Content uploaded by Nancy C Kula
Author content
All content in this area was uploaded by Nancy C Kula on May 28, 2015
Content may be subject to copyright.
A preview of the PDF is not available
Syntactic gradients in compounding: Bemba associative nominals vs. Italian prepositional and deverbal compounds
Abstract
In this article we compare three classes of nominal constructions: Bemba so-called 'associative nominals', a class of nominal constructions found in several Bantu languages (though we will essentially concentrate on Bemba), Italian so-called 'prepositional compounds' (or 'phrasal com-pounds'), a class of nominal constructions common to other Romance languages (such as French and Spanish), and a specific class of preposition-less deverbal compounds that is peculiar to Italian and is not found in the other Romance languages. The reason for comparing such geographically and typologically distant languages is that Bantu associative nominals and Romance compounds share some important properties. As is well known, Romance noun-noun compounds (see also Basciano et al. 2011, this volume) differ from Germanic root compounds in a number of respects. First of all, Romance noun-noun compounds are left headed, whereas Germanic root compounds are typically right headed. Also, in Romance noun-noun compounds, the two nominals are sometimes separated by a phonologically independent preposition-like element which seems to contribute in a non trivial way to the meaning of the whole compound. Germanic noun-noun compounds sometimes feature a so-called 'linking element', occurring between the two nouns. However, this element has the form of a nominal inflection marker, it is a bound form, and does not seem to contribute in any substantive way to the meaning of the whole construction. Moreover, some recent contributions (see Delfitto & Melloni 2009, 2011) have shown that Romance prepositional compounds (that is, those noun-noun compounds featuring a preposition-like element) do not always obey the restrictions dictated by the Lexical Integrity Hypothesis and, therefore, are less morphologically encapsulated than Germanic root compounds. Interestingly, Bemba associative nominals pattern together with Romance prepositional compounds in that they are left headed, they are composed of two nominals separated by a phonologically independent marker which seems to restrict the set of possible interpretations of the whole construction, and, arguably, they lie outside the scope of the Lexical Integrity Hypothesis. The goal of this article is therefore to compare in greater detail Bantu associative nominals and two specific varieties of Romance com-302 pounds in order to further investigate their differences and similarities from a theoretical perspective. The paper is organized as follows. In section 1 we discuss the main structural and interpretive properties of Bemba associative nominals and consider a number of tests enlightening their degree of morphological encapsulation. In section 2 we consider Italian prepositional compounds, their structural and interpretive properties, and, by applying the same tests adopted for Bemba associative nominals, their degree of morphological encapsulation. In section 3 we discuss the preliminary results of the analysis. Section 4 offers an overview of the properties of Italian deverbal compounds, shortly considering their relationship with the two classes of nominal constructions discussed in the preceding sections. The general conclusions are found in section 5.
... What these studies have not addressed is the nature of the noun class prefixes and locative prefixes inside the VN-compounds in Cinyanja. Similarly, studies on associative constructions in Bantu languages have discussed the interpretation and the syntax of these constructions (Mugane, 1997;Delfitto et al., 2011) but none of these studies has discussed the nature of the associative marker in the associative compounds in Cinyanja. In this thesis, I address this gap. ...
... In the compound words above, the two nouns in (1a) are njinga 'bicycle and moto 'fire' linked together by an associative marker ya 'of' while in (1b) the two nouns are mmisiri 'expert' and nyumba 'houses' linked together by the associative marker wa 'of'. These nominal compounds are mostly referred to as associative nominals or associative constructions in the Bantu literature (Mugane, 1997;Delfitto et al., 2011;Kula, 2012), and they seem to have the same structure as regular associative nominal phrases which are generally syntactic as shown in example (2) because, by assumption, they include the agreeing associative marker. Further, Delfitto et al. (2011) argue that such associative constructions are syntactic because they allow modification of internal constituents. ...
... These nominal compounds are mostly referred to as associative nominals or associative constructions in the Bantu literature (Mugane, 1997;Delfitto et al., 2011;Kula, 2012), and they seem to have the same structure as regular associative nominal phrases which are generally syntactic as shown in example (2) because, by assumption, they include the agreeing associative marker. Further, Delfitto et al. (2011) argue that such associative constructions are syntactic because they allow modification of internal constituents. In (2) above, the regular associative nominal phrases, like the compounds with the associative marker, also include two nouns; the head noun and the possessor 9 linked together with an associative marker. ...
Phrasal compounds such as two-for-the-price-of-one sales pose problems for linguistic theories because they behave like words, but clearly include syntactic phrases. The main aim of this thesis is to analyse two types of phrasal compounds in Cinyanja, the verb-noun (VN)-compounds such as cigona-mubawa, 'drunk' (lit. 'it sleeps-in-bar') and associative (Assc)-compounds such as njinga ya moto, 'motorbike' (lit. 'bicycle of fire'). I argue that these two types of compounds differ in whether or not the phrasal parts inside the compounds are accessible to syntactic rules, i.e. whether or not they are subject to the Lexical Integrity Hypothesis. I show that the phrasal part of VN-compounds obeys
Lexical Integrity and is not accessible to syntactic rules such as modification, extraction, or gapping. From this, I conclude that VN-compounds are formed by a process called reification (Harley, 2009), where a phrasal structure is reanalysed as a simple root. In contrast, the phrasal parts inside Assccompounds are accessible to syntactic rules such as pronominalisation, modification and gapping, in violation of Lexical Integrity. I, therefore, analyse the Assc-compounds as (partly transparent) phrasal idioms. My analysis is based on previous work by Carstens (1991, 2008, 2018) on the syntax of Bantu
noun phrases, Harley’s (2009) analysis of phrasal compounds in the Distributed Morphology
framework, and Jackendoff’s (1997) theory of lexical licensing, as applied to phrasal idioms. My results inform our understanding of the processes of compound formation as well as the analysis and lexical representation of phrasal idioms.
... What these studies have not addressed is the nature of the noun class prefixes and locative prefixes inside the VN-compounds in Cinyanja. Similarly, studies on associative constructions in Bantu languages have discussed the interpretation and the syntax of these constructions (Mugane, 1997;Delfitto et al., 2011) but none of these studies has discussed the nature of the associative marker in the associative compounds in Cinyanja. In this thesis, I address this gap. ...
... in the examples below: In the compound words above, the two nouns in (1a) are njinga 'bicycle and moto 'fire' linked together by an associative marker ya 'of' while in (1b) the two nouns are mmisiri 'expert' and nyumba 'houses' linked together by the associative marker wa 'of'. These nominal compounds are mostly referred to as associative nominals or associative constructions in the Bantu literature (Mugane, 1997;Delfitto et al., 2011;Kula, 2012), and they seem to have the same structure as regular associative nominal phrases which are generally syntactic as shown in example (2) because, by assumption, they include the agreeing associative marker. Further, Delfitto et al. (2011) argue that such associative constructions are syntactic because they allow modification of internal constituents. ...
... These nominal compounds are mostly referred to as associative nominals or associative constructions in the Bantu literature (Mugane, 1997;Delfitto et al., 2011;Kula, 2012), and they seem to have the same structure as regular associative nominal phrases which are generally syntactic as shown in example (2) because, by assumption, they include the agreeing associative marker. Further, Delfitto et al. (2011) argue that such associative constructions are syntactic because they allow modification of internal constituents. ...
Phrasal compounds such as two-for-the-price-of-one sales pose problems for linguistic theories because they behave like words, but clearly include syntactic phrases. The main aim of this thesis is to analyse two types of phrasal compounds in Cinyanja, the verb-noun (VN)-compounds such as cigona-mubawa, 'drunk' (lit. 'it sleeps-in-bar') and associative (Assc)-compounds such as njinga ya moto, 'motorbike' (lit. 'bicycle of fire'). I argue that these two types of compounds differ in whether or not the phrasal parts inside the compounds are accessible to syntactic rules, i.e. whether or not they are subject to the Lexical Integrity Hypothesis. I show that the phrasal part of VN-compounds obeys Lexical Integrity and is not accessible to syntactic rules such as modification, extraction, or gapping. From this, I conclude that VN-compounds are formed by a process called reification (Harley, 2009), where a phrasal structure is reanalysed as a simple root. In contrast, the phrasal parts inside Assc-compounds are accessible to syntactic rules such as pronominalisation, modification and gapping, in violation of Lexical Integrity. I, therefore, analyse the Assc-compounds as (partly transparent) phrasal idioms. My analysis is based on previous work by Carstens (1991, 2008, 2018) on the syntax of Bantu noun phrases, Harley’s (2009) analysis of phrasal compounds in the Distributed Morphology framework, and Jackendoff’s (1997) theory of lexical licensing, as applied to phrasal idioms. My results inform our understanding of the processes of compound formation as well as the analysis and lexical representation of phrasal idioms.
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.
This is a book about the meanings of words and how they can combine to form larger meaningful units, as well as how they can fail to combine when the amalgamation of a predicate and argument would produce what the philosopher Gilbert Ryle called a 'category mistake'. It argues for a theory in which words get assigned both an intension and a type. The book develops a rich system of types and investigates its philosophical and formal implications, for example the abandonment of the classic Church analysis of types that has been used by linguists since Montague. The author integrates fascinating and puzzling observations about lexical meaning into a compositional semantic framework. Adjustments in types are a feature of the compositional process and account for various phenomena including coercion and copredication. This book will be of interest to semanticists, philosophers, logicians and computer scientists alike.
This chapter deals with the relevance of some cases of prepositionless genitives in Romance for a general theory of genitive assignment. It examines the peculiar properties of a specific class of Italian N+N compounds, identifying in the juxtaposition genitives of Old French, Occitan, and crucially Old Italian, the possible diachronic antecedent for such construction, which-it argues-is strictly intertwined with Romance construct state. © individual chapters their various authors 2009. All rights reserved.
The present collection includes papers that address a wide range of syntactic phenomena. In some, the authors discuss such major syntactic properties as clausal architecture, syntactic labels and derivation, and the nature of features and their role with respect to movement, agreement, and event-related constructions. In addition, several papers offer syntax-based discussions of aspects of acquisition, pedagogy, and neurolinguistics, addressing issues related to case marking, negation, thematic relations, and more. Several papers report on new findings relevant to less commonly investigated languages, and all provide valuable observations related to natural language syntactic properties, many of which are universal in their implications. The authors challenge several aspects of recent syntactic theory, broaden the applicable scope of others, and introduce important and provocative analyses that bear on current issues in linguistics.
This landmark volume is the first work specifically designed to explore the extent to which striking surface morpho-syntactic similarities between Bantu and Romance languages actually represent similar syntactic structures. In particular, it explores the timely and much debated issues of verbal morphology and agreement, the structure of DPs, and word order/information structure, with the goal of providing a better understanding of the structure of the different languages investigated, and the implications this holds for syntactic theory more generally. All of the papers draw on data from both Bantu and Romance languages, providing a framework for much-needed further comparative research on the nature of linguistic structure, its diversity and constraints, and the implications this has for learnability/acquisition. The volume also provides an important precedent for incorporating insights from Bantu linguistic structure into mainstream of syntax research.
I propose a phrasal-movement analysis of word order in Hebrew and Arabic noun phrases. I argue that the positioning of nominal modifiers with respect to the head of simple nouns cannot be adequately handled by an N-raising derivation. This is so because nouns fail to move as heads in the grammars of Hebrew and Arabic. The only heads that seem to be able to move are those which do not assign genitive case. When modifiers (adjectives, numerals, demonstratives, etc.) appear to the right of the noun, it is the noun phrase itself which has raised to the left of the modifier, moving from specifier to specifier and pied piping all the material on its right. In this manner, inverse or mirror image ordering of post-nominal material is accounted for.Some modifiers have the option of appearing pre-nominally as well as post-nominally. I argue that these elements can be merged either as heads governing their complement and assigning genitive case to their sister or to the specifier of their sister, or as specifiers of projections containing a phonetically null head. In the former case, the modifier’s sister is frozen in position, being licensed by genitive case. In the latter case, XP-movement of the sister of the null head is triggered, and agreement or concord is manifested on the modifier. I develop a configurational theory of agreement, which accounts for these phenomena. Many of the signature properties of the Semitic Construct-state fall out from the phrasal movement analysis.Finally, action nominalizations are dealt with. I provide arguments for a syntactic derivation in which a VP is associated with a nominalizing head configured in the clausal hierarchy as a functional head. I show that the position of the nominalizing head is above certain aspectual projections and lower than others. This explains why some adverbials, but not others, may occur inside a derived nominal.
Possessive nouns phrases are minimally made up of at lest two Noun phrases (NPs) that are possessively ralated to each other. In this article we argue that the syntactic order in which these noun phrases appear is not as rondom as it may look on the surface. Semantic constrains in a large measure, determine the syntactic order of the NPs constituting possessive phrases. Some such constrains include the Personal Animacy (PAH,) generic and par-whole relationships of nouns among others that will be explored in the article. It shall also be argued that the said semantics are not syntactic rules as such but are strong cross-lingustics tendencies to which the linear ordering of the possessive phrase more often than not conforms and especially within the performance as opposed to the competence domain of language ability