AdpType: adposition type
Czech has neither postpositions nor circumpositions but there are several forms of prepositions that this feature distinguishes.
Prep: (normal) preposition
Examples
- v “in”, na “on”, o “about”, z “of”, s “with”, do “into”, k “to”, pro “for”, za “behind”, po “after”
Voc: vocalized preposition
Some Czech prepositions are non-syllabic and their form has to be changed in some contexts to facilitate pronunciation. Moreover, some syllabic prepositions are altered too, if the following word starts with certain consonant clusters.
Examples
The first line shows examples of vocalized preposition forms, the second line shows corresponding base forms.
- ve, se, ze, ke, ode, beze, ku, skrze, přede, nade
- v, s, z, k, od, bez, k, skrz, před, nad
Comprep: dependent part of compound preposition
This value marks dependent first part of a compound preposition. This word cannot occur alone.
Not all compound prepositions contain words marked Comprep.
Many compound prepositions consist of two normal prepositions and a noun
(an example is na rozdíl od “in contrast to”).
Sometimes there are just two words, the second one is normal preposition and the first one
is a secondary preposition (etymologically some other part of speech, but it has been frozen as
a preposition).
Examples
- vzhledem k(e) “due to”, nehledě na “regardless of”, narozdíl od “in contrast to”
Treebank Statistics (UD_Czech)
This feature is language-specific.
It occurs with 3 different values: Comprep, Prep, Voc.
145943 tokens (10%) have a non-empty value of AdpType.
227 types (0%) occur at least once with a non-empty value of AdpType.
117 lemmas (0%) occur at least once with a non-empty value of AdpType.
The feature is used with 1 part-of-speech tags: cs-pos/ADP (145943; 10% instances).
ADP
145943 cs-pos/ADP tokens (100% of all ADP tokens) have a non-empty value of AdpType.
ADP tokens may have the following values of AdpType:
Comprep(286; 0% of non-emptyAdpType): vzhledem, nehledě, narozdílPrep(134515; 92% of non-emptyAdpType): v, na, o, z, s, do, k, pro, za, poVoc(11142; 8% of non-emptyAdpType): ve, se, ze, ke, ode, beze, ku, skrze, přede, nade
| Paradigm v | Prep | Voc |
|---|---|---|
| Case=Acc | v | ve |
| Case=Loc | v | ve |
AdpType seems to be lexical feature of ADP. 91% lemmas (106) occur only with one value of AdpType.
Relations with Agreement in AdpType
The 10 most frequent relations where parent and child node agree in AdpType:
ADP –[mwe]–> ADP (569; 53%),
ADP –[case]–> ADP (17; 63%),
ADP –[conj]–> ADP (13; 93%),
ADP –[foreign]–> ADP (1; 100%).