Donald Davidson's Triangulation Argument

A Philosophical Inquiry

Robert H. Myers, Claudine Verheggen

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781134641291
• Edition: 1
• Topic: Philosophy
• Subtopic: Ethics & Moral Philosophy

According to many commentators, Davidson's earlier work on philosophy of action and truth-theoretic semantics is the basis for his reputation, and his later forays into broader metaphysical and epistemological issues, and eventually into what became known as the triangulation argument, are much less successful. This book by two of his former students aims to change that perception. In Part One, Verheggen begins by providing an explanation and defense of the triangulation argument, then explores its implications for questions concerning semantic normativity and reductionism, the social character of language and thought, and skepticism about the external world. In Part Two, Myers considers what the argument can tell us about reasons for action, and whether it can overcome skeptical worries based on claims about the nature of motivation, the sources of normativity and the demands of morality. The book reveals Davidson's later writings to be full of innovative and important ideas that deserve much more attention than they are currently receiving.

Part One

Language, Thought and Knowledge

Claudine Verheggen

1 Davidson’s Triangulation Argument

1. Introductory Remarks

As has often been remarked, there is no such thing as “the” triangulation argument explicitly laid out in Davidson’s writings. The closest he comes to laying out an explicit argument is in “The Second Person” (Davidson 1992), the first paper in which he develops and defends the idea of triangulation,¹ which I shall refer to in what follows as the initial article. But the argument in the initial article is at some points difficult to interpret, and subsequent discussions only add to the complexity. Most commentators and critics have concluded that there are actually (at least) two triangulation arguments to be unearthed in Davidson’s writings—one for the conclusion that triangulation is required to fix the meanings (i.e., propositional contents, on which more soon) of one’s thoughts and utterances, the other for the conclusion that triangulation is needed to have the concept of objectivity, which, for Davidson, is also needed to have a language and thoughts (in addition, that is, to the obvious requirement that meanings be fixed). As we shall see, separating the arguments in such a way has made it much easier to reject them. However, though I grant that Davidson’s writings sometimes support this interpretation, I think that they are better understood, both exegetically and philosophically, as supporting a different kind of reading, according to which there is only one complex, multifaceted argument for the general conclusion that triangulation is needed for the possession of language and thoughts. This conclusion is reached by keeping the two requirements tightly connected, and thus by arguing that the meanings of one’s thoughts and utterances can be fixed only if one has the concept of objectivity, possession of which requires triangulation.² Here is a passage in which Davidson makes the connection pretty apparent:

A grasp of the concept of truth … depends on the norm that can be provided only by interpersonal communication; and of course … the possession of any propositional attitude … depends on a grasp of the concept of objective truth.

(Davidson 1994, 124)

Thus, according to my reading, the elements of what are often taken to be two arguments for two different conclusions are meshed in a single, integrated argument. As a result, the triangulation argument becomes much harder to resist, or so I hope to demonstrate in this chapter.

I shall start with a few clarifying remarks about what Davidson means, first, by “concept of objectivity” and, second, by “propositional content”.

Davidson refers to the concept of objectivity interchangeably as the concept of objective truth or, sometimes, as above, just truth, or the concept of belief, or the concept of error, or the belief-truth contrast.³ To have the concept of objectivity is to have the “awareness, no matter how inarticulately held, of the fact that what is thought [or said] may be true or false” (Davidson 1995a, 4). It is to be aware of “the distinction between what is believed or seems to be the case and what is objectively so” (Davidson 2001c, 294), that is, what is so independently of anyone thinking or speaking about it, between “what is asked or demanded and what is answered or done” (Davidson 1998, 86).⁴ So Davidson’s argument, as I read it, will be that the meanings of one’s thoughts and utterances can be fixed only if one has this sort of awareness of the contrast between belief and truth, and that one needs to have triangulated with others in order to have this sort of awareness.

I glossed the ‘meanings’ of thoughts and utterances above as their ‘propositional contents’, and I will use mostly ‘meanings’ to talk even of the contents of thoughts as a constant reminder that it is propositional thoughts that I am discussing—that is, thoughts with propositional content. Propositional content is not something whose nature Davidson ever made terribly explicit even though he acknowledged that different philosophers may have different conceptions of propositional thought (Davidson 2003, 698). But it is crucial for what follows to distinguish Davidson’s notion of propositional content from other notions of content, some of which have become much in vogue in the past few years, but all of which were largely neglected by Davidson. These are, to cite the dominant ones, the notions of non-conceptual content, intentional content, conceptual content, perceptual content, representational content and, perhaps the most recent, intuitional content. I will not even try to explain what these are, especially since very few philosophers agree on any one understanding of them.⁵ But I will try to say what Davidson understands by propositional content.

Note first that Davidson would use the notion interchangeably with that of conceptual or intentional content. However, though everyone may agree that all propositional content is conceptual and intentional, not everyone agrees that all conceptual or intentional content is propositional.⁶ Propositional content belongs, in the first instance, to propositional attitudes and thoughts. Davidson is indeed happy to say that propositions make up their content, just as they make up the meaning of sentences, provided, however, that no ontological commitment is implied by this (Davidson 1988b, 45). There is, for Davidson, a pretty strict parallel between the contents of propositional attitudes and thoughts and the meanings of sentences (Davidson 1989a, 57). The former are composed of concepts that contribute to their content, just as the latter are composed of terms whose meanings contribute to the meaning of the sentences in which they occur. Again, though, no ontological commitment concerning concepts or meanings follows from this. These have no existence apart from the communicative acts out of which they are abstracted.⁷ I think that, for Davidson, three features characterize propositional attitudes and thoughts and hence propositional content. (From now on, I will often drop ‘propositional’ in ‘propositional thought’ and ‘propositional content’.)

First, content of this kind is both attributed and constituted holistically. According to Davidson, all thoughts require beliefs, for the content of a thought is identified by its location in a pattern of beliefs, just as the meaning of a sentence is given by a pattern of sentences held true, i.e., believed (Davidson 1975, 162). Davidson uses ‘identify’ and ‘give’ here seemingly interchangeably because, I think, he means to be describing both the way content is attributed to someone’s thoughts and how it is constituted (fixed, determined) in someone’s thoughts. Thus, to use one of Davidson’s examples, one cannot attribute to another a thought about cats unless one can attribute to her some general beliefs about cats, such as the belief that cats can scratch or climb trees, some particular beliefs about cats, such as the belief that the cat seen running a moment ago is still in the neighborhood, and some logical beliefs, such as the belief that either the cat seen running a moment ago is still in the neighborhood or it is not (Davidson 1982, 99). No fixed list of beliefs of the kinds just described is necessary, but numerous beliefs of these kinds must be attributed before any thought about cats can be attributed. As Davidson summarizes it, “the system of such beliefs identifies a thought by locating it in a logical and epistemic space” (Davidson 1975, 157). And, to repeat, what holds of content attribution also holds of content constitution.⁸

The second feature that characterizes propositional attitudes and thoughts is that they can be expressed or reported using sentences that have truth-conditions and are thus true or false. According to Davidson, the core of meaning and content is truth-conditional. He never gave up the claim that reflecting on the construction of Tarski-style theories of truth, that is, theories from which we could derive the truth-conditions of a speaker’s utterances, is the key to understanding the philosophical nature of meaning. He did wonder “to what extent such theories can be made adequate to natural languages”, but never doubted that “they are adequate to powerful parts of natural languages”. It is not that he thought that “speakers and interpreters actually formulate such theories”, but rather he maintained that, “if we can describe how they could formulate them, we will gain an important insight into the nature of the intentional (including, of course, meaning)” (Davidson 1993a, 83–4). Thus we might say that the meaning of words, and what concepts they express, is, in the first instance, a matter of how they contribute to the truth-conditions of the sentences in which they occur, or of the beliefs these sentences express.

Now, for the sentences giving the truth-conditions of a speaker’s utterances to be truly meaning-giving, one must be careful how these truth-conditions are expressed. For instance, “ ‘Hesperus is identical with Phosphorus’ is true iff Hesperus is identical with Hesperus” would not be meaning-giving, for presumably the meaning of ‘Hesperus’ is different from that of ‘Phosphorus’, as revealed by the fact that someone may fail to believe that Hesperus is identical with Phosphorus. The axiom that eventually tells us how a name contributes to the truth-conditions of the sentences in which it occurs is extensional, in the sense that it picks out an entity in the world which, in my example, could be named either Hesperus or Phosphorus. However, if we want to make sense of someone’s saying that he just found out that Hesperus is identical with Phosphorus, we will make sure that the axiom for ‘Hesperus’ says that this word designates Hesperus, and not Phosphorus.⁹ This is to say that, for Davidson, meaning and content are, as might be expected, intensional, even though they may be fixed, at least in part, by items in the environment of speakers and thinkers or, as Davidson puts it, even though they “supervene on the observable and non-intentional” (Davidson 1993a, 84).¹⁰ This is of a piece with the third feature characterizing propositional attitudes and thoughts.

This is that sentences attributing propositional attitudes and thoughts are semantically opaque.¹¹ We cannot substitute co-referring expressions in these attributions without running the risk of changing their truth-value. Thus, though it is true that I think my cat is out, it may not be true that I think the most aggressive cat in the neighborhood is out, even though my cat is the most aggressive cat in the neighborhood. This third feature emphasizes the idea I expressed in the previous paragraph, which is that there is more to meaning and content, even of basic utterances and thoughts, than mere connections between items in the world and the words and concepts that refer to them. That is, we cannot fully account for meaning and content in extensional terms. As Davidson came to say repeatedly, the intensional cannot be reduced to the extensional,¹² even though he never thought that meaning and content could be explained in terms of intensions conceived of as abstract entities.

In short, then, propositional content is holistic, truth-conditional and irreducibly intensional. My intention so far has not been to defend these claims. Though I believe they have considerable intuitive appeal, they should for now be accepted at face value as characterizing the notion of content that Davidson is interested in and which the triangulation argument is concerned with. They will in any case become clearer and, I hope, be vindicated later, starting with my exposition of the triangulation argument in the next section of this chapter.¹³ I shall proceed with this exposition from a historical point of view, so to speak, tracing the argument back to the article in which it was first developed and following it through the main ensuing articles, for not all the ingredients were present in the initial piece. The second part of the chapter will be devoted to a discussion of the core objections that have been made to various facets of the argument. These are: first, that triangulation is not needed to fix the meanings of one’s thoughts and utterances; second, that triangulation is not needed to have the concept of objectivity; and, third, that one need not have the concept of objectivity in order to be a thinker and speaker. I shall end by addressing a more formal objection that has been made repeatedly to the argument, viz., that the account it gives of what makes thought and language possible is circular.

2. The Triangulation Argument

[Triangulation consists of] the mutual and simultaneous responses of two or more creatures to common distal stimuli and to one another’s responses.

(Davidson 2001a, xv)

This quotation suggests that triangulation is a rather vague affair, which may take different, more precise, forms in different contexts. This may explain in part why so many commentators have taken there to be different triangulation arguments, perhaps meant to establish different conclusions. The first thing to note is indeed that there are two kinds of triangulation that Davidson has in mind, one pre-conceptual and the other fully linguistic, between which he unfortunately does not always sharply distinguish. The first kind, which he often calls “primitive”,¹⁴ “does not require intensional attitudes” (Davidson 2001c, 292). “It is the result of a threefold interaction, an interaction which is twofold from the point of view of each of the two agents: each is interacting simultaneously with the world and the other agent” (Davidson 1999a, 128).¹⁵ It is the kind of triangulation in which non-linguistic creatures can engage, as in Davidson’s example of two lionesses trying to catch a gazelle and coordinating their behavior by watching each other and the gazelle and reacting to each other’s reactions (Davidson 2001b, 7). The second kind of triangulation involves full-blown linguistic interaction. It is the kind of situation in which the learner of a first language eventually finds herself, once she has the concept of objectivity. It is also the kind of situation in which the radical interpreter—in effect the learner of a second or additional language in situ—finds herself.¹⁶ We may add that it is the kind of situation in which we sometimes find ourselves, whenever ostension is needed to determine the cause, and hence the meaning, of a speaker’s utterance. Both kinds of triangulation are supposed to be necessary for someone to be a thinker and speaker, but only the second kind is sufficient—indeed, one might say redundantly so; we shall come back to this in Section 3. For now, let us ask why the first kind is necessary to begin with. What first prompted Davidson’s idea of triangulation?

2.1 Primitive Triangulation

The triangulation argument is premised on Davidson’s perceptual externalism, the view “that the contents of our thoughts and sayings are partly determined by the history of causal interactions with the environment” (Davidson 1991a, 200).¹⁷ Thus, take the thought, ‘there’s a cow’. “What determines the content of such basic thoughts (and what we mean by the words we use to express them) is what typically caused similar thoughts” (Davidson 1991a, 201). The question is, though, what are the typical (or “normal” or “usual” (Davidson 2001b, 4)) causes of our basic thoughts and sayings? Davidson claims that if we were to observe a person who had never interacted with another and the world they share—call this person a solitaire—we could not answer this question. And, crucially, neither could she, for the problem here is not that we could not verify what the causes of her thoughts and utterances are, but that such causes could not be uniquely determined (Davidson 1992, 119).¹⁸ Just as the stimulus causing the dog to salivate could be said to be the ringing of the bell or the vibration of the air close to its ears, and just as the stimulus causing the child to mouth ‘table’ could be said to be a table in its vicinity or the pattern of stimulations at its surface (Davidson 1992, 118), so the stimulus causing the solitaire to produce a sound apparently in response to her environment could be said to be “anything from the stimulation of [her] nerve endings to the original big bang” (Davidson 2001b, 4).¹⁹ And of course, in the absence of any determinate answer to the question of what caused someone’s thoughts or utterances, given perceptual externalism, there could be no answer to the question what she is thinking or saying, indeed, no reason to believe that she is thinking or saying anything. According to Davidson, matters change dramatically if the solitaire interacts with another person. They change dramatically if, specifically, the (no longer) solitaire interacts with an interlocutor as well as with objects or events in her environment with which her interlocutor is also interacting. Then it becomes possible to start answering the question what the causes of her utterances are. For any such utterance, the cause is what can be found at the intersection of the two lines that may be drawn from eac...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• Contents
• Acknowledgements
• Introduction
• Part I Language, Thought and Knowledge
• Part II Desires, Reasons and Morality
• Bibliography
• Index