Theory and Processing of Nominalizations

Guest Post by Artemis Alexiadou and Eva Wittenberg

Workshop Proposal:

Deverbal nominalizations are a fascinating slice of the language system, since they pose many questions that are common to both nouns and verbs. We will look at three of these questions from a theoretical point of view, and ask what and how psycholinguistic studies can contribute to a better understanding of nominal, and verbal, argument structure.

  1. Which arguments do nominalizations inherit from their underlying verbs? How do speakers represent these arguments, and how do they access them during language comprehension and production?
  2. How can theoretical architectures model the ambiguity of nominalizations? What determines the case-by-case or class-by-class resolution preferences? How do speakers resolve ambiguities during processing?
  3. How is nominal argument structure integrated into the sentence meaning when there are competing sources of argument structure? What are the cognitive processes behind the theoretical problems?

We aim to come up with ideas about experiments that can make use of the detailed theoretical insights while providing the right frame to tackle these questions.

Workshop Meeting Notes:

In the workshop, our aim was to discuss theoretical problems and to come up with an experimental design linking the linguistic model of a certain phenomenon to mental processes. Although there are an infinite number of questions to ask about nominalizations, we agreed to settle on one of the most basic topics: Which parts of the verbal argument structure does the mental representation of a nominalization include?

To this end, we focused on eventive vs object nominalizations, such as “construction”, “examinataion”, or “building”, which can be interpreted as ongoing process (the eventive reading), or as entities (the object reading).

There are two basic questions associated with these ambiguous nominalizations:

  1. Is there one basic reading from which the other reading is derived?
  2. How do readers systematically resolve this ambiguity?

For the second question, we came up with a sentence completion task (note that this does not answer the question of which mental processes the reader is employing to resolve the ambiguity; the result would only tell us which interpretation, in the
absence of context, is showing up more frequently).

Another way to tackle the second question would be a study using novel verbs and nominalizations, and looking at whether people (children and/or grown ups) have more or less difficulty interpreting the eventive or the object interpretation of the
nominalized novel verb.

As for the first question, we came up with an experimental design that partly draws on, among others, Baggiò et al, 2010, Bott, 2010, Pylkkänen & McElree, 2007, Frisson 2008, Kuperberg et al., 2010, Piñango & Zurif, 2001, and Todorova et al., 2000. These previous studies all investigated the contrast between “read a book” and “begin the book”. In “begin the book”, the literal Fregean composition is not enough; one has to understand that the event described in one of reading or writing. This process of understanding the right event type is called “coercion”. Analyses differ greatly as to what exactly is coerced, how coercion comes about, and where it is located in the linguistic system. Important for our purposes is the direction from noncoerced object reading to coerced event reading, since many linguistic theories would predict exactly the opposite path for contrasts like The student finished the examination (EVENT) vs. The student handed in the examination (OBJECT).

Glossing over details for now, such as the exact definition of “argument structure”, one can quickly determine three distinct predictions:

  1. If the event reading is basic, and the object reading is derived, a forced object reading should incur cost
  2. If both readings are derived from a root, the eventive reading should be more cognitively costly, since it presumably includes the verbal argument/event structure
  3. If the nominalization is stored with both meanings (or both meanings are stored as separate lexical items), then the processing should be equally costly for both meanings.

In sum, the contrasts we imagined for a future study of ambiguous nominalizations were:

  1. The student finished the examination (EVENT)
  2. The student handed in the examination (OBJECT)
  3. The student finished the book (EVENT)
  4. The student handed in the book (OBJECT)

We tentatively agreed also on using a self-paced reading paradigm, leaving other techniques open.

For another future experiment that might help to distinguish between a root-based nominalization representation vs the representation of a larger structure in the mental lexicon, we thought about cross-modal priming: In a contrast like The man finished the destruction vs The man looked at the destruction, one might expect reactivation of a semantically related probe word at different positions (see the work of Shapiro with aphasic patients, and follow-ups with healthy patients throughout the 1990s).

All in all, we felt this workshop was extremely productive and creative. Thank you all for your participation and interest, and we hope to further pursue together the projects we developed!

Eva & Artemis

Advertisements

Well Verbed, Everyone

For three days last week, we gathered a broad range of some of the most interesting people working on issues related to verbs and their argument structure in a small auditorium at CNRS-Pouchet. From the very first talk by Alec Marantz, with commentary by Victor Acedo Matellán, and the discussion that ensued, it was clear that we were in for a wide-ranging, productive, and inspiring conversation.

Over the next week or two, we’ll post reports from the facilitators of the brainstorming discussion groups we ended the meeting with, as well as handouts, slides and posters from the various presenters.

If you were at the meeting and would like to contribute a commentary on any issue that was raised, please let me know and I will be happy to post it. If you missed the meeting, we hope you’ll find these posts useful, and that you’ll join us for our next meeting (stay tuned).

Bienvenue/Welcome

Dates: 5-7 Septembre

Lieu/Place: CNRS Pouchet, 59 Rue Pouchet

Invités/Invited Speakers:

***

Le but de cet atelier est de rassembler des chercheurs d’horizons scientifiques et méthodologiques différents autour de la question de la représentation mentale de la structure argumentale du verbe. Cette thématique de recherche a en effet été abordée par divers champs de recherche (Développement linguistique, syntaxe et sémantique théoriques, lexicologie, typologie, psychologie, neurosciences cognitives, neuropsychologie) que nous espérons se faire se rencontrer dans cette plate-forme de travail où seront présentées des recherches actuelles, récentes ou en cours ; cela afin de permettre des échanges fructueux entre des chercheurs utilisant des cadres théoriques et des méthodologies variées et travaillant avec des populations diverses. Nous pensons que ces échanges peuvent être d’un grand bénéfice pour notre communauté scientifique. Les contributions des invités venant de champs théoriques différents recevront des commentaires croisés de chercheurs d’autres domaines. Une sélection des travaux présentés lors de ce workshop seront publiés dans un numéro spécial des Recherches Linguistiques de Vincennes (RLV)

****

The aim of this workshop is to bring together scientists working in different fields and with different methodologies to discuss the mental representation of verbal argument structure. This topic has been addressed by work in a variety of fields (language and conceptual development, theoretical syntax and semantics, lexicology, linguistic typology, psychology, cognitive neuroscience, neuropsychology, etc.). In this workshop we hope to create a cross-disciplinary platform to discuss ongoing research. We believe that an open exchange between scientists using different methods, working with different populations and within different theoretical frameworks will be of great benefit to the community. To this end we have invited speakers from different fields (language acquisition, neuro-imagery and aphasiology, theoretical syntax, theoretical semantics, sign language), each to be commented upon by a researcher from another discipline. A selection of papers presented in the workshop will be published in a special issue of Recherches linguistiques de Vincennes (RLV).

Read more…