Appendix 2: Tagging decisions of APPLYHYPEN

Notes

  1. "WIC means "Word-initial Capital".
  2. See Note 4, Appendix 1.
  3. The "Hyphen-List" consists of "class", "hand", "like", "price", "proof", "quality", "range", "rate", and "scale".
  4. See Note 5, Appendix 1.
    For words not ending in "s", if IN is one of the tags, tag the word NN JJ@; if VBN is one of the tags, tag the word JJ; if VBG is one of the tags, tag the word JJ NN VBG@; if NNU is one of the tags, tag the word JJB; if NN with "normal" probability is one of the tags, tag the word NN JJB; otherwise leave the tags unchanged.
  5. For words ending in "s", if IN is one of the tags, tag the word NNS; if VBG is one of the tags, tag the word NNS; if NNU is one of the tags, the tag is JJB; if NN with "normal probability" is one of the tags, the tag is NNS; otherwise retain tags that take "s" (see Note 5, Appendix 1). If there are none, then tag the word NNS VBZ.