Passives

Passives

adapted from a March 30, 1997 post by Paul M. Hoffman

Matt Pearson wrote:

Sort of. The conlang I worked on before Tokana, called Kosan, used stem modification to mark voice (active versus passive) - specifically, modification of the first vowel of the stem. [...] In the original drafts, I had a rather elaborate system for converting each of the six vowels into a different vowel or diphthong (e.g. "a" becomes "au", "e" becomes "oi", etc.). But I ran into the same problem that you describe above, so I decided to give it up. Instead, I chose to mark the passive by inserting an "i"-glide before the first vowel of the stem (unless that vowel was itself "i", in which case it became "oi"). Thus, with made-up examples (since I can't remember many actual verbs):
        Active          Passive

        taman           tiaman
        kelhan          kielhan
        qiman           qoiman
        duligan         diuligan

This is very elegant, and doesn't seem at all unrealistic -- I can see how this might easily arise, for example, from a passive infix -ih- that lost the h and dissimilated q-i-iman to q-o-iman. Did Kosan have any defective verbs, or verbs with suppletive passive forms?

I'm fascinated by how many different kinds of passive constructions there are in natural languages, and by the great variety of ways in which (more generally) changes in valency and back-/foregrounding are accomplished. In my language SeMpaiton, there is a small assortment of valency-altering suffixes, each of which displaces the subject or direct object (or both):

  suffix   kind of like...  displaces
  -------  ---------------  ----------------------------------------
   -ri     reflexive        subject of a transitive verb (when
                              coreferential with the object)
   -na     passive          subject of a transitive verb (when
                              *not* coreferential with the object)
   -the    detransitive?    direct object (must not be coreferential
                              with the subject)
   -tsuvu  impersonal       subject and direct object of a
             passive??        transitive verb
   -mua    impersonal       subject of an intransitive verb
             passive

The first four suffixes are used only with transitive verbs. (All verbs in SeMpaiton are lexically either transitive or intransitive; there aren't any that can be both, except some that have significant, and idiosyncratic, semantic differences when used in the two different roles). I don't recall the exact forms of the last two suffixes, so I'm pretending they're -tsuvu and -mua. Note that pronouns in SeMpaiton are ergative/absolutive and nouns are (essentially) nominative/accusative, but that these suffixes follow neither an erg/abs nor a nom/acc pattern. (At least none that I can see.)

BTW, I use the term "displace" loosely -- even though pronominal subjects and direct objects are morphologically part of the verb in SeMpaiton, these suffixes don't actually take the place of a deleted pronoun. Instead, they fill the slot preceding that for pronouns. The shape of a SeMpaiton verb is basically this:

  stem  aux*  aspect?  tense?  valency-alterer?  subject-and-object

Where aux = internal auxiliary, * = zero or more, ? = zero or one. The slot for subject and object might actually be considered two slots, but I lean toward treating it as a single slot for reasons which I won't get into here. Here are some illustrative examples:

  1a. Doloke na Tsekon.
      do   -(n/l)o -ke       na                Tsekon
      kill -PA.DEF -2SG.ERG  DEF.ART.PROP.ACC  Tsekon
      You killed Tsekon.

      [don 'kill' is used only when the subject and direct object are
        both animate]

  1b. Dolonan la Tsekon.
      do   -(n/l)o -na -n  la                Tsekon
      kill -PA.DEF -NA -N  DEF.ART.PROP.ACC  Tsekon
      Tsekon was killed. / Someone killed Ts. / They killed Ts.

      [na -> la is conditioned by the preceding nasal]
      [-n is an obligatory suffix used when all core participants
        (subject, direct object) are nouns (or absent) rather than
        pronouns; it fills the slot for pronouns]

  1c. Dolorin la Tsekon.
      do   -(n/l)o -ri -n  la  Tsekon
      kill -PA.DEF -RI -N      Tsekon
      Tsekon killed himself.

  1d. Dolorinte.
      do   -(n/l)o -ri -nte
      kill -PA.DEF -RI -2SG.ABS
      You killed yourself.

  1e. Dolotheke.
      do   -(n/l)o -the -ke
      kill -PA.DEF -THE -2SG.ERG
      You killed (someone).

  1f. Dolotsuvun.
      do   -(n/l)o -tsuvu -n
      kill -PA.DEF -TSUVU -N
      There was a killing/murder. / Someone killed someone.

  2a. Tirasan ti tiure.
      tira -sa   -n  ti                tiure
      fall -IMPF -N  DEF.ART.INAN.NOM  snow
      The snow is falling.

  2b. Tirasango.
      tira -sa   -ngo
      fall -IMPF -3SG.INAN
      It's falling.

  2c. Tirasamuan.
      tira -sa   -mua -n
      fall -IMPF -MUA -N
      Something is falling.

Sentence 1e looks funny, having an ergative pronoun and no direct object in sight, but that's how it is.

Now the thing I *really* like about all this is how it can be used with attributive verb forms.

An attributive verb in SeMpaiton has essentially the same structure as a predicative verb, except that it cannot have marked forms for tense or pronouns:

  stem  aux*    tense    aspect?  valency-alterer?  subject-and-object

                ^^^^^                               ^^^^^^^^^^^^^^^^^^

             always zero                                always -n

These are used either adjectivally (SeMpaiton has no real adjective class) or substantively. Here are some examples:

  3a. Korasan ta Muilise ni ndudine.
      kora  -sa   -n  ta  Muilise  ni                ndudine
      build -IMPF -N      Muilise  DEF.ART.INAN.ACC  hostel
      Muilise is building the hostel.

  3b. Ikhisankave dai dzo korasathen ta Muilise.
      ikhi-sa  -nka     -ve      dai dzo  kora -sa  -the-n ta Muilise
      know-IMPF-PA.INDEF-1SG.ERG not COMP build-IMPF-THE-N    Muilise
      I didn't know that Muilise built (things).

      [ikhin 'know' is used only in negative constructions; the tense
      in the subordinate clause is relative to that in the matrix
      clause]

  3c. Athiove nu korasathende.
      athe -o      -ve       nu                korasathen -nde
      see  -PA.DEF -1SG.ERG  DEF.ART.ANIM.ACC  builder    -PL.ANIM
      I saw the builders.

  3d. Athiove ni koranake.
      ...  ni   koranan  -ke
      ...  the  building -PL.INAN
      I saw the buildings.

  3e. Athiove ni korasanake.
      ...  ni   koranasan  -ke
      ...  the  building   -PL.INAN
      I saw the buildings [i.e., as they were being built].

Actually, the imperfective aspect is usually omitted when forming "agentive" nouns, so _korathen_ (one who has built things) is normally used instead of _korasathen_ (one who habitually builds things). To tell you the truth, I'm a little uncertain right now about aspect in SeMpaiton; I don't like having the perfective be the unmarked aspect.

The verb kan 'have' (one of a small number of what I call postclitic verbs, for lack of a better term) has been partially delexicalized in its -na form to take on (roughly) the meanings 'there is/are' and 'a(n)':

  4a. Li suen kave.
      li                suen  ka   -ve
      DEF.ART.INAN.ACC  book  have -1SG.ERG
      I have the book.

  4b. Li suen kanan.
      li   suen  ka   -na -n
      the  book  have -NA -N
      They have the book. / It's the book. / There's the book.

      [This is hard to translate well into English; in French and German
      it might be 'On a le livre' (which to me sounds too much like *we*
      have it) and 'Man hat den Buch' (which sounds strange to me).]

  4c. Li suenge kanan.
      li   suen  -nge      ka   -na -n
      the  book  -1SG.ABS  have -NA -N
      They've got my book. / It's my book. / *I have my book.

  4d. Li suenge karin.  [?? Equivalently, Li suen karinge. ??]
      li   suen  -nge      ka   -ri -n
      the  book  -1SG.ABS  have -RI -N
      I have my book.

  4e. Ba suen kanan li abosa.
      ba     suen  ka   -na -n  li  abo   -sa
      FOCUS  book  have -NA -N  the table -ADESSIVE
      There's a book on the table.

  4f. Athiove suen kanan.
      athe -o      -ve       suen  ka   -na -n
      see  -PA.DEF -1SG.ERG  book  have -NA -N
      I saw a book.

The contrast between 4a. and 4d. is pregnant with significance, but I'd better not elaborate here as this message is already too long. There's more -- at this point somewhat contradictory -- information at http://www-personal.umich.edu/~nkuitse/conlang/sptn/ under the heading "Quasi-coreference".

Return to Conlang-related topics|Back to FAQ page