Norwegian
Norwegian read in both its written forms
Norway has written itself in two official standards since 1885, so an archive box moves between Bokmål and Nynorsk from one document to the next, and a page from before 1938 may still spell å as aa. Evidano keeps each page in the spelling it was written in and tells the two norms apart rather than levelling them into one.
Sample pages
Real pages from public collections, shown beside their transcriptions. Pages whose transcription is still being checked are marked.
National Library of Norway (Nasjonalbiblioteket), 22 September 1833 · Source · Public domain
Transcription
The dateline, the salutation and the opening of the letter, from the transcription of the recto. The crop is tighter than the text: it clips the right-hand end of almost every line, and it runs out partway through the second paragraph, so the transcription completes lines the image cuts off. The sheet’s address panel, off to the left, and the closing lines on the verso are not included. A diplomatic transcription proofread on Norwegian Wikisource from the National Library’s digitisation, keeping Munch’s Danish-Norwegian spelling; a word he struck through is given here as written before deletion.
Christiania d. 22 Sept. 1833. L. S. Hr. Pastor Faye! Ved min Ankomst hertil efter Sommerferierne forefandt jeg Deres Folkesagn og venskabelige Brev, hvorfor jeg her hjerteligen takker Dem. Deres Ønske om at levere en Recention af Skriftet har jeg, som De af Vidar vil see, efterkommet. Om vi i flere Henseender ere uenige, om jeg endog maaskee skulde synes for haard i Bedømmelsen af Et og Andet, haaber jeg dog, at De derfor ei vil have noget mod min Person; det er jo muligt, at jeg kan have feilet, og modtager gjerne et Gjensvar og Berigtigelse af, hvad jeg har fremsat: men jeg smigrer mig dog med, at vi begge kunne lade vor videnskabelige Uenighed være fjærnet fra vort personlige Forhold. Har jeg af og til taget Munden fuld, da er det blot og lutter Oprigtighed; saalænge man gjør Complimenter, er der ingen fair play, og jeg vilde saa gjerne, at man nu ogsaa her til Lands skulde ophøre med de intelligente Complimenter ved saadanne Anledninger, og efterligne en Wolfgang Menzel i den ægte og redelige Fritalenhed, som nu ogsaa hersker i Maanedsskrift for Literatur og i den svenske Literaturtidende. Det er det eneste, som hos os kan ophjælpe sand Videnskabelighed. Hvis De, som rimeligt er, har Noget at erindre ved min Recension, tør vi vel smigre os med, at De under Vidar det. Hvad de Meddelelser angaaer, som De smigrer mig med at ønske fra mig, saa maa jeg tilstaae, at jeg i den senere Tid har haft og endnu har saa meget at bestille med mit paa mig tyngende Brødstudium, at faa ledige Stunde have tilbudt sig for mig til fortsat Studium. Jeg tvivler derfor paa, at jeg nu kan være til nogen Nytte. Hvad de gamle Egennavne angaaer, da troer jeg, at man jo før jo heller bør begynde med at indføre en rigtigere
Transcription from Brev fra Peter Andreas Munch til Andreas Faye av 1833-09-22, proofread page via Norwegian Wikisource (Public domain).
Stortinget (Norwegian Parliament), 2008 · Source · Public domain
Transcription
The title block, the opening paragraph of "Bakgrunn" and the "Forslag" section with its quoted 1814 clause, transcribed on Norwegian Wikisource from the Storting’s own published document and validated to its highest proofreading status; the intervening paragraphs of argument, also on this page, are omitted here.
Grunnlovsforslag nr. 3 (2007–2008) Grunnlovsforslag fra Øystein Djupedal, Ågot Valle og Rolf Reikvam Dokument nr. 12:3 (2007–2008) Til Stortinget BAKGRUNN Undertegnede fremmer dette forslag til endringer i Grunnloven slik at Norge kan endre sin formelle styreform fra monarkisk til republikansk. FORSLAG På denne bakgrunn fremmes følgende forslag: forslag: Overskriften endres fra «Kongeriget Norges Grundlov» til «Grundlov for Norges Rige». § 1 skal lyde: Norge er en fri, selvstændig, udelelig og uavhændelig Republik. Overskriften til kap. B skal lyde: Om den udøvende Magt. § 3 skal lyde: Alternativ 1: Den udøvende Magt er hos Regjeringen.
Transcription from Grunnlovsforslag 3 (2007-2008).pdf, proofread page via Norwegian Wikisource (Public domain).
National Library of Norway (Nasjonalbiblioteket), 21 January 1929 · Source · Public domain
Transcription being verified
No proofread transcription of this issue has been published; the National Library’s newspaper portal (nb.no) carries an uncorrected OCR layer of the same scan that a transcription could be checked against.
What the model is told to watch for: Produce a diplomatic transcription in Norwegian.
Source · CC BY-SA 3.0 · Photograph by Eirik
Transcription being verified
A short public inscription rather than an edited text; no separate published transcription exists, and the wording can be read directly from the plaque.
What the model is told to watch for: Produce a diplomatic transcription in Norwegian.
Bokmål, Nynorsk and the Danish-Norwegian page before both
Until 1814 Norway was under the Danish crown, and its clergy, courts and printers wrote Danish; the union with Sweden that followed did not touch the written language, which is why there is no single date at which "Norwegian" properly begins. Two answers to that inheritance now stand side by side as official norms. Bokmål grew out of the Dano-Norwegian usage of the towns, brought closer to Norwegian speech by the spelling reforms of 1907, 1917 and 1938 and renamed from Riksmål in 1929. Nynorsk was built the other way round: from the 1840s the philologist Ivar Aasen travelled the country recording rural dialects and in 1853 published a grammar, and in 1873 a dictionary, of the composite standard he called Landsmål, renamed Nynorsk at the same 1929 reform. Both are taught in every school, both appear in the Storting’s own papers, and a municipality can choose either as its working language, so a twentieth-century Norwegian archive mixes the two without warning.
The two norms share an alphabet of twenty-nine letters, and the last three, æ, ø and å, exist in no English keyboard layout and are the ones an uncalibrated model is likeliest to drop or garble. Å is comparatively new: before the 1917 reform made it standard, and 1938 made it compulsory in schools, the same sound was written aa, a spelling that survives today only in older personal and place names such as Aasen or Aalesund. Consonant doubling is the main vowel-length signal in both norms, so a single or double letter after a vowel can be the only thing that distinguishes two different words, and long compounds are written solid, with no space or hyphen between their parts. Nynorsk keeps a wider range of accepted forms than Bokmål, so kyrkje, kva and ikkje sit on the same shelf as kirke, hva and ikke, and neither spelling is a mistake.
Everyday Norwegian material spans all of this at once. A nineteenth-century letter is written in the Danish-based hand and spelling that came before either reform; a Storting document from this century can quote the 1814 Constitution’s original wording, selvstændig and uavhændelig, inside an otherwise ordinary modern Bokmål page; and a church or trail sign in a Nynorsk municipality reads frå and kyrkje where the same sign further east would read fra and kirke.
Why ordinary OCR struggles here
Two norms treated as one language
A model calibrated on generic "Norwegian" text has usually seen far more Bokmål than Nynorsk, and quietly corrects kyrkje to kirke or kva to hva, or blends the two into a hybrid that neither norm actually allows. Both spellings need to be recognised as valid outputs in their own right, not as competing guesses at the same word.
Æ, ø, å and the ghost of aa
These three letters sit outside the alphabet an English-trained model expects, so æ is read as a stretched a, ø as an o with a stray mark, and å as a or aa. Material written before the 1938 reform genuinely does spell å as aa, which is correct for its date and needs to be kept rather than silently modernised.
Solid compounds with no natural break
Norwegian compounds are written as one unbroken word, sometimes running to five or six elements, and layout analysis trained on English tends to insert a space wherever the shape of the word suggests one, or to read the whole string as noise. Getting the word right means recognising Norwegian morphology, not just Latin letterforms.
Quoted old spelling inside a modern page
Because Norway’s founding texts are reproduced verbatim in later law, a Storting document from this decade can carry a clause in the 1814 Constitution’s own spelling, such as selvstændig and Regjeringen’s older forms, set beside ordinary contemporary Bokmål on the same page. A transcription that modernises everything indiscriminately erases exactly the distinction the document is drawing.
Who works with this material
Local and family historians
Bygdebok projects and genealogists work through parish registers, newspaper notices and letters that cross the Danish-Norwegian, Riksmål and modern Bokmål or Nynorsk periods within a single family archive. They need names and place names kept in the spelling of their date, since a farm name modernised to today’s form can break a match against the land register.
Municipalities and public offices
Norwegian law lets residents write to their council in either Bokmål or Nynorsk and expects a reply in kind, so offices digitising decades of correspondence need the two norms told apart rather than folded into one. A transcription that keeps kva as kva, not silently turned into hva, keeps that record usable for its own purpose.
Newspaper and archive digitisation
Regional papers such as Vikebladet ran for a century in much the same close-set column layout, moving through several spelling reforms without ever changing their masthead. A project indexing decades of issues needs the reforms themselves to survive the transcription, so a search for aar and for år can be understood as the same word at different dates.
Students of Norwegian language history
Coursework on the Bokmål-Nynorsk split works from primary material: an 1830s Danish-Norwegian letter, Aasen’s own Landsmål writings and a Storting document quoting the 1814 Constitution’s wording beside modern prose. None of that comparison holds if the transcription quietly normalises the older spelling to match the newer.
Getting the best transcription
Say which norm to expect
Tell the prompt whether the page is Bokmål or Nynorsk, or that it may switch between the two, and ask for the spelling to be reproduced as written rather than corrected to the other standard. That one instruction is what keeps kyrkje from becoming kirke.
Set a rule for æ, ø, å and aa
Decide whether å should be kept as å or, for material that predates the 1917–38 reforms, whether the older aa spelling should be preserved as written or given alongside a modernised form. Fix the rule once and it applies across the whole batch.
Flag the document’s date
A page from 1833 needs to be read as Danish-Norwegian, not corrected towards modern Bokmål, while a modern Storting document that quotes the 1814 Constitution needs both spellings kept apart on the same page. Naming the date range in the prompt stops the model from levelling the two together.
Test on a mixed page first
Before running a full archive box, try a page likely to hold more than one register at once, such as a form that quotes older law or a scan that pairs a Nynorsk sign with a Bokmål caption. If the norms come back distinguished there, the more uniform pages need only a spot check.
Frequently asked questions
- Can the transcription tell Bokmål and Nynorsk apart?
- Yes. Each is recognised as its own valid norm rather than a misspelling of the other, so kva, ikkje and kyrkje are kept as written on a Nynorsk page and hva, ikke and kirke on a Bokmål one. If a scan mixes both, each stretch can be labelled by norm.
- What happens to æ, ø and å?
- They are transcribed as themselves. For material written before the å reform, the historical aa spelling is kept as printed; ask for a modernised å alongside it if you want both forms for searching.
- Will an 1830s Danish-Norwegian letter come out as modern Bokmål?
- No, not unless you ask for it. The default keeps the letter’s own spelling, forms such as Grundloven and selvstændig included, since normalising them would erase the evidence of when the document was written.
- Can it separate a quoted historic clause from the surrounding modern text?
- Yes. Where a modern Norwegian document quotes older wording, such as the 1814 Constitution’s original phrasing, that stretch can be marked and kept in its own spelling rather than folded into the rest of the page.
- Does it split up Norwegian’s long compound words?
- No, compounds are transcribed as the single unbroken words they are. If you need them broken into their component parts for indexing, ask for that as a separate, labelled step rather than the main transcription.
Related scripts and pages
Transcribe your Norwegian documents
Upload a letter, a form or a photographed sign, say whether it is Bokmål or Nynorsk, and check the first page before running the rest.