Logo: MPI
Valency classes in the world's languages
August 14-15, 2013


Michela Cennamo (University of Naples Federico II): The syntax and semantics of anticausativization in Italian: diachronic insights and synchronic variation
This paper discusses the role played by the interplay of the aspectual template of verbs, the verb’s inherent meaning (the 'root'), and the nature of the –P subject (e.g., animacy and control) in determining the distribution of the different strategies available in contemporary Italian and early Italo-Romance (Old Florentine) to mark anticausativization, the reflexive pattern and the active intransitive. [Read more...]


Bernard Comrie & Zaira Khalilova (MPI-EVA): The Antipassive Alternation in Bezhta
The so-called antipassive in Bezhta presents interesting analytical problems with regard to the question of valency alternation. Canonical transitive (and ditransitive) verbs have a canonical antipassive, where the A (ergative) of the basic form corresponds to the S (absolutive) of the antipassive and the P (absolutive) of the basic form (or the T of the ditransitive) shows up in the instrumental. Unergative verbs follow a parallel pattern with respect to their sole unergative argument, which appears in the ergative in the basic form and in the absolutive in the antipassive. But other intransitive verbs have their single argument in the absolutive in both the basic voice and in the antipassive, thus calling into question whether there is a valency alternation – perhaps a vacuous alternation? – in this case. What unites these "antipassives" is their semantics, namely iterative aktionsart. But even here there is one exception: The antipassive of transitive 'wash' has reflexive ('wash (oneself)') rather than iterative meaning.


Denis Creissels & Alain-Christian Bassène (University of Lyon & University of Dakar): Valency patterns for bivalent verbs in two West African languages: Mandinka (Mande) and Jóola Banjal (Atlantic)
The valency patterns of bivalent verbs that do not encode prototypical transitive events may be identical to the valency pattern <A, P> whose definition refers to verbs encoding prototypical transitive events, or different, and the proportion of bivalent verbs with valency patterns other than <A, P>  shows important cross-linguistic variations. The main questions that arise are : (1) Among the bivalent verbs that do not encode prototypical transitive events, is it possible to define semantic types particularly predisposed to being used with the same valency pattern <A, P> as verbs encoding prototypical transitive events, and others that tend to occur in distinct valency patterns? (2) Are there semantic regularities in the choice of a particular coding (A-like, P-like, or other) for the arguments of bivalent verbs that do not encode prototypical transitive events? In our talk, we will compare the situation of two West African languages (Mandinka and Jóola Banjal) in this respect. Jóola Banjal (Atlantic) illustrates the type of situation found in most of the language families of West Africa, with a very limited set of bivalent verbs using valency patterns other than <A, P>, whereas Mandinka (Mande) belongs to a language family characterized by more diversity in the valency frames used by bivalent verbs.


talk cancelled

Nicholas Evans (Australian National University): Experiencer objects in Nen (Southern New Guinea)
In this presentation I discuss the subclass of Nen verbs which encode ‘experiencer objects’ – a phenomenon widespread in the languages of Sahul, both Papuan (e.g. Kalam (Pawley et al 2000), Yagaria (Windschuttel 2012, Sougb (Reesink 2005)) and Australian (e.g. Iwaidja and Mawng (Evans 2004, Singer 2007), Murrinh-Patha (Walsh 1987)). They are also found in some African languages such as Yoruba (Atoyebi 2013). [Read more...]


David Gil (MPI-EVA): A Wolf in Sheep's Clothing: Valency Classes and Valency Preference Classes in Jakarta Indonesian
Part 1 presents a brief sketch of valency in Jakarta Indonesian, arguing that it has but a single valency class encompassing almost all words in the language, including those refering to activities, properties, things, and entities of other ontological classes. (This is the wolf.)

Part 2 shows that if one differentiates between constructions that are preferred and frequent and those that are dispreferred and infrequent, distinct valency classes emerge, resembling the valency classes of more familiar languages. (This is the sheep's clothing.)

Part 3 attempts to place Jakarta Indonesian in typological space. Results from an ongoing cross-linguistic experiment examining various aspects of clause structure suggest that while Jakarta Indonesian is located somewhat towards the flexible end of the scale with regard to valency classes, it is actually less of a typological outlier than languages of the Standard Average European type, which are exceptional with respect to the degree of rigidity of their valency classes. (There may be more wolves out there.)

Part 4 suggests that the valency preference classes of Jakarta Indonesian may constitute a model for the diachronic and perhaps even phylogenetic precursors of the more rigid valency classes of most other languages. (The wolf with sheep's clothing gradually becomes a sheep.)


Csilla Kász (University of Kiel): Passives in Modern Standard Arabic
Modern Standard Arabic uses inflectional passive paradigms as well as derivational means of forming passive stems from active stems (passives are most frequently formed by the derivational pattern VII and occasionally by V and VIII).
The aim of my paper is to explore syntactic, semantic and pragmatic differences between these valency changing devices, with a special focus on those lexemes, to which both processes – flection and derivation – can be applied.


Andrej Malchukov (University of Mainz): Valency classes cross-linguistically: parameters of variation
The first part of the talk addresses the issue of universality of valency classes, identified by coding frames and valency alternations (both coded and uncoded). Classification by coding frames is clearly not uniform cross-linguistically since the inventory of coding frames (e.g. case inventories) varies widely across languages. Nevertheless, both variation and universal tendencies can be captured by a semantic map approach (semantic maps can also be seen as multidimensional versions of Tsunoda’s (1981) Transitivity Hierarchy). Uncoded alternations can be represented on a semantic map as areas of overlap between individual constructions (e.g., transitive vs intransitive). Similar hierarchies/semantic maps can be profitably applied for coded alternations (“voice alternations”) as well (cf. Wichmann et al. forthcoming), but in this case cross-linguistic variation is greater, as transitivity is just one of the several factors predicting availability of voice/valency-changing constructions. One complicating factor is the phenomenon of ‘voice ambivalence’, which refers to polysemies of voice markers dependent on the verb type to which these markers apply (e.g., causative-passive or causative-applicative polysemies). Voice ambivalence, discussed in the second part of the talk, is not only challenging for valency classification but also deserves to be studied in its own right. It will be argued that an approach relying on the concept of (local) markedness and shared syntactic features can go a long way in explaining recurrent patterns of such polysemies.


talk cancelled

Igor Nedjalkov (St. Petersburg): Verb valency classes in Evenki in the comparative perspective
The present paper presents an overview of valency patterns and valency alternations in Evenki, a North-Tungusic language, from a comparative perspective. The paper is structured as follows. Section 2 provides information on case paradigms in Tungusic languages (TLs) relevant to issues of valency classification. Section 3 describes main valency patterns in Evenki. Section 4 addresses uncoded argument alternations (“case alternations”), while section 5 describes verb-coded alternations (“voice alternations”) in Evenki in the comparative perspective in TLs.
[Read more...] [additional material]


Ronald Schaefer & Francis Egbokhare (Southern Illinois University Edwardsville & University of Ibadan): Argument, adjunct or neither in Emai serial verb constructions
This investigation of southern Nigeria’s Edoid language Emai examines a subset of its serial verb constructions. They have in common the identification and coding of a co-participant that is obligatorily human but not required by core verb argument structure. All reflect classic serial verb properties (Aikhenvald and Dixon 2006): a verb sequence sharing tense, aspect and polarity under a single intonation contour acting as a single predicate with no overt marking of clausal dependency.
[Read more...]


Fernando Zúñiga (University of Bern): Some valency-neutral applicatives in the Americas
The literature on voice has occasionally noted the existence of applicatives that appear not only to add new objects to the syntactic core but also to add non-objects (or at least non-canonically marked objects) and even, most strikingly, to clearly reduce the valency of the predicate. [Read more...]