The upper part of the table follows the main organizing principles of the UD taxonomy such that rows correspond to functional categories in relation to the head (core arguments of clausal predicates, non-core dependents of clausal predicates, and dependents of nominals) while columns correspond to structural categories of the dependent
(nominals, clauses, modifier words, function words).
The lower part of the table lists relations that are not dependency relations in the narrow sense.
* The advmod relation is used for modifiers not only of predicates but also of other modifier words.
Individual languages may define more specific relations as subtypes of the universal types defined here.
A subtyped relation always starts with the basic type, followed by a colon and the subtype string.
In general, subtypes are language-specific and optional. However, some subtypes are assumed to apply
to many languages and they should be considered semi-mandatory: If the language has the phenomenon that
the subtype focuses on, then the subtype should be used. The following subtypes currently have the
semi-mandatory status:
Alphabetical listing
(This is not the list of all relation types used in all languages.
It is the list of relations that happen to have a documentation page available globally, as opposed to language-specific pages.
So it contains all main types plus the semi-mandatory subtypes discussed above, plus an arbitrary selection of other subtypes.
To see all subtypes that occurred in the most recent UD release, go to Relation Subtypes in the Data.)