|EDITOR: Lyngfelt, Benjamin; Solstad, Torgrim
TITLE: Demoting the Agent
SUBTITLE: Passive, Middle and Other Voice Phenomena
SERIES: Linguistik Aktuell/Linguistics Today 96
PUBLISHER: John Benjamins
Amaya Mendikoetxea, Departamento de Filología Inglesa, Universidad Autónoma de
Passives and related constructions involving 'agent demotion' (e.g. middles and
antipassives) have been a core area of research in contemporary linguistics.
These structures are typically analyzed in contrast with active constructions.
The active-passive contrast raises interesting issues concerning the
lexicon-syntax interface and the syntax-discourse interface. Consequently,
different perspectives have been adopted in different theoretical frameworks.
Formal linguists have focused mostly on issues at the lexicon-syntax interface:
the contrast between active and passives (and related constructions) regarding
their semantic representation (event structure, argument structure) and their
syntactic realization. Functionalist approaches, on the other hand, have
concentrated mostly on the expression of discourse participant prominence and
distinctions to do with other aspects of information structure, as well as
specifying the circumstances in which agents can be omitted. The aim of this
volume is, as stated by the editors in the preface, to bring together different
perspectives on passives and related constructions: different theoretical
approaches, different languages and different voice-related phenomena.
This book contains 11 articles based on talks presented at the workshop
'Demoting the agent: Passive and other voice-related phenomena' held at the
University of Oslo in November 2004, which are preceded by an introductory
chapter by the editors of the volume and followed by a Language Index, a Name
Index and a Subject Index.
The introductory chapter ''Perspectives on demotion'' by Torgrim Solstad and
Benjamin Lyngfelt, editors of the volume, is devoted to the definition of agent
demotion, which has been used in many different ways in the literature. In line
with much recent research, the editors adopt a wider approach to agent demotion,
including passivization but also other phenomena in the domain of what is
commonly referred to as middle voice. A key concept in the analysis of agent
demotion related constructions is transitivity in its broad sense (e.g. Hopper &
Thompson 1980), which in turn involves notions such as action, telicity,
volitionality and affectedness. They show that passive, and voice phenomena in
general, belong to two different paradigms. In the 'pragmatic paradigm',
passives are analyzed in relation to information structure, along with other
constructions used to express (relative) prominence: topicalization,
intonational patterns, clefts and so on. In the 'event-semantic' paradigm,
passives are analyzed along with middles and other non-transitive constructions,
in which agents are omitted. The tendency in the literature is to focus on
either one or the other. The editors' belief is that one should take both
perspectives into consideration for a better understanding of voice and related
In ''Semantic and syntactic patterns in Swedish passives'', Elisabet Engdahl
discusses the ways in which the two passive forms (the morphological passive
with the suffix -s and the periphrastic passive with the copula _bli_ 'become')
are used in present-day written Swedish, in contrast with Danish and Norwegian.
A quantitative analysis based on Laanemets (2004) reveals that the -s passive is
the unmarked form: it is more frequent and has a wider distribution (syntactic
contexts, types of verbs) than the _bli_ passive, which is subject to additional
restrictions (it is used when the subject of the passive clause has control over
the situation described or may influence it). There is also a discussion on the
extent to which a demoted agent is syntactically and/or semantically present. A
corpus analysis reveals that in the vast majority of passive sentences the agent
remains unexpressed. A possible syntactic analysis of passive forms is given in
the last section of the paper based on Engdahl (2001), within the context of
Head-Driven Phrase Structure Grammar, and an alternative analysis in terms of
Construction Grammar is outlined.
Andrey Filtchenko's paper ''The Eastern Khanty locative-agent constructions: A
functional discourse-pragmatic perspective'' is concerned with the identification
of certain grammatical resources with semantic-pragmatic properties associated
with voice constructions in Eastern Khanty, one of the native languages of the
indigenous hunter minorities of north-western Siberia. The author concentrates
on the occurrence of agented passives and so-called ''ergative'' constructions
with locative-marked agents in Eastern Khanty narratives. Both types of
constructions are similar in that they are chosen to mark a temporary shift in
the status of discourse participants - the parenthetical establishment of a
secondary topic - expressed by the foregrounding of the non-agent referent and
the backgrounding of the agent referent, which nevertheless maintains a high
activation (topicality) status. The difference between both constructions has to
do with the semantic properties of the proposition: the role of participants and
the underlying transitivity of the event.
In ''Agent back-grounding as a functional domain: Reflexivization and
passivization in Czech and Russian'', Mirjam Fried focuses on the functional
properties of 'be' passives and the passive reflexive in Czech and Russian. The
author argues that despite the formal similarities between the two structures,
they occupy different parts of the functional space associated with the notion
of agent back-grounding, i.e. they are different constructions in the sense of
Construction Grammar. Though both constructions presuppose a backgrounded agent,
the 'be' passive serves the purpose of highlighting the end result of an action.
This is also true for the Russian reflexive, while the Czech reflexive has an
existential, event-reporting function, which highlights the event, rather than
its participants. This is motivated by the pragmatic function of the reflexive,
which is a marker of unexpected (diminished) referential status of the agent.
The relevant criteria for the distinction between the two constructions include
animacy of the agent, the interpretation of the back-grounded agent
(indefiniteness vs. genericity), inherent verb meaning, aspect and the potential
for semantic extensions.
In ''Invisible arguments: Effects of demotion in Estonian and Finnish'', Elsi
Kaiser and Virve-Anneli Vihman examine the properties of the implicit agent in
two constructions: the 'impersonal construction' and the 'zeroperson
construction'. The authors put forward what they refer to as the dissociation
hypothesis, by which demotion may take place at the discourse level, at the
semantic level and/or at the syntactic level, which accounts for the differences
observed between Estonian and Finnish impersonal constructions, whereas in the
zero person construction both languages appear to behave similarly: the implicit
argument is present on the semantic and syntactic level, but fails to project a
salient discourse entity,
The aim of Dalina Kallulli's ''Argument demotion as feature suppression'' is to
define the role of the special morphology in constructions involving demotion,
such as passives and anticausatives. Data from Albanian and English suggest that
both structures are the result of the same operation: suppression of an
event-type feature in the functional head 'v' in the syntax. This is based on
the hypothesis that anticausatives have causative semantics. The analysis
questions the validity of typical tests for the presence of implicit agents such
as adverb-oriented adjectives, purpose clauses and by-phrases.
In ''A comparative view of the requirement for adverbial modification in
middles'', Marika Lekakou addresses a well-known contrast regarding the presence
of adverbs in middle constructions: middles in Germanic (e.g. English, Dutch and
German) require adverbial modification, as opposed to middles in other languages
such as Greek and French (see Fagan 1992). The author argues that this is a
syntactic or a structural restriction, rather than a semantic/ pragmatic one,
against some existing proposals in the literature. In particular, the role of
the adverb is to recover the implicit agent of middles through identification
with the experiencer argument contributed by the adverb, in those languages
where the latter is not syntactically represented. This, together with the
dispositional semantics ascribed to middles, makes predictions about the set of
appropriate middle modifiers.
Joan Maling's paper ''From passive to active: Syntactic change in progress in
Icelandic'' reports on the results of a nationwide study designed to investigate
the properties of a 'new' construction in Icelandic, which has surface
properties of both the standard passive and the active voice and is widely
accepted by adolescents, but regarded as ungrammatical in the standard use. The
results of the survey support the hypothesis that what looks like a
morphological passive is well along the way to being reanalyzed as a
syntactically active construction with a phonologically null indefinite subject.
Such reanalysis parallels a similar diachronic development that has occurred
independently in both Polish –no/to constructions and Irish autonomous
impersonals. A key factor in understanding why this change is happening in
Icelandic but not in other Germanic languages is that even in the standard
language passive morphology is associated with a human agent reading.
Anneliese Pitz's ''The relation between information structure, syntactic
structure and passive'' investigates the occurrence of passive constructions in
translations from German into Norwegian and vice versa, with data drawn from the
German-Norwegian subcorpus of the Oslo Multilingual Corpus. The study shows that
change of voice is to a large extent determined by information structure:
adhering to the theme-rheme sequence of the source sentence, but its use
constitutes an essential structural device to avoid interpretation problems and
Eva-Maria Remberger's ''Syntax and semantics of the deontic WANT-passive in
Italo-Romance'' discusses several constructions in Romance varieties (e.g.
Calabrian and Sardinian) that have the verb 'want' as an auxiliary, with special
reference to the passive construction. She provides a minimalist syntactic
analysis in which the implicit argument of passive constructions is both
semantically and syntactically present, as the element PRO in the specifier of
the Passive Pr-head, which can either be arbitrary or controlled by a
prepositional adjunct. The semantic analysis is concerned with the semantic path
from volitional modality to deontic modality, as a process of grammaticalization.
In ''Agentivity and the virtual reflexive construction'' Nola M. Stephens argues
that virtual reflexives (e.g. 'This problem solves ITSELF') constitute an
independent construction, independently from middles and other related
constructions. A lexical semantic analyses of the verbs in virtual reflexives
indicates that they involve a volitional agent and, like middles, signify that a
property of the patient contributes to the agent's actions. However, unlike
middles, virtual reflexives obtain this interpretation via a ''metaphorical
device'' by which agentivity is partially transferred to the patient. This
accounts for some differences between middles and virtual reflexives regarding
the class of verbs which may appear in these constructions.
Thomas Stroik's paper ''Arguments in middles'' contributes to the current debate
in the formal literature as to whether middle formation is a lexical process or
a syntactic process. In previous work, the author has defended a syntactic
approach, parallel to passive formation (see Stroik 1992, 1999), as opposed to
pre-syntactic analyses, in which a ''middle verb'' is detransitivized in the
lexicon (notably Fagan 1992 and Ackema and Schoorlemmer 1995, 2003). Here, he
discusses two additional issues, which are best accounted for under a syntactic
approach: (i) the optional presence of a _for_-PP middle in English and (ii) the
mandatory presence of a reflexive pronoun in German. Under this account, middles
involve both demotion of the external argument and promotion of the internal
argument. Towards the end of the paper, Stroik attempts to show there is a
connection between several interdependent properties of middles: the presence of
reflexives, the event responsibility property assigned to the promoted object,
the promotion of the logical object, the demotion of the external argument and
the absence of a vP in middles. The analysis, though highly speculative seems
promising because it offers a way to explain several interdependent properties
All the articles in this volume make a significant contribution to the ongoing
discussion on voice and voice-related phenomena. The book provides an overview
of the relevant data in a variety of languages in which middle and other
voice-related constructions have received considerable attention (mostly
Germanic languages such as Swedish, Norwegian, Icelandic, English and German,
but also other languages such as Polish, Czech, Russian, Greek etc.), as well as
offering new data from other languages (e.g. Eastern Khanty and Italo-Romance
varieties) and/or from constructions which have figured less prominently within
the general phenomenon of agent demotion (e.g. Remberger's 'want'-passive in
Italo-Romance varieties or Stephens' virtual reflexives in English).
The volume brings together different theoretical perspectives, with different
research questions and methodologies. This is for the most part enriching, but
can create some confusion, as authors differ even when giving definitions of
what constitutes a passive or a middle construction, ranging from analyses in
which voice phenomena are analyzed along an active-passive continuum (e.g.
Flitchenko's paper, in line with Shibatani 1985 and Givón 2001) to those in
which constructions involving agent demotion are structurally defined (e.g. in
Kallulli's, Remberger's and Stroik's papers based on classical analyses of
passivization in generative grammar such as Baker, Johnson and Roberts 1989 and
more recent accounts of transitivity within the Minimalist Program such as
Bowers 2002). Authors also vary in how carefully they define the phenomena under
study. I found Filtchenko's paper particularly difficult to follow: the
definitions used are not always clear and the examples from Eastern Khanty are
often difficult to interpret. On the whole, most authors take care to be precise
in their definitions and examples are well chosen (see, for instance, Kaiser and
Vihman's comprehensive definition of agent demotion).
Notions such as transitivity, agentivity, telicity, volitionality, and
prominence are present throughout the volume. There are also recurrent themes,
such as the interpretive bias towards a human interpretation of the subject in
'impersonal' agent-demotion constructions, the suitability of standard
agentivity tests such as agent-oriented adverbs and by-phrases, the issue of
adverbial modifications in middles, and so on. This provides unity to the volume
and, though the overall picture is a generally coherent one, perhaps the editors
could have made more of an effort in bringing all these issues together by, for
instance, providing more guidance for the readers in the introductory chapter
and by making more use of cross-referencing,. Unusually for a volume of this
type, the editors do not provide in their introductory chapter a state-of-the
art overview of research into voice, nor do they summarize the contents of the
papers included in the volume (though there are frequent references to the
articles contained in it). This, in itself, is a good decision, due to the
enormity of the task at hand, but more of an attempt could have been made to
highlight the central issues covered by the different authors and emphasize what
the different analyses have in common and in what they differ.
In spite of this, and the varying degree of technicality of the different
papers, the book reads well and will be of interest to linguists from different
theoretical frameworks, who can choose to read each paper individually or as a
contribution to the volume as a whole. Reading this volume, one gets an idea of
the vast literature on the topic covering a wide range of approaches (syntactic,
semantic, morphological, pragmatic, cognitive) and the wealth of (often
conflicting) terminology devised to analyze passives, middles and related
constructions. This only adds to the challenge of trying to uncover the nature
of the different aspects involved in voice-related phenomena. As M. Fried says
in the introduction to her paper ''capturing the essence of passive-like patterns
or agent demotion in their various formal and functional manifestations is not a
trivial task even within a single language'' (p. 83).
Ackema, Peter and Maaike Schoorlemmer. (1995) Middles and non-movement,
_Linguistic Inquiry_ 26, 173-197.
Ackema, Peter and Maaike Schoorlemmer. (2003) Middles. In Martin Everaert and
Henk van Riemsdijk (eds.), _The Blackwell Companion to Syntax_, Vol III, CD-ROM.
Baker, Mark C., Kyle Johnson and Ian Roberts. (1989) Passive arguments raised.
_Linguistic Inquiry_ 20, 219-251.
Engdahl, Elisabeth. (2001) Scandinavian passives in HPSG. In A. Holmer, J-O
Svantesson and Å Viberg (eds.), _Proceedings of the 18th Scandinavian Conference
of Linguistics_, Vol 2. [Travaux de l'institut de linguistique de Lund 39:2],
23-36. Lund: Universitetstryckeriet.
Fagan, Sarah. (1992) _The syntax and semantics of middle constructions_.
Cambridge: Cambridge University Press.
Givón, Talmy. (2001) _Syntax: An Introduction_. Amsterdam: John Benjamins.
Hopper, Paul J. and Sandra Thompson. (1980) Transitivity in grammar and
discourse. _Language_ 56, 2, 251-299.
Laanemets, Anu. (2004) Dannelse og anvendelse of passiv i dansk, norsk og svensk
[Nordistica Tartuensia 11], Tartu University.
Shibatani, Masayoshi. (1985) Passives and related constructions. _Language_ 61,
Stroik, Thomas. (1992) Middles and movement. _Linguistic Inquiry_ 23, 127-137.
Stroik, Thomas. (1999) Middles and reflexivity. _Linguistic Inquiry_ 30, 119-131.
ABOUT THE REVIEWER
Amaya Mendikoetxea is a lecturer in English Syntax in the Department of English
Philology at the Universidad Autónoma de Madrid (Spain). Her research interests
include, among others, aspects of the lexicon-syntax interface in Romance and
Germanic languages (passivization, impersonal construction , middles,