Lest we forget

While working on a short text on a very specific little detail in Swedish phonetics with colleagues David House and Christina T?nnander , we found a pointer to a Swedish dissertation from 1693, containing some notes that would likely be the earliest written attestation for this in Swedish. After a little searching, it turns out the book was digitized and available for download. The quality is good, but the book is in Latin, and Latin written by a Swedish 17th century scholar... so that seems like a pretty real challenge for a language model. This is what OCR made with a random page:

é 2 o dubitari pofiir, maximc dc confonantibus il W/ f)/ q/ $/ num in alphabeto Suecano in numcrum & cenfum literarum diftin- ?arum fint referenda:?hic dubiratioiftacc expendenda de tollenda cft. §. Primum circa je/ de (t>e/ dubitari potcft, an omnino pro diftindhs literis a vocali i/ dc u ( vel b J fint habend?, ac proprio chara?erc(cribend?,cum potius, ut apud Latinos, eadem litcra modo vo? calis,-modo confonans elfe queat, juxta vulgarem Latinorum Grammaticocum regulam; J de V.cum lyllabamincoant, fcqnente vocali, fiunt confonantcs. Ad hoc refpondemus : primum quidem.quod ad literam j regulam iftam in lin- gua Latina quodammodo procedere pofle; fed cum diftinftionem inter voca? les, dc confonantcs, qvac in qualibet lin* gua, quo certior de evidentior fit, co fa* ciliorem reddat linguac pronunciaiioncm, confandcrc videatur; dfhnino re&ius con- ionans )/ prodiftin&a litera habetur, de pecuh'ari charadtere a vocali ( diftingui* tur. Neqve enim apparet ccrta ratio,cur in iingua nota magis conlonans tt) ab t> vocali,?

I do not know latin at all, and am limited to what I can guess from my familiarity with related languages. Which in this case is very little. So I massaged an LLM to accept the oddness of the text and to try to make the best English and Swedish translations it could muster. Here they are. In English:

It may be doubted, especially regarding the consonants w, f, q, s, whether they should be considered as distinct letters from the vowel i and u (or v) and given their own character, or rather, following the common Latin grammar rule, whether the same letter can function as both a vowel and a consonant. Regarding the letter y, it is considered a consonant when it follows a vowel to form a syllable. To this we respond: firstly, that this Latin grammar rule can to some extent be applied in the Swedish language, but since the distinction between vowels and consonants, which makes pronunciation easier and clearer in any language, it may be advisable to consider them as distinct letters. Secondly, it is more correct to consider the consonant y as a distinct letter, with its own character different from that of the vowel i.

And in Swedish:

Det kan ifr?gas?ttas, s?rskilt g?llande konsonanterna w, f, q och s, om de ska betraktas som s?rskilda bokst?ver fr?n vokalen i och u (eller v) och ges en egen karakt?r, eller om samma bokstav kan fungera som b?de vokal och konsonant, i enlighet med den vanliga latinska grammatikregeln. Vad g?ller bokstaven y betraktas den som en konsonant n?r den f?ljer en vokal f?r att bilda en stavelse. Till detta svarar vi: F?r det f?rsta kan denna latinska grammatikregel till viss del till?mpas i det svenska spr?ket, men eftersom skiljelinjen mellan vokaler och konsonanter, vilket g?r uttalet l?ttare och tydligare i alla spr?k, kan det vara l?mpligt att betrakta dem som separata bokst?ver. F?r det andra ?r det mer korrekt att betrakta konsonanten y som en separat bokstav, med sin egen karakt?r som skiljer sig fr?n vokalen i.

I find this pretty impressive. I want to point out that I'm not suggesting that we take LLM translations at face value. These translations look legit, but that's exactly what an LLM will do: make it look legit, whether it is or not. But this dissertation is long, and I definitely do not mind getting an unreliable preview before I expend favours or moneys on a translation by someone who actually knows 17th century Swedish scholarly latin.

And... at least from a layman perspective, I would say that this is a pretty good indication that Ericus Aurivillius was on the right side of history as far as the Swedish alphabet is concerned. And more importantly, if we really want to find out, we now know where to look.