Guided tour to Águia

Águia (meaning "eagle" in Portuguese) is a search tool for treebanks, operating on the Web and using the IMS Corpus Workbench as the underlying corpus encoding tool. It was developed by Linguateca, www.linguateca.pt in the context of the Floresta Sintá(c)tica project, and can be accessed here. This page intends to give an idea of Águia's functionalities in a nutshell.

Please note that the query tool is still under development, so we are not worried (for the moment) about user friendliness or robustness, but with finding out the query expressiveness required. Any feedback about kinds of queries would therefore be most welcome (while comments about error messages or problems encountered are not relevant at this moment).

Basically, in addition to the standard questions you can ask of an annotated corpus , you should be able to get information that hinges on phrase or clause structure, and tree depth, etc. which are only available when your underlying linguistic objects are trees (or graphs):

In case you want to have quantitative information about the treebank,
- what kind of clauses are most frequent?
  distribution of STA, QUE, EXC and UTT in terms of constituents
- what kind of syntactic objects have the function question?
  [funcao="QUE"] distribution in terms of phrase distribution
- what is the most frequent verb in each kind of clause?
  /fcl[classe,"P:v.*"]
- what is the most common function of a finite clause?
  "fcl" distribution of function...
If one believes some categories should not be there, one could inspect them a little better
/fun_fcl['SC'] concordance
In case you want to look for specific examples of special cases
- in how many cases do adverbs occur in relative clauses? Definition of relative clause: ( N<:fcl followed by a pron-indp() )
- Find all clauses including an adverb as immediate constituent
  /ass_fcl['pron-indp .* adv .* ']
- Find all noun phrases including relative clauses in which the pronoun has the subject role
  /np[classe,SUBJ:pron-indp]
- Find all NP's including relative clauses in which the pronoun has the object or dative role
  /np[classe,ACC:pron-indp]
- Find all finite clauses starting with subject
  /ass_fun_fcl['SUBJ .*']
In case you want to look at the underlying generative grammar
- what is the generation grammar of a particular phrase?
  "ap" distribution of immediate constituents
- what is the generation grammar of a particular function?
  [funcao="SC"] distribution of immediate constituents
In case you want to determine the grammatical properties of a lexical item
- what is the valency grammar of a particular lexical item? (verb, preposition)
- given a particular class of adverbs, in which patterns they occur?
Other questions that include more than one query
- what is the depth of the embedding (find finite clauses under finite clauses)
- how many PPs are not directly attached to the preceding phrase?

Last update: 22 August 2006.
Author: Diana Santos.

Comments and suggestions