Novgorod Codex - Preservation and Reading Method

Preservation and Reading Method

Preservation of the tablets presented unique challenges, as the usual preservation method for wood would have destroyed the wax layer, and vice versa. The method eventually decided on called for careful separation of the wax layer, and preserving each material separately. The newly exposed wood under the removed wax was found to have been extensively scratched by the stylus cutting through the thin wax. It took the research team several weeks to realize that some symbols could be discerned in the scratches.

Famed Russian linguist Andrey Zaliznyak, one of the foremost experts on the early medieval Novgorod dialect, has taken tremendous effort to reconstruct so far only a small portion of the texts preceding the basic text. The main difficulty with this task is the fact that the feeble traces of dozens of thousands of letters left by the stylus, often hardly discernible from the natural shading of the soft lime wood, have been superimposed on each other, producing an impenetrable labyrinth of lines (Zaliznyak speaks of a “hyper-palimpsest”). Consequently, ‘reading’ a single concealed text of one page can take weeks.

According to Zaliznyak, reading the concealed texts in the scratches is a unique challenge unlike anything attempted by any research team previously. The very compact surface of the four writing surfaces contains traces of thousands of texts, estimated to have been written over several decades. As such, the stylus traces form a constant mesh of lines across the entire surface. To complicate the process, they are also all written by a single hand, making handwriting analysis impossible. As such, Zaliznyak does not call the process 'reading'; instead, he calls it 'reconstruction'. Instead of asking himself the question, 'what's written on this line', Zaliznyak approaches the problem as 'is a phrase A or a word B possible among everything written in this sector'.

The reconstruction is therefore done letter by letter, starting from an arbitrary position, usually somewhere at the top of a 'page'. After analyzing a meshwork of scratches and identifying some of the letters in a given spot (which can realistically number hundreds or even thousands), Zaliznyak then moves some distance to the side and begins identifying symbols at the next position. After several positions are discerned in that way, most letter combinations are discarded as senseless jumble, and possibly meaningful words are identified. Zaliznyak then moves to the next position and attempts to locate subsequent symbols that would complete the word or a sentence. As the text, in a typical fashion for the writing of the time, is written with no spaces between words, identifying these chains becomes somewhat easier compared to if it had been written with spaces.

After careful examination of each position, Zaliznyak creates symbol chains that continue to grow in size. The search often branches off into false leads, where at a certain symbol the chains switch off to a different text fragment. Sometimes these false branches are identified after only a few symbols, but sometimes the false branches can take several words, sentences, and even longer to be discounted. Such false leads can take several days or even weeks to identify.

Another interesting specific of the texts is that many of them have been written multiple times, for reasons that can only be guessed at. Due to the previous copy being erased before a new copy is made, each repetition is written somewhat shifted compared to the previous copy. It is unknown whether the copies were made right after each other, or weeks, months or even years apart. Multiple copies of the same text make identifying false chains easier.

The process remains exceedingly hard to peer-review. Only small portions of Zaliznyak's texts have been peer-reviewed to this time, as no research team came forward that was willing to learn and repeat the process over the length of a large text. Linguist Izabel Vallotton of Geneva cooperated with Zaliznyak on some portions of the reconstruction, where Zaliznyak identified a portion of the chain and passed it on to Vallotton, with both of them then continuing to independently reconstruct the text. In the experiment, both Vallotton and Zaliznyak ended up with completely identical chains, matching to the letter, but the chains were admittedly short, only 20-30 symbols long.

Finally, a problem Zaliznyak considers unsolvable is identifying spelling errors or Russisms in the Church Slavonic. Often, the position where an error or deviation from Church Slavonic is possible, a correct symbol or symbols will also be present alongside an erroneous one, in which case Zaliznyak always assumes the original text was correctly written. In some cases, such assumptions will of course be incorrect. These multiple possibilities may also be the original author's correcting himself by erasing a mistake with his stylus and writing in a correct symbol.