| AUTHOR: Artemis Alexiadou
TITLE: The Semantics of Nominalizations across Languages and Frameworks
SERIES TITLE: Interface Explorations [IE] 23
PUBLISHER: De Gruyter Mouton
Kilu von Prince, ZAS Berlin and HU Berlin
This volume is a collection of papers presented at the workshop ''Nominalizations
across Languages'', in Stuttgart, Germany in December 2007, as follow:
1. Monika Rathert and Artemis Alexiadou: Introduction.
The two editors of the volume give a short overview of the major research
questions concerning the semantics of nominalizations as well as a short summary
of the individual papers.
2. Chris Barker: Nominals don't provide criteria of identity.
In his contribution to the volume, Chris Barker claims that ''Nominals don't
provide criteria of identity''. The opposite position has been held by
researchers like Geach (1962) and Baker (2003), who claim that criteria of
identity are part of the lexical meaning of nouns.
Barker first dismisses cases which involve pairs of count nouns and mass nouns
such as ''river'' and ''water''; he argues the apparent differences in identity
assignment can in fact be accounted for by a proper definition of the mapping
relation between the count noun and the corresponding mass noun.
More challenging examples, Barker argues, involve cases where identification is
based on a per-event interpretation rather than a per-individual interpretation
as in the following pair:
(1) Easyjet served 10 million passengers last year.
(2) Easyjet served 10 million people last year.
Barker observes that deverbal nouns with agentive ''-er'' or ''-ee'' form a
productive class of nouns with a per-event interpretation as the default
reading. This is because the referents of such nouns are defined by their
engagement in a given activity; thus, the same individual may be counted more
than once if they engage in the activity more than once.
Barker then endeavours to show that even for these cases, lexically given
criteria of identity are neither necessary nor plausible. Instead, he argues,
whether a noun receives a per-event interpretation or a per-individual
interpretation depends on lexical criteria of application as well as pragmatic
3. Regine Brandtner and Klaus von Heusinger: Nominalization in context --
conflicting readings and predicate transfer.
Brandtner and von Heusinger address deverbal nouns derived with ''-ung'' in German
which can have a variety of sortal readings such as events, results or objects,
depending on the context. To discuss how these different readings are determined
compositionally, they start out with the following observation: modifiers and
predicates of the same noun can have conflicting selectional restrictions for
the sortal reading of the noun. Thus, in the following clause, the adjective
triggers an event reading (EV), while the verb requires a result reading (RE).
The noun apparently fulfils both requirements at the same time:
(1) Die [wiederholten]EV Messungen [belegen]RE, dass keine Besserung eingetreten
''The [repeated]EV measurements [show]RE that there has not been an improvement.''
The authors show that previous assumptions about how the sortal reading is
determined fail to account for such cases. Instead, the authors suggest that
the noun in fact receives only one reading, which is fixed by the first
expression that comes with selectional restrictions -- which would be the
adjective in the case of (1). Then, they argue, the second expression with
selectional restrictions undergoes a predicate transfer as defined by Nunberg
(1995). This approach was developed to account for clauses like ''I am parked out
back'', where the meaning of the predicate [parked out back] is transferred to
[the owner of a car that is parked out back], according to Nunberg.
4. Liesbet Heyvaert: A cognitive-functional perspective on deverbal
nominalization in English. Descriptive findings and theoretical ramifications.
The contribution by Liesbet Heyvaert advocates the advantages of a
cognitive-functional approach to nominalizations over other treatments,
especially over generative syntactic approaches. After outlining the principles
of her take on language, the author turns to case studies of ''-er'' and ''-ing''
For ''-er'' nominalizations, Heyvaert gives a brief summary of her work in
Heyvaert (2003). She reviews the treatments of ''-er'' nominalizations by, among
others, Rappaport Hovav and Levin (1992) and concludes that they do not deal
with, and cannot account for, cases profiling non-subject arguments such as
''stroller''. Based on a corpus study of non-agentive ''-er'' nominalizations,
Heyvaert concludes that lexicalized ''-er'' nominalizations, including ''stroller'',
''broiler'', ''teacher'', ''drinker'' and ''toaster'' systematically draw on a modal
meaning, which could be given the following form: according to the properties of
x it is possible that P(x). By contrast, ad hoc nominalizations as in ''the
city's destroyer'' typically refer to the agent of a particular event.
The section about ''-ing'' nominalization also sums up some of Heyvaert's earlier
work. She discusses two issues: first, the problem of how to distinguish between
factive action nominalizations as in (1) and gerundives as in (2):
(1) They say [Saddam's targeting of Israel] did not achieve its objective
(2) My husband speaks very well, but his job involves [my answering the phone on
his behalf quite a bit of the time]
The second issue concerns the question of why certain items such as adjectives
and demonstratives are no longer available for gerundive nominalizations.
5. Tibor Lackzó: A new account of possessors and event nominals in Hungarian.
Tibor Lackzó addresses the syntax and semantics of Hungarian possessive
constructions and introduces his approach within a Lexical Functional framework.
He identifies two problems with classical approaches within a GB framework. The
first problem is with the semantic representation commonly assumed for the
possessive construction, as given in (1):
(1) λxλy[N(x) & R(x,y)], where N is the predicate denoted by the head noun and R
is the (possessive) relation between the head noun and its possessor. E.g. in
''John's hat'', N would be λx[hat(x)] and R would be λx. owns(John)(x).
Lackzó claims that the possessive relation is subordinate to the predicate
denoted by the head noun and that a semantic representation should represent
this subordinate relation.
The second problem the author seeks to amend concerns relational and deverbal
nouns. In the influential account by Szabolcsi (1994), the possessive suffix to
a noun always assigns a theta role to the possessor. As a consequence the
possessor of a deverbal noun as in (2) is assigned two theta roles, one by the
verbal root, and one by the possessive suffix to the deverbal noun.
as ö megérkez-és-e
the he.NOM arrive-DEV-Poss.3SG
Lackzó suggests instead that the possessive suffix always forces a noun to be a
two-place predicate. Relational and deverbal nouns then assign one of their
lexical roles to the possessor noun.
Finally, Lackzó discusses some of the empirical challenges to his approach.
6. Fabienne Martin: The semantics of eventive suffixes in French.
Fabienne Martin starts her treatment of the French nominalizing suffixes ''-age'',
''-ment'' and ''-ion'' by showing that French speakers do apply them productively
and systematically, even though many of the nouns they occur with have been
inherited or loaned from Latin.
The author restricts her study to eventive readings of the deverbal nouns under
consideration and establishes a test for diagnosing eventive readings. She then
discusses a number of hypotheses about how the distribution of the different
suffixes is determined. Arguing that no single factor can account for all the
different cases, she proposes a multi-feature analysis.
Martin contrasts the suffixes in pairs, starting off with the contrast between
''-age'' and ''-ment'', and identifies several features to account for the
differences in each of the three pairs. As a result, Martin comes up with a list
of five semantic features for which the three suffixes appear to have differing
values. These features then are suggested to determine the distribution of the
suffixes in productive use.
7. Chiara Melloni: Action nominals inside: lexical-semantic issues.
The paper by Chiara Melloni deals with deverbal action nominals and the variety
of interpretations they can receive. Action nominals have been divided into
argument-structure (AS) nominals and referential nominals, depending on whether
they preserve the argument structure of their verbal stem or not. This
morpho-syntactic distinction corresponds to the semantic classification of
action nominals into 'event nouns' and 'result nouns'.
However, Melloni argues, the term 'result noun' is not descriptively adequate
for a variety of non-AS, non-event nouns such as 'connection', 'distraction' or
'administration': Instead, these nouns rather refer to the means, the source or
the group of agents of the respective event.
The hypotheses Melloni puts forward is that such differences in readings can at
least in part be systematically derived from the lexical meanings of the verbs
involved. For her investigation, Melloni focuses on nouns derived by the Italian
suffixes ''-mento'', ''-zione'' and ''-tura''. Claiming that these suffixes are
semantically interchangeable, she goes on to identify those lexical features of
verb meanings which are responsible for the different readings available for the
Starting out with actual result nouns such as 'construction', 'translation' and
'puncture', Melloni suggests that those readings are related to the presence of
a positive [Loc] feature in the lexical entry of the verb. This feature is
assumed to indicate that spatial configuration plays a role in the meaning of
verbs like 'stay' or 'remain'.
Secondly, if a referential nominal refers to a participant of the event, that
participant has to be non-sentient. Finally, a binary aspectual feature
[dynamic], which is part of the verb meaning, will determine whether a
referential nominal will refer to the whole of the event [+dynamic] or its
target state [-dynamic].
8. Antje Roßdeutscher and Hans Kamp: Constraints on the formation and
interpretation of -ung-nouns.
German deverbal nouns with the suffix ''-ung'' are the topic of Roßdeutscher's and
Kamp's contribution to the volume. The authors try to identify the main factors
that determine whether a verb can serve as a productive base for an ''-ung''
nominalization or not.
The first major factor they identify is the distinction between verbs with
bi-eventive projections vs. verbs with mono-eventive projections. For example,
the verb 'to clean' implies both an agent doing something to an object and the
object getting clean as a result of that. By contrast, 'to wipe' only specifies
the activity of the agent. ''-Ung''-nouns can only be derived from bi-eventive
verbs. The authors then endeavour to derive this constraint from two more
general constraints about structural requirements of ''-ung'': they suggest that
''-ung'' must be inserted immediately above a minimal vP node and that it requires
as input a structure containing a condition of the form 'e CAUSE s'.
Elaborating on the consequences of these two hypotheses, the authors also
consider verbs which allow for the prefix ''be-'' and verbs derived from nominal
Finally, the authors turn to the variety, and often ambiguity, of the
interpretations available for ''-ung'' nominals. Classifying them into four
categories, they relate the meanings of the nouns to the semantics of the verbal
bases. One major factor accounting for the readings available for the noun,
according to the authors, is the distinction between 'property' and 'sortal
roots': The former are related to roots describing properties of a noun as in
'to weaken -- weak', whereas the latter refer to actions which contribute to, or
modify the existence of an entity such as 'to collect'.
9. Melanie Uth: The rivalry of French -ment and -age from a diachronic perspective.
In the final paper of the volume, Melanie Uth addresses again the two French
nominalizing morphemes ''-ment'' and ''-age''. Basing her investigation on
diachronic data, she proposes an alternative to Martin's account in the same volume.
Retracing the development of the two suffixes, Uth concludes that a single
semantic difference has determined their contrasting historic roles and is still
responsible for the differences in their productive use. She suggests that
'''-age' nominals focus on the property of subject referents to take part in the
event designated by the base verb''.
By contrast, ''-ment'' is suggested to be diachronically related to the Latin
morpheme ''-to'' which was used in the formation of past passive participles and
nominalizations referring to the result of an event. Correspondingly, Uth
claims, today's use of ''-ment'' can be characterized as nominalizing ''the
property of Theme arguments to participate in the state resulting from the base
In her final section, Uth discusses the semantic differences between ''-age''
nominals and ''-ment'' nominals in modern French, including those observed by
Martin in the same volume. She argues that in general, these differences can be
accounted for by the diachronically established difference mentioned above.
The collection offers a range of perspectives on the semantics of
nominalizations and does so, as the title promises, 'across languages and
frameworks'. As such it certainly makes for an inspiring resource for anyone
working on related issues.
Having said that, I would like to point out several shortcomings, some of which
are rather widespread for this type of publication, while some others are more
specific to this particular volume or individual contributions therein.
First of all, the authors as well as the editors could have made some effort to
somehow relate their different approaches to each other. Except for Melanie
Uth's paper, obvious connections between the papers are not commented on at all.
In particular, the observations and problems dealt with by Martin, Melloni,
Roßdeutscher and Kamp, and, to a lesser extent, Brandtner and von Heusinger,
appear to have significant overlaps, but are approached from very different
directions and with very different assumptions. While on the one hand, this
variety of approaches to similar problems is interesting in itself, it would
have been of great service to the reader if the authors and/or editors had made
the effort to comment on their differences.
Another point of critique concerns the papers by Barker and by Brandtner and von
Heusinger: While they introduce some intriguing observations and arguments, it
does not seem entirely clear to the reviewer to what extent their treatments are
specific to nominalizations in contrast to simple nouns such as 'passenger' in
the case of Barker or 'newspaper' in the case of Brandtner and von Heusinger. A
short elaboration on that would have been welcome.
The individual papers are frank about the fact that they are to a large part
sketchy or at least preliminary. The reader will not find definite answers to
general questions, but rather a number of interesting observations relevant to
very specific discussions.
Baker, Mark. 2003. Lexical Categories: Verbs, Nouns, and Adjectives. Cambridge:
Cambridge University Press.
Geach, Peter. 1962. Reference and Generality: An Examination of Some Medieval
and Modern Theories. Ithaca, NY: Cornell University Press.
Heyvaert, Liesbet. 2003. Deverbal -er suffixation as the equivalent of the
clausal Subject-Finite unit. WORD 54(1): 39-68.
Nunberg, Geoffrey. 1995. Transfers of Meaning. Journal of Semantics 12(2): 109-132.
Rappaport Hovav, Malka and Beth Levin. 1992. -ER nominals: implications for the
theory of argument structure. In Syntax and Semantics. Vol. 26: Syntax and the
Lexicon, T. Stowell and E. Wehrli (eds.), 127-153. New York: Academic Press.
ABOUT THE REVIEWER