Numbers in Faroese

[gdshaw]

I have some questions about how cardinal numbers are written in Faroese, for an Open Source machine translation system that I'm developing (http://www.babel.org.uk/ if anyone wants to know more). They may not have a right answer, but I want to at least be sure that what I'm doing isn't wrong. I'm using the base-20 variant of the number system, and would be most grateful for any help that you can give.

Firstly, for numbers such as 21, I've seen both 'ein og tjúgu' (three words) and 'einogtjúgu' (one word). Is one or other of these formats preferred?

Secondly, for 22, I've seen both 'tvey og tjúgu' and 'tvey og tjúga'. Similarly for 23 to 29, and in the 30s 'tretivu' versus 'tretiva'. Are both of these forms acceptable, and if so, which is preferred? Is there any equivalent process applicable to 40 or higher? I'd also be interested to know whether this is just an ad hoc rule that applies to these particular numbers, or a manifestation of a more generic inflectional process.

Finally, it would appear to be the rule that tens and units are always preceded by 'og' if they are part of a larger number, but I've also seen instances in other contexts. To take a particular example, is it better to write 'eitt túsund eitt hundrað', 'eitt túsund og eitt hundrað' or 'eitt túsund og hundrað', or are all three acceptable? Is there a general rule?

With thanks,
Graham Shaw

[Mulder-21]

Even though you're using base-20, I'm still gonna cover both forms, just in case.

I'm not sure about the first one myself, but I'm going on a limb by saying, that writing the numbers and 'og' separated is preferred: Fimm og tjúgu. Sjey og fjøruti. Níggj(u og) hálvfems. Base-10 and what would be the standard/preferred Faroese position system completely avoids these: tjúgu fimm, fjøruti sjey, níti níggju. But again, I'm unsure how to properly write them with letters. Looking at the forms right now, I guess they should be written as one word, but I honestly don't know. In practice you're better off with just writing the numbers: 25, 47, 99.

The -a forms are completely non-Faroese. I don't know where you got them, but they look like someone transliterated someone's speech. The base-20 numbers are: tíggju, tjúgu, tretivu, fjøruti, hálvtrýss, trýss, hálvfjerðs, fýrs, hálvfems. They are all non-inflectional. Faroese only inflects the numbers 1, 2 and 3. (Note: 1 is also inflected in plural number). For completeness sake here are the base-10 numbers: tíggju, tjúgu, tretivu/tríati, fjøruti, fimmti, seksti, sjeyti, áttati, níti/níggjuti.

When reading large numbers, 'og's are only put between hundreds and tens (and ones if using base-20), so a number like: 485,743,582 is read: fýra hundrað og fimm og fýrs milliónir sjey hundrað og trý og fjøruti túsund fimm hundrað og tvey og fýrs. Base-10: fýra hundra og áttati fimm milliónir sjey hundrað og fjøruti trý túsund fimm hundrað og áttati tvey. (alternative, 'tveir' and 'tvær' if referring to masc. resp. fem. objects)

In your case, you're referring to the number 1,100. We'd say 'túsund eitt hundrað', but a much more common way to say this is 'ellivu hundrað' (eleven hundred). The year 1982 for instance is referred to as 'nítjan hundrað og tvey og fýrs (og áttati tvey)'. Numbers over 2,000 are, while they can be referred to as 1,100 - 1,900, usually read number by number: 2,045 (tvey túsund og fimm og fjøruti (and not tjúgu hundrað og fimm og fjøruti))

Hope this shed some light.

JP

[gdshaw]

Thank you most kindly for your very helpful reply.

Examples of 'tvey og tjúga' that I've encountered include http://mitglied.lycos.de/juronjaure/scandnumb.html and http://www.framtak.com/info/faroese.html. However I've not seen any reasoned explanation for this effect, so I'm fully willing to believe that it is just a mistake (and mistakes on the Internet do tend to propagate).

I take it from your later example 'tvey túsund og fimm og fjøruti' that an 'og' is placed between the thousands and the tens/ones if there are no hundreds.

Writing 1100 as 'eleven hundred' seems to be common practice in quite a few languages, including English, but I'd put off doing anything about it because my translation engine made it very difficult to implement this without also turning 1001100 into 'one million eleven hundred' (which would certainly be non-standard English, and I presume Faroese too). It's about time I fixed that, so I have.

I understand what you mean about writing large values as numbers rather than words, but rather than force any particular policy I'm going to leave that decision to the author of the source text.

Should anyone want to have a play, I've set up a demonstration as a web app. It currently supports 22 languages and can be found at http://www.babel.org.uk/demo.cgi?lang=fo.

[Mulder-21]

Thank you most kindly for your very helpful reply.

Examples of 'tvey og tjúga' that I've encountered include http://mitglied.lycos.de/juronjaure/scandnumb.html and http://www.framtak.com/info/faroese.html. However I've not seen any reasoned explanation for this effect, so I'm fully willing to believe that it is just a mistake (and mistakes on the Internet do tend to propagate).

Yes, those are typos, though the Lycos site seems simply be wrong.

I take it from your later example 'tvey túsund og fimm og fjøruti' that an 'og' is placed between the thousands and the tens/ones if there are no hundreds.

Thinking about it 'og' is always added before the tens/one. 1.000.045 would be read "ein millión og fimm og fjøruti", for instance.

Writing 1100 as 'eleven hundred' seems to be common practice in quite a few languages, including English, but I'd put off doing anything about it because my translation engine made it very difficult to implement this without also turning 1001100 into 'one million eleven hundred' (which would certainly be non-standard English, and I presume Faroese too). It's about time I fixed that, so I have.

I'm unsure. Ein millión og ellivu hundrað doesn't sound wrong. So I guess it's acceptable.

I understand what you mean about writing large values as numbers rather than words, but rather than force any particular policy I'm going to leave that decision to the author of the source text.

Of course.

Should anyone want to have a play, I've set up a demonstration as a web app. It currently supports 22 languages and can be found at http://www.babel.org.uk/demo.cgi?lang=fo.

Interesting site indeed.

Hopefully, I shed some light here.

JP