Noun Incorporation

En las frases posesivas del wayuunaiki, el poseído antecede al poseedor y exhibe un prefijo que concuerda con éste: nü-tüna Luuka 'el brazo de Lucas'. En las frases preposicionales, la preposición antecede al complemento y exhibe un prefijo que concuerda con éste: nü-maa Luuka 'con Lucas’. Existe ASCENSO DEL POSEEDOR, donde éste asciende de constituyente frasal a constituyente clausal (con el poseído integrando un verbo complejo). Existen APLICATIVOS, donde el complemento de la preposición deviene objeto (con la preposición integrando un verbo complejo). Ambos fenómenos resultan de un mecanismo general de incorporación. En las lenguas analizadas en Baker (1988), las construcciones aplicativas ocurren cuando las FNs cumplen roles semánticos de dativo, beneficiario, instrumental o locativo. La incorporación preposicional guajira contrasta porque los objetos aplicados desempeñan cualquier rol semántico, produciéndose así aplicativos irrestrictos. Estos objetos aplicados son accesibles también a la pasivización, originándose abundancia de seudopasivas. El aplicativo explica también la facilidad de relativizar una FN en cualquier función gramatical (Keenan & Comrie, 1977). La incorporación promueve frases nominales de posiciones no relativizables a posiciones relativizables (sujeto/objeto).

Se describen las principales construcciones posesivas en guajito/wayuunaiki: (a) diferencias morfológicas entre nombres absolutos (alienables) y nombres relativos (inalienables), los prefijos para indicar la persona y número del poseedor, los diversos sufijos posesivos usados con los nombres alienables, ausencia de tales sufijos en los nombres inalienables por la existencia de una forma indefinida; (b) verbos posesivos y privativos con los prefijos ka- y ma- (y alomorfos) y formas derivadas a su vez de estos temas; (c) construcciones posesivas con la preposición -ma'ana; (d) incorporación del nombre cabeza de frases posesivas al verbo para formar un verbo complejo (Baker, 1988), permitiendo al ascenso del poseedor de constituyente de frase a constituyente de cláusula, y consecuencia de este ascenso para la relativización.

Esta obra es un compendio de la gramática de la lengua wayuu donde expongo de manera sucinta las nociones más importantes de su morfología y su sintaxis. Se desprende de dos anteriores publicaciones mías, pero he revisado definiciones y explicaciones, acortado secciones, eliminado ejemplos redundantes, modificado gráficos, tablas y esquemas, reordenado la secuencia de varias secciones, eliminado apartes no estrictamente gramaticales y, en fin, he procurado dar una visión de conjunto un poco más simple, pero sin dejar de ser rigurosa. Las secciones se presentan como en unidades manejables de corta o mediana extensión para permitir cubrirlas en tiempos relativamente cortos. Como es una obra de consulta rápida, no se incluyen ejercicios, pero sí abundantes ejemplos obtenidos mediante colaboración de hablantes nativos, o extraídos de los textos escritos por connotados autores wayuu, así como también de las traducciones al wayuunaiki de El Nuevo Testamento, y de la obra literaria clásica El Principito. Se ha respetado siempre la variante dialectal usada en esos textos, pero normalizando la ortografía. Con la finalidad de contrastar las lenguas, en la mayoría de las secciones se presentan primero las nociones gramaticales y ejemplos en español, para luego desarrollarlas en wayuunaiki. Mi deseo es que esta obra sirva de base a los instructores de wayuunaiki para que puedan preparar con facilidad sus lecciones para la enseñanza de la gramática. Con el estudio de esta obra, los interesados en esta lengua podrán alcanzar un conocimiento sólido que les permitirá trabajar con una buena base en las urgentes tareas de promoción de esta maravillosa y rica lengua. Deseo expresar mi respeto y admiración a los autores de gramáticas del wayuunaiki realizadas en distintas épocas, con diferentes audiencias en mente y con diferentes metodologías. Esta gramática circula primeramente en versión digital, pero confío en que pronto se publicará en forma impresa. En esta versión digital he colocado fotografías de amigos wayuu que me han acompañando en este trabajo durante muchos años.

The volume is aimed at preserving invaluable knowledge about Ainu, a language-isolate previously spoken in Hokkaido, Sakhalin, and Kurils, which is now on the verge of extinction. Ainu was not a written language, but it possesses a huge documented stock of oral literature, yet is significantly under-described in terms of grammar. It is the only non-Japonic language of Japan and is typologically different not only from Japanese but also from other Northeast Asian languages.
Revolving around but not confined to its head-marking and polysynthetic character, Ainu manifests many typologically interesting phenomena, related in particular to the combinability of various voice markers and noun incorporation. Other interesting features of Ainu include vowel co-occurrence restrictions, a mixed system of expressing grammatical relations, which includes the elements of a rare tripartite alignment, nominal classification distinguishing common and locative nouns, elaborate possessive classes, verbal number, a rich four-term evidential system, and undergrammaticalized aspect, which are all explained in the volume.
This handbook, the result of unprecedented cooperation of the leading experts of Ainu, will definitely help to increase the clarity of our understanding of Ainu and in a long-term perspective may provide answers to problems of human prehistory as well as open the field of Ainu studies to the world and attract many new students.

Resumen: En este trabajo se analiza el español coloquial hablado desde el punto de vista de la tipología morfológica que distingue entre lenguas aislantes, aglutinantes y flexivas. Se compara el español coloquial espontáneo con lenguas consideradas aislantes como el chino, aglutinantes como el turco o el euskera y con las lenguas esquimales, consideradas polisintéticas. La conclusión del trabajo es que el español coloquial es una lengua tan aglutinante como pueda serlo el turco, el vasco o el yupí. Son las convenciones ortográficas típicas de las lenguas europeas las que han llevado a los lingüistas a pensar que hay lenguas aglutinantes y polisintéticas y no polisintéticas. Para apoyar el análisis presentado en este artículo se recurre a propuestas recientes de Ö Dahl (2004) y W. J. de Reuse (2009). Palabras clave: tipología morfológica, morfología del español, aglutinación, polisíntesis Summary: In this paper spoken colloquial Spanish is analyzed from the point of view of the traditional morphological typology that distinguishes isolating, agglutinative and inflectional languages. Colloquial Spanish is compared with isolating languages such as Chinese, agglutinative languages such as Turkish or Basque and with the so-called polysynthetic languages such as the Eskimo languages. It is argued that colloquial Spanish presents a degree of agglutination comparable to that of Turkish, Basque or Yupi. The reason why Spanish and other European languages are not usually considered as agglutinative lies in the fact that linguists rely on the orthographic conventions used to write these languages. In order to support the arguments given in this paper some recent proposals and analyses by Ö. Dahl (2004) and W. J. de Reuse (2009) are discussed and adopted.

Gumuz is a Nilo-Saharan dialect cluster spoken in the river valleys of northwestern Ethiopia and the southeastern part of the Republic of the Sudan. There are approximately 200,000 speakers, the majority of which reside in Ethiopia. This study is a phonological and grammatical analysis of two main dialects/languages: Northern Gumuz and Southern Gumuz.
The study provides an overview of the Gumuz people and culture, including historical accounts of the language(s) and migration patterns. Most major aspects of the language are described and analyzed in detail: phonology, nouns, pronouns, demonstratives and other noun phrase constituents, verbs and verbal morphology, noun incorporation, verbal classifiers, noun categorization, basic clauses, and subordinate clauses. Northern and Southern Gumuz varieties are contrasted throughout.
Gumuz tone has two levels, High and Low, with tonal downstep of High. The tonal melody on bound pronominals on verbs indicates transitivity.
    Nouns are divided into two basic types: relational and absolute. Relational nouns have an inherent relationship with another nominal element, either within a noun-noun compound or with a (historical) possessive affix. Two sets of relational nouns —
attributive and relator nouns— obligatorily take an inherent possession suffix if not in a compound.
Gumuz has two noun-noun constructions: the Associative Construction and the Attributive Construction. The first is left-headed with ‘noun of noun’ semantics. The second is right-headed with the initial noun expressing an inherent quality of the second.
    Certain body part terms have grammaticalized as a variety of other morphosyntactic categories, in particular as relator nouns, verbal classifiers, and class morphemes, the final two of which are noun categorization devices. Many of these same body part terms can be incorporated into the verb or form part of lexicalized verb-noun compounds.
      Deverbal nominalizations with /ma-/ are found throughout the language structures. These /ma-/ nominalizations serve as both subject and object complements. They are also commonly found in other subordinate clauses such as relative and adverbial clauses. Purpose clauses are formed with the dative preposition plus a /ma-/ nominalization. Finite purpose clauses take pronominal inflection and have further grammaticalized as future tense main clause verbs in Southern Gumuz.

""This article investigates the synchronic status and diachronic origin of an incorporation-like construction in Japhug, a polysynthetic Sino-Tibetan language of Eastern Tibet. This construction constitutes the intermediate stage on a path of development from denominal derivation to incorporation, the opposite of the usual path of development from incorporation to denominal derivation.

Additionally, this article shows that similar phenomena exist in other languages, and that coalescence between noun and verb is not the only attested diachronic origin of incorporating verbs.""

The Proceedings of the 38th Siouan and Caddoan Conference follows the meeting of the 38th Siouan and Caddoan Language Conference from June 7-9, 2018 at Northeastern Illinois University in Chicago, IL.  This volume was published on July 12, 2019 in Chicago, IL by the Northeastern Illinois University Linguistics Department and edited by Ryan M. Kasak.

This work represents the first time in over three decades that the conference produced a proceedings volume. The papers are copyright of their original authors and are published under the Creative Commons Attribution 4.0 License (CC BY 4.0).

Ainu is a typical polysynthetic language in the sense that a single complex verb can express what takes a whole sentence in most other languages. A single verb form may include more than one heavy element: up to two applicative prefixes (out of three), two causative suffixes (out of five), two incorporated objects, one lexical prefix (out of two originating in nouns ‘head’ and ‘bottom’), one verbalizing suffix (originating in the verb ‘make’), as well as reciprocal, reflexive and general object (=antipassive) prefixes and agreement affixes for the first/second person subject and object. The degree of combinability of various voice markers and noun incorporation is spectacular. Nevertheless, it has been claimed that Ainu deviates from more typical polysynthetic languages such as Mohawk in that it has less freedom of word order, interrogative phrases in situ, and unrestricted morphological causatives (Baker 1996). The present paper aims to distinguish what Ainu shares with other polysynthetic languages from what is truly unique to Ainu.

This thesis examines the phenomenon of ‘object incorporation’ (aka ‘noun incorporation’) in relation to Uzbek nominals in the object position.
Based on previous cross-linguistic research of object incorporation, as well as novel data presented in the current work, I argue that certain Uzbek direct objects, specifically bare singular nominals, do not engage in ‘true incorporation’ (von Humboldt, 1836; Kroeber, 1909; Sapir, 1911; Mithun, 1984; Sadock, 1980; Baker, 1988; van Geenhoven, 1998), but rather undergo the process known as ‘pseudo-incorporation’ (Massam, 2001; Dayal, 2003; Farkas & de Swart, 2003; Espinal & McNally, 2011; Öztürk, 2005; Kamali, 2015; Modarresi, 2014; Krifka & Modarresi, 2016).
In accounting for the various morpho-syntactic and semantic properties of Uzbek bare objects, I suggest that they get pseudo-incorporated into verbs by means of predicate restriction, whereby the meaning of a verb is restricted to the meaning of a bare nominal via the non-saturating RESTRICT mode of semantic composition (Chung & Ladusaw, 2004; Modarresi, 2014).
As such, this thesis contributes to a deeper understanding of such topics as (pseudo-) incorporation, bare nominals in the object position, discourse anaphora and number-neutrality in a relatively understudied Turkic language.

One overlooked and highly polysemous English noun phrase form is the bare singular, i.e., the null determiner with a singular count noun complement. Occurring in all grammatical positions, this constituent shape is used in English for multiple functions. Examination of naturally occurring English data shows the conditions under which bare singulars are used as generics (like bare plurals), as deictics (like definite articles, demonstratives, and possessive determiners), and as components of a predicate conveying a stereotypical activity (like indefinite articles).  The English bare singular aligns with morpho-syntactic options—including preposition choices, different nominal constructions, and case marking—that provide this same range of referential options in other languages.

Polysynthesis presupposes the existence of 'words', a domain or unit of phonology and syntax that is extremely variable within and across languages: what behaves as a 'word' with respect to one phonological or syntactic rule or constraint may not behave as such with respect to other rules or constraints. Here we develop a system of variables that allows cataloguing all verb-based domains in a language in a bottom-up fashion and then determining any potential convergence of domains in an empirical way. We apply the system to case studies of Mapudungung and Chintang. These confirm earlier observations that polysynthetic languages do not operate with unified units of type 'word' in either phonology or syntax.

This thesis is an examination of a number of related morphosyntactic constructions in Wubuy (formerly named ‘Nunggubuyu’ by Heath (1984)), a polysynthetic Gunwinyguan language from south-east Arnhem Land. In particular, focus is placed on possessive constructions and how these interact with other areas of the morpho-syntax, such as verbal agreement and noun incorporation. Additionally, this thesis separates Wubuy possessive constructions from possessive interpretations of benefactive applicatives. In Wubuy, possession behaves in a way that is reminiscent of ‘external possession’, or ‘possessor raising’, constructions (B. Baker, et al. 2010; M. Baker 1996; Payne and Barshi 1999a). However, empirical data demonstrates that current typologies and formal analyses of EP and NI cannot adequately account for all of Wubuy’s possessive mechanisms. This thesis will outline why these accounts are unsatisfactory, and it will suggest a new classification of Wubuy’s possessive constructions that is more successful in capturing the intrinsic grammatical and semantic contrast between different types of possession. A possible avenue for the development of a new approach to Wubuy possessive constructions within the framework of Lexical Functional Grammar (Lødrup 2009; Nordlinger and Sadler 2008; Schrock 2007) will also be proposed. Data is taken from a variety of sources, including my own fieldwork with speakers.

This paper is devoted to S- and P-argument noun incorporation in a Chukchi variety spoken in Amguema (Far-Eastern Chukotka). The paper is based on my fieldwork in Chukotka as well as on the text corpus collected by other HSE and MSU linguists. Special attention is paid to the constraints on incorporation and the possibility of 'promotion' of another participant to Absolutive position 'vacated' by incorporation.

Ese Ejja is an Amazonian language that displays two types of noun incorporation. The first type is typical of the Amazonian area: it occurs within verb predicates, is restricted to inalienable nouns and has no influence on the valency of the verbal predicate. The second type is unusual in that it involves adjectival predicates and results in adjectives, semantically comparable to the English derived adjectives ‘blue-eyed’ or ‘red-haired’. However, the phenomenon is far more productive in Ese Ejja: adjectives do not only incorporate body-part terms but also possessed nouns, and even NPs and verbal roots. Their modifying function compensates for the quasi absence of attributive adjectives in the language. The analyses presented here are based on first-hand data collected in a Bolivian community.

This paper explores the polyfunctionality, grammaticalization, and typological relevance of applicatives in Ainu. Applicatives are derived by the valency-increasing prefixes which are generally defined here as instrumental e-, dative ko-,
and locative o-. The referential range of the respective constructions stretches over several semantic roles and the exact role is attributed to the interaction between the semantics of the prefix and verb. The typologically unusual properties of Ainu applicatives include the ability of e- applicatives to add the roles of Theme and Content, the ability of the so-called unaccusative intransitives to host applicative prefixes e- and ko-, the possibility of e-ko- and ko-e- double applicatives, the absence of non-applicative paraphrases for some applicatives, and the possibility of applicative object incorporation.

This paper argues that wordhood in the polysynthetic Inuit language is predictable from syntactic structure and that words correspond to the domains of CP and DP. This entails that Inuit’s morphological component need not be any more complex than that of more isolating languages and that individual morphemes are not idiosyncratically specified as affixes. As evidence for our approach, we contrast a variety of free and bound elements, showing that in every case, subparts of words are smaller than CP/DP and full words correspond to CP/DP. We also discuss “stem” ellipsis, which we argue is further evidence that the elements which are usually bound in Inuit are not genuinely affixes.

This thesis is devoted to constructionist and force-dynamic analysis of argument structure constructions involving noun incorporation in Amguema Chukchi. I argue that the restrictions on the possibility and syntax of NI constructions can be explained without syntactic relations and semantic roles. Instead, the semantic maps, diachronic development pathways and order of participants in a causal structure is employed.
Supplementary typological study on the basis of 38 languages is conducted to collect evidence on the semantic map and development pathways.

"I will argue, following Singh and Dasgupta’s (2003) and Starosta’s (2003a) analysis of compounds, and Starosta’s (2003b) analysis of incorporation, that NI is a not-so-special case of derivation by way of a suffix that looks like a noun, and that as such it is amenable to a word-based treatment, contrary to what Anderson (1992: 292 ff.) says (cf. Starosta 2003a). If we treat NI as just another case of lexical subordination (1c) rather than as a combination or translation of words or category-bearing stems – as in (1a) or (1b), which we could expect to be very productive and semantically compositional – then the limited productivity and semantic idiosyncrasy of incorporation in Innu will follow naturally.
(1) a. [X]V. + [Y]N.
b. [[XYt]V. [tN.]NP]VP
c. /X/V. ↔ /Xabc/V.
Where abc = some specified phonological material
resembling a noun.
Using mainly data from Drapeau (1999) and forms elicited from a native speaker, I would like to offer a word-based lexicalist account of NI in Innu that goes a step further than Mithun (1984: 847) by adopting a truly word-based theory of morphology in which it is impossible in principle to say that NI is nearly syntactic. What is more, this account also has the advantage of bringing together under the same type of formal process all three types of incorporated nominals (“medials”) recognized by Wolfart (1973: 66-67).
"

In this paper we examine the interaction of a number of grammatical phenomena in Wubuy, a polysynthetic language from northern Australia, and show how they can be given a comprehensive analysis within the framework of LFG. While each of these phenomena–noun incorporation, verbal agreement, coordination and external possession–has received various treatments within the LFG literature, no one study has addressed the compatibility of these analyses under interaction, despite the fact that they frequently co-occur in the ...

An investigation of the first language acquisition of productive nouns in Inuktitut (Inupiaq) is presented. This study begins with descriptions of noun incorporation, relevant aspects cl the ntructure of Inuktitut, and working criteria of productivity. Acquisition data from Inuktitut and corroborating data from Greenlandic are outlined and contrasted with evidence from Nohawk. Several structural and sociolinguistic explanations for tht seemingly early acquisition of noun incorporation in Inuktitut in relation to its acquisition in Mohawk are offered: (1) placement of verbal affixation in relation to the incorporated noun differs significantly in the two larguages; (2) the criteria for use of noun incorporation are more restrictive or clearer in Inuktitut; (3) noun incorporation is considered the most usual way to represent the cmcept in Inuktitut, and is therefore learned earlier; and (4) the Inuktitut living environment is more conducive to language learnIng than the Mohawk environ...

Northwest Caucasian (NWC) languages have usually been considered to lack noun incorporation (NI). Indeed, they do not feature structures considered prototypical NI as familiar from such languages as Iroquoian or Chukotkan, i.e. optional productive compounding of the verb with the root of its P or S argument under specific discourse conditions with consequences for transitivity. However, I argue that NWC still shows phenomena amenable to an NI analysis, i.e. the so-called spatial preverbs, many of which not only historically go back to incorporated nouns, but may be analysed as such synchronically as well, albeit under a broad understanding of NI. In my talk, I shall focus on two types of incorporated nouns in Abaza, a member of the Abkhaz-Abaza branch of NWC, i.e. body-part nouns and non-relational nouns, discussing similarities and differences in their semantics and morphosyntactic behaviour, drawing parallels from other languages of the family. I shall argue that Abaza and, more broadly, NWC spatial preverbs form a cline from incorporated nouns to fully grammaticalised lexical affixes, suggesting several successive stages of incorporation and ensuing affixalisation.

In this paper I show that multiple non-polysynthetic Uto-Aztecan (UA) languages display properties that have been attributed to polysynthesis, and I suggest that these properties led to polysynthesis in Nahuatl, rather than the reverse.  I argue that Nahuatl polysynthesis developed in a discrete series of stages, each of which has attestation elsewhere in the family.
The proposal made here is consistent with divergent work in both grammaticalization and more formal approaches to language change.  The gradualist approach links a gradual syntactic change, the development of obligatory NI with the fixing of subject clitic placement, to the abrupt change required by approaches to diachronic syntax focused on child language acquisition (e.g. Lightfoot 1991).  If this gradualist approach is correct for the polysynthesis parameter, then it could probably also be applied to other proposed macroparameters as well.  Thus, this line of research suggests the need to search for potential ‘precursors’, in the form of retentions, in other languages and language families with no written record, in order to find (and reconstruct) likely sources for the ‘grammaticalization’ of particular ‘macroparameters’.

Despite superficial differences, constructions with the verbalizer peTuka in Malayalam can be shown to derive from a uniform structure.  Parallels could be drawn between the way peTuka functions to verbalize STATIVE nominals  and its role as an independent verb in the Malayalam; and the verbalization  therefore  is amenable to an analysis involving  incorporation of STATIVE nominals to peTuka.  This denominalization in Malayalam could perhaps represent the overt counterparts to the well known instances of  zero derived locative and locatum denominal verbs in English(as proposed in Hale and Keyser, 2002 and earlier), albeit as intransitives.  On transitivization, these denomInals result in two kinds of structure: one, conforming to the 'change of location' verbs, and the other, the 'change of possession' verbs, i.e. double object ditransitives.

The talk deals with peculiar syntax of Hittite existential quantifiers which are consistently immediately preverbal. I argue against their in-situ construal by (Huggard 2015) and suggest their linear position can be best explained by positing a dedicated eQuantifier Projection. It is shown that solutions proposed for superficially similar phenomena such as DOM/DSM, pseudo-incorporation or object shift do not hold for Hittite.

В кн.: Индоевропейское языкознание и классическая филология - XXII. Материалы чтений, посвященных памяти профессора Иосифа Моисеевича Тронского, 18-20 июня 2018 г. / Отв. ред.: Н. Казанский. СПб. : Наука, 2018. С. 1063-1069.

The paper discusses 10 Old Frisian verbal constructions that can be considered noun-incorporated verbs. Nominal incorporation is wide-spread in Indo-European. In Germanic, nominal incorporation as applied to verbs is not productive as a word-formation tool, except for Frisian. S. Dyk, a linguist and an expert in Frisian, has carried out a comprehensive research on noun-incorporation in Modern Frisian (Dyk 1997). Some incorporated verbs are part of Middle Frisian texts. Yet, as the author states, no words following this word-formation pattern had been attested in Old Frisian. This paper present new data achieved within a PhD-thesis on compounding in Old Frisian (including all the attested lexis, which amounts to ca. 11,750 lemmas). The findings are 10 Old Frisian lexical constructions that might be treated as noun-incorporated (proto-)compounds due to a set of reasons. The arguments for considering these words to be noun-incorporated compounds are: (1) ‘terminological’ specification of their semantics; (2) proven evaluation of this word-formation pattern into a productive and frequent mechanism in Modern Frisian through Middle Frisian. The arguments against considering these words to be noun-incorporated compounds are: (3) no conjugational paradigm present in the actual contexts, i. e. the 10 words occur exclusively as substantivized infinitives and do not function as finite verbs yet; (4) the 10 words are not frequent in terms of being attested in various sources distinguished by chronological, spatial and genre-based criteria. Moreover, the paper discusses some limitations of the words’ possible interpretations, and their formal and semantic features are described. The 10 words are: bon-skelda ‘to impose a fine’, brond-skatta ‘to commit arson’, hēr-plokkia ‘to pull at someone’s hair’, holt-sāgia ‘to cut wood’, hreg-breka ‘to break someone’s back’, mes-lūka ‘to pull a knife’, rēd-slā ‘to give advice’, stēn-drega ‘to carry stones around the town (as a punishment)’, stēn-fēra ‘to move stones’, wax-drāia ‘to produce wax candles’. Semantically, most of the words refer to criminal, legal actions; two of them (‘to cut wood’ and ‘to produce wax candles’) are designations of highly frequent occupational actions. One of them, ‘to give advice’, is well-known for having cognates in other Germanic languages. These meanings might have been rendered through a noun-incorporating pattern for a reason: they denote some actions so frequent and collocational that they were bound to form ‘terminological’ items and develop into a productive ford-formation model.

This paper examines a type of noun incorporation and its related patterns of doubling in Budai Rukai (Austronesian, Taiwan). An incorporated nominal root is indefinite and forms a close morphological bond with the incorporating verb. The nominal root modifies the predicate semantically and does not act as a direct object. The resulting forms describe various activities and exhibit properties of (in)transitivity and (a)telicity. Noun incorporation of such is formed by an array of verbal prefixes and nominal roots, while doubling is limited to a handful of complex verbs.

A heritage language grammar is an incomplete grammar resulting from partial or interrupted acquisition due to insufficient exposure to that language in childhood, or attrition caused by reduced input and usage later in life. It is common in immigrant and aboriginal communities, where a minority language is spoken alongside a majority language (such as English in Ontario). Heritage language grammars can make important contributions to the study of language by providing new sources of data for linguistic theory, or new perspectives on first and second language acquisition and bilingualism (Montrul 2008, Benmamoun, Montrul and Polinsky 2013). The linguistic variables in heritage language grammars differ in vulnerability to incomplete acquisition or attrition. Morphology, in particular inflectional morphology such as verb and noun agreement, is more vulnerable than syntax (Håkansson 1995, Montrul 2004, Keijzer 2008). Within inflectional morphology, nominal morphology such as gender, num...

A heritage language grammar often develops through insufficient exposure in
childhood (incomplete acquisition) or insufficient use later in life (attrition), and manifest
as deficits in grammar, both in syntax and – even more so – in functional morphology
(see Benmamoun, Montrul & Polinsky, 2010 for a review). This paper presents the
results of a study investigating a particular structure, noun incorporation (NI), in Inuktitut
heritage speakers living in Ottawa. It is the first study of noun incorporation in adult
heritage language grammar, as this phenomenon does not exist in more commonly
studied heritage languages.
Example (1) illustrates NI in Inuktitut. The properties of NI in Inuktitut are: (1) the
verb is a light verb (Johns, 2007), (2) the incorporated noun precedes the verb in the
verbal complex, (3) the incorporated noun does not have case, and (4) the modifier of
the incorporated noun (e.g. ‘green’ above) appears with –mik case, which is the case
found on antipassive objects.
(1)
Ilisaiji saa-taaq-tuq (uujaujar-mik)
teacher table-get-3s.part green-MIK
'The teacher got a (green) table.’
How is case in noun incorporation represented in a heritage grammar? Fluent
speakers (n=11) and heritage speakers of Baffin dialects of Inuktitut (HS, n=13) living in
Ottawa participated in a grammaticality judgment task. In this paper we focus on the
following conditions: preferred NI and non-NI structures, case on incorporated and non-
incorporated nouns, and case on adjective modifiers of incorporated and non-
incorporated nouns.
The fluent speakers preferred both structures most of the time (75%), while HS
were more likely to choose just one of the structures, with the choice between NI and
non-NI seemingly random (30% each).
HS were not significantly different from fluent speakers in correctly accepting
caseless incorporated nouns (91% vs. 97% accuracy), but they rejected incorporated
case-marked objects significantly less frequently than fluent speakers (75% vs. 98.5%).
To test their general knowledge of case we also included sentences with non-
incorporated nouns with and without case. They correctly accepted case-marked non-
incorporated objects most of the time (88.5%), while they correctly rejected caseless
objects less frequently (70.5%).
HS showed a similar pattern with correct acceptance of -mik on modifiers of
incorporated objects (86% vs 95.5% for fluent) and correct rejection of missing -mik on
modifiers (76% vs 95.5% for fluent), and performed better on modifiers of incorporated
objects than on modifiers of non-incorporated objects. For the latter, HS performed
much lower than fluent speakers in both the correct case-marked condition (66.7% vs.
93.9%) and incorrect caseless condition (53.6% vs. 92.4%).
These results suggest that HS are sensitive to the nature of the verb, the
syntactic status of the object (incorporated or not), and -mik case marking requirements
for incorporated and non-incorporated objects and their modifiers. Even though noun
incorporation does not exist in their dominant language (English), at least some aspects
of it are correctly represented in HS grammar, although its processing is inconsistent,
sometimes resulting in a failure to reject unlicensed case marking.