• Research
• Research Clusters
• Research Centres & Groups
• Staff Research Interests
• Research Students
• Research Strategy
• Recent Books
• Current Research Activities
• Research Activities Archive
• Search Activities by keyword

Referential Hierarchies in Morphosyntax

«Back

Summary: The RHIM project, funded by the AHRC as part of the ESF Babel initiative, explores morphosyntactic systems that are based on a hierarchy of referents - first and second person ranking over third, humans over non-humans, and known referents over unknown ones. This hierarchy is known to influence the structure of grammatical relations (the basic "who does what to whom in an event"), giving rise to e.g. inverse morphology or differential argument marking. There have only been few comparative studies on these phenomena, and most languages displaying these patterns (found e.g. in the Americas, the Himalayas, and Australia), are seriously endangered. Based on fieldwork and documentation corpora, we aim at a better understanding of such systems from a typological and diachronic perspective.

Project Description

The complete project consists of five sub-projects, namely:

1. "Differential agreement vs. differential case" (Balthasar Bickel)

2. "Sahaptian and the evolution of hierarchical systems" (Spike Gildea)

3. "The Movima inverse from a cross-linguistic perspective" (Katharina Haude)

4. "Hierarchical ranking and argument encoding in three participant clauses" (Anna

Siewierska)

5. "Mapudungun and Blackfoot: inverse alignment with special attention on three-participant

clauses" (Fernando Zúñiga).

The Lancaster contribution will involve mainly project 4 which is described below:

a) Aims and objectives

This IP seeks to examine the role played by referential hierarchies in the encoding of the non-agentive arguments (NAAs) in three participant clauses. The investigations of three participant clauses carried out to date, suggest that the encoding of their NAAs is less often affected by the referential properties of the arguments than in the case of dual argument clauses (see. e.g. Haspelmath 2005; Malchukov et al 2007; Siewierska & Bakker 2007; Haspelmath 2007, Bickel et al 2008). In the case of clauses involving the verb „give", for example, most languages extend the treatment of the transitive patient argument (P) either to that of the recipient (R), i.e. exhibit secundative alignment, or to that of the theme (T), i.e. manifest indirective alignment, and less often treat both the T and R in the same way as the P (see e.g. Siewierska & Bakker 2007). Nonetheless, we do come across languages in which some aspect of the encoding of the T and R, be it order, case marking or agreement marking, are sensitive to the hierarchical status of the two. For example, in the Yuman language Jamul Tiipay whether agreement in ditransitive clauses is with the T or the R depends on which is higher on the person hierarchy of 1 > 2 > 3. Thus, in (1a) since the R outranks the T the person prefix marks the R, while in (1b) the T outranks the R, and consequently it is the T that is marked.

(1) Jamul Tiipay

a. xikay ny-iny-ma

some 1:2-give-prom

`I'll give you some.'

b. nyaach maap Goodwill ny-iny-x

I:sub you Goodwill 1:2-give-irr

`I'm going to give you to Goodwill.'

In French it is the order of the R and T proclitics relative to the verb which is hierarchically determined. When the R is first or second person, the R precedes the T but with third person Rs the T precedes the R. Compare (2a) with (2b).

(22) French

a. IL me=le=donne.

he 1sg(R)=3sg(T)=give

'He gives it to me.'

b. IL le=lui=donne.

he 3sg(T)=3sg(R)=gives

'He gives it to him.'

And in Araki, an Oceanic language, the patterns of agreement and case marking depend on the animacy of the T. If the T is inanimate, as is generally the case, the R is bound to the verb while the T occurs as the object of the instrumental/oblique preposition ni/ini or lo. But if the T is human it may take priority over the R with respect to attachment to the verb. In such a case the R is marked by a different preposition, namely sa/isa. Compare (3a) with (3b).

(3) Araki

a. Na sile-ko ne-re presin.

I give-2sg obl-some present

'I feel like giving you a present.'

b. Na pa sle-ko sa-n ramare

I seq give-2sg to-cst devil

'I will give you to a devil.'

Furthermore, as argued by Haspelmath (2007), ditransitive clauses also manifest effects

completely parallel to inverse patterns in mono-transitive clauses be it via a different form of marking. Whereas in languages with a direct/inverse distinction in mono-transitive clauses, the inverse, i.e. the situation where the P is referentially higher or equal to the A, is typically expressed via verbal marking (see the examples given in the other IPs associated with this project), the ditransitive inverse, where the T is higher or equal to the R, is expressed via the form of the personal pronouns used for the T and R; in direct marking both the T and R may be reduced forms, while in inverse marking the R occurs in its less reduced form. Note the ungrammaticality of the inverse (4b) in Shambala in which the clitic forms are used for both the T and R as compared to (4c) in which the R is an independent pronoun.

(4) Shambala

a. (1>3) A-za-m-ni-et-ea.

3sg.sbj-pst-3sg.thm-1sg.rec-bring-appl

'S/he has brought him/her to me.'

b. (3>1) * A-za-ni-mw-et-ea.

3sg.sbj-pst-1sg.thm-3sg.rec-bring-appl

'S/he has brought me to him/her.'

c. A-za-ni-eta kwa yeye.

3sg.sbj-pst-1sg.thm-bring to him/her

'S/he has brought me to him/her.'

An alternative solution adopted by some languages, again parallel to what happens in mono-transitive clauses, is to prohibit altogether the expression of some inverse constellations, such as those involving a 1st person T and a 2nd person R, as is the case, for example in Modern Greek.

The above typological observations on the hierarchical effects in ditransitive clauses are based in the main on constructions involving the verb give. This is due to the fact that typically constructions with give are the only ditransitive constructions considered in descriptive grammars. However, if, as argued by Borg & Comrie (1984) and Kittilä (2006), constructions with the verb give are by no means necessarily representative of the ditransitive constructions of a language, let alone of the three participant clauses that it may display, we may well expect to find more or even other hierarchical effects on the encoding of NAAs once we extend our range of enquiry to constructions with predicates which allow for arguments with a different constellation of semantic and referential features.

The present project will concentrate on three participant constructions which, given examples such as those in (1) -(4) above, may be assumed to be most likely to exhibit hierarchical effects of some type, namely constructions in which the two NAAs are human or animate. While two human or animate NAAs are highly atypical of the predicate give, they are considerably more common with predicates such as those in the English examples in (5).

(5) a. He introduced me to his wife/*his wife me.

b. I will present your friend to the panel/*the panel your friend.

c. He recommended me to the jury/* the jury me.

d. You promised the child to me/ me the child.

e. She offered me to him/ him me.

f. They showed him to me/me him

As these examples suggest, in English there is some variation in relation to the encoding possibilities of the T and R with human NAAs (see e.g. Bresnan and Nikitina 2007) most of which conform to the generalizations advanced by Haspelmath (2007; see below). However, it remains to be seen whether his generalizations hold on a cross-linguistic basis and in particular what type of reflections of hierarchical ranking are to be found in ditransitive clauses with human NAAs in languages manifesting mono-transitive inverse systems such as those investigated in the other IPs of this larger project, namely, namely Algonquian, Sahaptian, Caribian, Tupian, Movima. Mapudungun and Kiranti. To the best of my knowledge, there is no study of ditransitive constructions with a special focus on such languages.

This IP will therefore rely in the first place on the language data stemming from the other IP projects in combination with data collected from additional grammar based research. On the basis of a cross-linguistic sample of 200 of the world's languages, the project will establish (i) which languages have basic or even derived predicates which allow for two NAAs participants, (ii) to what extent the encoding and syntactic behaviour of the relevant arguments is subject to hierarchical factors, (iii) whether and if so how the patterns of marking displayed differ from those found in give-based constructions and (iv) in the case of languages with hierarchically based mono-transitive systems whether and how the patterns of marking found differ from those obtaining in mono-transitive constructions. We will be particularly concerned with determining the validity of the three universals proposed by Haspelmath (2007) presented below:

Universal 1: Special ('indirective' or 'dative') R-marking is the more likely, the lower the R is on the animacy, definiteness, and person scales.

Universal 2: Special ('secundative') T-marking is the more likely, the higher the T is on the animacy, definiteness, and person scales.

Universal 3: If a language shows any ditransitive inverse patterns, on the scale

of decreasing harmony of person-role association the upper end is expressed by a

simpler construction, and the lower end is expressed by a more complex

construction.

We will also seek to determine whether:

a. there is any evidence from ditransitive clauses for more than one person hierarchy, i.e. if the 2nd person is ever ranked higher than the 1st, as sometimes suggested, for example, in Algonquian mono-transitive clauses;

b. there are any interactions between person and number and potential distinctions of honorificity in hierarchical rankings

c. ditransitive alignment is primarily predicate as opposed to argument based.

Purpose of Research

Academic Research - Externally Funded - The purpose of the research is to provide a description of syntactic phenomena found mainly in endangered languages. Sensitivity to referential hierarchies is one such phenomenon which is reflected in a wide variety of syntactic structures. The part of the project based at Lancaster concentrates on hierarchical effects in ditransitive clauses.

Project Funder

AHRC

The funding for this project by the AHRC is part of the ESF sponsored BABEL (Better Analyses Based on Endangered Languages)initiative which aims to bring endangered language data to bear on theoretical linguistic issues.

Associated News Stories

International workshop on referential hierarchies in three-participant constructions

Date: 6 June 2011

The Workshop on Referential Hierarchies in Three-participant Construction was held in Lancaster on 20-22 May 2011 at the Lancaster House Hotel. It ... Read more»

Paper at 43nd Societas Linguitica Europea

Date: 3 September 2010

At the 43rd Societas Linguistica Europea meeting to be heldin VIlnus 3-6th September 2010, together with Eva van Lier and Katherina Hauden I will be p ... Read more»

Invited paper at workshop on Local pronouns

Date: 7 June 2010

I will be giving the opening paper entitled Local persons in 3-participant scenarios" at the workshop "Between you and me: local pronouns across modal ... Read more»

Presentation at Workshop on Argument Structure

Date: 27 May 2010

Together with Eva van Lier I will be presenting a paper entitled "Non-prototypical ditransitives" at the Workshop Variation & Change in Argument R ... Read more»

Associated Events

International workshop on referential hierarchies in three-participant constructions

Date: 20 May 2011 Time: 09.00 - 06.00 pm

The workshop is part of the EuroBABEL project on Referential Hierarchies in Morphosyntax (RHIM). It brings together language(-family) specialists a ... Read more»