Sovetskaya Etnografiya) proposed that Maya glyphs were logosyllabic — combining logograms (word signs) with syllabic signs in a system analogous to Japanese. This approach was rejected by Thompson (who insisted glyphs were ideographic and that Maya writing did not encode spoken language) but was vindicated by Tatiana Proskouriakoff (1960: demonstrated that Piedras Negras inscriptions recorded historical events and royal lineages, not astronomical cycles) and by the phonetic readings achieved by David Stuart, Linda Schele, and others from the 1970s onward. By the 1990s, ~85% of Maya glyphs could be read.

2. CREDIBLE CLAIMS (Tier 2 — Academic / Debated but Supported)

• Undeciphered scripts: the Indus Valley script (Harappan civilization, ~2600–1900 BCE) remains undeciphered — short inscription length (average ~5 signs), unknown language family, and the absence of a bilingual text are major obstacles. Farmer, Sproat, and Witzel (2004) controversially argued that the Indus signs are not a writing system at all but non-linguistic symbols — a claim rejected by most specialists (Parpola, 2009; Wells, 2015). Linear A (Minoan, ~1800–1450 BCE) uses many signs shared with Linear B (suggesting similar phonetic values), but the underlying language (Minoan) is unknown, preventing translation even when individual signs can be "read."
• Computational approaches: machine learning and statistical methods are being applied to undeciphered scripts. Luo et al. (2019, Proceedings of ACL) used neural sequence-to-sequence models to assist in deciphering Ugaritic (a known script, as a test case) and Linear B. Revesz (2020) applied entropy and Zipf's law analysis to argue that the Indus script has linguistic properties. These approaches show promise but have not achieved breakthroughs on truly undeciphered scripts.
• Rongorongo: ~25 wooden tablets with incised glyphs from Easter Island (Rapa Nui). Writing direction is reverse boustrophedon (alternating left-right, with upside-down lines). The script has not been deciphered — there is no bilingual text, the language (Rapanui) is not certainly the inscribed language, and the corpus is small (~14,000 glyphs total, many repetitive). Researchers believe it is not a true writing system but a mnemonic system.

3. SPECULATIVE CLAIMS (Tier 3 — Possible but Unverified)

• Whether AI/machine learning will achieve decipherment of Linear A, Proto-Elamite, or the Indus script in the near future is uncertain — the fundamental obstacle (unknown language) is not easily overcome by statistical methods alone.
• Whether undiscovered bilingual inscriptions exist (e.g., an Indus-Mesopotamian bilingual) is plausible but unconfirmed.

4. DUBIOUS CLAIMS (Tier 4 — No Credible Source / Contradicted by Evidence)

• Claims of decipherment of the Indus script that have not been peer-reviewed or replicated. Numerous proposed "decipherments" have been published, attributing the script to Sanskrit, Dravidian, and other languages — none have gained scholarly consensus.
• Claims that the Phaistos Disc (a unique Minoan artifact with stamped symbols) has been deciphered. With only one specimen and ~240 signs from 45 distinct types, statistical constraints make reliable decipherment impossible.

Counter-Arguments & Criticisms

The Indus Script Debate: Writing System or Non-Linguistic Symbols?

• Steve Farmer, Richard Sproat, and Michael Witzel (2004, EJVS) argued that the Indus signs are not a writing system at all but non-linguistic symbols (comparable to heraldic devices, potter's marks, or religious icons) — key evidence: the average inscription length is only ~5 signs, far shorter than any known writing system; no inscription exceeds 26 signs; and the corpus shows no evidence of the syntactic patterns expected in written language
• Asko Parpola (Deciphering the Indus Script, Cambridge UP, 1994) and Bryan Wells (2015) reject the Farmer-Sproat-Witzel thesis, arguing that the signs show combinatorial regularities consistent with language encoding, that the short inscription length reflects the functional context (seal stamps on trade goods), and that Dravidian language provides the best candidate for phonetic values
• This debate remains unresolved and illustrates a fundamental problem in decipherment: without a bilingual text or a known candidate language, the question of whether a sign system is writing cannot be answered definitively from internal analysis alone

The Thompson-Knorozov Controversy and Ideological Bias in Scholarship

• J. Eric S. Thompson (1898–1975), the dominant figure in Western Maya studies for decades, insisted that Maya glyphs were ideographic (representing ideas, not sounds) and that Maya writing did not encode spoken language — he actively suppressed Yuri Knorozov’s 1952 phonetic reading, dismissing it as Soviet propaganda during the Cold War
• Thompson's authority delayed the correct decipherment of Maya by approximately two decades; the phonetic approach was only widely accepted after Tatiana Proskouriakoff (1960) demonstrated that inscriptions recorded historical events and after younger scholars (Linda Schele, David Stuart, Peter Mathews) systematically confirmed Knorozov's syllabic readings in the 1970s–1980s
• This episode is a cautionary example of how institutional authority and political context can impede legitimate scholarly progress in decipherment

Limitations of Computational Decipherment

• While Luo, Cao, and Barzilay (2019, ACL) demonstrated neural decipherment on Ugaritic (a known script, as validation), the method requires a known related language as an anchor — for truly undeciphered scripts like Linear A, Proto-Elamite, or Rongorongo, where the underlying language is unknown, computational methods alone cannot solve the core problem
• Entropy and Zipf's-law analyses can suggest that a sign system has linguistic properties but cannot distinguish between writing and other structured symbolic systems (accounting notation, calendrical records) — the claim that statistical methods will "crack" undeciphered scripts oversells current capabilities

IMAGES

No images assigned yet.

BIBLIOGRAPHY

1. Robinson, Andrew | 2007 | ∅ | The Story of Writing: Alphabets, Hieroglyphs, and Pictograms | ∅ | ∅ | London: Thames & Hudson | 2nd | doi:10.1108/09504120810859909, isbn:9780500286609 | ∅ | ∅ | ∅
2. Champollion, Jean-François | 1822 | ∅ | Lettre à M. Dacier relative à l'alphabet des hiéroglyphes phonétiques | ∅ | ∅ | Paris: Firmin Didot | ∅ | doi:10.3406/piot.1921.1813 | ∅ | ∅ | ∅
3. Chadwick, John | 1967 | ∅ | The Decipherment of Linear B | ∅ | ∅ | Cambridge: Cambridge University Press | 2nd | doi:10.1017/s1478572211000296 | ∅ | ∅ | ∅
4. Pope, Maurice | 1999 | ∅ | The Story of Decipherment: From Egyptian Hieroglyphs to Maya Script | ∅ | ∅ | London: Thames & Hudson | Rev. | isbn:9780500281055 | ∅ | ∅ | ∅
5. Coe, Michael | 2012 | ∅ | Breaking the Maya Code | ∅ | ∅ | London: Thames & Hudson | 3rd | isbn:9780500289532 | ∅ | ∅ | ∅
6. Knorozov, Yuri | 1952 | "Ancient Writing of Central America" | Sovetskaya Etnografiya | ∅ | 3::100–118 | ∅ | ∅ | ∅ | ∅ | ∅ | ∅
7. Proskouriakoff, Tatiana | 1960 | "Historical Implications of a Pattern of Dates at Piedras Negras, Guatemala" | American Antiquity | ∅ | 25.4::454–475 | ∅ | ∅ | doi:10.2307/276633 | ∅ | ∅ | ∅
8. Rawlinson, Henry | 1846 | ∅ | The Persian Cuneiform Inscription at Behistun | ∅ | ∅ | London: Royal Asiatic Society | ∅ | ∅ | ∅ | ∅ | ∅
9. Parpola, Asko | 1994 | ∅ | Deciphering the Indus Script | ∅ | ∅ | Cambridge: Cambridge University Press | ∅ | isbn:9780521430796 | ∅ | ∅ | ∅
10. Farmer, Steve, Richard Sproat; Michael Witzel | 2004 | "The Collapse of the Indus-Script Thesis: The Myth of a Literate Harappan Civilization" | Electronic Journal of Vedic Studies | ∅ | 11.2::19–57 | ∅ | ∅ | ∅ | ∅ | ∅ | ∅
11. Singh, Simon | 2000 | ∅ | The Code Book: The Science of Secrecy from Ancient Egypt to Quantum Cryptography | ∅ | ∅ | New York: Anchor Books | ∅ | isbn:9780385495322 | ∅ | ∅ | ∅
12. Houston, Stephen, Oswaldo Chinchilla Mazariegos; David Stuart (eds.) | 2001 | ∅ | The Decipherment of Ancient Maya Writing | ∅ | ∅ | Norman: University of Oklahoma Press | ∅ | isbn:9780806133907 | ∅ | ∅ | ∅
13. Luo, Jiaming, Yuan Cao; Regina Barzilay | 2019 | "Neural Decipherment via Minimum-Cost Flow: From Ugaritic to Linear B" | Proceedings of the 57th Annual Meeting of the ACL | ∅ | ∅ | In 3146 3155 | ∅ | doi:10.18653/v1/P19-1303 | ∅ | ∅ | Florence: ACL
14. Ferrara, Silvia | 2022 | ∅ | The Greatest Invention: A History of the World in Nine Mysterious Scripts | ∅ | ∅ | New York: Farrar, Straus and Giroux | ∅ | isbn:9780374601635 | ∅ | ∅ | ∅
15. Stuart, David | 1987 | "Ten Phonetic Syllables" | Research Reports on Ancient Maya Writing | ∅ | 14::1–14 | ∅ | ∅ | ∅ | ∅ | ∅ | ∅
16. Daniels, Peter T.; William Bright (eds.) | 1996 | ∅ | The World's Writing Systems | ∅ | ∅ | Oxford: Oxford University Press | ∅ | isbn:9780195079937 | ∅ | ∅ | ∅

CROSS-REFERENCE INDEX

Generated from V4 expansion plan. Last Updated: Apr 12, 2026