Technical results

Preliminary Version,
January 2016

The annotation

This part of the Report is currently under construction.
The the products of the annotation may be seen in the following temporary addresses:

Navigation experiment
Manual

Linguistic searches with Corpus search

In order to evaluate the possible outcomes of this annotation, a few searches for patterns have been conducted, with two aims.

First and foremost, to make sure the proposed annotation is adequate for the conduction of syntactic research with automatic tools, as was the aim of the project. Second, to see if some of the hypothesis in the main project could already be confirmed or disproven on the basis of the initial study of this text.

In order to do this, a specific set of the full data was selected: nouns phrases projected by nouns, in high argumental positions (i.e., subjects and accusative objects), in main clauses. Over this selected data a few experiments of automatic searches with Corpus Search were conducted, focusing on observing the patterns of order – marked in the previous syntactic annotation – in combination with the referential properties marked in the present annotation.

As regards the first aim of this battery of tests – checking the adequacy of the annotation for automatic research – the results could be classified as very promising. The combination of the previous syntactic annotations with our sub-annotation for reference status works very well, and can be conducted in two ways – either the queries can target the internal syntactic structure and then analyze the reference ID tags that surface, or they can target combinations of reference IDs, short-codes for constructions, and patterns marked in the syntactic annotation. The latter style is the most promising one, and it based the first results below. A few of the queries conducted are available for consultation here.

As regards the second aim, the first results of the searches in terms of linguistic relevance are exposed in the section Linguistic Results.

Anúncios