PEOPLE AS OBJECTS OF STUDY

In a study disciplined by science rather than by grammar, all theoretical statements must be supported by evidence from the senses or on the basis of valid reasoning from the results of evidence from the senses. This necessitates not only setting aside the unsupported assumptions of the linguistics of language but also holding in abeyance any possibly interfering preconceptions concerning the nature of thought, ideas, mind, will, consciousness, sensation, perception, and the like, and concepts of culture, institutions, conventions, and the like. Theoretical statements in these areas do not necessarily need to be set aside for other purposes, but should any of these be needed in the foundations of a scientific linguistics, they would have to be carefully and explicitly justified on the basis of evidence. Without such a Galilean stance toward the tradition we have no protection against possible confusions and fallacies originating in that tradition or in our understanding of it. Science starts with official doubt—the suspension of belief or credulity.

Leaving all such preconceptions aside, we begin by taking as our objects of study only real objects, which exist independently of theory and about which we can obtain observational and experimental evidence from different points of View. We will be concerned with a linguistic point of view, that is, a point of view consonant with the observational scope of linguistics outlined in chapter 2.

As suggested earlier, the objects of study in human linguistics are primarily people, individually and collectively. Their existence is well confirmed by observation. They can be observed not only from a linguistic point of view, but from other points of view as well. Unlike the objects of language they do not need to be introduced by a special assumption. They are already there in the real world.

Sound waves are also involved. Bloomfield noted that individuals in a human society cooperate by means of sound waves (1933:28). He gave as an example Jill seeing an apple in a tree, making a noise with her larynx, tongue, and lips, and Jack vaulting the fence and getting the apple for her. Such transactions are easily observable and can even be made the basis of experiments.

Light and other forms of energy must be included, for gestures and facial expressions cannot be ignored, nor the mechanical energy of a tap on a shoulder or the grip of a handshake. These forms of energy can also be observed and measured, although in some cases the techniques of physical analysis may not yet be as well developed as they are for the sound waves of speech and their physical production and reception.

Physical objects and the physical environment sometimes play an important role, such as lanes, fences, apples, and trees, and sometimes objects with marks on them like newspapers, books, and keep-out signs. These are also objects given in advance that can be studied from other points of view.

In sum, then, the objects of study in human linguistics are primarily people, but also the means of energy flow, and sometimes other physical objects and the physical environment. All these are relevant to a scientific study of people from the point of view of how they communicate.

The next question is how to conceive of the task of human linguistics. Recall that the task of the linguistics of language in business as usual is the scientific study of language, which we know to be incoherent. The task of human linguistics is phrased instead as the scientific study of how people communicate. This rejects the goal of studying language, but retains the goals of being scientific and of seeking explanations in terms of people.¹ In this statement of the task, scientific is understood in the same sense as in the physical and biological science. People refers, as in ordinary language, to real flesh and blood people. And communicate refers to the scope of phenomena of interest in linguistics as briefly indicated in the beginning of chapter 2. This term does not have a special technical sense here either. It simply points to the initial scope of the questions we are asking.

SYSTEMS IN HUMAN LINGUISTICS

In order to avoid a possible source of confusion, an explicit terminological distinction is maintained in human linguistics between the major objects of theory and the real objects that they are a theory of. Such a distinction is always honored in science, but not often carried explicitly by means of separate terms. The distinction is especially clear when a model or theory of some real object is tested against observations of the real object. Thus a linguistic model or theory of a real person is called a communicating individual, sometimes simply an individual; and a linguistic model or theory of a group of real people, large or small, along with the communicatively relevant aspects of their environment, is called a linkage.

A terminological distinction will not be enforced, however, for nontechnical terms such as person, people, group, and so on. Thus the term person will sometimes be used for a real-world person, and sometimes for our everyday-language concept of a person. Care will be taken at all times, however, to make clear whether the real object or our concept or theory of that object is being discussed.

The distinction between communicating individuals and concepts of persons can be understood in terms of how much is covered. Communicating individuals are defined in terms of properties that reflect only those aspects of real persons that are relevant to understanding how they communicate. Similarly, linkages are defined in terms of properties that reflect only those aspects of groups of people and the environment that are relevant to how they communicate. Communicating individuals and linkages, therefore, are abstractions. They include only properties of relevance for linguistics, leaving aside any nonlinguistic properties of people that would belong in theories of real people from other points of view. In this way we exclude from direct consideration in the theory that which is irrelevant to the primary concerns of linguistics.

Sometimes it is of interest to consider how a person communicates in a particular group rather than how he communicates in general. For this we define a participant as an abstraction of a communicating individual that includes only those linguistic properties relevant for understanding how the person communicates in that group, i.e., relevant in respect to a particular linkage. For example, a participant representing a person communicating in his work group would exclude the many properties of the individual relevant to how he communicates in other groups but not relevant to how he communicates in his work group.

Thus a participant is an abstraction of a communicating individual just as a communicating individual is an abstraction of a concept of a person. Again this excludes from consideration that which is irrelevant for a given purpose.

A participant is a constituent of a linkage. Linkages are composed of constituents. Other types of linkage constituents are channels, props, and settings. Channels are abstractions representing the physical means by which the sound, light, or other communicative energy is carried. Props are abstractions representing any physical objects that may be relevant, such as clocks that tell the time, doors that are knocked on, and signs that are read. Settings are abstractions representing relevant aspects of the physical environment, such as living rooms, ticket counters, and conference rooms. These constituents of linkages are also constructs of theory which are distinguished terminologically from the corresponding real objects.

Communicating individuals, linkages, and linkage constituents are treated as systems in a way familiar in science. A system is a representation in theory of a given physical object or group of objects. It is separated from its environment by arbitrary boundaries. Setting up such systems is our first theoretical act. It formalizes our scope of inquiry at the theoretical level by enclosing a limited domain that we think is coherent and reasonably self-contained. If the objects modeled in the system are not entirely isolated and self-contained, in that they are affected by or affect other objects in the surroundings, then we postulate inputs and outputs representing physical flows of energy such as sound or light that carry these influences across the boundaries. And if the objects modeled are parts of larger objects, then the inputs and outputs represent the influences that cross the boundaries between the objects modeled and the larger objects. This concept of system is quite different from the concept of system found in the linguistics of language following Saussure and others in that it represents a theory or model of a physical reality, as is the case in the physical and biological sciences (Yngve 1985).

Setting up a system in science is justified on the basis that we are free to decide what part of nature we wish to study. The system boundaries formalize that free decision on our part. There may be many ways of setting up a system, with boundaries differently placed. A wise choice in the placement of system boundaries may simplify the ensuing theory and help to provide insight into the phenomena; an unwise choice may complicate the theory and hinder easy understanding.

Communicating individuals and linkages have external and internal boundaries.

The external boundary of a communicating individual usually coincides with the conceptual boundary of the physical organism, but it may include such objects as pencils, eyeglasses, clothing, and other extensions of the person, and it may exclude parts of the body such as fingers used for counting. The external boundary of a linkage represents the boundary of a cohesive group like a customer and clerk, a family around the dinner table, a committee in a meeting room, the citizens of a town, and so on. Here again, the boundaries are arbitrary and set on the basis of convenience for theory in confronting the observed phenomena being studied. Similar considerations hold for the other types of systems.

The internal boundaries are internal to the person or group. Human linguistic theories are not complete theories or models of people; they are only theories of how they communicate. Any scientific theory is limited to certain aspects of some portion of reality. Physics, for example, concentrates on the physical properties of matter, while chemistry concentrates on the chemical properties. A more complete model or theory would be concerned with both. In terms of Bloomfield’s Jack and Jill example, human linguistics would introduce two communicating individuals as models of the two persons, and a linkage to model the pair. These would be set up to show how they communicate, but they need not be concerned with modeling why Jill wants the apple or why Jack is willing to get it for her. Such questions could well be relegated to other disciplines. The internal boundaries of our systems, then, are the boundaries within the persons and within the groups that separate the linguistic properties of the persons and groups, the properties of direct concern to linguistics, from the nonlinguistic properties, the properties which are not of direct concern to linguistics.

The units of analysis in human linguistics, then, are not sentences with properties of declarative or interrogative; not words with properties of case, gender, or tense; not phonemes with properties of consonantal or grave; and not features. These would need to be introduced by assumption. They are instead people, individually and collectively, conceived in linguistic theory as communicating individuals and linkages having properties modeling how people communicate. These properties are postulated so as to account for the evidence, and they must be carefully tested against observations by the senses of real people.

Thus the framework of human linguistic theory is basically quite different from the framework of the linguistics of language. Yet inasmuch as it has to account for the same evidence, and more besides, one would expect the two theories to exhibit certain parallels and correspondences. These can be of great help in developing human linguistic theory as long as we take care not to be confused by the psychological and social reality of grammar fallacies, for words, sentences, and the like, or grammars and lexicons, are by no means properties of people.

THE LAW OF COMPONENTIAL PARTITIONING

Since we are laying the foundations of a new theoretical structure, it will be necessary to justify everything with the greatest of care so as to guard against errors at the very beginning that might have unwanted repercussions later. Therefore it is appropriate to ask whether there really is proper justification for defining communicating individuals, participants, and linkages in terms of properties.

The usual initial tactic in the linguistics of language is to introduce speech-sounds, forms, meanings, and other objects of language. Bloomfield, for example, said that the linguist focuses on the speech-signals, and for this he introduced his fundamental assumption of linguistics: “In every speech-community some utterances are alike in form and meaning” (1933:78). Distinctions of “the same” or “different” were to be obtained from the everyday knowledge of the linguist, by trial and error, or from someone who knows the language. Application of a method of sames and differents then led to postulating the properties of these linguistic objects that had been introduced by assumption. Although this tactic gave the methods the look of being empirical and scientific, it did not lead to a coherent linguistic science.

Human linguistics, not accepting any assumptions of the objects of language, departs from the usual sorts of analysis. Bloomfield himself provided an important hint when he noted that “the difference between distinctive and non-distinctive features of sound lies entirely in the habit of the speakers.” (1933:77). It is not clear what Bloomfield meant by “habit”, and in any event he did not pursue the matter, for he took the distinctions of “the same” and “different” made by speakers and hearers not as evidence about the habits of the speakers and hearers but as evidence about features of speech-sounds in an assumed domain of language.

Human linguistics, on the other hand, takes the distinctions made by speakers and hearers directly for what they are: evidence for properties reflecting aspects of the speakers and hearers themselves that are involved with the production and reception of sound waves and other forms of energy when they communicate. A method of sames and differents is then applicable to working out the properties representing aspects of these real objects. By this straightforward and obvious tactic the need for assuming the objects of language is eliminated and the psychological and social reality of grammar fallacies are avoided.

Let us examine this program in a bit more detail.

We start with an observational fact: different people show communicative differences and the same person shows communicative differences at different times. The same is true for groups. Different groups show communicative differences and the same group shows communicative differences at different times. Perhaps in reality no two people or no two groups are ever exactly alike, and no person or group is ever twice the same, but this does not need to be assumed. All that is needed is the easily verifiable fact that differences are widely observed. It means that the theory will have to confront the possibility of a very large number of systems, all different.

The way in which the theory can confront a very large number of systems, all different, stems from another observational fact: different people show communicative similarities and the same person shows communicative similarities at different times. The same is true for groups of people: different groups show communicative similarities and the same group shows communicative similarities at different times. Again it may well be that any two people will show similarities, and any two groups will show similarities, but this does not need to be assumed either. All that is needed is the easily verifiable fact that similarities are widely observed.

Evidence of communicative similarities and differences of people can be obtained from informant methods, field methods, experimental methods, and other observational methods. By these means data can be obtained on the communicative similarities and differences of different people and of the same person at different times, and of different groups and the same group at different times.

We interpret these observations of similarities and differences as flowing from underlying similarities and differences in the real people and groups themselves. On this basis we postulate properties of the communicating individuals and linkages. The properties are theoretical constructs set up to represent the real similarities and differences and thus to account for the observations of similarities and differences. Some of the properties are relatively permanent. Others are changeable, and it is their changes that account for the observed communicative behavior. Some of the properties parallel direct observation, but many are set up on the basis of indirect evidence. A person may be represented as a communicating individual with properties of having a high voice, or low, or as having a contrast between caught and cot, or between pin and pen; or a student in a classroom may be represented as raising his hand. A group may be represented as a linkage with properties of discussing a particular topic, or there may be eye contact between a teacher and a student.

The justification for postulating properties, then, is the observational evidence of communicative similarities and differences of people and groups of people, and the usual presumption of causes found everywhere in science, and indeed in everyday common sense. That is, if we observe similarities and differences of people, the similarities and differences must be caused by underlying similarities and differences in the people themselves, which we model in terms of properties.

On the basis of these considerations, we are justified in proposing an observationally based law, the law of componential partitioning (Yngve 1984):

The communicative aspects of a person, or of a group and the communicatively relevant parts of its environment, can be represented as a communicating individual, or a linkage, in terms of a set of component properties in respect to which different individuals or linkages show partial similarities and differences and in respect to which the same individual or linkage shows partial similarities and differences at different times.

This is the most general law of communicative behavior. It underlies the whole of human linguistics in much the same way as Bloomfield’s assumption underlies the whole of the linguistics of language. But rather than being an unjustified special assumption, it is an empirical law supported by the publicly available and reproducible evidence of observed similarities and differences of real objects.

This law reflects the most basic way in which communicating individuals and linkages are structured—in terms of properties. The structuring in terms of properties is of great significance for the conduct of research, for it holds out the hope of reasonably independent studies of single component properties or small sets of properties and their interrelations.

PROPERTIES OF PEOPLE

Human linguistics is not unique in the way it analyzes its subject matter in terms of properties. There are parallels in many other sciences. For example, the chemical properties of elements and compounds in chemistry are theoretical constructs postulated on the basis of observations of similarities and differences in the behavior of different aggregates of real matter in chemical reactions. In classical genetics, constituents of biological organisms (genes) are postulated as theoretical constructs representing certain properties that account for the observations of similarities and differences of real plants and animals, and how the similarities and differences propagate through different generations.

These examples and others that could be cited are strictly parallel to the way in which human linguistics postulates properties of communicating individuals and linkages on the basis of observations of similarities and differences of different people or different groups, and the same person or the same group at different times.

Formalizations in human linguistics are thus based on properties. Two sets of properties are initially defined, a dynamically changing set of properties called conditional properties and a more static set of properties called categorial properties.

The conditional properties reflect the momentary state or condition of the individual at any instant of time; their changes accompany communicative behavior in regular ways. For example, a person may have the turn in conversation at one moment, and at the next moment not, or at any instant he may be on the spot to answer a question or not, or he may have referred to a particular person rather than to some indefinite person.

The categorial properties represent the categories or dimensions along which these changes take place. For example, there would be categorial properties of being able to have the turn in conversation or not, and of being able to be on the spot to answer a question or not. Categorial properties are relatively unchanging. The categorial property of being able to have the turn in conversation or not will have developed in childhood and remain relatively unchanged throughout life, while the conditional property of now having the turn, and now not having it, will change many times as the person communicates.

It will be seen shortly that there are other types of properties, and that the properties of individuals and linkages are structured in complex ways.

It is worth noting that binary properties of individuals and linkages are completely sufficient in principle for reflecting how persons or groups are the same or different with respect to a battery of observational tests. It is well known that complex tests can be resolved into simpler binary tests, and it is well understood that continuous variables can be characterized to any degree of accuracy by means of binary variables.

If the principle of correspondence with the linguistics of language suggested on p. 41 holds here, most properties can be expected to be naturally binary. However some properties such as phonetic tongue height may seem to require continuous variables. But in this case neither the speaker nor the listener can distinguish phonetically more than a certain number of tongue heights, and a discrete representation would be perfectly satisfactory. In practice, however, it may sometimes be convenient to use nonbinary or even continuous notations in the theory, recognizing that they could be reduced to binary notations.

Another important point is that with only a few properties many different individuals can be distinguished. Although it might require thousands or millions of properties to fully characterize a communicating individual, yet with only 300 binary properties it would be possible to distinguish more than 10⁹⁰ different individuals. If the properties were not uniformly distributed over the individuals, more properties would be required, but in any case it appears that a very large number of different individuals can be distinguished by means of relatively few properties.

Thus we see that the unit of generalization in human linguistics, as in other sciences, is the property. By means of properties we can actually make generalizations even if no two people or no two groups are ever the same or no one person or group is ever the same at different times.