p.69

14

Causation: Partly Emergent and Partly Metaphorical

We have seen in our discussion of grounding that there are directly emergent concepts (like up-down, in-out, ob­ject, substance, etc.) and emergent metaphorical con­cepts based on our experience (like the visual field is a

CONTAINER, AN ACTIVITY IS A CONTAINER, etc.). From the

limited range of examples we have considered, it might seem as if there were a clear distinction between directly emergent and metaphorically emergent concepts and that every concept must be one or the other. This is not the case. Even a concept as basic as causation is not purely emergent or purely metaphorical. Rather, it appears to have a directly emergent core that is elaborated metaphorically.

Direct Manipulation: The Prototype of Causation

Standard theories of meaning assume that all of our com­plex concepts can be analyzed into undecomposable primi­tives. Such primitives are taken to be the ultimate "building blocks" of meaning. The concept of causation is often taken to be such an ultimate building block. We believe that the standard theories are fundamentally mistaken in assuming that basic concepts are undecomposable primitives.

We agree that causation is a basic human concept. It is one of the concepts most often used by people to organize their physical and cultural realities. But this does not mean that it is an undecomposable primitive. We would like to suggest instead that causation is best understood as an

p.70

experiential gestalt. A proper understanding of causation requires that it be viewed as a cluster of other components. But the cluster forms a gestalta whole that we human beings find more basic than the parts.

We can see this most clearly in infants. Piaget has hypothesized that infants first learn about causation by re­alizing that they can directly manipulate objects around them---pull off their blankets, throw their bottles, drop toys. There is, in fact, a stage in which infants seem to "practice" these manipulations, e.g., they repeatedly drop their spoons. Such direct manipulations, even on the part of infants, involve certain shared features that characterize the notion of direct causation that is so integral a part of our constant everyday functioning in our environment---as when we flip light switches, button our shirts, open doors, etc. Though each of these actions is different, the over­whelming proportion of them share features of what we may call a "prototypical" or "paradigmatic" case of direct causation. These shared features include:

The agent has as a goal some change of state in the patient. The change of state is physical. The agent has a "plan" for carrying out this goal. The plan requires the agent's use of a motor program. The agent is in control of that motor program. The agent is primarily responsible for carrying out the plan. The agent is the energy source (i.e., the agent is directing his energies toward the patient), and the patient is the energy goal (i.e., the change in the patient is due to an external source of energy).

The agent touches the patient either with his body or an instru­ment (i.e., there is a spatiotemporal overlap between what the agent does and the change in the patient). The agent successfully carries out the plan. The change in the patient is perceptible. The agent monitors the change in the patient through sensory perception.

p.71

There is a single specific agent and a single specific patient. .

This set of properties characterizes "prototypical" direct manipulations, and these are cases of causation par excel­lence. We are using the word "prototypical" in the sense Rosch uses it in her theory of human categorization (1977). Her experiments indicate that people categorize objects, not in set-theoretical terms, but in terms of prototypes and family resemblances. For example, small flying singing birds, like sparrows, robins, etc., are prototypical birds. Chickens, ostriches, and penguins are birds but are not i. central members of the category---they are nonprototypical birds. But they are birds nonetheless, because they bear sufficient family resemblances to the prototype; that is, they share enough of the relevant properties of the pro­totype to be classified by people as birds.

The twelve properties given above characterize a pro­totype of causation in the following sense. They recur to­gether over and over in action after action as we go through our daily lives. We experience them as a gestalt; that is, the complex of properties occurring together is more basic to our experience than their separate occurrence. Through their constant recurrence in our everyday functioning, the category of causation emerges with this complex of prop­erties characterizing prototypical causations. Other kinds of causation, which are less prototypical, are actions or events that bear sufficient family resemblances to the pro­totype. These would include action at a distance, nonhu-man agency, the use of an intermediate agent, the occur­rence of two or more agents, involuntary or uncontrolled use of the motor program, etc. (In physical causation the agent and patient are events, a physical law takes the place of plan, goal, and motor activity, and all of the peculiarly human aspects are factored out.) When there is an insuffi­cient family resemblance to the prototype, we cease to characterize what happens as causation. For example, if there were multiple agents, if what the agents did was re­mote in space and time from the patient's change, and if

p.72 

there were neither desire nor plan nor control, then we probably wouldn't say that this was an instance of causa­tion, or at least we would have questions about it.

Although the category of causation has fuzzy boundaries, it is clearly delineated in an enormous range of instances. Our successful functioning in the world involves the appli­cation of the concept of causation to ever new domains of activitythrough intention, planning, drawing inferences, etc. The concept is stable because we continue to function successfully in terms of it. Given a concept of causation that emerges from our experience, we can apply that con­cept to metaphorical concepts. In "Harry raised our morale by telling jokes," for example, we have an instance of cau­sation where what Harry did made our morale go up, as in the happy is up metaphor.

Though the concept of causation as we have charac­terized it is basic to human activity, it is not a "primitive" in the usual building-block sense, that is, it is not unanalyzable and undecomposable. Since it is defined in terms of a prototype that is characterized by a recurrent complex of properties, our concept of causation is at once holistic, analyzable into those properties, and capable of a wide range of variation. The terms into which the causation pro­totype is analyzed (e.g., control, motor program, volition, etc.) are probably also characterized by prototype and ca­pable of further analysis. This permits us to have concepts that are at once basic, holistic, and indefinitely analyzable.

Metaphorical Extensions of Prototypical Causation

Simple instances of making an object (e.g., a paper airplane, a snowball, a sand castle) are all special cases of direct causation. They all involve prototypical direct manip­ulation, with all of the properties listed above. But they have one additional characteristic that sets them apart as instances of making: As a result of the manipulation, we

p.73

view the object as a different kind of thing. What was a sheet of paper is now a paper airplane. We categorize it differentlyit has a different form and function. It is essentially this that sets instances of making apart from other kinds of direct manipulation. Even a simple change of state, like the change from water to ice, can be viewed as an instance of making, since ice has a different form and func­tion than water. Thus we get examples like:

You can make ice out of water by freezing it. This parallels examples like:

I made a paper airplane out of a sheet of newspaper.

I made a statue out of clay.

We conceptualize changes of this kind---from one state into another, having a new form and function---in terms of the metaphor the object comes out of the substance. This is why the expression out of is used in the above examples: the ice is viewed as emerging out of the water;

the airplane is viewed as emerging out of the paper; the statue is viewed as emerging out of the clay. In a sentence like "I made a statue out of clay," the substance clay is viewed as the container (via the substance is a con­tainer metaphor) from which the object---namely, the statue---emerges. Thus the concept making is partly, but not totally, metaphorical. That is, making is an instance of a directly emergent concept, namely, direct manipula­tion, which is further elaborated by the metaphor the ob­ject COMES OUT OF THE SUBSTANCE.

Another way we can conceptualize making is by elaborating on direct manipulation, using another metaphor: the substance goes into the object. Thus:

I made a sheet of newspaper into an airplane.

I made the clay you gave me into a statue.

   Here the object is viewed as a container for the material. The substance goes into the object metaphor occurs

p.74 

far more widely than in the concept of making. We con­ceptualize a wide range of changes, natural as well as man-made, in terms of this metaphor. For example:

The water turned into ice.

The caterpillar turned into a butterfly.

She is slowly changing into a beautiful woman.

The object comes out of the substance metaphor is also used outside the concept of making but in a much more limited range of circumstances, mostly those having to do with evolution:

Mammals developed out of reptiles.

Our present legal system evolved out of English common law.

Thus the two metaphors we use to elaborate direct manip­ulation into the concept of making are both used indepen­dently to conceptualize various concepts of change.

These two metaphors for change, which are used as part of the concept of making, emerge naturally from as funda­mental a human experience as there is, namely, birth. In birth, an object (the baby) comes out of a container (the mother). At the same time, the mother's substance (her flesh and blood) are in the baby (the container object). The experience of birth (and also agricultural growth) provides a grounding for the general concept of creation, which has as its core the concept of making a physical object but which extends to abstract entities as well. We can see this grounding in birth metaphors for creation in general:

Our nation was born out of a desire for freedom.

His writings are products of his fertile imagination.

His experiment spawned a host of new theories.

Your actions will only breed violence.

He hatched a clever scheme.

  He conceived a brilliant theory of molecular motion.

Universities are incubators for new ideas.

  The theory of relativity first saw the light of day in 1905.

p.75

   The University of Chicago was the birthplace of the nuclear age.

Edward Teller is the father of the hydrogen bomb.

These are all instances of the general metaphor creation is birth. This gives us another instance where a special case of causation is conceptualized metaphorically.

Finally, there is another special case of causation which we conceptualize in terms of the emergence metaphor. This is the case where a mental or emotional state is viewed as causing an act or event:

He shot the mayor out of desperation.

He gave up his career out of love for his family.

His mother nearly went crazy from loneliness.

He dropped from exhaustion.

   He became a mathematician out of a passion for order.

Here the state (desperation, loneliness, etc.) is viewed as a container, and the act or event is viewed as an object that emerges from the container. The causation is viewed as the emergence of the event from the state.

Summary

As we have just seen, the concept of causation is based on the prototype of direct manipulation, which emerges directly from our experience. The prototypical core is elaborated by metaphor to yield a broad concept of causa­tion, which has many special cases. The metaphors used are the object comes out of the substance, the sub­stance GOES INTO THE OBJECT, CREATION IS BIRTH, and

causation (of event by state) is emergence (of the event/object from the state/container).

We also saw that the prototypical core of the concept causation, namely, direct manipulation, is not an unanalyzable semantic primitive but rather a gestalt consisting of properties that naturally occur together in our daily ex­perience of performing direct manipulations. The pro-

p.76

totypical concept direct manipulation is basic and primitive in our experience, but not in the sense required by a "building-block" theory. In such theories, each concept either is an ultimate building block or can be broken down into ultimate building blocks in one and only one way. The theory we will propose in the next chapter suggests, in­stead, that there are natural dimensions of experience and that concepts can be analyzed along these dimensions in more than one way. Moreover, along each dimension, con­cepts can often be analyzed further and further, relative to our experience, so that there are not always ultimate build­ing blocks.

Thus there are three ways in which causation is not an unanalyzable primitive:

---It is characterized in terms of family resemblances to the prototype of direct manipulation.

---The direct manipulation prototype itself is an indefinitely analyzable gestalt of naturally cooccurring properties.

---The prototypical core of causation is elaborated metaphorically in various ways.