Construction-based corrections

From Icelandic Parsed Historical Corpus (IcePaHC)
Jump to: navigation, search

1. Run query 99.q and fix errors with dashes and dots (99.q)

2. fix "stundum": Should be (NP-TMP (NS-D stundum)) (stundum.q)

3. fix "móti": (PP (P0) (NP (N-D móti)) (moti.q)

4. Fix "annars" when stands alone (= otherwise): Should be (NP-ADV (OTHER-G annars)) (annars.q)

5. fix "er": BEPI/C (er.q)

6. Fix "Er, eru, ertu.." that are wrongly tagged as VB* (Er.q)

7. Fix unsplit imperatives (VBI/BEI..), split upp "*rtu/*rðu" into verb+ "NP-SBJ $tu/ðu" (rtu.q)

8. Fix "sé", make sure that is correctly tagged as BEPS (where applies) (sje.q)

9. Fix "meðan" Should be a P projecting a CP-ADV (medan.q)

10. Þegar: Move CP-ADV under PP and IP-SUB under CP-ADV. (thegar.q)

11. Þó: Change to PP and make a CP-ADV (and IP-SUB) (tho.q)

12. Fix nouns and articles that have to split up (slitinngreinir.q)

13. Make sure that "aðeins, einungis, jafnvel, alleina, alleinasta" are tagged as FP (FP.q)

14. Split up RP's and verbs: "upp*, á*, af*, að*, fram*, frá*, fyrir*, inn*, í*, niður*, með*, of*, til*, um*, úr*, út*, við*, yfir*" (Rp.q)

14. Sem: Prepare CP-REL, move under noun