The negative form of do when used
as present- or simple-past-tense pro-form is treated as textual ellips;
i.e. the lexical verb is copied onto the place of the negative
do-node. The inserted node inherits all relevant
attribute values from the original verb. The negation is rendered as a
#Neg.RHEM -node governed by the copied node. The node
for do is attached to the newly copied node as
its auxrf. The copied verbal node is assigned the same valency frame as
the original occurrence of the given verb. Obligatory modifications are
completed into the frame, having the appropriate t-lemma
substitutes.
Do you know him? "I <don't>
{know[negation=neg1]
#PersPron}.
#PersPron corefers to "him". The
copied node has the grammateme [negation=neg1]. The
same applies to will, would, should,
be as passive auxiliary, have as
perfect-tense auxiliary etc. (even when negated).
"Will you go there?" "I <will>
{go #OblFm.DIR3}.
The positive form of do, used as pro-form, however, is a different case. E.g. in
Did Peter take a plane to New York today? His wife did.
the lexical verb (take) is
not to be copied into the second sentence as elided since mechanical
insertion of the lexical verb on the surface behind the positive
do might modify the meaning of the sentence (e.g.
by moving the focus). Instead, do is attached as
auxrf to a newly inserted node with the -lemma substitute
#VerbPron. This new node corefers to the given verbal
antecedent. The valency frame of the antecedent is
assigned to the #VerbPron node and its obligatory modifications are inserted.
All modifications governed by the original do
naturally stay in their places, and they are assigned functors according
to the valency frame inherited from the antecedent verb. Coreference
arrows point from the present inner participants to their
antecedents.
Whenever do has object it is treated as lexical verb:
