Pacifique’s conjugation 6: “Transitive animate”

The long-awaited “transitive animate” paradigm! To figure out how to conjugate a form, find the person of the subject by going down to the left-side column to find the right row; next move over to the column for the object you want. For example, a first person exclusive subject (13) acting on a second person plural object (2PL)––e.g. for “We saw you guys”––gets the ending -ulneg.

↓ Subj / Obj→ 1 13 12 2 2PL 3 3PL
1 (refl) X X -ul -ulnoq -(V)’g -(V)’gig
13 X (refl) X -ulneg -ulneg -(Ve)g’t -(Ve)g’jig
12 X X (refl) X X -ugg -uggwig
2 -i’lin -i’lieg X (refl) X -(V)’t -V’jig
2PL -i’lioq -i’lieg X X (refl) -(V)oq -(V)oqig
3 -i’lit -ugsieg -ugsi’gw -(V)’sg -ugsioq (refl) X
3PL -i’lijig -ugsieg -ugsi’gwig -(V)’sgig -ugsioq X (refl)

Before we get into the patterns themselves, a few points should be made: diagonally from top right to lower left you see cells marked “(refl)” for “reflexive”; these forms will get a special marking, not included here. In addition to the reflexive forms, there are certain cells that are impossible. Note that these are cells where features “overlap”. Meaning, if there is a first person specified anywhere in the subject, it can’t be anywhere in the object; same for second person. This rules out forms like “You all saw you”, and “We saw us”. In Mi’gmaq, third persons can act on third persons, but this is where obviation comes into play, which we will save for a future post.

Also, note that some forms begin with either a -(V) for vowel or a -(Ve). With a stem like nemi- “see”, the final vowel is kept in the stem when the (V) appears, but dropped elsewhere. For example: nemul “I see you”, but nemi‘g “I see him”. In other verb forms, the vowel doesn’t appear. Take taqam- “hit”: taqamul “I hit you” and taqam‘g “I hit him”. It would be interesting if we could find regularities here.

(I should preface what I am about to write by saying: the next part might be fun for people interesting in finding patterns here. That said, understanding all of it isn’t necessary to being able to use these forms! However, if you are interested and something I’ve written is unclear or just plain wrong, please post comments.)

Looking now more carefully at the table above, some basic patterns emerge. First, we can assign feature values to morphemes (pieces) as follows:

-i’li – 1 object; -ul(n) – 2 object; -a – 3 object; -ugsi – “participant” plural object; -eg – 13; -gw – 12; -oq – 2PL; -ig – 3PL; -n – 2 subject; -t – 3 subject

Now how do we figure out how to put all of these together? Though it becomes more clear in the negative forms (to be posted), we can think of these as involving 2 slots, plus an additional slot if there is a third person plural involved. So we can think of our template as looking roughly like this:

Verb – Slot1 – Slot2 – (3PL)

The first step is to figure out what goes in slot 1; this is the easier slot, because we can basically just think of it as object agreement: see the object forms listed above. A couple of tricks apply here. -i’li is triggered if a first person is part of the object, including first person plural; this means that a 13 object gets -i’li in a 2>13 form like -i’lieg. However, if we find a third person subject acting on a participant plural object (12, 13, or 2PL), we get -ugsi in the first slot. Finally, note that in the 3rd person object cases, we don’t always see the -a. It will show up for us in negated forms though.

Next step: what goes in slot 2? Well, if no participant plural morphemes are involved, life is easy: just agree with the subject––note that the subject agreement forms (-n and -t) look familiar from the VAI paradigm. This means that a form like nemi’lin “You see me” can be straightforwardly broken down into nemi-i’li-n: see-1obj-2subj. Also as in the VAI paradigm, there is no clear subject agreement marker; rather, a 1st person subject is indicated by the absence of marking, as in the form nem-ul “I see you”, where we only have agreement with the second person object.

If plural participants (again, 12, 13, or 2PL) are involved as either the subject or the object, they will occupy the second slot. If a first person plural is involved anywhere, the second slot will agree with it (-gw for 12, -eg for 13). If neither of these is involved, agree with second plural, -oq. This means that sometimes the subject will not get to realize any of its features. Take for example 1>2PL, “I saw you”: nem-uln-oq. First we have the stem, then we have second person agreement with the object (-uln), then we have second person plural agreement again with the object (-oq). First person is nowhere to be found. (Why isn’t this form ambiguous? Remember, if a 3rd person subject is acting on a local plural object, you get a special object form -ugsi in slot one.)

Finally, note that almost all forms involving 3PL––whether in the subject  or object––end in -ig, and this appears outside of the second slot. (Remember that [t] turns into [j] before an [i], accounting for alternations like -it/-ijig.) Interestingly, this morpheme also comes outside of tense when tense is involved. Negation also intervenes between the 1st and 2nd agreement slots, making it clear that these really are separable. A full template, which needs more work, looks like this:

Verb – Slot1 – (Neg) – Slot2 – (Past) – (3PL)

Lots remains to be worked out! In some cases some unclear sound changes take place, but again, stay tuned for negative forms to see the pattern described above a bit more clearly. That said, everything here is subject to revision––please don’t hesitate to post comments, questions, or suggestions.

Some particular questions:

  • What is -ugsi? It seems like it can also be used in VAI forms, and is listed in Pacifique as a passive morpheme in some contexts. Is it possible that the -ugsi forms are really passives?
  • Some speakers dislike -ig endings for 3 plural subjects when a plural object is also involved––is there a pattern here? Do some sound worse than others with -ig?

5 thoughts on “Pacifique’s conjugation 6: “Transitive animate”

  1. It’s quite striking how different this overall paradigm is from what we find in Pacifique, DeBlois, and Dawe-Sheppard. In those systems, basically we find the object morphology (-i [1], -ul [2], -a [DIR], -‘(gw) [INV])—more or less consistently evident in combination with the familiar non-TA endings (-(an) ‘1sg’, -(‘)n ‘2sg, -t/-g ‘ANsg’, -eg ‘1pl’, -‘gw ’12’, -oq ‘2pl’, -‘ti-t ‘ANpl’, -‘li-t ‘ANobv’, -m’g ‘IDF’). From here there are a few contextual allomorphic adjustments (some historical-phonological, like -ul-t/g realizing as -‘sg; others more, well, morphological, i.e. the -nam’t element of 1pl in the context of ANsg/pl). Beyond this, the choice of which non-TA endings is used brings in all the issues Jess discussed above.

    The system as we see it in this paradigm appears to have two main innovations.

    First is the replacement of the old -i and -ul -based forms (-i-nam’t, -ul-‘gw, -ul-oq) for the AN>1pl/12/2pl with -ugsi -based forms. This is in fact a recapitulation of the Inverse system for AN>1/2 seen in the Independent paradigm of other Algonquian languages (and still extant in the Subordinative in Mi’gmaq). Something similar occurs in dialects of Ojibwe (Valentine 1990:295), and in Wampanoag (Goddard and Bragdon 1988b:556), though there the -ugsi equivalent is also used with 1s and 2s objects as well.

    In origin, the -ugsi is clearly from PEA *-əkʷ-əsi ‘INV-RFLX’, i.e. the Inverse plus the (AN) Reflexive. This collocation regularly forms lexical stems off of TAs, chiefly TA verbs of perception. These are found in Mi’gmaq as well: -am-ugsi ‘AN appear/look/seem…’, from TA -am-(a) ‘look at AN’ (e.g. milamugsit ‘AN is varied in appearance, comes in lots of forms/looks various ways); welgnnuksin ‘you are very huggable’ (Metallic, Cyr, & Sévigny 2005:354) from TA g’nn- ‘hold AN (2sg’. Notice too that the other main historical uses of -ul, i.e. in 3sg/pl>12, 3sg/pl>2pl are now out of the picture due to the -ugsi innovation. (Though negative concord forms may offer still further examples.)

    Still missing from this chart are obviative forms, i.e. ANprox>ANobv and vice versa (along with plurals). Also needed are Indefinite/Impersonal Agent forms, and inanimate Agent forms.

    And then from there, well, the negative forms of all of these. Which should mostly be these with -(u)(w) interpolated immediately after the -i, -ul, -a, -‘(gw), -ugsi elements, again, secondary morphophonological effects notwithstanding.

    The -ig issue is probably due to a difference in how intransitive ANpl subjects are realized vs. transitive ones. The intransitive system realizes ANpl subjecthood by simply pluralizing the ANsg forms with -ig. The transitive system instead brings in the -‘ti-t element, which is the ANsg -t with a *preceding* pluralizing element -‘ti (related historically to the Reciprocal -‘ti, which in turn also derives the Extended Plurals -ulti, -‘ti, -:C’ti). Finally, the -ig itself is generally a proximate ANpl; -i is the corresponding obviative ANpl, so we may expect -i instead of -ig when there are structurally/thematically distinct ANs floating around, as is often the case in transitive configurations. There is some complexity to that, though, in that nominal morphology tends to realize obviative ANpl the same way as proximate ANpl—i.e. as -(‘)g—which then makes it possible, I would imagine, to either keep the -ig vs. -i prox/obv distinction on verbal morphology, or lump them both together as -ig. Even then, though, we would see this -ig as the ANpl obviative object, not the subject, in Direct transitives (and vice versa in Inverse).

  2. Oh, and the -ugg form is interesting: direct sound correspondence would predict a simple -‘gw, and this is what Pacifique reports. Perhaps it’s a back-formation from the negative?

  3. For an example of the -i-nam’t dialect type, see wegwayuw-i-nam’j-ig () ‘those who were/are angry with us (1pl)’ at

    Note that the immediately preceding form, teli-apiksiktaqajig ‘how we forgive them’, is a nice example of several bits of morphophonology: the TA Final -uw combining with the 1pl>AN element -‘g’t as -a-g’t (i.e. -uw-‘…- > -a…), with that then backing -g to -q, then the -a- vowel harmonizing over the continuant -q- onto the following /’/ vowel, giving -aqat-. (The same treatment of TA -uw- occurs with -‘g ‘1s>AN’ and -‘t ‘2s>AN’, i.e. -a-g, -a-t.) This -a- looks sneakily like the Direct -a-, but it’s not quite that, even though it is in a Direct context: in fact, in the corresponding negatives, you will see the Direct -a-…but now preceded by the -uw- element in its original form. Since the -uw- now no longer directly contacts the initial vowel of the -‘g, -‘g’t, -‘t elements. This is part of a broader series of treatments of /-‘G’-/ sequences, i.e. schwa-glide-schwa sequences, that result in single vowels, often /-a-/ or something related thereto.

  4. Sorry, somehow deleted the original forms:

    wegwayuw-i-nam’j-ig () ‘those who were/are angry with us (1pl)’

    teli-apiksikta-qaj-ig ‘how we forgive them’

    These actually both have TAs in final -uw, but the first one realizes as -uw before -i ‘1sg/pl object’, while the other realizes as -a in combination with -‘g’t ‘1pl>AN’.

  5. Hi Conor––this wasn’t clear to me in your first comment: is it possible that what I’ve glossed as first person -i’li is really two morphemes?

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>