This document discusses some unclear cases in the tagging.
Plural
Mari has 4 plural suffixes.
- пӧрт-влак = Plural +Pl
- пӧрт-шамыч = Plural +Pl
- Трондмыт = Trond and his friends / family etc. +AssocPl
- яллаште, яллашке, ялла гыч, … +LocusPl
- ял = village, -ла is a plural used in combination with local
cases. -влак/-шамыч could be used here as well, but -ла is
sort of preferred
Difference between (1) and (2) is dialectal. (3) is
associative plural. (4) is referring to several locations.
The plurals 1-3 trigger Pl3 agreement in verbs, the fourth one does not.
Derivational genitive
We have no tag of the derivational genitive
- мемнан-влак = наши.
- мый мемнаным ужам = I see ours
- Шке шонымашым шуктыман, а еҥыным огыл. - You have to carry out your own
wishes, not those of others. еҥ-ын-ым: other.person-GEN-ACC
мемнаным = мый+Pron+Pers+Pl+Gen+Der/Poss+N+Sg+Acc
мемнаным = мый+Pron+Pers+Pl+Der/Poss+N+Sg+Acc
"<мемнаным>"
"мый" Pron* Pers Pl Der/Ellipsis N Sg Acc
"мый" Pron* Pers Pl Der/MWN N Sg Acc
пурышо лӱддымӧ разведчик шуко годым шкенжын-шамычше дене кылым телефон але
Abessive
Productive in Western, sporadic in Eastern, forbidden by normative grammars.
Vocative
ава to авай = vocative with -и for a class of nouns (Jeremy should deliver a
list)
- авай = авай N Sg Nom (not optimal because no info on vocativity)
- авай = ава N Sg Voc (bad because of case stacking)
- авай =
- ава+N+Der/Voc+N+Sg+Nom (morphology)
- “ава” N* Der/Voc N Sg Nom (input to syntax, which will be @Vocative
It seems (3) is the best.
ӱдыр to ӱдырем = for all (other) words - do not mark
- a. Keep the morphology as Px, add @Vocative in the syntax
- b. Have an analysis Der/Voc parallel to Px and then disambiguate
It seems (1) for удыр and (3) for авай is best.
Possessive suffixes
- PxSg1, PxSg2 ok
- What about PxSg3, which is also used as what some call a clitic
- a. two tags: PxSg3 and +Cl
- b. one tag PxSg3 and two syntaxtic tags
Superlative
- tag it for one word
- not tag it
==> Tag it
Cardinality
Skip Card? Skip Ord?
Moods
Throw out +Cond
+Imprt: We want all attested forms, also Sg1?
+Inf may have +Dat
We want to have all +Prc as fused tags
- +ActPrc
- +PassPrc
- +FutPrc
- +NegPrc
Then +Act and +Pass do not exist.
+Foc/я is missing, add it. (e.g. тол-я Oh come, please or something like
this)
The description of gerunds is not yet optimal. Mari has the following gerunds,
with all but the gerund in -н being attached directly to the verbal stem - with
stem simplifications/alternations, if necessary. The gerunds of ышташ (-ем)
to do:
- Affirmative instructive gerund: ыштен
- Negative gerund: ыштыде
- Gerund of prior action: ыштымек(е)
- Gerund of future action: ыштымеш(ке)
- Gerund of simultaneous action: ыштышыла
Some open questions:
- How do we want to gloss them? My suggestion would be - and in parentheses how
they are glossed now:
- +AffGer (+Ger+Gen)
- +NegGer (+Ger+Abe)
- +PriGer (+Ger+Prf)
- +FutGer (+Ger+Imprf)
- +SimGer (+Ger)
- The software does not yet produce/recognize the short forms of PriGer and
FutGer. These always exist - every verb has got them. The only overt
difference between the short and long forms is that the long forms
can take possessive suffixes (see below).
(Furthermore, NegGer has a long form -дегеч (here then ыштыдегеч), not
included in prescriptive materials, but worth including.)
- PriGer, FutGer, and SimGer can take possessive suffixes (the first two
cannot). For PriGer and FutGer, they’re attached to the long forms by regular
means. Strange stuff makkens with SimGer: the possessive suffix occurs between
the two elements (-шы & -ла) of the gerund. Furthermore, possessive suffixes
are NOT used in 1PL and 2PL. So the forms
we want are - again of ышташ (-ем) ‘to do’:
- 1Sg толшемла
- 2Sg толшетла
- 3Sg толшыжла
- 1Pl -
- 2Pl -
- 3Pl толшыштла
Derivational morthology
+Der/ .. a lot of tags just not implemented yet.
Here’s what’s missing from 100% productive morphology - everything else we can
leave to the lexicon, I’d say.
Verb > noun
All of these suffixes are attached to the verbal suffixes, all
the usual simplifications/alternations occur (as did with participles)
- -маш: nominalizer. тол- to come > толмаш coming
- -дымаш: negative nominalizer. тол- to come > толдымаш not coming
Noun > adjective
- -ан (-ян, -н): possessive adjective, having X: вӱд water > вӱдан
watery
- -(ы)сЕ (E = vowel-harmonic е/о/ӧ): relational adjective,
spatially/temporally related to X: вӱд water > вӱдысӧ aquatic
- can follow the plural suffix -ла discussed above: пӧрт house > пӧртла
houses > пӧртласе of/in the houses
- -дымЕ: privative adjective, -less: вӱд water > вӱддымӧ waterless
- -лык: purposive adjective, for X: корно road > корнылык
for the road
- also adjective > noun to form abstract nouns: пашадыме unemployed >
пашадымылык unemployment
Noun/adjective > verb
- -аҥ (-яҥ, -ҥ): translative verbs (conj 1), to become X-y: вӱд ‘water’ >
вӱдаҥаш (conj 1) to become watery
- -ем (эм, -м): translative verbs, (conj 1), to become X: сай good >
саемаш (conj 1) to improve (intr.)
- only in combination with these two suffixes productive deverbal verbal
suffix (conj 2) -д, which is a transitivizer:
вӱдаҥаш (conj 1) to become watery > вӱдаҥдаш (conj 1)
to make watery, саемаш (conj 1) to improve (intr.) >
саемдаш (conj 2) to improve (tr.)
Verb > verb
- -алт (-ялт): reflexive/intransitive/passive/impersonal (conj 1):
ышташ (conj 2) to do > ышталташ (conj 1) to be done
- There is also a sporadically used variant -ылт, but I think we can leave that
one to the lexicon …???
- -(ы)кт: causative (conj 2): ышташ (conj 2) ‘to do’ > ыштыкташ (conj 1)
to make someone do
- Maybe we can have three slots for these derivational suffixes, so that we can
have stem-DER-DER-DER-(inflectional morphology), to allow for all
combinations: ышталтыкташ to make something be done, ыштыкталташ
to be made to do, … (one does find combinations of these suffixes, but
anything beyond three of them at a time is probably just silly.)