A conventional database record structure

Computational lexicography is often, without much reflection on the topic, concerned with the major European languages and other languages which promise heavy funding. However, in many ways the need -- and linguistic, as well as applicational interest (for example in prosthetic devices for the handicapped) -- is higher for other languages of the world. The following lexicographic database extract for Anyi, a language of the Kwa/Tano family spoken in the Eastern Ivory Coast, is taken from a comparative lexicographic database of Anyi dialects which the author is preparing in cooperation with the Université de Cocody, Abidjan, Ivory Coast. The following example record is formatted in Shoebox notation: characters are all ASCII, xx tags are attribute, field or column names, followed by values for the field in question (lexeme, canonical form, English gloss, French gloss, German gloss, phonological representation, part of speech, date); the record always begins with the lx attribute.

lx AKO
lc
ge chicken
gf poulet
gd Huhn
ph HH
ps N
dt 09/03/1998

 

Field no. 1 2 3 4 5 6 7 8 Field name

Record 1 AKO chicken poulet Huhn N 09/03/1998

Record 2 ... ... ... ... ... ... ... ...

 

Glose française Forme de base Anyi

aimer k`ul´o

allumer s´

ami mija g´

 

 

Table 5: Extract from full Anyi comparative database. 

No.	[No.]	French	Anyi baseform	Indenié	Bona	Moronou	Tiassalé	Sanwi
9.	[116]	aimer	k` ul´ o	k` ul´ o	k` ur´ o	k` ul´ o	k` ul´ o	k` ul´ o
10.	[109]	allumer	s´	s´	s´	s´	s´	s´
11.	[31]	ami	mij a g´	mij a g´	mij a g´	mij a g´	mij a g´	mij a g´

The record structure, and a specimen record, are shown in Table 2.

The kind of microstructural information required in a comparative dialect lexicon, a variety of multilingual database, is shown in Table 3. In fact, this is a highly generalised database, containing only the base forms which are common to all Anyi dialects.

The database of common base forms is supplemented, as any database can be, by output filters; in the present case, these are dialect realisation rules, of which a selection is given in Table 4 (the names are of Anyi dialects).

Using the base form database, and the dialect rules implemented with UNIX tools (see Section 4.1, the more conventional dialect database excerpted in Table 5 was generated automatically.