Introduction | The structure of the language | Variation | Vocabulary | A few conclusions
Introduction
We’re dealing with an unknown language. It is actually a corpus, a limited number of sentences and contexts. We may be very well speaking of a natural language, once alive, or of a cult language, a code used only in writing, of an language artificially created with the purpose to demonstrate something, or of a mere sequence of senseless parts, all randomly created. In order to find out which of the possibilities above is true, we need a very serious language analysis. One that starts with the thought that such an experiment is worth its time, even with the sole benefit of theoretical linguistics, and not that starts with the idea that such an endeavor is futile or compromising. From our point of view, a scientist never discredits himself/ herself through his work. He does it only by not doing what he could have done, by not having the courage to walk that line.
Is this a real language or some kind of mumbo-jumbo? Is it a natural language or a formed one? Firs of all, any language is a living system and each of its elements is defined in context with the other elements; so, in order to see if the texts engraved on these plaques are randomly generated or the results of a certain system, we have to prove that such elements do or do not exist. In other words, discover their structure. Second of all, any natural language implies the existence of a certain variation, so we have to prove that such a variation, be it diachronic, diatopic, or even diastratic and diaphasic. These two co-ordinates - structure and variation – represent the key to defining the language on the plaques as a natural one. The enterprise must firstly be synchronic and descriptive. If we are facing a random generation of phonetic sequences, this thing should immediately become clear, as big of an effort its creator may have invested in cheating the apparent, for there is no system. If we’re facing the probability of a code created with a purpose, this should become clear quite soon for there is no natural variation (artificial variation is rather easily identifiable).
If we agreed that it is a natural language, we have to find out what are the most adequate methods to find out the meaning of the texts. Does the present situation resemble the one that the ones that deciphered the hieroglyphs, cuneiforms or Linear B alphabets? Romalo enumerates the most important precedents, trying to find a method that will be suited to the language on the lead plaques. We must first set straight two things. The first one: in this case we are not talking of an unknown writing, the predominant alphabet being the Greek one. There are entire excerpts and texts written in unknown signs, but they do not reach more than 20% from all the texts. So the methods applied in former cases to identify the phonetic value of an unknown sign, although valid in this case, it’s not decisive, because most of the signs are already known. [1].
The second issue: the Geto-Dacian history is an integrated part of the European history, their civilization was contemporaneous with that of the Greek – Roman one, and they were some of the Europeans barbarians that the Greeks and Romans currently admired. The Geto-Dacian people truly made history later, and due to this fact we have plenty of information on them. We cannot say the same about the Egyptians, Hittites or Greeks in the B Linear age. Civilizations flourished long ago, when numerous external sources were rare and distant. In consequence, we have the advantage of an unknown language presently, but also a language whose elements surely perpetuated until nowadays. This does not mean that taking into account previous methods is not useful; it means that we have to fully consider the present situation.
The method used by Dan Romalo is the etymological one, through a comparison between the words in the texts and formally similar words from other languages. In our opinion the results we obtained are relative for a series of reasons. Firstly, Mr. Romalo’s results was influenced by his premises about the genealogical belonging of the language and its reports diachronic with the Romanian language [2], in other words he started a subjective debate, under some prejudice, looking for resemblances of this language with Latin, with other Indo – European languages, but mostly, Romanian. Secondly, he was influenced and forced by his linguistic background: he made comparisons to the old French, English and German, but not the Greek, Slavic, and other Indo –European languages or directly to the common Indo-European language (in whose existence he does not believe; still, the reconstruction is made using the comparative method, an extremely rigorous one). Thirdly, he ignored the importance of correspondences (laws), admitting too much phonetic, morphologic, and synonymic variation, an explicit assumed freedom (romalo 2005: 42), and one that led to the diminishment of the method’s strictness degree.
For example, he has given the meaning “water” to the aeo, aepo, agua, eou, ako, apo sequence, the meaning “now” to the acho, aku, na, nau sequence, the meaning “wife” to the sotia, yxo sequence, the one for “mouth” to the osye, guro sequence etc.; or on the contrary, he ignored this variation even when it was fully justified (for example he recognizes the word sargeto as a hydronym, but to the sargecio variant he gives the meaning “camp”). Then, for the sequences he did not find formal resemblances to, in other languages, gave a meaning that will contribute to the coherence of the text (procedure he named “semantic motivation”), an operation not always objective, because the coherence of the text was already revealed through relative methods, and the sequence remaining was forced to take the meaning that the rest of the text had imposed.
In the end, the consistency that Mr. Romalo tried to keep when applying the etymological method worked sometimes against him because once he has given a meaning to a word, a wrong one, it perpetuated in all the contexts, altering and constraining the translation. The etymologic method implies the knowing of a very large number of languages with which the lexeme from the corpus be compared to, so that there are the least doubt regarding the veracity of the comparison, but even so, the results are unsure. Comparing to only a few contemporary close-by languages is not enough, because each language has its own millennia-long history, and, most often, it is brought by its speakers from great distances.
In Romalo’s case the number of languages used in the translation was small, fact that often forced him to arbitrary segment the texts, written the most often in scriptio continua. For example: font io diagr... "make that priestess" (in comparison to fr. diacre), instead of Fontio de Agripa (reference to Fonteius Agrippa), in plaque 027, grama tyio “arbor vitae (for) bark” (in comparison to the Latin gruma “bark”), instead of gramatyio < gr. grammateus, in plaque 013, “do tana thyo on bono nio" “from a pond to tan in a new house” (in comparison to the French tenner “to tan”, to the German wohnen “to reside”), instead of do Tanathyo on Bononio (in reference to the Tanatis and Bononia toponyms) in plaque 123 etc. To this we may add the low quality of the photos that led very often to misguided readings, as anio neu, tranplaqued as “new year”, instead of Antoneu “Antonius (Hybrida)”, in plaque 118, xtaretos instead of tetaretos in plaque 107, yrachil instead of trako in plaque 091, on songeo tranplaqued as “in blood”, instead of onsonteo in plaque 069, zoa tranplaqued as “the zone”, instead of noa, in plaque 039, temnisoe “dungeon” instead of tenisoe in plaque 025, pud toreo “bridge with towers”, instead of pudyo rep…, in plaque 015 etc. An instructive example is the case of the Etruscan language, which was used in comparison with a huge amount of languages (from Hittite and Armenian to the Aztec and Hungarian language), but the etymological method had feeble results (bonfante 1995: 96).
From our point of view, without any bilingual texts that may offer the decoding key, there is only one method that can lead to sure results, to which we may add two auxiliary ones, already used by Romalo. We are talking about the intern method (combinatory or distributive analysis), that was applied with sufficient success in the case of the Etruscan language, and with some results in Linear A. Given the fact that we do not have an entire language, but only a corpus, and that this language is not known a priori (only to some extent), a situation like this can be an exceptional research field for structuralism methods, but firstly for the distributive analysis method, that cannot be fully applied on known languages, without entirely eliminating from analysis sense and description. In consequence, all coocurrent should be inventoried, distributive relations analyzed and all types and distribution classes be established on all levels (phonetic, morphologic, syntactic, lexical). Romalo empirically applied this method in some cases and obtained reliable results. For example when he noticed that the cotopolo sequence is frequently followed by Deceneu’s name, that it’s never replaceable (Deceneu’s name is followed only by this word), that it can be also followed by Vezina’s name or other first names, and that sometimes is followed by the megaso sequence – a possible Greek borrowed word, and, knowing that Vezina and Deceneu had the great priest function, and that megas in Greek means “great”, he concluded that the cotopolo sequence must mean “priest”, so he defined this element by relating it to connected elements. When one of the elements to which the comparison is being made is already defined through external reports, as described above, the task is much simpler. In the case of the texts found on the plaques, identifying toponym and anthroponyms was the first step in Romalo’s work, being the first secure sequences from a semantic point of view. Defining the other elements is firstly done by reporting to them. The biggest drawback this method presents is that it gives pretty vague values. Happily, in our case we are not dealing with a totally unknown, without history and culture, language; we know names, institutions, relations, events, even words and structures that perpetuated in time and that can be identified in the Romanian language. Moreover, there are a large number of illustrations that form a second code, which can help decoding the first one. As a result we can use adjacent methods that bring new views and complete the results of the combinatory method.
The first of the auxiliary methods is based on the presumption (as logical as it can be) that the numerous plaque representations are described, “told” in texts, or, in reverse order, that the texts are explanations, glosses of the respective images. In reality we have two codes, kind of a ‘bilingual” situation, but only one of them is linguistic. The representation of a scene – often with many details – limits, to our help, the possibilities from which we could chose the meaning of a word, collocation or enunciation.
Not for a few times, the texts are distributed in such a way on the plaques, so that they enter under or besides the figures they mention. Romalo used this method in some cases, with success, but in some he did not pay attention to this precious clue, which resulted in a discrepancy between some decoding and images in that plaque. For example, starting with the interpretation of the word rumuno as “pro-Romans”, and not “Romans”, as it should have been normal, in item 065 is understood that Decebal is surprised by the pro-Romans, who betrayed him (firstly mentioned Vezina and Diegis) and he is beheaded. But we can clearly see from the image that Decebal, represented on his horse, is returning victorious from a battle with the Romans, carrying a signum captured from the enemies. He is received by priests when entering Sarmizegetusa. Those considered pro-Romans are actually Roman prisoners, guarded from behind by Dacian soldiers. Decebal’s killing does not harmonize with the other sources. Also, in plaque 094, Diegis and Vezina are considered traitors in Romalo’s interpretation and called to be judged by Decebal, who was the great priest in Diurpaneus’ times. But from the illustration we see another situation: the two leaders seem to return victorious from a battle against the Romans, from whom they captured two manipulus signs. Decebal with his priests greats the victorious with wreaths. The character depicted on the throne is certainly Decebal, recognizable from the initial D. Attributing this initial to Diurpaneu by Romalo is not sustained by the text, where the name of this king does not appear. Even more, on the sides of the character sitting on the throne is written Mato Dacibalo.
In the end, as we have seen, the most famous is the “quasi-bilingual” method also applied on the Etruscan language (bonfante 1995: 97). It implies situational external references and resembling models – patterns, formulas, foreseeable contexts or realities known from other sources. For example, we can presume that at the end of the inscriptions is a signature or an ending formula, that around numerals are signs for temporal or monetary units etc. In the special case of the lead plaques, the existence of a large enough number of information regarding the Getae and Dacians is of great importance, even if they are indirect. This method was successfully applied by Mr. Romalo, who made comparisons with information from other sources (historiography, archeological, etc.) especially for first names, positions and institution names, certain links between the characters identified etc. Actually, this was the first step he took when starting, and the most important, which allowed him to start the initial text segmentation. Still, we consider that, in their vast majority, this external information is exact and we have to respect them (we don’t find plausible that, for example, Decebal was killed by Vezina and by his brother Diegis).
In conclusion, Romalo tried to decipher the texts without trying to understand the structure of the language in which they were written. The fragility of his etymologic method is due firstly to the fact that this language does not look Indo – European and (at least in our opinion) its structural resemblance to Latin is not sustainable (at least they resemble lexically, but this thing does not ensure a connection between Latin and the Dacian language), and he started from the Indo – European premise and of a great resemblance between the two languages [3].
Secondly, the translation was made somehow “in the dark”, with a lot of approximations and without visibility to the system. Still, its results cannot be ignored, but must be validated by comparing them to results obtained by different methods. Knowing the limits of his method, the author declares in his book that the action should be firstly regarded as a challenge for breaking the indifference that lasted almost a century, and secondly, as a challenge to contest his work hypothesis and translations (romalo 2005: 288). Assuming the risk of a work that may be seen as a felix culpa, Dan Romalo has done almost all he could, in a sad comparison to those, that with their training, could have done much more, but did nothing.
From our point of view, the inventory of the elements based on distributive analysis must foresee translation, this being necessary in revealing the system. Of course, we do not understand this as a strict chronological sequence of the two steps, but as a permanent priority given to objective analysis and understanding of this language’s structure, that has as a result the association of each element’s value or identified relations. Auxiliary methods (external references and text-illustration correlation) will come in handy when association these values. After the identified element has a value, only then the etymologic method can test and validate the choice we made by appealing to the external comparison.
[1] Still, Romalo did some mistakes, for example giving the OY sequence the oiu value, which appears mostly in texts that do not include the V sign (that had the u value), and that in our opinion, as it does in Greek, notes the u vowel. Starting from this faulty appropriation, Romalo considers this sequence as an augmentative suffix, as in Romanian for căsoi, pisoi. We must relate to the fact that the plaques that do not have the V sign, noted u as OY, for example plaque 021, where the name of the capital is noted as SARMIGETOYZO, or plaque 120, where toponyms are mentioned as GENOYCLO, CARSEOY or NOBYDOYN, from which we may clearly conclude OY’s phonetic value.
[2] “From both linguistic and archeologic reasons the presumption that the Geto–Dacian people language was of Indo – European origin is almost certainly sure. At the same time we addopted the work hypothesis – unverified directly and unaccepted by most Roman fans – that the primitive Romanian language evolved directly from the Geto–Dacian, under the powerfull presion of popular Latin. Given these conditions, the etymologic criteria remembered above will be applied only to segment texts on the plaques in words that present a maximum resemblance, as much in the form as in the meaning it takes, firstly with those from the Romanian language (...)”(romalo 2005: 41).
[3] The arguments that were brought in favor of the resemblance of the two languages (and implicitly for the cancellation of the linguistic Romanization idea as it is perceived by most philologists) are borrowed from N. Densusianu and, in our opinion, none of them can be proven. For example, the fact that Ovid complains at a certain time that he forget to speak and write Latin and that he is afraid that some barbaric words have found their way in his poems, does not prove under any circumstances a resemblance between the two languages. After six years of exile during which he could not speak his native tongue, it is normal that the poet became uncertain about his language knowledge. The phenomenon happens to each man that leaves his country for many years, during which he does not use his native tongue at all. More over, the ones appealing to this verse, seem to forget the verses in which Ovid is lamenting that no one speaks Latin in Tomis and that he is forced to understand the locals by using signs, or those verses that mention that, after many years after staying with the barbarians he managed to learn their language. We can mention the Latin inscription in which INTERPREX DACORUM appears, probing the existence of some translators used by the Romans to communicate with the Dacians. Examples could continue.