Some edges of a tectogrammatical tree do not represent dependency. We establish such non-dependency edges in order to represent parataxis and some other specific syntactic relations in the tree.
A non-dependency edge is:
the edge between the root node of a sentence and the technical root node of the tectogrammatical tree (nodetype=root).
It is an auxiliary (technical) edge without a linguistic interpretation.
the edge between the effective root node of an independent clause and its mother node.
The functors for the effective root nodes of independent clauses (PRED, DENOM, PARTL, VOCAT, PAR) express non-dependency and they tell us of what type the clause is. This edge simply integrates the particular nodes (subtrees) into the tectogrammatical tree.
For more see Section 4, "Verbal and non-verbal clauses" and Section 1, "Functors for the effective roots of independent clauses".
edges in paratactic structures:
the edge between the paratactic structure root node (nodetype=coap) and its mother node (nodetype≠coap),
the edge between the paratactic structure root node and a direct element (member) of the paratactic structure,
the edge between the paratactic structure root node and the effective root node of a shared modifier.
As for modifications within a paratactic structure, their dependency is always represented by two edges at least. For example the dependency of a terminal element of a paratactic structure on its governing node is indicated in a simple, non-embedded paratactic structure by a combination of edges of type a) and b). In an embedded paratactic structure, the dependency of a terminal element of the paratactic structure on its governing node is indicated by a set of b) type edges and one a) type edge. The dependency of a shared modifier on the terminal members is indicated by a combination of edge type b) and c).
For more details on paratactic structures (including definitions of the terms) see Section 6.1, "Representing parataxis in a tectogrammatical tree".
edges in list structures:
the edge between the root of the list structure (nodetype=list) and its mother node.
the edge between the root of the list structure and a list item (nodetype=fphr) or the effective root node of an identifying expression (functor=ID).
the edge between the root node of a list structure and the effective root node of a modifier of the list.
Edges between list structure nodes have various meanings depending on the type of the list structure. In list structures for foreign-language expressions, type b) edges only gather individual nodes in the list (they do not express dependency) while type a) edges express dependency of the entire list structure on its governing node. In identification structures (which are list structures, too), dependency is represented by both type a) and b) edges. Edge type a) expresses dependency of the entire identification structure, edge type b) expresses dependency of the effective root node of the identifying expression. Edge type c) represents (in both cases) dependency on all items of the list as a whole.
For more on list structures for foreign-language expressions see Section 9, "Foreign-language expressions".
For more on list structures for identifying expressions see Section 8.1.3, "Identification structure".
the edge between an atomic node (nodetype=atom) and its mother node.
Edges connecting atomic nodes integrate these nodes into the tree. Their meaning varies according to the functor of the particular atomic node. Edges connecting atomic nodes with the functor RHEM determine the position of the rhematizer within the deep structure and it defines its scope. Edges connecting atomic nodes the functors of which are ATT and MOD express potential dependency - for more details see Section 11.1.3, "The semantic scope of modifications with the functors MOD and ATT". Edges above atomic nodes with the functor PREC integrate the sentence into the preceding context.
For more on functors of atomic nodes see Section 7, "Functors for rhematizers, sentence, linking and modal adverbial expressions". For more details on rhematizers see Section 6, "Rhematizers".
the edge between a node with the functor DPHR, CPHR or CM and its mother node.
A node the functor of which is DPHR, CPHR or CM expresses the fact that it constitutes a single lexical item together with its mother node. Therefore, the edge expresses that the two parts belong together - it does not signal dependency.
For more details see Section 9.3, "Complex predicates", Section 8, "Idioms (phrasemes)" and Section 16.1.2, "Conjunction modifiers".