The following discussion of Peirce's pragmaticism and semeiotic will focus on five main topics. First, it will be helpful to sketch the way in which pragmaticism implies semeiotic. Second, the anticipation of semeiotic in the early series of articles, “Certain Faculties Claimed for Man” and “Some Consequences of Four Incapacities,” will be considered in order to indicate the way in which Peirce's development of his anti-Cartesianism and his departure from Kant served as a basis for his semeiotic. Third, we shall see how the keys to pragmaticism point us toward Peirce's conception of the three conditions of sign action, or what Peirce refers to by the term semeiosis. These conditions will then be discussed. Fourth, it will be found that one of the conditions of semeiosis, the object, is presemeiotic, or preinterpreted, and at the same time semeiotic. A distinction Peirce makes between two kinds of object, the immediate and the dynamical, will then be considered briefly. Finally, Peirce's classification of signs will be outlined.

Thoughts Are Signs

The implications of pragmaticism for semeiotic should be clear. Pragmaticism is a theory of meaning. The meanings of general concepts or terms, it will be recalled, are dispositions, habits, or laws that can be formulated in linguistic expressions. Insofar as these meanings are indeterminate with respect not only to vagueness but also to generality, they remain to be interpreted with respect to the particular consequences that follow from acting in accord with them. They must be interpreted with reference to the patterns of consequences that they represent.

The conception of Peirce's envisaged architectonic can be projected beyond the context of the late nineteenth and early twentieth centuries. The projection with which I am concerned springs from two architectonic conceptions dealt with in the earlier chapters: semeiotic and evolutionary realism. These conceptions may be applied to current interconnected controversies over the possibility of deciding between metaphysical antirealism and realism and between relativism and antirelativism in interpretation theory, which for Peirce falls under the domain of semeiotic.

Our main responsibility will be to see how Peirce's philosophical outlook avoids some of the consequences of recent attacks on the possibility of affirming realism. These considerations, in turn, have a bearing on the question of whether interpretation (in art criticism as well as in theoretical investigation) has a grounding outside texts and theories.

In the following remarks I shall respond from a Peircean perspective to the challenge that metaphysical assumptions about reality or extralinguistic objects and systematic speculation about all actual and possible experience are nonsensical, futile, or at least outdated. Philosophies that have dominated Western thought until the early twentieth century – and, indeed, have persisted in some circles in this century – presuppose a metaphysical or epistemological realism or an objective idealism (whether they are set forth in terms of these labels or appear as integral to rationalisms, empiricisms, and philosophies that suppose that there are noumenal or extraexperiential conditions of thought). These views share the assumption that there is something sufficiently objective to serve as a foundation that warrants rational argument and possible conclusions about which single perspective on experience or the world is more intelligible than another.

UNIVERSALS AND POSSIBILITY

It is central to immanent realism to conceive of a universal as an ontologically significant feature that can occur in many places at the same time. If a universal can occur many times or just once, it is equally possible for it not to occur at all. Nevertheless a sentence containing a predicate says something irrespective of whether the universal corresponding to the predicate occurs or not. For example, a sentence that denies that a certain object is regular chiliagon-shaped says something, even though that shape probably never has occurred and probably never will occur. Such sentences can be true, and we can understand what is being said, even though no object instances the universal concerned.

It appears, then, that talking about and referring to a universal is just as possible whether it occurs many times or not at all. It seems that, in one sense, whether “there is a universal” is independent of whether it occurs, and it also seems that the meaning of the corresponding predicate is also independent of whether the universal occurs.

In his book on universals, Armstrong took the position that it is important to separate the theory of meaning and the theory of universals, but later, in What Is a Law of Nature?, he took a slightly different view: “It may well be that it is impossible to explain the use of general words without postulating universals. ” This seems to be right.

The aim of this book is to give an account of which things are the “real constituents of the world”. The account is based, to begin with, on a characterization of the notion of universal. It also attempts to decide which particulars are real constituents of the world, and in doing so argues against events, and things like events, in a number of different ways. The focus is on causality, particularly the notion of causal relation, as a guide to what is real.

The central theme is that the natural world is a world of particulars and universals as understood by immanent realism. In order to make that more precise, I argue that it is special sorts of universals – namely, basic universals – and special sorts of particulars – namely, unified particulars – which are the real constituents of the world. It is not part of my intention to show that immanent realism is the correct theory of universals, since there are other works that do that, notably Armstrong's. In the course of the discussion, however, arguments will be given that will show the superiority of that theory. The work should be regarded as being in the area of metaphysics or general ontology; epistemology and semantics will, on the whole, be avoided. My interest is in how things are, not in how we come to know how they are. I hope the conclusions can serve as metaphysical foundations for scientific realism, while avoiding attachment to any particular scientific theory.

A BASIS FOR PREDICATION

Although the notion of particular is in need of some refinement, it can perhaps be accepted that the basic idea of a particular is familiar and not controversial. Certainly anyone today who wished to deny that there are particulars would have a lot of explaining to do. It is different with universals. The immanent realists' notion of universal is a metaphysician's idea, which not everyone regards as intuitively acceptable. But if, with Russell and Moore, we regard universals as “real constituents of the world”, along with particulars, we shall have to give an account of the notion of universal. It will have to be an account that brings out the way in which a universal is a real thing, something ontologically significant.

Frege notes that fundamental notions such as “concept”, and we may add “universal”, cannot have proper definitions. With such fundamental notions, “there is nothing for it but to lead the reader or hearer, by means of hints, to understand the words as intended”. What can be said about fundamental notions Frege calls an ‘explanation’; alternatively, it could be called a ‘characterization’. Our first aim, then, is to discuss a number of ways of characterizing the notion of universal that bring out the way in which a universal is a real constituent of the world.

THE DISTINCTION

There are two sorts of particular, those that have a natural principle of unity and those that do not. The principle of unity is whatever it is that gives us a reason for thinking we have a single particular. It is a natural principle of unity if the principle of unity is there in the world independently of human thought, human conceptualization, or human decisions. It is, of course, controversial whether there are in fact natural principles of unity. With respect to this issue Michael Ayers uses the word ‘realist’ for those who support his position, as I do, that there are particulars with natural principles of unity, and the word ‘conceptualist’ for those who oppose it: “Adopting, then, a ‘realist’ tone of voice, let me say first that physical objects are natural unities or natural structures which come into existence, continue to exist and cease to exist quite independently of any conceptualizing on our part.” It is difficult to know what picture of reality conceptualists are presenting us with; they seem to see reality itself as something obscure or unknowable: “The picture of reality as an amorphous lump, not yet articulated into discrete objects, thus proves to be the correct one, so long as we make the right use of it.” It is also very difficult to argue against this position in general terms.

BASIC UNIVERSALS

I have distinguished statements that say “there is” a universal from statements that say a universal occurs. If “there is” a universal, then it is possible that it occur any number of times, or not at all. The central characterization of the notion of universal is in terms of this possibility of multiple occurrence. This notion of occurrence is a primitive, intuitive notion based on a commonsense view of the world, but it is by no means alien to a scientific view of the world. The shape or mass to be found here, instanced by one object, is also to be found there, instanced by a second object. The shape or mass occurs in the sense that it can be “found”.

In order to elucidate the notion of occurrence, or being found, it is natural to turn to causality. If a certain shape occurs, or is found, then we can interact with it in some sense, and it should be regarded as causally significant. If the shape of a body is changed from spherical to ellipsoidal, for example, then the particular concerned will interact causally in a different way. It is likely therefore that causality will be some sort of guide to the ontological status of universals.

There are some predicates that appear to signify universals that are significant for causality, but that are dependent on other universals.

THE LOGICAL FORM OF SENTENCES ABOUT CAUSALITY

One reason the problem of causality is difficult is that language and ordinary thought appear to presuppose what we can call the linguistically natural picture of causality. It appears to suppose that there are objects such as a, b, c, d,… that are causes and effects, and that there are relations between them that are causal relations. If there is a causal relation between a and b, then that causal relation is a matter of a's producing b, or a's being responsible for b, though these phrases merely further express our intuition. It is further supposed that causes and effects occur in causal chains: a causes b, b causes c, and so on. Russell ascribes this view to Mill.

Events are appealing as causes and effects, since they are at least particulars and are easily arranged into chains. They seem to fit the linguistically natural picture of causality. The main problem with events is that of finding a causal relation to link one event with another. This is a serious problem for events, as without causal relations to link them together events cannot be taken seriously as causes and effects. Physical objects, as we have shown, can enter into causal relations; but the problem with physical objects is that it is difficult to see them as causes and effects in the hoped for chain of causes and effects.

REAL RELATIONS

We are happy to say that certain particulars exist and are real, and this lends credibility to the idea that basic attributes are real constituents of the world. The reason is that basic attributes inhere in particulars and are thought of as in some way constituents of particulars, contributing to their reality. Relations, on the other hand, seem to have a shadowy sort of existence; they seem to hang between particulars. Intuitively, it is difficult to regard relations as real constituents of the world.

According to Gottfried Martin this intuition found expression in mediaeval metaphysics. For Aquinas the being of an accident (i.e. an attribute) was a matter of its existing in an individual being “Esse accidentis est inesse” – for an accident, to be is to inhere: “A relation, considered as a real accident, would be an accident which stood so to speak with one foot in one substance and the other in another, and this contradicts the idea of an accident as inherent in a particular individual being.” This point of view leads to the doctrine that all relations are “relations of reason”, in other words that they are things produced by thought and that there are no such things as “real relations”.

THE SYNTACTIC PRIORITY THESIS

Some philosophers have held that language, which is taken to include logic, is the only guide we have to the nature of the world. Russell divided philosophers into three classes on the issue of the relation of language to the world. The second and third classes are not very promising, but the first consists of “those who infer properties of the world from properties of language”, “a very distinguished party” Russell called them. He then says: “If, therefore, we are confined to the above three alternatives, we must make the best of the first. ”

Presumably it is for such reasons that Russell concluded that facts are objects we come across in the world, and that Wittgenstein concluded the world is composed primarily of facts. Facts are things in the world that correspond to sentences in language. The reason for thinking that facts are ontologically significant is, presumably, that sentences are linguistically significant.

Later the principle that Russell understood as characterizing the first class of philosophers became known as the syntactic priority thesis. It arises in the discussion of Frege's philosophy, and it is regarded as how Frege characterized, and presumably justified, the concept–object distinction, and also, at a lower level, how he determined which classes of things were classes of objects. It can therefore have two roles.

SETS AS ARBITRARY PARTICULARS

Although the criterion of identity for a set may be clear, it is not clear what a set is. It is not clear in what sense a set is one thing, or what it is that binds its elements together to make it one thing. A number of attempts have been made to answer this question, but unfortunately none of them appears to be very satisfactory. By reviewing some of these answers I shall try to show that sets of particulars deserve to be called arbitrary particulars. The basis of my discussion will be Max Black's discussion of the nature of sets as collections, as abstract objects, and as examples of plural reference.

The elements of a set have to be sharply demarcated things. It has to be made clear that it is the copse that is to be the element of the set, and not the five trees as trees that make up the copse; and it has to be made clear that it is the material object as such and not its constituent parcel of matter that is to be the element of the set. Each element must therefore have associated with it an unambiguous criterion of identity. And since each type of particular has its own criterion of identity, if follows that an element of a set enters into membership only as a particular of a certain type.

DETERMINATES AND DETERMINABLES

Determinate attributes of particulars can be gathered into types; there are mass determinates, temperature determinates, length determinates, and so on. Not only do all mass determinates have something in common, but all particulars that have a mass determinate have something in common – they are all massive particulars. It follows that at least one type of universal other than determinate attributes is required.

There have been a number of accounts of what the extra common factor is, the two most obvious being what I call the second–order theory and the second–level theory. According to the second–order theory, all determinates of a certain type have a second–order attribute in common, a property of a property; what the particulars that have determinates of that type have in common has to be explained in terms of the second–order attribute. According to the second–level theory, all particulars that have determinates of a certain type have another attribute in common, a second–level attribute; what the determinates of a given type have in common is explained, in my version of this theory, by a relation of “subordination” between the determinate and a second–level attribute. The term ‘determinable’ is often reserved for what I have called the second–level attribute, which is a determinable in the classical sense, though the problem, irrespective of proposed solution, is usually called the problem of determinables. Although both determinate and determinable are first–order attributes, the relation between them of “subordination” is a second–order relation.

The logical operators have been defined over all implication structures. It is possible to develop a concept that is the generalization of the notion of a truth-value assignment that is also applicable in all structures, despite the fact that most of them do not have truth-bearers as members. It is a notion that reduces to the familiar truth-value assignments on those structures that are associated with the usual logical systems, and it yields a general framework within which one can study the extensionality (or nonextensionality) of the logical operators on arbitrary structures.

Extensionality and bisection implications

Suppose that S is a nonempty set, and T = 〈K, L〉 is a bisection on it. Form the bisection implication relation “⇒T” that is associated with T, and let IT = 〈S, ⇒T〉. Since “⇒T” is an implication relation, we can study the behavior of the logical operators with respect to it.

Consider the structure IT. Since CT(A, B) ⇒TA and CT(A, B) ⇒TB, and is the weakest to do so, it is possible to compute the distribution patterns for CT(A, B) in the sets K and L, given the distributions of A and B in K and L. Conjunctions. Suppose that A is in K. Then, since CT(A, B) ⇒TA, CT(A, B) cannot be in L.