Skip to main content

Modal determiners

Final Report Summary - MODALDET (Modal determiners)

Modal Determiners (PIEF-GA-2013-622311): Final Publishable Report

Project’s website:

Contact details:

Paula Menéndez-Benito
Department of Translation and Language Sciences
Universitat Pompeu Fabra
Edifici Roc Boronat, office 53.708
Roc Boronat 138
08018 Barcelona, Spain

1. Project Objectives

Modal expressions in natural language allow us to talk about unrealized possibilities. For instance, (1), with the modal "might", describes a possible scenario compatible with our evidence. (2), with "should", evokes potential situations in which John fulfills his obligations.

1) (Given what we know), Jones might be the murderer.
2) (Given his parents’ orders), John should be in bed by ten.

The ability to construct discourses about the non-actual is a fundamental cognitive skill and has been claimed to be one of the design features of human language (Hockett 1960). Accordingly, modality has been extensively studied in linguistics and philosophy (see, e.g. Portner 2009 for an overview). Until recently, work on modality had concentrated primarily on verbal elements. Over the last few years, however, the semantics literature has begun to recognize the existence of a class of functional nominal items –modal determiners– that convey modality. The sentences in (3) and (4) illustrates the phenomenon. The use of the indefinite "some or other" in (3) contributes information about the speaker’s epistemic state (that she does not know what book Juan bought). Thus, this item expresses epistemic modality, like the modal "might" in (1). The Spanish indefinite "uno cualquiera" in (3) expresses random choice modality. It compares what the agent did (e.g. taking card a) with other options (e.g. taking card b, taking card c), and says that, for the agent, all the potential actions count as equivalent to the actual action.

3) Juan bought some book or other.

4) Juan cogió una carta cualquiera.
Juan took a card cualquiera.

This project aims to improve our understanding of natural language modality through the study of modal determiners. Most work on modal determiners has been driven by research agendas orthogonal to the study of modality. Research that deals specifically with modality typically doesn't discuss modal determiners, the implicit assumption being that the study of these items cannot contribute in significant ways to a general theory of modality. The project is driven by the conviction that modal determiners will advance our understanding of modality, as they raise issues that are relevant for current debates on the topic and open up new questions on this domain. Some of the core research questions that the project addresses are:
Q1. How can we characterize the modal meanings conveyed by modal determiners? To what extent do they differ from modal meanings in the verbal domain. Why?
Q2. Modal determiners impose constraints on their interaction with verbal modals. How can these constraints be characterized? How can they be derived?
Q3. Modal determiners do not instantiate all the modal flavours expressed by auxiliaries. What explains these gaps in the paradigm?

1.2. Description of main results

A systematic investigation of random choice determiners has led to the discovery of new patterns of interaction between modal determiners and verbal modals, and to an analysis of those patterns that involves the novel claim that modal determiners are anchored to events (as has been argued for verbal modals in recent work). Overall, the picture that emerges from this investigation is that modal determiners employ the same basic modal projection mechanisms as verbal modals. Another significant finding in this connection was that the interpretation of random choice modality within the determiner domain involves mechanisms that have been claimed to be at work for weak necessity modals like "ought" or "should".

Investigating epistemic determiners has also uncovered new connections between verbal modals and modal determiners. To date, work on this area had centered on items that signal description-based knowledge (i.e. that signal that the speaker knows/does not know a description that uniquely identifies the referent). The project’s results indicate that Spanish cierto conveys a different type of knowledge, signaling that the speaker has familiarity with the referent. This enriches the typology of epistemic indefinites and makes a connection with the epistemic distinction encoded in the two verbs of knowing in languages like Spanish or German ("saber/wissen" vs. "conocer/kennen"), which is evocative of Russell’s classic distinction between knowledge by description and knowledge by acquaintance.

The project has also yielded empirical discoveries in the area of verbal modality: (i) that the epistemic future in Romance displays embedding patterns different from those attested for epistemic modal auxiliaries (like "might"), (ii) that there is a sub-class of doxastic attitude verbs (including canonical doxastic verbs like "believe") that are sensitive to directionality of reasoning, and (iii) that embedded modals can reverse this directionality effect in some cases. These findings are unexpected under current analyses of modal displacement, and should therefore inform a general theory of modality.