The Format for Valency Representation

For the computer processing of verb valencies we need to define a format to describe the valency expression schema. The format stores the grammatical information such as the grammatical case of a noun group, the preposition of the noun group or the type of a clause and its conjunction. The information is kept in a form suitable for any automatic valency processing.

The format we work with is called the BRIEF format. It uses an attribute-value notation. A valency expression is written as a string of pairs of an attribute character and its value in upper case letter or in braces. Four important attributes are the following:

1. attribute h -- type of expression (basic semantic features). Its possible values are:

   P -- person	L -- location
   T -- thing	A -- direction from
   Q -- quality	F -- direction to
   R -- reflexive pron.	D -- which way
   M -- amount	W -- time

2. attribute c -- grammatical case (grammatical features). There are seven possible values of this attribute:

   1 -- nominative, 2 -- genitive, 3 -- dative, 4 -- accusative, 5 -- vocative, 6 -- locative and 7 -- instrumental

3. attribute s -- type of clause and the conjunction of the clause (syntactic features). The values of the attribute s are codes of the conjunctions that are used in our list of valencies:

   I -- infinitive	P -- conj. at' (I wish)
   C -- conj. az (as soon as)	R -- rel. clause (which)
   D -- conj. ze (that)	U -- conj. aby (so as to)
   F -- conj. zda (if)	Z -- conj. jak (how)

4. attribute r -- preposition of a noun group (syntactic features). The value of this attribute is the particular preposition in curly braces.

When we write down the valency list for a verb, we separate the valency expressions that form a valency with a dash and all valencies of a verb with comma.

In the BRIEF format the representation of the list of valencies of the verb informovat looks as follows::

informovat <v>hPc4,hPc4-hPTc6ro,hPc4-sD