CONTENTS

1   INTRODUCTION - THE PROBLEM

2   ONTOLOGICAL DEPENDENCE

3  EMERGENT PROPERTIES AND EMERGENT OBJECTS

4   TRANSCENDENCE

5   CONCLUSIONS

3   EMERGENT PROPERTIES AND EMERGENT OBJECTS

As with our examples of ontological dependence, so emergence can be characterised by “getting out more than you put in”. This time, at least initially, we are talking about features rather than objects. Various features of the world are postulated as fundamental, but as a result of these, under special circumstances, new features appear which are quite different from the original ones. In a subset of the cases, we may find ourselves wanting to speak about “emergent objects” as the bearers of these new features. In this section we try to develop this notion of emergence, first by giving examples, then asking more generally about the conditions under which we can say that emergence has happened, and finally by looking at the conditions for the emergence of new objects into the ontology of the world.

3.1       Examples of Emergence

To develop the notion of emergence, let us look at the main examples of this phenomenon we have encountered in our scientific study of the world. In each case the emergence is associated with the presence of a new scientific discipline dedicated to study the new features. This shows how the new features can be sufficiently different from the old ones that they merit a mode of study all of their own. (Historically these separate disciplines have grown up to deal with the features as discovered phenomenologically, with their status as being emergent features of something more fundamental being discovered only later.)

The first example is the emergence of chemistry out of physics. At the fundamental level we have physical objects, the fundamental particles (which at a still more fundamental level are seen to be emergent features, namely quantised excitations, of the fields). When enough of these particles are aggregated, they begin to act together as  continuous substances, and it is a good approximation to describe the material causation of macroscopic physical objects in terms of being made of one or more of these substances.

These chemical substances can be treated as objects (in the generalised, Fregean sense) in their own right. Unlike the physical objects which they make up, they are abstract objects, not dwelling in space, and not being characterised by size and shape. As well as actual lumps of sulphur, for example, we have sulphur the chemical substance. It is characterised by both physical properties, such as density and thermal conductivity, and the chemical properties, such as its propensity to react with oxygen to form sulphur dioxide. The physical properties are intrinsic to the substance itself; they can be defined without reference to any other substance. The chemical properties are about reacting with or turning into other substances.

Once we speak of chemical substances we are operating within the realm of discourse called “chemistry”. This grew up as an independent discipline to deal with the phenomenology of chemical substances. It was later discovered that the chemical substances are not continuous at all length scales. They are eventually made of physical particles which are themselves not made of the chemical substance. Sulphur is made of sulphur atoms, but these atoms are not made of sulphur. The physical and chemical properties of sulphur the substance depend upon the quite different sorts of property of the sulphur atoms, which arise from the quantum mechanical pattern formed by sixteen electrons when they bind to a nucleus with an electric charge of +16. Looking at it the other way, the properties of the substance are emergent features of the atomic properties, becoming manifest when a large number of the same atoms cluster together. The emergent properties are attributed to emergent objects: namely the chemical substances.

Our next example is the emergence of organic chemistry out of chemistry. A striking feature of chemistry is that the chemistry of carbon is as big or bigger than that of the rest of the elements put together, and that it is treated as being a separate subject. The reason why carbon chemistry is different in this way is because the carbon-carbon bond can be iterated indefinitely. With other elements we can formally iterate the bonds using the rules of valency, but in reality the resulting molecules rapidly become unstable. In inorganic chemistry we have to discover empirically whether a molecule exists or not. By contrast, in organic chemistry, subject to some weak constraints (that the bonds are not too highly strained and that the atoms are not forced to occupy the same space), any molecule you can draw can be made. Thus we have as an emergent feature a new framework of substances and relationships between substances, in which chemistry has become in effect part of the branch of mathematics called graph theory.

Within the iterations of organic chemistry we can take small sets of molecular subgroups and iterate them indefinitely to form long linear sequences of the subgroups. Out of the interactions between two such families of sequences: DNA and proteins, comes our next example, the emergence of life out of organic chemistry. Among the radically new features associated with life, the most striking and important is Darwinian evolution. This provides biology with its own emergent dynamic, and will be discussed in more detail in Section 4 below, under the heading of “transcendence”. The new features are attributed to emergent objects, the most obvious of which are the biological organisms and the biological species.

Somewhere along the evolutionary process we begin to observe the emergence of sentience out of life. All organisms react to properties of their environment, but the more complicated animals begin to centralise the representation of information about the environment, and react, at least in part, to this representation, instead of directly to the environment. Somewhere along this line as well, there begins to emerge a centralisation of the state of self, dependent upon the state of the environment, and this begins to take over as the primary cause of behaviour. The problem of characterising this new property of sentience is as much a philosophical as a scientific one, because here we have the beginnings of subjectivity, which begins in turn to take us outside the boundaries of the competence of science.

Full subjectivity is achieved with the emergence of personhood out of sentience. A new dynamic, that of the personal will and of moral obligation, emerges to challenge the imperatives of Darwinism. Persons are the emergent objects which are the bearers of these new dynamical features. Because the new features constitute a new dynamic, they will be discussed further below as an example of transcendence.

Emergence does not stop with personhood. Just as when many atoms or molecules aggregate the result is a chemical substance, so when many persons congregate in a society, there are emergent features of that society. Just as subjectivity emerges out of the purely objective physical world, so out of the subjectivity of personhood can emerge social features which are effectively objective. One example of this second objectivity is that of moral law, . Another, perhaps less elevated, but no less important, is monetary value. The basis of monetary value is subjective: things have this value when people believe them to have it. However because the final value is the result of a large number of individual subjective valuations, it becomes effectively an objective feature of the world, obeying its own inscrutible and often chaotic laws of motion. From the point of view of the single individual, the price of food and the price of housing are as much objective features of the external environment as the temperature and the rainfall.

In complex societies, the economic value of things is determined by markets. In the behaviour of markets we see the unwilled outcomes of a multiplicity of willed actions. We can perceive in this behaviour some features of lawlike behaviour, but the laws are not teleological in nature. Instead they are more akin to the efficient laws found in the physical world, though they differ from them in detail. We can characterise this as the emergence of unteleological system from a teleological substrate, the inverse of what happens with the emergence of persons.

Although the behaviour of markets is non-teleological, it still bears the mark of its teleological substrate, namely that it is conceptually reflexive; knowledge of the system has causal power. There is feedback from the non-teleological whole to its teleological components. People taking part in the market can learn about the market movements as a whole, and modify their behaviour accordingly. This feature is an important component of the overall emergent dynamics of the system.

3.2            Conditions for Emergence

Having looked at some examples, let us now try to understand what these examples have in common, to develop a general account of the emergence of new features. A key word in any discussion of emergence is “different”. The new features have to be sufficiently different from those put in by hand. This means we can speak of a spectrum of types of emergence, according to how different the new features are. We can speak of weak emergence when the difference is not great, and strong emergence when the new features are dramatically different from the old. The word “transcendence” is used to describe emergence of the strongest type.

In this discussion, I want to reserve the word “emergence” for those cases where certain strength conditions are satisfied. I conjecture that three conditions which define a useful concept of emergence are the following;

• emergence occurs only under special conditions;
  
  
• emergence involves the coming into being of a new causal framework, ontologically dependent on but distinct from the old causal framework;
  
  
• although the emergent features are ontologically dependent on the old, there is a possible world in which they are the fundamental features.

The requirement that emergence takes place only under special conditions rules out simple material causation as an emergence. The composite objects are always present when the constituents are present; no further conditions are required to hold. The special conditions which help characterise emergence might occur only at a special time, as happened with human evolution, or only in possible worlds where the laws of nature allow the emergence as a logical consequence, as is the case with organic chemistry, which is dependent upon the laws of quantum mechanics giving a unique strength to the carbon-carbon bond.

The second condition stipulates that emergence is more than the appearance of new features at certain historical epochs. It is not just about what happens, but the general way in which things tend to happen. Truly emergent features are like laws of nature. Moreover, says the third condition, they are sufficiently independent of their ontological substrate that it is conceivable that, in different circumstances, they could be introduced as fundamental features of the world. We can conduct much of our discussion of them without having to mention their ontological substrate. For example, most of our discussions of genetics can treat the four bases as fundamental objects, without mentioning the atoms out of which they are made.

Given these conditions, we can revisit the example of waves in a continuous dynamical system and ask if they then count as an example of emergence. The first condition is obeyed: only certain types of dynamical system can sustain wavelike behaviour. The third condition is also obeyed. We can discuss much of wave behaviour without making explicit reference to the underlying continuous system, which means we could imagine a possible world in which the fundamental objects had most of the properties of waves. The second condition is the hardest to satisfy in this case: is the equation for the propagation of waves sufficiently distinct from the equation of motion for the constituents of the continuum to count as a separate law of nature? Even if the waves did pass this test they would still have to be characterised as a weak example of emergence, in that their laws of motion are not dramatically different from those of the underlying substrate.

3.3           Emergent Objects

Having investigated the conditions for the appearance of emergent features, let us now shift attention to the question of when emergence is accompanied by the appearance of new objects. We encountered this question first in Section 2.5 above, when discussing whether waves should be added to our ontology as objects. Now we return to it, examining in more detail the issues surrounding the ontology of objects.

A problem with Fregean ontology is to maintain the objectivity of object reference, especially for those objects which are ontologically dependent on others. Object reference must not depend on our choice of mode of expression - for example on whether we choose to speak in terms of the predicate “... runs” or in terms of the abstract noun “running”. We need an objective criterion determining whether or not “running” refers to an object, and indeed whether “activity” denotes a category of objects to which running belongs. If they do exist, all these activities will be ontologically dependent, because there cannot be an activity without something to undertake the activity.

The question of the existence of objects in Fregean ontology is addressed at some length by Dummett. The danger which Dummett seeks to avert is that “Frege’s notion of a proper name, and, with it, his notion of an object, would be reduced to absurdity because serious objects were overwhelmed by frivolous and spurious ones”. This is an important requirement. Whereas we should not be dogmatically narrow in saying what can count as an object (say restricting objects to physical ones, which is contrary to Frege’s introduction of the notion of object as a logical one, well before the development of a physics to describe the contingencies of this world), the notion of object would be destroyed if objects were allowed to proliferate without restraint.

Without analysing Dummett’s own answer with the care it deserves, we note that his general conclusion is that to be a serious candidate for being an object, the putative name-referent has to belong to a class for which we have an extensive vocabulary with which we can talk about this referent. “About a supposed object which can only be referred to in one way we can perforce say very few things; and this is what underlay our earlier feeling that we do not take seriously the use of such abstract nouns as names of objects.”

For a realist, this criterion seems too subjective. Maybe we can say very little simply because we have not chosen to develop the vocabulary with which to do so. Just because objects of a given type exist does not mean we are compelled to talk of them. But conversely, just because we choose not to talk of them does not mean they do not exist. Of course it is possible that the reason why we have such a poor vocabulary is that it is not possible to construct such a vocabulary. Here we are getting nearer to what might be an objective criterion, but to use it we would have to show that all attempts to construct such a vocabulary collapse.

The general principle against ontological extravagance is of course Occam’s razor. We are told that “ objects should not be multiplied without necessity”. The whole question then comes down to what counts as “necessity” in this case. The realist answer is that it is necessary to posit an object if and only if it exists, but this begs the real question, namely how are we to tell when one of these objects exists. From this realist perspective, Occam’s razor might be seen as a defence of realism, saying that the multiplication of objects is not at our disposal, but should be undertaken only when the world actually possesses these objects.

A tentative answer to what might count as the objective necessity for the introduction of new objects was given in Section 2.5 above, in the context of the discussion of waves as objects. Some emergent features are not able to take the old objects as their bearers – in the case of waves, transverse motion was the feature which could not be attributed to the constituents of the continuous dynamical system. The image of the process whereby new objects emerge is the following. We start with the names of the fundamental objects fitting into the argument places of the initial predicates. The emergence of new features is reflected by the appearance of new predicates with new argument places. Sometimes these new “holes” which open up are fitted by the names of the old objects, but sometimes this is not possible. In the latter case, where the new “holes” require the names of new objects, our ontology must admit the existence of new, emergent objects.

To provide an actual example of the process described metaphorically here, let us simplify our example of waves in a continuum, and consider a single oscillator. We can talk of this system being excited, but when there is only one oscillator, there is no grounds for speaking of an object called an excitation. However when there are two coupled oscillators, one can, and perhaps must, begin to speak of the transfer of an excitation from one oscillator to the other. This transfer is the simplest example of the process which ends up, in a continuous chain of oscillators, as wave propagation. Notice that the transition from “can” to “must” here is important; this turns a manner of speaking into objective existence.

We can speak of the excitation because conservation of energy allows us to say that the new excitation in the second oscillator is the same one as was earlier seen in the first oscillator. We can turn “being excited” into “possessing an excitation”. This advent of an identity criterion establishes the possibility of a new object, and it is a first step towards what Dummett characterises as having an extensive vocabulary to talk about the object. This having a vocabulary is not merely a matter of choice; there have to be features in the world to which this vocabulary refers. Moreover when we talk purely about the oscillators we can say that they are (timelessly) coupled, and that each of them has such-and-such an excitation history, and even that these histories are correlated in certain ways. However to speak in terms of a transfer requires that something other than the oscillators be transferred. Relative to the introduction of the concept of transfer, the emergence of excitations as the objects transferred is compulsory. The question then remains as to whether the concept of transfer has objective reference in the first place, and is therefore in this sense required to be spoken of (recalling that we are not compelled to speak of everything that exists).

A typical sentence describing transfer of the excitation might be

“Excitation X of oscillator A is transferred to another oscillator.”

Here we see both the original reference to oscillator A as that which is excited, and the new reference to excitation X as the thing transferred. This shows that the emergent feature is still a feature of the old objects, but in addition there are new objects associated with the new argument places. It is then possible to suppress reference to the underlying oscillators, and talk simply of the transfer of X. The presence of this reference to the oscillators is implicit in the sense of  the term “excitation”, even though this aspect of the sense may no longer be explicitly expressed by the form of words chosen. The same form of words without this implicit component to the sense would describe the possible but counterfactual world in which the excitations were the fundamental objects.

The question now is: under what conditions must we associate new argument places with new objects? A candidate for this condition is:

new objects emerge if and only if the emergent features contradict the essential features of the fundamental objects.

This contradiction prevents us identifying the bearers of the emergent properties with the fundamental objects - it provides the necessary identity barrier. In the example of the pair of oscillators, the basic property of the oscillators is that they move about their respective centres. One oscillator can not move from its position to that of the second one. The transfer of excitations contradicts this basic property and must therefore be a feature of a different object. In the case of the waves, the longitudinal propagation contradicts the property of the elements of the continuous system that they have only small transverse motions about their central position.

As a second example of an identity barrier, let us go from a very simple case to perhaps the most complex form of emergence, namely that of the person. The more one develops the conceptual structures necessary to understand what happened with the creation of the human person, the more different the human person appears to be from everything that went before. It is not only that the original framework of efficient causation has been overlaid with an emergent framework of final causation, which explains the behaviour of the persons. The persons are the source of this new causation. When comparing this with the situation in the physical world, the persons are less like the physical objects in the temporal theatre of history, and more like the timeless set of physical laws which govern what happens in history (note that to make this analogy we have to have an ontology which recognises the objective existence - in some sense of “existence” - of these timeless laws). In the world of persons, we individual persons are more like modal than historical structures.

The physical body is by contrast purely historical. The modal features of the person contradict this, and so can not be attributed to the body. It is not a matter of choice that we speak of the person as something different from the body, it is an objective consequence of the type of emergence which happens when a human body is assembled. What remains from the fact that the properties are emergent from (rather than being emergent properties of) the human body is, in place of identity, the relation of ontological dependence.

Click here to go to Section 4