Tuesday, December 15, 2009
Data Mining in Snorri's Prose Edda
Written by Snorri Sturluson in the last half of the 13th century, this text, the Poetic Edda, and our collection of Norse poetry comprise the bulk of our knowledge about the 'mythology' of the Old Norse. Snorri wrote his text, not only so that some of the old stories could be preserved, but so that this type of poetry and storytelling would be better understood and also preserved for posterity.
The fact that the bulk of the text is about gods and goddesses is not news, nor does one have to be an expert in Norse literature or the Prose Edda to know that. However, I do think that once a lot of the medieval documents, both manuscripts and the many different textual editions that have appeared over the years, data mining will be a useful tool, not only to those of us who are trying to become experts, but to those who already are. It will be particularly useful when comparing different editions of the same text, as it will be very easy to see how many authors/translators have differently interpreted one section, and what language they have chosen to use in the translation. I did not think to undertake anything so ambitious. I thought I would just see what a newbie could find out about the place of gods and goddesses in the text, and what a very simple kind of analysis might yield.
In this case, I used a text that was available on Project Gutenberg. It was much easier to read the HTML version, though the other versions which show what the actual physical book looked like could be used to look at the placement of words on the pages, and so could be used to take the analysis a step further than anything I had attempted.
I chose to look at one of these texts, not only out of personal interest, but also because taking something that was originally in another language, I felt perhaps the use of certain vocabulary in the text could not only tell you something about the original work, but might also tell you the differences in thought pattern between English translators and translators in other ancient or modern languages. I decided that I would see what I could find out if I looked at the instances of the word 'god' in the Snorri's Prose Edda, as translated in 1897 by Rasmus B. Anderson. In that way I can see what the text is telling me, keeping in mind that there will be an overlay of 19th century scholarship. This can be seen especially in the extracts that have been chosen from the text, they are heavy on the mythology and if things are left out it is more of the instruction about how to write the poetry.
And, despite the rudimentary nature of my data mining I still found out things that were interesting about at least Rasmus's version of this text. I used the Concordance tool on Project TaPor. The first thing I found out is that the there are only 57 instances of the word 'god' in the text, but there are 155 instances of the word 'gods'. This makes sense, as you use the vocabulary to describe them as a whole, and that if they are talked about individually you might use their name. In face most of the instances of just 'god' are in the glossary at the end of the text. Likewise, there are 15 instances of the word 'goddesses,' but there are only 11 instances of the word 'goddess,' and mostly that is in the glossary. It comes as no surprise that there are more instances of the word god than goddess, but it is interesting that goddesses is usually part of the phrase 'the gods and goddesses,' and that they are not always part of the generic 'gods' when Snorri is describing the actions of the whole group. And the fact that they are used mostly at the beginning fits with the nature of the text, as that is more heavy on mythology and less about textual construction.
If I were doing a broad project about mythology this would be incredibly useful, because then I would know where to start reading, and that the first section of the text is the most useful. However, for anyone who knows anything about the text this is not really new, and reading the text straight through would lead you to the same conclusion. What is the best selling point for me is that, once you have read the text and made a conclusion or a thesis, you would not have to re-read the whole thing to pick out all the instances or evidence. Instead you use this tool to lead you to the location of the evidence, and then you mine it out, and use it in your essay.
The picture above is from the Penguin Edition, much newer than the one found on Project Gutenberg.