Noun Incorporation
Mithun (1984) described four types of noun incorporation that are observable in various languages around the world. In Type I there is nothing more than the lexical compounding of nouns with verbs. In Type II this compounding is accompanied by the promotion of an otherwise oblique argument to the status of a core argument. In Type III the incorporation is used in discourse as a way of backgrounding less salient nominal referents; one can speak of pragmatic incorporation. In Type IV incorporated nouns are used as semantic classifiers; there is classificatory incorporation.
Mithun hypothesized a path of historical development whereby each of these four types might evolve into the next type, although a language might also stop at any stage. This path would imply, for example, that a language with classificatory incorporation would also exhibit pragmatic incorporation. With Mithun’s concurrence, however, I would like now to suggest that it is possible for a language to develop classificatory incorporation without having first developed pragmatic incorporation, and that Cherokee is just such a language. If that is correct, we can interpret the pragmatic use of incorporation in the Northern Iroquoian languages as a florescence of the more limited type illustrated by Cherokee, which may lie closer to Proto-Iroquoian in this respect.
Classificatory incorporation in Cherokee is limited to a small number of semantic classes that involve shape and substance. It can be illustrated with examples like the following from North Carolina Cherokee (Cook 1979:186):
Otherwise, incorporation in Cherokee is limited to body parts (Cook 1979:133, 134,136):
There seems to be no good evidence that pragmatic incorporation ever existed in Cherokee, so it is plausible to suppose that Cherokee (and probably Proto Iroquoian) reflects a language type in which incorporation is limited to classificatory nouns, body parts, and clothing.
Mithun’s stages III and IV may thus be reversible, with the Cherokee classificatory type (IV) evolving into the Northern Iroquoian pragmatic type (III). Incorporation is much more widespread in the Northern Iroquoian languages. I will draw on a particular text for examples, an autobiographical account by a Seneca man. In this account there were 240 verbs, and 75 of them, or 31%, contained incorporated nouns. I suspect that that is a typical proportion, though more genres and more speakers obviously need to be examined.
When one first encounters this widespread use of incorporation there is a natural tendency to think that any of the nouns that occur in some particular core syntactic or semantic role — say, any patient nouns — can be incorporated with any verb roots whose argument structure accepts that role. One quickly finds, however, that some nouns never incorporate and some (transitive) verbs never allow incorporation. But even with elements that do sometimes take part in incorporation, it is by no means an unconstrained process. It is instructive to look in some detail at examples of noun incorporation in Seneca, where pragmatic incorporation seems to have reached its zenith, as in the other Northern Iroquoian languages. I will take these examples from the autobiographical account just mentioned.
The term “pragmatic incorporation” implies that we need to consider the discourse function of this device which, to quote Mithun, is “used to background known or incidental information in discourse” (Mithun 1984:867) If, for example, a salient referent is being introduced for the first time, the noun root that refers to it is not incorporated. For example, this speaker told how his family would sometimes go to town to buy shoes:
but he did not. Expressing the idea of the shoes as a separate word (as well as placing it at the beginning of the clause) makes clear its function as newsworthy information, whereas its incorporation within the verb, as in (5), would have downplayed the shoes by saying that his family went to town for the more diffuse activity of ‘shoe-getting’.
What kinds of things are incorporated in a Northern Iroquoian language? Like Cherokee, Seneca does exhibit classificatory noun incorporation. But whereas in Cherokee such classification is limited to a small number of shapes and substances, the possibilities in Seneca are a great deal more diverse. They include, for example, categories like ‘domestic animal’. In telling about foods that were available to his family when he was a child, the same speaker said:
Not only is there greater diversity in the classifications that can be made, but the incorporated nouns can occur as separate nouns in their own right, in this case as (with the necessary prefix and suffix):
In Cherokee the classifiers do not occur independently in this way.
Another example of the flexibility of classificatory incorporation in Seneca can be seen in something the speaker said close to the beginning of his account, when he was placing himself in the social world:
Both of the analyzed words contain the verb root-o‘de˛- ‘be a certain way’, with which are incorporated-ya‘d- ‘physical self’ and-‘se- ‘clan’. These two noun roots express larger semantic categories: ‘being Seneca’ is an instance of one’s physical identity, and ‘belonging to the turtle clan’ is of course an instance of clan membership.
Like Cherokee, Seneca also exhibits the incorporation of body parts:
This word is actually the way one says ‘I put on my shoes’, showing a nice difference in the way one conceptualizes the same physical event in Seneca and English. Putting on one’s shoes is of course a familiar, even automated activity, and the way of expressing it illustrated in (9) is a highly routinized combination of noun root with verb root (as well as middle voice). It thus exemplifies a high degree of lexicalization. As we look at further examples of noun incorporation we will see reasons to regard them, too, as lexicalized combinations, but combinations more diverse than those found in Cherokee. As Mithun has emphasized, pragmatic incorporation “is still a lexical process; if speakers use a new combination, they are creating a new word and are aware of that fact. The productivity of the process is governed both by lexical accident and by pragmatic considerations” (Mithun 1984:863).
One commonly encountered type of noun incorporation is found in proper names. We already saw one example in (8), the Seneca word for ‘Seneca’, which means something like ‘pertaining to the great hill’. It has the following structure:
Obviously a proper name is an established way of expressing a familiar idea, in this case the identity of the people themselves. The frozen status of this particular name may have removed any sense of its original composition, as evidenced by two alternative ways in which it is sometimes pronounced. More often than not one hears the truncated form n˛odowá‘ga·‘ without the initial pronominal prefix o-. I have also occasionally heard n˛odawá‘ga·‘ with a instead of o in the second syllable, where the etymology of the word — specifically the presence of the verb root meaning ‘great’ — is no longer recognizable.
At another point early in his account the same speaker said:
In the first word of (11) the name of Newtown, a settlement on the Cattaraugus Reservation, is a highly lexicalized morphological formation expressing an extremely familiar idea, the place where the speaker lives. But (11) also illustrates, in the last word, a lexicalized combination of a different kind: no longer a proper name but an idiom whose meaning has taken on a life of its own, a meaning that cannot be predicted in any systematic way from its literal meaning. The literal meaning ‘where our fire is laid’ has acquired the idiomatic meaning ‘where our home is’. This idiom would appear to go back to precontact living where each family within a longhouse was centered around a fire.
An example of a similar sort is:
The experience of getting through winters provided a handy way of counting years in the days before European calendars.
A third type of noun incorporation that exhibits lexicalization can be illustrated with the following passage. Speaking of a doctor who made house calls, the speaker said:
To understand the last word in this example it is necessary to know that with certain kinds of referents involving large-scale environmental phenomena like hills, lakes, rivers, the earth, the sky, towns, and houses, Seneca requires that the noun roots expressing them must be incorporated with what may be called a positional verb root such as-de- ‘be in place’,-ye˛- ‘be laid out’, and-ot- ‘be upright’. Thus, a hill is on˛odade‘, a town is gan˛odaye˛‘, and a house is gan˛óhso·t:
The word at the end of (13) is the word for ‘house’ with the so-called cislocative prefix t- added, yielding a meaning more like ‘where the house was’.
(It would appear from the first two examples that the Seneca noun roots meaning ‘hill’ and ‘house’ are identical. But that is a historical accident resulting from the fact that in Seneca a became ˛o after n. Thus an earlier-nada- meaning ‘town’ (which appears in the name Canada) became Seneca-n˛oda-, falling together with the noun root meaning ‘hill’.)
Verb roots like the-de-,-ye˛-, and-ot- illustrated in (14) occur very frequently with incorporated noun roots, and not only in expressions of environmental phenomena. In fact we saw two examples of the-ye˛- verb root earlier. In (11) it was the fire that was ‘laid out’ in the word niy˛ogwatshísdaye˛‘, and in (6) the same verb was used in the word ˛ogwán˛osgwa·ye˛‘ with an animate patient pronominal prefix and an extended meaning that is translatable as ‘have’ (in this case a domestic animal).
These positionals are thus favorite verb roots for incorporating. Seneca has certain noun roots that are also favorite candidates for this construction. Probably the greatest favorite of all is a noun root whose earlier form was *-rihwa-, with a meaning reminiscent of Latin res, which I learned in my high school days to translate as ‘thing, affair, matter’. In Chafe (1967) I glossed this Seneca root as ‘thing, matter, affair, cause, subject, message, news, word, speech, conversation, story, tradition, custom, ceremony, ritual, assignment’, and in fact each of these translations (and more) is appropriate in one context or another. One example of its use in this speaker’s account is:
It seems appropriate in (15) to translate this noun root as ‘what was going on’. Another example of its use appears in:
Here the same noun root refers to the mother’s household chores.
This speaker discussed at some length the way his family ate when he was a child; because of that, the incorporated noun root-khw(a)- ‘food’ occurred a number of times. In an expression parallel to (6) we find it with the verb root-ye˛-, here again with the meaning ‘have’:
Another verb root that incorporates very frequently is-˛oni- ‘make’, and it appears with the ‘food’ noun root in a combination with the idiomatic meaning ‘cook’, literally ‘make food’:
When the middle voice marker-ade- is added to the construction in (18) the resulting meaning is ‘eat’. In (19) we find that construction, but also another word in which the ‘food’ noun root is incorporated with a verb root meaning ‘be hot’:
The following sentence shows not only the ‘hot food’ word, but also lexicalized expressions for ‘fry-bread’ and ‘fried potatoes’, so that the example contains three examples of incorporation:
These examples show the extent to which noun incorporation in Seneca is involved in lexicalized combinations: familiar proper names like Seneca and Newtown; familiar states like having a home and being a certain age; familiar environmental phenomena like hills, towns, and houses; familiar activities like cooking and eating; and familiar foods like fry-bread and fried potatoes. Can one, nevertheless, make up new incorporated constructions that do not have this lexicalized status, combinations that have never, perhaps, been said or heard before?
There are a few situations where such new formations do seem possible. One involves counting things. The verb root-æge- can be glossed ‘how many’, and it is used in constructions like the following:
Counting years is a familiar concept, but there seems to be no restriction on the use of incorporation for this purpose so long as the noun root is one that can be incorporated in the first place.
The following shows the combination of the noun root meaning ‘car’ with the same verb root meaning ‘how many’, this time in the form-age-:
When I visit the Allegany Reservation, I rent a car in Buffalo and the physical properties of the car can vary from visit to visit. That provides occasions to create words like:
I do not know whether and to what extent these are lexicalized combinations or combinations made up on the spot. That kind of knowledge is available only to a fluent native speaker of Seneca. I do know, however, that I have sometimes made-up noun-verb combinations that have elicited a reaction of a sort that suggests I have created a neologism. During meals it is often appropriate to say that something tastes good, and it is possible to incorporate a noun that will specify the kind of food that tastes good:
Although the first of these words is often used and is certainly lexicalized, the second is something I made up one time when I was eating in a restaurant. My companion reacted in a way that showed she had paid more attention to the way I had said this than to its content, remarking, “That was a cute thing to say.”
But obviously if there has been a florescence of noun incorporation and if the ideas of new cultural items like cars can be incorporated, new formations must be possible. How do they come about? I believe that incorporation is active enough in the language that speakers know how to create new instances, but that when they do so they know that they are creating something new, and their listeners know that too. In fact people often remember when they heard a particular example of incorporation for the first time, as well as the context in which they heard it.
I found myself recently using in English the word man-madeness. I imagine that any sensitive native speaker of the English language would recognize that it was a word I made up. And so it is with noun incorporation in Seneca. The device is available for people to create neologisms, but they are recognized as such. As time goes by they may or may not become established in the language and be used in a variety of contexts. That will depend on how well they capture ideas that need to be expressed, but it may also depend on the status of their creator. I do not have the status to create new instances of noun incorporation in Seneca, but real speakers of the language, and particularly those whose style of speaking is admired, may have more success in that regard.
Mithun (1984:872) summarizes this issue of productivity in the following way:
To a casual observer of languages like Mohawk, [noun incorporation] may appear to be a syntactic process simply because of its tremendous productivity; the number of[noun-verb] combinations that occur seems unlimited. However, speakers are keenly aware of the lexical status of all such combinations. They know not only which constructions are possible, but also which of these actually exist — i.e. which are lexicalized. They immediately recognize those that are not. Speakers remember who uses a word not used by others, even when it is a perfectly transparent combination of two highly productive stems. A Mohawk speaker’s lexicon can be enormous, because of the high productivity of word formation processes like [noun incorporation]; but it is well-defined.
Is it possible to pinpoint some feature of Northern Iroquoian culture that might be responsible for the florescence of noun incorporation? I would suggest that what lies behind the flowering of this device is the high degree of prestige attached to skill in oratory. Writing in the middle of the nineteenth century, with the biases of that period, Lewis Henry Morgan, in his pioneering ethnography of the Iroquois, said the following:
The Indian has a quick and enthusiastic appreciation of eloquence. Highly impulsive in his nature, and with passions untaught of restraint, he is strongly susceptible of its influence. By the cultivation and exercise of this capacity, was opened the pathway to distinction; and the chief or warrior gifted with its magical power could elevate himself as rapidly, as he who gained renown upon the warpath. (Morgan 1851:107)
The skill involved in creating new instances of noun incorporation may be considerable, especially in a highly fusional language like Seneca, and it is no wonder that those possessed of this skill to a high degree have been accorded a special prestige. To quote Mithun again:
[Noun incorporation] can be a difficult process, often requiring special linguistic virtuosity, especially in languages with complex phonological alternations. Speakers who do it well may be especially admired, while marginal speakers may not be able to do it at all. Noun incorporation can be a powerful, but fragile, linguistic resource. (Mithun 1984:891).
