Proposed set of grammatical relations

[Brian MacWhinney]

Dear Info-CHILDES,

   Alon Lavie, Kenji Sagae, and I have been working on the development of a
system for the automatic computation of grammatical relations in the CHILDES
database.  The work has two goals.  The first is the complete automation of
the calculation of IPSyn and DSS on the basis of a fully tagged %mor line
and a complete tagging of grammatical relation on a %grr line. For this
work, we have to compute exactly the grammatical relations required by these
measures.  The second goal is the computation of all grammatical relations
that play a significant role in current thinking and research about
syntactic development.  For this purpose, we are seeking community-wide
input.

   Our proposal for a set of grammmatical relations is given at
http://childes.psy.cmu.edu/grasp.doc

We are hoping that people who are interested in the automatic analysis of
the grammatical relations in CHILDES can read this proposal and send
feedback to Brian MacWhinney at macw at cmu.edu if they have suggested changes
in this list or well-specified alternative proposals regarding the general
issue of parsing of the CHILDES database.  Commentary on this proposal will
be collected and eventually posted both at that web site and eventually on
the info-childes at mail.talkbank.org list and in the IASCL newsletter.

  Many thanks
  Brian MacWhinney