Person and Number morphemes

Daniel Abondolo abondolo at DIAL.PIPEX.COM
Sun Nov 4 11:11:34 UTC 2001

In many Uralic languages the morphemes encoding person and number agreement
are to this day close to transparently separate (and many of the portmanteau
suffixes may be reconstructed as the result of fusion of subsequent and
separate morphemes, e.g. Hungarian <vagyunk> 'we are' from earlier
*[woD-m-uk], with [-m-] first person and [-uk] plural).

It is a question of the degree of morphonological analysis the investigator
is willing to countenance. For example, the preterite paradigm of Udmurt
myny- 'go' (<y> = hi back unrounded vowel),

s1 myni
s2 mynid
s3 myniz
p1 mynimy
p2 mynidy
p3 mynizy

can easily be analysed as

s1 myn-i-m m> myni
s2 myn-i-d
s3 myn-i-z
p1 myn-i-m-y
p2 myn-i-d-y
p3 myn-i-z-y

wherein -i- = preterite, -m/-d/-z = 1st/2nd/3rd person subject, and -y =
plural (subject). The only cost is a morpheme-sensitive rule which converts
word-final  m to zero; the cost is mitigated by the usefulness of the rule
elsewhere in the grammar, in both the verbal and noun paradigms.

Because so much person-and-number marking in proto-Uralic and the subsequent
daughter proto-languuges can be reconstructed as agglutinative, it is not
surprising that parallel synchronic analyses are conceivable for parts of
the verb paradigms of Mansi, Mordva, and Mari, as well as for person-marked
forms of nouns in Mordva and Saamic. In the case of Saamic languages, to
take an admittedly extreme example, one might reckon with separate
(possessor) person morphemes  1st/2nd/3rd -mA/-tA/-se and number
morphemes -0/-n/-k (or -t, depending on the language). I think that whether
one entertains these as examples of what Jordan Lachler is referring to
depends on the degree to which one is willing to invoke allomorphy and/or
rich morphonological alternations.


Daniel Abondolo

----- Original Message -----
From: "Jordan Lachler" <lachler at UNM.EDU>
Sent: Sunday, November 04, 2001 3:02 AM
Subject: Person and Number morphemes

> Greetings,
> I am looking for references to languages which have both person and number
> argument marking on the verb, and/or person and number marking for
> possessor on the noun, but where these categories are not combined in the
> same portmanteau morpheme, and are instead expressed by separate
> This is the case in Keresan languages, for example, where the transitive
> verb (and the possessed noun) has the following basic structure:
> Person + Object/Patient # + Stem + Subject/Agent #
> So, for instance, consider the following forms:
> (Note: + is used to represent a barred-i)
> si.u.m'eek'um'e            "I am watching him"
> si.u.m'eek'um'e.sraty+     "we (dl) are watching him"
> si.u.m'eek'um'e.shi        "we (pl) are watching him"
> Here, the prefix si- indicates 1st person subject and a third person
> (as well as Factual mode), the prefix -u- indicates a singular object,
> followed by the stem -m'eek'um'e- "watch someone/something", followed by
> either "zero" for singular subject, -sraty+ for dual subject, or -shi for
> plural subject.
> Likewise with the verb stem -kacha "see someone/something".
> si.u.kacha "I saw him/her"
> si.y'uu.kacha "I saw them two"
> si.w'aa.kacha "I saw them (3+)"
> Here, the prefix si- again indicates a first person acting on a third
> person in the Factual mode, followed by either -u- for a singular object,
> -y'uu- for a dual object, or -w'aa- for a plural object, followed by the
> stem -kacha.
> A similar pattern is found with possessed nouns.
>'a    "his/her buttock(s)"
> g.u'u.much'a   "their (dl) buttocks"
> g.uw'a.much'a  "their (pl) buttocks"
> Here, the prefix g- indicates a third person possessor, followed by either
> the prefix -au- for a singular possessor, -u'u- for a dual possessor, or
> -uw'a- for a plural possessor, followed by the noun stem -much'a
> (These patterns hold pretty well for transitive verbs and possessed nouns,
> but intransitive verbs are a different story.)
> If anyone knows of references to similar phenomena in other languages, I
> would appreciate hearing about them.  I'll post a summary of responses to
> the list.
> Thanks very much!
> -<>-<>-<>-<>-<>-<>-<>-<>-<>-
> Jordan Lachler
> Department of Linguistics
> University of New Mexico
> lachler at

More information about the Lingtyp mailing list