Issues in the Semantics and Pragmatics of Disjunction

Mandy Simons

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781315520315
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

First published in 2000, this book is about sentences containing the word or, dealing primarily with sentences in which or conjoins clauses, but also some cases in which it conjoins expressions of other categories. The author aims to give an account of the discourse properties and felicity conditions of disjunction, and to use this account in explaining the behaviour of presupposition projection and of anaphora in disjunctive sentences. The author begins by giving an account of the discourse properties and felicity conditions of disjunction before turning to the presupposition projection problem. The final two chapters discuss anaphora and its interactions with disjunction.

Chapter 1
Introduction

1.1. Overview

1.1.1. What this dissertation is about

This dissertation is about sentences containing the word or. In the sentences I discuss, or occurs in a main clause and is not embedded under any other operators. I will deal primarily with sentences in which or conjoins clauses, but also some cases in which it conjoins expressions of other categories.

The starting point of the dissertation was the observation that not all clauses can sensibly be disjoined. Sometimes, the disjunction of two independently acceptable clauses produces an unacceptable result. Sentence (1), below, is acceptable, but (2) is not. (3) is also odd, but in a different way from (2). (I use the symbol "#" to indicate that a sentence is unacceptable in some non-syntactic way.)

• (1) Jane owns a red truck or she owns a blue truck.
• (2) #Jane owns a truck or she owns a red truck.
• (3) #Jane owns a truck or it is raining in Tel-Aviv.

I began with the intention of characterizing the permissible relations that may hold between disjuncts, and of explaining why disjunction is constrained in this way. At the same time, I began to think about other constraints on the assertion of disjunctions. (4) and (5) are also very odd things to say in a normal conversation.

• (4) Jane owns a truck. #Either she owns a truck or she owns a station wagon.
• (5) Jane doesn' own a truck. #Either she owns a truck or she owns a station wagon.

The oddity of these sentences seems to have something to do with their being redundant, but I wondered whether there could be a connection between the infelicity of these and the infelicity of (2) and (3).

While working on these questions, I began to think that if I could answer them, then I would also be able to resolve some other puzzles about disjunction. One of these has to do with the way presuppositions project in clausal disjunctions. Usually, a disjunction inherits all of the presuppositions of its disjuncts. The exception is when the presupposition of one disjunct is incompatible with the content of another. So (6) inherits from its second disjunct the presupposition that Jane is in town, but (7) does not.

• (6) Either George had a particularly good day, or he knows that Jane is in town.
• (7) Either Jane isn't in town or George knows that she is.

The question is why presuppositions should be "filtered" or "canceled" in just these conditions.

In the framework which I adopt, a presupposition is a proposition that must be assumed as part of the background context in which the presupposing sentence is uttered. So when a speaker asserts a presupposing sentence, what she does is, in a way, equivalent to what she would do by asserting the presupposition, and then asserting the sentence (or a non-presupposing version of it). If I say out of the blue:

• (8) George knows that Jane is in town

the communicative effect is equivalent to my saying:

• (9) Jane is in town, and George knows that she is.

Now, suppose that (7) were to inherit the presupposition. Then to say (7) would be equivalent to saying:

• (10) Jane is in town. Either Jane isn't in town, or George knows that she is.

But this is something a speaker would never say. So if we adopt a view of presupposition which allows presupposition projection to be sensitive to considerations of felicity, then the presupposition projection problem for disjunction comes down to explaining when and why a disjunction can be felicitously asserted.

Another puzzle about disjunction has to do with possible anaphoric relations between an indefinite in one disjunct and a pronoun in another. The problem is illustrated by the contrast between (11) and (12), a contrast originally observed by Barbara Partee and already much discussed in the literature:

• (11) Either there's no bathroom in this house, or it's in a funny place.
• (12) #Either there's a bathroom in this house, or it's in a funny place.

Researchers working on unbound anaphora have been puzzled by the question of why anaphora is possible across disjunction in (11 ), but not in (12). This seems to have something to do with the presence of negation in the first disjunct. But this just makes the matter more mysterious, for normally negation itself blocks anaphora:

• (13) There's no bathroom in this house. #It's in a funny place.

It seemed to me, however, that the infelicity of (12) does not indicate that the pronoun cannot be anaphoric on the NP in the previous clause. The observed infelicity arises because, as the pronoun in fact is anaphoric on that NP, the disjunction (12) has an interpretation equivalent to:

• (14) #Either there's a bathroom in this house, or the bathroom in this house is in a funny place.

(14) itself is infelicitous in the same way as (2) above, but this has nothing to do with anaphora. It is a consequence of the constraints which determine the allowable relations between disjuncts. Thus, if this approach is right, then the solution to the anaphora puzzle is the same as the answer to the question about why certain pairs of clauses cannot be disjoined. But to give this answer, I also had to have a way of explaining how the pronoun in (12) comes to be interpreted as "the bathroom in this house," that is, I needed some theory of cross-clausal anaphora.

The project undertaken here is thus to give an account of the discourse properties and felicity conditions of disjunction, and to use this account in explaining the behavior of presupposition projection and of anaphora in disjunctive sentences. This project, then, would seem to have as its goal the formulation of some set of special properties pertaining to the word or.

However, the project has led in a quite different direction. At the heart of my account is the claim that there is nothing very special about or, and that the only semantic information that must be specified about this lexical item is that it functions as a logical operator equivalent to Boolean join (inclusive disjunction). I will argue that the properties observed can be accounted for in terms of very general principles governing assertoric contributions to discourse, and their interaction with the truth conditional properties of or.

This position echoes the familiar views of Paul Grice. Grice ( 1967) argued that many of the properties of logical operators in natural language are to be explained in terms of general principles governing all cooperative rational interaction, including conversation. The general principle, which he dubs the Cooperative Principle, he formulates as follows:

The Cooperative Principle

Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.

This principle subsumes four Maxims of Conversation: the Maxims of Quality, Quantity, Relation and Manner. The brief formulation of these maxims is as follows:

Quality: Try to make your contribution one that is true.

Quantity: Make your contribution as informative as is required (for the current purposes of the exchange), and no more informative than is required.

Relation: Be relevant.

Manner: Be perspicuous.

Grice urged that those aspects of the behavior of an expression which can be explained in terms of speakers' compliance with the Cooperative Principle not be treated as part of the semantic content of an expression.

Grice's proposal in part informed the model of presupposition and assertion developed by Robert Stalnaker in a series of papers written in the 1970's. The central idea of this model is that every discourse takes place against the background of the assumptions shared by the participants: their common ground. The purpose of an assertion is to change the common ground, to add to the set of shared assumptions. This purpose determines the appropriateness of a given utterance in a given discourse context. Following Grice, Stalnaker suggests that the kinds of information update allowed may be constrained by certain general principles. The Stalnakerian model thus suggests a reframing of the Gricean Maxims as conditions on information update, and also provides a framework for a precise formulation of these conditions. Stalnaker's model will provide the central framework for the development of the ideas in this dissertation.

In Chapter Two, I will use the Stalnakerian model to give a formal characterization of the Gricean Maxims of Quantity and Relation, and will use these as the basis of im account of the discourse properties and felicity conditions of disjunction. The characterization of Relation will require an extension of Stalnaker's original model, with which some of that chapter will be concerned. The conclusions reached in that chapter will provide the foundations for the discussion of presupposition projection and anaphora which follow.

I turn to the presupposition projection problem in Chapter Three. The account is, essentially, that the presuppositions of disjuncts are inherited by the disjunction whenever this does not lead to infelicity. The kinds of infelicity that may be caused by inheritance of presuppositions are just those which are described and explained in Chapter Two. To spell out the account, of course, some theory of presupposition is needed. This is provided, again, by the Stalnakerian model, but makes crucial use of proposals due to Van der Sandt (1992). I argue that the projection properties of disjunction provide evidence for the pragmatic approach to presupposition advocated by Stalnaker, for on this approach it is to be expected that general considerations of felicity will constrain presupposition.

In Chapter Four, I move to the anaphora puzzle. Following Groenendijk and Stokehole (1990), I dub the data discussed in this chapter internal anaphora, as what is involved is anaphora between disjuncts. Once again, the goal is to account for the data in terms of the felicity conditions of disjunction, but in order to do this, I must give an account of the anaphora itself. I thus present a version of the E-type account of cross-clausal anaphora, based closely on that of Neale (1990), and apply it to the disjunction data. This view of anaphora will support the felicity based account of the basic data which I sketched above. I will argue that disjunction imposes no special constraints on anaphora, but that certain cases of anaphora across disjunction result in infelicity. The usefulness of the E-type account, though, will go beyond its facilitation of this kind of explanation. In discussing some more complex instances of internal anaphora, we will find cases in which it is indeed not possible to establish an anaphoric link between an indefinite in one disjunct and a pronoun in another, as in (15).

• (15) #Either most people own a car;. or it;'s in the shop.

This failure of anaphora is predicted by the E-type account adopted, and is entirely parallel to the failure of anaphora in (16):

• (16) Most people own a car_i. #It_i's in the shop.

So the E-type view will enable me to maintain my principal claim, namely, that disjunction imposes no special constraints on anaphora.

In Chapter Five, I depart somewhat from my main theme to pursue further the interaction of disjunction and anaphora. In that chapter, I will discuss the external anaphora data. These data involve anaphora between a disjunction of NPs and a pronoun in a following sentence (as in (17)), and anaphora between NPs contained inside a clausal disjunction and a pronoun in a following sentence (as in (18)).

• (17) A soprano or an alto will sing. She will be accompanied on the piano.
• (18) A soprano will sing Mozart, or an alto will sing Schubert. She will be accompanied on the piano.

The account of these data will not rely on the felicity conditions of disjunction, as do the accounts in Chapters Three and Four. But these data will provide an opportunity to investigate further the E-type account introduced in Chapter Four. In fact, the external anaphora data will lead me to propose a new version of the E-type account, in which the interpretations of E-type pronouns are derived compositionally from the content of the antecedent clause. This proposal constitutes a significant departure from existing formulations of the E-type strategy.

1.1.2. Context change: Pragmatic vs. semantic approaches

The Stalnakerian model in terms of which I will frame my account was originally conceived as a pragmatic model, intended to provide a basis for stating "some general rules of conversation" (Stalnaker 1973: 450). This is how I shall understand the model, and how I shall use it.

In other frameworks, the Stalnakerian idea is transmuted into a semantic theory. These dynamic semantic theories are developments of the proposals of Kamp (1981) and Heim (1982). The central thesis of these theories is that the meaning of an expression resides in its potential to change the information state of a hearer. The theories are thus closely related to the pragmatic Stalnakerian model, but differ from it in seeing context change as a semantic phenomenon. In Dynamic Semantics, the potential context change effect of an expression constitutes its semantic value. Consequently, the job of the semantics is to specify context change operations as values for expressions.

Dynamic semantic theories have been proposed primarily to account for cross-clausal anaphora and for presupposition projection, two issues on which this dissertation focuses. The dissertation thus offers an opportunity for a critical discussion of the semantic view of context change. Throughout the dissertation, I will discuss dynamic semantic accounts of the data, and will compare and contrast my own accounts with them. I will focus on the difference between the kinds of explanation offered by dynamic semantic theories, and the kinds of explanation which can be given on the basis of a pragmatic view of context change. I will argue that the pragmatic view provides more satisfactory accounts of the phenomena in question.

In the remainder of this introduction, I will present the theoretica...

Table of contents

• Cover
• Half Title
• Title
• Copyright
• Original Title
• Original Copyright
• Dedication
• Contents
• Preface
• 1. INTRODUCTION
• 2. DISJUNCTIVE SENTENCES IN DISCOURSE
• 3. PRESUPPOSITION PROJECTION
• 4. INTERNAL ANAPHORA
• 5. EXTERNAL ANAPHORA
• CONCLUDING REMARKS Grice, Stalnaker and Dynamic Semantics
• Appendix
• Bibliography
• Index