Writing system

The inscriptions are in an alphabet similar to that used to write the Etruscan language and the older Phrygian inscriptions, all derived from Euboean scripts (Western Greek alphabet, alphabets of Asia Minor). These scripts are ultimately of West Semitic origin and were adapted by various peoples from the 8th century. Characters similar to those used in Lemnos Stele inscription are also found on some pottery fragments on Lemnos.

Classification

Lemnian and Etruscan

A relationship between Lemnian and Etruscan, sometimes grouped together as Tyrrhenian or Aegean languages is largely accepted because of the strong connections between vocabulary and grammar. For example, both Etruscan and Lemnian share two unique dative cases, masculine *-si and feminine-collective *-ale, shown both on the Lemnos Stele (Hulaie-ši "for Hulaie", Φukiasi-ale "for the Phocaean") and in inscriptions written in Etruscan (aule-si "To Aule" on the Cippus Perusinus as well as the inscription mi mulu Laris-ale Velχaina-si "I was blessed for Laris Velchaina"). They also share the masculine genitive in *-s and a simple past tense in *-a-i (Etruscan <-e> as in ame "was" (< *amai); Lemnian <-ai> as in šivai "lived"). Such strong evidence such as this provides little doubt of their true kinship.

Like Etruscan, the Lemnian language appears to have had a four-vowel system consisting of "i", "u", "a" and "e". Having a contrast between front and back vowels, it would appear to lack a high back vowel (a "u"-like sound) which is curious because this defies the linguistic universal of contrast maximization. Since vowel systems such as these without "u" are rare, it is strongly likely that what we transliterate as "o" from the symbol omikron was in fact meant to record a high, back, rounded vowel instead (written in IPA as /u/). This is not unusual considering that different languages may take the same letter to transcribe different sounds. Note for example how "u" in English is used to write a front vowel in French muet, a sound that does not exist in standard English. Note also that in English, the "o" may also denote /u/, e.g. in the word "to". Also, there is the argument concerning the origin of Tyrrhenians in general. If it is true that they originate from Lydia, it is rather coincidental that the languages neighbouring this region, namely Hittite and Akkadian, also happen to have the same four-vowel systems lacking "o". This suggests early areal influence.

Eteocypriot, Eteocretan

Beyond Etruscan and Rhaetian, further relationships to Lemnian become more tentative and highly debated. There is a possible affiliation of Eteocypriot to the above Tyrrhenian grouping. Texts in Eteocypriot (few are known, making it difficult to determine language affinity) have been found in the vicinity of the Lemnian language sphere, amidst the Aegean islands.

Debate continues on concerning the relationship of Eteocypriot, Eteocretan and Minoan to this family. The Amathus bilingual written in Eteocretan shows important structural similarities bearing what appears to be a genitive in -O-SE (Etruscan <-as> and Lemnian <-š>) as well as a 3ps animate pronoun A-NA (Etruscan <an> 'he, she'). The meager text however makes it difficult to prove a kinship for certain. Eteocretan likewise shows grammatical similarities and vocabulary terms but again the number of texts are meager. Since Minoan texts are also few and far between, any grammatical similarities with Etruscan are always tentative. However it has been noted by some online that the oft-repeated Minoan U-NA-KA-NA-SI and U-NA-RU-KA-NA-SI may bear resemblance to what would be written in Etruscan as *unχva cenase "bearing libations" which is surprisingly reasonable considering that the objects on which this is consistently written are in fact libation tables. (The value of <un> as 'libation' is proven by its repeated usage in the Liber Linteus.) Time will tell whether these connections bear fruit.

Indo-European

Scholars now generally agree that these Tyrrhenian languages are not members of the Indo-European family and any connections with the Anatolian languages in particular are probably due to areal influence at best. At the present time without strong evidence one way or another, most academics remain conservative about external connections and consider Tyrrhenian, however it may be eventually defined, an isolate family.

Some modern scholars have claimed that the Tyrrhenian family as a whole is distantly related to the Indo-European languages (IE), citing similarities in grammatical endings and vocabulary. With the paucity of complete texts, this is merely conjecture at present. For now, many remain conservative and consider Tyrrhenian to be an isolate group. A connection with IE is merely one of the strongest possibilities so far but not proven satisfactorily. Some contend that the <-s> and <-l> genitival endings seen in Tyrrhenian languages are evidence of substrate influence from the Anatolian languages (which are part of the Indo-European family), acquired during a time when Tyrrhenian languages were still centered around Asia Minor.

Classical sources

The ancient Greek historian Herodotus referred to the pre-Greek population of Lemnos as the Πελασγοι 'Pelasgoi' (see Pelasgian). However, Herodotus may well have incorrectly lumped two distinct non-Greek peoples together. According to other authors like Thucydides, the pre-Greek population of Lemnos were called Τυρσενοι 'Tyrsenoi' (alternatively, Tyrrhenoi) (see Tyrrhenian).

In his Natural History (1st century AD), Pliny wrote about Alpine peoples: "The Rhaetians and the Vindelicans border with these [Noricans], all distributed in numerous cities. The Gauls maintain that the Raetians descend from the Etruscans, pushed back under the leadership of Raetus." Based on this and linguistic data it's clear that Etruscan ought to be related to Raetic. However, beyond these known facts, there is ample debate and hearsay that follows.

The Lemnos stele, Kaminia

The Lemnos stela

The stela was found built into a church wall in Kaminia and is now at the National Museum, Athens. The 6th-century date is based on the fact that in 510 BC the Athenian Miltiades invaded Lemnos and Hellenized it. The stele bears a low-relief bust of a helmeted man and is inscribed in an alphabet similar to the western ("Chalcidian") Greek alphabet. The inscription is in Boustrophedon style, and has been transliterated but had not been successfully translated until serious linguistic analysis based on comparisons with Etruscan, combined with breakthroughs in Etruscan's own translation started to yield fruit.

The inscription consists of 198 characters forming 33 to 40 words, word separation sometimes indicated with one to three dots. The text consists of three parts, two written vertically and one horizontally. Comprehensible is the phrase avis sialchvis ("aged sixty", B.3), reminiscent of Etruscan avils maχs śealχisc ("and aged sixty-five").

Transcription:

front:

A.1. hulaieš:naφuθ:šiaši

A.2. maraš:mav

A.3. sialχveiš:aviš

A.4. evisθu:šerunaiθ

A.5. šivai

A.6. aker:tavaršiu

A.7. vanalasial:šerunai:murinaic

side:

B.1. hulaieši:φukiasiale:šerunaiθ:evisθu:tuveruna

B.2. rum:haraliu:šivai:eptešiu:arai:tiš:φuke

B.3. šivai:aviš:sialχviš:marašm:aviš:aumai

Translation of the Lemnos Stele

In order to properly translate the stele, one must sift through a sea of hearsay and speculation that abounds about this cloudy text. Some words attract an especially inordinate amount of controversy, yielding multiple and conflicting translations for the same word. We need to obtain a more accurate picture of what this text is telling us. The only way to do this is through a balanced analysis of the smallest details while keeping sight of the larger context at the same time. Let's undo some of the myths that continue to rear their ugly head.

No, <mav> is not a numeral

One of these overly debated words is <mav>. The word is seen in A.2 and assumed by some to be part of a phrase <mav sialχveiš> identifying the age of a deceased person with further comparison to Etruscan . Thus <mav> is often translated as "five" with slight resemblance to Etruscan <maχ> 'five' and <muvalχ> 'fifty'. However this is certainly false because the age is repeated twice in the text, once on line A.3 and another on B.3 where <mav> is nowhere to be seen. Unable to accept this fact, some further imagine that <marašm> must be a scribal error for *mavašm. Again, however this is unlikely because the so-called error is repeated twice (<maraš> on line A.2). More likely is that a would-be translator has fallen prey to his or her own imagination without paying proper attention to these important text patterns. Whatever the value of <mav> really is, it is certainly not a number.

The value of <šivai>

Most have already seen that <šivai aviš sialχviš> is surely relatable to well attested phrases in Etruscan, most notably <zivas avils LXXVI> 'lived 76 years' (inscription known as TLE 880). Since <ziv-> is without a doubt 'to die' in Etruscan, there is strong likelihood that Lemnian šivai means 'died' and thus the person to whom this stele is dedicated had died at the age of 60. But who then was this person?

The name of the person to whom this was dedicated

As one would expect, the person being celebrated is very likely at the beginning of the text on A.1 (<hulaieš>) with a genitive suffix -š attached, meaning 'of' as it does in Etruscan. The name is repeated again at the very beginning of line B.1 (<hulaieši>) with a dative suffix -ši meaning "to" or "for", which is again comparable to Etruscan <-si>. Immediately after we find <φukiasiale> with another recognizable dative suffix from Etruscan, -ale. Thus the name of the person deceased is most likely Hulaie Phukiasi. Both the first and last names are given the dative suffix on line B.2. This may seem odd to English speakers, but this is the pattern we see also in Etruscan -- <mi laris-al meminiie-s> means 'I am of Laris Meminiie', not 'I am of Laris of Meminiie', written on a cup in Campania (ETP 30). So this last name on the stele is then further compared with <φuke> (line B.2) presumed to mean Phocaea, an important region in Asia Minor in those times. This suggests that he was Phocaean, and thus called literally Hulaie the Phocaean.

Suffixes <-m> & <-c>

The comparison between the instances of both <marašm> and <maraš> helps us to properly identify a suffix -m which when compared to Etruscan is the phrasal conjunctive meaning 'and'. A phrasal conjunctive is a suffix used specifically to link two sentences together. An example of phrasal conjunction is "I went to school and I listened to the teacher". In Etruscan, phrasal conjunction is distinguished from nominal conjunction (eg: <apa-c ati-c> "both the father and the mother", Pillar of the tomb of Claudii in Cerveteri, 4th c.BC) where two nouns, not two phrases, are linked with another suffix, <-c>. Surprisingly, we see that in Lemnian -c may indeed be also used as a second conjunction suffix (note line A.7 <murinai-c>)

Is Hulaie's age "60" or "40"?

Another important controversy involves the value of <sialχveiš> itself. Some say it's 'of 60'; some say it's 'of 40'. This confusion stems from the interpretation, or rather misinterpretation, of the Tartaria dice, an important Etruscan find. On the dice we find the numbers '1' to '6' just as we find on modern dice, except they have been kindly written out in letters for us by the person who made them long ago. Many presume that the pattern of the numbers as they are arranged on the six sides of the dice is such that the value for each side when added to the value on the opposite side equals seven. See Etruscan numerals for further information on this debate.

Assuming this pattern is real for a moment, if we find <ša> on one side and <ci> on the other, and if we are certain from other texts that <ci> is "three", then <ša> must be "four" since 3 + 4 = 7. Unfortunately, these are only assumptions based on even more assumptions. Other Etruscan dice have been found which do not show the same pattern. We can't assume. For all we know the Tartaria dice could in fact show another equally valid pattern: Maybe all the sides when subtracted give three (eg: <ša> "six" minus <ci> "three" perhaps).

Also, based on the first pattern, if <ša> is really "four", <huθ> must be "six" but it is known that the pre-Greek name of Tetrapolis (meaning 'Four-cities' in Greek) was <Υττηνια> "Yttenia", thought to be a Tyrrhenian name containing the numeral <huθ> and dating to a time when Etruscans were still in Asia Minor and the Aegean islands (nb. Herodotus' account in Histories). Thus some insist the opposite, that <huθ> means "four" and <ša> means "six". Which value should we give this number?

Some common sense is in order by directing our attention to the phrase <hulaieš naφuθ>. Since Etruscan <neftš> is known to mean 'grandson', it stands to reason that Hulaie is a grandfather. If he were truly '40', we can't explain the mention of his grandson here who is surely old enough to have participated in the funeral to have special mention! Further while there was much hardship in ancient times, nobility were known to live much longer than common folk. Forty would still be a little young for someone with enough money to erect a stele with his name on it.

So in all, it's likelier that Hulaie was in fact sixty at the time of his death, meaning that <sialχveiš> has the value of 'of 60' and that Etruscan <ša> must mean "six".

References

Dieter H. Steinbauer, Neues Handbuch des Etruskischen 1999, pages 357-366.