For the last few weeks I have been trying to follow a lead on the origin of the idea that the Greeks made armour by gluing layers of linen together. Everyone who believes this theory today seems to have got it from the late Peter Connolly, but some of my American friends have found versions as early as 1869 (if you know of an earlier text linking glue and armour, please say so in the comments!) I think I can link it to Isaac Casaubon and another famous 16th century scholar, and show how between 1868 and 1875 their theory of linen soaked in vinegar until it became like felt turned into Connolly’s theory of linen soaked in glue until it became like a mask of bandages soaked in plaster. But my case for that will appear in a footnoted article not a blog post, and today I want to make a larger point which is useful even if you have never spent 10 minutes ranting about silly theories of armour construction.
Everyone with a browser and an uncensored Internet connection is two clicks away from every book in a great library. And if you chose to learn to use it, you can discover wonderful things known to very few people in this world. There are rooms full of books which which are interesting to some community today which have either been forgotten, or were never brought to the attention of that community because it did not exist in 1881. Armour in Texts might seem impressive, but most of the works there were quoted or summarized in about three books published before I was born. I did not find most of them by reading sources, I found most of them by reading people who had read sources and noted down which were useful for understanding armour. The farther back I dig into scholarly books on armour, the more interesting sources I find which nobody seems to read.
It helps if you can read even a little bit of any major language other than English, and if you know a little bit about 19th and early 20th century culture to spot the Edwardian equivalent of Osprey books and self-published treatises on how mainstream science is totally wrong. But I know plenty of people without a lot of university education or knowledge of other languages who have still found and copied useful things. This work is too big for me: I have a dissertation to finish, and I do not love every kind of learning equally.
Google Books and archive.org are the best known collections of digitized books, but even more useful are French projects like Persee and Gallica and German projects by the Bayerische Staatsbibliothek and the University of Kassel. These projects are run by libraries and universities, and librarians are expert in putting books away in a place where they can be found again, and in warning people about issues like the different forms of letters used before the 20th century. (Google rushed to scan books and refused to listen to librarians, so about a third of their books are mis-catalogued and many have transcriptions which make basic blunders like confusing ʃ and f … and it is much more expensive to correct these mistakes after they have been scanned and processed than it would have been if they had moved more slowly and done it right the first time). But having any of these resources is a treasure, and it gives you powers which were once limited to people living in Vienna or Paris or London.
Further Reading: If you want to read about why linguists, librarians, and philologists are concerned about the Google Books corpus, some good articles are:
- Sarah Zhang, “The Pitfalls of Using Google Ngram To Study Language” (2015) https://www.wired.com/2015/10/pitfalls-of-studying-language-with-google-ngram/
- Geoff Nunberg, “Google Books: A Metadata Train Wreck” (2009) http://languagelog.ldc.upenn.edu/nll/?p=1701
- Jessamyn West, “Google’s Slow Fade with Librarians” (2015) https://medium.com/tilty/googles-slow-fade-with-librarians-fddda838a0b7
It is much harder than it should be to find their scan of the first edition of a book that was printed for a hundred years, or the 8th edition of the Encyclopedia Britannica (not the 9th, the 11th is right out) and their badly labelled books crowd out search results.
A 2012 paper in the Journal of Library Metadata found that 36% of 400 randomly chosen items in Google Books had incorrect metadata http://www.tandfonline.com/doi/abs/10.1080/19386389.2012.652566
The scholarly paper on the ngrams tool is Penchenik et al., “Characterizing the Google Books Corpus,” PLOS One (2015) https://doi.org/10.1371/journal.pone.0137041