Publishing Partner: Cambridge University Press CUP Extra Publisher Login

New from Cambridge University Press!


Revitalizing Endangered Languages

Edited by Justyna Olko & Julia Sallabank

Revitalizing Endangered Languages "This guidebook provides ideas and strategies, as well as some background, to help with the effective revitalization of endangered languages. It covers a broad scope of themes including effective planning, benefits, wellbeing, economic aspects, attitudes and ideologies."

We Have a New Site!

With the help of your donations we have been making good progress on designing and launching our new website! Check it out at!
***We are still in our beta stages for the new site--if you have any feedback, be sure to let us know at***

Review of  Complex Predicates

Reviewer: Aroldo Leal de Andrade
Book Title: Complex Predicates
Book Author: Leila Lomashvili
Publisher: John Benjamins
Linguistic Field(s): General Linguistics
History of Linguistics
Issue Number: 23.557

Discuss this Review
Help on Posting
AUTHOR: Lomashvili, Leila
TITLE: Complex Predicates
SUBTITLE: The syntax-morphology interface
SERIES: Linguistik Aktuell/Linguistics Today 174
PUBLISHER: John Benjamins
YEAR: 2011

Aroldo Andrade, State University of Campinas, Brazil


This book explores the morphosyntax of a subset of complex predicates, involving
causative and applicative constructions in three polysynthetic languages of the
South Caucasian (Kartvelian) language family. Lomashvili’s monograph applies the
Distributed Morphology framework in order to explain the different possibilities
of complex predicate formation in these languages.


The book is divided into four chapters. Chapter 1 is intended to familiarize the
reader with the preliminaries to the study (introduction, basic theoretical
assumptions and specific grammatical information). Chapter 2 presents the types
of causative constructions and their derivations, with focus on the possibility
of causative alternation by syntactic-semantic verb class. Chapter 3, named
“Applicatives as complex predicates,” discusses different applicative
constructions, their derivations and productivity restrictions. Finally, chapter
4 includes the conclusions of the study.

Chapter 1 begins by enumerating the research goals. The main objective of the
book, the discussion of different causative and applicative predicates,
understood as complex predicates (in the sense of Baker 1996, i.e., verbal
structures containing at least two morphemes, each marking a phrasal argument in
the theta-grid) is framed by the author’s interest in the mismatches that arise
between the syntax-semantics and the morphology of these constructions.
Theoretical assumptions are presented from section 1.2 to 1.6. First,
Distributed Morphology is contrasted with the lexical-semantic approach, in that
the syntax of word and sentence derivations is treated in the same syntactic
structure in the first approach. The assumptions concerning the argument
structure of causative and applicative constructions are then presented: the
projection of a Cause head which adds an event argument to the structure (but
only increases the valency of the verb upon its bundling with a Voice feature);
and the projection of an Applicative head, either below the VP (when the
argument expresses a static possession, recipient, or source), or above it (when
the argument undergoes the effects of an event expressed by the verb). Apart
from these assumptions, driven from Pylkkännen’s (2002) work, the author also
considers Cuervo’s (2003) additional idea that psychological verb constructions
may be interpreted as high applicatives. Another facet of Pylkkännen’s (2002)
work, the variation regarding the type of complement that Cause takes (root, VP
and vP), is also assumed. The mentioned configurations vary regarding the
creation of mono- or bi-eventive structures, which can be verified by checking
the scope of VP-modifying adverbials; and the same reasoning is then applied to
low and high applicatives. Other more basic theoretical assumptions are then
presented, such as the Distributed Morphology framework, the notion of syntactic
phases, the connection between contextual allomorphy and phasehood (Embick
2010), and the different flavors of little vs (v-do, v-be and v-go, Cuervo
2003). Sections 1.6 and 1.7 present specific grammatical information about
Kartvelian syntax: the status of templates and Case and agreement. This
presentation is centered on Georgian, a language with three positions for
morphemes before the verbal root and up to eight positions after it. In order to
account for its order, Lomashvili rejects Baker’s Mirror Principle, which would
be too restrictive to account for the data, especially with respect to the
post-verbal morphemes, with some of these recurring for no obvious semantic
reason (cf. ‘-eb’ and ‘-in’):

(1) The perfective series verb
‘I would have caused X to make Y.’

Discontinuous bleeding (Noyer 1997), a central theme of the book, proposes that
a single terminal node may be associated with two positions of exponence. This
can also be observed in (1), where the subject-person morpheme occurs in the
second slot, whereas the object-person morpheme occurs in the final slot. The
author then discusses the functioning of the screeve system (conjugation pattern
for one specific combination of Tense, Aspect, and Mood), the possibilities of
case marking and the properties of the aspectually-conditioned split ergativity.
The contexts where the inverse agreement pattern shows up are relevant, i.e.,
all those where subjects are assigned dative case (the perfective series,
adversity causatives, causatives of internally-changed verbs, some applicatives
etc.). Finally, section 1.8 includes an outline of the following chapters.

Chapter 2, “The morphosyntax of causative alternations,” begins with the
enumeration of assumptions already outlined in chapter 1. Lomashvili adopts
Pylkkännen’s (2002) approach in that the Cause and Voice heads may project
separately, with the external argument being realized in the SpecVoiceP. This
separation permits one to dissociate the presence of a Cause head from the
projection of an external argument, as happens with adversity causatives. At
this point the text describes the main types of verbs (sections 2.3 to 2.5).
First, causatives of inchoative verbs are formed by either an ‘a-’ or zero
morpheme inserted in the Cause head. The insertion of such a Vocabulary Item is
root-conditioned. The ‘a-’ morpheme co-occurs with non-syllabic roots, a fact
that indicates the Cause head is attached at a low position in the clause.
Second, in causatives of unergative verbs the Cause head is cyclic due to its
phase-selecting attachment and therefore shows only the ‘a-’ morpheme, selected
by a transitive v. In this sense, Lomashvili adopts Hale & Keyser’s (2002)
theory of conflation in order to explain unergatives formed by noun or adjective
incorporation. For iterated causatives, a second Cause head is inserted in the
structure, realized as ‘-in’, which co-occurs with ditransitive verbs. These
adjective-incorporated causatives with the reading “make X do V” are
bi-eventive, as the tests with VP-adverbials and depictive modifiers attest.
Third, with causatives of transitive verbs, two causative morphemes are realized
(‘a-’ and ‘-(ev)in’), but only in the perfective series (“make X do V” sequences
do not normally show an additional causative morpheme). The occurrence of
‘-evin’ instead of ‘-in’ is attributed to an erstwhile thematic marker ‘-eb’
turned into ‘-ev’ due to root-conditioned allomorphy, i.e., with non-syllabic
roots. The presentation then turns to the discussion of specific constructions
(sections 2.6 to 2.8). First, the discussion on adversity causatives is
important to illustrate that a causative does not necessarily involve a new
argument, its main feature being the expression of a causing event: in this
construction, the Voice head marked [nonactive] feeds the first Cause head to be
zero, and the second to be expressed as ‘-(ev)in’. Due to the fact that the
causer (external argument) is not projected, the corresponding sentences behave
as mono-eventive. Second, constructions with the meaning “make X pretend to be
Adj/N” are taken up. These include the grammaticalized item “pretend” (realized
as ‘-un’ or zero, depending on the number of syllables) and involve the
incorporation of a noun or an adjective, with a derivational affix and a
reflexive morpheme. The two causative heads are merged over the vP corresponding
to the “pretend” morpheme, implying a bi-eventive interpretation. Third, the
behavior of psych verbs is analyzed. These are classified into three types,
according to the expression of a state, dynamic passive, or activity. Only this
last class can alternate with both causatives and passives. In the causative of
activity psych verbs, only one or two Cause heads may show up, the first option
indicating a reflexive causative (“make X love …self”). Finally, there is a
section with the basic facts on causative predicates in the related languages
Mengrelian and Svan, with data collected from grammars and dictionaries.

Chapter 3 explores the morphosyntactic properties of applicative constructions.
The morphological realization of the Appl head may not be uniform due to
contextual allomorphy. In 3.1 the author presents the theoretical goals and
assumptions, together with an outline of the chapter. From sections 3.2 to 3.6,
low applicatives are analyzed in their manifold cases. First, recipient and
source applicatives are presented in terms of allomorphy among ‘i-’, ‘u-’ and
zero, according to substantive features, except when the verbs are basic
ditransitives. The third semantic type, low applicatives of stative possession,
is presented as being derived from activity state verbs, where a dative can be
replaced by a genitive argument. Second, applicatives occurring with
unaccusative and inchoative verbs are set apart due to their novelty with regard
to previous accounts in the literature. For instance, Cuervo (2003) mentions
affected applicatives formed from inchoatives, which are higher than v-be. For
Lomashvili, Georgian has examples of true low applicatives with inchoatives,
where the nominative DP is not projected as an external argument, but as a
theme, and the applied argument occurs in SpecApplP, as usual. In these
structures, the Voice head is realized as ‘e-’, and the Appl head is realized as
zero. The best solution for this would be a dependency between the [nonactive]
feature in the former, and Vocabulary Insertion in the latter. Third, Noun- and
Adjective-Incorporated predicates also realize low applicatives with a dative
benefactee subject and a nominative theme argument. Both the Appl head (‘i-’ or
‘u-’, according to the person feature) and the Voice head (‘-d’) are realized.
Fourth, reflexive applicatives show a benefactive relation between the theme and
an external argument. This expected relation is solved by means of the
postulation of a PRO in SpecApplP, coindexed with the external argument. An ‘i-’
reflexive morpheme is inserted in the same preverbal slot used for Voice and
Appl. Fifth, a section is devoted to possessor datives, in which a verbal root
is inserted under v-do, v-be or v-go. It is noticeable that some stative verbs
do not accept a dative, and allow instead a genitive. The explanation for this
restriction is that only predicates in the form of activity psychological verbs
allow the dativization. In section 3.7, four-place predicates are analyzed,
either as a combination of high and low applicatives in the same structure, or
as a high applicative together with a PP. A correlation between case marking (or
selection of postposition) and semantic interpretation is put forward: inanimate
datives are marked with ‘-ze’, indicating a location, whereas animate datives
are marked with ‘-tan’, indicating a recipient. In section 3.8, high
applicatives, i.e., applicative arguments that relate to the event expressed by
the verb, are discussed under three types: state unaccusatives, other
unaccusatives (denoting a change of state, and therefore under v-go) and dynamic
activity verbs. However, not only high applicatives are described, but also the
very distinction between high and low applicatives. A last section includes
comparative data about applicatives on the related languages Mengrelian and Svan.

Chapter 4 presents the main conclusions, where the complex predicate status of
the constructions is reaffirmed, and the main theoretical solutions of the book
are summarized.


This is a valuable book, for it includes a rich discussion on data from
different Caucasian languages. The analysis of under-represented languages is
always very welcome in generative linguistics, for it broadens the scope of the
theories and at the same time it permits us to test these in typologically
diverse systems. Another very interesting aspect of the book is the questioning
of previous proposals, necessary to accomplish the challenge taken up by the
author: to give a coherent formal account of complex predicates in Georgian, and
to apply these ideas to Mengrelian and Svan, whenever possible. Therefore the
following criticisms are not intended to invalidate the bulk of the proposals,
but to point out some weaknesses arising from the lack of clarity and
completeness of the presentation.

Although the author does introduce theoretical background to her readers, in
some points the presentation lacks clarity. First, in the literature review,
some works are presented as assumptions; but in fact they are theoretical
background that is still worked out during the proposal, such as in the
presentation of case valuation of low applicatives on page 116, without having
presented the proposal based on Lomashvili and Harley (2011) before. Second,
there are contradictions between some parts of the presentation, at least at
first glance: “High and low applicatives show a distinction in terms of
mono-/bi-eventiveness” (page 13) and “The simpler event structure of high and
low applicatives and the size of the complement that both Appl heads take means
that the resulting structure will always show mono-eventive properties with
respect to adverbial scope” (page 14). Third, there are minor typing and
numbering errors (mismatches between the examples and the references to them)
that may also render the text less straightforward for those less familiar with
the latest theoretical developments. In some cases, as on page 151, correlations
between features and exponents are switched (in the case of the morphemes ‘-ze’
and ‘-tan’).

The argumentation could also be more complete and stronger. First, although
expressly dismissed by the author as a task outside the scope of the book, the
reader feels as a relevant gap that no postsyntactic rule is presented in order
to explain the correct linearization of the morphemes. Second, some other
aspects of the book should be better explored, such as decisions regarding the
projection of arguments. For instance, in the representation of the causative of
an unergative (“X makes Y scream”) on page 50, the only argument of ‘scream’ is
projected as a theme sister of RootP, although it would usually be projected as
the external argument in SpecvP. In the same vein, no explanation is given for
why the dative causer argument is included under a supplementary SpecApplP
together with the genitive argument marked with a postposition ‘Gias-tvis’ (“for
Gia”), on page 116. Third, more explanation should be given with respect to the
minimalist framework assumed, as for page 17. One example is the apparent
violation of the Inclusiveness Condition deriving from the adoption of the
Principle of Phi-activation in Lomashvili and Harley (2011), according to which
“a probe acquires an active phi-feature bundle when it is merged into a domain
which contains a Case-active DP with marked phi-features” (page 118). Other
examples relate to the use of coindexation of PRO to the external argument in
order to account for reflexive applicatives in section 3.7, and restriction to
the projection of multiple specifiers, used to explain why stative verbs would
disallow the projection of applied arguments.

In conclusion, the problems pointed out above make this book more difficult to
read and also give a feeling of an unfinished task. Nevertheless, I consider
these observations to be mitigated in face of the difficulties related to the
application of this formal approach to under-represented languages. I recommend
it to all linguists interested in either polysynthetic languages or in causative


Baker, M. 1996. The Polysynthesis Parameter. New York: Oxford University Press.

Cuervo, M. 2003. Datives at Large. PhD dissertation. Cambridge, MA:
Massachusetts Institute of Technology.

Embick, D. 2010. Localism versus Globalism in Morphology and Phonology.
Cambridge, MA: The MIT Press.

Hale, K. & Keyser, S. J. 2002. Prolegomenon to a Theory of Argument Structure.
Cambridge, MA: The MIT Press.

Lomashvili, L. & Harley, H. 2011. Phases and templates in Georgian agreement.
Studia Linguistica 65:3, p. 233-267.

Noyer, R. 1997. Features, Positions and Affixes in Autonomous Morphological
Structure. New York: Garland.

Pylkkännen, L. 2002. Introducing Arguments. PhD dissertation. Cambridge, MA:
Massachusetts Institute of Technology.

Aroldo Andrade is a postdoctoral fellow at the State University of Campinas. His research focus is the relation between morphosyntactic change and information structure. For his PhD he studied complex predicates formed by raising/control and Exceptional Case Marking verbs, in connection with the realization of clitic climbing in the history of Portuguese.