MYS0CWRBGTW47P1·Mystery · script

Cypro-Minoan

Mikhail Ryazanov · CC BY-SA 3.0

Where: Cyprus (Enkomi, Kition, Kalavasos, Kourion, Palaepaphos); Ugarit; Tiryns⁠[1]
Date: c. 1550 to 1050 BC⁠[1]
Corpus: About 250 inscribed objects, about 4,000 signs⁠[2]
Longest text: Clay cylinder from Enkomi (##097), 217 signs⁠[11]
Editions: Olivier 2007 (217 inscriptions); Ferrara 2012 and 2013 (about 244)⁠[15]
Sign inventory: Olivier 96 signs; Ferrara 74; Valério 57 to 70; Polig and Donnelly 88⁠[7]
Unicode: Encoded in Unicode 14.0 (2021), U+12F90 to U+12FFF, 99 characters⁠[2]
Open corpus: None; the sign images used in computational work are in copyright⁠[17]
Status: Partly explained
Questions: 0 of 5 answered

People Jean-Pierre OlivierEpigraphist; HoChyMin corpus and sign list Miguel ValérioLinguist; signary and proposed values Silvia FerraraPhilologist (Bologna); corpus volumes and INSCRIBE Cassandra M. DonnellyArchaeologist; potmarks, writers and the integrated signary Emilia MassonEpigraphist; the CM1, CM2, CM3 classification

All 7 people ›

Richard JankoClassicist; Enkomi cylinder as an Eteocypriot account Thomas G. PalaimaAegean epigraphist; critique of decipherment schemes

Catalog keyMYS0CWRBGTW47P1

Cypro-Minoan is the writing of Bronze Age Cyprus. Arthur Evans named it in 1909 for its likeness to Linear A on Crete, from which it is thought to derive, and it later gave rise to the Cypriot syllabary used for Greek and Eteocypriot in the first millennium BC⁠[1]⁠[2]. About 250 inscribed objects survive from about 1550 to 1050 BC: clay tablets, small clay balls, cylinders, vessels and potmarks, metal bowls and ingots, ivory and stone. Most come from Cypriot sites, above all Enkomi, Kition and Kalavasos-Ayios Dhimitrios, with a group from Ugarit on the Syrian coast and single finds as far as Tiryns⁠[1]⁠[2]. The whole corpus holds about 4,000 signs, against more than 7,000 for Linear A and about 30,000 for Linear B when it was deciphered⁠[2].

Nothing in it can be read. The script is agreed to be a syllabary of open syllables, from the number of signs and the analogy with its relatives, and some signs look like Linear A or Cypriot syllabic signs with known values; but there is no bilingual, the texts are short, and the language is unknown⁠[1]⁠[3]. Palaima concluded in 1989 that every decipherment scheme then on offer was improbable, and no later one has changed that verdict⁠[3]⁠[2].

The work of the last two decades has been to fix the corpus and the sign list. Olivier's 2007 edition gave 217 inscriptions and 96 signs; Ferrara's 2012 and 2013 volumes added 27 more texts and argued that the three "scripts" of the standard classification are one; Valério's 2016 thesis built a signary of 57 to 70 syllabograms and proposed values for 60 sign forms⁠[4]⁠[5]⁠[6]⁠[7]. A 2022 neural-network study of 2,899 sign images supported a single script whose shapes vary with the medium, and in 2021 the script entered Unicode on Olivier's list⁠[8]⁠[2]. Of the 5 questions on this site, 0 are answered. The most likely answer to “Is Cypro-Minoan one script or several?”, not yet confirmed: cypro-Minoan is a single script; the subgroups reflect writing medium, writer and date, not separate systems.

Key questionsset by the editors

Is Cypro-Minoan one script or several?Most likely: cypro-Minoan is a single script; the subgroups reflect writing medium, writer and date, not separate systems

What language or languages does Cypro-Minoan write?Most likely: the language is unrelated to any known language, and the evidence cannot yet identify it

All 5 questions

The object

The corpus is a set of objects, not an archive. Eight tablets or tablet fragments carry over 2,000 signs, half the total: the small archaic tablet no. 1885 and three large tablets from Enkomi, and four from Ugarit⁠[2]⁠[9]. About 90 small clay balls of 1.5 to 2.3 cm from Enkomi and Kition carry three to five signs each; six clay cylinders, one from Enkomi and five from Kalavasos, carry the longest texts, the Enkomi cylinder with 217 signs⁠[10]⁠[11]. The rest are short marks on vessels, ingots, bowls and seals, many of them one or two signs, which makes it hard to say where Cypro-Minoan ends and potters' marks begin⁠[12]⁠[13].

The large Enkomi tablets (the CM2 group)Jastrow (Marie-Lan Nguyen). Public domainArtifact

The texts are known from three printed editions. Olivier's Édition holistique (HoChyMin, 2007) numbered 217 inscriptions ##001 to ##217 and set out 96 signs; Ferrara's two Oxford volumes (2012, 2013), which grew out of her 2006 UCL thesis, analysed 243 and added later finds⁠[14]; new pieces, such as a Canaanite jar from Tiryns with a previously unseen sign, appear in Kadmos⁠[4]⁠[15]⁠[6]⁠[16]. None of this is available as open data. The 2022 Bologna study built a set of 2,899 sign images from the printed drawings but could not release it because the drawings are in copyright; its code is public under a non-commercial licence⁠[8]⁠[17]. The INSCRIBE project at Bologna (2018 to 2024) set out to build the first complete digital corpus of the three undeciphered Aegean scripts with 3D models; what is public for Cypro-Minoan is two 3D models of clay-ball replicas⁠[18]⁠[19]. Polig and Donnelly's 2022 signary rests on a digital sign database and sub-millimetre 3D scans that have not been released⁠[20].

The script was encoded in Unicode 14.0 in 2021: 99 characters at U+12F90 to U+12FFF, named from Olivier's catalogue numbers on the understanding that the list is a catalogue, not a decipherment⁠[2]⁠[1].

Map of the regionOpen the full map

+−

EnkomiArchaeological site

Within 1 km

KitionArchaeological site

Within 1 km

CyprusIsland

A region

Kalavasos-Ayios DhimitriosArchaeological site

Within 100 m

Ugarit (Ras Shamra)Archaeological site

Within 1 km

MappedWithin 1 kmA regionNatural Earth

Timeline

Timeline

Squeezed where little happened. Each mark is a published claim, by how it stands:AcceptedMost likelyStill possibleDisputedRuled out

What is known

The script runs mostly left to right, separates words with a divider, has numerals like those of the other Aegean scripts, and has too few signs to be logographic and too many to be an alphabet, so it is taken to be a syllabary of open syllables⁠[1]. Its sign shapes link it to Linear A, which fits the archaic tablet no. 1885, the oldest text of any length, of Late Cypriot IB⁠[9]⁠[2]. A few signs carry over into the Cypriot syllabary with known values, which is the only external anchor for any reading⁠[7].

The classification has changed. Masson divided the texts into an archaic group and CM1, CM2 and CM3, treated as different scripts or stages; CM2 is essentially the three large Enkomi tablets (with a fourth long text), some 1,300 to 1,500 signs in a squarish style, and CM3 the small group from Ugarit⁠[21]⁠[1]. Ferrara showed that the three groups are contemporary, share statistical and combinatorial regularities and correlate with the material written on, and argued for one script⁠[5]. Valério built a signary from homogeneous subcorpora without the division and found 57 to 70 syllabograms; Polig and Donnelly counted 88 signs with 278 variants and also treat the corpus as one script⁠[7]⁠[20].

Inscribed clay balls from Enkomi and Kition (about 90)Jastrow (Marie-Lan Nguyen). Public domainArtifact

The function of some objects is partly known. The clay balls repeat sign sequences across finds and were probably tokens or labels in an administrative or allotment system; Janko reads the structure of the Enkomi cylinder as an account⁠[10]⁠[11]. Many vessel inscriptions are mercantile, and the script's users were traders more than palace scribes⁠[22].

What has been tested

The three-script division has been tested twice by measurement. Ferrara's statistics on sign frequency and combination found no difference between the groups that the writing medium could not explain⁠[5]. Corazza and colleagues trained an unsupervised convolutional network on 2,899 sign images from 213 inscriptions and found that the shapes cluster by support (clay tablet or other medium) rather than by subgroup, which they took as support for a single script⁠[8]. Both tests use the same printed drawings, so drawing conventions could shape the result.

Proposed sign values have been tested only by internal consistency. Valério assigned values by comparison with Linear A and the Cypriot syllabary, then checked them by transliterating the corpus and looking for recurring words and names; he regards nine values as confirmed and reads several personal names known from cuneiform in the Ugarit tablet RS 20.25⁠[7]. No proposed value has been confirmed by a bilingual or by a reading that predicts the content of an unread text. Palaima's survey of earlier decipherments found none that survived this standard⁠[3].

The readings

On what the script is, two positions stand. One is the older view that CM1, CM2 and CM3 are distinct scripts or stages, possibly writing different languages, with CM3 an adaptation used at Ugarit⁠[21]. The other, now the majority view, is that there is one script whose sign shapes vary with the writing surface and the writer, applied perhaps to more than one language⁠[5]⁠[7]⁠[20]⁠[8].

On language, nothing is established. Janko argues that the Enkomi cylinder is in Eteocypriot, the non-Greek language later written in the Cypriot syllabary; Valério allows for Eteocypriot on Cyprus and a Semitic language at Ugarit; other identifications have been proposed and none accepted⁠[11]⁠[7]⁠[3].

Published claims

Published claims and their status

Year	Who	Claimed	Status	Reason
1909	Arthur Evans⁠[23]	The Bronze Age signs found on Cyprus are a linear script derived from the Minoan writing of Crete, which he called Cypro-Minoan.	Accepted	Derivation from Linear A is the accepted view and the name has stuck, though the route of transmission is not documented⁠[1]⁠[7].
1974	Emilia Masson⁠[21]	The texts fall into an archaic group and three scripts or stages, CM1, CM2 and CM3, with CM3 an adaptation used at Ugarit.	Disputed	Ferrara showed the three groups are contemporary and share their regularities, and a 2022 network study grouped the signs by writing medium instead⁠[5]⁠[8].
1989	Thomas G. Palaima⁠[3]	Every past and current decipherment of Cypro-Minoan is improbable: the texts are too few and short, there is no bilingual and the language is unknown.	Accepted	No decipherment proposed since has been accepted, and Ferrara reached the same verdict in 2012⁠[5]⁠[2].
2007	Jean-Pierre Olivier⁠[4]	The corpus holds 217 inscriptions written with 96 distinct signs.	Disputed	Later signaries merge many of his signs as variants and count 57 to 70 or 88, though his numbering is the one Unicode adopted⁠[7]⁠[20]⁠[1].
2012	Silvia Ferrara⁠[5]	CM1, CM2 and CM3 are one script whose sign shapes differ with the material written on, not three scripts.	Accepted	Valério's signary, Polig and Donnelly's sign database and the 2022 network study all reached the same result⁠[7]⁠[20]⁠[8].
2016	Miguel Valério⁠[7]	Cypro-Minoan has 57 to 70 syllabograms once variants are merged.	Disputed	Polig and Donnelly count 88 signs from the same corpus, so the inventory is still disputed⁠[20].
2016	Miguel Valério⁠[7]	Values for 60 sign forms can be proposed from Linear A and the Cypriot syllabary, nine of them confirmed, and the Ugarit tablet RS 20.25 contains personal names known from cuneiform.	Still possible	The values rest on comparison and internal consistency; no bilingual or newly found text has yet tested them⁠[7].
2017	Silvia Ferrara and Miguel Valério⁠[10]	The inscribed clay balls were tokens or labels in an administrative or allotment system, as their repeated sign sequences show.	Most likely	The repetitions across balls found together are documented, but no ball has been read⁠[10].
2020	Richard Janko⁠[11]	The 217-sign cylinder from Enkomi is an accounting document, and its language may be Eteocypriot.	Still possible	The account structure is an argument from layout; the language cannot be tested until sign values are secure⁠[7].
2022	Michele Corazza, Fabio Tamburini, Miguel Valério and Silvia Ferrara⁠[8]	A neural network trained on 2,899 sign images without prior knowledge groups the signs by writing medium, not by Masson's subgroups, so Cypro-Minoan is one script.	Most likely	No study has challenged it, but it uses the same printed drawings as earlier work, so drawing conventions could shape the result⁠[8]⁠[20].
2022	Martina Polig and Cassandra M. Donnelly⁠[20]	An integrated signary of the whole corpus has 88 signs with 278 variants.	Disputed	Valério's count of 57 to 70 merges more of these forms, and the 3D scans behind the count have not been released⁠[7]⁠[20].
2024	Cassandra M. Donnelly⁠[22]	Cypro-Minoan was written mostly by merchants on goods and vessels rather than by palace scribes, and it outlived the collapse of about 1177 BC.	Still possible	It is an interpretation of the find contexts that has not been tested against an alternative⁠[22]⁠[12].

What would settle it

An open sign-level corpus is the first need: every inscription transcribed in Olivier's or Unicode sign numbers with object, site, context and medium, so that frequency, position and repeated-sequence statistics can be re-run and the single-script result tested on new drawings or on the 3D scans rather than on one set of published drawings⁠[17]⁠[20]. Sign values would be settled by a bilingual, most likely from Ugarit with Akkadian or Ugaritic, or by a set of values that turns the Ugarit names and the clay balls into consistent sequences and predicts the reading of a newly found text⁠[7]⁠[10].