Difference between revisions of "Splitting and joining words"

Revision as of 10:59, 21 July 2010

Items treated as compounds

Items that are split

Definite article (determiner).

(NP-VOC (N-N gæska$-gæska)
        (D-N $n-hinn)
        (NP-POS (PRO-N mín-minn)))

Suffixed þú 'you' on finite verbs. -du, -ðu-, -tu is always NP-SBJ:

	      (ADVP-RSP (ADV þá-þá))
	      (VBPI veis$-vita)
	      (NP-SBJ (PRO-N $tu-þú))
	      (NP-OB1 (PRO-A það-það)

Verbs and particles written as one, that is combinations of particles (RP) and verbs (VB) are split off.

Items treated as unitary

Items of this kind may often either be written as one word or more. When they are written as one, they get a simple POS tag but when written apart each part gets its own numbered POS tag. Together they project one tag:

(NP-OB1 (NS-D (N21-A kapal-kapall) (NS22-D hestum-hestur))))))

The first number (2) is the number of parts in the item, the second one (1 / 2) shows each part's place within the sequence.

Note that in the case of nouns, adjectives and pronouns, different parts of these items usually don't have the same case, as in the example above; the first part usually gets accusative or genitive since it modifies the last part.

Genitive modifiers in sequences like these do not project NP-POS.

Names like Skalla Grímur, usually written as one orthographic word, are treated a little bit differently from regular nouns like manns bani (which is also usually written as one):

(NP-SBJ (NPR-G Skalla) (NPR-N Grímur))

(NP-SBJ (N-N (N21-G manns) (N22-N bani))

Treatment of individual items and parts

NÉ EITT (NEITT): (NP-SBJ (Q-N (NEG21 né-né) (ONE22-N eitt-einn)))

Ó: (ADJP (ADJ-N (NEG21 ó-ó) (ADJ22-N heilagur-heilagur)))

SYSTUR SON EGILS:

(NP-PRN (N-N (N21-G systur-systir) (N22-N son-sonur))
			  (NP-POS (NPR-G Egils-egill))))