Wikidata:Lexicographical data/Documentation/Languages/dum
| Subclass of | Low Franconian  |
|---|---|
| Country | Holy Roman Empire |
| Follows | Old Dutch |
| Replaced by | Dutch |
| Has grammatical gender | feminine, masculine, neuter |
| Writing system | Latin script |
Normalised spellings, lemmas and representations
editSpelling in Middle Dutch was highly variable. Middle Dutch dictionaries such as VMNW and MNW use a normalised spelling. There is no "standard" normalisation; the two mentioned dictionaries differ in their normalised spelling. The following normalization is suggested for Wikidata, based mostly on MNW and on what is already established on English Wiktionary (see wikt:WT:ADUM):
- Use g, never gh.
- Use gg rather than cg.
- Use sch rather than sc or sk.
Use normalised spellings in lemmas and in forms. Any spelling that is found in a manuscript can be a form, but the normalised spelling should also be a form even if it's not in a manuscript. The lemma should contain only the normalised spelling, not any other attested spellings, to avoid having a huge number of lemmas. Mark the language of normalised spellings, in both forms and lemmas, with -x-Q56669831 (with normalised spelling (Q56669831)).
Vowel length
editVowel length can be inferred from the normalised spelling, but not necessarily from spellings attested in manuscripts. However, Middle Dutch has two different kinds of long a, e and o (see w:Middle Dutch#Phonology), a distinction which is not obvious even from normalised spellings. These are denoted with a circumflex ^ and macron ¯ respectively:
- â/âe, ê/êe and ô/ôo reflect Old Dutch long vowels.
- ā/āe, ē/ēe and ō/ōo reflect Old Dutch short vowels that were lengthened, or borrowed long vowels.
The normalised spelling does not include these marks, as they were never actually written in manuscripts. They are used only in reference works and grammars. It is suggested to add the normalised spelling including these marks with a second representation alongside the normalised spelling, with the language tagged using -x-Q7249970 (with pronunciation respelling (Q7249970)).
In principle, long i/ij and u/uu are always î/îj and û/ûu (reflecting Old Dutch long vowels), while ue is always ūe (reflecting a lengthened vowel). These could also added to the pronunciation respelling in theory, for completeness if nothing else, but this is somewhat less useful.