Sunday, June 5, 2011

A new map of Middle Minoan Crete - Assessing the place-names on vessels inscribed with Linear A

Greetings to all my readers once again. In the previous series of posts, we have seen ample examples of place-names on the Linear A tablets. But miscellaneous objects inscribed in Linear A were not yet discussed. As we shall see, Cretan vessels - both religious and profane vases - will turn out to be a real treasure trove of Minoan toponyms. And since they were found all across the island, they can be used to map out where these places were actually located!

There are a number of reasons why jars or amphores would be inscribed. The most practical reason is that it makes them easier to administer, just think about trade! This is the very reason why so many vases were found from the Mycaenean era, containing in scriptions in Linear B. Kim Raymoure has a nice collection of such jars-texts on her website. These texts are typically short, consisting almost exlcusively of names: anthroponyms (personal names), toponyms (place-names) or a combination of both, with some titles mixed in. In case only a single word is inscribed, it is most frequently a name of a town. Obviously, they describe the provenience of the vessel or the producer of its contents. Finding such jars on one place inscribed with the name of another is a clear indicator of trade relationships; and can be used to map out ancient trade routes.

Religious vases and altar-stones were inscribed for different reasons. Given that most ancient (and modern) temples acquired their prosperity and material wealth through donations, many objects that are inscribed contain the name of their donors. The more important sanctuaries could amass a respectable amount of goods through the centuries or millennia. It is enough to take a look at the ruins of the Oracle at Delphi; where most of the ruins enclosed within the temenos wall belong to treasuries from various polities. Athens erected a separate building for them, so did Sparta, Argos, Thebes and Corinth. Even smaller polities, like Siphnos or Sicyon had their very own treasury constructed, and the sanctuary received items from as far as Knidos, the opposite end of the Aegean Sea. Obviously, the "attraction radius" of a sanctuary was proportional to its imporance: minor temples might have received donations only from their immediate surroundings. Given this tradition of state (or polity) gifts, finding toponyms on materwork Minoan vessels that once served as libation cups or portable altar-stones (the so-called libation tables) is the least surprising discovery.

First of all, take a look at the stone libation vessel found at Apodoulou (slightly north-west from Phaistos). This cup has a number of interesting phrases on it (see figure). The key word is I-PI-NA-MA, that is repeated in the lower line as I-PI-NA-MI-NA-TE (restored reading). But it is a toponym that is most intesting. One of the words very clearly reads I-KU-PA3-NA-TU-NA-TE (the first sign has only its corner visible). This very name returns as KU-PA3-NA-TU (without the *i- prefix and the *-ate suffix) on Haghia Triada tablets HT47 and HT119. The latter tablet probably lists people by places (it was not included in my previous lists due to the ambiguous topic). It is very unlikely that the name would be independent of KU-PA3-NU, another putative place-name: rather, it just seems to be a regular variant. On the tablets, KU-PA3-NU very frequently groups with genuine western Cretan names (e.g. KU-DO-NI); this could mean if KU-PA3-NU and KU-PA3-NA-TU are one and the same (or two, directly next to each other), they should definitely lie west of Phaistos, probably in west-central Crete.

References to a place SU-KI-RI-TA are encountered on vases found at Phaistos and Haghia Triada. This time we have an easier job: SU-KI-RI-TA is not only commonly mentioned in Linear B at Knossos (when it was apparently a local province capital of some sort), but the place is extant: it is none else than Classical Sybrita, modern Syvritos. Its location south-west of the Idaian range can explain the distribution of its references reasonably well.

Turning to the Linear B documents, apart from SU-KI-RI-TA and PA-I-TO, there is a third place that is mentioned as having its very own QA-SI-RE-WI-JA (local chieftaindom): SE-TO-I-JA. Despite the obvious similarity with the name of modern Sitia (ancient Séteia), this identification is not necessarily straightforward or correct. SE-TO-I-JA never groups with eastern Cretan places on the Knossos tablets, and in Linear A, it is mentioned only on a libation table found at Prassas, next to Knossos. A second, doubtful mention could be on the libation table found in the Psychro cave (PS Za 2), where a word [?-?-?]-JA-TI was restored as SE-TO-I-JA-TI by Gareth Owens, based on the length of the missing fragment and the rarity of other place-names in Linear A ending with -JA. Nevertheless, his identification of SE-TO-I-JA with Archanes is questionable: Why would a town so close to Knossos be a local province capital? Judged by the considerable distance of Sybrita (westwards) and Phaistos (southwards) from Knossos, it is more plausible that SE-TO-I-JA was a key city on eastern Crete, perhaps lying at Mallia or even further to the east. These local centres are seldom mentioned on Linear B place-name lists, making their localization difficult by groupings alone. Therefore I do not yet discard the original hypothesis of placing SE-TO-I-JA to Sitia (i.e. the Minoan site of Petras, near Sitia).

On a mundane amphore from Tylissos, another interesting term can be read. The text consists of a single word: A-[*]-KI-TA-A: The sign originally standing at position * was probably *306 (it is still partly visible), before it was erased and changed to *301. Because we already have a putative toponym from the Haghia Triada tablet HT122 in the form [?]-*306-KI-TA2, the correcture of the scribe was likely erroneous. While the reading of Lin A *306 is officially "unknown", it very closely corresponds to Linear B sign *42, that is, WO. Such a phonetic value is not unlikely in Linear A, either, because semivowels, including approximants are commonly seen in word-initial positions. Yet even if the stem word was indeed WO-KI-TA, it is hard to identify it with any Cretan place. Well, unless "Luke" is a "wookie" [StarWars pun intended], in which case *Wúkita could be Mycaenean Lukitos (Lin B RU-KI-TO), modern Lyttos. Unfortunately, while the change of laterals into approximants is common in all languages of the world, I have no idea if the reverse process could ever happen. But at least at some rare borrowings, the initial *w- can change over to other consonants, as the example of the Behistun inscriptions show: the Old Persian name Wishtashpa (= Greek Hystaspes, the father of Great King Darius I) is repeated in the form Mishtashba in the Elamite text. (Use of the same cuneiform sytem makes a scribal error extremely unlikely.) Therefore I do not discard this theory, that gets further support from the Knossos archives: Lyktos is probably the most commonly mentioned place on all tablets. It likely also had strong trade connections with Tylissos, as these places are frequently mentioned together. This could easily explain why excavations have discovered a vase at Tylissos, imported from Lyktos. And at least on a single tablet, Lyktos is also listed together with Daos (perhaps Haghia Triada itself). But even without direct identification, the distribution of places where *306-KI-TA was mentioned would place *Wúkita somewhere into central Crete, close enough to Lyktos anyway.

Speaking about Haghia Triada, it is an interesting fact that none of the place-names mentioned on the HT tablets can be equated with the town itself. From the Linear B material we know however, that Phaistos did have a sister-town called Daos (DA-WO, almost always paired with PA-I-TO). Yet by sheer luck, there is a single fragment of a libation vessel found at Knossos (KN Za 10, see figure), that contains a severely damaged series of signs; the term DA-WA-[SI?] is still readable. If interpreted correctly (I am uncertain if putting an "arbitrary" divisor to the broken off segment was right), that could mean that it was *Dawa, that is Daos (i.e. Haghia Triada) that donated this stone plate to the temple of Knossos and the offerings it once held to please the gods.

The original Minoan name of Tylissos (Linear B TU-RI-SO) is similarly difficult to find out. In this case, the documents supply us with not one, but two candidates. One of them is DU-RE-ZA, a toponym on the clay tablets found at Khania and Zakros; the other one is a certain TU-RU-SA mentioned on a vessel at Kophinas and also at Knossos (in the form of A-TU-RI-SI-TI). I have no idea whether *Duletsa or *Tursa is a better match for Tylissos: I leave it to the reader to decide which one looks like a better candidate.

The site of the mountain-sanctuary of Syme also bestowed us a number of probable toponyms. Out of these many, the term PA3-NI stands out. This place is also frequently mentioned at Haghia Triada, as a donor of specialized agricultural goods (such as figs, several types of grain, malt, etc.). The occurrance of this term at Syme (one time securely on SY Za4, and possibly another time on SY Za7) hints that this was a place at mid-eastern Crete. Perhaps it is not an overtly bold step to search for PA3-NI in the Hierapetra region. A large settlement at that time: Gournia is definitely a good candidate. It is also worth to note that PA3-NI is very frequently paired with DI-RI-NA (*Drina) on the Haghia Triada tablets. Eerily enough, there is a small town called Prina roughly 15kms west from the excavation site of Gournia, but I am not sure if that town's name is ancient.

Some of the terms provisionally transliterated by John Younger (whom I cannot thank enough for providing on-line access to the original images) need slight corrections. One of the terms, now read as JA-PA-RA-JA-NA-SE is especially interesting, as it seems to recall the same stem as the famous historical polity Praisos has. The town of Praisos has at least 4000 years of history: Several middle and late minoan ruins were excavated in the region, at Zou and at Praisos itself. Later into the classical era, Praisos was one of the last strongholds of Eteocretans and their poorly understood language. Nevertheless, the term *A-PA-RA-JA (*Apraya) looks like an un-derived original version (i.e. without the *-(i)ssos ending), but it is suffixed similarly to the Eteocretan ethnic term Phraisona.

Two eastern cretan locations: Petras and Palaikastro supply us with references to a place called *KI-TA-NA. Because this very term also appears on a number of hieroglyphic seals (e.g. from Mochlos and Sitia), it must have been a place of paramount importance. There is one more-or-less obvious candidate on the eastern end of Crete: the palatial site of Zakros. This toponym is not seen anywhere in the Linear B archives, which is explained well by the fact that the town of Zakros and many other places lay in ruins and were completely uninhabited by that time. Note that the name of Palaikastro (Greek: "Old Castle") is not ancient, either, but I was unable to find any reliable reference to that in Linear A materials (yet it is likely that the town is mentioned in Linear B). In contrast to that, the name of another mid-eastern Cretan town: Polychna (perhaps modern Vryses, near Mallia) returns as PU-RE-KA-NA on one of the Hieroglyphic seal impressions found at Knossos. The name *Pulekna (which I initially incorrectly assumed to be a personal name) shows a very nice correspondence with Polychna, the latter one seems to be a hellenized version (as *pule- is meaningless in Greek, but *poly- would mean "many"). Simiar warpings can be found in the names Aptera, earlier Aptawa, Linear B A-PA-TA-WA (πτερος = "wing") and Hierapetra, previously Hierapythna (πετρα = "stone"). These un-systematic changes testament the process how certain, originally non-Greek names became established in Hellenic dialects. But the very same fact makes their reconstruction difficult. Yet we have seen it is not impossible, fortunately for us.

As a special gift, I made a supplementary figure - a map to show all places we have talked about at once. Although I did not discuss that before, it also displays the potential location of KU-DA (HT122), likely the same as classic Kytaion (Lin B KU-TA-I-TO ?), as well as DA-RE (potentially Tarra, on south-western Crete).


  1. TU-RU-SA to me looks like "Troy" (Hittite Taruwisa). So that would seem to leave DU-RE-ZA as Tylissos.

    However since I have my eyes set on a tall two-handled vase (KN Zb 5) with A-TU-RI-SI-TI (= *Atúlisi-ti 'in Tylissos') written on it, I'll have to wager that DU-RE-ZA is yet a third place which I render as *Θureza or *Θuleza with initial dental fricative. However I have no clue yet what it might be. Note DU-RE-ZA-SE [ZA 10.a5, 10.b1] with added -SE.

  2. I think both A-DU-RE-ZA and DU-RE-ZA are far more likely A-DU RE-ZA and DU RE-ZA, as far as parsing goes. RE-ZA may even be subordinate to (A-)DU in some way.

    My reason for suggesting this lies on HT 88. Line 1 has A-DU VIR+KA and RE-ZA split only by a single quantity. That we find these two digrams on the same tablet is hard to overlook when considering how best to parse A-DU-RE-ZA.

    What do you think?

    Glen's comment about DU-RE-ZA-SE is an interesting one. It calls to my mind the ZO-xx-xx-SE anthroponyms of Cypriot that I've encountered in the Palaeolexicon index, though I admit I've spent very little time wandering through Cypriot sign groups. Thoughts?

  3. I hadn't seen that Younger added online access to the images! I must send him a thank you note :) I started begging for that ... gosh, a lot of years ago. *happy dance*

  4. I very much like your argument for KI-TA-NA as Zakros, and promise to keep that in mind in my own studies and let you know if I encounter any evidence or counter-evidence. Great post, András!

  5. I red all your article, unfortunately Andras, everybody knows that EteoCretan language had had anything to do with IndoEuropean languages' family. Any connection with Greek language it must be seeing as a substrata word in Greek. Also Linear B connection still under investigation.
    Further then, I start my research paper from Beekes'description, and I can say, it is a good point of view.