10. a elaborated example (cont'd)
Step
4:
(CONSTITUENT
((:SYNTACTICALFORM :MCS))
(AGGREGATION COREDAGGOBJ
((CONSTITUENT "Anheben"
(CONSTITUENT "der"
(CONSTITUENT "Bauchdecke"))
(CONSTITUENT "mit"
(CONSTITUENT "zwei"
(CONSTITUENT "Backhausklemmen")))))
(CONSTITUENT "Einführen"))
(CONSTITUENT "des" (CONSTITUENT "Kameratrokars")))
(CONSTITUENT "Einbringen"))
(CONSTITUENT "der"
(CONSTITUENT "Optik")))))
:ELLIPSIS-NP T))
The
morphosyntactical expressions have now generated a syntax tree where only
one node (Aggregation)
is not yet evaluated. This function now
examines
the three noun phrases and detects that the two latter can be aggregated
since the heads of the noun phrases “Einführen”
(“Insertion“)
and
“Einbringen”
(“Installation”) are synonymous. Thus, the head is retained and the attributes
(“des Kameratrokars”, “der Optik”) are submitted to
the
Enum-function. Together with the first noun phrase the resulting conjunct
is then again submitted to Enum.
Step
5
(CONSTITUENT
((:SYNTACTICALFORM :MCS))
(CONSTITUENT "Anheben"
(CONSTITUENT "der"
(CONSTITUENT "Bauchdecke"))
(CONSTITUENT "mit"
(CONSTITUENT "zwei"
(CONSTITUENT "Backhausklemmen")))))
#\,
(CONSTITUENT "Einbringen"))
(CONSTITUENT "des" (CONSTITUENT "Kameratrokars")))
"sowie"
(CONSTITUENT "der"
(CONSTITUENT "Optik"))))
Please
note that the original syntax tree has to hold more information than shown
above. For the detection of synonymity, for instance, the original
lexemes
have to be kept along with the inflected forms.
Hopefully
it has become clear what Suregen-2 is intended to do.
Still
a long way has to be gone, comments are welcome.
-
previous: example (cont'd)
up (home) next: appendix
-