Thought and Language

J. M. Moravcsik

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781315524115
• Edition: 1
• Topic: Filosofía
• Subtopic: Lenguaje en filosofía

Originally published in 1990, this book centres on a certain way of surveying a variety of theories of language, and on outlining a new proposal of meaning within the framework set by the survey. One of the key features of both survey and proposal is the insistence on the need to locate theories of language within a large framework that includes questions about the nature of thought and about general ontological questions as well. The book deals in an interconnected way with both very general and specific issues. At one end of this spectrum there are discussions of the contrast between realist and nominalist ontologies, while at the other are analyses of specific lexical items of English.

PART ONE PROBLEMS

CHAPTER I

Ontology

Tersely stated, ontology attempts to answer the twin-questions: ‘What exists?’ and ‘What is it for something to exist?’ These questions surface also in the form in which ‘exists’ is replaced by ‘is real’. Not surprisingly, some philosophers think that the two formulations are equivalent, while others disagree. These questions are not so far removed from the level of common sense as it might seem at first glance. Dreams, bad conjectures, and superstitions, among other things, often make us ask about some putative entity: ‘Is this real?’ At times we end up with a contrast between what is real and what only seems to be, and at other times we juxtapose what exists but is only a surface phenomenon with a fundamental underlying element of reality. We raise these questions and form these contrasts in contexts of specific investigations or in particular psychological states. Ontology borrows the contrast of appearance and reality from common sense and projects it across the cosmos.

One can trivialize the ontological question by answering it with long lists of specific times and kinds. The philosopher, on the other hand, wants as answers not endless lists of specifics, but broad categories of reality. The basic ontological categories, though any one of these can be found at times denounced as not genuine and at other times passionately defended, remain surprisingly constant through the history of western philosophy. These include: the abstract and the spatial, the recurrent and the non-recurrent, the atemporal and the temporal, the actual and the possible or necessary. We shall concentrate on only a few of these debates, examining those that are most relevant to the theories of thought and language that will be presented in the subsequent chapters.

The cosmic projection of the appearance versus reality distinction leads philosophers not only to debates about what exists and what does not, but also about what are more fundamental and what are less fundamental elements of reality. The former are supposed to explain the latter. Thus some ontological theories are not mere inventories but consist of proposed explanatory patterns. Modern ontologies like the realism of Moore and Russell, dealt with at pp. 138–40, are more like inventories, while the early classical ontologies of Plato and Aristotle concentrate on ontological explanatory structures. Plato’s Forms are not only real, but also have a fundamental ontological role. Their existence and nature explain and make possible the existence of everything else. Aristotle’s substances have the same status in that ontology. But even in the more inventory-like ontologies, positing two or three fundamental categories of reality within which one is to capture the diversity of what there is, amounts to an attempt at clarification, making the many one, and thus bringing order into the chaotic.

We shall introduce the notion of ontological priority to capture the distinction between the more and less fundamental. Entity A is ontologically prior to entity B if and only if the existence of B depends on the existence of A, but not the other way around. This notion brings out the sense in which Plato took Forms to be prior to all else while Aristotle assigned the same honorific place to his substances. The philosophical dependency proposals are conceptual. Hence there may be situations in which one conceptual dependency relation makes A prior to B, while another one reverses the order. Another problem with this notion is that not all conceptual relations count as ontologically relevant dependency relations. There is, however, no general way of characterizing the conceptual dependency relations that all philosophers regard as ontologically relevant. Different schools have different views on this issue.

Though the two questions: ‘What exists?’ and ‘What is it for something to exist?’ are conceptually distinct, philosophers tended to link the issues together. If a philosopher is convinced that only material things exist, it is easy to lapse into a position according to which to exist is to be material. Or, to take an historical example, to think that to exist is to be perceived. It is clearly desirable, however, to keep the questions distinct, and to formulate an account of what it is to exist that is neutral with regard to the various ontological theses. The advantage of doing ontology in this way is that one has a neutral conceptual background against which rational examination of competing ontological theses can take place. If, for example, materialists and anti-materialists mean different things by their uses of ‘exists’, then their ontologies would be difficult to compare, and could even turn out not to be in conflict with each other. Such an interpretation would, however, distort the historical facts. Such distortion would also be involved if we said that materialists and anti-materialists argue about what it is to exist. Few if any such arguments can be found in the historically influential presentations of the materialist position.

To some the task of explicating what it is to exist seems too formidable. They advise us to try our hand at the seemingly more modest enterprise of providing criteria for what it is for a theory to be ontologically committed to the existence of certain kinds of entity. Additional criteria can then be added to evaluate theories in terms of more or less adequate ontological commitments. This is the programme embraced some time ago by Quine. In his earlier writings he attempted to give an explication of ontological commitment that was supposed to be neutral between rival ontological theories and would not commit us to entities that everyone would not accept as real. Quine’s criterion, however, only applies to theories that can be translated into his ‘canonical language’ which is, roughly, first-order predicate calculus with certain interpretations (Quine, 1953a). Unfortunately, even this seemingly more modest enterprise ran into trouble. Quine’s proposal was tied to certain technical notions such as that of a variable and the notion of quantification, which make sense only within certain formalisms. Not surprisingly, the difficulties turned out to be technical as well. Detailing these would take us too far from the main purpose of this work. Some of the problems centred on the issue of whether the proposed criterion was as ontologically neutral and uncommitted to abstract entities as Quine thought it to be (Chomsky and Scheffler, 1958). Other difficulties emerged when one considered the proposed criterion in the light of formal results in logic about translatability. It turned out to be the case that in the light of these results and Quine’s criterion, the best ontology one in which only numbers were acknowledged as real – a conclusion that pleases the Pythagoreans, if there are any left, but hardly Quine. In his subsequent writings, Quine introduced various further technical notions to avoid this conclusion; the jury is still out on that one (Quine, 1969).

Let us return to the project of formulating a neutral criterion of what it is to exist. We shall carry out the project in an informal intuitive way, not tying the results to formalisms. This is, after all, the way in which most of the important ontologies have been presented in western thought. In fact, most of the ontological positions were invented before the formalism of symbolic logic within which Quine’s proposal is couched came into existence. The intuition guiding the proposal is that what exists must have a qualitative nature, and likewise, what has a qualitative nature must exist. Thus what exists must be the subject of true propositions other than simply existential ones. Furthermore, what does not exist cannot be the subject of true qualitative propositions. If we had the collection of all true singular propositions, then we would also have a handle on all that exists. The proposal is, therefore, that a necessary and sufficient condition for what it is to exist is to be the subject of some true propositions. Giving this characterization of existence in terms of a necessary and sufficient condition does not commit us to further views about the full meaning of ‘exists’. It is easy to see that this analysis is indeed neutral between materialism and its foes, or an ontology of abstract entities and its critics.

Unfortunately, the proposal faces several difficulties only some of which can be met by adding qualifications. First, there are many views about what it is to be a subject. There are grammatical, semantic, and logical criteria, and these do not coincide. What is the grammatical subject of a singular sentence may not be its subject on the basis of semantic analysis (for details see Russell, 1905). Again, a theory of logical reconstruction might yield a still different notion of subject, or eliminate the notion altogether. One might try to cut the Gordian knot by saying that the subject of a proposition in our sense will be whatever the proposition is about. Unfortunately, about-ness is not a sufficiently clear semantic notion to bear such a conceptual burden. Philosophers do not always agree what a sentence, or the proposition it expresses, is about. Consider, for example, ‘John smokes Marlboros’. This could be construed as being about John, or about Marlboros, or about smoking. More sophisticated explorations of this concept have not yielded satisfactory results either (Goodman, 1972, chapter VI). Thus the proposal needs to be supplemented by the lame disclaimer ‘pending a satisfactory analysis of the notion of subject’.

Further difficulties emerge in connection with negative existential propositions. Our proposal requires a reinterpretation of such propositions in such a way that the non-existent entity does not emerge as the subject. Such interpretations are available but are not beyond controversy. Finally, the proposal is not meant to exclude formal ontologies within which domains of nonexistent objects are admitted in order to have the tools for formulating adequate analyses of propositions of all kinds with abstract and non-abstract objects, qualitative, existential, and non-existential content, etc., for such theories too admit the difference between what exists in reality and what does not (e.g., Parsons, 1980; Zalta, 1983).

Let us see now how one would compare different ontological theories in terms of our criterion. Realism claims that abstract recurrent entities, sometimes called universals, exist, while nominalism claims that no recurrent entities exist. According to our criterion, for nominalism to be successful it will have to reconstruct all of the truths that seem to be about recurrent entities to be actually about combinations of non-recurrent entities. Likewise, a materialist who denies that there are purely mental entities will have to show how all truths that seem to be about mental entities can be shown to be really about material entities. Even if at any given time we do not know how to carry out these programmes of reinterpretation, one could argue against an ontological thesis by showing that the favoured category or categories suffer from obscurity or even incoherence. Of course, obscurity has to be understood here in a logical sense. Common-sense intuitions, especially of the crude variety, are no guides to what is or is not a legitimate part of reality. (I am indebted to Neil Delaney for the felicitous phrase: ‘the podaic projectibility criterion’, which says: ‘if you can kick it, it exists, if you can’t, it ain’t.’) As we shall see, there are general criteria of adequacy for ontological categories that transcend historical and cultural limitations, unlike notions like what is ‘odd’ or a ‘strange entity’.

Ontological proposals and their assessments must be kept separate from epistemological issues. For example, we proposed to examine claims about what are subjects of true propositions, but we did not specify that the propositions in question must be known or even knowable. We might have reasons to believe in the reality of a certain entity even if it can be shown that the nature of this entity cannot be known by us. For example, we might be able to prove that it is reasonable to admit the existence of certain very large numbers, or of God, or of faraway galaxies. In particular, epistemological claims such as those of empiricism, according to which all of our informative non-tautologous knowledge of reality is based on information received through the senses, must be laid aside when assessing ontological claims. Empiricism proposes some constraints on the realm of human knowledge. One cannot deduce from that alone limitations on the realm of what exists. To be sure, in a complete philosophy one would like to see harmony between one’s ontology and epistemology. But to rule out from the start any ontology that seems to cause trouble for one’s epistemology is like fitting the glass slipper to the lady’s foot by chopping off her toes.

In this chapter we shall review ontological debates about universals and non-recurrent entities called particulars. One might hold that only universals exist, or that only particulars exist, or that both types exist. Within the latter view there are those who do and those who do not want to argue also for the ontological priority of one of these categories. We shall also review ontological debates about the reality of modalities such as necessity and possibility. The actualist denies the reality of these modalities, though he could admit them as mind-dependent conceptual entities playing key roles in the way – human conditions being what they are – humans organize experience. We shall then turn to the examination of two kinds of particular: material objects and events. We shall consider views according to which events do not exist, or are at least ontologically posterior to material objects. There are many other ontological debates. Some are about facts, others about time, space, and still other aspects of reality. We shall concentrate on the issues enumerated because these provide the right background for the questions about thought and language that we shall raise. Can we have an adequate analysis of thought without events? Can we have an adequate analysis of language and its syntactic and phonological constituents without positing the ontological reality of universals? Can we account for verb- and aspect-semantics without assuming that in many sentences of subject–predicate form verbs introduce events? The debate between materialists and dualists is also relevant to our concerns, but we shall leave the discussion of that matter to the second part.

Before launching into an outline of the debate between realism and nominalism, another problem involving another sense of ‘realism’ needs to be mentioned, only to be laid aside. This problem, discussed primarily by Kant, concerns the alleged gap between reality as such and reality as known through human experience. We are asked to consider the possibility that all of the categories mentioned – particular, universal, space, time, etc. – are constituents of the phenomenal world of human experience only, and that we might not have any a priori reasons to ascribe these also to a mind-independent reality. The ‘realists’ in this debate think that the basic categories of experience correspond to those of the real world, while the phenomenalists are sceptical about this. Since all of the issues that we shall consider would emerge regardless of which side one chooses, we shall not devote much space to this important topic. But at least the bare outlines of a moderate realism are worth mentioning. According to this view, we might not have a priori reasons to show that realism in this context is right, but that there may be good empirical reasons to suppose that realism is likely to be true. One could cite the success of the human mind in coming to understand reality, and suggest that one of the contributing causes might be our ability to approximate with our categories those of reality. There are several other arguments to the effect that even if one cannot prove the realist to be correct, one need not embrace phenomenalism. Realism is not an all-or-nothing proposition. A moderate version of it can turn out to be more reasonable than the available alternatives.

Realism and Nominalism

The paradigmatic examples of particulars traditionally have been entities with both spatial and temporal or at least only temporal location. Thus material objects in the broad sense, such as living things, mountains, buildings, stars, and islands are particulars, and so are events such as the Battle of Waterloo, the birth of Jesus, and Vladimir Horowitz’s last concert. Mental events such as Cleopatra’s thinking of Anthony at a particular time are also particulars, and according to some philosophers these have only temporal locations. The paradigmatic cases of universals are attributes such as being a human, being a triangle, or being made of wood. Many but not all philosophers regard numbers also as universals. This dichotomy is not claimed to be exhaustive of reality. Sets, types such as the 1982 Volkswagen, or letter-types, and various other candidates for reality might not fit into either slot (see Cartwright, R., 1962).

Philosophers do not rest content with these intuitive and commonsensical examples. They point out that there are particulars not contemplated by common sense. For example, the object whose parts are all of the red surfaces in the world is a particular, and so are things made up of all the greens, blues, browns, etc. It is fairly easy to extend our intuitions to these large-scale particulars that are designated by what the linguists call ‘mass terms’, i.e. terms that do not pluralize but behave otherwise as general terms (for discussion see Goodman, 1951; Moravcsik, 1973.) There are also universals outside of common sense. Many of these are not introduced by a single word of some Indo-European language. An example would be the universal of being either a mathematician or six feet tall, or both a barber and a bridge-player, etc.

These are huge collections of items, and not all of these correspond to common-sense intuitions. Hence the urgency for definitions. This task, however, turns out not to be easy. Traditionally, the key difference between universals and particulars has been claimed to be the fact that universals are recurrent while particulars are not (e.g., Stout, 1923; Moore, 1923). The attribute of being intelligent, for example, recurs repeatedly in instances such as Newton, Semmelweis, and Einstein, but a particular made up of all of the water on earth can have only parts not instances and does not recur. It is interesting to note that nothing in this characterization says that particulars must be in space and time. Thus additional burden is placed on the notions of recurrence and being a part.

Recent attempts at clarifying the notions of a universal and a particular involved working out the ontological theories in formal presentations. Nominalists attempt to clarify the notion of a particular, and then try to show that all truths can be presented without assuming the existence of universals (Lesniewsky, 1931; Goodman and Quine, 1947; Goodman, 1951). These enterprises need to meet the following conditions for metaphysics. First, the formal accounts must link up to the pre-theoretic intuitions of what particulars are. Second, the systems have to meet formal criteria of consistency and clarity. Thirdly, the undertakings have to deal with the key examples of human knowledge, such as mathematics. Among the various formal presentations of both realism and nominalism, we shall select Nelson Goodman’s Structure of Appearance because of its avowed attempt to meet these criteria, and because nominalism has a greater need to be seen in this way, while there have been many good, less formal, defences of realism.

In...