Theiling Online    Sitemap    Conlang Mailing List HQ   

Diving In...

From:Doug Barr <dbarr@...>
Date:Tuesday, October 30, 2001, 6:45
Hello, all!

I just joined the list a week or so ago - I've been lurking to see what's what,
and I'm very happy to have found a group of kindred linguistically-minded
maniacs. I feel quite at home. <smile>

I wanted to jump in and give my $0.02 on a couple of topics.

The distinction between "active" and "middle/medio-passive" voice - at least in
the Indo-European language sphere, with which I'm most familiar - turns on the
object of the action. In Proto-Indo-European, and its immediate descendants -
Ancient Greek, Sanskrit, etc. - the active voice was used when the action was
performed by one entity on another, i.e. where the subject and object of the
verb are not the same. For example, "I [subject] wash the dog [object]." The
middle or medio-passive voice is used when the subject is also the object or
the beneficiary of the action. This can come out in translation as a reflexive
- "the dog washes itself," i.e. the dog is both subject and object of the verb
"wash" - or a passive, "the dog was washed."

Another example is the verb yaj- "sacrifice" in Vedic Sanskrit. Using the third person
singular "he/she" in the present tense, the active-voice yajati "he/she
sacrifices" is used of the priest who actually performs the actions of
sacrifice in Vedic ritual. The person for whom the ritual is done - and who
paid for it to be done but whose role in the ritual is limited to being present
in the ritual space - is referred to by the middle-voice yajate "he/she
sacrifices (i.e. has a sacrifice performed for his/her benefit."

True passives didn't really exist in Proto-Indo-European or its immediate descendants.

Polysynthetic vs. agglutinating languages:

Both types of languages tend to make long words; if I recall correctly, the
difference between the two types is based not so much on phonetics as on
freedom of compounding.

In the agglutinating languages I have some familiarity with, Turkish and
Japanese, the suffixes are added on in a definite order and their number is
usually limited, at least towards the end of the word. In Turkish, for example,
I can say evsizliklerinden "from their homelessness." This breaks down as:

  a.. Ev - "house" 
  b.. -siz- - "without" 
  c.. -lik- - "-ness," "-hood" (makes abstract nouns or sometimes nouns of place)

These suffixes have to be added on in the order given - you can't say *evliksiz
("without house-ness") - and then come the grammatical suffixes, which have to
be added on in a definite order:

 a.. -leri- - "their" (this can also represent the plural suffix -ler- and the
third person singular possessive "his/her" suffix -i.
 b.. -den- "from" (the -n­- is a phonetic infix that links third person
possessive endings with case endings).

The order for endings is always the following in Turkish:
[plural]-[possessive]-[case]. They don't all occur in every word, of course,
but when they do, they always occur in that order and there can only be one of

There may also be constraints on which suffixes can follow each other: in Japanese,
you can take the basic word kamu "bite" and make a causative kamaseru
"make/cause/incite to bite," a passive kamareru "get bitten," and a passivized
causative kamaserareru "be made/caused/incited to bite," but it is
grammatically impossible to make a causativized passive *kamaresaseru
"make/cause/incite to be bitten." This is not allowed in Japanese grammar, the
causative suffix cannot follow the passive suffix.

In the polysynthetic language I have some familiarity with - the Iqaluit dialect
of Inuktitut (the so-called "Eskimo" language, "Eskimo" is pejorative), the
common image used in teaching materials is of a train: you have an engine (the
root) and a caboose (the grammatical ending), but between those two morphemes
you can infix a theoretically infinite number of "post-bases" or train-cars in
any order that has meaning.

For example:

  a.. Sana- - "make; do; work" - a verb-stem 
  b.. -qati- - "partner, mate" - sanaqati "workmate, colleague" - now a noun-stem 
 c.. -gi- - "have as a ." (refers to relationship or object of emotion) -
sanaqatigi "have [someone] as a colleague, be a colleague to [someone]" - now a
verb-stem again
 d.. -junnaq- [NB. J is pronounced as in German or Dutch, i.e. like the English
y] - "be able to" - sanaqatigijunnaq "be able to have [someone] as a colleague,
be able to work with [someone]" - still a verb-stem
 e.. -juma- [the q of junnaq and the j of juma fuse into r] - "want" -
sanaqatigijunnaruma "want to be able to work with [someone]" - still a
 f.. -juq - third person singular instransitive verb ending "he/she/it does
something" OR a "nominalizer" converting a verb-stem to a noun-stem -
sanaqatigijunnarumajuq- "[act of] wanting to be able to work with [someone]" -
now a noun-stem again: the -juq- in this word cannot be the third person
intransitive ending because -gi- needs an object, therefore a transitive verb
ending; it is the nominalizer required by the following suffix)
 g.. ­-mmari- - "true; real" (adjectival post-base added to noun-stems) -
sanaqatigijunnarumajummari "true [act of] wanting to be able to have [someone]
as a colleague" - still a noun-stem
 h.. -aluk- - "big" (adjectival post-base added to noun-stems) -
sanaqatigijunnarumajummarialuk "definite ("big true") [act of] wanting to be
able to have [someone] as a colleague" - still a noun-stem
 i.. -u- "to be" or a "verbalizing" suffix [deletes the final consonant of the
preceding root] - sanaqatigijunnarumajummarialuu "really truly want to be able
to work with [someone]" - now a verb-stem again
 j.. -lauq - (past tense) [deletes the final consonant of the preceding root] -
sanaqatigijunnarumajummarialuulauq "really truly wanted to work with [someone]"
- still a verb-stem
 k.. -nngit- "not" [NB. Nng represents a double ng sound as in "singing"]
[deletes the final consonant of the preceding root] -
sanaqatigijunnarumajummarialuulaunngit "didn't really all-that-much want to be
able to work with [someone]" - still a verb-stem
 l.. -taa - transitive verb suffix, third person singular subject to third
person singular object "he/she [does something to] him/her" -
sanaqatigijunnarumajummarialuulaunggittaa - an actual verb "He/she didn't
particularly want to be able to work with him/her."
Notice that this word has changed back and forth between being a verb and a noun;
this is perfectly fine and common, and can happen any number of times within a

Also, the post-bases could have been added on in a different order:

 a.. Sanaqatigijumajunnarummarialuulaunngittaa "He/she wasn't particularly able
to want to work with him/her." Here "ability" and "wanting" are reversed.
 b.. Sanaqatigijunnarumalaunngittu*mmarialuujaa* "He/she really truly didn't
want to be able. * = jaa and juq become taa and tuq after consonants. Here
"really truly" modifies "not" rather than "want."
 c.. Sanaqatigijunnaqtummarialuujuma. ". didn't particularly want 'really to be
able' ." Here "really truly" modifies "ability."
 d.. Sanaqatigijunnanngittumajummarialuu. ". really truly wanted not to be
able." Here "not" modifies "ability" (I am not 100% of the form -nngittuma-
"want not [to do something]," the phonetic behaviour of -juma- is very
And so on - any order that has meaning is acceptable. Some post-bases have pretty
esoteric or specialized meanings, for example -nnguaq "image or small-scale
representation" - inuk "human (or specifically, 'Inuk person')," inunnguaq
"doll," nuna "land," nunannguaq "map."

Another simpler example:

  a.. Ui- - "husband" - noun-stem 
  b.. -qaq- - "have (a) ." - uiqaq "have a husband; be married to a man" - verb-stem. 
  c.. -tuq/-juq "he/she." third person singular intransitive ending. 

  a.. Uiqaqtuq "She has a husband" 
  b.. Uiqarumajuq "She wants to have a husband." 
  c.. Uiaqarumajunnaqtuq "She wants to be able to have a husband." 
  d.. Uiaqarunnarumajuq "She is able to want to have a husband." 
  e.. Uiqanngittuq "She does not have a husband." 
  f.. Uiqarumanngittuq "She does not want to have a husband." 
  g.. Uiqarunnarumanngittuq "She does not want to be able to have a husband." 
  h.. Uiqarumajunnanngittuq "She is not able to want to have a husband." 
  i.. Uiqanngittumajuq "She wants not to have a husband." 
  j.. Uiqarumanngittumajuq "She wants not to want to have a husband." 
  k.. Uiqarumanngittunnaqtuq "She is able not to want to have a husband."
 And so on more or less ad infinitum.

Inuktitut also has the useful feature of a "null" base pi- used to affix post-bases to.
It has no meaning in and of itself, it just serves to attach post-bases to,
when they aren't included in the main word or are used by themselves as the
relevant feature of a word, especially in answering questions.

  a.. Pialuujuq.  "It is big." 
  b.. Pimmariujuq.  "It is true. " 
  c.. Piqaqtuq.  "He/she has one/some (of whatever is being referred to)." 
 d.. Sanaqatigijunnaqpaa? "Can he/she work with him/her (i.e. can he/she have
him/her as a workmate?)," parsed: sana-qati-gi-junnaq-paa
[work]-[partner]-["have ... as]-[ability]-[third person singular transitive
subject, third person singular direct object, question].
 e.. Aakka, pijunnanngittuq. "No, he/she can't," parsed: pi-junnaq-nngit-tuq
[null "header"]-[ability]-[negative]-[third person singular intransitive,
Not sure if this will help or hinder - hopefully it will be slightly helpful to someone somewhere.



Christophe Grandsire <christophe.grandsire@...>
William Annis <annis@...>
Yoon Ha Lee <yl112@...>