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):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?Active Passive taman tiaman kelhan kielhan qiman qoiman duligan diuligan
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 passiveThe 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-objectWhere 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 -nThese 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".
Copyright © 1997, Peter Clark,
Last updated: July 14, 1997