Essays On Pragmatic Naturalism: Discourse Relativity, Religion, Art, and Education

2015

Essay I

Discourse Relativity and Specialization

1

Pragmatic Specialization and Rationality

At primitive levels of sign interpretation, pragmatic functions are combined. For the grazing gazelle, an odor can be associated with the sight of an approaching lion, arouse fear, and lead to a fleeing response. For the hunting predator, tracks or odors may be associated with the sight of the prey, arouse the pleasurable anticipation analogous to what for us is hope, and direct further tracking. Assuming these responses by prey and predator are instrumentally learned and not innate and reflex, a single event such as an odor is a sign that is the object of combined cognitive (the association of one event with another), emotive (the arousal of the emotions of fear or hope), and dynamic (fleeing or approaching) interpretations. Similar combining typically occurs for signaling within animal groups. The cry of a baboon sentry can warn of an approaching predator, express and arouse fear in members of the tribe, and signify the response of fleeing. Both its use by the sentry and interpretation by others combine cognitive, emotive, and dynamic pragmatic functions. At the signaling level there is, however, a form of specialization that is introduced. This is in the form of calls whose principal function is that of establishing contact between members of a group or between members of mating pairs. The function of specialized grunts recorded by Cheney and Seyfarth in their studies of vervet monkeys is that of enabling recognition that a grunt’s source is a friendly group member and not an antagonistic outsider.¹ Signature bird calls and signals transmitted by dolphins and whales seem to perform a similar transactional function of enabling identifications for those outside visual contact.

Lost forever are the means of communication used by early hominids intermediate between signals used by infra-human species and what can be formed from the resources of natural languages accessible to linguists. At some stage signals used to signify kinds of objects were combined with those signifying qualities and actions to form proto-sentences with a distinctive referring function for subjects and ascribing function for predicates. This was accompanied by the development of the syntactic categories of nouns, verbs, and adjectives, and of rules for combining them into sentences. Syntactic rules for word ordering and verb form then enabled distinctions of sentence mood and specialization of cognitive, emotive, and dynamic pragmatic functions in a way impossible for isolated signals. The single word run can in isolation have combined descriptive, expressive, and prescriptive uses, but word ordering in English (verb forms in many other languages) removes ambiguity of use for The man runs, Would that I could run, and Run to the store. With sentence formation also come introduction of proper names and personal pronouns as communicative transactional elements, making it possible to indicate sources of information, expressions of emotions, and commands, and to address these to specific individuals, as when we use the name Bill to indicate the addressee of Bill, the house is on fire. Ethnologists cite evidence that chimpanzees show some capacity for combining counters with some of the referring and ascribing functions of subjects and predicates. This reinforces the view that sentence formation, along with an understanding of the syntactic rules required for it, are not unique to the modern human species.

Certainly unique, however, is the human capacity to combine sentences into discourses in the form of conversations, detailed instructions, and storytelling. The central features of discourse formation and the problem of assigning topics to specific discourses will be discussed in the next chapter. But at this stage we note only the importance of discourses in which a conclusion is combined with premisses to form an inference. Those who can use and interpret these inferences are said to have the capacity to reason, that is, are endowed with that rationality distinguishing our species from all others. Inferences bring with them still further developed means of pragmatic specialization through variety of types of relations between their premisses and conclusions, a form of specialization that eventually leads to the development of social institutions that prepare its members to use inferences employed within them. We shall be outlining below, first, the principle forms of inference occurring at the discourse level, and, secondly, the social role performed by philosophy in describing these forms and acting as a counter-balance to harmful effects of specialization.

These effects have been obscured by the faculty psychology prevalent during the Enlightenment, which through the influence of Immanuel Kant has distorted views within all branches of philosophy. Kant distinguished in his Critique of Pure Reason the mental faculties of understanding and reason by describing them as capacities for judging the truth of propositions expressed by two different types of sentences. Analytic sentences Kant defines by the relation of predicates to subjects. They are those whose predicates are contained in their subjects, that is, having predicates appearing after analysis of noun subjects into their component meanings. The sentence All bachelors are unmarried would be classified as analytic, assuming its subject bachelor can be analyzed into the complex term unmarried man, a term in which the predicate unmarried occurs. Replacing the original sentence with its analyzed form, the sentence becomes the trivially true and uninformative All unmarried men are unmarried. Synthetic sentences, in contrast, are those sentences not analytic, that is, those sentences in which analysis of their subjects will not reveal their predicates. Sentences express propositions that are judged to be true or false in two different ways for Kant. Those evaluated relative to what we experience through our senses are a posteriori sentences; those not capable of being evaluated in this way are classified as a priori. Kant then uses this classification to distinguish the faculties of understanding and reason. The understanding is the faculty directed towards propositions expressed by synthetic a posteriori or empirical sentences, sentences not true because of the meanings of their subjects and whose truth or falsity are established on the basis of sensory observation. They include both the descriptions of everyday life such as The dogwood tree by my house is in full bloom and the generalized descriptions of the natural sciences. Reason, in contrast, is directed towards a priori propositions, and these may be expressed by either analytic or synthetic sentences. Those like All bachelors are unmarried that are analytic a priori provide no new information, but only some reiteration of meanings of subjects.

The focus of Kant’s attention in his Critique was on the more interesting synthetic a priori sentences that purport to convey information, and this attention has had a pernicious influence on all later philosophic attempts to describe our reasoning capacity. This is because the concept of the synthetic a priori is totally indeterminate in application, a concept defined in terms of what it is not. A synthetic sentence is defined by Kant as one that is not analytic, one for which its predicate does not occur after analysis in its subject. Similarly, the a priori classification is reserved for those propositions that are not empirical, those not evaluated relative to perceptions. We are thus not given by Kant any positive characterization of the synthetic a priori, but only two contrasts. Since the faculty of reason as capable of providing new information is directed towards this type of sentence, it too was left indeterminate. It can take the form of speculative reason directed toward the propositions of dogmatic metaphysics, and in this form Kant denies such a capacity. But it can also take the form of the capacity to apprehend propositions as diverse as the theorems of number theory and the principles of morality, both regarded by Kant as being expressed by the synthetic a priori and within human capacities to apprehend. As we shall see in Chapter 5, when we turn to normative forms of discourse, this indeterminateness opened the way for the artificial contrast between reason and emotion assumed by Enlightenment rationalism.

Such indeterminateness is avoided by recognizing that to reason is to use an inference, and that there are as many forms of inference as forms of discourse. We use inferences when we explain a past event and predict a future one, when we trace what is implied by a story or interpret a religious narrative, when deciding what we should do, and when making judgments on the justice of political and legal systems. Such inferences are formulated in different forms of discourse and governed by different rules, all capable of being specified relative to the purposes of a distinctive kind of reasoning. These forms of inference define both theoretical reasoning conducted within institutionalized specializations such as mathematics and the natural and social sciences and practical reasoning employed within legislative branches enacting legislation and a legal system responsible for interpreting and applying it. Still different forms are employed within institutions entrusted with selecting from, transmitting and interpreting, and evaluating our religious, literary, and artistic heritage. The capacity for rationality that seems to be unique to our species is thus not that of apprehending and evaluating the synthetic a priori as a single form of indeterminate proposition. It is instead a multi-faceted capacity to use and interpret the great variety of inferences used within the cultures of developed societies, and takes as many forms as there are types of inference. This capacity is distributed unequally through a society in ways that depend on the aptitudes and educational backgrounds of its members and the specialized institutions to which they belong.

Of special interest to us here are contrasts and relationships between three broad categories of forms of discourse that we label as descriptive, transactional, and normative. For descriptive forms with sentences uniformly in the indicative mood, we can readily assign discourse topics, and differentiation of topics can be a source of specialization. Chemists describe the properties of molecules, while physicists address the topics of atoms and sub-atomic particles. Chemistry is further divided into branches describing organic and inorganic molecules, and groups of chemists further sub-divide into those studying an increasing variety of molecules in both branches. Members of these groups become specialists whose assertions are accepted by outsiders lacking the expertise to independently test them, based instead on trust that methods employed by trained specialists will provide reliable information about their discourse topics. For transactional and normative forms of discourse, however, the situation is different. Mastery of the transactional is shared by all members of a natural language community capable of using personal pronouns, though some will have greater facility than others in the nuances of address and deference expected in social relations. And as we shall see more clearly in Chapter 2 and 3, the unity of normative discourse used to guide conduct is not provided by common topics of a kind that admits of specialization. Unity here is derived instead from the purposes of the discourse and relations between premisses and conclusions of inferences. Within the legal profession, there is specialization in such areas as corporate, federal, state, and international law, where training is required for understanding and applying enacted laws and past judicial decisions. But all members of a community share an interest in determining whether a given law promotes a policy that they approve of, and whether the legal system that applies and enforces laws is just. For answers to questions about the goodness of policies or justice of laws and political institutions raised within normative forms of discourse there are no specialists.

The contrast between forms of discourse used in theoretical and practical reasoning is accompanied by a contrast in forms of relativity. Chapter 3 describes how the meanings of key terms such as true, exists, real, and identity are relative to general topics of descriptive discourses. Failure to recognize this relativity is responsible for attempts throughout the history of philosophy to impose features of one form of discourse on others used for very different purposes. I shall argue in Chapter 4 that the traditional opposition of materialism and idealism is the result of such impositions. This form of theoretical relativity is, however, very different from the practical form. Practical inferences include premisses expressing wants, aversions, and preferences, and these wants and preferences can vary with the history and circumstances of different societies and at different stages in their development. Such variation affects how a society sets its moral standards and how its members weight the sometimes competing demands of security, freedom, and equality in their assessments of social justice.

So much for this introductory summary of some of the central issues to be discussed in this essay. We begin with an outline of some basic features of discourses in general and then proceed to an examination of some of the ways they become specialized and relativized.

2

Varieties of Discourse

The most basic form of discourse is conversational speech, which typically combines sentences used to establish contact between speaker and interlocutors, provide information, issue commands or requests, and express a variety of attitudes. Conversations may also include arguments in which conclusions are inferred from premisses. This occurs when we reason to what we should do, explain why an action was performed, predict a future action, or determine the consequences of obeying a command. While discourse at this conversational level is typically a hybrid mixture of sentences with different uses, it can be specialized in order to accomplish purposes such as describing, prescribing, or expressing feelings and emotions. Philosophy has historically concentrated on indicative sentences used to describe, and there has been a tendency either to ignore other forms of sentences or to assimilate these forms to descriptions. This tendency has been reinforced by the success of applications of modern predicate logic to deductive inferences, applications that are the topic of this chapter’s third section. Finally, we will turn to the problem of distinguishing the roles of linguistics and philosophy in describing features of discourses.

2.1. Conversational Speech

The familiar has inherent complexities, and for this reason is usually the most difficult to understand. This may explain why the philosophy of language has tended to avoid discussing forms of discourse formulated within natural languages in everyday use. The following phone conversation between Bill and Betty illustrates this complexity.

Bill: Hello, Betty, this is Bill. How are you?

Betty: Hi, Bill. I’m fine. Where are you?

Bill: It’s raining here, and so I’m at home. I think there’s also a chance for rain tomorrow, but I hope for good weather. Then I should go shopping, as I want some printer paper. I’ll get some for you if you need it.

Betty: Please do. I’ll reimburse you.

Prominent at the beginning are transactional elements, including greetings, addresses that identify speaker and hearer, and the personal pronouns you and I as proxies for the addressing names Betty and Bill. Then follow a description (It’s raining here), expressions of a belief (that it might rain tomorrow), hope (for good weather), and desire (for printer paper), a prudential normative (I should go shopping), a conditional resolution (I’ll get some [paper] if you need it), and finally a request (Please do [get some paper]) and a promise to reimburse. Included also are enthymemes as inferences with unexpressed premisses. The first provides Bill’s explanation of why he is at home (because it is raining, with the implication that he doesn’t want to get wet). The second is a practical prudential inference giving the reason for why he should go shopping (his want for paper), with the implicit means-end assumption that the best way to satisfy this want is to go shopping. Such a discourse is typical of those used in our daily lives, and is mirrored in our streams of consciousness—what Plato called the silent discourse of the soul with itself—as we plan and anticipate encounters with others.

Hybrid conversations between speaker and audience feature the personal pronouns I and you used in conjunction with mental terms such as believe, think, hope, and want. Contrasted to such hybrids are pure descriptive discourses sanitized of such elements, discourses combining indicative sentences, and illustrated by the sequence Bill walked down the street; then he walked into the store, and bought a tie and the disjunctive syllogism Bill must be at home, since he is either at home or at the store, and he isn’t at the store. It is usually possible to assign a common topic to such discourses, and in these examples it is obviously the individual Bill. This specialization of descriptive sentences makes them suitable as the paradigm used in discourse analyses of linguistics. As occurring in discourses used within the sciences, combinations of descriptions are commonly the model for analyses of language found in philosophical accounts of scientific method. In science, mental terminology common in face-to-face speech is either totally absent, as in physics and biology, or tends to become replaced by descriptions to which the evaluative standards of science can be applied, as occurs in psychology and sociology. In contrast, mental terminology is preserved, along with personalized addressing forms, in religious narratives describing encounters between the human and the supernatural. This borrowing of familiar features of transactional conversations enables these narratives to be readily understood by all members of a linguistic community. Relationships between the personalized and mental terminology of everyday conversation and religious narratives, on the one hand, and specialized descriptive discourses sanitized of these elements, on the other, are the focus of the issues of metaphysics to be discussed below in Chapters 4 and 5. Of special interest to us later will be the persistent tendency in philosophy to impose features of descriptive discourse used to provide information on forms of discourse used for different purposes.

Besides practical inferences with a want premiss and should or ought conclusion (as in Bill’s conclusion that he should go shopping), within conversational speech we find imperative inferences used to determine the consequences of a command relative to circumstances. These include those from a general imperative to a particular application, as in Pick up all the boxes. This is a box. So pick it up, a type of inference used in applying laws in specific situations. Storytelling expressing hopes and fears also occurs at this conversational level, and can be used both to entertain and to convey moral instructions. All speakers of a natural language, including members of small, isolated tribes, are able to understand and use such specialized fragments. This is only a preliminary, however, to the specialization accompanying population increases in restricted regions and the advent of writing. With writing, there is communication over extended distances that is much more reliable than relayed speech, an extension that is the mark of what we label a civilization. At this stage, institutions begin to develop using increasingly specialized forms of discourse, and schools develop for training prospective members of these institutions in their use. Thus develop academies for training in the sciences, mathematics, law, engineering, religion, and interpretive studies of literature and the arts. Forms of discourse used within them develop standards of evaluation appropriate to the purposes they fulfill, along with technical vocabularies that diverge from the shared vocabulary of the natural language in common use. Then understanding of relationships between these forms becomes increasingly difficult, and philosophy emerges as the discipline attempting to provide this understanding. Often the form philosophy takes within a civilization either reflects the prestige of an institution within it or of conflicts over the priority to be assigned one institution relative to another. The pervasive influence of religious institutions, especially in the earlier stages of a civilization’s development, usually guarantees that their forms of discourse enjoy a prominent position in discussing problems arising from these conflicts.

There is a reason for the puzzling nature of these philosophical problems and their persistence in different guises over many centuries. Nearly all of us are very adept at both using the shared natural language of our speech community, with its transactional addressing expressions and mental terminology used in aiding and sympathizing with others, and evaluating and reacting to others’ beliefs and wants. We are also adept at using the complex combinations of sentences typical of conversational speech, as illustrated by the conversation between Bill and Betty. But we differ greatly in the interests and aptitudes necessary for acquiring competence in using the specialized forms of discourse of diverse social institutions. The talented novelist is often a dunce in mathematics; the respected judge or inspiring religious leader may show no interest in or aptitude for experimental science; those successful in the sciences often lack either interest in interpretive studies of art and literature or an interest in expending the time and energy required to acquire the background necessary for understanding them. These differences in interests and talents and restrictions on time and energy make it increasingly difficult to understand relationships between specialized forms of discourse. This translates into a tendency to impose features characteristic of one form of discourse on others used for different purposes. The intractable disputes of metaphysics reflect the associated tendency of writers to assign priority to forms of discourse for which they happen to have an aptitude and interest.

Philosophy has claimed for itself the role of relating the evaluative standards employed within social institutions. As a second-order inquiry into the central activities of a civilization, it encompasses studies of the empirical standards of the sciences. Its inquiries are also directed towards the evaluative standards used in assessing the moral actions of individuals, the policies of a government, and the worth of literary and artistic contributions. Philosophy has thus included within its scope epistemology, with its discussions of justified belief, and the normative disciplines of ethics, political philosophy, and aesthetics. In its early history, individual philosophers such as Plato, Aristotle, Locke, and Kant could write in all these areas. The growth of specialization in the nineteenth century has made this increasingly difficult, and it has been common for philosophers to restrict themselves to a single area. But even with this relatively recent restriction, most philosophers have continued to write for the general educated public, and have employed, with some exceptions for a technical vocabulary unique to their discipline, the vocabulary of the natural language shared by members of their intended audience, including personal pronouns and the mental terminology characteristic of conversational speech. In discussions of justified belief, philosophers may include examples from highly technical areas of theoretical physics, but the basic framework of epistemology is designed to include both the familiar beliefs of everyday life and those of scientists using their special methods of justification. Moral philosophers may discuss complex issues related to the use of stem cell research or rationing of medical care, but they apply to these issues moral principles that derive from the more tractable decisions that we face daily.

There are thus good reasons for conversational speech formulated within a natural language (or ordinary language, as it is often called) being the focus of recent philosophy. As the common currency within a community, it is the ideal focus for attempts to understand relationships between forms of discourse with specialized vocabularies. Its hybrid forms include practical inferences that combine the descriptive with expressions including mental verbs such as believe and want used in conjunction with the pronouns I and we, along with the normative terms ought, should, and may. An understanding of such combinations within practical inferences is valuable in relating the descriptions of science to the expressive and normative. And finally, philosophy’s use of the vocabulary of conversational speech relates the discipline to the expressions of religious faith found in theology. Indeed, both philosophy and theology have staked out claims to provide a comprehensive theory that encompasses the major human activities and serve society as its primary means of social integration. Sometimes they have pursued these claims in collaboration, sometimes in opposition. By making conversational speech its focus, philosophy enables its audience to assess the relative merits of these claims when they overlap and perhaps create conflict.

2.2. Some Historical Notes

Inferential discourse has been the topic of philosophic discussion since the ancient Greeks. Inferences have included those in the indicative mood, including deductive and inductive inferences with descriptive premisses and conclusions, and what Aristotle termed practical syllogisms in which the conclusion expresses a resolution to act in a way that realizes a goal. Despite such attention to inferences, the focus of philosophy in the past has been on smaller constituent elements of discourses. These were initially words, especially nouns and adjectives, with much attention to what words stand for, whether ideal forms (Plato), forms of substances (Aristotle), psychological concepts (Medieval philosophers), or ideas (Locke, Berkeley, and Hume) as either images accessible to introspection or concepts. Central importance was attached to the problem of determining the relation between words and minimal elements out which the more complex is formed. Later attention was to shift to sentences as combinations of subjects and predicates, with subjects referring to objects and predicates ascribing attributes to these objects. These subject-predicate combinations then became the objects of psychological acts of judgment evaluating them as either true or false. Sentences were assigned primary importance because words have meanings, it was realized, in the context of sentences. This is obvious from the different meanings of large and small in Felix is a large mouse, Mice are small, Dumbo is a small elephant, and Elephants are large, for independent of such sentence contexts, the adjectives large and small lack meaning. With the shift of attention to sentences, primary importance became shifted to an understanding of the relation between a sentence and whatever it represents or corresponds to.

It was not until the past century that discourses with sentence constituents became a central focus, and it was realized that sentences in turn have meaning in the contexts of the discourses in which they occur. The initial impetus for this came in the breakdown of the analytic/empirical distinction for sentences within the context of scientific theories. We noted above that analytic sentences were regarded as those sentences whose truth or falsity can be determined solely on the basis of the meanings of their constituent terms, while empirical synthetic sentences (synthetic a posteriori sentences) were those evaluated relative to observational evidence. The breakdown of this distinction was originally noted by Henri Poincaré and then developed by W. V. O. Quine. Poincaré’s example was the formula of Newtonian gravitational theory The force exerted on a body is equal to its mass times its acceleration, or F = ma. Does this provide definitions of the terms force, mass, and acceleration, and is therefore analytic? It doesn’t seem so, for it is used to assert relationships in nature. But then is it empirical as a description of a relation in nature between a force exerted and a body’s mass and acceleration? In fact, it is not derived from direct observations of force, mass, and acceleration, as is indicated by the lack of a constant in the equation. The answer is that, considered in isolation, the formula is neither analytic nor empirical, and that only Newton’s theory as a whole can be so classified. The formula relating force, mass, and acceleration is combined in Newton’s theory with two other assertions: the law of inertia that the acceleration of a body is zero if no force is applied to it (a = 0 if no external F) and the inverse square law that the force exerted between two bodies is directly proportional to the product of their masses and inversely proportional to the square of the distance between them (Mm/r²). This discursive combination can be observationally tested, and is therefore empirical, or synthetic a posteriori. This breakdown of the analytic/synthetic distinction applied to sentences suggests that if philosophy is to be conceived as the discipline that analyzes meaning, its units for analysis should be discourses such as Newton’s gravitational theory, not individual sentences.

The shift of focus from sentences to discourse combinations of them was brought on by other factors. One was the use of paraphrases of modern logic, which in the process of evaluating inferences replaced single sentences with combinations of sentences related through logical symbolism. These paraphrases resulted in attention being focused on subjects of inferences as wholes rather than on subjects of sentences, and on pronouns as means of determining relations between the constituent sentences of a discourse. These features of modern logical symbolism will be discussed in the next chapter. A second factor was the analysis of combinatorial syntactic rules initiated by the linguist Noam Chomsky. For Chomsky, sentences with prepositional phrases and relative clauses are the products of embedding what are termed sentence strings. For example, the sentence The boy with the baseball runs down the street that is wide and narrow is constructed from the sentence strings the boy runs down the street, the boy has a baseball, the street is wide, and the street is narrow. In learning the rules of English, speakers learn how to construct prepositional phrases and relative clauses from the relevant sentence strings. Chomsky thus showed how complex single sentences should be regarded as discourses or discourse fragments that for stylistic reasons have not been broken up into discrete parts by punctuation.

During this transition from the word to sentence and finally to the discourse as the unit of analysis, descriptive language has usually served as the paradigm form of language, the model to which all other uses of language must conform. Prior to the last half of the twentieth century, philosophy had been indicative bound, to use R. M. Hare’s expression, and in the process had sidelined imperatives and expressions of feelings and emotions by either dismissing them as meaningless or by assimilating them to indicatives used to describe. This took the form of claiming that the content of an imperative such as Close the door is the proposition expressed by the corresponding indicative The door is closed describing the result of obeying the command. This has the effect of denying a distinct logic of imperatives, for every imperative inference would be translated into an indicative inference expressing propositions. The imperative inference from Close the door and open the window to Close the door is then thought to derive its validity from the corresponding indicative inference The door is closed and the window open; therefore, the door is closed of the form A ⋀ B ∴ A.

But this assimilation of the imperative to the indicative is easily shown to be mistaken. In standard indicative logic we can infer by contraposition from the conditional A→B to the conditional ~B→~A with negated antecedent and consequent. For example, we can infer from If it rains, the door is closed to If the door is not closed, it isn’t raining. But from the conditional imperative If it rains, close the door we cannot infer the imperative If the door is not closed, then make it not rain. In the absence of control over the weather, this is a command that is impossible to obey. There are thus special features of imperatives and the discourses in which they occur that resist their assimilation to indicative descriptions.

Much attention was focused by the ordinary language philosophers on individual sentences used for other than descriptive and prescriptive purposes. These included performatives with illocutionary force indicators—sentences such as I promise that the package will arrive, I warn you it will arrive, and I guarantee it will arrive—by which we make promises and issue warnings and guarantees. As J. L. Austin emphasized, such first person sentences are not self-descriptions. I am not describing myself as promising when I say "I promise that p" but instead making a promise, performing a speech act different from making a statement. The descriptive content of the promise is provided by the sentence radical p, while the prefix I promise that expresses the illocutionary force of the sentence conveying how it is to be understood, in this case as a promise, as contrasted, say, to a warning or guarantee. Austin extended his analysis to psychological verbs in the first person prefixed to sentence radicals, noting that sentences such as "I believe that p and I know that p" also should not be regarded as used by a speaker to describe her psychological state. The belief sentence is instead used to indicate hesitancy about the truth of p and to invite possible correction by others, in contrast to "I know that p" as a personal guarantee of p’s truth. Such an analysis can be extended to the sentences I want my daughter to have a college education and I hope she will have a college education of the forms "I want (or desire) that p and I hope that p. Like the first person believe and know" sentences, these are used by the speaker to express a desire and a hope that p come about, not to describe a mental state. As a psychological prefix, the use of hope conveys the expression of p in a way that contrasts with a use of wish. Hope implies that p has a basis in justified belief, while wish lacks such an implication. In contrast to hopes, wishes may be idle or fanciful.

Austin and those influenced by him directed their attention to individual sentences used for other than descriptive purposes. Applications to discourse combinations of sentences came when Elizabeth Anscombe, G. H. von Wright, and Max Black turned their attention to practical inferences. Consider the inference of the previous section’s conversational sample used by Bill. With premisses made explicit, this might be formulated as the inference

I want some paper

If there is good weather tomorrow, then the best way to get some paper is to go shopping.

The weather will be good tomorrow

I’m able to go shopping

It follows that I should go shopping tomorrow.

The inference is from the want expressed in the first premiss (an abbreviated form of I want that I have some paper), the assertion that going shopping (rather than, say, ordering through the internet) is the best means to realize the want if the weather is good, and the assertion that it will be good (won’t rain) to the conclusion that the means should be performed. Philosophers have in the past been puzzled by the meaning of normative should or ought sentences. Are they descriptions of special kinds of normative facts? Or are they expressions of attitudes? As occurring in the context of this inference, it seems evident that they are neither. The meaning of the normative should or ought is determined by its role in an inference in which there is an expression of a desire combined with a description of a means-end relation and relevant circumstance. Much as Poincaré and Quine demonstrated that we cannot classify Newton’s formula F = ma in isolation as analytic or synthetic, consideration of practical inferences shows that sentences of the form I should (ought to) do M are neither descriptions nor emotional expressions, but derive their meanings from the inference contexts in which they occur. Again, the proper unit for analysis is the discourse as a whole, not individual sentences occurring within it.

2.3. Assigning Topics and Logical Paraphrase

A discourse is commonly defined as a combination of sentences about a common topic or topics. This definition does apply to a wide range of discourses that combine sentences in the indicative mood, as for a conversation about the weather, an account of a meeting between two friends, a scientific treatise describing the properties of a newly discovered plant, or a novel about the interactions between a restricted number of central characters. For such pure indicative mood discourses, the anaphoric pronouns he, she, and it function as devices linking constituent sentences, along with relative pronouns such as who, which, and that used to embed sentence strings. As we shall presently see, defining a discourse is straightforward for indicative mood inferences of the kind that can be paraphrased and then represented by the symbolism of modern predicate logic. It seems also possible to apply the definition to sequences of imperatives such as Pick up all the toys in the room. Then pick up the papers, and mop the floor. Here we can perhaps assign as the common topic the action of cleaning up the room. But the definition does not apply, certainly not in an obvious way, to hybrid discourses combining sentences with different uses. What is the topic, for example, of the practical inference formulated at the end of the previous section? Is it the end of getting paper, the means of going to a store, or the circumstance of it not raining? Because there is a description only of the circumstance, we don’t seem able to assign a topic. And what is the topic of the speech conversation between Bill and Betty? In it Bill identifies himself, greets Betty, describes his whereabouts, and promises to help, while Betty promises to reimburse. It seems arbitrary to assign as its topic the offer to help, rather than the participants in the dialogue or the weather. For such hybrid discourses, the unifying aspect seems to be their overriding purposes, and these can be complex and subject to interpretation. There may not be any one topic that the Bill/Betty conversation is about, nor even several topics, but its special combination of sentences can be interpreted as fulfilling the common general purpose of notifying of whereabouts and coordinating actions.

Some such assignment of topic, even though vague and subject to different interpretation, seems necessary to set boundaries for a given discourse. A phone conversation between two people may last for over an hour. Rather than describing this as a single discourse, it is usually useful for purposes of analysis to divide it up into portions for which topics can be assigned. We can then assign boundaries between those parts of the extended conversation that deal with the weather, those informing of mutual activities, those providing instructions, or the myriad purposes served by the conversation. We then define a discourse block as a combination of sentences with a common topic or topics, with the concession that this assignment of topics as a means of assigning boundaries is subject to differing interpretations. The indefiniteness of such assignment poses a problem for discourse analysis as a branch of linguistics, as we shall see in the section to follow.

Topic assignment can be made very precise, however, for inferential discourses represented by the symbolism of modern predicate logic. In classical Aristotelian logic, it was possible to assign a common subject only for so-called immediate inferences, valid deductive inferences with a single subject such as the inference All men are mortal; therefore, some men are mortal of the form All S is P ∴ Some S is P. Here the subject S designates the topic of the inference, what the inference is about. But for syllogisms such as All men are mortal. All Greeks are men. Therefore, all Greeks are mortal, the so-called Barbara syllogism of AAA mood in the first figure, there is no single subject and therefore no common topic. The inference is represented by

with two subjects, the middle term M and the subject of the conclusion S, and two different topics, men and Greeks. The major innovation of modern predicate logic as founded by Frege, Peirce, Russell, and others was to paraphrase sentences of the form All S is P and Some S is P in such a way as to provide a single subject for deductive inferences. Paraphrases of such sentences introduced a more general subject common to these inferences and converted original subjects into predicates.

This procedure can be applied to the Barbara syllogism. The sentence All men are mortal is paraphrased as Everything is such that if it is a man then it is mortal, with thing (or material thing) introduced as subject and men converted to the predicate is a man. It is then symbolically represented by the universal quantifier ∀x standing for every thing and the material implication → standing for if … then. This representation as applied to All men are mortal thus yields ∀x (Mx→Px). The syllogism as a whole is represented symbolically by

The variable x within the quantifier is said to have as its extension a domain, and this domain becomes the common topic of the syllogism. As applied to Some man is mortal of the form Some S is P, paraphrase produces There is some thing (at least one thing) such that it is S and it is P of the form ∃x (Sx ⋀ Px), with the existential quantifier ∃x representing some thing (or there is at least one thing, there exists something), the conjunction sign ⋀ standing for and, and the second and third occurrences of the variable x representing the anaphoric it. The immediate inference from All men are mortal to Some men are mortal is then represented by

The effect of the paraphrase is to convert the original subject of the inference men into the predicate is a man, and introduce the new all-inclusive subject thing. This introduced subject now enables a shift from men to (material) things as the inference’s common topic.

The specialized task of predicate logic is to evaluate deductive inferences as valid or invalid. The technique of paraphrase of modern logic and the introduced symbolism had the effect of greatly extending to the scope of this logic beyond Aristotelian syllogisms with three constituent subject and predicate terms and two premisses to inferences with any number of terms and any number of premisses. It also permitted the extension of logic beyond sentences ascribing attributes such as being mortal or being a man to sentences stating relations, sentences such as Every elephant is larger than some lion used to state relations between two or more kinds of objects.

But beyond this more specialized function of logical evaluation, this technique had wider implications for philosophy. A logical paraphrase of a sentence S1 eliminates the original subject by converting it to a predicate and introducing a more general subject to produce a sentence S2 that is claimed to preserve the meaning of the original by insuring that S2 is true if and only if S1 is true. Such a meaning-preserving paraphrase is a reduction of S1 to S2. Applied to inferences in which S1 occurs in combination with others, these paraphrases become a method of reducing, not only individual sentences, but whole discourses to those with different subjects. Evaluating the success or failure of such discourse reductions then becomes an important means for philosophy to determine relationships between forms of discourse. The generality of logical paraphrase makes possible these determinations. Standard applications are to descriptive empirical sentences such as All men are mortal, All copper conducts electricity, or "All bodies in the vicinity of the earth accelerate at the rate of s = 16t²", sentences whose truth or falsity is determined relative to observational evidence. But logical paraphrase can be extended generally to any indicative sentence that can occur in an inference with reoccurring subject and predicate terms and can be evaluated as valid or invalid. These include fictional sentences such as "All the heroes of Homer’s Illiad are protected by some god, the (false) mathematical sentence Every prime number is evenly divisible by some positive integer, and the theological sentence of Christianity Some human is the son of some god. None of these sentences is evaluated in the manner appropriate for Copper conducts electricity", but there are procedures that enable readers of the Illiad to assess assertions about heroes and gods, mathematicians to reach consensus about assertions about prime numbers, and Christian theologians to agree about divine origins. When occurring in the context of inferences, they can be paraphrased, represented by the appropriate symbolism, and used in the process of evaluating the relevant inferences. The scope of logic thus extends across very different forms of discourse, and this makes possible its use as a tool for making comparisons.

At this preliminary stage, it is important to offer a warning. Reductions of one form of sentence to another through logical paraphrase are conducted relative to forms of discourse with special features. They are restricted to sentences within inferential contexts. Typically these sentences are in the indicative mood expressing true or false propositions, though extension of logical paraphrase to imperatives seems possible. But expressions of wants and preferences as constituents of practical inferences seem immune to paraphrase, as are the normative conclusions derived from them. There is also a wide range of sentences for which it is difficult to arrive at consensus about the truth or falsity of the propositions they express. These include the constituent sentences of interpretive narratives characteristic of the arts, literature, and religion. Logical paraphrase thus has restricted application. But those areas in which it can be applied provide valuable models for extending conclusions from the relatively well known to the amorphous and indefinite. Relationships between the principal forms of discourse in use within developed civilizations are of bewildering complexity. By isolating and examining the most tractable of these relationships with the help of logic we have the means for understanding at least some of the basic features of what is inherently more complex.

2.4. Discipline Boundaries

Philosophy has been in the unique position of first originating and then spinning off to separate institutionalized professions the principal areas of the sciences. This began with the first speculative theories of physics of the Greek early cosmologists and Aristotle, theories which became the basis for experimental physics as developed in the sixteenth and seventeenth centuries by Galileo and Newton. This was followed in the eighteenth and nineteenth centuries by the incorporation of biology into the natural sciences and its separation also from natural philosophy. In the nineteenth and early twentieth centuries, psychology became an experimental science after origins with Locke, Berkeley, Hume, and Kant, and at about the same time sociology separated from its original home in Hegelian idealism. In the late nineteenth century and throughout the twentieth, philosophical attention was directed to language, first by Peirce, Frege, Russell, and the early Wittgenstein of the Tractatus, then later by the ordinary language philosophers—Ryle, J. L. Austin, and the later Wittgenstein of the Investigations. As before for physics, biology, psychology, and sociology, many of the results of these philosophical discussions have been incorporated into the science of linguistics through its formulation of syntactic, semantic, and pragmatic rules.

Discourse analysis as a branch of linguistics has also been directed towards conversational discourse, and this presents the problem of determining what portions of this topic remain in the more indefinite and controversial discipline of philosophy, what have been relegated to a specialized discipline making steady progress in reaching consensus on features of discursive language. That discourse analysis lacks the exactness of other branches of linguistics is indicated by the difficulty linguists encounter in specifying what it is about. They agree it is directed towards actual speech and writing and that within extended portions of speech and writing it is possible to isolate discourse fragments with boundaries that separate off one fragment from others. These bounded fragments— what I have been referring to as discourse blocks—then become the basic units of analysis for linguistics, replacing sentences as the objects of descriptions of syntactic and semantic rules. For discourses like our Bill and Betty sample, discourse blocks are portions of a conversation between two or more participants, and for many of these blocks transactional and pragmatic features that control speaker/hearer interchanges are prominent. For single turn monologic discourses the transactional may have minimal importance.

The problems raised earlier in this chapter apply to linguists’ attempts to establish boundaries between separate blocks. What determines whether a conversation is to be itself a discourse unit for analysis or is to be divided into two or more? For some cases of descriptive language we seem to have an answer, as when a conversation begins by describing the weather and then shifts to a description of a couple’s marital problems familiar to the interlocutors. Where this occurs, topics can be assigned to discourse fragments, and used to mark the boundary between two distinct discourse blocks. Brown and Yuul claim that such shifts of topic can be used in this way, though they concede that these techniques provide only a guide.¹ Difficulties of assigning topics are emphasized by Michael Stubbs in his example of the opening of Faulkner’s The Sound and the Fury: Through the fence, between the curling flower spaces, I could see them hitting. They were coming toward where the flag was and I went along the fence.² Background knowledge leads the astute reader to assign golf as the topic of this passage, for interpreting the word hitting as hitting golf balls and flag as golf hole flag enables us to understand the subsequent occurrence of the word caddie. This suggests that topic assignment is an interpretation made to establish coherence and connection between the constituent sentences of a discourse, and that such an interpretation may vary from one hearer/reader to another in a way that depends on their backgrounds. Where nondescriptive language is used, as in our opening sample conversation between Bill and Betty, we saw in the previous chapter that shift of topic cannot be used in establishing boundaries between blocks. We must instead appeal to the purpose for which a given discourse fragment is used, and this is even more obviously subject to different interpretations. Such subjective variation in isolating units of analysis reinforces the view that discourse analysis as a branch of linguistics faces problems in reaching consensus not found where linguistics restricts itself to describing rules of sentence formation. Many of the conclusions of philosophers of language such as P. F. Strawson, Austin, and H. P. Grice have indeed been incorporated into pragmatics as a branch of linguistics in the form of rules of presupposition, interaction within conversation, and the performance of speech acts. This incorporation has required shifting from individual sentences as the units of analysis to wider discourse contexts in which the sentences occur. But this incorporation has not produced an exact science whose ability to reach consensus contrasts with the indeterminacy that marks much of philosophy.

Also, the development within linguistics of pragmatics and discourse analysis should not obscure very different interests that philosophy and linguistics have in the study of language. Discourse analysis is directed towards regularities in discourses formed within natural languages as the shared languages within speech communities.³ These are contingent regularities described by generalizations about how language as an instrument of communication happened to have evolved within human communities on our planet. In contrast, philosophy’s attention has been directed towards relationships between the natural (ordinary) language of a speech community and various forms of specialized languages used within the institutions of a civilization after the advent of writing. It has also been directed to contrasts and comparisons between natural language sentences and discourse blocks and more primitive forms of communication within animal and insect societies and means of receiving information from and reacting to a changing environment. Philosophy thus conceives both sentences and discourse blocks as signs to be compared to other more primitive forms of signs. In the process of describing such comparative sign features, philosophy attempts to isolate necessary features of discourses used both within conversational speech and specialized institutions, features that such discourses must have in order to carry out the pragmatic functions characteristic of their sign level.⁴ This focus on necessary features rather than on contingent regularities distinguishes the philosophy of language from linguistics. We shall return to the problem of distinguishing the contingent from the necessary in Section 8.2 below.

The primary tool used in this search for the necessary has been logic. As noted, logical representation applies to any sentence, and is neither restricted to sentences of certain natural languages, nor to sentences that have a certain type of subject matter. Inferences by means of logical symbolism include those from daily conversation and the specialized discourses of mathematics, the natural sciences, and law. The scope of logic thus extends to the levels of both natural and specialized languages. The principles on which deductive logic is based also are means of understanding relationships between the uses of individual sentences and sentences as occurring within the contexts of deductive inferences. The individual sentence John is sitting can at one time be used to make a true statement; at another, when John stands up, the statement would be judged false. But as occurring in the inference, John is sitting; therefore, he is sitting the truth value of the proposition being expressed must be assumed to remain constant. This requirement is formulated as the principle of identity, one of the two foundations of deductive logic.

The other foundation is the principle of bivalence, the principle that a proposition must be either true or false. It cannot be applied to individual sentences, for we may have insufficient evidence for asserting with confidence a sentence such as It will rain tomorrow, and can only assign to it a probability. Physicists are currently not able to assert or deny Dark matter is causing the expansion of the universe, and thus cannot confidently judge this to be true or false. But when evaluating the validity of deductive inferences in which these sentences occur, we must assign them only these values. This is because deductive validity is defined as a relation between premisses and conclusion that holds under the condition that if the premisses were to be true, the conclusion must be also. Evaluation for validity assesses only a relation between premisses and conclusion, not their factual truth or falsity. Since the hypothetical if used in evaluation does not require any assertion of premisses or conclusion, adequacy of evidence and the probability of an event are irrelevant to the task of evaluation. Similar considerations apply to isolated imperatives and those embedded within inferential contexts. It is possible for someone to neither obey nor disobey the imperative Lift up the car, as he may lack the strength or the opportunity for carrying out the command. But these violations of bivalence as applied to imperatives are irrelevant to the evaluation of the inference Lift up the car and place it on the rack; therefore, lift up the car. In assessing this inference from a conjunction to a conjunct as valid we are determining only that if the premisses were obeyed, the conclusion would have to be also. Considerations of ability and opportunity are irrelevant to any determination of this entailment relation.

The centrality of logic to philosophy as itself a special discipline is derived from the descriptive role that philosophy performs within contemporary society, the role of determining relationships between different institutionalized specialties and of relating features of the uses of natural language with the languages used within these specialties. For this, the abstractness of logic provides a valuable tool.

3

Descriptive Frameworks

Discourse frameworks are the structures of specific discourses and the means by which we classify them. We turn now to an outline of some features of descriptive frameworks that provide these structures for descriptive conclusions. Frameworks in general can be characterized in different ways. The most basic is in terms of the ways by which the assertion or acceptance of conclusions is justified, that is, by the rules used to derive conclusions from premisses, what Wilfrid Sellars terms the logical space of reasons.¹ We can use this characterization to distinguish descriptive frameworks from normative frameworks appealing to desires and aversions in assertions of what actions ought to be performed, frameworks to be discussed below in Chapters 6 and 7. We can also distinguish descriptive frameworks by the purposes for which they are used, their means of introducing nouns used as subject terms and identifying their referents or designata, and by the vocabularies (lexicons) employed within them. As an example, the framework for discourses used within the natural sciences is characterized by the purpose of providing information about our environment, by their methods of term introduction and reference, and by application of empirical methods of evaluation using inductive rules of inference. The discourses themselves are combinations of sentences formed from lexical items, some of which are unique to a specialty, others shared with specializations within other sciences and occurring in natural languages. In these respects, the framework differs from those for mathematics and fictional discourse. Within the framework of the sciences we can distinguish the sub-frameworks of such specific sciences as physics and zoology, each with its own distinctive (though overlapping) vocabulary and means of term introduction and reference. Frameworks are specializations of features typically combined within natural language discourses in complex ways. These combinations result in overlaps between descriptive frameworks proper and what in the second section of this chapter I distinguish as overlapping frameworks with distinctive standards of justification. One of them, the mental transactional framework in which we describe the mental states of ourselves and others, overlaps in ways that create controversies to be discussed in the next chapter.

3.1. Carnap’s Internal/External Distinction

The concept of a discourse framework was introduced by Rudolf Carnap in his seminal paper Empiricism, Semantics, and Ontology, though his terminology is different from that employed here.² Carnap applies this concept to sentences within inferential contexts as represented by the symbolism of modern predicate logic. Consider the sentences All men are mortal and All prime numbers are divisible by themselves. The first would be paraphrased, as we have seen in Section 2.3, by All (material) things are such that if they are men, they are mortal and represented by ∀x (Mx→Px), with the first occurrence of the variable x representing the introduced subject material thing and the second and third occurrences the anaphoric pronoun they. Such a representation guarantees a common topic for any inference in which the sentence might occur. This sentence is formulated within what Carnap refers to as the material thing language. The second sentence, All prime numbers are divisible by themselves, would be paraphrased by All numbers are such that if they are prime, they are divisible by themselves and represented, let’s say, by ∀x (Px→Dx). Here the first occurrence of x represents the subject number as introduced by the paraphrase. Such a sentence is formulated within what Carnap referred to as the number language. As noted above, general subjects such as material thing and number are introduced by paraphrase in order to insure that there is a