To: Michael Sperberg-McQueen 312 996-2477 -2981 <U35395@UICVM.uic.edu>
Subject: Word class tags - summary
Date: Wed, 09 Jan 91 13:00:50 EST
From: anderson@sapir.COG.JHU.EDU

Michael -

I've summarized our results on how to tag words below. Well, it's not
just a summary, because I've managed to sneak in a few changes that
occur to me as I go over the whole things. This list includes stuff from
your notes, and also from the discussion Geoff and Nicoletta and I had
while you, Gary and Terry were solving the real problems.

It seems to me that if the formalism supports it, it's worth imposing a
bit more structure on this: in particular, adding properties that are at
the same level as "wordclass", like "form" below. So I've divided this
summary into three parts: (1) a set of properties that recur within the
specification of multiple classes, though generally subject within any
given langauge to some class-specific restrictions implemented via the
prologue mechanism you were describing to me; (2) the overall tag for a
word, which at present contains an internally-structured class and a
specification of word-form type; and (3) the content of the individual
word classes.

Of course I didn't prepare this in the form of an SGML-tagged
document.....hope you can read it anyway. I'll be in touch again when I
have a chance to go through grammars of Danish and Russian with this
scheme in mind.

Steve

------------------(included file tei_wordtags.txt)------------------------

                   TEI tags for words

1. Some properties that recur in more than one word class:

Person = 1st | 2nd | 2Polite | 2Familiar | 3 | indefinite | unspecified
    [add 1Inclusive | 1exclusive ?]

Number = sg | pl | unspecified
    [add | du ?]

Gender = masculine | feminine | neuter | common | invariant | unspecified

Animacy = animate | inanimate | unspecified

Case = nom | gen | dat | acc | voc | instr | loc | subj | obj | obl |
 unspecified

Degree = positive | comparative | superlative

Definiteness = definite | indefinite | unspecified
    [add specific | nonspecific | generic ?]

Deixis = proximal | distal | remote | unspecified

Affect = diminutive | augmentative | pejorative

2. An overall tag for words:

Word

    Wordclass = Noun | Pronoun | Adjective | Verb | Adverb | Preposition
    | Coordinator | Subordinator | Particle | Interjection | Punctuation

    Form = phrase | compound | full | reduced
    | clitic | proclitic | enclitic

3. Properties of the various classes:

a. Noun:
number
gender
animacy
case
proper = proper | common
definiteness
[add deixis for French "cette femme-ci/l`a" ?]
affect
[ What's "w.i.c.", and does it still belong here? -sra]

b. Pronoun:
person
number
gender
case
deixis
type = personal | demonstrative | indefinite | expletive | | relative |
    interrogative | possessive | reflexive | reciprocal | partitive |
    locative | propredicate| zero
pro-form = emphatic | disjunctive | conjunctive

c. Adjective:
number
gender
animacy
case
degree
numeral = cardinal | ordinal
pronominal = possessive =[ person
                           gender
                           number ]
             | relative | interrogative | demonstrative | indefinite
declension = strong | weak | long | short

d. Verb:
agreement = [person
             number
             gender
             case] (we omitted this, but participle agreement may need it - sra)
tense = present | past | future | imperfect | aorist | perfect | imperfect |
    pluperfect | future-perfect | conditional
aspect = perfective | imperfective
mood = indicative | subjunctive | imperative | conditional
voice = active | middle | passive | mediopassive
polarity = positive | negative
verb-form = finite | infinitive | gerund | supine |
            participle = [ participle-type = present | past | future ]
p-incorporation = [ do = []
                    io = []
                    partitive
                    locative ]
verb-type = lexical | auxiliary | modal | copula

e.Adverb:
degree
deixis
directionality = static | dynamic    (e.g. Danish ind vs. inde)
interrogative = + | -
[add type = locative | temporal | manner | .... ?]
[add incorporation = something, for German dabei, darauf, herein, etc.?]

f.Preposition:
art-incorporation = [ number             (for Italian agli,
                      gender                  French du,
                      case                    German zum, etc.)
                      definiteness]
pro-incorporation = [ number             (for Portugese, maybe also
                      gender                  German damit, etc. ?
                      case ]                  if so, need more properties)

g. Coordinator:
(no internal structure)

h. Subordinator:
(no internal structure - except that some Dutch and spoken German dialects have
agreement with the lower subject marked on some complementizers.....)

i. Particle:
(any uninflectable item that doesn't belong somewhere else - no general internal
structure motivated)

j. Interjection:
(no internal structure)

k. Punctuation:
orientation = open | close | matched | unary

---------------------

Existing TEI tags <formula>, <foreign>, <number> seem to be usable for those
objects without extension.