Linear A

• Linear A: The Undeciphered Language of Minoan Crete
• Could Linear A Encode a Semitic Language?
• What Gordon Actually Saw in Linear A
• Could Linear A Preserve Semitic the Way Linear B Preserves Greek?
  • Linear B : Greek :: Linear A : proposed Semitic
• Appendix: Ancient Traditions Connecting Judeans with Crete and Sparta
  • Herodotus: Palestine as a Geographical Region
  • Tacitus: A Tradition That the Judeans Came from Crete
  • 1 Maccabees: Spartans and Judeans as Brothers of the Same Genos
  • Crete and Sparta
  • Judeans and Relevance to Linear A
  • Chronology: From Bronze Age Crete to Later Judea
• The Serious Objection to Gordon

Linear A: The Undeciphered Language of Minoan Crete

Linear A is one of the major unresolved linguistic problems of the ancient Mediterranean. It was used primarily on Crete during the Bronze Age, approximately 1800–1450 BCE, within the civilization conventionally called Minoan. Hundreds of inscriptions survive on clay tablets, sealings, vessels, ritual objects, and other materials, particularly in administrative and religious contexts.

The writing system itself is not completely mysterious. Its signs can be catalogued, recurring words and formulas can be recognized, and many signs can be assigned probable phonetic values through comparison with Linear B, the related script later used to write Mycenaean Greek. This allows scholars to approximately sound out portions of Linear A without necessarily understanding what those words mean.

The great unresolved question is therefore not simply how to read the signs, but what language those signs represent. Unlike Linear B, whose underlying language was demonstrated to be Greek, no identification of the language encoded by Linear A has achieved scholarly consensus.

Among the proposed solutions is a particularly interesting hypothesis developed by Cyrus H. Gordon: that beneath the Minoan syllabic writing system lies a Semitic language.

Could Linear A Encode a Semitic Language?

Yes. As a hypothesis, Linear A encoding a Semitic language is entirely possible at the level of writing-system mechanics, and Cyrus H. Gordon spent decades arguing precisely that. The claim remains disputed and has not become the prevailing decipherment, but it is not equivalent to saying “the symbols look vaguely Phoenician.” Gordon’s argument was linguistic: he proposed that the language written in Linear A was Northwest Semitic, and he tried to demonstrate this by reading recurring Linear A sequences through values inherited into Linear B and comparing the resulting words and grammatical patterns with Ugaritic, Akkadian, Hebrew, and related languages. Gary Rendsburg summarizes Gordon’s position straightforwardly: after studying the connections among Greece, Ugarit, and Israel “with Crete as the hub,” Gordon concluded controversially that Minoan Linear A encoded a West Semitic dialect. (⁠RCNi Company Limited)

That distinction between script and language is crucial. Linear A is not an alphabet like Phoenician or the later Paleo-Hebrew script. It is a logo-syllabary: many signs represent syllables, apparently often simple vowels or open consonant-plus-vowel syllables such as ka, ma, ti, and so forth, while other signs function as commodity signs or logograms. Linear B, which was adapted substantially from Linear A, uses exactly this sort of syllabic machinery to write Mycenaean Greek. (⁠OUP Academic)

But a syllabic writing system (like Linear B) does not always imply Greek language (obviously). That is the opening Gordon exploited. A Semitic language can perfectly well be forced into a syllabic writing system; Akkadian is the obvious ancient example, because a Semitic language was written for millennia using cuneiform syllabograms rather than a consonantal alphabet. Accordingly, Linear A could theoretically encode Semitic phonemes very differently from later Phoenician or Paleo-Hebrew writing even if the underlying language were related. The later consonantal alphabet would not need to descend graphically from Linear A for both systems to have represented related languages.

Linear B primarily uses CV syllables—consonant plus vowel—alongside vowel-only signs. Because Mycenaean Greek contained consonant clusters and final consonants that the script could not represent directly, Linear B often omits consonantal information or introduces vowels required by the syllabary. Thus Greek Knōsos, for example, has to be represented approximately by the resources of the syllabary. In this sense, the written surface can disguise the actual phonological structure of the spoken language. Linear A could therefore produce spellings that look rather unlike how the same roots would later appear in a consonantal Semitic alphabet. (⁠OUP Academic)

What Gordon Actually Saw in Linear A

Gordon first presented the Semitic case prominently in his 1957 article “Notes on Minoan Linear A.” He initially leaned toward West Semitic, while acknowledging the possibility of East Semitic/Akkadian influence. One of the things that attracted him was the similarity between Linear A administrative vocabulary and recognizable Semitic forms. He later wrote that a Linear A sequence he read as ga-ba could be compared with Akkadian gabba, “all,” in a totaling expression. (⁠Cambridge University Press)

Another frequently discussed example is Linear A KU-RO, which regularly occurs at the ends of administrative lists and is widely understood from context to mean something like “total.” Gordon connected this with the Semitic root kl / kull-, “all, total.” Even one of his contemporary critics conceded that kuro/kulo could plausibly be compared with the common Semitic root kl, while objecting to Gordon’s handling of some sign values elsewhere. (⁠Cambridge University Press)

This is worth emphasizing because it shows why Gordon’s proposal cannot simply be dismissed as visual pattern matching. He was looking at recurrent vocabulary in recurrent administrative positions, and asking whether Semitic morphology and lexical roots explain it.

His broader proposal became stronger rather than weaker in his own work. By 1975 he was arguing that both the Bronze Age Minoan language written in Linear A and the much later Eteocretan inscriptions of Crete belonged to Northwest Semitic. He explicitly said he wanted to make the case from the linguistic evidence rather than relying principally upon historical legends, divine names, or place names. (⁠Cambridge University Press)

That later Eteocretan component is particularly interesting for the question you are asking because it potentially gives Gordon a continuity bridge. Linear A disappears around the middle of the second millennium BCE. Centuries later, from roughly the sixth through fourth centuries BCE, eastern Crete still contains inscriptions written with Greek letters but not in Greek. These are conventionally called Eteocretan—essentially “true Cretan.” Their language remains undeciphered, and modern scholarship does not presently accept that its relationship to Linear A has been demonstrated. But Gordon argued that both were Northwest Semitic. (⁠OUP Academic)

And Gordon’s idea has not disappeared completely. A 2016 study of an inscription from Dreros reconsidered the possibility of a Greek–Semitic bilingual and explicitly said that the non-Greek portion might corroborate Gordon’s identification of a Northwest Semitic word in the text. That does not prove Gordon’s whole decipherment, but it does show that the Semitic hypothesis remains capable of generating scholarly discussion rather than merely surviving in fringe literature.

Could Linear A Preserve Semitic the Way Linear B Preserves Greek?

This is the central linguistic possibility behind the Semitic hypothesis for Linear A.

Linear B preserves an early form of Greek, but it does not write Greek the way the later Greek alphabet does. Linear B is principally a syllabic writing system. Its signs normally represent open syllables such as pa, ko, ti, or me, while Greek itself contains consonant clusters, final consonants, and other phonological structures that the script cannot represent directly.

Consequently, a Mycenaean Greek word can look substantially different in Linear B from the form familiar from later alphabetic Greek. The script is effectively filtering the spoken language through a restricted inventory of syllabic signs. Consonants may disappear from the written representation, consonant clusters may be broken apart, and vowels may appear as part of the mechanics required to represent the underlying Greek word.

This provides a useful analogy for the Semitic hypothesis concerning Linear A.

If Linear A encoded a Semitic language, there is no reason to expect its words to look like later Phoenician or Paleo-Hebrew spellings. Those later scripts primarily represent consonants, whereas Linear A appears to represent language predominantly through syllables. The same underlying vocabulary could consequently have a very different written appearance.

A simplified hypothetical example illustrates the principle. A Semitic root structured around consonants K-L might appear in a consonantal alphabet principally through K-L, while a syllabic writing system would necessarily express those consonants through syllables containing vowels: something resembling KU-RO, KU-LO, KA-LA, or another form depending upon the actual pronunciation and the capabilities of the script.

This is precisely why Linear A KU-RO is so interesting.

KU-RO repeatedly occurs in administrative documents in a position indicating a total or summation. Its function is unusually secure even though its linguistic derivation is not. Cyrus H. Gordon compared it with the widespread Semitic KL / KLL word-family meaning “all, whole, entirety,” including forms such as Akkadian kalû/kullatu and Northwest Semitic kl. Other scholars have independently recognized that KU-RO functions as a totaling term while disagreeing over whether Gordon’s Semitic etymology explains it.

That creates an intriguing structural comparison:

Linear B : Greek :: Linear A : proposed Semitic

The analogy does not require the scripts used for later Greek and later Semitic languages to descend directly from Linear B and Linear A respectively. It concerns how an earlier syllabic script can preserve an underlying spoken language in a written form substantially different from the alphabetic spelling used for related language forms centuries later.

Linear B provides the demonstrated case. Mycenaean Greek is unmistakably Greek once the conventions of the syllabary are understood, even though many written forms initially look strange when compared directly with alphabetic Greek.

The Semitic hypothesis asks whether Linear A presents the same kind of problem.

Perhaps researchers looking for recognizable Semitic words should not expect a later consonantal spelling at all. They should instead ask what a Semitic word would look like after being passed through the phonological restrictions of the Minoan syllabary.

That was an important part of the opportunity recognized by Cyrus H. Gordon. Gordon proposed that Linear A encoded a West or Northwest Semitic language and attempted to identify Semitic vocabulary and grammatical structures beneath its syllabic spellings. His proposal remains disputed, but Linear A itself remains undeciphered; there is presently no universally accepted alternative identification of the underlying language.

The hypothesis can therefore be stated cleanly:

Linear B demonstrates that an Aegean syllabary can preserve a recognizable language while substantially obscuring its familiar phonological form. Linear A may similarly preserve a Semitic language whose words look unfamiliar precisely because they have been forced through a Minoan syllabic writing system rather than the later consonantal alphabet through which Northwest Semitic languages are usually encountered.

That proposition is testable. The question is not whether individual Linear A words can occasionally be made to resemble Semitic words. A convincing decipherment would need recurring phonetic values, vocabulary that matches the archaeological context, predictable grammatical morphology, and readings that continue to work when applied to inscriptions that were not used to construct the decipherment.

Gordon believed the Semitic hypothesis could meet that test.

Whether it ultimately does remains an open and disputed question—but Linear B shows exactly why the underlying idea is linguistically possible.

Appendix: Ancient Traditions Connecting Judeans with Crete and Sparta

The proposed identification of Linear A as a Semitic language is a linguistic question and does not depend upon later traditions concerning the origins of the Judeans.

Nevertheless, those Judean traditions become relevant to the larger historical question because ancient sources themselves preserve explicit claims connecting Judeans with Crete and Judeans with Sparta. These sources are considerably later than the Bronze Age and cannot by themselves establish where the speakers of Linear A went after the disappearance of the script. They do, however, demonstrate that Cretan and Spartan connections to Judean origins or ancestry existed within ancient historical traditions.

Herodotus: Palestine as a Geographical Region

Before considering the origin traditions themselves, Herodotus provides an important fifth-century BCE geographical reference for Παλαιστίνη / Palestine. His terminology places Palestine within the larger geographical category of Syria while distinguishing it from Phoenicia.

The expression Συρίη ἡ Παλαιστίνη καλεομένη means literally “Syria, the one called Palestine.” Herodotus is therefore using Syria as the broader geographical designation while distinguishing a territory called Palestine between Phoenicia and Egypt.

This reference should not itself be treated as evidence for Cretan origins. Its importance here is chronological and geographical: Παλαιστίνη is already an identifiable geographical designation in surviving Greek literature of the fifth century BCE, independently of the considerably later Cretan-origin tradition preserved by Tacitus.

Tacitus: A Tradition That the Judeans Came from Crete

One of the clearest surviving statements connecting the Judeans with Crete occurs in Tacitus, Histories 5.2, written around the beginning of the second century CE. Tacitus presents several competing accounts of Judean origins; the Cretan account is the first one he records.

His Latin is explicit:

---

Tacitus, Histories 5.2, ca. 100–110 CE

Iudaeos Creta insula profugos novissima Libyae insedisse memorant, qua tempestate Saturnus vi Iovis pulsus cesserit regnis. argumentum e nomine petitur: inclutum in Creta Idam montem, accolas Idaeos aucto in barbarum cognomento Iudaeos vocitari.

“They record that the Judeans, fugitives from the island of Crete, settled at the furthest parts of Libya, at the time when Saturn, driven out by the power of Jupiter, withdrew from his kingdom. The argument is taken from the name: in Crete is the celebrated Mount Ida; its inhabitants, the Idaeans, were called Judeans through an expanded foreign form of the name.”

---

Tacitus therefore preserves an ancient tradition with an explicit geographical movement:

Crete → migration of Cretan fugitives → settlement toward the eastern Mediterranean → Iudaei.

He additionally reports an ancient etymological argument connecting Idaei, inhabitants associated with Cretan Mount Ida, with Iudaei, Judeans.

This should not be represented as Tacitus independently demonstrating that the Judeans historically descended from Minoan Cretans. He is reporting an origin tradition, and the proposed Idaei → Iudaei etymology is not accepted as an established modern etymology. What the passage securely establishes is that an explicit Cretan-origin tradition for the Judeans circulated in antiquity and was considered sufficiently noteworthy for Tacitus to preserve it in his ethnographic account.

The chronological distance is also important. Tacitus wrote more than a millennium after the disappearance of Linear A. Consequently, this passage cannot independently bridge the missing centuries between Bronze Age Crete and historical Judea. In the context of a proposed Cretan–eastern Mediterranean population connection, however, it is evidence that the association itself is ancient rather than a modern invention.

1 Maccabees: Spartans and Judeans as Brothers of the Same Genos

A separate ancient tradition connects the Judeans and Spartans. It is preserved in 1 Maccabees 12, where Jonathan sends representatives to Sparta and other places to renew an earlier relationship. The text then reproduces what it presents as an earlier letter from the Spartan king Areus to the Judean high priest Onias.

The central passage is unusually direct:

The crucial Greek is:

The word γένος is significant. It can denote a lineage, stock, race, family, kind, or descent-group. The text is therefore not merely describing the Spartans and Judeans as diplomatic “brothers.” It follows that statement with an explicit assertion of shared γένος associated with Abraham.

The tradition appears again within the same chapter when Jonathan addresses the Spartans:

Thus the claimed relationship is not confined to a stray phrase within the quoted letter. The author of 1 Maccabees incorporates the Spartans-as-brothers relationship into the narrative framing of the correspondence itself.

Whether the purported letter of Areus is an authentic third-century BCE Spartan document, an authentic diplomatic tradition subsequently transmitted in Judean literature, or a later literary construction has been debated. That question should remain separate from what the surviving ancient text actually says. 1 Maccabees unequivocally preserves a tradition in which Spartans and Judeans possess a shared ancestry and are “brothers,” specifically placing both within the γένος of Abraham.

Crete and Sparta

Sparta itself was not located on Crete. Historical Sparta occupied Laconia in the southern Peloponnese. Nevertheless, Crete and Sparta were closely associated within Greek traditions concerning institutions, laws, education, communal dining, and social organization, and both regions contained historically important Dorian-speaking populations.

Consequently, the two traditions should not be collapsed into the claim that “Spartans were Cretans.” Rather, they provide two separate ancient connections surrounding the later Judeans:

Tacitus preserves a tradition of Judean origins in Crete, while 1 Maccabees preserves a tradition of common ancestry between Judeans and Spartans. The historically recognized relationships between Cretan and Spartan cultural institutions make the combination noteworthy, but they do not by themselves establish a single migration or genealogy connecting all three populations.

Judeans and Relevance to Linear A

The relevance of later Judeans to Linear A runs from Crete forward in time to later Judea. Bronze Age populations on Crete used Linear A; much later, Tacitus preserves an origin tradition in which the Iudaei/Judeans came from Crete. The question is therefore whether some later Judean populations could preserve ancestry from earlier Cretan populations, and whether the language once written in Linear A has any bearing on the linguistic history of those later populations—including their subsequent use of Proto-Hebrew/Paleo-Hebrew and the closely related Phoenician script and language, Greek, and Aramaic.

This becomes particularly interesting because Cyrus H. Gordon independently argued from the linguistic evidence that Linear A encoded a Semitic language. His proposal does not depend upon Tacitus: it begins with the Linear A inscriptions themselves and attempts to explain their vocabulary and grammar through Semitic languages. The later Cretan-origin tradition therefore approaches the problem from a different direction. If Gordon’s linguistic identification were ultimately demonstrated, a Semitic-speaking population would already be present in Bronze Age Crete long before the historical appearance of the Judeans.

The later evidence adds another possible line of continuity. Non-Greek Eteocretan inscriptions survive on Crete many centuries after Linear A disappeared, and Gordon argued that Eteocretan too was Northwest Semitic. The relationship between Eteocretan and the language of Linear A remains disputed, but if such continuity could be established, it would narrow substantially the chronological gap between the Bronze Age Cretan language and the later Mediterranean populations known from historical sources.

Chronology: From Bronze Age Crete to Later Judea

The proposed relationship becomes clearer when the evidence is placed chronologically:

• ca. 1800–1450 BCE — Linear A on Crete: Minoan Crete uses Linear A to record a language that remains unidentified. Cyrus H. Gordon later argues that this underlying language was Semitic, particularly West or Northwest Semitic.
• ca. 1450 BCE onward — Linear A disappears: Linear A ceases to be the principal administrative writing system, while Linear B, encoding Mycenaean Greek, appears prominently at Knossos. The disappearance of Linear A establishes the disappearance of the writing system, not necessarily the disappearance of its spoken language or population.
• Late Bronze Age–Early Iron Age — Mediterranean population changes: The Aegean and eastern Mediterranean undergo substantial political disruption, migration, settlement, cultural exchange, and population reorganization. This period provides a possible—but presently undocumented—historical interval through which earlier Cretan populations could contribute to later Mediterranean populations.
• First millennium BCE — Phoenician and related Northwest Semitic languages: Alphabetic Northwest Semitic languages are documented extensively in the eastern Mediterranean. Later Hebrew belongs to this Northwest Semitic linguistic environment, while its early alphabetic writing belongs to the closely related Phoenician/Paleo-Hebrew script tradition.
• ca. sixth–fourth centuries BCE — Eteocretan: Crete still contains inscriptions written in a non-Greek language, now using the Greek alphabet. Cyrus H. Gordon argues that Eteocretan is also Northwest Semitic and therefore potentially preserves linguistic continuity with Linear A; this identification remains disputed.
• Hellenistic period — Judean–Spartan kinship tradition: 1 Maccabees 12 preserves correspondence asserting that Spartans and Judeans are brothers and are both ἐκ γένους Ἀβραάμ, “from the genos of Abraham.”
• ca. 100–110 CE — Cretan origin of the Judeans recorded by Tacitus: Tacitus, Histories 5.2, preserves an explicit ancient tradition that the Iudaei/Judeans were fugitives from Crete, together with an ancient attempt to connect their name with the Idaei of Cretan Mount Ida.

Read chronologically, the hypothesis therefore sequences as: Bronze Age Crete → Linear A population and language → subsequent Mediterranean population movements and linguistic changes → later Judean and Spartan populations → surviving ancient traditions connecting those later populations back toward Crete and the Aegean.

Linear A potentially supplies evidence from the beginning of this sequence. If Gordon’s Semitic identification is correct, it would place a Semitic-speaking population in Bronze Age Crete independently of the much later Cretan and Spartan traditions. Those later traditions would then become separate historical evidence to compare against the linguistic hypothesis rather than evidence from which the Linear A identification itself was constructed.

The Serious Objection to Gordon

The strongest objection to Gordon’s Semitic identification of Linear A is not that a Semitic language could not have existed on Bronze Age Crete; nothing about the writing system makes that impossible. The serious objection is whether Gordon demonstrated the identification with enough internal consistency and grammatical evidence. Maurice Pope’s contemporary critique identified specific cases in which Gordon changed or selectively transferred phonetic values from corresponding Linear B signs in order to obtain proposed Semitic readings. Pope argued that this creates a methodological problem: if a sign can be assigned one value for one proposed cognate and another value elsewhere, apparent matches can be produced without establishing a reproducible decipherment. Another contemporary assessment likewise criticized several of Gordon’s proposed word readings as incorrect or contextually unsuitable and observed that the early argument had not yet demonstrated the grammatical structure expected if the underlying language were actually Semitic. Importantly, that critic explicitly allowed that Gordon might nevertheless be correct in his general conclusion that Linear A could represent a Semitic language. (Cambridge University Press)

The unresolved test is therefore larger than finding plausible words such as KU-RO. A persuasive Semitic decipherment would need a stable set of phonetic values that works across inscriptions, recurring Semitic morphology and syntax, meanings independently supported by the administrative or ritual contexts, and successful readings of texts not used to construct the hypothesis. Later assessments have similarly noted that lexical correspondences alone could represent borrowing or coincidence and that Gordon drew comparisons from more than one Semitic language rather than demonstrating a single consistently recoverable linguistic system. (MDPI)

Thus the serious rebuttal to Gordon is not “Linear A cannot be Semitic,” but “the Semitic hypothesis has not yet been demonstrated at the level required for a decipherment.”

• Bottom Line: Linear A remains undeciphered, leaving Gordon’s proposal possible but unproven.