- #1
- 2,285
- 3
Our considerations of the problems of consciousness thus far have repeatedly brought us to questions about causation. In chapter 8, we saw arguments that we should be realists about causal connections. Given that causal constraint is a real feature of the world, what can we say about it, and what consequences does it have for our understanding of nature?
The obvious starting point in our investigation of causation is physics. Physics gives us powerful methods by which to describe and predict the evolution of physical phenomena, but it does not appear to tell us the whole causal story. We can readily be realists about the various entities described by physics but still be Humeans about causation, because the mathematical machinery of physics does not explicitly refer to anything like causal connections or causal constraint. The minimal theoretical obligation physics places upon us in this regard is that we hold that there are certain regularities that occur among physical phenomena; we need not view these regularities as the results of causal processes in order to interpret physics in a logically coherent fashion. If we choose to regard physics as a causal theory, it is only because we are projecting something extra into the theory, not because the theory itself forces us to take this point of view. Thus, if we accept the argument from chapter 8 that the world must have real relations of causal constraint, it follows that physics is not telling the whole causal story; it seems to tell us part of the causal story, while leaving the rest to reside implicitly in the background. Our task, then, is to unearth the portion of the causal story left implicit by physics and make it explicit, such that we cannot be realists about our complete theory of causation without also being realists about connections of causal constraint.
It seems that accounts of causation in philosophy have also been inadequate. Philosophers have historically considered causal responsibility the core explanatory target for a theory of causation, where the study of causal responsibility is the study of how a given event, process, agent, or fact can be credited as being the productive cause of a certain effect. But attempts to assign causal responsibility seem inevitably to involve intentional and interest-relative features, and so fail to be completely objective. As Rosenberg puts it, these problematic aspects of accounts of causal responsibility "create a striking portrait of a convenient explanatory construct rather than an objective natural relation" (p. 148). A natural theory of causation should strip away these intentional and interest-relative aspects and reveal the underlying objective basis, the causal significance of phenomena in nature.
Rosenberg proposes the following definition for causal significance:
In other words, the minimal fact that will be true in a world featuring real causal constraint is that certain things within it will place restrictions on how other things can be. Thus, we can think of causation minimally as an operator on a space of possibility. This minimal construal of causation is a more general version of our usual concept of the term. We usually think of causation as an asymmetric process occurring locally in space and forward in time, whereby a certain cause produces a certain effect. On the above definition, however, we are not committed to the conditions of asymmetry, locality, or directionality; a causal relation could be symmetric and non-local and still have causal significance. Causal significance and causal responsibility also differ in that for the former we characterize causation negatively, as a constraining influence, and the latter positively, as a productive influence.
To see how we could have a relation of causal significance that violates our notions of causal responsibility, imagine a world wherein two coins must always be flipped together, and they share the causal constraint that they must always land either both heads or both tails. The landing state of one coin has causal significance for that of the other, and vice versa, so this is a symmetric causal relation; there is no clear grounds for causal responsibility here whereby we could name the state of one coin the cause and that of the other, the effect. There is just a mutual condition of constraint on the possible joint states of the coins. In fact, we need not conjure up an imaginary world for this exercise; the states of entangled quantum particles in our world have a relation of symmetric causal constraint analogous to that of our imaginary coins. The shared causal significance of entangled particles is also non-local, thus presenting us with another violation of normal concepts of causal responsibility that is easily accommodated by a more general theory of causal significance.
Effective and Receptive Properties
Causal significance must arise as a result of certain properties of phenomena in nature that are causally relevant. Rosenberg names these causally relevant properties the nomic content of an individual, and proposes two basic kinds of such properties: effective and receptive.
Effective properties are those properties that have the capacity to place causal constraints on the space of possible ways the world could be. These are precisely the kind of properties studied by physics. Physics is ultimately an empirical endeavor driven by observation, and we only observe things (either directly or indirectly) by means of the roles they play in a causal chain of events that culminates with the systematic activation of our biological senses. In other words, all phenomena that can be the objects of objective empirical study must be effective properties or complexes thereof, because such effective causal properties are precisely the things that subserve objective observation in the first place.
Receptive properties are those properties that allow the causal constraints of effective properties to be placed. Receptive properties do not figure heavily, if at all, into contemporary notions of causation, but they seem to be a logically necessary counterpart of effective properties; asserting the existence of properties that can place causal constraint already presupposes the existence of properties that can 'feel' those causal constraints (even if only implicitly). Likewise, the notion of receptive properties already presupposes the existence of effective properties. In this way, effective and receptive properties are circularly defined, interdependent entities, but they are also distinct. We can think of the relationship between effective and receptive properties as analogous to the relationship between the front and back of a wall. The two seem to be dual aspects of the same fundamental thing, and although the existence of one logically requires the existence of the other, they are not identical and one does not supervene on the other.
We can get a firmer grasp on their distinctness by considering some hypothetical examples where one is at play but not the other. The classical conception of God as the unmoved mover, and Newton's absolute space and time, are both examples of hypothetical entities that have only an effective aspect; they can place causal constraint, but cannot themselves be changed or affected. The conception of consciousness as an epiphenomenon is an example of an entity with only a receptive aspect; an epiphenomenal consciousness can receive causal constraint from brain events, but cannot affect the brain, or anything else, in turn.
Rosenberg proposes to model receptivity as a connective property that binds various effective properties together, rather than as a monadic, one-place property. If we were to view receptivity as a monadic property, we would still have the problem of accounting for how an individual's receptive property connects to effective properties in various circumstances. Modeling receptivity as the connection itself avoids this problem, and also allows for further explanatory power. As we'll see, modeling receptivity as a connection will allow us to 1) produce an account of what levels of nature are and how they emerge, 2) produce conditions of substantial metaphysical unity upon which we can ground facts about natural individuals and natural individuation, and 3) sketch the outlines of how it could be that spacetime is not a fundamental entity, but rather arises from causal circumstances.
The Theory of the Causal Nexus
Let us now delve into Rosenberg's detailed theoretical proposal for what causation is and how it works. Recall that we characterize causal significance as an operator of constraint on a space of possibilities. Thus construed, causation solves the determination problem, which is stated as follows: given that the world contains many properties with many potential states, what is it that creates a world with determinate instantiations of these properties?
In the spirit of the determination problem, we can view effective properties as determinables, which are kinds of properties that may take on a range of particular values or forms, called the determinates. General examples of determinables include redness, a kind of property whose determinates are shades of red, and shape, a kind of property whose determinates include circularity, triangularity, etc. Fundamental effective properties in nature such as mass, charge, and spin are also determinables. Insofar as sufficiently unconstrained effective properties are not determinate, we can say they are incomplete, and they approach completion by becoming more and more determinate (i.e., by having more and more of their possible determinates excluded from possibility by conditions of causal constraint). An effective property is defined as being complete if and only if it is in a completely determinate state (i.e., all but one of its possible states have been excluded by conditions of causal constraint).
We can also consider receptivity in terms of completeness and incompleteness. Receptivity, modeled as a connective property, is a neutral essence with the ability to bind to effective individuals. We can think of this ability to bind to effective properties metaphorically as a kind of open 'slot' in a receptive property into which an effective property could fit. In this sense, an instance of receptivity is analogous to the plastic binding that holds together a six-pack of Coke cans, where the cans are the analogue of effective individuals. A complete instance of receptivity, then, is one whose 'slots' are completely filled, and an incomplete instance of receptivity is one that can bind to further effective properties. (Rosenberg builds a theory where all instances of receptivity have a discrete and finite number of 'slots,' but does not take this discreteness and finiteness to be a necessary feature of the theory.)
The six-pack metaphor is useful for loosely conceptualizing what an instance of receptivity is, but the proposed binding relation between effective and receptive properties is a more substantial one than that implied by the metaphor. What it means for effective and receptive properties to bind is that in some sense they become part of one another's natures, and in the process approach completion. In particular, if two effective properties EP1 and EP2 are bound to the same instance of receptivity R, then EP1 and EP2 infuse part of R's nature, and part of R's nature is taken up by EP1 and EP2. Thus, by transitivity supported by R, EP1 and EP2 become part of each other's natures, and so have a means of placing causal constraints upon each other's determinable states.
When a single instance of receptivity binds two or more effective individuals, the resulting property complex is called a causal nexus. Causal nexii are governed by causal laws, which are laws that describe the compatibility, incompatibility, and requirement relationships between effective properties within a nexus.
Natural Individuals
In order to characterize causation at various levels of nature, it is useful at this point to introduce the notion of natural individuals (which I will also refer to in the following simply as 'individuals'). Singular instances of 'primitive,' unbound effective properties and receptive properties are defined as level zero natural individuals. These should be taken as useful abstractions; lone instances of effective properties that are not bound to any receptivity, and lone instances of receptivity that are not bound to any effective properties, are not things that actually exist in nature. A level one individual is defined as a completed level zero receptivity that binds level zero effective properties. Level one individuals are the most fundamental kinds of things in existence, whether those turn out to be fundamental particles, or strings, or something else. A level two individual is constituted by an instance of a level one receptivity binding two or more level one individuals. An example of a level two individual would be a causal interaction between two colliding fundamental particles. In general, a level N individual (for N > 0) is defined as a set of natural individuals of level N-1 that are bound together by a common level N-1 receptive connection. Each instance of receptivity at each level is taken to be an ontologically novel and irreducible entity.
As with the primitive effective and receptive properties, we can evaluate an individual at level one or higher in terms of completeness, as follows: a level N individual (for N > 0) is complete if and only if all of the level N-1 individuals bound by its receptivity are complete. It is important to evaluate individuals in terms of completeness, because following causal significance and the determination problem, the process of causation is just the process of individuals becoming more determinate, or more complete. If a given individual is complete, then it is completely determinate, and no more causal work can be done on it (although it may place causal constraints on other incomplete individuals). If a level N individual X is incomplete, then its effective state is not yet determinate; there is a range of possible values that X's effective state can take, and it can approach completeness by binding with other level N individuals to form a level N+1 individual. The new level N+1 individual is a causal nexus in which X can receive constraints on its effective state from the other individuals to which it is bound (we say that the individuals with which X shares a common receptivity are within its receptive field). The exact nature of the constraint placed on X by the causal nexus is determined by the causal laws of the nexus and the effective states of the individuals within X's receptive field.
Causation at Work
To get a better hold on these concepts, imagine a world in which two individuals, A and B, are bound by a common receptivity R to form a causal nexus C. To evaluate the causal constraints placed on A and B by the nexus and its causal laws, we must first discern the possible effective states A and B could have as a result of their own internal causal relations, considered independently from the constraining influences of their causal environment. These are the independently possible states of A and B; it is these independently possible states that constitute the space of possibility which the causal relations inherent in C can operate upon and constrain.
Suppose that A has the independently possible states 1 and 2, whereas B has the independently possible states 1, 2, and 3. Considered independently, the possible joint states of A and B is the Cartesian product of their independently possible states, i.e. (1, 1), (1, 2), (1, 3), (2, 1), (2, 2), and (2, 3). Further suppose that there is a causal law on the nexus C such that the sum of the values of all its individuals' effective states cannot exceed 3. This causal law immediately excludes B's independently possible state 3, and also excludes the independently possible joint state (2, 2). As a result of this causal constraint, B has had one of its independently possible states filtered out, and thus has become more determinate (although it has not become completely determinate). Because A and B both condition each other in this scenario, the nature of the causal relation between them is symmetric.
Now suppose we tweak the situation slightly, such that A's only independently possible state is 2. In this case, we say A is determinate when considered independently; its own internal causal relations are sufficient to determine completely its effective state. As before, suppose B has independently possible states 1, 2, and 3. In this case, two of B's independently possible states, 2 and 3, are excluded by the causal law of the nexus, in conjunction with A's fixed state 2. B's only remaining possible state is 1, and thus B has become fully determinate as a result of its receptive binding with A. Note that although A placed effective constraint on B in this example, A's effective state was already fixed, and so it could not receive any causal constraint from B. Thus, this is an example of an asymmetric causal relation, since A conditioned B but not vice versa.
Note that the creation of higher-level individuals is essentially a recursive matter of the binding of lower-level individuals, with more and more constraints being placed as more and more individuals 'nest' within each other; thus, the route to completeness is via the creation of higher-level individuals. There is no in principle limit to the depths of this recursion; depending on the circumstances, it might take only a layering of two levels of nature to result in a complete (level two) individual, or it might take two hundred.
At this point we are faced with the significant question of what are the laws that govern the emergence of higher-level individuals, specifically, their configurations. Why would an individual of type A tend to bind with type B instead of type C, assuming such tendencies even exist? One rule that seems to follow directly from the model of causal significance considered here is a negative one: higher-level individuals cannot be formed unless there is indeterminateness at a lower level. If we are given a set of complete, level N individuals, there is no determination problem to solve, and thus no need for the creation of a level N+1 individual. The positive rules governing the formation of higher-level individuals seem to be a more difficult matter, and Rosenberg hesitates to offer a concrete proposal. However, he does nominate two principles which might guide how nature chooses to configure natural individuals. The first is the principle of maximal completion, which states that individuals tend towards completeness; under this principle, nature might favor the creation of those individuals which place the greatest constraints on their lower-level constituents. The second is the principles of thermodynamics; under this principle, nature might favor the creation of those individuals whose states have the highest entropy.
The obvious starting point in our investigation of causation is physics. Physics gives us powerful methods by which to describe and predict the evolution of physical phenomena, but it does not appear to tell us the whole causal story. We can readily be realists about the various entities described by physics but still be Humeans about causation, because the mathematical machinery of physics does not explicitly refer to anything like causal connections or causal constraint. The minimal theoretical obligation physics places upon us in this regard is that we hold that there are certain regularities that occur among physical phenomena; we need not view these regularities as the results of causal processes in order to interpret physics in a logically coherent fashion. If we choose to regard physics as a causal theory, it is only because we are projecting something extra into the theory, not because the theory itself forces us to take this point of view. Thus, if we accept the argument from chapter 8 that the world must have real relations of causal constraint, it follows that physics is not telling the whole causal story; it seems to tell us part of the causal story, while leaving the rest to reside implicitly in the background. Our task, then, is to unearth the portion of the causal story left implicit by physics and make it explicit, such that we cannot be realists about our complete theory of causation without also being realists about connections of causal constraint.
It seems that accounts of causation in philosophy have also been inadequate. Philosophers have historically considered causal responsibility the core explanatory target for a theory of causation, where the study of causal responsibility is the study of how a given event, process, agent, or fact can be credited as being the productive cause of a certain effect. But attempts to assign causal responsibility seem inevitably to involve intentional and interest-relative features, and so fail to be completely objective. As Rosenberg puts it, these problematic aspects of accounts of causal responsibility "create a striking portrait of a convenient explanatory construct rather than an objective natural relation" (p. 148). A natural theory of causation should strip away these intentional and interest-relative aspects and reveal the underlying objective basis, the causal significance of phenomena in nature.
Rosenberg proposes the following definition for causal significance:
pg. 150 said:The causal significance of a thing is the constraint its existence adds to the space of possible ways the world could be. A successful theory of causal significance should lay bare an objective base of facts on which less objective facts about causal responsibility might rest.
In other words, the minimal fact that will be true in a world featuring real causal constraint is that certain things within it will place restrictions on how other things can be. Thus, we can think of causation minimally as an operator on a space of possibility. This minimal construal of causation is a more general version of our usual concept of the term. We usually think of causation as an asymmetric process occurring locally in space and forward in time, whereby a certain cause produces a certain effect. On the above definition, however, we are not committed to the conditions of asymmetry, locality, or directionality; a causal relation could be symmetric and non-local and still have causal significance. Causal significance and causal responsibility also differ in that for the former we characterize causation negatively, as a constraining influence, and the latter positively, as a productive influence.
To see how we could have a relation of causal significance that violates our notions of causal responsibility, imagine a world wherein two coins must always be flipped together, and they share the causal constraint that they must always land either both heads or both tails. The landing state of one coin has causal significance for that of the other, and vice versa, so this is a symmetric causal relation; there is no clear grounds for causal responsibility here whereby we could name the state of one coin the cause and that of the other, the effect. There is just a mutual condition of constraint on the possible joint states of the coins. In fact, we need not conjure up an imaginary world for this exercise; the states of entangled quantum particles in our world have a relation of symmetric causal constraint analogous to that of our imaginary coins. The shared causal significance of entangled particles is also non-local, thus presenting us with another violation of normal concepts of causal responsibility that is easily accommodated by a more general theory of causal significance.
Effective and Receptive Properties
Causal significance must arise as a result of certain properties of phenomena in nature that are causally relevant. Rosenberg names these causally relevant properties the nomic content of an individual, and proposes two basic kinds of such properties: effective and receptive.
Effective properties are those properties that have the capacity to place causal constraints on the space of possible ways the world could be. These are precisely the kind of properties studied by physics. Physics is ultimately an empirical endeavor driven by observation, and we only observe things (either directly or indirectly) by means of the roles they play in a causal chain of events that culminates with the systematic activation of our biological senses. In other words, all phenomena that can be the objects of objective empirical study must be effective properties or complexes thereof, because such effective causal properties are precisely the things that subserve objective observation in the first place.
Receptive properties are those properties that allow the causal constraints of effective properties to be placed. Receptive properties do not figure heavily, if at all, into contemporary notions of causation, but they seem to be a logically necessary counterpart of effective properties; asserting the existence of properties that can place causal constraint already presupposes the existence of properties that can 'feel' those causal constraints (even if only implicitly). Likewise, the notion of receptive properties already presupposes the existence of effective properties. In this way, effective and receptive properties are circularly defined, interdependent entities, but they are also distinct. We can think of the relationship between effective and receptive properties as analogous to the relationship between the front and back of a wall. The two seem to be dual aspects of the same fundamental thing, and although the existence of one logically requires the existence of the other, they are not identical and one does not supervene on the other.
We can get a firmer grasp on their distinctness by considering some hypothetical examples where one is at play but not the other. The classical conception of God as the unmoved mover, and Newton's absolute space and time, are both examples of hypothetical entities that have only an effective aspect; they can place causal constraint, but cannot themselves be changed or affected. The conception of consciousness as an epiphenomenon is an example of an entity with only a receptive aspect; an epiphenomenal consciousness can receive causal constraint from brain events, but cannot affect the brain, or anything else, in turn.
Rosenberg proposes to model receptivity as a connective property that binds various effective properties together, rather than as a monadic, one-place property. If we were to view receptivity as a monadic property, we would still have the problem of accounting for how an individual's receptive property connects to effective properties in various circumstances. Modeling receptivity as the connection itself avoids this problem, and also allows for further explanatory power. As we'll see, modeling receptivity as a connection will allow us to 1) produce an account of what levels of nature are and how they emerge, 2) produce conditions of substantial metaphysical unity upon which we can ground facts about natural individuals and natural individuation, and 3) sketch the outlines of how it could be that spacetime is not a fundamental entity, but rather arises from causal circumstances.
The Theory of the Causal Nexus
Let us now delve into Rosenberg's detailed theoretical proposal for what causation is and how it works. Recall that we characterize causal significance as an operator of constraint on a space of possibilities. Thus construed, causation solves the determination problem, which is stated as follows: given that the world contains many properties with many potential states, what is it that creates a world with determinate instantiations of these properties?
In the spirit of the determination problem, we can view effective properties as determinables, which are kinds of properties that may take on a range of particular values or forms, called the determinates. General examples of determinables include redness, a kind of property whose determinates are shades of red, and shape, a kind of property whose determinates include circularity, triangularity, etc. Fundamental effective properties in nature such as mass, charge, and spin are also determinables. Insofar as sufficiently unconstrained effective properties are not determinate, we can say they are incomplete, and they approach completion by becoming more and more determinate (i.e., by having more and more of their possible determinates excluded from possibility by conditions of causal constraint). An effective property is defined as being complete if and only if it is in a completely determinate state (i.e., all but one of its possible states have been excluded by conditions of causal constraint).
We can also consider receptivity in terms of completeness and incompleteness. Receptivity, modeled as a connective property, is a neutral essence with the ability to bind to effective individuals. We can think of this ability to bind to effective properties metaphorically as a kind of open 'slot' in a receptive property into which an effective property could fit. In this sense, an instance of receptivity is analogous to the plastic binding that holds together a six-pack of Coke cans, where the cans are the analogue of effective individuals. A complete instance of receptivity, then, is one whose 'slots' are completely filled, and an incomplete instance of receptivity is one that can bind to further effective properties. (Rosenberg builds a theory where all instances of receptivity have a discrete and finite number of 'slots,' but does not take this discreteness and finiteness to be a necessary feature of the theory.)
The six-pack metaphor is useful for loosely conceptualizing what an instance of receptivity is, but the proposed binding relation between effective and receptive properties is a more substantial one than that implied by the metaphor. What it means for effective and receptive properties to bind is that in some sense they become part of one another's natures, and in the process approach completion. In particular, if two effective properties EP1 and EP2 are bound to the same instance of receptivity R, then EP1 and EP2 infuse part of R's nature, and part of R's nature is taken up by EP1 and EP2. Thus, by transitivity supported by R, EP1 and EP2 become part of each other's natures, and so have a means of placing causal constraints upon each other's determinable states.
When a single instance of receptivity binds two or more effective individuals, the resulting property complex is called a causal nexus. Causal nexii are governed by causal laws, which are laws that describe the compatibility, incompatibility, and requirement relationships between effective properties within a nexus.
Natural Individuals
In order to characterize causation at various levels of nature, it is useful at this point to introduce the notion of natural individuals (which I will also refer to in the following simply as 'individuals'). Singular instances of 'primitive,' unbound effective properties and receptive properties are defined as level zero natural individuals. These should be taken as useful abstractions; lone instances of effective properties that are not bound to any receptivity, and lone instances of receptivity that are not bound to any effective properties, are not things that actually exist in nature. A level one individual is defined as a completed level zero receptivity that binds level zero effective properties. Level one individuals are the most fundamental kinds of things in existence, whether those turn out to be fundamental particles, or strings, or something else. A level two individual is constituted by an instance of a level one receptivity binding two or more level one individuals. An example of a level two individual would be a causal interaction between two colliding fundamental particles. In general, a level N individual (for N > 0) is defined as a set of natural individuals of level N-1 that are bound together by a common level N-1 receptive connection. Each instance of receptivity at each level is taken to be an ontologically novel and irreducible entity.
As with the primitive effective and receptive properties, we can evaluate an individual at level one or higher in terms of completeness, as follows: a level N individual (for N > 0) is complete if and only if all of the level N-1 individuals bound by its receptivity are complete. It is important to evaluate individuals in terms of completeness, because following causal significance and the determination problem, the process of causation is just the process of individuals becoming more determinate, or more complete. If a given individual is complete, then it is completely determinate, and no more causal work can be done on it (although it may place causal constraints on other incomplete individuals). If a level N individual X is incomplete, then its effective state is not yet determinate; there is a range of possible values that X's effective state can take, and it can approach completeness by binding with other level N individuals to form a level N+1 individual. The new level N+1 individual is a causal nexus in which X can receive constraints on its effective state from the other individuals to which it is bound (we say that the individuals with which X shares a common receptivity are within its receptive field). The exact nature of the constraint placed on X by the causal nexus is determined by the causal laws of the nexus and the effective states of the individuals within X's receptive field.
Causation at Work
To get a better hold on these concepts, imagine a world in which two individuals, A and B, are bound by a common receptivity R to form a causal nexus C. To evaluate the causal constraints placed on A and B by the nexus and its causal laws, we must first discern the possible effective states A and B could have as a result of their own internal causal relations, considered independently from the constraining influences of their causal environment. These are the independently possible states of A and B; it is these independently possible states that constitute the space of possibility which the causal relations inherent in C can operate upon and constrain.
Suppose that A has the independently possible states 1 and 2, whereas B has the independently possible states 1, 2, and 3. Considered independently, the possible joint states of A and B is the Cartesian product of their independently possible states, i.e. (1, 1), (1, 2), (1, 3), (2, 1), (2, 2), and (2, 3). Further suppose that there is a causal law on the nexus C such that the sum of the values of all its individuals' effective states cannot exceed 3. This causal law immediately excludes B's independently possible state 3, and also excludes the independently possible joint state (2, 2). As a result of this causal constraint, B has had one of its independently possible states filtered out, and thus has become more determinate (although it has not become completely determinate). Because A and B both condition each other in this scenario, the nature of the causal relation between them is symmetric.
Now suppose we tweak the situation slightly, such that A's only independently possible state is 2. In this case, we say A is determinate when considered independently; its own internal causal relations are sufficient to determine completely its effective state. As before, suppose B has independently possible states 1, 2, and 3. In this case, two of B's independently possible states, 2 and 3, are excluded by the causal law of the nexus, in conjunction with A's fixed state 2. B's only remaining possible state is 1, and thus B has become fully determinate as a result of its receptive binding with A. Note that although A placed effective constraint on B in this example, A's effective state was already fixed, and so it could not receive any causal constraint from B. Thus, this is an example of an asymmetric causal relation, since A conditioned B but not vice versa.
Note that the creation of higher-level individuals is essentially a recursive matter of the binding of lower-level individuals, with more and more constraints being placed as more and more individuals 'nest' within each other; thus, the route to completeness is via the creation of higher-level individuals. There is no in principle limit to the depths of this recursion; depending on the circumstances, it might take only a layering of two levels of nature to result in a complete (level two) individual, or it might take two hundred.
At this point we are faced with the significant question of what are the laws that govern the emergence of higher-level individuals, specifically, their configurations. Why would an individual of type A tend to bind with type B instead of type C, assuming such tendencies even exist? One rule that seems to follow directly from the model of causal significance considered here is a negative one: higher-level individuals cannot be formed unless there is indeterminateness at a lower level. If we are given a set of complete, level N individuals, there is no determination problem to solve, and thus no need for the creation of a level N+1 individual. The positive rules governing the formation of higher-level individuals seem to be a more difficult matter, and Rosenberg hesitates to offer a concrete proposal. However, he does nominate two principles which might guide how nature chooses to configure natural individuals. The first is the principle of maximal completion, which states that individuals tend towards completeness; under this principle, nature might favor the creation of those individuals which place the greatest constraints on their lower-level constituents. The second is the principles of thermodynamics; under this principle, nature might favor the creation of those individuals whose states have the highest entropy.
Last edited: