Internal grammar structure

Lexicon

A data structure for accessing efficiently several elementary trees. For each elementary tree there exists a lexical anchor, to find the rule consistently in this lexicon. The lexicon structure is recursive, because the lexical anchor could be a expression sequence of arbitrary length. Assumed we are seraching for a elementary tree with a multipart anchor. In that case the first part of the sequence would be searched and returns a sub-lexicon for all elementary trees, where the lexical anchor starts with that first part.

Elementary tree

An entry of the lexicon in TIG. It consists of a list of context free grammar rules (also Layer). The production rules together implicitly describe a tree as defined in the TIG. To transform such a tree into production rules, the tree is traversed in a breath-search manner. For each inner-node a production rule is created by transforming the categorie of the node as left side nonterminal. The children become either nonterminal with the categorie as label, if it is a inner-node, or terminal, if it is a leaf-node. The the label of the terminals depend on the different node-types. Furthermore it has some more properties, like an unique index and the previously mentioned lexical anchor.

Layer

Common production rules of a context free grammar extended by a Gorn-number indicating the position in the elementary tree. ###Gorn-number Unambiguously numeration in a tree for finding a specified layer.

Lexical anchor

For finding the rules which are required for parsing a sentence, it is usefull to create some index for them. This index is the lexical anchor for the for the TIG-rule. It consists of a sequence of expressions in the language.