Concepts of structural underspecification in Bantu and Romance

Abstract

The paper explores parallelisms between Bantu (specifically Otjiherero) and Romance (through Latin and Spanish) with respect to left and right peripheries, and subject and object clitics. The analysis is formulated in Dynamic Syntax (DS, Cann et al. 2005) and centrally involves notions of structural underspecification. Through providing detailed analyses of different word order possibilities in the Bantu and Romance languages discussed, we show how DS concepts of structural growth over initially underspecified tree relations, such as the building of linked structures and unfixed nodes, provide a uniform basis for analysis for word order variation across the two language groups. We then extend our analysis to include Bantu subject/ object markers, which we analyse by employing the same formal tools as used in the analysis of Romance (object) clitics, namely unfixed nodes which have to be construed within a tightly locally restricted domain. Empirical support for our analysis comes from restrictions on the presence of object markers in passive and locative inversion constructions in Otjiherero, which we show to follow from independent constraints of the availability of unfixed nodes within a given domain. The analyses of Bantu and Romance presented show that despite differences in surface morphology between the two language groups, both exhibit a striking parallel with respect to the way lexical information and general structure building principles of DS interact. The difference between Romance clitic systems and the agglutinative morphology of Bantu subject and object makers is thus seen to be comparatively superficial, while the DS analysis brings out the strong structural parallelism between the two language groups. 1

The Bantu-Romance-Greek connection revisited: Processing constraints in auxiliary and clitic placement from a cross-linguistic perspective

“…The introduction of fixed structure by the auxiliary follows the analysis of the English copula be provided by Cann (2011) in which be is assumed to project fixed predicate-argument structure. This account also reflects the historical origin of Bantu auxiliary forms in main verbs which are modelled in similar terms (see, for example, Marten et al 2008 andGibson 2012 for DS analyses of Bantu auxiliaries from this perspective). The fixed structure introduced by the auxiliary enables the establishment of a fixed tree node address for the information introduced by the infinitival verb.…”

“…In this vein, researchers have observed similarities in Bantu and Romance languages with regards to weak object pronominals (Labelle 2008;Marten et al 2008), the structure of the DP (Carstens 2008;Zamparelli 2008), as well as information structure (Costa & Kula 2008;Frascarelli 2008). This paper continues this endeavor, extending the connection from yet another perspective: we show how different processing strategies that can be encoded as lexical triggers for parsing, appear in similar ways in Bantu, Romance and -adding in this respect a third language family to the emerging picture -Greek.…”

Information structure and agreement: Subjects and subject agreement in Swahili and Herero

“…Particular structural choices can be, or are even routinely associated with specific contexts, and then lead to information structure effects. However, by defining information structure effects as resulting from the interaction of structural choices and context information, gradient, underdetermined pragmatic effects are expected, as could be seen both in Swahili and Herero, and thus the approach presents a more nuanced conception of the syntax-pragmatics interface (see Marten, 2007;Marten et al, 2008;Cann et al, 2005 for further discussion).…”

“…This is possible since the subject marker decorates a locally unfixed node, and this will often, but not necessarily merge with the logical subject node in the tree. However, in locative inversion, and indeed in passives, the locally unfixed node merges not at the subject node, but at the object node, thus allowing for the decoration of the subject node with information from the post-verbal NP in locative inversion, or with a weak 'agent' annotation supplied by the passive morphology in passives (see Marten et al, 2008 for a detailed discussion of this analysis). From the point of view of the present analysis, what is important is that in locative inversion, the locative subject marker is routinely associated with a tree update process in which the logical subject is annotated with information which is only provided after the predicate has been constructed.…”

“…Also note that strictly speaking, the class 1 subject marker should be annotated with a case-like feature as it differs, in contrast to other classes, in morphological shape from the corresponding object marker -m-. 7 SeeCann et al (2005) andMarten et al (2008) for further discussion of this analysis. The analysis might well be extended to other TAM markers in Bantu, or to auxiliary verbs in Zulu, as well as to auxiliaries for example in English, but space prevents further discussion of this.…”