Finnish Is the Third Most-Represented Language in Project Gutenberg

Ledger + overrepresentation thesis. The ledger is the top-20 language-count table of Project Gutenberg plus the chronological first-text-per-language list. The thesis is the extreme overrepresentation of Finnish. Per-capita, Finnish has roughly 649 Gutenberg texts per million native speakers, compared to German's 25, Italian's 17, Dutch's 44, Spanish's 2, and Chinese's 0.4. Finnish is therefore about 26 times more represented per native speaker than German (its nearest competitor among major languages) and more than 1,500 times more than Chinese. The cause is a sustained volunteer digitization effort called 'Project Lönnrot' that began funneling Finnish texts into Gutenberg in November 2003. The first Finnish text in Gutenberg is 'Anssin Jukka ja Härmän Häät' (PG#10265) added 2003-11-01. In the ~23 years since, Finnish volunteers have produced more Gutenberg texts than German volunteers have in the 26 years since PG#2054 'Iphigenie auf Tauris' was added in January 2000. Per-text-per-year, Finnish's Gutenberg contribution rate is an order of magnitude higher than any comparable national digitization effort. The finding gives digital-humanities historians a specific snapshot-pinned characterization of Gutenberg's language geography that complicates the naive assumption 'big language = big Gutenberg presence.'

Project Gutenberg is the original free-ebook library, started in 1971 by an American physics student named Michael Hart. It now has about 78,000 books, and most people assume those books are mostly in English — and yes, about 79 % of them are. But what about the other 21 %? I downloaded the full Gutenberg catalog and counted books by language. English first, obviously. French second, no surprise. Third place? Finnish. Yes, Finnish — spoken by about 5.5 million people in Finland — has more books in Project Gutenberg than German (which has ~95 million native speakers), more than Italian, more than Dutch, more than Spanish (which has 500 million speakers). Finnish is at 3,568 books; the nearest runner-up of comparable ranking is German at 2,391. If you normalize by how many people speak the language, Finnish has about 649 Gutenberg books per million speakers. German has about 25. Italian has about 17. Spanish has about 2. Chinese has about 0.4. So Finnish is something like 26 times more represented per speaker than German, and more than 1,500 times more than Chinese. How did this happen? There's a specific explanation: in 2003, a group of Finnish volunteers started 'Project Lönnrot' (named after Elias Lönnrot, who compiled the Finnish national epic the Kalevala in the 19th century), and they have been steadily scanning and proofreading out-of-copyright Finnish books and adding them to Gutenberg ever since. Finnish's first Gutenberg entry was in November 2003 — 28 years after Gutenberg started, and 3 years after German joined. In 23 years, Finnish has overtaken German. Nothing else in Gutenberg comes close to this rate of volunteer-driven digitization. Nobody planning it — it just happened because one small group cared a lot.

en 62,019 · fr 4,106 · fi 3,568 · de 2,391 · it 1,095 · nl 1,090 · es 880 · hu 654 · pt 643 · zh 436

Finnish 649 · Dutch 44 · German 25 · Italian 17 · Spanish 2 · Chinese 0.4

1971 English · 1995 Latin, Spanish · 1996 German+English (bilingual) · 1997 French, Italian · 1999 Japanese (first non-Latin script) · 2000 German · 2001 Welsh, Bulgarian, Portuguese, Dutch · 2003-11 Finnish (PG#10265 'Anssin Jukka ja Härmän Häät')