This monograph is based on the author's Ph.D. thesis (2007, University of Verona, Italy) but also incorporates her more recent work on the same topic (e.g. Melloni and Bisetto 2010). It is a study of deverbal nouns that regularly have both a process/state (=event) reading and a variety of more ''referential'' readings. Using data from Italian, Melloni (henceforth M) describes the polysemy of event (E) and result (R) nominals within the approach to the semantics of derivational affixes mainly developed by Lieber (cf. Lieber 2004).
The issue tackled by M is the following: How does one describe the ambiguity of so-called Action Nominals from a morpho-semantic perspective? This question has had a distinguished place in the recent history of linguistic theory and obviously continues to do so (from Chomsky 1970 and Comrie 1976 to the present). The basic facts can be summarized by the following examples:
a. Speakers NOMINALIZE verbs constantly.
b. The speakers' constant NOMINALIZATION of verbs…
c. The frequent NOMINALIZATION *(of verbs) (by speakers)…
d. ''Nominalization'' is a NOMINALIZATION.
e. ''Nominalization'' is a NOUN.
The deverbal noun NOMINALIZATION (related to the verb NOMINALIZE) can be used in a noun phrase, as in (1b), that is parallel to the sentence in (1a), with both the internal and the external argument and an aspectual modifier, but it can also be used like a non-derived, non-relational noun (cf. 1d-1e). One also needs to consider examples like (1c) that show a structural resemblance between the noun phrase and a passive sentence, but where the object (''of verbs''), and not the subject (''by speakers''), is obligatory. Semantically, NOMINALIZATION can denote either the result of the process, as in (1d), or the process, as in (1b-c), that is denoted by NOMINALIZE in (1a); pragmatically, NOMINALIZATION can refer to a thing, as in (1d), or an event, as in (1b-c). In (1d), NOMINALIZATION seems to denote the result of the process denoted by NOMINALIZE. Thus, it appears that the same deverbal noun may sometimes have some verbal properties (presumably inherited from its base), and sometimes may simply be a noun.
I will first summarize each of the six chapters of the book and then discuss some questions raised by this study.
Chapter 1: Aims and Orientation
The first chapter presents the basic facts from the perspective of Generative Grammar: the class of Action Nominals consists of morphologically complex lexemes that are the result of merging an affixal head and a verbal base. Nominalizations are thus endocentric constructions. Nominalizing suffixes are deemed to be ''transpositional'', i.e., they only change the syntactic category while preserving (most of) the properties of the base verb. The distinction between E and R nominals hinges almost exclusively on the presence (E) versus the absence (R) of argument structure, since R nominals form a heterogeneous class denoting not only results, but also locations, and collectives, among others.
Despite the wide range of meanings of E/R nominals, which has led some prominent morphologists to claim that their interpretation is only constrained by world-knowledge (i.e. pragmatics) and linguistic context, M maintains that there are significant semantic regularities that do not follow from such a pessimistic stance and that can be described within Lieber's (2004) ''descriptive framework and formal apparatus for derivational semantics'' (p. 10). In this framework, the meaning of a lexical item or of an affix consists of a ''skeleton'', representing only semantic properties that have a morpho-syntactic reflex, and a ''body'', representing the encyclopedic knowledge necessary to use the item properly.
M wants to characterize the type of ambiguity of E/R nominals, the default meaning, the range of possible meanings, and the reason for the presence or absence of argument structure in these nominals.
Chapter 2: Generative Approaches to Nominalization
As M correctly points out, the analysis of nominalizations in generative literature has focused on deriving the verbal properties of E nominals. In this chapter, she reviews some of the influential approaches to the topic (on the lexicalist side of the debate), giving special attention to solutions that involve features of lexical meaning, as opposed to solutions that rely more heavily on the contribution of syntactic configuration to the interpretation of nominalizations (the latter being one of the most important insights of Chomsky 1970, with the introduction of X-bar theory).
In Grimshaw (1990), the ambiguity of E/R nominals results from the ambiguity of nominalizing affixes, which may bind non-thematic arguments that are either referential (i.e. real-world entity that is referred to) or eventual (i.e. situations as entities). This is possible because Grimshaw's framework includes a level of representation called lexical conceptual structure (LCS), where these E and R arguments are represented but may or may not be mapped onto the intermediate representation that specifies the argument structure of the lexical item. The suffixes are, in essence, polysemous, and, depending on the selection of an E or R argument at the level of LCS, different morpho-syntactic reflexes will follow. M's analysis will essentially follow that of Grimshaw, while trying to remedy some of its flaws, like the ad hoc treatment of the optionality of the arguments of E nominals, simply renamed ''adjunct-arguments'' by Grimshaw.
In her review of ''lexico-semantic'' approaches, M adopts the concept of inherent polysemy (Pustejovsky 1995, based on Weinreich 1970), where the meanings of a word are different without being mutually exclusive (e.g. complementary: ''BANK crisis'', ''food BANK''; exclusive: ''a river's BANK''). The polysemy of E/R suffixes is thus also found in non-derived words and is of the complementary type (see also Apresjan 1974). Lexico-semantic approaches also draw attention to the semantic contribution of the base in word-formation, emphasizing, for example, that different types of situations (e.g. ''states'', ''activities'', ''accomplishments'', and ''achievements'', cf. Vendler 1967, Dowty 1979) give different kinds of interpretations in nominalizations.
Chapter 3: Suffixes and Co-indexation
In this chapter, M first briefly presents the data, where the main focus is on three Italian nominalizing suffixes (i.e. -mento, -zione, -tura) that have cognates in English and many Romance languages. Those three suffixes are the most productive and least specialized forms and are taken to be semantically equivalent, filling the same paradigmatic cell.
Then, Lieber's (2004) decompositional semantics is introduced. In it, the skeleton of lexical items and affixes is represented as a hierarchical list of features coupled with an appropriate number of argument slots for items of a predicative nature. In affixation, affixes bind one of the base's arguments (normally the highest one) through a mechanism of coindexation, and they together form more complex semantic representations by compositionally subordinating the base's skeleton. M proposes that E/R suffixes have a ''double representation'' corresponding to a single lexical entry (p. 64):
(2) Lexical Entry of Nominalizing suffixes
E skeleton: [-material, +/-dynamic ([ ]E, <base>)]
R skeleton: [+/-material, -dynamic ([ ]R, <base>)]
In her discussion of the mechanism of coindexation, M makes a number of novel claims. Echoing Grimshaw and Higginbotham (1983), among others, she argues that ''E suffixes'' bind a non-thematic event argument, which is something that isn't present in Lieber's (2004) system, and is meant to explain the presence of syntactic satellites with E nominals (bound arguments cannot be expressed syntactically). She also argues that ''R suffixes'' may coindex any participant (thematic or not) present in the LCS of the base verb and that this takes care of most of the denotations possible for R nominals.
Chapter 4: Base Verbs and Semantic Constraints
The next chapter addresses the contribution of the base in the interpretation of R nominals. M's aim is to explicitly state the limitations on R interpretations in order to reflect the fact that they are less productive than E nominalizations (with the latter being the default).
To do so, verbal bases are almost exhaustively classified using criteria developed by Pustejovsky (e.g. 1995) and Levin & Rappaport-Hovav (e.g. 2005). In order to obtain an R reading, the base verb must have some participant in its semantic representation (LCS and/or qualia structure) which satisfies the constraints imposed by the suffix's R slot: it must be NON-SENTIENT, and is preferably INCREMENTAL or EFFECTED. The last two constraints are violable but, when they are satisfied, they correspond to PATH or PRODUCT readings, which are themselves sub-types of ENTITY-IN-STATE readings (a cover term for most R readings).
Thus, by virtue of their semantic representation, creation verbs correspond to product nominals because they include an effected participant that can fill the R slot of the suffix. Path readings are obtained by binding an incremental argument of the base verb. In the case of state verbs, which lack effected or incremental arguments, the binding of the internal argument results in an entity-in-state reading.
Crucial to this analysis is the idea that the R slot of the suffix may be filled by ANY participant present at some level of semantic representation. Notice, however, that it does not bind an abstract argument, as is the case for the E suffix (this important formal distinction is not discussed by M).
Chapter 5: Verb Skeletons and Co-indexation
The last chapter details how the expanded system of semantic representation works. First, M shows how coindexation is implemented between the skeleton of the base, which now contains additional non-thematic arguments and a D-PATH argument (corresponding to the incremental LCS participant), and the skeleton of the suffix in a way that in principle should reflect the possibility of overt arguments for E nominals. In particular, M finds support for the non-thematic event argument through the fact that Lieber's Coindexation Principle would wrongly coindex the external (i.e. subject) argument of the base if it were indeed the highest one. M's approach is more consistent with the fact that the subject can be optionally expressed as a by-phrase with E nominals.
Second, M shows how the mechanism of inheritance fixes the value of underspecified features in the skeleton of affixes (i.e. [+/-dynamic] for E suffixes; [+/-material] for R suffixes). Stative bases confer a minus value and active bases a plus value for the feature [dynamic], and the base argument that is coindexed with the R slot transfers its value for the feature [material].
Finally, M demonstrates how the semantic representation of the base and the suffix are merged to form a single expression that describes the grammatically relevant meaning of E and R nominals.
Chapter 6: Conclusion
This short epilogue summarizes the previous chapters and underlines the defining characteristics of M's approach: a focus on the MEANING of E/R nominals, not just the morpho-syntax; a systematic treatment of R nominals' varied interpretations; the novel integration in Lieber's (2004) framework of Pustejovsky's (1995) qualia structures as a formalization of the body of lexical bases; and finally, the empirical coverage of her analysis.
The double representation of nominalizing suffixes accounts for E versus R interpretations, with the former being relatively unconstrained, as is expected of the default option. Within R interpretations, the uncovered semantic constraints are said to account for “product”, “means”, “path”, and “psych-stimulus'' readings. There remains a small number of possible non-E meanings that cannot be directly described in M's scheme: “agentive-collective” and “locative” are said to result from ''paradigmatic sense-extension'', a functional notion already invoked by Lieber (2004) in connection with a different affix; “manner”, “temporal”, and “factive” are said to be triggered by predicative context (some kind of coercion effect.)
Unfortunately, this short summary cannot do justice to the level of detail, breadth, and theoretical refinement of M’s study of the semantics of Italian nominalizations and nominalizing affixes in Generative Grammar.
The theoretical problems tackled in this book are at the center of most formalist approaches to the description of language, in particular: 1) the nature of syntactic categories and the types of mappings between syntactic categories and one (or many) levels of semantic representation; and 2) the (apparently) different ways in which an utterance can refer to something worldly (for events and concrete/abstract entities particularly) and the related issue of the status of various kinds of semantic “features” (perhaps beginning with Katz & Fodor 1963, Weinreich 1970).
These questions also reflect a renewed interest in the semantics of word-formation (frequently defined as “category-changing morphology”) on the part of morphologists, with a sense that there is new ground to cover in this area. In his closing words at the last International Morphology Meeting on “Morphology and Meaning”, Hans-Christian Luschützky was talking about the tools we use to describe and understand meaning in morphology and said something to the effect that “we are still eating our semantic soup with a fork, but a nice fork nevertheless”.
In this regard, M’s Chapter 2 (Generative Approaches to Nominalization) provides a detailed and insightful overview of the current state of the field, highlighting the fact that basic assumptions and biases (e.g. focus on syntactic configurations or on lexical semantic content) give rise to different, perhaps complementary, analyses, and showing that influential treatments of the subject, such as Grimshaw (1990) and Borer (2003), fall short of an explanation of the polysemy of E/R suffixes because they fail “to account for the evident semantic relation between the interpretation of a nominalization as a situation and as a product or effect of the same situation” (p. 43).
M’s claim that E/R exhibit logical/inherent polysemy (different but not mutually exclusive meanings) is a welcome attempt to improve on such work. However, it appears that in her analysis, the opposite is true; each suffix binds EITHER a referential argument OR an eventual argument. In other words, the E and R readings are mutually exclusive. As with Grimshaw (1990) and Borer (2003), M does not explain the “natural connection” between E and R readings, but simply puts several definitions under a unique label. As Rainer (2010: 15) rightly remarks: “the use of disjunctions in the definiens makes it possible to ‘unify’ anything”. Instances of regular polysemy do not need an independent, language or item-specific characterization (Apresjan 1974: 17) because they are explained by independent and more general (universal) principles. For example, the possible metonymic shift between ‘container’ and ‘content’, as in “drink a GLASS of wine”, is probably explained by a principle that is not specific to the English language.
In the study of nominalizations, one of the fundamental questions seems to be: “What does it mean for a noun to denote an event?” Verbal Nouns, as Comrie (1976) calls them, do just that, and I think that formalists would benefit from the insights of as many perspectives as possible. Among many other titles, there are interesting leads in Langacker (1987), for example, who suggests that temporality, or its absence, in the construal of situations plays an important role in understanding the concept of “nominality”. Another option that should be entertained is Bhat’s (1979) proposal that the categorial ambiguity of nominalizations stems from a more primitive distinction, familiar to philosophers of language, between NAMES and DEFINITE DESCRIPTIONS.
This book will be of interest to graduate students and researchers working on areas such as morphological theory, lexical semantics, argument structure, and Italian/Romance derivational morphology.
Apresjan, Juri D. 1974. Regular Polysemy. Linguistics 142.5-32.
Bhat, D. N. Shankara. 1979. The referents of noun phrases. Pune: Deccan College.
Borer, Hagit. 2003. Exo-skeletal vs. Endo-skeletal Explanations. In: The Nature of Explanations in Linguistic Theory, ed. by J. Moore and M. Polinsky, 31-67. Chicago: CSLI and University of Chicago Press.
Chomsky, Noam. 1970. Remarks on Nominalizations. In: Readings in English Transformational Grammar, ed. by R.A. Jacobs & P.S. Rosenbaum, 184-221. Waltham, Mass.: Ginn.
Comrie, Bernard. 1976. The Syntax of Action Nominals: A Cross-Language Study. Lingua 40.177-201.
Dowty, David R. 1979. Word Meaning and Montague Grammar. Dordrecht: Reidel.
Grimshaw, Jane. 1990. Argument structure. Cambridge, Mass.: MIT Press.
Higginbotham, James. 1983. The Logic of Perceptual reports: An Extensional Alternative to Situation Semantics. Journal of Philosophy 80: 100-127.
Katz, Jerrold J. & Jerry A. Fodor. 1963. The Structure of Semantic Theory. Language 39.170-210.
Langacker, Paul. 1987. Nouns and Verbs. Language 63.53-94.
Levin, Beth & Malka Rappaport-Hovav. 2005. Argument Realization. Cambridge: Cambridge University Press.
Lieber, Rochelle. 2004. Morphology and Lexical Semantics. Cambridge: Cambridge University Press.
Melloni, Chiara & Antonietta Bisetto. 2010. On the Interpretation of Nominals: Towards a Result-oriented Verb Classification. In: Language and Cognition: Traditional and New Approaches, ed. by O. Souleimanova, 165-178. Frankfurt: Peter Lang.
Pustejovsky, James. 1995. The Generative Lexicon. Cambridge, Mass.: MIT Press.
Rainer, Franz. 2010. Sobre Polisemia en la Formación de Palabras. Hesperia. Annuario de Filología Hispánica 13.7-52.
Vendler, Zeno. 1967. Linguistics in Philosophy. Ithaca, N.Y.: Cornell University Press.
Weinreich, Uriel. 1970. Explorations in Semantic Theory. In: Current Trends in Linguistics v. 3, Theoretical Foundations, ed. by T.A. Sebeok, 395-477. The Hague: Mouton.