I did, and as linguistics is my hobby I probably should have said indo-european. Never mind though... Finnish is about as agglutinating as Swahili, I know.
Not necessarily... to take the example nakisoma "I read it", if the dictionary listed only the verb "soma" and the affixes "na" and "ki" separately, and didn't contain any information on how they were combined, then it would be fairly useless, since it either wouldn't recognise "nakisoma" as correct (since it doesn't know they can be glued together) or it would also recognise "kinasoma" which is meaningless (since it wouldn't know the rules about the order things can be fixed together). Any spell checker must also know derivational rules (and thus some grammar), such as in english any regular verb takes -s in the 3rd person singular, unless you list every possible derivation in its word list (difficult in swahili, less so in english).
I'm surprised that the spell checking code in Open Office is flexible enough to cope with Swahili spell checking, given how different the european languages it was designed to handle are from the bantu languages (eg swahili) in structure. European languages generally have rather poor derivational morphology, and nouns and verbs inflect by taking a single suffix only. Compare this to Swahili, where verbal derivational morphology is quite rich, and the verb takes a tense prefix along with subject and object concord prefixes, and a mood suffix... I would have expected the spell checking code to choke when they tried to specify which prefixes a verb can take in which order, etc. An example to prove my point, with the english and swahili sentences broken down into their constituent parts:
English:
I read it (the book)
Cannot be broken down
Swahili:
nakisoma
ni-a-ki-som-a
I:subj-simple:present-it:ki:class-read-indicative
English:
He who leaves
He who leave-s
Swahili:
anayetoka
a-na-ye-tok-a
3p:subj-present-3p:relative-leave-indicative
And so on. There are very few english, french or spanish (languages I speak a little of as well) words which approach this level of agglutination, the obvious example being antidisestablishmentarianism.
Slashdot Mirror
I did, and as linguistics is my hobby I probably should have said indo-european. Never mind though... Finnish is about as agglutinating as Swahili, I know.
Not necessarily... to take the example nakisoma "I read it", if the dictionary listed only the verb "soma" and the affixes "na" and "ki" separately, and didn't contain any information on how they were combined, then it would be fairly useless, since it either wouldn't recognise "nakisoma" as correct (since it doesn't know they can be glued together) or it would also recognise "kinasoma" which is meaningless (since it wouldn't know the rules about the order things can be fixed together). Any spell checker must also know derivational rules (and thus some grammar), such as in english any regular verb takes -s in the 3rd person singular, unless you list every possible derivation in its word list (difficult in swahili, less so in english).
I'm surprised that the spell checking code in Open Office is flexible enough to cope with Swahili spell checking, given how different the european languages it was designed to handle are from the bantu languages (eg swahili) in structure. European languages generally have rather poor derivational morphology, and nouns and verbs inflect by taking a single suffix only. Compare this to Swahili, where verbal derivational morphology is quite rich, and the verb takes a tense prefix along with subject and object concord prefixes, and a mood suffix... I would have expected the spell checking code to choke when they tried to specify which prefixes a verb can take in which order, etc. An example to prove my point, with the english and swahili sentences broken down into their constituent parts: English: I read it (the book) Cannot be broken down Swahili: nakisoma ni-a-ki-som-a I:subj-simple:present-it:ki:class-read-indicative English: He who leaves He who leave-s Swahili: anayetoka a-na-ye-tok-a 3p:subj-present-3p:relative-leave-indicative And so on. There are very few english, french or spanish (languages I speak a little of as well) words which approach this level of agglutination, the obvious example being antidisestablishmentarianism.