----- Original Message -----

From: llama_nom

To: norse_course@yahoogroups.com

Sent: Tuesday, April 04, 2006 6:59 PM

Subject: [norse_course] Transcription methods for digitized Old Norse texts

What preferrences do people have for different methods of
transcribing Old Norse texts for the Internet?

Those at Netútgáfan [ http://www.snerpa.is/net/index.html ] are a
mixture. (1) Some follow Modern Icelandic spelling, others (2)
approximate more or less towards the standardized medieval
orthography (samræmd stafsetning forn), except that certain features
are modernized: oe > æ; ø > ö, e; sk, zk > st.  (3) Some of the
Netútgáfan texts are basically in modern spelling, except with -r
for -ur (where this was originally -r), but still with modern -r for
older -rr.  The practice of Fornaldarsögur Norðurlanda is like (2)
above, as explained in Guðni Jónsson's preface [
http://server.fhp.uoregon.edu/norse/ ], [
http://www.heimskringla.no/original/fornaldersagaene/index.php ].

Við höfum eigi hirt um að samræma ætíð orðmyndir og færa þær þann
veg til fornara máls. Koma því t. d. oft fyrir nútímamyndir í
sagnbeygingum, eins og er margvíða í handritunum sjálfum. Við ritum
æ og ö alltaf á einn veg og st í miðmynd. Stafsetning vísna og kvæða
er höfð sem líkust því, sem tíðkast í meginmálinu. Yfirleitt höfum
við reynt að láta málið á sögunum bera svip og yfirbragð síns tíma
og forðazt að fyrna það að nauðsynjalausu.

"We haven't taken care to always standardize word forms and thus
bring them into line with those of the earlier language.  So, for
example, modern forms often appear in conjugations, as is common in
the manuscripts themselves.  We always write æ and ö in just one way
[as opposed to æ : oe, and ö : ø] and st in the middle voice. 
Verses and poetry are spelt, as far as possible, in accordance with
the forms most generally found in the body of the text.  We've tried
in general to give the language of the sagas the character and
appearance of of their time [of origin] and avoided unnessecary
archaicisms."

A different approach is taken by Gordon's Introduction to Old Norse
and by Norse Course, both of which use a more archaic standard. 
Faulkes's edition of Snorra Edda follows a similar practice to
Gordon, but with forms like 'gera', 'gersimi' instead
of 'gøra', 'gørsimi'.  Cleasby & Vigfússon's dictionary has single
values for æ and ö.  Zoega compromises, with modern forms for the
middle voice, but etymologically useful distinctions between oe and
æ, and ö (printed for hooked o) and ø.

But which is most useful for online texts?

The more archaic spelling is useful for beginners.  It makes clear
certain etymologies and connections between words that later become
less apparent; it's harder to get the hang of making distinctions
when you're used to not doing, than it is to blur distinctions that
you're used to making.  Searching Google for archaic word forms can
save time because it eliminates more modern texts from the search. 
In some poetry, the rhymes depend on more archaic forms than are
preserved in the maniscripts.  More specific searches can be made. 
An archaicly spelt text can be converted automatically to modern
spelling more easily than a text in modern spelling can be converted
into older spelling which does have oe, etc.

But, in favour of a more modern orthography, searching may be easier
when all letters can be typed simply by setting your keyboard to
Modern Icelandic.  Many of the texts, though medieval, were written
after the distinctions between oe and æ, and ö and ø had
disappeared.  In some later poetry, rhymes depend these changes. 
It's easier for people more familiar with modern spelling to
transcribe without introducing accidental inconsistencies.

Either way, it helps if a consistent standard is adopted for
searching.  I have an preferrence for making things look old-
fashioned (maybe because it helps to learn about the early history
of the langauge), but for search purposes, there's a lot to be said
for adopting modern spelling or a compromise old spelling with
certain distinctions blurred.  If a lemmatized text is desired, it
might be best to adopt the spelling system of a particular
dictionary, but probably not a high priority, as differences could
be programmed around.  Does anyone else have any thoughts on this?

Llama Nom