Rational causation. The remaining problems about the nature of causation arise in the branches of science known as psychology and social science.

Psychology has to do with the explanation of individual behavior, and that is problematic mainly because we know too much. As rational beings, we have a special way of seeing into the minds of other rational beings (and subjective animals generally). We ordinarily explain individual behavior by the reasons that the individual has for it, that is, the beliefs, intentions, desires and the like that are responsible for it, or subjectivistic understanding, as we have been calling it. There are two problems,

One problem in this field is that rational explanations do not seem to be the kind of explanation that a branch of natural science ought to be seeking.

But another problem is that, even if they are, they do not seem to be reducible to the kinds of explanation given in physics.

The social science have to do with the explanation of social phenomena, or what has been explained here as the behavior of spiritual animals. We know that human societies are different from other groups of animals, because our capacity for subjectivistic understanding gives us an “inside view,” so to speak, of the phenomena. However, that view is not based on perception and, thus, is not from the vantage of natural science. Thus, there is a problem about the nature of the object that is being studied by the social sciences.

The problem about reductionism in this case is just opposite to the other cases considered here. Though there have been social scientists, like Comte and Durkheim, who thought that societies are not reducible to the individuals, that view is not common these days. Contemporary naturalists tend to assume that social phenomena must somehow be explained in terms of the individuals who make up human societies, because they do not see how there could be any relevant causes that arise from the nature of society as a whole.

The main philosophical problems about the nature of causation in social science has to do, therefore, with showing how social phenomena can be explained as a result of the nature of the individuals, the regularities in their behavior, and the situations in which they act. The project of explaining social phenomena in that way is called “methodological individualism,” and its most popular current form is sociobiology, which bypasses individual psychology and tries to explain social behavior by genes that have evolved in individuals.

The ontological explanation of the nature of change provides, however, a solution in all of these cases. Though the laws of nature (or regularities) discovered in psychology and social science may not be reducible to the laws of physics, they are reducible to the ontological causes recognized by spatiomaterialism in a world like ours. Once again, the reason is the failure to recognize that the global regularities are caused ontologically by the wholeness of space and other substances contained by it, both basic and derivative, like material structures and reproductive cycles. Indeed, all the basic phenomena investigated by both psychology and social science have already been explained in tracing the course of evolution by reproductive causation. What follows here is just a reminder of their relevance.

Psychology. In the first instance, psychology is based on our ordinary way of understanding human beings. That is to explain individual behavior and beliefs by the reason which are responsible for it, or what I have been calling “rational explanation.” For decades now, it has been is called “folk psychology” in epistemological philosophy of science, because it is generally assumed that such explanations depend on learning the relevant “laws of nature” as a normal part of the process of growing up in human society. But it has been explained here as subjectivistic understanding.

Subjectivistic understanding is part of the cognitive capacity I have been calling “reason,” for it is the use of rational imagination to think about the causes of beliefs and behavior in subjective animals like us. Reason has been explained here as a capacity that derives from the use of psychological sentences, for that is what enables the subject to represent and, thus, reflect on the psychological states that are involved causally in the process by which their animal behavior guidance system. That is the basis of the subject’s capacity to use the theoretical and practical reasoning that takes place in his own brain to simulate the reasoning going on in the brains of others, and thus, it is what enables the subject to see into the minds of other subjects.

Naturalistic understanding is another part of the capacity of reason. It is the use of rational imagination to think about the causes and effects of states of objects in space, or the kind of imagination that first evolved in primitive spiritual animals, which had only the use of natural sentences (with a subject-predicate grammar). The use of natural sentences gives the subject the concept of a state of affairs (or event) in nature, and since reason uses a faculty of imagination that is built on the spatio-temporal imagination of mammals and the structuro-temporal imagination  of primates, it involves the ability to understand efficient causes and their effects (both those that depend on these basic aspects of the spatial structure of the world and those that are learned from experience of other regularities in the natural world).

Hermeneutics. By “hermeneutics,” I mean the belief that the best that science can do in the way of explaining individual beliefs and behavior is to give rational explanations.

This view is now most commonly defended in the philosophy of social science. There seems to be no hope explaining social phenomena unless the beliefs and behavior of individual can be explained. Even the gathering of statistics about individuals, as in economics and sociology, depends on being able to start with the ordinary explanations of their beliefs and behavior. Thus, those who are eager to have the social sciences recognized as a form of genuine knowledge about the world seem forced to accept a hermeneutical understanding of individual behavior.

Hermeneutics is also the foundation of most social psychology and clinical psychology for the same reason. But in psychology, there are attempts to give a deeper explanation of individual behavior, which would make it clear that psychology is a branch of natural science and, thus, no less entitled to claim that its conclusions are science. They will be considered next.

The main problem with simply accepting rational explanations as scientific explanations is that the empirical method does not lead to general agreement about what is true, at least not in a way that is comparable to using the empirical method with efficient-cause explanations.

This problem with the empirical method was discussed when the empirical method was introduced (in Method). The empirical method is the attempt to discover what is true by inferring to the best explanation of what is observable in the natural world, and as we noted, it is a method that can, in principle, be used in conjunction with various kinds of explanation: efficient-cause explanations, rational-cause explanations and ontological-cause explanations. The way that it leads to agreement in the case of efficient-cause explanations has made natural science a spectacular success in the attempt to discover the true. Its use in conjunction with ontological-cause explanations is the foundation of ontological philosophy, where it may also lead to general agreement, this time about the basic substance constituting the natural world. But in the case of rational-cause explanations, it fails to lead to agreement about what is true. Different rational subjects trying to explain the same behavior (or the same beliefs) of some individual often wind up with different conclusions, and no matter how much they consider one another’s rational explanations, there does not seem to be any way for them to reach agreement.

The problem about reaching agreement on rational-cause explanations is sometimes called the hermeneutical circle, because the attempt to resolve disputes about what an individual intends or believes in a particular case depends inferring to the best rational cause explanation. Since one standard of the best explanation is explaining the widest range of phenomena, the widest range in this case is the range of the individual’s behavior. But for other instances of the individual’s behavior to be relevant in judging which explanation is best, they must also be explained rationally, and thus, the same problem arises about explaining them. The rational explanation of one instance of behavior depends on the rational explanation of the other, and that instance on yet another, so that in the end, all the behavior has to be interpreted. The rational explanation of the part thus depends on the rational explanation of the whole, and as it happens, even when all relevant behavior is included, there are still differences among the subjective scientists.

The reason for these disputes can be explained, as we did earlier, by the nature of rational explanation. It comes down to disagreements among the subjective scientists themselves in basic their beliefs about the world, especially their most basic and general beliefs, such as moral and religious beliefs. An inference to the best rational explanation is an inference to the fewest and simplest psychological states that will explain the widest range of behavior, but it depends on a judgment about which alternative explanation is the most coherent, that is, rational selection. And since rational explanation involves using one’s own process of practical and theoretical reasoning to simulate the reasoning of others, the judgment about which alternative set of psychological states is simplest and fewest depends on using one’s own desires and beliefs (including beliefs about what is good) as the background in which they are compared. Since that background varies from one subjective scientist to the next, subjective scientists tend to disagree about which is the best rational explanation.

Inferences to the best efficient-cause explanations are not subject to this kind of dispute, because naturalistic understanding involves only beliefs about the natural world which are ultimately based on perception. No judgments about what is good and bad, or what is meaningful, or how one feels is relevant in natural science. But they are the stuff of the subjective sciences.

Insofar as such disagreements about the best rational-cause explanation are not resolvable, it is apparent that the conclusions of subjective science are not objective. The ontological explanation of the nature of reason shows that there is a good deal of validity in rational explanations, because the animal behavior guidance systems of rational subjects do work in basically the same way. Thus, to some extent, they can be used to discover the true, though the range in which they are trustworthy may be limited to more immediate intentions is rather well defined social situations. However, rational explanations will lead to much greater agreement about the reasons behind individual behavior when ontological philosophy evolves in philosophical spiritual animals, because there will be a great deal more agreement about background beliefs and values.

However, a genuine science of individual needs more than rational explanation, because psychology must be integrated as a branch of natural science. Thus,  naturalists are on the right track in attempting to reduce rational cause explanations to the kind that is used in natural science.

Naturalism. There have been various attempts to reduce rational-cause explanations to efficient-cause explanations, and as a way of showing the relevance of the ontological explanation of the nature of change to issues about causation, let me mention the main varieties here.

Behaviorism. The original attempt to turn psychology into science is behaviorism, that is, the attempt to discover a law of nature describing the regularities about individual behavior so that it would be possible, in principle, to explain particular actions by efficient causes. These first attempts tried to reduce behavior to what is now called “respondent conditioning,” exemplified by Pavlov’s dog, in which behavior that is already triggered by some stimulus is conditioned so that it comes to be triggered by another stimulus. It was followed by the theory of operant conditioning, developed mainly by B. F. Skinner. Operant conditioning is based on the law of effect. When kinds of behavior that are generated spontaneously or randomly are reinforced, they are more likely to generated again, especially under similar stimulus conditions.

Functionalism. Behaviorism has been replaced in psychology by cognitive psychology. It departs from its predecessor by recognizing that behavior is mediated by internal states, and thus, it takes the project of psychology to be to discover the internal states that are responsible. But cognitive psychology does not attempt to discover the physical properties of internal states. Instead, it attempt to discover them in terms of their causal connections to input states and output states of the organism. That leads to what is called “functionalism.”

Neurophysiology. The other thriving trend in psychology is the attempt to reduce rational explanations to neurophysiology, that is, to the states of the brain. (Though brain states may still be defined functionally, the functions are physiological functions, and thus, involve descriptions that are more closely tied to physics.)

The significance of ontological philosophy for each of these projects is implicit in what has been said in tracing the course of evolution as a global regularity caused by reproductive cycles and the wholeness of space.

Neurophysiology. The problems of neurophysiology have been addressed by this ontological explanation of the course of evolution by tracing the stages of animal evolution from somatosensory through manipulative animals to rational subjects (stages 4-9). The nervous system was explained as an animal behavior guidance system, but the biggest departure from received neurophysiology comes from the recognition of levels of neurological organization and what each contributes to the animal system of representation. That functional explanation shows how structures in the nervous system serve as a faculty of imagination, that is, a mechanism in which covert behavior calls up sequences of images from memory in the sensory input system to represent the effects of motion on the relations of objects in space, of manipulation on the geometrical structures of objects in space, the causal relations among states of objects in space, and the causal relations among psychological states.

This is a different kind of neurophysiological explanation of behavior than is expected by the current defenders of neurophysiology, such as Paul and Patricia Churchland, for they are eliminative materialists, who expect rational explanations (or “folk psychology,” as they call it) to be replaced by neurophysiology. By contrast, this explanation of how the brain works explains the validity of rational explanation by showing not only how they are valid explanations, but also by explaining how it is possible for rational subject to give such explanations of beliefs and behavior.

As a functional explanation of those structures in the brain, however, it leaves a great deal yet to be explained. Indeed, all the detailed mechanisms that are required to serve these functions remain to be explained. But those nervous mechanisms are quickly yielding to the astonishing progress of empirical neurophysiology. Since they are coming at it from opposite directions, what ontological philosophy implies and what empirical neurophysiology is disclosing should converge on a single, complete explanation of how the brain works before long.

Behaviorism. What made it possible to explain the stages of neurological organization by reproductive causation was the recognition that the faculty of imagination does not require the mechanism of embryological development (that is, the multicellular biological behavior guidance system) to provide the detailed structure of the brain. It needs to provide only the basic systems of the faculty of imagination, because its structure makes possible a contained form of reproductive causation in which the behavioral schemata behind covert behavior can evolved by reinforcement selection. That is, given that there are random variations on behavioral schemata, the learning of new ways of behaving and thinking can be explained by a memory circuit that strengthens the synapses of neurons involved in generating behavior of that kind when they are successful by genetically determined criteria (such as success in getting around in space or success in social relations mediated by linguistic behavior). Thus, the brain has a built-in structure that internalizes structures of the world, from the spatial structure of the natural world to language and the capacity for reflection.

This is an explanation of the validity of operant conditioning, at least, in mammals and beyond. The law of effect is true, on this functional explanation of the faculty of imagination, because the regularity it describes is the evolution of behavioral schemata by reinforcement selection within the mammalian brain. (The memory circuit works in a similar, but far more limited way in non-mammalian vertebrates, and thus, the learning in pigeons was limited enough to stand out in the kinds of experiments that Skinner conducted.)

But this neurophysiological explanation of operant conditioning reveals that it is not as open ended and unstructured as Skinner believed, because it is the evolution of behavior schemata that operate as various faculties of imagination (spatio-temporal, structuro-temporal, naturalistic and subjectivist imagination). That is, behind the overt operant behavior, including verbal behavior) is a covert operant that calls up sequences of images of a certain kind, and thus, from the point of view of the subject, the behavior is generated in world from an understanding of the world that sees the actual against the background of the possible.

Functionalism. The neurophysiological structures in the nervous system have been explained by the functions of various systems at a series of level of neurological organization. That is a functional explanation in the strong sense that is entailed by reproductive causation and the recognition that evolution is progressive, increasingly sophisticated ways of serving as an animal system of representation are what causes each higher level of neurological organization. And these functional explanations of the levels of neurological organization include functional explanations of various nervous structures in the brain, such as the behavior generator, the local image, the object image, and the like.

These functionally described states are not quite what cognitive psychology is looking for. In the first place, they are tied certain neurophysiological structures in the brain, and thus, internal states are no explained exclusively in terms of their causal connections to (sensory) input and (behavioral) output. Secondly, the functions that are ascribed to internal states are not merely that of representing aspects of the world, but as representing objects, representing them as being located in space, as having geometrical structures, as being efficient causes, and even as having reasons. And the functions of such states depend on them being parts of a faculty of imagination.

This departure from received functionalism in psychology solves the problems that have been encountered, and by considering them more closely, those who are interested can see how.

Intentionality. The philosophical problems about the nature of mind arise from certain aspects that seem to be incapable of explanation by the basic laws of physics. One of those problems is consciousness, or the subjective aspect of experience, such as the phenomena appearance of the natural world in perception. The foundations for the ontological explanation of consciousness were discovered in Properties, and the way in which it explains the unity of mind, or the fact that many qualia appear to the subject at the same time, was explained as part of the discussion of the mammalian brain (Stage 6). The other main problem, which will be discussed here, has to do with intentionality. ^[1]

The problem about intentionality is how there can be psychological states that are about the world. We know there are psychological states about the world, because they are what we use to give rational explanations of behavior (and beliefs) of rational subjects (and other subjective animals). Though the mind obviously depends in some way on the nature of the brain, it does not seem that that the aboutness of psychological states can be explained by the basic laws of physics. Functionalists believe, however, that they can be explained as functional states. No one denies that it is plausible to suppose that the intentionality of psychological states involves a system of representation built into the brain. But how can states of the brain be representations? How can they be about the world?

Intentionality cannot be explained as something we read into the phenomena, as if it were just a useful way of describing or summing up what happens in nature.^[2] That would be to deny the reality of the phenomenon, at least as part of the natural world. And it is hard to see how even that is possible without contradicting oneself, because no one who holds that we are reading things into nature (or describing them in certain ways) can deny there are intentional states in the world. Those very interpretations are about objects in the natural world. The only way to avoid self-contradiction, therefore, is to hold that one’s own mental states are not part of the natural world, and that is, ontologically speaking, a form of mind-body dualism. It implies that there are two basically different kinds of substances in the world: natural entities without real intentional states, and beings like us, who must have them, since we do refer to other objects and ascribe intentional states to them. This is a disastrous kind of dualism, for there is no way to explain how substances whose natures differ as mind and body are related to one another as a single world. And even if there were, it would be to give up naturalism and, thus, ontological philosophy.^[3]

Though it is generally agreed among naturalists that intentionality is to be explained functionally, there is little agreement about what such a functional explanation would involve. There are two main schools of thought about the nature of "functionalist theories" of psychological states, and both would explain intentionality in terms of representations in the brain. One theory holds that the most science can do is give functional descriptions of the brain. The other holds that natural science can give functional explanations of the brain, although it is based on an analysis of functional explanations (the etiological theory) that precludes their reduction to the ontology of naturalism. A brief account of these theories will provide a sense of the obstacles that intentionality poses for a naturalistic metaphysics.

Intentional states as functional states. The still dominant view of psychological states is called "functionalism,” a philosophy of psychology inspired by the analogy between minds and computers a quarter century ago.^[4] The idea is that psychological states can be understood as internal states in a complex system whose kinds can be distinguished in terms of the causal roles those states play in mediating between input and output, much as internal states of computers explain its output in response to certain kinds of input because of how internal states are related by the program. Thus, the goal of psychology is supposed to be giving a functional description of the mind/brain, much as one would a computer, that is, by describing a system of interconnected internal states that tells how all possible inputs would affect output. Two points about functionalism of this kind should be noticed,

First, it denies the possibility of reducing functional systems to the kinds of physical processes that realize them. According to the deductive-nomological model of explanation, the reduction of one theory to another depends on establishing a necessary connection between the terms used by one theory and the terms used by the other, and functionalists deny that there is any such type-type identity between functional states and their physical realizations in the brain.^[5] That is, the functional properties of a system are thought to "supervene" on its physical properties. One of the deepest convictions functionalists have is that, just as physically different kind of computers can perform the same computations, so physically different kinds of brains or brain states can realize the same psychological states. Functionalists are quick to point out that they are not denying materialism (or physicalism). They need not believe in the existence of anything but entities of the kind mentioned by the basic laws of physics. They admit that functionally defined states are identical to the physically defined states that realize them in each specific case. The agree that if a physical system of some kind realizes a functional system, then another physical system of the same (relevant) kind must also realize it. But they believe that there is only a token-token identity between functional and physical properties. They deny there is any necessary connections between the types of these tokens, because they believe that indefinitely many different kinds of physical systems can realize a functional system.

Second, the very form of functionalist psychology precludes any explanation of intentional states in terms of representations of the world. Psychological states are ordinarily classified not only by the propositional attitudes involved (that is, depending on whether they function as beliefs, desires, intentions or the like), but also according to content (or what they are about, beliefs about water, say, being different from beliefs about alcohol). Though the former kinds are plausibly explained by their casual role in mediating between input and output, the latter cannot be, for any correspondence to objects/states in the world would lie outside the functional system. The only way of distinguishing psychological states according to their content within the functional system is by differences in the representations themselves, that is, by the so-called formal aspects of the states (which are analogous to syntax, as opposed to semantics, in linguistic analysis). They have, in the jargon of this field, "narrow content,” but not "wide content.” They cannot have a content that depends on a relationship to objects/states in the rest of the world, because the only relationship of the system's internal states to the rest of the world is by way of its input and output, and functional theories abstract from how input and output connect to the rest of the world. Thus, since functionalist theory cannot connect the mind with real objects/states in the world, it cannot explain the intentionality of psychological states — that is, explain how and why they are about the world. It cannot, for example, say which beliefs are true. It cannot even explain what makes true beliefs true.

The leading proponent of functionalism, Jerry Fodor, argues that these two points are connected. He argues that psychology cannot explain psychological states by how physical states correspond to objects/states in the rest of the world, because functionally described states supervene on physically described states and the physical states on which they supervene are in the brain. This doctrine he calls "individualism.”^[6]

Fodor does not, of course, deny that the internal states of functional systems do sometimes refer to objects/states in the world. But he proposes to account for the "wide-content" of our ordinary psychological explanations by supplementing his functionalist theory of mind with a "causal theory of reference.” The referents would be picked out as certain more or less remote causes of input to the functional system that are regularly related to the internal states. That is supposed to account for the intentionality of psychological states, but even Fodor recognizes that such a causal theory of reference has trouble accounting for some kinds of references.^[7] And there are more basic philosophical objections to such a theory, which Fodor does not acknowledge.^[8] However, neither class of problems is relevant here.

For our purposes, the problem is that, if the intentional content of psychological states can be explained only by tacking a causal theory of reference onto a functionalist theory, then far from explaining intentional states in terms of the ontology of naturalism, functionalist psychology actually makes intentionality more puzzling. Even if all the references we take psychological states to be making did turn out to have causal relations to the world, it would show, at most, that there is an objective regularity about our ascriptions of references to psychological states. But it would not explain why psychological states are about the world. need to tack a causal theory of reference onto a functionalist theory of mind would still suggest that the intentionality of psychological states is something accidental.

What Fodor’s functionalism is leaving out can be seen with the help of our ontological explanation of the function of the animal behavior guidance system. Because animals acquire their free energy by ingesting other objects in space, they need, in addition to their biological behavior guidance system, a system to guide behavior that acts on other objects in space. Thus, animal behavior is different from biological behavior, because it must direct behavior at other objects in space, rather than just at the world as a whole (or merely oriented in a gravitational or electromagnetic field). Thus, what makes animal behavior guidance systems more powerful is the evolution of a subsystem, the animal system of representation, which uses an interaction between sensory input and behavioral output to represent the objects toward which its behavior is directed.

Behavior is generated by the structure of the organism as an irreversible structural global regularity, but as animal behavior, it can make events occur regularly in its territory that are otherwise quite improbable only by acting other objects in the region. That is, what coincides with the geometrical structures of region’s thermo­dynamics flow of matter toward evenly distributed heat to do work is not an unchanging material structure, like a region-wide machine, but rather animal behavior, that is, behavior in which, typically, the animal moves around in the region and acts on other objects (as in chasing prey and ingesting them). But that requires animal behavior to be guided in relation to objects in space, and thus, a system evolves in the animal behavior guidance system  to represent the object, or what we have called the animal system of representation. The animal stages of evolution are all increases in animal power that comes from the animal system of representation representing the nature of the world in which its behavior must act more completely.

The animal system of representation evolves first in telesensory animals. (The somatosensory animal has only an implicit representation of the object, because it uses the location of the sensory input in the body to locate the object for purposes of directing behavior at it, for example, as the hydra’s tentacles sting prey that touch it and contract to draw the prey into its gastrovascular cavity.) Embryological development constructs a nervous system in telesensory animals that uses the regular changes in sensory input as a function of behavioral output to represent the object in such a way that it can guide locomotion in relation to the object. The function of this brain structure depends on how the animal interact with other objects in space, and that is the basis of the relationship of representation between the states of the animal system of representation and the objects in space.

Functionalism abstracts from this functional explanation. To insist that such internal states be defined strictly in terms of the internal causal relations by which they mediate between sensory input and behavioral output is to cut off from consideration all the structural effects outside the body that are involved in doing the non-reproductive work of controlling relevant conditions. The culprit here is the computer analogy, and there are two ways in which it cuts psychological states off from any deeper explanation.

First, on the computer model, the only context that is relevant in a functional system is the input to the system and its output, and thus, functionalism abstracts from the part of the structural effects outside the organism. That cuts the animal behavior guidance system off from any coincidence with the thermo­dynamic flow outside the organism, including any relevant conditions the behavior it is generating might be controlling.^[9]

Second, on the computer model, the internal states of a functional system are defined only in terms of the causal relations among them that are responsible for mediating between input and output, and thus, functionalism also abstracts from the structural global regularities that occur within the animal behavior guidance system. When functionalists abstract from the "physical realization" of the functional system, they are abstracting from the material structures that channel the flow of free energy in the animal behavior guidance system.

This abstraction is necessary, functionalists would insist, because there are different kinds of structural causes that could generate the same kind of structured thermodynamic order. That may be true of computers, but it is not true of biological mechanisms, because in products of reproductive causation, there is a necessary connection between functions and traits. The kind of structural effects that serve any function are determined by that function, because they are the most powerful way of controlling that relevant condition that is possible for organisms of their kind when they evolved. That necessary connection makes a type-type reduction to naturalist ontology possible.

Both kinds of abstraction are appropriate for computers, because their input and output is strictly linguistic (or digital), and many different machines can be built that manipulate the syntax of linguistic or mathematical representations. But animal behavior guidance systems are structural causes that have evolved by reproductive causation to guide behavior in a world of objects in space, not just syntax manipulators designed by human ingenuity to work in a linguistic environment. Given our definition of "functions,” therefore, neither kind of abstraction — neither from the objects in space outside the brain nor from the physical nature of the brain itself — is appropriate.

Functional explanations of intentional states. This brings us to the other received theory of the intentional content of psychological states, the one that would explain representations by their function, rather than just describe them by their causal roles as internal states in a functionalist system.

Ruth Millikan (1989, p 282) rightly challenges Fodor's assumption that the status of an inner state "as a representation is determined by the functional organization of the part of the system that uses it,” pointing out that there is no such a thing "as behaving as a representation without behaving like a representation of anything in particular.” The relationship to objects/states in the world is essential, she insists, to any explanation of intentional states in terms of representations. She is also correct to insist that such a system can be explained functionally, and not merely described functionally. But her theory fails to reduce psychological states to naturalist ontology, because she accepts a theory of functional explanations, the "etiological theory,” that takes accidentalism for granted. And as a result, she overlooks an essential ingredient in any adequate explanation of the nature of psychological states.

Let us call Millikan's kind of explanation the "teleological theory" of representations. It holds that what makes an inner state a representation is that its function is to represent.^[10] According to the etiological analysis, representations are states of an organism that correspond to certain objects/states of the world and that were selected to be parts of the organism because they correspond to those objects/states in the world. That makes the correspondence part of the explanation of the intentional state something more than what happens to be true of it or what we read into it, because the state's correspondence to the world is responsible for the organism having been able to do something that was (and perhaps still is) required for its success in reproduction.

There are, for example, bacteria that use tiny magnets (magnetosomes) to guide their locomotion. What they represent is not, however, the direction of magnetic north, which causes their orientation, but rather the direction of oxygen-free water, because magnetosomes were selected for their correspondence to oxygen-poor water. That correspondence causes their reproductive success by enabling them to avoid the toxic, oxygen-rich water near the surface, and thus, the magnetosomes have the function of representing oxygen free water.^[11]

The teleological theory of psychological states is closer than Fodor's functionalism to the explanation entailed by this ontological explanation of the course of evolution, because instead of tacking a causal theory of reference onto a functional system, it gives a functional explanation of the correspondence between inner representations and objects/states in the world. But the teleological theory of representations nevertheless agrees, in effect, with the other abstraction involved in functionalism, for it still assumes that there is no necessary connection between intentional states and the physical states that realize them. The accidentalist assumptions of the contemporary Darwinist explanation of about the course of evolution lead to the etiological analysis of functional explanations, and since that precludes explaining course of evolution by the functions that are possible, it does not seem possible to explaining psychological states ontologically. Both assumptions of accidentalism are relevant.

First, though inner states of an animal may have the function of representing something, what they represent is contingent. Since natural selection is imposed by changes in the environment, what inner states correspond to depends on environmental changes or conditions that could be different. There may be a historical explanation of the natural selection of intentional states, but since what is represented is contingent, no ontological reduction of psychological states is possible.

Furthermore, even if the selection pressure responsible for psychological states were given and the nature of the correspondence were determined, psychological states would still not be reducible to the ontology of naturalism, because the etiological theory has nothing to say about the mechanisms that would serve that function. The kinds of inner states and how they are made to have the required correspondence would depend on which random variations happened to be available at the time the selection pressure was imposed. Thus, the teleological theory of representations does not offer an account of intentionality that reduces psychological states to the ontology of naturalism.

The examples used to illustrate states with representational functions, such as the magnetosomes in bacteria mentioned above, seem to confirm accidentalism. Though they might guide some bacteria to oxygen free water, they might guide other animals in seasonal migrations. But such examples are misleading, because they implicitly assume that the representational functions of inner states are tied directly to the control of rather specific conditions. And this may be true in somatosensory animals and simpler animals, since they do not have animal systems of representation. And since the accidentalists assumptions of contemporary Darwinism keep teleological theorists from trying to trace the course of evolution, they do not notice that the evolution of greater power in higher animals comes from serving a more universal function in behavior guidance, namely, the representation of objects for the purpose of adapting behavior to the spatial aspects of the world. That is, they overlook the inevitability of the evolution of the animal system of representation in multicellular animals. ^[12] 

The animal system of representation has a necessary neurological structure in telesensory animals because of how behavioral output must be combined with sensory input to locate objects in space for purposes of guiding behavior. There are, of course, different ways of serving this function, as we have seen, with the greatest differences arising from the fundamental difference between proterostome and deuterostome embryological development. But the inevitability of the neurological structure of the system for representing the objects of animal behavior at later stages of evolution, because they use higher levels of neurological organization to represent additional aspects of the spatial structure of the world. Spatio-temporal and structuro-temporal imagination give the animal subject internal states that correspond to the world in a way that does not depend on the selection pressure that happen to have been imposed on the animal. It evolves because evolution is progressive. In order for animal to have more power to control relevant conditions, their behavior guidance systems must have animal systems of representation that represent objects as being located in space and as having geometrical structures. 

These forms of imagination in animals are the foundation, as we have seen, for the evolution of naturalistic and subjectivistic imagination in primates with the use of language. But those forms of imagination are also inevitable, and they involve a correspondence between brain states and the states in the world, including other subjects, that is also necessary.

This solves a problem that functionalist explanations encounter when they try to explain correspondence with nothing but causal connections between input and output within the organism. The correspondence is not just a constant conjunction between telesensory input and the object in space that is involved in reference, as Fodor seems to mean by calling it a casual connection, but an isomorphism between geometrical structures in the brain and the geometrical structures of the locations of objects in the space around the telesensory animal.