These are parsed using the normal grammar, as described in Bender (2002).

What we should do is parse numbers and pass them through using SMAF.

We set ChaSen to join numbers together:

(連結品詞 ((名詞 数))
          ((記号 アルファベット)))
(COMPOSIT_POS ((名詞 数))
          ((記号 アルファベット)))

Until we have a proper preprocessor, we just add all numbers from 0--999, using the script utils/make-num.perl. We add fullwidth and halfwidth (marked -a for ascii) versions. We could add more as needed.

n_card-42 := card_2_no_valence &
 [ STEM < "４２" >,
   SYNSEM.LKEYS.KEYREL.CARG "42" ].

n_card-42-a := card_2_no_valence &
 [ STEM < "42" >,
   SYNSEM.LKEYS.KEYREL.CARG "42" ].

• Bender, Emily M. 2002. Number Names in Japanese: A Head-Medial Construction in a Head-Final Language. manuscript (comments welcome)