Ancient Writing Scripts: The Linear A Writing System

Linear A is a writing script used mainly on the island of Crete from the beginning of the Middle Minoan IIb period circa 1825 BC (Rehak & Younger, 1998) until the end of the Late Minoan I period circa 1490 BC on the island of Crete, with scattered survivals of the script outside of Crete surviving as late as Late Minoan IIIa1-2 period circa 1320 BC (Duhoux, 1998: 8). As you can tell from the dates mentioned, the Linear A writing script predates the Greek alphabet by about 1000 years and the writing script was in use during the same time period as Middle Egyptian hieroglyphs and Akkadian cuneiform writing scripts. In other words, it is a very old writing script.

The Linear A writing script has also been discovered on other Aegean Islands like Kea, Kythera, Melos, Thera, Samothrace, mainland Greece (Mycenae, Tiryns, Argos) (Davis, 2010), and even the Levant (Tel Haror, Tel Lachish) (Finkelberg et al., 1996).

Sir Arthur Evans has been credited with discovering the writing script in 1900 during his excavations of the palace in Knossos in Crete. Sir Arthur Evans named the writing script ‘Linear A’ after the writing script’s distinctive feature of utilising mainly combinations of lines (Robinson, 2009: 8). Many scholars have drawn parallels between Linear A and the Cretan ruins with the myth of a ‘Minoan thalassocracy’ that gained popularity among Classical Greek writers such as Thucydides (Starr, 1955). As the word thalassocracy implies, Minoan Crete were portrayed as a naval empire and among the many stories that emerged about the island from Greek mythology is the story of King Minos and his labyrinth, and the island being the birthplace of Zeus. Minoan wares and architecture were famed in the Bronze Age societies with mentions of the island found on several Middle Egyptian hieroglyphs and Minoan-style architecture found all over the Mediterranean.The discovery of the language behind Linear A would be fascinating as it can potentially re-write early western history as we know it. It could tell us who lived and built the great Minoan civilisation, the precursor of the famous Ancient Greek civilisation.

Linear A was not the only writing script found by Sir Arthur Evans in 1900. Alongside Linear A, the archaeologist discovered Cretan Hieroglyphs and the Linear B writing script. Cretan Hieroglyphics was used during the Middle Minoan period circa 2000-1700 BC (Rehak & Younger, 1998) and the writing script may be the progenitor of Linear A, although this proposal cannot yet be ascertained (Davis, 2010). Linear B is a writing script used from about 1450 BC - 1200 BC that was found to encode the earliest known form of Greek, Mycenaean Greek (Steele, 2017). The writing script was deciphered by Michael Ventris in 1952. Due to their graphemic similarities, Linear B is seen as a daughter script of the older Linear A writing script. Another daughter script of the Linear A writing script is the undeciphered Cypro-Minoan script found on the island of Cyprus.

Linear A remains an undeciphered writing script, and the main reason for this is due to the paucity of Linear A materials and lack of existing bilingual texts (Hooker, 1975) (Olivier, 1986: 383). Bilingual texts such as the Rosetta Stone and Behistun Inscriptions are how Egyptian hieroglyphs and the cuneiform writing systems were eventually deciphered. As a comparison, Linear B which was deciphered in 1952 by Michael Ventris appears on more than 4600 artefacts with signs occurring 57398 times while Linear A only appears on 1427 artefacts with signs appearing around 7362-7396 times (Younger, 2020: Section 5). With the limited sample size and the possibility that many words identified within the Linear A corpus could just be linguistic borrowings as well, this would severely reduce the possibility of discovering what language is being encoded within the writing script.

Use of Linear B Sound Values to Read Linear A Texts 

One step researchers used to decipher Linear A is to use Linear B sound values to ‘read’ Linear A texts (Hooker, 1975) (Olivier, 1975, 1986: 382-3). The Linear B writing script was used between 1450-1200 BC and its replacement of the Linear A script on Crete, which fell out of use shortly after Linear B’s introduction, is associated with the Mycenaean Greeks assuming palatial control over the island from the Minoans around 1450 BC. The methodology of using Linear B sound values has been criticised on the basis that 80% of Linear A signs are unique to Linear B (Younger, 2020: Section 7b) and correspondingly produces meaningless ‘words’ when read this way. A similar methodology employed to decipher the Carian language by comparing Carian symbols with Greek alphabets have shown that incorrect phoneme-grapheme inferences between writing systems can lead to incorrect reading (Adiego, 2006: 176-187). As Linear A represents a separate writing system, phonological adaptation and realignment must have taken place between the two writing scripts (Steele & Meissner, 2017).

Nevertheless, there are strong arguments to support the use of Linear B sound values to read Linear A texts. Statistical experiments by Packard (1974: 72-102) and Davis (2014: 246ff), the identification of place names within the Linear A corpus, and the argument for the acrophonic principle in writing scripts (Steele & Meissner, 2017) all lend evidence for the use of Linear B sound values to read Linear A symbols. Recent epigraphic studies for example showed that 64 out of the 89 signs are shared between Linear A and Linear B signs, giving a grapheme correspondence rate of 72% which is comparable to the relationship between late 3rd Century Latin alphabets and Old Italic alphabets which produced a 74% grapheme correspondence rate (Steele & Meissner, 2017). Sound values from related scripts have also been shown to be tenacious and remain largely constant over a period of time (Steele & Meissner, 2017). For example, the Linear B da (/da/) sign was borrowed into Cypriot Syllabary as ta (including /ta/, /tha/, /da/) while Linear B ti (/ti/) and to (/to/) also ended up in the t-series of Cypriot Syllabary but not di and do (See Fig. 1). This means that while the exact phonetic values of Linear A cannot be recovered with precision, the shared values would be close enough to be reconstructed approximately.

Alas, using Linear B phonetic values to read Linear A script produces mostly meaningless sounds that does not attribute to any known languages. Using the scant Linear A material researchers have to work with, a number of languages has been hypothesised to be behind the Linear A writing script ranging from Indo-European Sanskrit (Owens, 1999), Greek (Nagy, 2002), Semitic (Rendsburg, 1996) and Etruscan (Facchetti & Negri, 2003). Each one of the researchers assertions have their merits and justification, however, given the paucity of the linear A corpus and lack of known bilingual texts, it is not conclusive as to which language is behind the writing script.

Linguistic Research

One vital step towards a systematic study of the Linear A language is the consolidation of Linear A artefacts into an updated and scientifically redacted corpus (Oddo, 2007). Leaving aside Sir Arthur Evans ambitious but incomplete attempt, the first scientific publication of a Linear A corpus was done by Giovanni Pugliese Caratelli in 1945 with subsequent innovations, revisions and updates to the corpus providing researchers with better means and alternatives for analysing Linear A (Oddo, 2007). Currently, Godart and Olivier’s (1976 - 1985) GORILA corpus are regarded as the most authoritative work for any scientific analysis on Linear A although Raison and Pope’s (1980, 1994) corpora also offers alternative methodologies and approach that are preferred by certain researchers (Oddo, 2007). With these vital foundations in place, an internal and external analysis of the language of the writing script can be done.

From what we know about the writing script so far other than the fact that the script is linear and looks similar to Linear B, the Linear A corpus, like Linear B, has been found to mostly contain economic tablets (Palmer, 1995) (Petrolito, Petrolito, Winterstein, & Perono Cacciafoco, 2015: 96). This means that the Linear corpus would consist of mostly personal or place names. The most common combination of signs in the corpus is ‘KU-RO’, which in Semitic languages has the meaning of ‘all’ or ‘total’. Researchers have also managed to deduce the tallying system used in the tablets, thus we are able to tell how much of a certain object or person is being tallied, although we may not know what the object exactly is. The writing script also appears to have an added religious and artisanal use as evidenced by inscriptions on libation tables (Palmer, 1995) (Eu Min, Perono Cacciafoco & Cavallaro, 2019).

For more insight into the current state of linear A research, you may check out Younger’s website that aims to provide a general overview of Linear A researches and to refer to GODART’s GORILA corpus. As new linear a materials are found once in a while, it may be sooner than we think that the writing script be deciphered.