Dr Albert Tiller
EVIDENCE OF A HEXAMETER IN LINEAR B?
Following its first publication by Gottlieb von Eisenmann in Minos (1961), the Chania tablet KH Kr 144 attracted no special attention from researchers for nearly two decades until the 1978 article of Mirko Bezruč and Karl Grmek,[1] who pointed out the frail foundations of some of von Eisenmann's conclusions. It is to be borne in mind that von Eisenmann's interpretation still enjoys the reputation of being universally accepted, though the fact that Bezruč's entirely different assumptions lead to an internally consistent analysis should, and did, call for caution. Thus the existing interpretations, including the two articles of Miss Eleanor Dilley[2] and the hypothesis by Hector Etherington,[3] are no more than tentative in nature; which fact leads to the necessity of reviewing all existing hypotheses (as this is the name they properly deserve, rather than being universally accepted opinions). First of all, let us return to an inspection of the original tablet.
Fig. 1. Tablet KH Kr 144. Courtesy of the Εθνικό Μουσείο, Athens
For a photograph of the tablet, see Fig. 1. The tablet is in a gravely damaged state. The surviving part of the tablet (145 × 47 mm, 523/32 × 127/32 inches) consists of eight small fragments bearing three lines of text. Eisenmann's transcription is as follows:
ku-ru-ti a2-re ke-ka-u-je-u-ka a-i-[..] wa-na-ka-to mulier vir
[.....] ko-wo si-ri-to [..] a-we-re-e [...] vir
[.................] -we-jo cornu
In broad terms, the history of the interpretations is as follows. In his editio princeps, Gottlieb von Eisenmann gave a cautious interpretation. According to him, ku-ru-ti is a place name in the dative-locative, and a2-re, a proper name in the nominative. He also mentioned a possibility - phonetically feasible, but still dubious - of identifying the name with Ἄρης. He did not make an attempt at a full interpretation of the next word, ke-ka-u-je-u-ka, which strongly resists identification, although he implied a possibility of it being an anomalous case of a profession name in -ευς of the root of καίω (< *καϜjω), preceded and followed by modal particles - this last point was perhaps the least well grounded supposition of Eisenmann's entire interpretation. The next word in the series, a-i-, is damaged (about two characters illegible), so Eisenmann attempted at tentatively restoring it as a form of the verb αἴθω,[4] presumably the predicate verb, meaning the action done by a2-re the ka-u-je-u. The last word of the line, wa-na-ka-to, actually the only clear word in the whole tablet, is doubtless the genetive singular Ϝάνακτος; the anomalous spelling -ka-to (instead of expected -ko-to) can be explained as the scribal habit of modelling the whole declension of the word after the nominative wa-na-ka.[5] Eisenmann assumes that the ideogram of man at the end of the line is connected with the man called a2-re. The start of the next line is damaged (some six to eight characters missing); ko-wo evidently stands for κόρϜος, although the case is not absolutely clear; si-ri-to, according to Eisenmann, should be understood as the adjective describing the ko-wo - but cf. Hector Etherington's interpretation, which sees here a prefixed form of the verb λείπω (< *likw), a supposition which lacks probability and has never been accepted, especially after Miss Dilley's criticism.[6] The last legible word, a-we-re-e, is almost certainly the infinitive ἀϜείρεεν ('αἴρειν') - but again cf. Etherington's counter-suggestions,[7] which were exploded by Eleanor Dilley. The ideogram of man which closes the second line, after a significant gap, is obviously connected to ko-wo. The fragmentary word we-jo in the last line, according to Eisenmann, could be an adjective in -ειος.
Thus, according to Eisenmann, the whole tablet is an account of the cremation of corpses, possibly of a king's daughters and sons, performed by the cremator called a2-re (Ares ?) in the place called ku-ru-ti.[8]
Shortly after the first publication by Eisenmann, an article appeared by a young archaeologist, Dr Kemal Yigitoglu,[9] who made an attractive hypothesis that the tablet could be an invocation to the god Ares; unfortunately, this skilful but premature guess did not receive any response, especially after the severe criticism of Eisenmann, whose authority was considered unshakable. However, Eisenmann could not make any formal objections to Yigitoglu's reading of the first two words as κλῦθι, Ἄρης, except, as he said, that such an invocation would be 'very atypical' stylistically.
An entirely different approach was taken by Grmek and Bezruč, two Czech scholars whose article, originally written in 1978, was not known to western scholars until 1989, when it was translated into English and published in Acta Minoica. Grmek and Bezruč first of all pointed out the dubious interpretation of ku-ru-ti as a place name in the dative-locative, for this purpose assuming that the ending -ει should be expected in that case rather than -ι, on the basis of di-we = ΔιϜεί etc., so the dative ending -ι should be considered anachronistic here.[10] This led the two Czech scholars to an essentially different concept, that ku-ru-ti could be a female name of the -ιδ stem in the nominative, which they somewhat hastily interpreted as Γρῦπτις, 'the Bent Lady', a name describing some kind of an obvious physical defect, presumably a hunchback. Another difference in Grmek and Bezruč's reading from that of Eisenmann was that they took ke-ka-u-je-u-ka as a verb in the perfect tense, as obviously exhibiting the perfect reduplication and the suffix -κα. The Czech scholars also suggested that the verb could be an anomalous reduplicated form of χέ(Ϝ)ω, and as an argument they presented anomalously reduplicated Homeric forms. The most daring innovation of Grmek and Bezruč was, however, a different conjecture as to the missing characters after a-i- as -qo-ro, thus producing a-i-qo-ro 'αἰπόλοι' (but cf. Miss Dilley's observation). The difficult word si-ri-to also received a new interpretation from the Czech scholars. They interpreted the syllable si- as the abbreviation of ἱερός (<*σισερός?), although they did not clearly state whether it was just a scribal abbreviation, or also a phonetic one (as implied by their examples of Homeric δῶ < δῶμα, κρῖ < κριθή). The word a-we-re-e was understood by the Czech scholars, in concord with Eisenmann, as ἀϜείρεεν, which, they believed, had some religious meaning in the context.
Thus, the Czech scholars' interpretation resulted in the following translation: 'Gruptis [and] Ares (represented by the respective ideograms), the king's shepherds (αἰπὸλοι), shed [a libation?] on the altar-stone (σι[ερὸς] λίθος) of the ever-young (κόρϜος) [god].' The last fragmentary word of the tablet, to their opinion, could have been an epithet of the 'ever-young' deity. Both Miss Dilley[11] and Hector Etherington pointed out the artificiality of the translation of the second half of the text; even Grmek and Bezruč themselves described it as 'proposed' and 'tentative'.
The source of such aberrant interpretations, in our opinion, was the attempt of the researchers to fit the facts to the theory, which states that Linear B tablets can contain only simple sentences or nominative phrases, where there is no place for imperatives, invocations or any other traces of a style and syntax even slightly above simple lists of goods and personnel. In contrast, we decided to look at the tablet with a fresh eye; and what first attracted our attention was the striking plausibility of Dr Yigitoglu's proposed reading of ku-ru-ti as the athematic imperative κλῦθι, a form attested 12 times in Homer.[12]
This first step led to a completely new understanding of the whole text. Another important milestone was a sudden realisation that ke-ka-u-je-u-ka does not necessarily have to be one word; thinking about the reason of the absence of the word boundary, we came to the conclusion that it would be justifiable to take it as a way of indicating an elision.[13] The part ...e-u-ka bears enough semblance to the root εὐχ- (attested in Linear B as e-u-ke-to, etc.), so we decided that the word (more exactly, the phrase) might be separated into ke-ka-u-j' e-u-ka. The first of these words bears a striking resemblance to an elided feminine perfect participle, cf. attested a-ra-ru-ja, while the second one is perspicuously a case of εὐχά. Although the participle form γεγαυῖα is not attested in alphabetic Greek, the Homeric poems mention the prefixed form ἐκγεγαυῖα six times;[14] which fact, in our opinion, satisfactorily justifies the supposition.
Thus, the present state of the text gives us an invocation (κλῦθι) addressed to someone described by the participle γεγαυῖα. Since the latter is feminine, and we still supposed, following Yigitoglu, that a2-re must be vocative - that is, the person invoked - we had to reject Ares and search for another reading. We could not reach any satisfactory results until we abandoned the idea of a2-re being a vocative. The ending -e could represent -ει, that is, a dative-locative inflection; so we have ἁλεί 'in the sea' - although not attested, but to our opinion satisfactorily plausible, cf. attested di-we = ΔιϜεί (= Attic Διί), etc.
Thus, so far we had:
Κλῦθι, ἁλεὶ γεγαυῖ', εὐχᾶς αἰ[...] Ϝάνακτος
'Listen, sea-born, to the prayer... of the King...'
What caught our eye was the semblance of a metrical pattern among the first few words. The fragment κλῦθι, ἁλεὶ γεγα... was very strikingly two dactyls. This encouraged us to try to work out the metrical scheme of the whole text, so far deciphered. The result was as follows:
- ∪ ∪ | - ∪ ∪ | - - | - - | [ - ∪ ? ] ∪ | - ∪
There seems no doubt that we are dealing here with a hexameter line. The word Ϝάνακτος seems to be a plausible hexameter ending,[15] as well as the word κλῦθι, which always occurs at the start of a line in Homer. Therefore the two missing characters after a-i- should constitute one long and one short syllable to complete the hexameter line.
How could the fragmentary a-i- be restored? Very often the word ἄνακτος occurs in Homer together with the proper name of the king.[16] Could it be supposed that it is so in this case too? Miss Eleanor Dilley mentioned that the remainder of the first missing character was suggesting of ke; thus, a-i-ke-?[17] The idea of the name Αἰγεύς thrusts itself forward, so that the line can be tentatively completed:
Κλῦθι, ἁλεὶ γεγαυῖ', εὐχᾶς ΑἰγῆϜος Ϝάνακτος.
The reproach could be made here that the final -ς of ΑἰγῆϜος meets the initial Ϝ of Ϝάνακτος, and the syllable remains short. However, this seems to be satisfactorily explanable, assuming that the syllable boundary here could have been taken as Αἰ-γῆ-Ϝο-σϜά-..., thus keeping the syllable Ϝο short.[18] Indeed, such a view can be illustrated by examples from Homer, in lines which are surely not later additions:
Κοπρῆος φίλον υἱόν, ὃς Εὐρυσθῆος ἄνακτος Ο 639
ἤτοι μὲν γὰρ ἔναντα Ποσειδάωνος ἄνακτος Υ 67
Here we see that the final -ς of the proper name meets the reconstructed digamma (not printed in traditional texts) of (Ϝ)ἄνακτος, and the syllable preceding Ϝα- remains short. In fact, there is no other way of juxtaposing the genitive of a proper name of the third declension to the word (Ϝ)ἄνακτος without creating the consonant cluster -σϜ-. However, as said, that should not be considered as an argument for the 'inauthenticity' of such a line.
Now let us look at the remainder of the text: ko-wo is certainly κόρϜος, although not necessarily in the nominative case. The following word is si-ri-to: as it seems Bezruč was very close to the truth, having seen traces of λίθος here. The etymology of λίθος itself unexpectedly gives the explanation of the si-, which had been puzzling scholars for a long time. Frisk associates λίθος with Lithuanian root of slidùs;[19] so it is to be concluded that in Greek, the root once must have had an initial s- too, viz., *σλίθος - not attested in alphabetic Greek, but cf. *σλαμβάνω, σμικρός, etc.[20]
As it seems, Bezruč was also right about a-we-re-e = ἀϜείρεεν. Thus, the remainder of the text makes up what could also be considered fragments of a hexameter line:
...κόρϜωι σλίθον [..]ἀϜείρεεν...
[- ∪ ∪ | - ? ] - | - ∪ ∪ | [- ∪ ?] ∪ | - ∪ ∪ | [- ∪ ?]
The two missing characters between σλίθον and ἀϜείρεεν apparently have to be a disyllabic (- ∪) word starting with a vowel, as implied by the metre.
It is hardly possible to tell anything about the fragmentary third line. If we accept that wejo is -ειος (as Eisenmann pointed out decades ago), then it should fit as the end of a hexameter line. The meaning of the cornu ideogram remains unclear.
Thus, in this tablet we have perhaps the earliest known fragment of written poetry.
At this point it should be said that this is not the first time that the presence of verse has been postulated in Linear B, at least hypothetically. The remarkable article of Ion Ciorbea[21] can serve an example, where he not unmotivatedly claims to have discovered a stinging satirical epigram in the tablet PY Ep 704 and, partially, in other tablets of the PY Ep series. The tablet, which since the time of Ventris has been believed to be a land property document, according to Ciorbea contains an epigram against a covetous priestess who collects a rich income from the worshippers of her deity, although she publicly swears to have no profit from her cult. This is Ciorbea's proposed translation of the tablet: Eritha la prêtresse jure qu'elle a un dieu vain, mais le peuple dit qu'elle a plus de bénéfice [de lui] que de toutes les parcelles de terre sacrée.
It is also to be mentioned that, although in general Ciorbea's hypothesis presents a fresh and not groundless view, some of his further conclusions deserve less trust. First of all, what he believes to be the 'strophe minoenne' strikes us by its oddity: as deciphered, it consists of one line of anapaestic hexameter and two lines of iambic pentameters, a form of poetry not attested in later Greek.[22] On the other hand, the presence of a metric pattern, especially the anapaests in the first line (Ἐρίθα ἱέρεια ἔχει <ε>ὔχετοί τε hετώνιον...), can be instantly seen, and it is surprising that it has not attracted anyone's attention before. However, Ciorbea's suggested synizesis of the gen. pl. -άων, as well as some other things, still provoke doubt.
Another, and probably major, reproach to Ciorbea's decipherment, as well as the traditional one, of the 'Eritha' tablet would be that, to our conviction, there are no grounds to consider e-ri-ta as the name of the priestess. In Homer, the word ἔρῑθος 'hired labourer, servant' is attested twice in the Iliad.[23] Would it not be more logical to suppose that here we are dealing with the feminine of the word, that is, a 'female servant'? Thus we have a very prosaic ἐρίθα ἱερείας 'servant of the priestess' - compare i-je-re-ja do-e-ra, universally believed to be ἱερείας δοέλαι (PY Ae 303, etc.). Such an arrangement, however, would spoil Ciorbea's perfect anapaests, since the ι in Homeric ἔρῑθος is known to be long, and so is the genetive ending in ἱερείᾱς. Nevertheless, Ciorbea's hypothesis (not to say discovery) deserves the close attention of mycenaeologists, especially his statistics on the repeating metric patterns in the Ep tablets of Pylos.
To come back to the Aigeus tablet, the question of the precise significance of the deciphered fragment now needs to be considered. First, does it actually refer to Aigeus, and in particular to the legendary king of Athens? Little of substance is known about this shadowy figure. Nearly all of what is known is stated in Plutarch's life of Theseus, compiled many centuries later from sources which are now lost. Plutarch, among numerous others,[24] recounts the legend (oddly prefiguring the story of King Arthur) of the son who shows his lineage by lifting a stone to reveal his father's sword and shoes. The three surviving words of the second line can certainly be read as alluding to this legend.
If the fragment does refer to the Aigeus-Theseus legend, can anything be made of the appeal to a deity in the first line? Plutarch mentions only Aigeus' visit to the Delphic oracle, which in any case is probably a later interpolation into the myth. However, Pausanias mentions that Aigcus established at Athens a cult of Aphrodite Ourania,[25] a Palestinian avatar of the goddess. 'Sea-born' naturally suggests Aphrodite, though we should remember that this fragment, however uncertain its date (13th century BC, according to Eisenmann), is at least five centuries older than Hesiod's Theogony, and it would be unwise to jump to conclusions.
Leaving aside the question of the identity of the person addressed, if we accept that these lines constitute the earliest reference to the legend of Aigeus and Theseus, what can this tell us about the nature and purpose of the tablet itself? Clearly, since it was found at a site in Crete, it cannot be a votive tablet of Aigeus himself (even assuming that he had a real existence). Yet, if our reading is accepted, it is a direct appeal to a deity. This can only mean that its purpose is narrative: it is a historical or mythical account of the life of Aigeus. Further, its apparent verse form suggests that it is part of a narrative poem.
A narrative poem in hexameters is, of course, not necessarily an epic. But epics such as those of the Trojan cycle must have grown from accretions of rhapsodic ballads, presumably in hexameter form. It seems reasonable to suppose that early examples of these existed before the fall of the Mycenaean civilisation in the 12th century BC, some four centuries before the generally presumed dates of Homer. As is well known, usually at this time these ballads would have been orally transmitted, but we know from the Iliad[26] that writing was used at least for messages before the Iliad itself was written down.
That leaves the question of the purpose of making the tablet, about which it is impossible to say anything with any confidence. It could be supposed that the tablet represents a scribal training exercise, which consisted of rewriting lines of well known poems (which nevertheless were usually delivered only orally); or that it served as an amulet, inscribed with sacred formulas; or that it was indeed a votive tablet, supposed to be read by the deity; or that it was used as a memory aid at some ceremony, when Aigeus' invocation had to be solemnly declaimed.
The above is highly speculative reasoning, but it must be admitted that it provides a more credible interpretation of this puzzling tablet than the suggestions of von Eisenmann or other scholars. If it is accepted, what we have in this fragment of clay is nothing less than the earliest piece - by many centuries - of the entire epic tradition.
[1] Bezruč, M. and Grmek, K., Gruptis and Ares, the King's Shepherds, AM XXVI.2, 1989, 17 ff.
[2] Dilley, E., A Few Remarks Concerning the Interpretation of Selected Chania Tablets, Revue..., XLII, 1961, 27 ff.; A Reconsideration of the Chania Tablets, Minoica, 2 (154), 1990, 35 ff.
[3] Etherington, H., Evidence of Rituals at Chania, Act. APM, LIV.3, 1962, 95 ff.
[4] But confusingly, Miss Dilley, who herself was Eisenmann's student, mentions in her first article that after close inspection of the original tablet she concluded that the remainder of the missing character suggested ke rather than any other character. Unfortunately, the original tablet was not available to us, so we could base our own interpretation only on the present photograph. As can be seen, the photograph is not of much help on this question.
[5] Cf. also the mention in: Hooker, J. T., Introduction, 1980, 58 ff.
[6] Etherington was trying to prove that the si- element could have been the prefix meaning separation, in the lines of Latin se-. Miss Eleanor Dilley aimed her criticism at two main directions: first, it is unlikely that the initial s- had been preserved without changing into h-; second, such a hypothetical prefix is not attested in later Greek.
[7] Etherington tends to read the last word a-we-re-e as some form of a denominal verb related to αἴλινος (= *ἀϜείλινος), according to his supposition), 'lamentation song'. Miss Dilley points out the impossibility of restoring *ἀϜει- from αἰ- in this case, as the element αἰ here, as she points out, is cognate to the interjection αἴ (Lat. vae); so only *Ϝαι- can be restored here, but under no circumstances *ἀϜει-. Compare Frisk, GEW s.v. αἴλινος.
[8] It is also to be mentioned here that, working on Eisenmann's 'funeral' interpretation, Miss Eleanor Dilley was attempting to put the contents of the tablet into the context of the Niobe myth, but this daring idea of hers did not receive any positive response, not even from Eisenmann himself. Paradoxically, this sway of fantasy has put her remarkably close to Hector Etherington, whose rash conclusions she so deservedly criticised in support of Eisenmann. On the other hand, Etherington's 'tentative translation' was not so far from the general lines of Eisenmann's version, except that, from his own phonetic interpretation, he invented a son who was 'left orphan' and 'conducted the lamentation ceremony'. See Etherington H., Funeral Ceremonies in the Greek Bronze Age, Arch. Quart., XLVII, 1966, 45 ff.
[9] PCDTTU, 25, 1970, 13-20.
[10] But compare also Miss Dilley's second article, where she not unmotivatedly defends Eisenmann's view.
[11] Miss Dilley particularly emphasised that other abbreviations of such a kind were not attested in the known corpus of Linear B; the phonetic interpretation of si- should also raise doubts, bearing in mind the attested forms such as i-je-re-ja etc.
[12] Α 37, Α 451, Ε 115, Κ 278, Π 514, Ψ 770, β 262, γ 55, δ 762, ε 445, ζ 324, ι 528. It is also remarkable that the word always occurs at the start of a line.
[13] In our view, this method could be successfully applied to deciphering numerous other unidentified words in Linear B inscriptions. To our knowledge, no attempts have yet been made.
[14] Γ 199, Γ 418, δ 184, δ 219, ζ 229, ψ 218. The word, five times applied to Helen, and once (ζ 229) to Athene, is used to describe their birth from Zeus.
[15] The word ἄνακτος occurs 71 times in Homer; with very few exceptions, it is the last word in the line, cf. A 75, A 529, B 373, B 566 etc.
[16] Cf. B 373, B 566, B 624, B 672, etc.
[17] From this point, it is strange that neither Dilley nor Eisenmann had attempted to find traces of αἶγες here; it is also to be borne in mind that the spelling a-i- for the diphthong αἰ- is anomalous; the character ai-, or just a- would have been expected.
[18] The line α 5, ἀρνύμενος ἣν τε ψυχὴν καὶ νόστον ἑταίρων, is not relevant here, because -ος ἣν is to be reconstructed as *-os swân, so in fact here we have two s's which actually make the syllable long, and not s+digamma.
[19] GEW s.v. λίθος: Scheftelowitz Festgabe H. Jacobi (Bonn 1926) 28: zu lit. slidùs 'glatt'. A different opinion given by Boisacq (Dict. étym., 1916, 581): 'Le rapport avec lit. lẽdas 'glace, pl. grêle' (Fick, Vergl. Wörterbuch der idg. Sprachen, 3. Aufl. Bd. I, 536) est conjectural.' Thus the material of the present tablet confirms Scheftelowitz's hypothesis, given by Frisk.
[20] For the Indo-European etymology of (σ)λίθος, cf. also English slide, etc.
[21] Ciorbea, I. Tablette Ep 704 de Pylos-la genèse de l'épigramme satyrique? EGP, xxvi, 1969, 11 ff.
[22] Although occasionally 'elegiac' distichs composed of a hexameter line and one of iambics are found in the alphabetic Greek inscriptions. Ciorbea cites this Thessalian gravestone inscription in his support:
Μνᾶμ' ἐμὶ Πυρ[ρ]ιάδαο, ὃς οὐκ ἠπίστατο φεύγειν,
ἀλ[λ]' αὖθε πὲρ γᾶς τᾶσδε πολ[λ]ὸν ἀριστεύων ἔθανε.
(as quoted by Ciorbea; = IG IX, 2, 270)
[23] Σ 550 and Σ 560.
[24] Wrede-Grischkat J., Das Stein-Motiv in der Theseus-Sage, AGRm, 817 ff., 1891 (Bd II).
[25] πλησίον δὲ ἱερόν ἐστιν Ἀφροδίτης Οὐρανίας. πρώτοις δὲ ἀνθρώπων Ἀσσυρίοις κατέστη σέβεσθαι τὴν Οὐρανίαν, μετὰ δὲ Ἀσσυρίους Κυπρίων Παφίοις καὶ Φοινίκων τοῖς Ἀσκαλῶνα ἔχουσιν ἐν τῇ Παλαιστίνῃ, παρὰ δὲ Φοινίκων Κυθήριοι μαθόντες σέβουσιν· Ἀθηναίοις δὲ κατεστήσατο Αἰγεύς, αὑτῷ τε οὐκ εἶναι παῖδας νομίζων - οὐ γάρ πω τότε ἦσαν - καὶ ταῖς ὰδελφαῖς γενέσθαι τὴν συμφορὰν ἐκ μηνίματος τῆς Οὐρανίας. τὸ δὲ ἐφ' ἡμῶν ἔτι ἄγαλμα λίθου Παρίου καὶ ἔργον <Φειδίου>... I.14.7.
[26] Ζ 168-9, traditionally thought to be one of the later parts of the epic.