LRNOM

Descriptions:

Nominalizations

Fields:

FieldNameNotes
1EUIEUI of nominalization
2BASBase form of nominalization (noun)
3SCASyntactic category (noun) of nominalization
4EUIEUI of ref. nominalization
5BASBase form of ref. nominalization
6SCASyntactic category (adj or verb) of ref. nominalization

Algorithm:

  • Load EUI|spVars into a hash table

  • For noun: find all records with nominalization (" nominalization_of=nomCit|nomCat|nomEui")
  • For adj & verb: find all records with nominalization (" nominalization=nomCit|nomCat|nomEui")
  • Leave nomEui blank if no EUI information is available
  • Assign nomCat to noun if no nomCat is available and cat is verb or adj
  • Print out base|cat (noun)|EUI|nomBase|nomCat (verb or adj)|nomEui only if nomEui exists in the LEXICON, including follows:
    • citation|cat (noun)|EUI|nomCit|nomCat (adj or verb)|nomEui
    • spVars|cat (noun)|EUI|nomCit|nomCat (adj or verb)|nomEui

    • citation|cat (noun)|EUI|nomSpVars|nomCat (adj or verb)|nomEui
    • spVars|cat (noun)|EUI|nomSpVars|nomCat (adj or verb)|nomEui

PS. spVars of expansion are only available when the expansion is in LEXICON (with EUI).

Theoretically, all nominalizations should be symmetric (noun vs. adj and verb). Thus, the size of LRNOM file are doubles from all nominalizations. A unique sort (sort -u) program should be used to run through LRNOM after it's generated.

A sophisticated test should be performed to ensure the symmetric characteristic of nominalization.