Mike Holliday

Section 4: Emergence, Existence, and Explanation

1) Emergence

I said at the end of the previous section that the alternative account I had given of our relationship with the world was one characterized by emergence. But this is a notion that can take a number of different forms; for example, Robert van Gulick (2001) distinguishes between no less than ten different types of emergence. Therefore I obviously need to provide more details about the way in which the developing metaphysical account may be said to be an emergent account.

On the one hand, there are what might be termed 'strong' versions of emergence. For example, van Gulick describes a variant that he terms 'radically emergent causal powers', where a higher level system has features that are not in any way determined by the properties or relations belonging to its constituents. The suggestion is that the higher level has causal powers that act independently of the causal powers that operate at lower levels. For example, one might imagine a 19th century biological vitalist, who explains that, although living beings are certainly not immaterial, the fact that they have 'life' cannot in any way be accounted for by the arrangement of their material parts; no matter how much we cut-up and delve down into the organs and sinews, we will never find the nature of a living being. Instead, suggests the vitalist, we must hold that the vital essence of life, that which makes an animal move and exhibit behavior, is something that somehow arises out of, but is not understandable in terms of, certain types of material objects. Now, there seems to be a mystery inherent in this sort of emergence, because it is difficult to see how the novel powers that appear at the higher level (e.g. 'vitalist' powers) can be made consistent with the belief that the micro-physical world is a closed causal system. As van Gulick reminds us, the emergent level arises out of the base level, and the two levels are therefore not completely independent, as would be the case with a fully dualist conception. This implies the possibility that the separate higher level causal powers might have an effect on events at the lower levels; in other words, that there might be some form of downward causation. But now we seem to be forced to choose between two possible causes of the behavior of an element at the base level (for example the behavior of an atom within a molecule) - is the behavior determined by the physical laws that apply at the level of individual particles, or is it determined by the causal laws that apply at the emergent level as they affect the lower, base level?1

However, there is no mystery about other sorts of emergence. For example, consider the mass of a block of lead. This mass (e.g. precisely 1 kilogram) is emergent in the sense that it is something that is possessed by the block but which is not possessed by any of the block's parts, for example its molecular constituents. In fact, this seems so innocuous that one might be inclined to query the use of the term 'emergent' in this instance, but the point to appreciate is that there is nothing that is necessarily mysterious about the fact that a whole can have properties that do not belong to its parts.

Given this range in potential interpretations of the concept of emergence, I need to clarify the sense in which I have used the term. Let me start by considering the distinction, made by van Gulick (2001) among others, between ontological emergence and epistemic emergence. Ontological emergence concerns real-world items such as objects, properties, events, and processes, and their relations such as identity, composition, and realization. Epistemic emergence concerns explanatory notions that are to do with how we predict, explain, and understand real-world items; hence it deals with relations of derivability, conceptual necessity, and pragmatic equivalence. Now the account I gave in the section An Alternative Metaphor suggests that no clear demarcation between ontological emergence and epistemological emergence can be sustained. What objects might be said to exist depends upon our perceptions and actions as subjects and participants. For example, in Maturana's view (1970, p. 10) a 'system' or a 'unity' only exists as such relative to an observer: 'Strictly, the identity of a unit of interactions that otherwise changes continuously is maintained only with respect to the observer, for whom its character as a unit of interactions remains unchanged.' The concept of the system as a unity must therefore be understood as a unity with respect to the observer, rather than as a unity in some sort of metaphysical sense. The system can be seen by an observer as a simple unity or as a composite unity, but not as both at the same time - attempting to do so can lead to erroneous descriptions and explanations:

Frequently the dual existence of living systems in particular, and of systems in general, is obscured by the notion of emergent properties. By treating the features that an observer distinguishes in a system as if they were intrinsic to it, the notion of property obscures the relational nature of these features. ... So, to speak of emergent properties in the constitution of a system is both a mistake and misleading. As a system is constituted as a totality, a new domain arises, the domain in which the system exists as that totality. To say that autopoiesis is an emergent property would be a mistake. To say that the constitution of an organism gives rise to emergent behavior would also be a mistake ... (Maturana 2002). 

The science journalist John McCrone (2001) would agree with Maturana here, since he writes:

A gas is orderly within a physical context. Under certain gravities, states of containment, etc, it has an orderly statistical behaviour once it has reached equilibrium. So it gains emergent properties such as temperature, pressure, combustibility. But this order is only being appreciated from the outside. ... all the observing ... is external to the system. The gas can only meaningfully be considered 'a gas' - to have a bounded identity - from the standpoint of the environmental context that creates whatever order exists and does the responding to that order. Its emergent properties are quite literally only there in the eye of the beholder.

El-Hani and Pihlström (2002) make a similar point. Noting Putnam's arguments against reductive, physicalist notions of causation and in support of the idea that causation is a normative, interest-relative concept, they conclude that we should be careful not to reify emergence with a 'metaphysical essence'. It is not possible, suggest El-Hani and Pihlström, to draw a clear line between ontology and epistemology because ontology relates to a conceptualized reality. Since all properties are the result of human interpretation, or of our transactions with our environment, nothing is non-emergent.

Therefore ontology, as the description of what sorts of items can be said to exist, has an irreducible epistemic element to it - necessarily so, one might argue, in order to avoid making unjustified assumptions at the start of our metaphysical analysis. We might bring this back to my conception of an emergent metaphysics as follows: (i) it is precisely objects or types that are emergent, as emphasized by Smith, Maturana, and Pattee; (ii) objects are identified as such by means of their practical powers and causal effects (in particular their effects on the perceiving entity); (iii) therefore the establishment of the object also institutes some form of 'science' of the behavior of the object; (iv) it is this establishment of the concept of an object of a particular type and of the concepts we use to describe its behavior that leads to the possibility of a link between that object and those entities that we consider to be its constituents and the concepts that describe their behavior (such a link being considered to be either a deduction upwards and/or a reduction downwards); (v) without both of these ontologies, the ontology of the whole and the ontology of the parts, we cannot deduce or reduce - for example, how can I deduce anything about the behavior of water from the behavior of hydrogen and oxygen, unless I already possess the concept of water (and accordingly already know a great deal about its behavior)?; (vi) thus objects and their properties are 'emergent' in the sense that they have to be conceived of as objects and properties in order for reduction and deduction to take place. The notion of emergence therefore has a strong relationship with that of existence, i.e. what it is to say that some thing exists, as well as to reduction and deduction, i.e. what it is to explain something.

2) Existence and Hypostatization

A consistent theme throughout the previous two sections, Mirroring Reality and An Alternative Metaphor, was that 'objects' do not exist in some manner that is completely independent of human thought or activity.2 This suggests the possibility that we may on occasions inappropriately treat something as an object. Improperly treating a concept or abstraction as a concrete object is termed reification or hypostatization. For example, if someone believes that a mental image of a tiger has to have a determinate number of stripes, then they would be reifying mental images by assuming that they have all the attributes of physical objects. Another frequently used example of reification is the belief that, because we can try and measure intelligence using a single IQ number, intelligence must therefore be a single trait in humans, as opposed to a multi-faceted concept relating to human abilities. A third example would be that of treating nations as if they are individual persons that are capable of actions such as decisions and judgments, for example 'France has decided it doesn't want anything to do with ...'. Of course, sometimes such a usage is unexceptional, because we know exactly what the speaker means. The problem comes when what is initially used as a metaphor to get across a limited point is then utilized as a premise in an argument to derive further conclusions. An example of this might be someone who claims that, because 'France has decided on something', then anyone wanting to prevent France from acting on that decision is depriving France of its rights. The metaphor has outrun its usefulness.

One reason behind the temptation to reify would seem to be the prevalence of objects in the physical realm that we inhabit as humans, or 'medium sized, dry goods' as J. L. Austin termed them. We can unthinkingly transfer the sorts of properties that we associate with such objects (e.g. essences, causal effectiveness, and temporal-spatial coherence) to the conceptual classifications that we make when discussing other realms.

Given that the existence of objects is relative, hypostatization is relative as well, and the correctness or otherwise of treating something as an object is itself context sensitive. This can be seen in the discussion of reification in the The New Fontana Dictionary of Modern Thought (ed. Bullock & Trombley 1999):

An analysis of any relationship in a complex world involves a process of simplification through a set of abstractions in which certain aspects of a given phenomenon are selected and stressed for heuristic purposes. These abstracted elements of reality may be reduced to an ideal type or a conceptual model. If they are taken as a complete description of the real phenomenon and the resulting abstractions endowed with a material existence of their own, the process exemplifies what [Whitehead called] 'the fallacy of misplaced concreteness', which is in effect a special case of the fallacy of reification.

I want to use the term hypostatization, rather then reification, because I think that the former term provides the nuance that we take something that is initially conceived of as an activity and then incorrectly treat it as an object that is in some way separate from the activity with which we started. Consider an example referred to by Susan Hurley in her book Consciousness in Action (1998, note on p. 412). It is quite possible for a system to exhibit rational behavior, even if it isn't possible to describe any of the components of the system as rational; rationality might be an emergent property of the entire system. Now rationality is a generic description of certain human activities, and to assume that we must be able to trace rationality to certain components of our cognitive system (i.e. that we are rational because something in our brain is rational) is, I suggest, a hypostatization of the concept of rationality. If you like, people can be rational, but their brains cannot be. Similarly, psychologists do not normally look for something called 'anger' but rather they seek the mechanisms and processes that produce the behavior that we classify as expressing anger.

Of course, there is a point at which the hypostatization of activities does make sense. At some level of analysis, we will no doubt be justified in taking a hypostatized term to be an entity. Whether a hypostatization is appropriate depends on the purpose for which the attribution is made and at which level of description. A particular temptation is to reify the concepts that we generate as part of our scientific heuristics. Husserl, for example, considered that certain 'methodological devices of the sciences, chiefly idealisation and objectification, have been misunderstood such that their objects are thought to yield the natural world as it is in itself ...' (Moran 2000, p. 143). As Husserl suggests, we must be careful not to confuse scientific methodology with ontology. 

The context sensitivity of whether something can be said to be 'incorrectly' treated as an object suggests that there is a sense in which all objectification is reification. Consider, for example, Quine's comments (1995, pp. 24, 36-37) to the effect that the origin of our processes of objectification lies in our primitive perceptual capacities:

The conjunction ['Black and dog'] describes any scene in which black and dog are both salient, whereas 'Black dog' requires that they be situated together, the black patch engulfing the canine patch. The predication expresses the compact clustering of visual qualities that is characteristic of a body. Bodies are our first reifications ... . It is in analogy to them that all further positing of objects takes place. ... Typically a body contrasts with its visual surroundings in color and in movement or parallax, and typically it is fairly chunky and compact. ... [A further step is] the transcending of the specious present. ... To begin with, we recognize a recurrent body merely by strong perceptual similarity. ... Next we distinguish fixed bodies from vagrant ones by the locomotive effort on our own part that accompanies the disappearance and recurrence of a fixed body. ... Our science at length progresses to where qualitative indistinguishability is neither necessary nor sufficient for identity. A body can grow, shrink, discolor.

So hypostatization can be seen as a natural, but potentially problematic, cognitive trait. The specific sort of misplaced usage of hypostatization that I have in mind is, in general terms, as follows: (i) we classify certain activities as being of a particular type; (ii) this leads to the concept of that sort of activity; (iii) we then treat that concept as a hypostatization; (iv) we start arguing as if the hypostatization is available as an explanation of the activities that we considered at the outset; (v) but the hypostatization is ontologically superfluous and circular as an explanation.3 To use Brian Cantwell Smith's terminology, it is an inscription error.

3) Hypostatization and Metaphor

On this account, hypostatization is closely related to the use of metaphor, which increases our ability to understand a particular situation or a domain of thought by relating it to another situation or domain with which we are more familiar. Max Black (1993) provides a good description as to why we often resort to metaphor: 'the available literal resources of the language [are] insufficient to express our sense of the rich correspondences, interrelations, and analogies of domains conventionally separated; and ... metaphorical thought and utterance sometimes embody insight expressible in no other fashion.' On the one hand, we can emphasize the positive side of metaphors, that they assist us in articulating novel ideas, including theories that our existing literal language may not be flexible enough to allow (Boyd 1979). On the other hand, we may mistakenly take metaphors as being themselves explanations, rather than useful (re)descriptions, and in such cases they may give us an unwarranted feeling that something has been fully understood simply by use of the metaphor (Pylyshyn 1979).

The psychologists George Lakoff and Mark Johnson have suggested that much of our cognitive activity depends upon metaphoric thought processes, even if the metaphorical element is not immediately apparent (Lakoff & Johnson 1980, Lakoff 1987, Lakoff & Johnson 1999). They explain (1980, p. 115) that '[b]ecause so many of the concepts that are important to us are either abstract or not clearly delineated in our experience (the emotions, ideas, time, etc.), we need to get a grasp on them by means of other concepts that we understand in clearer terms (spatial orientations, objects, etc.).' Of particular interest is Lakoff and Johnson's description of ontological metaphors, whereby we view events, emotions, and activities as if they were physical objects; as such they can be referred to, categorized, and hence reasoned about (1980, p. 25). As an example, they give the experience of rising prices, which can be metaphorically viewed as an entity by terming it inflation. Then we can say 'inflation is lowering our standard of living', 'we need to combat inflation', and so on. Lakoff and Johnson (1980, p. 28) say that it is so natural to use such metaphors that we often take them as direct descriptions, especially where they relate to mental phenomena, e.g. 'he broke down':

If you say 'The odds are against us' or ' We'll have to take our chances,' you would not be viewed as speaking metaphorically but as using the normal everyday language appropriate to the situation. Nevertheless, your way of talking about, conceiving, and even experiencing your situation would be metaphorically structured (1980, p. 51).

In Lakoff and Johnson's view (1980, pp. 56-57) the ultimate grounding of all metaphorical descriptions is our direct experience, of which our perception of physical objects forms a major element, and the only concepts that are not metaphoric are basic experiential categories such as those of space, and of our own motor functions - in other words, categories that we live in a fundamental manner. This is not to say that our experiences themselves are metaphoric. If we compare the examples Harry is in the kitchen with Harry is in love, then the emotion described by the latter is just as much a basic experience as the visual experience of spatial reality in the former. But our conceptualization of the emotional in the latter example is metaphoric (1980, p. 59).

Or rather, our conceptualizations contain substantial metaphoric elements: for there is no clearly drawn distinction between directly experienced concepts and metaphorical concepts. Indeed, we tend to experience things as a gestalt, as a set of properties that belong together (1980, pp. 53-54, 69-71). So our concepts of our own experiences may never be wholly non-metaphoric. But metaphors hide certain aspects of our concepts, namely those that don't fit into the metaphor (1980, p. 67).

There is an underlying dissonance inherent in our use of metaphors as part of our conceptual and cognitive apparatus. Any one set of metaphors cannot hope to fully capture the complexity of reality and of our lives; we therefore tend to use multiple metaphors, and these can contradict each other in certain respects (Lakoff & Johnson 1999, pp. 71, 78). Accordingly, we frequently need to shift from one set of metaphors to another set (Lakoff & Johnson 1980, pp. 220-221). However, we normally view our experiences in terms of an objectivist model; by doing so we can more easily draw non-conflicting inferences and expectations, and develop suggestions for future actions. But this type of model does not allow for the multiple descriptive schemas that are implied by our having to use more than one metaphor.

Lakoff notes that the traditional objectivist view is based on the notion of truth; accordingly, in a situation where we have more than one way of describing a situation, or of modeling the world, each of which fit the facts, the objectivist view cannot find any way to assess the appropriateness or otherwise of the different descriptions or models (Lakoff 1987, p. 201). In Women, Fire, and Dangerous Things (1987), Lakoff is primarily concerned with showing that this traditional, objectivist, view of meaning and rationality is incoherent. The traditional view is summarized by Lakoff (1987, p. xii) in a manner that is very similar to Putnam's characterization of metaphysical realism: (i) rational thought consists of the manipulation of symbols; (ii) these symbols get their meaning from a correspondence with the world, and therefore represent the things in the world; (iii) the world is construed as objective, in the sense that it has certain properties independent of our understanding of it. Central to this view is the notion of a category, since words and concepts usually represent types of things, rather than individual things, and the traditional view is that we place things in the same category if (and only if) they have certain properties in common. Lakoff emphasizes that this last point is a philosophical position, based on a priori grounds, and that the psychological, linguistic, and anthropological evidence shows that this position is not empirically supported, causing major difficulties for the traditional or 'objectivist' view (see Lakoff 1987, pp. 12-154, for his review of this evidence).

Nevertheless, we frequently tend to behave as if the objectivist view of truth were always appropriate. One aspect of this is our tendency to overuse our metaphors, resulting in the sort of hypostatization that I described above. A key element in hypostatization is that we take a metaphoric expression that isn't obviously such and then draw implications as if it were fully literal, thereby ignoring other aspects of the situation that are hidden by the hypostatization, and arriving at spurious conclusions by assuming that the hypostatized entity behaves like an object in all respects.

In Lakoff and Johnson's opinion (1999, pp. 118-129), metaphors are conceptual in nature - a view which they contrast with the normal opinion that metaphors are simply a particular way of using words and language, rather than a method of thinking. I take Lakoff's main thesis to have a philosophical underpinning: having argued that metaphysical realism is false, he seeks an alternative explanation for how we categorize things, and is led by the psychological evidence to the conclusion that categorization must utilize an embodied method. Lakoff and Johnson term this method primary metaphor, and stress (1999, p. 57) that it has a developmental, rather than a linguistic, basis: '[it] is not the result of a conscious multistage process of interpretation. Rather it is a matter of immediate conceptual mapping via neural connections.' One example of this at work in the young is their conflation of everyday experiences, which provides numerous basic metaphors that link their experiences of sensorimotor activity with their subjective experiences and judgments (Lakoff & Johnson 1999, pp. 48-49).

Lakoff and Johnson suggest that there are a number of basic concepts that seem to emerge directly from our activity as conscious, embodied humans, and which we utilize in the categorization of our experiences. These include (i) causation, (ii) participants, (iii) parts, (iv) stages, (v) linear sequence, (vi) purpose. If we have a gestalt with these dimensions, then we can perceive coherence in our experiences by understanding them in terms of such gestalts (1980, pp. 81-82). 'Domains of experience that are organized as gestalts in terms of such natural dimensions seem to us to be natural kinds of experience [and] are a product of our bodies ..., our interactions with our physical environment ..., [and] our interactions with other people in our culture ...' (1980, p. 117). So if we think in terms of a 'something', a hypostatization, that has these dimensions, it will appear to us as if we have a suitable framework for understanding the area that we are interested in. And concepts that we comprehend in a metaphoric manner tend, therefore, to be understood in terms of the interactional properties of our more basic concepts; that is, those properties that have to do with perception, motor activity, purpose, and function (1980, pp. 119-122).

The discussion so far suggests that we might consider an 'object' to represent a set of implications which is typical, or (perhaps better) which is seen to be reliable from the perspective of our everyday expectations or the methodology of science (whereas a metaphor might be considered to represent a much looser set of implications that is of use in imaginatively examining possibilities). The typical implications of an object will often consist of its actions or the results of those actions. So we might consider it to be the case that things are what things do.

4) Parts and Wholes

Much of our modern, Western understanding can be characterized as a reductionist understanding, whereby one thing is reduced to some other thing(s). I want to reconsider our concept of reduction in light of the above discussion concerning hypostatization and metaphor, and the best place to start, I think, is with the notions of parts and of wholes. The discussion so far suggests that 'parts' and 'wholes' do not exist in and of themselves, independent of our perception and description of things. To quote James H. Austin (1998, p. 550): 'We and the pebble are still integral parts of one and the same ongoing universe. It is still a universe undivided. It has never been split up by such artificial distinctions as minutes, first names, rock names, or place names. It is, in its suchness.' These distinctions and objectifications depend on our cognitive processes and on our interactions with the world. Therefore the idea that objects are 'put together' in some way from their parts is not the only way of viewing the relationship between them. Brian Cantwell Smith (1996, p. 270) voices an alternative when he says that 'The pieces are (partially) sedimented or extruded from the whole; the whole is not put together from the pieces ... .' This is because, for Smith, individuality is not, ontologically speaking, a fundamental feature of the world.

George Lakoff (1987, pp. 47, 51) believes that Tversky and Hemenway (1984) provide a suggestive explanation as to why our normal conception of the relationship between parts and wholes is such a strong one. An object's functions are usually strongly related to what we perceive to be its parts; the way we perceive an object depends on its shape, which is in turn related to the object's parts; we interact with objects through their parts; and so on. So our basic knowledge of physical objects is largely correlated with the division between wholes and parts. For basic objects, says Lakoff, the distinction between different parts, and the resulting interactional properties, will appear to us to be objective.

This might be interpreted as a claim that analyzing things in terms of their parts is somehow a 'natural' form of explanation. But I want to block such an understanding at the outset. After all, Lakoff's whole approach is based on the use of metaphor, even if the ultimate grounding of our metaphorically derived concepts and explanations is our existence as bodies in a physical world. Rather, I want to suggest that the metaphorical nature of our explanations and our concepts means that we can never have a single, straightforward understanding of the world. This means that there will always be a difference of some sort between considering something as a whole and considering its parts. Maturana (1970) seems to make the same point when he says:

In principle a part should be definable through its relations within the unit that it contributes to form by its operation and interactions with other parts; this, however, cannot be attained because the analysis of a unit into parts by the observer destroys the very relations that would be significant for their characterization as effective components of the unit. Furthermore, these relations cannot be recovered through a description which lies in the cognitive domain of the observer and reflects only his interactions with the new units that he creates through his analysis. Accordingly, in a strict sense a unit does not have parts, and a unit is a unit only to the extent that it has a domain of interactions that defines it as different from that with respect to which it is a unit ... . 

I interpret Maturana as saying that, if we consider something as a part, then we are objectifying it as such and consequently understand the total situation from a different perspective than if we had considered as an object the whole to which the part belongs. And if we try to understand the part and then understand the whole, we will find that we generate problems for ourselves, because our cognitive relationship to what we are considering is different in the two cases. As Roberto Poli (1999) describes it: 'In principle it is always possible to take a part of [a] whole and analyze it as a whole. This means that the element is analyzed in and of itself, without being considering in terms of its connections / functions / dependences vis-à-vis the whole with which we started.'

Poli's view of the relationship between wholes and their parts derives from a view of the universe as flux that owes much to Brian Cantwell Smith. Accordingly, he views objects as regions of relative stability within this flux and emphasizes the importance for ontology of processes and relationships. He therefore sees objects as necessarily complex and their parts as representing the object's significant relationships, both internally and externally:

To take up and generalize a remark by Bohm, we should 'give up altogether the notion that the world is constituted of basic objects or "building blocks". Rather, one has to view the world in terms of universal flux of events and processes' ... [This] assumption of anti-atomicity entails a number of important consequences. Firstly, every object, precisely because it is complex, is a whole with parts (both as components and as functional parts). The complexity of the object also determines the fact that the parts of the object interact with each other according to various kinds of dependence internal to the object. This means that an object has a structure and consequently structural stability (Poli 1999).

5 Reduction and Explanation

These conclusions concerning the nature of the relationship between wholes and their parts affect our understanding of analysis and of reduction.4 For example, in order to provide a reductive explanation of something, I have to have a description of its parts and then relate those parts to the aspects of the whole that I wish to explain. But my conception of a whole has to be understood in a somewhat different context than my conception of its parts (i.e. the parts as they are themselves and not as merely elements that are dependent on the whole). This suggests the likelihood that I will encounter difficulties in fully accounting for the relation between the whole and its parts. From this perspective, both complete epistemological reduction and complete ontological reduction are problematic. It is not the case that I can fully understand the whole in terms of its parts, nor is it the case that the properties of the whole can be fully reduced to the properties of the parts.

Perhaps it is the case that we can perform analysis and reduction precisely because we ourselves are a part of the whole, embedded in the universe, and therefore have a perspective from which parts and wholes are meaningful categories by which to parse the world, (which in turn suggests that analysis and reduction would be impossible if we had the perspective given by a 'view from nowhere'). I referred above to John McCrone's suggestion that the behavior of a gas is only evident if we think of the gas as an entity that is in some way not identical to 'a lot of particles'; if we concentrate on just the particles, then that is all we have - a lot of individual particles whizzing around. Of course, as cognitive entities, we can deduce the behavior of something that we term 'a gas', but we should be wary of moving from this directly to ontological conclusions that simply reify our own epistemological activity. In a similar vein, Michael Silberstein (2001) comments on how our hierarchical view of nature may actually be a reflection of our epistemology:

... the microscopic and the macroscopic may only be contextually separable from one another. Such properties warn us not to reify the layered conception of the world. The standard divisions and hierarchies between phenomena that are considered fundamental and emergent, simple and aggregate, kinematic and dynamic, and perhaps even what is considered physical, biological and mental are redrawn and redefined. Such divisions will be dependent on what question is being put to nature and what scale of phenomena is being probed. It is true that science is divided into hierarchical descriptions and theories, but given mereological emergence, this might be only an epistemological artifact of scientific explanatory practice and not a fact about the world.

More generally, I want to emphasize here the idea that explanation is part of our human involvement in the world, and should not itself be hypostatized into some form of transcendent knowledge of the universe 'as it really is'.5 The significance of reduction and analysis for our understanding of the world derives instead from the universe's dynamic and relational nature, but precisely because of this it can produce (as Maturana and Poli suggest) only a partial understanding.

Having worked through some of the general implications of an emergent metaphysics, I now want to turn to an area where there may be more specific lessons: that of the explanation of our own conscious experience, which remains one of the most contentious areas of Western philosophy.

This version started: July 2004

Last updated: July 2005

Here are links to:

To send me an email, click here.


[1] For a more detailed discussion of some of the issues involved here, see van Gulick 2001, pp. 17-20; Kim 1998, especially  pp. 39-47; and Loewer 2001.

[2] To give a further (and rather extreme) illustration: if we identify some large group of atoms as 'a star', then we are placing importance on atoms that are near to other atoms - but why should we do this? There is no logical reason to do so, outside of our own activities and interests as embodied and active participants in the universe. Putnam (1987, p. 35) provides a similar example: there seems no logical reason why we shouldn't accept the existence of the discontinuous object consisting of my nose and the Eiffel Tower - 'This is an unnatural object to talk about, to be sure, but what has the 'naturalness' of an object to do with its existence?' ( Putnam actually uses this example to make an ad-hominem point against a supposed anti-mereologist who objects to merological entities as 'strange discontinuous objects', whereas I am utilizing it for my own, rather different, purposes.)

[3] Susan Oyama (2000, p. 63) gives an excellent description of the tendency to make this error in discussions concerning biology. She refers to the evolutionary theorist Ernst Mayer, who, she claims:

rightly criticizes the reification of processes, like 'life' and 'mind', [but] sometimes fails to heed his own good advice ..., turning regular, organized ontogenetic, physiological, and behavioral processes into things, programs, that are then used to explain the processes.

[4]  See van Gulick 2001 for a categorization of the different types of views that are labeled 'reductionist'. The whole area of reduction, and its implications for the sense in which anything may be called emergent, is a contentious one in the philosophical literature. A flavour of the debate can be obtained from Bickle 1998, Kim 1998, Silberstein 2001, and Huttemann 2004. One interesting topic is that of inter-theoretic reduction, for example of psychology to biology, or biology to physics. For example, Silberstein notes that a strict deductive account of inter-theoretic reduction cannot really be said to hold even for the examples usually described. One example is the reduction of classical thermodynamics to statistical mechanics: in this case, the reducing theory includes essential terms, such as statistical terms, that simply do not appear in the theory that is (purportedly) being reduced. Another example is quantum chemistry, which cannot be deduced from quantum mechanics because of, inter alia, the many body problem (Silberstein 2001). Modern theories of reduction instead concentrate on the causal mechanisms that can, from a pragmatic perspective, explain the higher level theory (see, for example, Bickle 1998). Another short, but useful, discussion is Kirk 1999, pp. 156-163.

More generally, there is a tendency for reductive explanations to have something of a bad name, since many see them as holding that the entity to be reduced is 'nothing but' something else. We must, however, distinguish between (i) ontological reduction, the question of whether the world is nothing but the fundamental constituents of reality, or at least determined by such constituents, and (ii) epistemological reduction, which relates to whether our scientific theories about the world can be reduced to or identified with our theories about the fundamental constituents (Silberstein 2001, p. 67). The type of metaphysics that I have outlined finesses many of the difficulties; the following section on Consciousness will, I hope, give a more specific example.

[5] From Maturana's perspective, we often misunderstand the nature of explanation because we have been misled by the standard metaphysical picture. This picture implied that a theory or an explanation 'mirrored reality' and accordingly ignored how the theory or explanation related to the observer who derived or adopted it. Therefore, most attempts at explanation assume a unique 'reality', which acts as a measure of the correctness of the explanation:

... in this explanatory path, the assumption by different observers of different kinds of independent entities as the ultimate source of validation of their explanations constitutively leads them to validate with their behaviour different, and necessarily mutually exclusive, universes, realities or domains of objective explanations (Maturana 1988a).

Accordingly, claims Maturana, we tend to believe that someone who does not accept an argument that we put forward must have done so simply because they are unintelligent, or acting in bad faith. But the domain in which an explanation is understood is the domain of the observer's life-praxis, and not the domain in which the phenomenon that is to be explained actually exists - 'explanations and descriptions do not replace what they explain or describe' (Maturana 1988a). Therefore as an observer I must have some criteria by which I accept a purported explanation as an explanation. If I feel that this criteria has been met, then my puzzlement or doubt ceases, and I stop asking the same question over and over again. We live in a number of different domains, says Maturana, but we move between them emotionally, not rationally; therefore we have no way in which we can ensure that another person accepts our argument unless they already accept the premises that are constitutive of the domain to which our argument belongs - for example, a scientific argument may appear completely besides the point to someone who lives mainly in the domain of the religious. The enquirer 'accepts or rejects a statement as a reformulation of a particular situation of his or her praxis of living ... [and] ...determines whether that statement is or is not an explanation' (Maturana 1988a). So if I am discussing a matter with someone and find that we fundamentally disagree, what we need is not a 'compelling argument' but rather to find a domain in which our positions can be mutually understood and accepted, or for our disagreement to be accepted in mutual respect (Maturana 1988a).