Noun incorporation in verb-based languages


I found this paper on noun incorporation I did for one of my classes a few years back. It might be too technical for some folks’ tastes, but some people here might enjoy it.

The languages used as examples here are polysynthetic languages, which means basically that they form sentences that are often one long complex word, based on the verb, with other elements incorporated into the verb. (Most Native American languages are polysynthetic.) The elements separated by hyphens when the word is broken down are combining elements and cannot act separately as independent words. This paper gives an idea of how that works.

Some abbreviations: 1S = 1st person subject case; 3S = 3rd person subject case; 1OB = 1 person object case; agr= agreement particle; imp = imperfective (ongoing action); I = indicative (statement). (ERG and ABS are a little too involved to explain here.)


Noun incorporation is not unknown in Indo-European languages. In English, we have many Obj+Verb compunds, such as “birdwatching,” “babysitting,” and “fingerprinting.” However, despite the Obj+Verb structure, this is not a case of the verb incorporating the object of the sentence… When used transitively, these verbs take other objects: “I have to babysit the neighbor’s kid” or “I’ll fingerprint the suspect.” When they do not have an external object, they are intransitive: “When I go birdwatching, all I see is mosquitoes” (IOW, “birdwatching” does not mean you necessarily see the “bird” incorporated into the verb).

But languages that use noun incorporation actually incorporate verbs with their noun arguments. Some can incorporate only objects (this seems to be the most common option); some incorporate subjects as well as objects; some can incorporate subjects only with stative verbs; some can incorporate objects or oblique cases; some can incorporate subjects, objects, or oblique cases.

In Mohawk, the incorporated noun can be understood only as the object and never as the subject.

   yaowira'a  yenuhsnuhwe'a    = the baby likes the house
   yao-wir-a'a ye-nuhs-nuhwe?-a
   pref-baby-suf  agr-house-like-imp         (lit: The baby house-likes)

In Tiwa the subject can be incorporated:

    wefanlurmi = snow is not falling

In Chukchee, a direct object or an oblique case can be incorporated:

[direct object:]

 totkoc'yntywatyn = I set a trap
 1S-trap-set-1S        (lit: I trap-set)

[oblique cases:]

 tynnytkerkyn = I smell of fish
 1S-fish-smell-pres.i      (lit: I fish-smell)

 tykeiñy'ejñekyn  =  I am shouting like a bear
 1S- bear- shout- pres.i  (lit: I bear-smell)            

[direct object & oblique case in same sentence:]

gymnan tygytkarkyplyn regokalgun = I hit the fox on the leg
 gymnan t-y-gytka-rkypl-yn reqokalgyn
  I.ERG 1SG-leg-hit-3SUB- fox.ABS              

[adverb as well as noun incorporated by verb:]

  tyjanraykoprentywatyrkyn + I am putting out the net separately
  t-y-  janra-y-   kopre-  ntywat-y- rkyn
  1S -separately- net- set-  pres.i      (lit: I separately-net-set)
 (the -y- element is for euphony)

In Cree, the subject, object, or an oblique case can be incorporated:

siipihkwaskwaa = The sky is blue
siipihk-    -was-    -kwaa                               -w 
 is blue-    -sky-     inanimate intransitive       agr

pahpawastimawiiw = He/she brushes the horse
pahpaw   -astimaw  -ii                              -w 
 brush-      -horse-    animate intransitive-   agr     (lit: he horse-brushes)      

 kohkoosimiiitisaw Otto = Otto eats like a pig
 kohkoosi-  miitisaw                                Otto
  pig-           -eats (animate intransitive)    Otto         (lit: Otto pig-eats)

Note that in “he brushes the horse” the verb is intransitive (he horse-brushes). In most NI languages, when object noun is incorporated, if the noun does not require a possessor, the verb (in spite of the incorporated object) is detransitivized. An example from Mbya (Brazil):

ay’u = I drink water
a- -y- 'u
I- water- -consume

However, if the object requires a possessor, the possessor commonly becomes the direct object, as in this Mbya example:

aipokwa = I tie his hands
a- i- po -kwa
I- him- hand-tie

Polysynthetic languages often have combining forms of basic elements that are different from their independent forms. Example from Mingo (an Iroquoian language still surviving in W VA):

hu’nöhtö'ö = “his arrow is lost”.

hu- “he/his”,
-‘nö- “arrow” (independent form: ka’nö’)
-htö'ö “it is lost” (independent form: uhtö'ö “it’s lost”)

Example from Nahuatl (Aztec):

nixochiteki = I cut flowers
ni- = I
xochi- = flower (independent form: xochitl; -tl is the suffix for independent nouns generally)
teki = cut

Kutenai is another example in which most nouns have combining stems and independent forms, the independent form is formed by an affix (usually the prefix a?k- or a?q).

hun’aslhamaa’lhne: = I have two blankets
hu- = I
n’as = two
-lhamaa’lh = blanket (independent form = probably either a’klhamaa’lh or a’qlhamaalh)
-ne: = indicative

Among languages of the Amazon, if there is noun incorporation, typically only those nouns which are obligatorily possessed (like relatives and body parts) can be incorporated, and they typically precede the verb root. Example from Wai Wai (Carib family):

oyewkrashi = it stuck into my eye
oy- ew-krashi
1P.OBJ - eye.stick.into - imm. past (lit: me eye-stuck-into)

Noun-incorporating languages divide on the question of whether modifiers can be “stranded,” i.e. whether an external adjective, numeral or whatever can modify an incorporated noun. In some (Eskimo languages, Mohawk, Tiwa) they can be while in others (Chukchee) they cannot.

Noun incorporation is rarely (if ever) obligatory, and almost all my sources included alternative non-incorporating ways of forming sentences, usually stressing that the choice to incorporate or not affected emphasis or shades of meaning. One source suggested that in Mohawk incorporation is used less in informal contexts and more in ritualistic contexts, and that Mohawk speakers use noun-incorporation much more when they want to be poetic.

Here are examples from the Nadeb language of Brazil of how different options in NI create difference in emphasis. (This uses just the schematic patterns, somewhat simplified.)

Subih house I make = I am making Subih’s house
Subih I house-make = I am house-making Subih (the benefit for Subih is emphasized)

Here a progression of zero to three incorporated nouns with accompanying shift on emphasis: (Incorporated nouns are underlined)

grandmother house mouth uncle closed = Uncle closed the door of grandmother’s house
grandmother house uncle mouth-closed (effect on door emphasized)
grandmother uncle house-mouth-closed (effect on house emphasized)
uncle grandmother-house-mouth closed (effect on grandmother emphasized)

Noun incorporation tends to have a fixed or at least a preferred order with respect to the verb and other elements. For example, if you take the person/number verb ending as representing the subject, Eskimo (Inuktitut) word elements almost always have an OVS order:

nergiksaqariaqarqet? = do you need food?

nerqiksa- qaria- qarq- et?
food- need- you- question

In fact, in virtually every noun-incorporating language I found, the incorporated noun, whether subject, object, or oblique case, preceded the verb. This is an area I think could be researched more: the function of morpheme order in noun incorporating languages and agglutinative/ polysynthetic languages overall. Although synthetic languages tend to have non-fixed word order, many if not most have a prescribed morpheme order within words. So the way that syntactic principles seem to be expressed through morpheme order deserves further research.

Anderson, J.O, Rules of the Aztec Language (University of Utah 1973)
Boas, Franz, Kutenai Tales (Bureau of American Ethnography 1918)
Dixon, R.M.W & Aikhenvald, Alexandra Y., The Amazonian Languages (Cambridge 1999)
Mithun, Marianne, The Languages of Native North America (Cambridge 1999)
Peacock, FW, Conversational Eskimo (Breakwater Books 1977)
Spencer, Andrew, Morphological Theory (Blackwell 1991)
Summer Institute of Linguistics (


Intriguing. Has there been more research on this subject since, as far as you know?

Another aspect that I noticed was the example on “closing the door” - after all, here an interesting shift apparently took place: it seems that it was not the door that got closed, but the dooropening (mouth), making me wonder if more such examples can help shed light on another way of thinking.