On Tue, Feb 4, 2014 at 2:04 AM, Alex Fink <[log in to unmask]> wrote:
> On Mon, 3 Feb 2014 16:53:22 -0800, Gary Shannon <[log in to unmask]> wrote:
>>Generalized some of the rules so that variables and words
>>(non-terminals and terminals) can be separated, with only words being
>>put in the dictionaries.
> That's very... normal, isn't it?  I liked the purity of the "the lexicon _is_ the grammar" note which you started out on, but this is kinda undoing that, to my eye: now you're calling some of the rules the lexicon, and the rest not.
> Alex

They are still rules, they are just in an abbreviated form.

The problem is that there needs to be a way for human users to find a
given word if they don't know the meaning of it. That means arranging
the words alphabetically. And since every adjective has an identical
rule, there's little point in repeating that rule 2,000 times for
2,000 adjectives.

To use an English example, imagine trying to look up a word in a rule like:

ADJ -> abject | ablative | able | abnormal | abominable | aboriginal |
abrasive | abrupt | absent | absolute |  ...

And that doesn't even include the English to Pandári translations of
each word in the list.

So I make each a separate rule:
ADJ -> absent
ADJ -> absolute

arrange them alphabetically and write them:
absent ADJ
absolute ADJ

Yes, it spoils the "purity: of it, superficially. But, it is still
STRICTLY, and only defined by the production rules, so in that sense,
it is still as pure as the driven snow. Only the superficial style of
displaying those rules has changed.