Encoding of semantic information

  we eat  -> one argument case  -> eat1a  we eat fish  -> two argument case  -> eat2a

The final alphabetic character is employed because there might be semantic variants with the same number of arguments:

  we gave fish to Felix  -> give3a
  
  we gave Felix fish  -> give3b

To simplify the lexicon we assume that eat1a, eat2a, etc. are mere logical constants to be found in our language of semantic representation, and we neglect the logical relationship that might exist between these constants. With such a simplistic view of semantics the lexicon just has to tell us what constants correspond to what words. This can be done with the featural apparatus we already have:

(Figure 7.12)

We have to specify our syntactic class abbreviations because we assume that every DAG splits into a syn-branch, a sem-branch and a mor-branch.

Macro syn_iV:
    <syn cat> = V
    <syn arg0 cat> = NP
    <syn arg0 case> = nom.

This leads to such lexical entries:

Lexeme die:
    syn_iV
    <sem> = die1a.

Lexeme elapse:
    syn_iV
    <sem> = elapse1a.

Lexeme eat:
    syn_iV
    <sem> = eat1a.

Lexeme eat:
    syn_tV
    <sem> = eat2a.

Lexeme give: 
    syn_tV
    <sem> = give2a.
Lexeme give:
    syn_dtV
    <sem> = give3a.

Lexeme give:  
    syn_datV
    <sem> = give3b.

Lexeme hand:
    syn_dtV
    <sem> = hand3a.

Lexeme hand: 
    syn_datV
    <sem> = hand3b.

Lexeme love:
    syn_tV
    <sem> = love2a.