[developers] prefixation of multi word lexemes
Stephan Oepen
oe at csli.Stanford.EDU
Mon Aug 8 10:08:34 CEST 2005
hei!
if only to document a (cute, in a sense) debugging case for a known bug
in both the LKB and PET. woodley and i discovered last night that for,
say, `let's sleep!', we get an extra analysis like this:
(127 hcomp 0.0 0 3
(126 punct_sqleft_rule 0.0 0 2 (106 lets_2 0.0 0 2 ("let s " 0 2)))
(117 punct_bang_rule 0.0 2 3
(116 bse_verb_infl_rule 0.0 2 3 (109 sleep_v1 0.0 2 3 ("sleep" 2 3)))))
using the following pieces:
punct_sqleft_rule :=
%prefix (!f '!f)
punctuation_sqleft_rule.
lets_2 := va_lets_imp_le &
[ STEM < "let", "s" >,
SYNSEM.PHON.ONSET con ].
all not too surprising, given the current treatment of prefixation, as
the preprocessor strips off the |'s|.
ann, what is the current status in the discussion of prefixation? i am
tempted to patch this to always attach prefixes to the left periphery.
cheers - oe
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (ILN); Boks 1102 Blindern; 0317 Oslo; (+47) 2285 7989
+++ CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++ --- oe at csli.stanford.edu; oe at hf.uio.no; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
More information about the developers
mailing list