Print

Print


On 20 Nov 2015 19:13, "Logan Kearsley" <[log in to unmask]> wrote:
>
> On 20 November 2015 at 08:44, And Rosta <[log in to unmask]> wrote:
> > By contrast, my method calculates the complexity value for phonological
> > forms, and on the basis of manually-specified basicness values in
lexical
> > entries assigns forms to lexical entries in a way consistent with the
> > tendency for more basic lexis to have more basic forms.
>
> While putting All The Features into one big package does seem a
> popular approach, this seems to me like yet another feature that could
> be easily, and best, implemented as separate tool that operates on
> pre-generated wordlists, and thus could be be combined with a bunch of
> different word generators.

I can see how, given as input a list of all possible forms paired with
their complexity value, a separate lexis-generator module/program could
then match forms to lexical entries. But I don't see how to calculate the
complexity values from the outputs of existing generators.

So an Everyword-like program that also calculates complexity values would
be one desideratum, and the lexis-generator another.

> > (Either approach would
> > fail to produce good lorem ipsum for inflected lgs, tho.)
>
> And that's where we need conlang-morphological-analyzers! Which I will
> get to eventually....

And something that generates inflected forms. I (putting myself in your
shoes) would be completely daunted at the prospect of trying to write a
program that generates inflected forms for any language. Or at the prospect
of a general-purpose conlang-morphological-analyser.

--And.