The possibility of linguistic level of neurological organization. In order to show the inevitability of spiritual animals in the course of evolution, however, it is necessary to show not only that language is functional and can control relevant conditions that were out of reach for animal societies, but also that language is possible. From its function, we can tell that language must involve some form of overt individual behavior that others can observe and that it must be able to reach deep into their brains and control their imagination. The overt linguistic behavior must be able, therefore, to convey what the speaker is imagining so that the perception of that behavior by others causes the same content to occur in their imaginations. The overt behavior must, in other words, be a representation of the content of imagination.

As we have seen, however, such representations would have to be more complex than the signs used by other animals. Animals use cries and screams as signals to communicate with one another, and though they can be quite varied, they are inherently incapable of assigning tasks to members of a group. Their meanings and references depend on the context of their utterance, and that makes them useless in assigning different roles to individual members in a plan of social level behavior. In order for animals to use verbal behavior to share a plan that divides up tasks among them, it must be possible, in advance of acting and regardless of its context, to refer to particular members and describe how each is to behave at a later time so that every member can understand who is to behave how and in what situations.

In order for cries and screams to distribute the same plan of social level behavior to all the members, they must become organized with a subject-predicate grammar. At least two elements must be combined in each linguistic act, for it must connect an individual member with a specific way of behaving without relying on the context to identify either element. Hominids could associate particular cries with different members or other objects and different ways of behaving or other properties, but such signs must be combined in a certain way to serve as representations of a plan for their behavior in some situation. This is the function of predication, that is, taking certain signs to represent the objects and other signs to modify them in some way. But little more than predication would be needed, since the listeners can retain one instruction after another in memory and combine them all as parts of a single representation of the plan, and a series of such instructions would give every member an understanding of the whole plan. But if they had the capacity to communicate verbally how they were combining signs as subjects and predicates at all, it would be possible to complicate such utterances so that they also indicate the other objects and individuals to which the behavior is supposed to be directed, where it is supposed to occur, when, under what conditions, and the like.

The complexity of the behavioral schemata required for a primitive language can be estimated from the grammar of natural sentences in our own language. Even in sentences with a simple subject-predicate grammar, predicates are often themselves rather complex, with a verb and its objects to indicate how the action is directed at other objects in the world. (For example, if behavior is to be directed toward an object in space, the predicate will be a verb that takes a noun as its grammatical object, for example, “picking up a certain ax” or “going to a certain location”.) And complexity increases as these basic elements are modified by adjectives, adverbs, phrases and clauses, as would be needed to make clear what the individual is supposed to do. For example, “Pick up the sharp ax carefully and put in where the tools are kept.”

Repeated and varied uses of the basic mechanism of predication within single utterances may explain how all these elements can be included in natural sentences, but only if there is some way for listeners to tell how the speaker is combining them, for listeners must be able to construct the same representation in their imagination as the speaker’s. There is an asymmetry about the grammatical subject and predicate, and not only must it be clear from what the speaker says which sounds are to be treated as subjects and which as predicates, but it must be clear in each of the other phrases, clauses and modifiers based on predication. In other words, the cries and screams must be made with grammatical markers indicating how their meanings are to be combined in the listener’s imagination.

The linguistic system of representation. Considering the enormous range of possible natural sentences in our language, even a primitive language, based on a simple subject-predicate grammar, may seem too complex to be tried out as a random variations during the evolution of hominid societies. Unless there is some simple way in which the capacity for grammar could be acquired, it will not be clear that the evolution of language is inevitable, and it will not be plausible to hold that the our capacity for language is the result of a long series of accidental selection pressures that would probably not occur on other planets.

There is, however, a basically simple mechanism that could be tired out as a random variation on the primate brain, and it would account for such a primitive language. It is possible, because primates had evolved a faculty of imagination that included both spatial and structural imagination. The ability to use of language comes from the evolution of a linguistic system of representation.

The linguistic system of representation is different from the animal system of representation, because its representations can exist as public objects, that is, as sentences that are generated by overt verbal behavior.

But such a linguistic system of representation is possible only because it is constructed out of parts supplied by the animal system of representation. The manipulative animal system of representation already has a faculty of imagination (both spatial and structural imagination), and the ability to speak and understand linguistic representations comes from a higher level of neurological organization within the faculty of imagination of the animal system of representation. (That is, linguistic behavioral schemata are on a higher level of organization than locomotor and manipulative behavioral schemata, and the images on which covert behavior operates in the sensory input system are naturalistic images, which are compounds of local images and object images)

This new mechanism in the faculty of imagination would account for the “deep grammar” of “universal grammar” that Chomsky identified beneath the surface grammar of all natural languages. Chomsky argued that the remarkable capacity of humans to learn a language, including the ability to generate an indefinitely large range of sentences permitted in it, is best explained by a genetically inherited capacity, could not be explained as a result of operant conditioning (as B. F. Skinner had insisted). The surface grammar of a natural language is, as we shall see, simply the conventions by which the speaker indicates the kinds of covert behavior he is using to construct some representation in his faculty of imagination so that the listener can construct the same representation in theirs. What is universal, in other words, is a faculty of naturalistic imagination.

This explanation might also confirm a more controversial aspect of Chomsky’s position. Chomsky has puzzled many of his followers by expressing doubts about whether this genetically inherited capacity can be explained by Darwinian natural selection. What is bothering Chomsky may be the assumption of contemporary Darwinists that the details of this brain mechanism must be explained as an accumulation of traits a series of caused by a series of externally imposed special selection pressures. But it is not necessary to accept accidentalism to explain the evolution of language. The capacity for language could be another of the more revolutionary episodes in the course of evolution that are not recognized by contemporary Darwinists. If it is provided by a higher level of neurological organization, the advent of language would mark the beginning of a new evolutionary stage. Like all such radical random variations, it would start off simple, uniform and weak, and as a result of reproductive causation, the capacity for language would gradually become more complex, diverse and powerful. And as we shall see, one stage leads to another such revolutionary episode in which a yet higher level of neurological organization makes it possible to use psychological sentences and leads to reason.

What makes it possible for something as simple as a higher level of neurological organization in the faculty of imagination to account for language is the way that imagination is a foundation for the kind of learning that has been explained as a contained form of reproductive causation that takes place during a late phase of embryological development.

The trait that needs to be explained, as Chomsky insisted, is the remarkable capacity of individuals to learn natural languages of vastly different kinds, and that is explained as the evolution of behavioral schemata for generating covert linguistic behavior within the faculty of imagination in each individual brain by a form of natural selection that depends on reinforcement. (Skinner was not entirely wrong to think that operant conditioning was involved.)

The higher level of part-whole complexity in covert behavior marks them a linguistic schemata. But just as neurological development internalized spatial causation as spatial imagination and structural causation as structural imagination, so this contained form of reproductive causation internalizes the natural language used by the spiritual animal in the developing brains of its new members.

And as a byproduct, it gives language users a faculty of naturalistic imagination by which they can think about state of affairs in the natural world and understand efficient cause explanations. That is, they can see states of objects in space against the background of what is possible by efficient causation, either their possible causes or possible effects.

Verbal and nonverbal sides of linguistic representations. The basic social function of language is to represent in public behavior a certain kind of image that is constructed privately in each faculty of imagination. The (potentially) public part of linguistic behavior will be called “verbal behavior,” the words and sentences that can be spoken. But since the images in imagination are a form of representation based on the animal system of representation, the verbal behavior involved is actually just a re-representation of some kind of higher level animal representation.

The linguistic representation has, therefore, both a verbal and a nonverbal side. To show how it is possible for the verbal side, that is, overt verbal behavior, to represent the nonverbal side (or nonverbal covert linguistic behavior and the naturalistic images on which it operates as a higher level animal representation), I will consider how naturalistic images can be constructed in the faculty of imagination and then show how the nonverbal covert linguistic behavior involved in doing so can be represented publicly in overt verbal behavior. I will begin by focusing on how language is handled in the brain of the speaker and then expand on how it works in the brains of listeners who can understand overt verbal behavior.

Let us assume that the mouth and larynx are able to form the broad range of sounds needed for language, though that capacity is undoubtedly among the traits that evolved gradually during the primitive spiritual stage. We can also assume that hominids had the capacity to associate specific (complex) sounds with particular local scenes, particular objects in space, kinds of objects, and ways of behaving as their meanings, since this is a capacity that chimps have. Though that is not enough to explain how words and meanings are connected in the linguistic brain, it does allow us to see how language could evolve in the primate brain. It means that hominids could call up appropriate images in imagination when they heard certain sounds, or make those sounds when the images occurred to them.

The basic mechanism that needs to be explained is the subject-predicate grammar of verbal behavior, and its function can be explained by the meaning that is supposed to be conveyed by it. That meaning is the nonverbal side of the linguistic representation, and since it is a higher level naturalistic image that can be constructed in the erstwhile primate faculty of imagination, let us begin by recalling what is involved in spatial and structural imagination.

Local images are the telesensory images combined (according to input about the bodily condition) as a representation of the local scene relative to the animal’s body. Spatial imagination is the capacity to understand the effects of motion on the relations of objects in the territory by calling up sequences of local images by covert locomotion as well as the effects of motion in the local scene by calling up changes in the telesensory images of which it is composed. Thus, if words (as distinctive sounds) can be associated with such images at all, they can refer to (and re-identify) particular locations by way of local images and to particular objects in space by way of telesensory images in the local images. They can also refer to kinds of objects, or properties, insofar as those kinds can be recognized by telesensory images, including ways in which bodies behave. And they can refer to spatial relations among objects by way of the understanding of the structure of space through spatial imagination.

Because of structural imagination, however, objects in a local scene can also be represented by object images, which are sequences of telesensory images that can be called up from memory by covert manipulation that represent what happens to the object when it is manipulated in various ways. Object images can also be the meanings represented by words, enabling the words to refer not only to particular objects, but also to kinds of objects.

Suppose the leader devising a plan of social level behavior wanted some members to chase deer into a ravine and other members to roll rocks down on them as the deer ran by. He would think about members of the first group by calling up object images of them in his imagination by covert behavior, connecting those images with the (perceived or imagined) local image to represent them as being located in the field outside the ravine where deer are grazing. He would then call up another object image, this time of bodies running and yelling, and his covert linguistic manipulation would combine this second object image with object images of the members in the field, orienting their running and yelling in certain directions relative to the deer. Such combinations of local images and object images would create naturalistic images, or representations of objects in space as being related, moving or interacting with one another in such ways. The structure of imagination together with memory would supply sequences of images representing the changes that can occur, given the naturalistic states being picked out, and thus, the leader would, for example, have images of deer locomotion in response to the images of disturbances being caused. He could anticipate where the deer would run, and so on, allowing him to see what would happen as a means to his goal.

It is, however, one thing to think a plan and another to communicate it, and the latter requires the leader’s verbal behavior to enable others to construct the same plan in their imaginations. Even if the leader had words associated with the local images and with the object images of the members, the deer, and the behavior of chasing, that would not suffice to represent the animal representation he has constructed by his nonverbal (covert) behavior, for the listener must combine the meanings called up from memory by each word in the same way and there would be no way to tell how to do that.

References to the field in the local scene and the particular deer could, perhaps, be conveyed by pointing to them and using such words as “deer” and “field”, because pointing is public behavior and that would enable listeners to could construct representations of those objects in the local scene as objects of such kinds. But the leader could not communicate his plan for social level behavior unless his verbal behavior could control their imaginations so that they combined object images of the members being assigned the task with object images of the task being assigned to them. That is the function of grammar.

If there were a mechanism that would make the leader utter the words in a way that indicated which meanings were to be treated as grammatical subjects and which were to be treated as predicates, the listeners would be able to combine their own object and local images in the correct way. They would call up local images for the words referring to locations and an object image for each member named, and they would combine those object images with images for the kinds of behavior named, so that the listeners would also have a naturalistic image representing each member as behaving in a certain way.

In this case, however, at least three object images would be involved each naturalistic image, because members are supposed to chase the deer. Two of the object images would be treated grammatically as subjects of predication (the member and the deer) and one object image treated as a predicate (the relation of chasing), and all three would be combined in a certain order (together with local images for any references to locations) as a single naturalistic image.

Speaking: nonverbal behavior determining verbal behavior. The nonverbal (always covert) part of linguistic behavior — combining image with another as a naturalistic image of some state of objects in space (or event involving them) — may require some special mechanisms to handle the grammatical subject differently from the predicate, but it surely requires only a minor modification of primate imagination. Chimps that learn to associate words with certain kinds of objects spontaneously combine different words for the same object as a single new word, and thus, it is just a matter of evolving a behavioral schema that keeps the covert locomotion and manipulations straight. The key to the use of language is, therefore, the mechanism that makes the verbal side of linguistic behavior indicate how the images associated with the words are to be combined as a naturalistic image representing some naturalistic state of affairs.

Such naturalistic images are already a higher level of neurological organization on the nonverbal side of linguistic behavior, because it is a new way of “manipulating” object images. Instead of covert manipulations that call up sequences of telesensory images representing the effects of rotating the object, bending it, or putting objects with different geometrical structures together, covert linguistic manipulation combines different images as representations of the same object in the local image at the same time.

The nonverbal side of linguistic behavior may start with one object image for the object in the scene (the grammatical subject) and superimpose another object image (for a kind of object or way of behaving) on the same object in the scene. The result would depend on the object images, but since both meanings are various sequences of telesensory images, it would involve a sequence of telesensory images that depends on some sequence of images from both meanings. That sequence of images would be a naturalistic image, the potential meaning of a sentence or what is called a “naturalistic proposition.”

The naturalistic image could then be used as the grammatical subject to which further object images are added in a similar way (compounding predication within a sentence), or it could be joined in the local image by object images representing other objects in the local scene (conjoining different sentences as a description of the scene).

Such nonverbal covert behavior would require linguistic behavioral schemata with a higher level of part-whole complexity, but that is not all that is required by this higher level of neurological organization, because verbal behavior must also represent how the object images are to be combined. That is, verbal behavior must indicate two things at once, both the images involved and how they are to be combined as naturalistic images. Words for all the images involved must be uttered as part of a single speech act, called a sentence, and the kind of covert manipulation involved must modify the words in some distinctive way, by their order, inflection, prefixes, suffixes, or the like in the sentence.

The capacity to generate verbal behavior of this kind requires a new brain mechanism that does two things: it must connect each image (or “meaning”) involved in the nonverbal linguistic act with its name, and it must register which images are subjects and which are predicates in its construction of the naturalistic image(s) (or “meaning of sentence’). The output of such a center would be able to determine motor commands for verbal behavior that would indicate both aspects of the nonverbal side of the linguistic act.

Understanding: verbal behavior determining (effects of) nonverbal behavior. Linguistic interaction requires listeners who can understand the verbal behavior in the sense of constructing the same naturalistic images in their own faculties of imagination. This would require the new neural mechanism for verbal behavior to work in reverse. That is, instead of using input about the images and covert manipulation of them to generate verbal behavior for words with grammatical markers, it would have to use telesensory input of words and grammatical markers uttered by other to call up the images and combine them in imagination in the indicated way. That would be to construct the speaker’s naturalistic image, or proposition, in the listener’s faculty of imagination.

The structure of the linguistic brain. The evolution of the use of language involves changes that are located in only one of the two hemispheres of the forebrain. It is usually the right hemisphere that continues to serve spatial and structural imagination, which means that bodily behavior, including locomotion and manipulation, can continue to be generated with at least as much skill as primates. The left hemisphere evolves a higher level of neurological organization which is dedicated to linguistic behavior, both the verbal and nonverbal sides. Nonverbal linguistic behavior is a form of covert behavior that manipulates multiple local and object images to construct naturalistic images. They are the meanings of natural sentences, and there are special mechanisms in the left hemisphere for representing this nonverbal behavior verbally (labeled “Wernicke’s area” and “Broca’s area” in the functional diagram of the linguistic animal behavior guidance system). How this higher level is related to spatial and structural imagination is suggested by the functional diagram of subsystems in the linguistic animal behavior guidance system. 

This diagram suppresses the arrows between the sensory input and behavioral output systems representing the basic connections that stem from the telesensory animal behavior guidance system (perception and the input to the local image about the current bodily condition). They are crucial to the mechanism for speaking sentences and hearing them as sounds, but they are not crucial to generating linguistic acts or understanding their meanings.

The higher level of neurological organization that constitutes naturalistic imagination occurs in both the behavioral output system and the sensory input system, repeating the kinds of changes that occurred in the evolution of structural from spatial imagination. But in this case, the new structure does not occur within the local image. Nor does it occur within the object image. It is, rather, a structure that includes both of them as parts, namely, the naturalistic image, which is the meaning of the natural sentence. (See the functional diagram of the linguistic animal behavior guidance system.)

In the behavioral output system, the higher level of organization is a kind of behavioral schemata used by the behavior generator that can linguistically manipulate images of all kind, including both local images and object images, in order to construct a compound representation in the sensory input system, the naturalistic image.

The naturalistic image is located in the sensory input system (mainly in a region of the inferior parietal lobe). The covert nonverbal linguistic behavior acts on this region in the same way as spatial and structural imagination (by way of the projection from the anterior to the posterior cingulate gyrus and the latter’s connections with the adjoining areas of non-cingulate neocortex).

Once again, the linguistic behavioral schemata used by the behavior guidance system are actually structure contained in the frontal neocortex, and it has access to them by way of a complete circuit (through the neocortex, corpus striatum and ventral anterior thalamus) which is not depicted in this diagram. This is at least one place where the structures internalized by the evolution of behavior schemata by reinforcement selection are contained.

From the function of language and the nature of the primate faculty of imagination, we have inferred the brain systems that are required for the kind of linguistic act that would enable a leader to distribute a plan of social level behavior to the members of a nomadic band. In addition to the higher level of neurological organization in the faculty of imagination, which make it possible for linguistic behavioral schemata to manipulate images of various kinds as the meanings of words used in constructing the meanings of natural sentences, the use of language requires changes in the nervous system set up by embryological development in order to generate the verbal side of linguistic acts and to recognize their meanings from telesensory images of another’s verbal behavior. They are the regions of the neocortex called Wernicke’s area and Broca’s area in the linguistic hemisphere of the brain.

Wernicke’s area has two different functions, translating the verbal behavior of other speakers that is registered as telesensory input into the naturalistic images that are their meanings, and translating the naturalistic images constructed by the speaker’s nonverbal linguistic behavior into words and grammatical markers so that they can be generated as the verbal side of linguistic behavior.

There is a major connection between Wernicke’s and Broca’s areas not represented in this functional diagram by which Wernicke’s area identifies the words and grammatical markers to be generated so that Broca’s area can, as part of the complete linguistic act, generate the motor commands for speaking the natural sentence.

The connections between Wernicke’s and Broca’s areas, as well as between them and the faculty of naturalistic imagination, can be seen better in the diagram of the linguistic brain. This diagram is looking down on the two hemispheres of the brain, with the linguistic mechanisms located in the left brain. The blue lines depict the causal connections involved in speaking, whereas the red lines in the left hemisphere depict the causal connections involved in listening and understanding sentences spoken by others.

Wernicke’s and Broca’s areas in the left hemisphere are known to play a central role in language, and the roles they have, as indicated by the effects of damage to each, suggest that they work together to serve the function required by this explanation of how language is possible. The verbal aspects of linguistic behavior are handled by Wernicke’s and Broca’s areas of the left (linguistic) hemisphere.

Wernicke’s area is located in a posterior region of the superior temporal neocortex, and damage to this area is known to make patients unable to connect words and their meanings. Given the meanings (or object images), such patients cannot call up their names, and given the words, they cannot identify their meanings.

Broca’s area is located in the frontal neocortex just anterior to the motor neocortex sending motor commands to the mouth, larynx, tongue, and lips, and damage to Broca’s area is known to make patients unable to arrange words with their grammatical markers as well formed sentences, though they continue to be able to understand sentences and suffer only limited loss of ability to produce the words for object images (meanings). 

These two areas are connected by the arcuate fasciculus, a large bundle of association fibers that also connections the inferior parietal neocortex to the frontal cortex. It apparently transfers information from Wernicke’s area to Broca’s area, because the effect of cutting it is basically the same as damage to Broca’s area.

Speaking. Wernicke’s area has the function of connecting the verbal and nonverbal sides of linguistic acts. The speech act is generated by a linguistic behavioral schema in the frontal neocortex (as part of the behavior generator). It generates covert nonverbal manipulation in the anterior cingulate gyrus, and (by way of its projection to the posterior cingulate gyrus and the latter’s connections with the inferior parietal area), it constructs a naturalistic image in the sensory input system. The inferior parietal area, where the naturalistic image is constructed, has two way connections with Wernicke’s area, and so it can informs Wernicke’s area about how the naturalistic image is being constructed in naturalistic imagination. Wernicke’s area not only connects the various images with their names, but also registers how the words are being related to one another as grammatical subjects and predicates in each naturalistic image. Wernicke’s projection to Broca’s area provides all the information needed to generate the motor commands for verbal behavior, using the words with proper grammatical markers. The projection from Broca’s area to the part of the overt body image sending motor commands to the mouth, etc., would enable it to generate overt verbal behavior directly. Verbal behavior is, therefore, a result of an act of nonverbal imagination, though the behavior generator (with its linguistic schemata) in the frontal cortex could help Broca’s determine the actual surface grammar used in verbal behavior, when there are alternative ways of marking the words grammatically.

Understanding. Wernicke’s area also connects the verbal and nonverbal sides of the linguistic act in the listener’s brain. The location of Wernicke’s area next to the auditory analysis region of neocortex in the superior temporal lobe makes it easy to receive input about which words and which grammatical markers are identified in the auditory telesensory images caused by another speaker’s verbal behavior, and thus, with its two way connections to the inferior parietal neocortex, Wernicke’s area plays the role of the posterior cingulate gyrus in constructing the naturalistic image. That is, it acts as covert nonverbal behavior, except that it is reaching into the brain to control imagination by way of the public linguistic interaction. The meaning of the sentence is the naturalistic image, which is projected to the behavioral output system (by way of the caudate nucleus, globus pallidus, and the ventral anterior nucleus of the thalamus), so that it becomes part of the local scene in which the subject responds.

Since Broca’s area is not involved in understanding the meaning of a sentence (the proposition), that would explain why damage to Broca’s area does not affect language comprehension. The ability of such patients to name objects (that is, to generate the right word in for each object) is affected, but not completely impaired by damage to Broca’s area. On this account, naming could be explained as repeating sounds heard in auditory imagination. Wernicke’s area would still connect the image and with the word, and so a reverse connection back to the auditory area would provide a telesensory image of the spoken word, enabling the patient to repeat the word without any help from Broca’s area..

Understanding a sentence heard would not require any covert nonverbal manipulation. (But it would be possible for the listener to reconstruct the naturalistic image in his imagination by repeating the sentence to himself, since the behavioral schema for the nonverbal manipulation to generate the sentence would probably be connected in memory with the verbal behavior in speaking it.)

Though understanding linguistic representations does not necessarily involve any covert nonverbal manipulation, language learning still depends on acquiring linguistic behavior schemata for speaking, including covert nonverbal manipulation, because the connections in Wernicke’s area between object images and words and between linguistic manipulations of them and grammatical markers are acquired as part of those linguistic behavior schemata (along with the capacity to generate the right motor commands for the verbal side of the speech act). Without learning to speak, memory might not even contain the images that are to be combined in the naturalistic image.

Understanding sentences would require additional learning, because Wernicke’s area must be able to use telesensory images of the words and grammatical markers used by others to call up the images and control how they are combined in the naturalistic image. However, such learning would be all but automatic, because in learning to generate the verbal behavior, one also hears it, and two way connections between Wernicke’s area and the local and object images would be established by the memory circuit.

The higher level of neurological organization required for language does not come, therefore, from the bilateral specialization of the two hemispheres of the brain. The higher level occurs in only one hemisphere of the brain. The higher level is the higher level of behavior schemata required for linguistic manipulations of object images and the higher level of organization in the naturalistic image itself, though perhaps the addition of Wernicke’s and Broca’s area for representing such covert nonverbal linguistic behavior as verbal behavior might also be considered evidence of a higher level of organization.

The specialization of one hemisphere for language has other functions. It leaves one hemisphere of the brain unencumbered with the mechanisms required for speaking. That also makes it easier for animals to talk and do other things at the same time, even though both activities require the involvement of imagination.

Furthermore, special desires must be attached to linguistic representations in order for the desire to submit to a leader to cause member to do what is described by the leader’s verbal behavior. Thus, if those desires affected the goal selection of the linguistic hemisphere, behavioral schemata for such social behavior could evolve in that hemisphere, while behavioral schemata for satisfying other desires could continue to evolve in the other hemisphere. This requires, of course, that the linguistic hemisphere be the dominant hemisphere, for otherwise the use of language could not reliable control behavior.

There are, of course, many details about how neural connections are made in order to serve these general functions, but this evidence about the structure of the human brain is sufficient for our purposes. I have described a kind of higher level of organization in brain mechanisms that would give primates the use of a primitive language of natural sentences, and since it is within the range of random variations tried out by hominid brains, I conclude that the evolution of language using primates is inevitable. As we expect from this ontological argument, there are such mechanisms in human brains. Learning a natural language can be explained as a contained form of reproductive causation, that is, a neurological development in which linguistic behavioral schemata evolve by reinforcement selection toward maximum holist power, controlling relevant conditions in all the situations where linguistic interaction are useful at all.

Naturalistic imagination. With the evolution of the capacity to use natural sentences, subjective animal evolve a faculty of naturalistic imagination, and like each earlier stage in the evolution of imagination, it gives the animal the conception of something in the world that it could not previously think about explicitly. It is the conception of a state of affairs in the natural world, that is, the state of object in space and how they change over time. With naturalistic imagination, subjects always think about actual states of affairs against the background of what other states of affairs are possible. The possible in this case is the kind of events that can cause or be caused by the state of affairs, and so it gives them an understanding of efficient causation.

Naturalistic imagination does not replace earlier forms of imagination, but is, rather, based on them and incorporates what they understand as part of it own understanding.

Animal behavior necessarily acts on objects in space, and with the evolution of telesensory animals, there is already a conception of the object, because there is an explicit representation of the object in the sensory input system of its animal behavior guidance system. Though telesensory animals are hardwired to adapt their animal behavior to the locations of objects, those objects are not explicitly represented as being located in space.

In subjective animals, objects are represented as being located in space. Spatial imagination gives them a conception of the structure of space, and so they can think about the spatial relations of objects in terms of the effects of motion on them. This is an intuitive understanding of spatial causation. Subjective animals always see the actual spatial relations among objects against the background of the other spatial relations that are possible as a result of motion. Though geometrical structures are included in their telesensory images of objects in space, the objects in space are not explicitly represented as having geometrical structures.

In manipulative animals, objects in space are represented as having geometrical structures. Structural imagination gives them a conception of geometrical structures, and thus, they can think about the geometrical structures of objects in terms of the effects of manipulation on them. This is an intuitive understanding of structural causation. Manipulative animals always see the actual geometrical structures of objects in space against the background of the other geometrical structures that are possible as a result of manipulation. Though states of affairs are included in their representations of the world, those states of affairs are not explicitly represented as states of affairs.

In linguistic animals, the natural world is represented as having states of affairs. Naturalistic imagination gives them a conception of states of objects in space, either as events taking place or conditions that simply hold at some time, and thus, they can think about states of affairs in terms of their effects and their causes. This is an intuitive understanding of efficient causation. Though the regularities behind the most basic causal connections are built into naturalistic imagination by how it is based on spatial and structural imagination, other regularities can be internalized from experience and used in a similar way, for example, how objects are changed by fire, plants grow or can be killed, how animals behave in certain situations, and even how their own children become adults. Linguistic animals can always see the actual states of affairs against the background of the other states of affairs that are possible because of how other states of affairs might affect them or work together with them to produce effects. This is the kind of understanding that is involved in sharing a plan of social level behavior, but it can be used to explain how other states of affairs are caused or to predict the effects they will have.

Truth and logic. Natural sentences can exist as public objects, namely, as overt verbal behavior. (It is just speaking at the primitive spiritual stage of evolution, though after the advent of writing at the next spiritual stage of evolution, sentences can also exist as permanent results of overt verbal behavior.) Such public objects correspond to states of affairs in the natural world. When the states to which they correspond actually exist, the natural sentences are true, and when the states to not exist, the natural sentences are false. But this relationship of correspondence between natural sentences and states of affairs in the world is not direct. It is mediated by the faculty of naturalistic imagination.

This theory about the nature of truth for natural sentences is one of the necessary truths of ontological philosophy. It follows from the inevitability of the evolution of the linguistic level of neurological organization, because the correspondence between sentences and world is mediated by the faculty of naturalistic imagination. The verbal side of linguistic representations is connected to the nonverbal side, as we have seen, by Wernicke’s area. But that means that words acquire their references to objects or kinds of objects by way of their meanings, that is, by way of the images that they name in naturalistic imagination (of all kinds, local, object, telesensory and tactile). Naturalistic images represent states of objects in space as states of objects in space because they are constructed in naturalistic imagination by superimposing one image on another as a representation of something about objects in the local scene. That act of imagination explains how the grammatical structure of a sentence is relevant to its truth.

Logical entailments among sentences are explained by their meanings. One sentence is entailed by another just in case its naturalistic image is already constructed in the process of constructing the naturalistic image of the other. Logical connectives, such as “and,” “or,” and “not” would be introduced to make clear in verbal behavior how naturalistic images are to be combined as parts of a plan or description of the natural world. And the gradual evolution of the capacity to use a language of natural sentences would make it possible to include many predications in a single sentence (by the evolution of more versatile grammatical markers that make it possible for members to learn how to use phrases, clauses, modifiers, and the like).

“If . . .then . . “ might also evolve at this stage. But it would probably not be used to represent logical entailments between sentences, since they would be too obvious. Rather, “if . . . then . . .” would be used to represent the causal relations between states of affairs that could be understood with a faculty of naturalistic imagination. Their understanding of efficient cause relations would include not only all the basic regularities about change of location by motion and constraints on change imposed by the geometrical structures of objects in space, but also all the regularities that they have observed in the natural world, and it would be useful at time to share a mutual recognition of them. Such shared understanding of causal connections is crucial to sharing a plan of social level behavior.

Meaning and reference. This explanation of the evolution of language clarifies something about the ontological assumptions on which this philosophical argument is based. In postulating substances as causes in explaining the world ontologically, we assumed that certain aspects are constituted by the existence of substances, such as their existence and essence. This was to explain properties ontologically, without having to postulate properties as existing in addition to substances. But I did not explain how specific aspects of substances could be picked out or distinguished from one another. I simply used language to refer to those aspects and talked about them as properties of substances. The missing explanation of that ability, or at least, the foundation for it, is provided by this explanation of the evolution of language. It shows how the semantic relation is mediated by images in the faculty of naturalistic imagination.

Each word has a meaning as well as a referent, and its meaning is responsible for its reference to an object or a kind of objects in the world. Its meaning is the image that is connected to it by Wernicke’s area, and it is because that image corresponds by way of the primate faculty of imagination to an object, location, kind of object, or relation in the world that the word refers to objects, properties and relations. Likewise the meaning of a whole sentence is the naturalistic image that is connected to the natural sentence by Wernicke’s area, and it is because the naturalistic image corresponds (by way of the meanings of its constituent words and their grammatical relations in the faculty of imagination to a state of affairs in the natural world) that the natural sentence refers to such a state of affairs in the natural world.

What has been explained by ontological philosophy are the substances and aspects of substances to which words and sentences refer. Particular objects are all constituted by space and matter in some way, and particular substances fall into kinds because they have aspects of the same kinds. Those aspects of particular substances have all been explained as aspects entailed by the existence of two opposite kinds of basic substances, space and matter, and how they exist together as the objects in question. But that ontological explanation depends on being able to see that the basic substances have aspects, including their relations as parts of the same world, and how they constitute objects with the properties in question.

The reason it is possible for us to understand such an ontologically deep explanation of the world is that language acquires its relationship to the world by way of the a faculty of imagination based on the primate animal system of representation, which represents the most basic aspects of the world, its spatial and spatio-temporal structure. Spatial imagination represents the world in a way that makes it possible, in principle, to keep track of objects by their locations in space over time. The location of an object in the local image is a location in the entire world, and thus, with the evolution of language, that makes it possible to refer to any particular object unambiguously by its location in space at some time. Local images also make it possible to refer to spatial relations among objects, and how sequences of images over time represent change over time makes it possible to refer to temporal relations. The object images added by structural imagination represent the geometrical structures of objects in space, and thus, words can refer to some of the most basic properties of substances in a spatiomaterial world, along with whatever properties are represented in telesensory (and tactile, including taste) images of them. Naturalistic images constructed out of such images in naturalistic imagination can correspond to states of affairs, which are aspects of the world (that is, substances having properties). Naturalistic images with a kind of object as grammatical subject correspond to complex properties (that is, properties having properties, or aspects of aspects). But it is the meanings provided by a faculty of imagination whose very structure represents the basic, spatial nature of the world that explains what words and sentences refer to.