BMCR 2021.04.30

Aegean linear scripts: rethinking the relationship between Linear A and Linear B

, Aegean linear script(s): rethinking the relationship between Linear A and Linear B. Cambridge classical studies. Cambridge: Cambridge University Press, 2020. Pp. 436. ISBN 9781108479387 $120.00.

This important book significantly advances our knowledge of Linear A (LA) and its relationship to Linear B (LB). Salgarella has done a great deal of very technical work which has value even beyond the conclusions she uses it to support. As she notes early on, Michael Ventris’s celebrated decipherment of LB was only made possible by the palaeographic work of Emmett Bennett and painstaking labours of Alice Kober. In spite of the evident success of focussing on palaeography and structure, however, few have undertaken such studies of LA. Salgarella has done much to remedy that.

The first of the book’s four chapters contains a useful review of the discovery and categorisation of LA and LB. When Arthur Evans found two ‘linear’ scripts at Knossos, he divided these into Class A and Class B (hence our LA and LB). In his conception they were ‘siblings’, each developed from an unattested common source. This was upended by the decipherment of LB in 1952: since it records Greek and LA does not, the two (so the thinking went) must be completely different, and LB has since been understood as a separate script developed from LA. Though this comes closer to the truth, Salgarella demonstrates that the line drawn between the two need not be so hard.

Guiding Salgarella’s study is a distinction between ‘script’ and ‘writing system’. A ‘script’ represents the repertoire of signs used, while a ‘writing system’ represents the language-specific rules by which those signs are read, including the notation of weights, measures, and numbers. The range of languages written with the Latin alphabet, with their range of required pronunciations, amply demonstrate the validity of this distinction. Salgarella argues that we should think of LA and LB not as two different scripts, but one script written according to two different, language-specific writing systems. The next two chapters are dedicated to demonstrating the structural and palaeographic continuity of the script between the two systems.

Chapter Two provides a comprehensive investigation of LA structure and palaeography. At the structural level, Salgarella emphasises that the primary distinction is between simple and composite signs. The former represents a distinct graphic unit, and can function phonetically (as syllabograms), logographically (as logograms), or both, depending on context. Composite signs, made up of two or more simple signs, are much less versatile, and can function only as logograms. Much of the chapter is concerned with determining the systems of juxtaposition and ligaturing used to create composition signs. This is all copiously illustrated, but assumes significant familiarity with the signs of LA.

Salgarella distinguishes between two types of juxtaposition, synthetic and analytic. In the former, an element is shared between the two signs; in the latter, they are completely separate. The signs in both seem to be combined according to the same principles: the left appears to be the marked position, and other signs (especially if smaller) are likely to be added on the right or above. Given these similarities, I wonder if the distinction is entirely meaningful. Salgarella points to A 591 (analytic) and A 592 (synthetic), composite signs on the same tablet (HT 27b) which differ only in their method of juxtaposition. As the same scribe wrote the two differently, so Salgarella argues, that difference should be significant. But as the context is very similar, it may be instead that the two signs were equivalent and interchangeable. But this would not seriously affect Salgarella’s arguments were it true.

Ligatured signs, Salgarella demonstrates, are typically constructed along the vertical axis. It is possible that these should be interpreted phonologically and read from the bottom up, if A 559, MA+RU is a sure model. This sign survived into LB as the logogram for wool and may be equated with μαλλός, ‘wool’, a possible Minoan loanword attested in Hesiod (Op. 234). Significantly, even though most composite signs are limited to one site, these rules hold across the island, suggesting a shared knowledge of the script’s underlying structure.

Salgarella turns next to the palaeography of LA, which has hitherto received less attention than it deserves. The essential work is GORILA, the corpus of LA compiled by Louis Godart and Jean-Pierre Olivier, but its tendency to standardise elides important distinctions.[1] It is these which Salgarella highlights, especially between sites. While many of the simple signs have a wide distribution, a significant number are limited to a single site, a trend even more marked for the composite signs. This regionalism plays an important part in her final argument.

This section is, unlike the previous, not illustrated in the book itself, but in a 364-page appendix available online. The decision is understandable, given the number of signs that must be illustrated, but the PDF is very buggy and cannot be used reliably. This shortcoming is more than mitigated, however, by the excellent SigLA website developed by Salgarella and Simon Castellan, which offers an easily searchable database of LA inscriptions and sign forms. [2] Though still under development, it is already an invaluable resource.

In an important palaeographic study, Thomas Palaima demonstrated that LA looks more like LB in its latest stage (Late Minoan IB) than at any point previous.[3] Salgarella, conversely, does not much consider chronology, but is able to demonstrate a geographical pattern: while many sign variants are shared across sites, Ayia Triada and Phaistos appear somewhat isolated palaeographically, with many variants not passed down to LB. Given the size and date (LM IB) of the LA deposit from Ayia Triada, these findings are quite significant.[4] Rather, LA as written at Khania and Zakros looks the most like LB, both in signs attested and how they are written. We shall return to this idea.

In the next chapter, Salgarella compares her LA data with two LB deposits from Knossos. These are the Room of the Chariot Tablets (RCT), the earliest major deposit of LB (LM IIIA), and the somewhat later deposit found in the North Entrance Passage (NEP) and North-West Insula (NWI). The addition of this second deposit offers a valuable diachronic perspective on the development of LB.

As LB does not create composite signs with the freedom or frequency of LA, Salgarella focusses her structural analysis on the syllabograms and logograms attested in these deposits. Syllabograms in the RCT are sometimes used as logograms, which Salgarella suggests may be a survival of LA practice. But it is striking how few of her examples are limited to the RCT (216-7): the practice appears to be much more standardised in LB than it was in LA. LB logograms show significant continuity with LA, though there is now far less ligaturing. As ligatured signs in LB typically have a phonetic interpretation, it is not surprising that there was no carryover from LA, where the same was presumably true. At the structural level, however, they are combined according to the same logic as LA ligatured signs were. This is also true of ligatured signs attested first in the NEP/NWI, and while their absence from the RCT may be an accident of preservation, Salgarella makes a compelling case that the principles of sign combination used in LA were fully integrated in LB.

Palaeography has a privileged place in LB studies but is rarely discussed explicitly. Salgarella’s review of methodologies and approaches is therefore very welcome (233-44). The identification of differences in how signs are drawn is the cornerstone of scribal attribution (as, of course, Mycenaean scribes did not tend to sign their work), but Salgarella focusses instead on those differences themselves. Her goal is to determine where these ‘variants’ originated: were they inherited from LA, or developed within LB? Unlike other recent studies, she uses signs from tablets not assigned to any known scribe, which provides a much more comprehensive dataset.[5]

The palaeography of the RCT is dispatched in a surprisingly brief four pages, with a focus on variants that look most like LA, and the word a-qi-ru (KN Ce 50), which attests a traditional Minoan spelling and whose qi is a LA variant not preserved in later deposits.[6] The NEP and NWI are treated more systematically, and Salgarella demonstrates that many LA variants are still preserved at this later date, including some unattested in the RCT (though this may be an accident of preservation). By demonstrating this degree of palaeographic continuity, Salgarella is able to prove that LB never had a standardised sign list. Rather, it seems, scribes who had written LA were the first to write LB, using the same variants they had before.

In Chapter Four, Salgarella ties all of her ideas together. Looking back to the regionalism of LA, Salgarella argues that we should think not of LA, but rather ‘LAs’, which share a core of signs but add new ones (both simple and complex) as needed, an argument she supports by contextual and palaeographic analysis of the site-specific LA signs. By establishing that most of these signs do not have (attested) Cretan Hieroglyphic parallels, she makes a strong case that new signs could be invented at individual sites to meet local needs, which has important implications for how we understand its regionalism.

Salgarella then makes her strongest claim: LB, with its shared core of signs and new additions, simply represents the final development of LA. It is not a regional variant, as the other ‘LAs’, but a chronological one. She proposes, therefore, to speak not of LA and LB but a single ‘Aegean Linear Script’. Once again, Salgarella points to the similarities between LB and the ‘LAs’ at Khania and Zakros and suggests ‘LB is likely to have started off under the influence of local areas from the [north and east of Crete], especially Khania and Zakros’ (361). Ultimately, however, it is Knossos where Salgarella places the development (362), which, frustratingly, has yet to yield a LA deposit from LM IB. This is the single biggest problem for anyone assessing the relationship of LA to LB. Salgarella’s focus on Khania and Zakros, therefore, seems like an attempt to triangulate the palaeography and repertoire of Knossian LA. Roughly equidistant between Khania and Zakros, and probably the centre of gravity in the region, that Knossos should be the common factor between these two and LB seems quite plausible.

Salgarella’s idea of a single ‘Aegean Linear Script’ relies on her conceptual division between the script (the constitutive elements) and the writing system (the rules which govern its use). I am quite convinced that the script itself was not significantly altered or standardised at the transition from LA to LB. But, as Salgarella herself acknowledges (373), there are significant structural changes in the use of signs between LA and LB administrations, entirely different systems of weights and measures, and distinct tablet layouts. A more holistic assessment, which considers these elements alongside palaeography, paints a more complicated picture. Moreover, LB does not show the same range of uses or regional variations that we see in LA(s). While the lack of Knossian evidence makes it hard to be certain about the extent and pace of change, there was at least a significant bottleneck in both variants and variation at the transition to LB. The most likely explanation, it seems to me, remains a deliberate programme of script reform.[7]

A global study of LB palaeography remains a desideratum, and (in addition to her stated goals) Salgarella has provided an excellent starting point. Of special interest will be comparison with LB on the mainland; while Salgarella’s work contributes to the growing consensus that LB was developed on Crete, how it arrived on the mainland has received less investigation. Of interest here is Christina Skelton’s work, which has shown that the early Hand 13 at Pylos preserves some LA variants not found in Knossian LB.[8] This suggests that the script made it to the mainland early and the so-called ‘mainland koine’ (360-1) may have evolved there, rather than reflect the introduction of the script at a later stage when it was already more standardised.

If I am not inclined to accept all of Salgarella’s conclusions, that is no indictment of the quality of her work. The quantity of evidence collected, and Salgarella’s sure guidance through it, mean this book should become a standard reference for Mycenologists. We can be very thankful to Cambridge University Press for publishing such a technical study.[9] As a final note, Salgarella’s approach and findings should have significant interest for everyone interested in the development of new writing systems from existing models (Secondary Script Development). Her focus on similarities in scripts across linguistic boundaries, and demonstration that palaeographic variants may be transmitted, provide both a model and a challenge.


[1] Godart, L. and J.-P. Olivier (1976-85) Recueil des inscriptions en Linéaire A, vols I-V, Études Crétoise 21: 1-5, Paris.

[2] Introduced here.

[3] Palaima, T. (1988) ‘The Development of the Mycenaean Writing System’, in Olivier, J.-P. and T. Palaima (eds.) Texts, Tablets and Scribes: Studies in Mycenaean Epigraphy and Economy offered to Emmett L. Bennett, Jr, Minos Supplement 10.

[4] That the LA from Phaistos looks little like LB is less meaningful, given its earlier date.

[5] As opposed to other studies, which ascribe one variant to each scribe and take no account of unattributed tablets, such as Firth, R. and C. Skelton (2016) ‘A Study of the Scribal Hands at Knossos based on Phylogenetic Methods and Find-Place Analysis’ Minos 39, 159-228.

[6] This is a masculine name often equated with later a-qi-ro; if that is correct, the original Minoan nominal ending (-ru) is assimilated in time to the Greek second declension.

[7] Salgarella cites in favour of a longer process the inscribed sign on the ‘kessel’ from Shaft Grave IV at Mycenae, but neither this nor the LA inscription from Ayios Stephanos in Laconia has anything much to do with LB. See Nash, T. 2021 ‘Cultures of Writing: Re-Assessing the ‘Spread’ and ‘Development’ of Scripts in the Bronze Age Mediterranean’, in Boyes, P., P. Steele and N. Elvira Astoreca (eds.) The Social and Cultural Context of Historic Writing Practices, 219-22.

[8] Skelton, C. (2010) ‘Re-Examining the Pylos Megaron Tablets’ Kadmos 48, 107-23.

[9] The text, however, maintains the appearance of a dissertation – 12-point font, Times New Roman, double-spaced – rather than a professionally typeset book, in contrast with other volumes in the same series, such as Anna Judson’s nearly contemporary The Undeciphered Signs of Linear B (CUP 2020). There are very few typos: I note only a comma for a period on 132, while the constituent parts of A 565 (A86 and A188) are not identified on 166. Real errors are limited to ma for mo at the bottom of 258 and a missing image of qe on 257. Foreign languages (French and Italian) are not translated, but the required knowledge of LA and LB epigraphy is a much higher barrier to accessibility than that.