[YG Conlang Archives] > [romconlang group] > messages [Date Index] [Thread Index] >
Henrik Theiling skrev:
Hi! Melroch 'Aestan writes:Henrik Theiling skrev: ...Ah, ok. So awefully many books I don't have. :-/Yes, same here. Are books awfully expensive in Germany too? When I was there around 18 years ago I thought they were at least significantly cheaper than in Sweden, but that's long ago and everything is relative.Well, they are expensive enough to prevent me from buying everything I want immediately. :-/Anyway "Romanisches etymologisches W�rterbuch" is high on my list of books I'd want an electronic version of, ...Yes, any lexicon searchable with a computer significant improves.
Yes, and I've been known to convert wordlists from various other file formats intto PDF just to be able to use Adobe Readers grep search feature on them! :-)
... BTW I find this <http://tinyurl.com/z8yau> online Latin wordlist handy: it is less buggy than the Perseus database, doessupport substring searches(*) and its output is relatively easy to save as plaintext for further processing.Ah, I will bookmark it. If I haven't missed a switch, it has the unfortunate drawback of many, many online resources of Latin of not marking vowel length, however.
Yes, but that is a drawback it shares with many, many Latin word lists, and texts. One will have to look up the words in a dead-tree lexicon which marks length. For the unlikely possibility that you have none I can recommend Langenscheidt's Latin "Taschenw�rterbuch". Its main drawback is that the German words are printed in Fraktur -- not that Fraktur isn't cool, but it is a pain to read in small sizes and smeary or badly photo-reproduced print. Even my Swedish-Latin dictionary, which is a photo-reprint of a 19th century work, uses Fraktur -- and old-style orthography -- for the Swedish, and it's oh so smeary in the reproduction.
But apart from that, it seems to be *much* faster and offers more ways of searching than Perseus, yes.So thanks for sharing the link! :-)... BTW I searched your example file to see how you determine Latin stress. Is this: # property latin_stress # first last # first, [ last ] # long, [ last ] # any, [ short ], [ last ] all there's to it?Currently, yes.Does this mean you don't apply the _muta cum liquida_ rule in �rj�trunn in spite of its 'classicism' (no blame, just wondering)? How would you preserve stress information after it is lexicalized? I guess one *would* have to introduce some sort of stressmark into the string, e.g.rule "stress mark"
>> syllable latin_stress >> (V:,W) > _'
where ' after a vowel works as stress mark.If I were to really use the stress but for very simple things at the beginning of the Latin import, then yes, definitely. However, �rj�trunn shifts the stress to the first syllable almost immediately.
Of course, I'm well aware of this.
Only very few things are done using Latin stress, and for that, the current rules are probably enough. Wheneven I shift Latin words withstress that does not obey the above rules, I currently have to do some things manually.
When I tried to write Old English sound changes I made a rule that caused initial syllables prefixed with a % xcharacter not to be stressed. One could probably use something similar to prevent whole words from being stressed. In my old Slvanjek sound change script I had a mechanism for excluding words that were marked for stress in the input from automatic stress assignment. I guess something like .* ' .* > _ might do the trick in SCH syntax. BTW it occurred to me that I need to mark secondary stress to get syncope right in R3. I guess something like rule "6" syllable any, [short], [ .* " .* ] V > V' where " is primary and ' is secondary stress would do the trick, if applied after the primary stress mark is introduced.
**Henrik
--
/BP 8^)>
--
Benct Philip Jonsson -- melroch at melroch dot se
"Maybe" is a strange word. When mum or dad says it
it means "yes", but when my big brothers say it it
means "no"!
(Philip Jonsson jr, age 7)