OT Re: Applying for a Neo Subtag

Mark Davis ☕ (sıʌɐp ʞɹɐɯ) mark at macchiato.com
Mon Mar 7 22:24:49 CET 2011


For computational purposes, grammatical markers expressed as suffixes (resp.
prefixes or infixes) are only completely regular when (a) they always have
the same form, no matter the word they attach to, and (b) any word with that
suffix is an instance of that grammatical marker. The former is important
for production, the latter for identification (mining). For example, even if
we had no strong verbs in English, and every verb took -ed for the past and
past participle, (b) wouldn't be true; there are many words ending with -ed
that aren't verbs, or that are verbs that aren't in the past tense. And
(say) Esperanto fails here also: -i is the infinitive verb ending, but not
every word ending with -i is an infinitive verb.

And even a (a/b) level of regularity doesn't do much for for parameter
substitution. Take a sentence in Esperanto, for example. The fact that verbs
don't change with their subjects is great. But changing a numeric
placeholder value in a sentence from 1 to 2 can still require changes in
nouns and adjectives in the rest of the sentence.

Mark

*— Il meglio è l’inimico del bene —*


On Mon, Mar 7, 2011 at 11:56, David Starner <prosfilaes at gmail.com> wrote:

> On Mon, Mar 7, 2011 at 11:49 AM, Mark Davis ☕ (sıʌɐp ʞɹɐɯ)
> <mark at macchiato.com> wrote:
> > No, you can't ditch them all; not saying you can.
> > Particles or word order are so very much simpler than inflections for one
> of
> > the key problems we are faced with: parameter substitution. That is,
> > substituting variables in sentences templates like "{person} viewed
> {number}
> > {things} on {date}". Inflexions would be ok if they were perfectly
> regular
> > and never overlapped with other words—but that never happens, so they are
> > extremely painful.
>
> In most artificial languages, they are perfectly regular. In most
> languages, inflexions would be easy if you were dealing with just one
> language. If you want to be correct on "a/an", you simply store the
> correct form beside the word. It's only when you're dealing with
> English and German and Finnish and Turkish and Chinese that it's a
> problem, in which case Esperanto and most other artificial languages
> are pretty easy.
>
> --
> Kie ekzistas vivo, ekzistas espero.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.alvestrand.no/pipermail/ietf-languages/attachments/20110307/b0e088fd/attachment.html>


More information about the Ietf-languages mailing list