Incorporation in Linguistics Research Paper

Traditionally, noun incorporation is understood as a choice of whether arguments are retained on the syntactic level or not. Argument incorporation refers to the realization of the verb’s arguments exclusively on the morphological level, i.e., verb internally. Here, no arguments can be identified in syntax, and the basic grammatical relations are realized at the word level. Noun phrases and other supplements have only adjunct status. Going a step further, functional items may be incorporated as well. In such cases of grammatical incorporation, various kinds of grammatical relations are realized verb internally. The term ‘incorporation’ will be used for all three types, regardless of whether the involved items are independent lexical units, i.e., words, or bound morphemes. Such a distinction becomes instrumental for differentiating the different types of incorporation.

1. Noun Incorporation

Since Kroeber (1909, p. 569) and Sapir (1989/1911, p. 254), noun incorporation has been characterized as stem combination, i.e., the combination of independent lexical items into a complex verb. The verb stem determines the newly built form categorically, while the incorporated noun loses its categorical features, as well as any grammatical markings such as plural, if such are employed, see Ex. (3b) below. Nouns do not incorporate as arguments, but as specifiers of the incorporating verb.

(1) (a) Mohawk

/wa?kyvtho? ojı:ja?/ ‘I planted a flower’

wa? aorist modal prefix

-k 1st person singular agent

-yvtho? ‘plant’

(b) Mohawk

/wa?kji?jayvtho?/ ‘I flower-planted’

-ji?ja ‘flower’

(Bonvillain 1972, p. 21)

Although the most general features of the construction might at first suggest so, noun incorporation is more than just another variety of compounding. Conspiciously, languages featuring regular and productive compounding, such as German, do not allow for incorporation, and vice versa, incorporating languages do not necessarily allow for compounding.

In many languages, noun incorporation is not a fully productive process, but restricted to certain types of nouns, most often related to the degree of animacy, or it may only be accomplished with suppletive forms. See, e.g., Allen et al. 1984 on Southern Tiwa, Mithun 1984, pp. 863, 876/77.

The incorporating process often entails a loss of referentiality of the incorporated noun and its being backgrounded. From the perspective of the speaker, incorporation represents an option, an alternative way of expression, a strategy to increase the informational content of a proposition, or a focus shift when proceeding with a discourse (cf. Mithun 1984, pp. 869ff. on Mohawk). Thus, the incorporated variant can always be seen in contrast to its nonincorporated counterpart. Being a matter of discourse and style, noun incorporation is expected to be a nonobligatory process. As has been pointed out frequently, it represents an elaborate register and, with bilingual speakers, characteristically is lost with decreasing fluency and competence (see, e.g., Mithun 1984, pp. 879ff.).

It is just this optionality, made possible by the basic lexical independence of the constituents which has motivated the interpretation of incorporation as a syntactic process, as a grammatical function-changing operation (Baker 1988). Incorporation is described as causing the loss of the argument status of the incorporated noun via its vacating its original position in syntax. Such manipulation involves the internal argument (direct object) predominantly.

If at all possible, the incorporation of an external argument typically is restricted to nonagentive noun phrases (Sapir 1989 1911, pp. 252, 266, Mithun 1984, p. 853, Gerdts 1998, p. 87, Cook and Wilhelm 1998).

(2) (a) Chukchee

 ej -k l l t lg -g i

hill-LOC snow.ABS thaw-3SG.S

(b) Chukchee

 ej -k l lg -g i

hill-LOC snow thaw-3SG.S

‘The snow thawed on the hill’

(Spencer 1995; p. 451)

In its weakest form, noun incorporation simply causes detransitivization, as in examples (1a) and (1b). In its most dramatic version, phrasal constituency is dissolved and the constituents of the former noun phrase become distributed inside and outside the complex verb—the head of the noun phrase is incorporated, while its modifier(s) remain in their external position. This phenomenon is often refered to as ‘external modification’ or ‘classifying incorporation.’

(3) (a) Southern Tiwa

Wisi seuanin bi-mu-ban

two man:pl 1sB-see-past

‘I saw two men’

(b) Southern Tiwa

Wisi bi-seuan-mu-ban

two 1s:B-man-see-past

‘I saw two men’

(Allen et al. 1984, p. 295)

2. Argument Incorporation

Baker (1996) extends his syntactic approach of noun incorporation to languages which realize their arguments exclusively verb internally, i.e., which do not exhibit arguments in syntax. Contrary to such an attempt, it must be kept in mind that it is exactly the distinction between syntax and morphology, and the interaction between these two levels, on which the specificity of noun incorporation as compared to other function changing processes, e.g., passives, rests.

The question as to the status of the incorporated noun, i.e., whether it originates as a syntactic argument or not, functions as a crucial and effective criterion for distinguishing noun incorporation from the other types of incorporation. If we consider the implications of the traditional notion ‘grammatical,’ it is appropriate to distinguish noun incorporation from what was initially called ‘pronominal incorporation’ by Sapir. Contrary to what the term might suggest, usually no pronoun proper is incorporated. The morphemes under consideration most often cannot even be reconstructed as historically independent forms, but are strictly bound morphemes, obligatory and paradigmatically ordered (Sapir 1989 1911, pp. 250–2). Following Evans (1999) it can also be assumed that their semantic properties are different from those of true pronouns. The more recent term ‘pronominal argument’ (Jelinek 1984) draws on the relation to syntax again: the arguments of the verb are fully and exclusively realized by bound morphemes, i.e., by morphological marking. In contrast to the optionality of noun incorporation, argument incorporation does not envisage such a choice. Since the arguments of the verb are realized as bound morphemes joined to the verbal root or stem, no lexical arguments are necessary or even possible outside the verb. The expression of grammatical relations is thus shifted from the syntactic level to the word level, as is illustrated by all examples to come.

As opposed to noun incorporation then, ‘pronominal,’ i.e., morphological, arguments are not potential syntactic arguments but represent a fundamentally different strategy for the realization of arguments. They fulfill the requirement of representing the fundamental grammatical relations of the sentence, and stand in contrast to inflection, which only reflects the relations constituted on the syntactic level. In order to decide whether a person-marking morpheme has mere inflectional status or whether it actually represents the argument(s) of the verb, careful investigation is needed to decide whether syntactic argument positions are available in a particular language. Such investigations have been carried out only for comparatively few languages, such as the Australian language Warlpiri (Hale 1983, Jelinek 1984, Simpson 1991, see also Austin and Bresnan 1996), the Iroquoian language Mohawk (Baker 1996), Salishan languages (Jelinek and Demers 1994, Jelinek 1995) the Athabaskan language Navajo (Jelinek 2000) and Inuktitut (Eskimo) (Nowak 1996a, 1996b). Discussions of Algonquian languages seem to point in a similar direction, but do not yet cover a broad range of relevant tests. Other languages suggest themselves as promising candidates, but have not been scrutinized, as is, for example, the case with most of the languages discussed by Mithun (1984).

Many morphological argument languages tend to feature a wealth of lexically and/or functionally bound morphemes which have no independent lexical counterparts, and the proportion of roots and stems to bound morphemes is dramatically changed in favor of the bound morphemes (Nowak 2001).

In Inuktitut (Eskimo), whole semantic fields denoting activities, states, events, processes, including ‘being’ and ‘having,’ as well as the indication of reported speech and epistemic specifications, are expressed exclusively by bound morphemes which are suffixed to the stem they modify or supplement.

(4) Inuktitut

qimmiqaqtunga

qimmiq-qaq–tunga

dog-have.itr-1s.itr ‘I have a dog/dogs.’

(5) Inuktitut

nirijumajara

niri–juma–jara

eat–want–1s.3s.tr. ‘I want to eat it.’

In the Salishan languages of the American Northwest Coast, bound morphemes cover the do-main of referential expressions and do so exclusively; so-called ‘field suffixes’ such as -ene, -ne, ‘ear’ in Ex. (6) lack independent counterparts.

(6) Kalispel

na cene he gets an ear-ring

(Vogt 1940, p. 53)

In cases where items with lexical content are incorporated, no matter whether they are free or bound morphemes, typically the morphological argument and the incorporated lexical items do not exhibit agreement. Yet some cases of such internal agreement relations have been reported. For Southern Tiwa, obligatory incorporation is reported. In this language, person marking on the complex verb also reflects the person, number, and class of the incorporated noun. (Allen et al. 1984, pp. 295–8). The same seems to be the case with Bininj Gun-wok, a language of North Australia.

(7) Bininj Gun-wok

(a) ba-ginje-ng gun-ganj

3/31-cook-PP IV-meat

‘(S)he cooked (the) meat.’

(b) Ba-ganj-ginje-ng

3/31-meat-cook-PP

‘(S)he cooked (the) meat.’

(Evans 1999, p. 259)

3. Grammatical Incorporation

Grammatical relations may also be realized verb internally. Fundamental syntactic operators such as auxiliaries, pronouns, adpositions, complementizers, and other syntactic function words often do not exist at the word level at all.

(8) Inuktitut

(a) niuvirvingmuaqtutit

niuvirvik -muaq–tutit

shop-go.to–2s.itr

‘You go to the shop.’

The term niu ir ik ‘shop’ is a lexicalized affix combination, as is -muaq‘go to’; niu ir ik is composed of niu iq-, ‘to acquire’ and a nominalizing affix ik ‘place where s.th. happens’; -muaqis a highly frequent combination of the (singular) terminalis case -mut ‘to’ and -aq‘go.’ Besides these lexicalizations, all affixes are perfectly productive (Nowak 1996a, p. 255).

(8)

(b) niuvirvingmut

niuviq–vik-mut

acquire-place.where-term.s.

‘to the shop’

Interestingly enough, the noun marked by the terminalis is not stripped of its plural when involved in such an incorporation.

(8)

(c) niuvirvingnut

niuviq–vik-nut

acquire-place.where-term.p.

 ‘to the shops’

(d) niuvirvingnuaqtutit

niuvirvik -nut -aq-tutit

shop-term.p-go–2s.itr

‘you go to the shops’

Morphological arguments in many instances are fused with other features such as the indication of clause linkage, often referred to as ‘switch reference,’ or ‘anaphoric coreference,’ as can be seen in (9).

(9) Inuktitut

tikiniraqtauvutit piqsiqtillugu

tiki(t)–niraqtau–vutit piqsiq -tillugu

arrive-say.that.itr–2s.itr blizzard -4s.vpart.itr

‘somebody said that you have arrived while it was blizzarding’

Grammatical incorporation thus can be seen as the morphological realization of functional relators, i.e., as the word internal establishment of relationships which would otherwise demand syntactic spell out. The results are highly complex forms equivalent to full propositions.

(10) Woods Cree

sipwıhcıhkasow

away by-foot AI pretend AI 3I

‘he pretends to leave’ (Starks 1992, p. 127)

(11) Warlpiri

Pingka-rlipa mata-ma-ninja-kujaku

slow-1pis tired-CAUS-INF-PURP-NEG ya-ni

go-NPST

‘We’ll go slowly so that it (walking a long distance) doesn’t tire us out.’

(Simpson 1991, p. 108)

4. Conclusion

It must be kept in mind that languages hardly ever represent ideal types. This applies to what has been said about syntax and morphology, and the degree to which they interact. The types of incorporation discussed here exemplify an increasing complexity, or configurationality, in morphology, illustrating a unidirectional relationship between morphological complexity and the morphological realization of arguments. Languages of high morphological complexity also feature the morphological realization of arguments, but it is not necessary that all languages realizing their arguments morphologically must be morphologically complex in other respects.

The morphological strategies employed by such languages range from concatenative (Eskimo) through template (Athabaskan) to morphophonemic and prosodic devices (Salishan), and often exhibit mixtures of all these types. A unified account of such complex and multilayered morphological patterns is a desideratum that must be left for future research.

Noun incorporation on the one side, and argument incorporation and grammatical incorporation on the other, are separated by the optionality vs. obligatoriness of incorporation: with the latter, the speakers do not have a choice. The last type, grammatical incorporation, extends morphological configurations to other than argument relations, i.e., to grammatical relations in general. This is the content of the traditional term ‘polysynthetic.’

Bibliography:

1. Allen B, Gardiner D, Frantz D 1984 Noun incorporation in Southern Tiwa. International Journal of American Linguistics 50: 292–311
2. Austin P, Bresnan J 1996 Non-configurationality in Australian Aboriginal languages. Natural Language and Linguistic Theory 14(2): 215–68
3. Baker M 1988 Incorporation. University of Chicago Press, Chicago
4. Baker M 1996 The Polysynthesis Parameter. Oxford University Press, Oxford, UK
5. Bonvillain N 1972 Noun incorporation in Mohawk. In: Forster M (ed.) Papers in Linguistics from the 1972 Conference on Iroquoian Research. Mercury Series, Ethnology Division Paper No. 10. National Museum of Man, Ottawa, pp. 18–26
6. Cook E D, Wilhelm A 1998 Noun incorporation. New evidence from Athabaskan. Studies in Language 22(1): 49–81
7. Evans N 1999 Why argument affixes in polysynthetic languages are not pronouns: Evidence from Bininj Gun-wok. Sprachtypologie und Universalienforschung 52: 255–81
8. Gerdts D 1998 Incorporation. In: Spencer A, Zwicky A (eds.) The Handbook of Morphology. Blackwell, Oxford, UK, pp. 84–100
9. Hale K 1983 Warlpiri and the grammar of non-configurational languages. Natural Language and Linguistic Theory 1: 5–47
10. Jelinek E 1984 Empty categories, case, and configurationality. Natural Language and Linguistic Theory 2: 39–76
11. Jelinek E 1995 Quantification in Straits Salish. In: Bach E, Jelinek E, Kratzer A, Partee B H (eds.) Quantification in Natural Languages. Kluwer, Dordrecht, The Netherlands, pp. 487–540
12. Jelinek E 2000 Navajo as a discourse configurational language. In: Fernald T, Platero P (eds.) The Athabaskan Languages: Perspectives on a Native American Language Family. Oxford University Press, Oxford, UK
13. Jelinek E, Demers R 1994 Predicates and pronominal arguments in Straits Salish. Language 70: 697–736
14. Kroeber A 1909 Noun incorporation in American languages. In: Heger F (ed.) Internationaler Amerikanisten-Kongress XVI. Hartleben, Vienna, Austria, pp. 569–76
15. Mithun M 1984 The evolution of noun incorporation. Language 60: 847–94
16. Nowak E 1996a Transforming the Images. Transitivity and Ergativity in Inuktitut (Eskimo). Mouton de Gruyter, Berlin
17. Nowak E 1996b Uber die organisation komplexer Aussagen. In: Harras G, Bierwisch M (eds.) Wenn die Semantik arbeitet. Festschrift fur Klaus Baumgartner. Niemeyer, Tubingen, Germany, pp. 29–56
18. Nowak E 2001 Lexical categories in polysynthetic languages. In: Cruse A, Hundsnurcher F, Job M, Lutzeier P R (eds.) Lexikologie-Lexicology. Handbucher zur Sprache und Kommunikation. Mouton de Gruyter, Berlin
19. Sapir E 1989/1911 The problem of noun incorporation in American Languages. In: Bright W (ed.) The Collected Works of Edward Sapir. Mouton de Gruyter, Berlin, Vol. 5, pp. 27–60
20. Simpson J 1991 Warlpiri Morpho-Syntax. A Lexicalist Approach. Kluwer, Dordrecht, The Netherlands
21. Spencer A 1995 Incorporation in Chukchee. Language 71(3): 439–89
22. Starks D 1992 Aspects of Woods Cree Syntax. Ph.D. thesis, University of Manitoba, Winnipeg, Canada
23. Vogt H 1940 The Kalispel Language. Det Norske Videnskaps Akademivi Oslo, Oslo, Norway