Philosophy of Language and Webs of Information

Heimir Geirsson

• Publisher: Routledge
• Year: 2013
• Print ISBN: 9780415640657
• eBook ISBN: 9781136180187
• Edition: 1
• Topic: Philosophy
• Subtopic: Linguistic Semantics

The nature of propositions and the cognitive value of names have been the focal point of philosophy of language for the last few decades. The advocates of the causal reference theory have favored the view that the semantic contents of proper names are their referents. However, Frege's puzzle about the different cognitive value of coreferential names has made this identification seem impossible. Geirsson provides a detailed overview of the debate to date, and then develops a novel account that explains our reluctance, even when we know about the relevant identity, to substitute coreferential names in both simple sentences and belief contexts while nevertheless accepting the view that the semantic content of names is their referents. The account focuses on subjects organizing information in webs; a name can then access and elicit information from a given web. Geirsson proceeds to extend the account of information to non-referring names, but they have long provided a serious challenge to the causal reference theorist.

1

Introduction and Overview

Suppose that back in the day when Ziggy Stardust was still performing, you and I attend one of his concerts. After the show we agree that Ziggy puts on a good show, that is, we both believe

1.	Ziggy Stardust puts on a good show.

There is a sense in which we believe the same thing, namely, we accept and believe the same proposition. When we start discussing the show, it turns out that, quite apart from the fact that our criteria for a good show diverge, we think about or represent Ziggy in different ways. That is something that we might have expected, as this was my first Ziggy concert, while you have been a fan for a long time and have attended a number of his events. While we seem to have the same belief, that is, while we seem to believe the same proposition, we seem to represent Ziggy in significantly different ways. There is a sense in which our beliefs are not the same, which we can capture by saying that we believe the proposition in different ways.

As we continue our discussion, you proclaim that David always puts on a good show. “David who?” I ask. You repeat,

2.	David Bowie puts on a good show.

As a relative newcomer to music, I do not know who you are talking about. I do not say anything, but conclude that you believe something that I do not believe, namely, that David Bowie puts on a good show. Later in the day you report to a mutual friend

3.	Heimir believes that David Bowie puts on a good show.

I hear you say this and object. You continue and claim to be right, since it is true that

4.	Heimir believes that Ziggy Stardust puts on a good show.

As it turns out, you are correct. The following pages comprise an attempt to provide a systematic account of why you are right. Doing so involves discussing the nature of propositions, the nature of reference, the nature of de re beliefs or beliefs about objects, as well as substitutivity of names.

Philosophers have generally argued that we can substitute coreferential names in simple sentences, such as (1) and (2), while preserving truth value. Following Frege, it was common to argue that we cannot substitute coreferential names in attitude contexts, such as (3) and (4), while preserving truth value. Frege’s argument against such substitutions depended on the intuitions he had about the cognitive significance of names and resulted in his view of names as having a descriptive meaning that they contribute to propositions expressed by sentences in which they occur.

The introduction of the causal reference theory gave philosophers some of the tools that allowed them to rethink the widely accepted Fregean view on substitutions. The initial impetus for rethinking substitution was that the causal reference theory provided a radically different account of the meaning of names; instead of the Fregean view that names have a descriptive meaning that secured reference, the causal reference view held that names do not have descriptive meanings that secure reference. Instead they secure their reference via a causal chain to the object named. That, in turn, seemed to imply that names could no longer be thought of as contributing descriptive contents to propositions. Instead, the suggestion was, they contribute to propositions the object referred to, giving us singular propositions. But the suggestion that names contribute to propositions the object referred to did not seem to provide the intuitive results regarding substitution for sentences like (3) and (4), and so the focus of attention quickly became substitution in sentences embedded in attitude contexts and, more specifically, belief reports. It is here, I believe, that we started down the wrong path. It looks as if we forgot what Kripke had pointed out in the preface to Naming and Necessity, namely, that even though the contribution of the names “Hesperus” and “Phosphorus” to propositions (if indeed we could retain talk of propositions) is the same, the sentence “Hesperus is Phosphorus” can sometimes be used to raise an empirical issue that the sentence “Hesperus is Hesperus” cannot be used to raise. That is, substituting coreferential names can affect our epistemic appraisal and attitude towards the relevant simple sentences (i.e., sentences that are not embedded in attitude contexts).

Most philosophers want to claim that we need to treat our intuitions against substitution for simple sentences differently than our intuitions against substitutions for embedded sentences. Consequently, a number of philosophers have spent a great deal of time devising a solution to the substitution problem for embedded sentences without considering that there is also a substitution problem for simple sentences. In some instances, the result has been that a solution that seems to be promising when dealing with embedded sentences does not transfer at all to simple sentences. Some attempted solutions postulate one kind of a proposition for belief context and a different kind of a proposition for simple sentences. Almost all attempted solutions have started by focusing on embedded sentences, either ignoring simple sentences or providing solutions that cannot readily be applied to simple sentences. This approach, I believe, is fundamentally wrong. The more natural approach is to start with simple sentences and try to provide a unified solution: a solution that works for both simple sentences and embedded sentences.

A second mistake, I argue, is to focus exclusively on names and their meanings. It seems evident that the substitution problems are not unique to instances that involve names (and indexicals), as the various puzzles can easily be reproduced using, for example, pictures as a referring device. While the typical problem cases involve substitution of names and so focus on linguistic items, a solution to the substitution problem needs to be applicable to nonlinguistic representations of various kinds. Consequently, any solution that, for example, depends on including linguistic items in propositions fails what I call the generality constraint.

Very briefly, we need to acknowledge that we represent objects in a variety of ways and not just linguistically. I suggest that we gather information that we take to be about the same object into a web of information about that object, and that we then elicit information from the web in various ways. The webs of information are similar to what some have called dossiers or files with the exception that the webs allow in an intuitive way for some information to be more central than other information. One (but not the only) way to elicit information from a web is via a name that we take to be of that object. In some cases, we have different webs for the same person or object, especially if that person plays significantly different roles, or has significantly different guises. For example, it is appropriate to use different webs of information for Ziggy Stardust and David Bowie even though I know that Ziggy is Bowie, or, more accurately, a character played by Bowie. Once we assume that people have similar information in their webs of Ziggy and Bowie, it makes a significant difference whether I use the name “Ziggy” or the name “Bowie” in a sentence, for the two names elicit information from different webs. If I tell a friend that Ziggy will never perform again, then she would be mistaken if she assumed that Bowie will never perform again. We can thus elicit significantly different information with two sentences that express the same proposition. Once we have this basic picture we can apply it to belief reports as well, thus providing a unified approach to simple sentences and embedded sentences.

While the general picture argued for is rather simple, a number of issues have to be addressed before presenting it. So, here is a brief outline of the chapters that follow.

2. Reference. This chapter includes a general introduction to reference and substitution that includes an account of Frege’s theory, his motivation for introducing descriptive meanings, and how he proposed that we deal with substitutions in both simple sentences and in attitude contexts. Following a critical assessment of Frege’s theory, the causal reference theory is introduced and discussed. Special attention is given to Kripke’s epistemic argument against the description theory and recent discussion of that argument.

3 Propositions: Structure and Objects. In the introduction to Naming and Necessity, Kripke wondered whether we could preserve propositions given some of the challenges raised by his rejection of the description theory. While philosophers generally agree that the causal reference theory does not force us to abandon propositions, Frege’s way of thinking about propositions came under intense pressure. Instead of names contributing descriptive meanings to propositions as Frege had envisioned, a popular and widely discussed alternative is that names do not contribute such meanings. But if names do not contribute descriptive meanings to propositions, then what do they contribute, and what is the nature of the propositions expressed by simple declarative sentences containing names? As it turns out, the solutions to a number of the problems that prompted Frege to introduce his brand of propositions, as well as solutions to the problems raised against the causal reference view, introduce different kinds of propositions. This chapter discusses Fregean propositions, as well as several types of propositions introduced by causal reference theorists when trying to solve some of the problems faced by the theory. The best alternative for the causal reference theorist remains to adopt direct reference, where a referring term in a simple sentence contributes to the expressed proposition the object referred to, giving us structured singular propositions consisting of ordered objects and properties.

4. Reporting Attitudes. Discussion of attitude reports demanded a good deal of interest and effort among those attacking and defending direct reference. This chapter explains why that is the case and critically evaluates several of the most prominent accounts of attitude reports, including Frege’s account; the pragmatic implicature account of Nathan Salmon; the linguistically enhanced propositions introduced by Mark Richard; the hidden indexical account advocated by Mark Crimmins; and Scott Soames’s descriptively enhanced beliefs and assertions. The accounts discussed, I argue, focus on trying to solve the wrong problem. We have antisubstitution intuitions for both belief contexts and for simple sentences, that is, sentences that are not embedded. The accounts discussed generally fail to extend to simple sentences.

5. Singular Propositions and Acquaintance. Advocates of the direct reference theory have made it very easy to believe singular propositions. Generally, they have held that if one sincerely assents to a sentence expressing a singular proposition, then one believes that proposition. But while the easy access to singular propositions is now the norm, it was not always like that. This chapter discusses an argument of Donnellan in which he focuses on cases where names are introduced without the person who introduces the name being acquainted with the object named, such as when Leverrier introduced the name “Neptune” as the name of the object that was the cause of the perturbations of the orbit of Uranus. I develop Donnellan’s claim and argue that it is much harder to believe singular propositions than most direct reference theorists want to admit. In order to believe a singular proposition, I argue, one needs to be acquainted with the object in the proposition in the appropriate way.

6. Beliefs and Belief Reports. According to naïve Russellianism, we can substitute coreferential names while preserving truth value. While endorsing naïve Russellianism, I acknowledge that we do have antisubstitution intuitions that have to be taken into account. Those intuitions do not focus on information contained in a sentence, but rather on information elicited by a sentence. Because the information elicited by a sentence can differ depending on which name of an object is used, we generally cannot and should not freely substitute names in simple sentences even though truth value is preserved when doing so. The framework that allows us to account for this includes webs of information in which one stores information one takes to be about an individual or a persona. The name used in a sentence helps one elicit information from a given web, and so it makes a difference whether, for example, one uses the name “Ziggy” or the name “Bowie,” as the different names will elicit information from different webs of information. The account is easily extended to belief reports. When reporting beliefs, we assume that other people have similar basic information in their webs as we do, and that affects our intuitions when we substitute names in belief reports. Even when someone is in the know about a relevant identity, I argue that, in general, we still should not substitute, and so the issue at hand is not one of mistaken evaluation, as David Braun and Jennifer Saul have argued. While it is to an extent indeterminate what information is being provided when uttering sentences and reporting beliefs, that can be explained by talk about what is compatible with the information in our respective webs of information and not in terms of one asserting a number of propositions with a single utterance, as Scott Soames has argued.

7. Empty Names. The naïve Russellian claims that simple sentences express singular propositions: propositions that have as a constituent the object to which the name in the simple sentence refers. Accordingly, the sentence “Kasparov is a grandmaster” expresses a proposition that contains Kasparov as well as the property of being a grandmaster. But what about sentences such as “Santa is fat?” The name “Santa” fails to refer and so there is no object that it can bring to the proposition expressed by a simple sentence in which it occurs. One plausible way of dealing with this is to claim that “Santa is fat” expresses an incomplete proposition. Several issues then remain. What truth value should we assign to such a proposition? Is there a difference between different types of empty names, namely, fictional names, mythical names, and names of imaginary objects? And how can we account for the cognitive difference between “Santa is fat” and “The Tooth Fairy is fat,” assuming that the two sentences express the same incomplete proposition? I will argue that incomplete propositions are neither true nor false, and will then extend the account from the previous chapter to account for the information value of sentences containing empty names.

8. Attitude Contexts: Belief and Justification. The resistance towards naïve Russellianism can be further reduced. It is common to argue that naïve Russellianism clearly gives us counterintuitive and wrong results when substituting in knowledge contexts. Clearly, the argument goes, the Babylonians did not know that Hesperus is Phosphorus even though they knew that Hesperus is Hesperus. Here I argue that, once the direct reference theorist accepts that one can believe propositions in different ways, she needs to abandon the view that that one is justified in believing a proposition simpliciter. Instead, one can be justified in believing a proposition when believed one way and at the same time not be justified in believing the same proposition when believed in a different way. While believing a proposition is simply standing in the appropriate relation to it, how it is believed affects justification. Because a proposition’s justificatory status, and hence whether one knows it, depends on how one believes it, the Babylonians could very well be justified in believing that Hesperus is Hesperus without being justified in believing that Hesperus is Phosphorus. Not surprisingly, the argument entails that we need to relativize a priori knowledge to how the given proposition is believed. The Babylonians knew a priori that Hesperus is Hesperus when the proposition was believed appropriately, while they did not know a priori that Hesperus is Phosphorus, given how they believed the proposition before discovering the identity.

It is important to note that this work is significantly less technical than other books on the issue. The focus is on the broad picture, and specific language and logic puzzles have to give way. As a result, this work should be significantly more accessible to the nonspecialist than most other works on these issues. While the target audience is professionals, advanced undergraduate students should also be able to benefit significantly from the text.

2

Reference

In the 1960s and 70s, a number of philosophers started to advance a view of reference and meaning of proper names that was strongly at odds with the generally accepted theory. The theory that was generally accepted at the time was the description theory of proper names, according to which reference was determined by the descriptive meaning of a name. The description theory holds that proper names have both meaning and reference and, further, that their reference is determined by their meaning. When Gottlob Frege initially presented this theory, he suggested that a proper name has a sense or a meaning, which is how we think of the object. The meaning is the cognitive content or the mode of presentation associated with the expression we use to refer to the object. While Frege’s view is compatible with various ways of thinking about an object, he is typically interpreted as having the mode of presentation associated with a name be a descriptive meaning: thus the description theory of names. For example, the meaning of the name “Alexander the Great” might be given by the description “Aristotle’s most famous student.” When I utter a sentence containing the name “Alexander the Great”, then the name is only a shorthand for its meaning. The sentence “Alexander the Great conquered the Middle East” thus means, when unpacked, the same as the sentence “Aristotle’s most famous student conquered the Middle East.” The description picks out the person who was Aristotle’s most famous student, and that is the person referred to with th...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright
• Dedication
• Contents
• Acknowledgments
• 1. Introduction and Overview
• 2. Reference
• 3. Propositions: Structure and Objects
• 4. Reporting Attitudes
• 5. Singular Propositions and Acquaintance
• 6. Beliefs and Belief Reports
• 7. Empty Names
• 8. Attitude Contexts: Beliefs and Justification
• Notes
• Index