Phrase Types

From Icelandic Parsed Historical Corpus (IcePaHC)
Revision as of 07:12, 29 April 2011 by Joelcw (Talk | contribs) (NP - Noun Phrases)

Jump to: navigation, search

This is supposed to be a list of all the phrase types found in the corpus. Most of those are supposed to be compatible with the annotation scheme used in the Penn historical corpora and the PCEEC (Beatrice Santorini). Everything Icelandic specific should be documented here. See also head types.

Clauses

IPs

CPs

Internal Structure of non-clausal Phrases

Verb Phrases

Verbs are attached to the IP

NP - Noun Phrases

  • NP (noun phrase)
  • NP-ADV for constructions such as one with HVOR; also for adverbial free relatives, see CP-FRL. Also adverbial genitives attached at the IP level (click on NP-ADV for an example). For some datives at the clausal level, usually with locative meaning (e.g. "öðru megin"). Not for instrumental datives; see NP-ADT.
  • NP-CMP is for dative NP comparatives like (ADJP (NP-CMP sögu) (ADJR ríkari)), (ADJP (NP-CMP gulli) (ADJR betri))
  • NP-PRN, appositive (ísl. viðurlag), as in Fagridalur [hinn eystri] 'Fagridalur the more.eastern', Henry [the king], John [the great], etc. NOTE: this is also used for expressions like "Bjarni" in "Þau Bjarni" (= "they-NEU, including Bjarni").
  • NP-OB2, indirect object, or second object in copular constructions.
  • NP-PRD for copular constructions (and naming constructions) in accordance with the YCOE guidelines, but different from the PPCME2, PPCEME.
  • NP-POS possessive NPs, and *all* genitive NPs inside another NP or inside an ADJP where the genitive indicates some particular relationship between two NPs that is possessive, akin to possession (e.g. "my father's house", e.g. "Rome's destruction") or a set-subset relationship between the two nouns (e.g. "fegurst fjallanna" 'most beautiful of the mountains'). Note that the set-subset category also includes cases where there is a genitive quantifier or number phrase (e.g. "önnur tveggja" = one of two).

There is no -POS extension on NPs where there is neither a possessive relationship between the two NPs nor a set-subset relationship, e.g. "(ADJP fagurt (NP álits))". This is *not* exactly the same as in the PPCME2,PPCEME, where some of these genitives would be labeled "NP-COM". See the guidelines for NP-COM. Click on the link for more detailed discussion.

  • NP-COM NP complements of N which are *not* genitive; see MÓTI; see PPCME2,PPCEME guidelines. This includes adnominal datives which indicate possession (these are not tagged NP-POS). We do *not* use the "NP-DPS" label at all, which is found in the PPCME2. For cases that would be parsed under the guidelines for NP-DPS in the PPCME2, click on the link for [NP-COM].
  • NP-ADT For clause-level dative adjuncts, e.g. instrumental datives. Not for the second object in copular constructions: cf. [NP-OB2]. Also *not* for bæjum skipað; see ADJP#Complements_of_ADJ. Also for types in the PPCME2, PPCEME guidelines.
  • NP-TMP noun phrases denoting a time of occurrence or a frequency of occurrence.
  • *CON* (conjunction reduction)

Extraction out of NP

ADJP - Adjective Phrases

  • ADJP (adjective phrase)

PP - Prepositional Phrases

ADVP - Adverbial Phrases

If the meaning of an adverbial phrase is directional, locative or temporal, it always has an extended tag, whether it is at the clause level or not.

  • ADVP (adverbial phrase)
  • ADVP-LOC, locative adverbial phrase, like þar 'there, in that place', víða 'all over the place, everywhere'
  • ADVP-TMP, temporal adverbial phrase, like 'now', forðum 'in earlier times'

CONJP - Conjunction Phrase

  • CONJP (conjunction phrase)

WH-Phrases

  • WNP (wh- noun phrase)
  • WPP (wh- prepositional phrase)

Extended Tags

  • -LFD (left dislocation) - associated with a resumptive element
    • PP-LFD ef ... þá 'if ... then'
  • -RSP (resumptive element)
    • NP-SBJ-RSP þeir sem deyja, þeir eru dauðlegir 'they who die, they are mortal'

When a clause is both direct speech and left-disclocated, the order of tags is -SPE-LFD

Other

  • LATIN A string of latin words is dominated by a latin node and if it serves a grammatical function in the clause that phrase node dominates the latin node