by Pross, Addy
We have used the term ‘patterns’ to describe what it is that the inductive method seeks, though scientists typically use other terms, such as hypotheses, theories, laws, to mention the main ones, the difference being primarily in the degree to which the pattern has been confirmed. Thus Newton’s Law of Gravity is uncontroversially considered to be a law due to the innumerable times apples and other objects have fallen, and the regularity with which the sun rises every day. However, the term ‘pattern’ with its inherent fuzziness, does have its advantages. In contrast to terms such as ‘theories’ and ‘laws’ which radiate some sense of absolute truth, the term ‘pattern’ is more subtle, less committed, less definitive, more open to modification. Even Newton’s laws, those pertaining to gravity and motion, have had to undergo revision following Einstein’s revolutionary insights. If we keep in mind that every hypothesis, theory, or law is ultimately just a pattern, the day that theory or law is modified or revoked will be less surprising, less disconcerting.
As to the underlying reason for the existence of those patterns, rules, laws, generalizations, or whatever we wish to call them, science is unable and does not pretend to address such questions. Despite the widespread view that the laws of nature are the explanation of natural phenomena, Ludwig Wittgenstein, the great twentieth-century philosopher, pointed out almost a century ago in his famous Tractatus (Latin for treatise) that ‘the whole modern conception of the world is founded on the illusion that the so-called laws of nature are the explanations of natural phenomena.’ There is no fundamental explanation for any phenomenon and the best we can do is to say that the pattern is the explanation. Patterns are the link between the underlying reality and our understanding of that reality. The basis for the patterns, those underlying laws of nature, are fascinating questions in their own right, but these are philosophical questions, beyond the strict scientific domain, and therefore outside the scope of this discussion. To quote Wittgenstein yet again: ‘whereof one cannot speak, thereof one must be silent’.
Given the above statements it can be appreciated that there are degrees to understanding, that understanding is to a significant extent subjective, because the process of pattern recognition is not always definitive. Pattern recognition is, to some extent, in the eye of the beholder. As the Nobel physicist Steven Weinberg lucidly pointed out, as good a way as any to establish whether a pattern is insightful is to see whether it induces an ‘Aha!’ from colleagues. Having said that, however, it is clear that the nature of understanding within physics, a more fundamental science, is quite different from its operation within biology, whose domain is the study of inherently highly complex systems. Within physics generalizations are invariably rigorously quantified, articulated in the language of mathematics so that exceptions to the rule are not tolerated and require a reformulation of that rule. Within biology generalizations are frequently qualitative and exceptions to the rule are not just tolerated, but accepted as normal. In any case, regardless of the field of endeavour, it should be emphasized that the same set of observations may on occasion be interpreted in different ways and so may lead to the recognition of different patterns.
This is particularly true when the observed patterns are statistical rather than absolute, as is common in the social sciences, or when the patterns are qualitative rather than quantitative in nature. It is for this very reason that historians frequently come up with quite different models for understanding a set of historic events, since those events may be successfully organized in more than one pattern. The extensive literature on the causes of the First World War exemplifies the way an unambiguous set of historical events can be understood and interpreted in different ways. Nor do patterns have to be mutually exclusive. Both a 2-year-old child and a theoretical physicist have some understanding of why apples fall, though their explanations differ markedly. Both see in the falling apple the manifestation of a more general pattern, though the physicist recognizes a pattern that is both broader and quantifiable. Significantly however, the child’s simple ‘falling object’ rule is sufficient to serve him extremely well on a day-to-day basis. So provided that the child has no immediate plans to launch a satellite into space or undertake space travel, then for all practical purposes the extra insight that Newton’s law of gravity and Einstein’s theories of relativity offer into the behaviour of matter, beyond that offered by the ‘objects fall’ rule, will be of little consequence. In fact, if one thinks about it, the physicist about to undertake some mountain climbing is most likely to be applying the ‘objects fall’ rule to guide him in his adventure, rather than string theory or special and general theories of relativity.
In conclusion, when a system can be patterned in more than one way, the question as to which pattern is better may well depend on the particular application. The title of a 2009 Woody Allen movie, Whatever Works, captures the essential idea. Yes, that sums it up nicely—whatever works. Ultimately, whatever one calls them—theories, laws, models, hypotheses, patterns—all efforts to find order in our universe can never fully capture the reality of nature. The patterns we uncover are merely reflections of that reality—some better, some worse, whose recognition brings us some sense of order to the complex world that we find ourselves in. The preceding discussion will now assist us in addressing a central issue in the continuing search for biological understanding—the issue of reduction versus holism.
Reduction or holism
We pointed out earlier that the inductive method—the seeking of generalizations, the recognition of patterns—is at the core of all scientific understanding. However, a particular kind of inductive thinking has proven to be of special value, the one termed reduction. The concept of reduction can itself be elaborated upon and split up into a number of subgroups, something philosophers of science have been exploring in recent years, but these more detailed ideas need not concern us here. The essence of the reductionist approach is simply: ‘the whole can be understood in terms of the interaction of its constituent parts’. For example, if you want to understand how a clock works then break it up into its component parts—wheels, cogs, springs, etc., and see how these work together to create the functional entity. Reductionist thinking of one kind or other has been instrumental in advancing scientific understanding from the earliest days of the scientific revolution.
In opposition to the reductionist view is a more recent school of thought termed holism, whose philosophy can be summarized by the simple statement: ‘the whole is more than the sum of its parts’, and so appears to negate the reductionist view. Holism contends that within complex systems in particular, unexpected emergent properties arise that cannot be derived by examining the individual components of the system (by emergent properties we mean that there are properties at the higher and more complex level that are not observed at lower levels). This approach has gained considerable influence in recent years, specifically with regard to the biological sciences, due to the extraordinary complexity of even so-called ‘simple’ biological systems, and has led to the establishment of a new branch in biology—systems biology. Carl Woese’s view of biological systems as ‘complex dynamic organization’, rather than as a ‘molecular machine’ whose behaviour can be understood from its component parts, illustrates this new ‘systems’ way of thinking.1
So which is the better approach for addressing biological problems—reduction or holism? That depends on who you ask. Jacques Monod10 offered a rather disparaging view of holism (and holists) with his comment: ‘A most foolish and wrongheaded quarrel it is, merely testifying to the “holists” [sic] profound misappreciation of the scientific method and of the crucial role analysis plays in it.’ The confusion surrounding the apparent conflict between reductionism and holism as applied to biological systems is a long-standing one and graphically illustrated in the proceedings of a conference entitled ‘Problems of Reduction in Biology’ attended by a group of leading biologists and philosophers, including Peter Medawar, Jacques Monod, and Karl Popper that took place in Septe
mber 1972, in Bellagio, Italy. At the end of that meeting June Goodfield was reported as saying:
I am overpowered by a feeling of déjà vu verging at times on the very edge of intellectual impotence. ‘Reductionism’ ‘anti-reductionism’ ‘beyond reductionism’ ‘holism’…. The issue is a very old one recurring in various forms with unfailing regularity throughout biological history, and the feeling of impotence arises because, after all this time, the issue never seems to get any clearer.21
Well, almost forty years on and little seems to have changed. Reduction and holism in biology seem as controversial now as then. A recent polemical essay by Denis Noble that comes down firmly on the side of holism, discusses the same dilemmas, though illustrated with examples from modern systems biology.22 Carl Woese, a reborn holist, puts it even more starkly:
Biology today is at a crossroad. The molecular paradigm, which so successfully guided the discipline throughout most of the 20th century, is no longer a reliable guide. Its vision of biology now realized, the molecular paradigm has run its course. Biology, therefore, has a choice to make, between the comfortable path of continuing to follow molecular biology’s lead or the more invigorating one of seeking a new and inspiring vision of the living world, one that addresses the major problems in biology that 20th century biology, molecular biology, could not handle and, so, avoided. The former course, though highly productive, is certain to turn biology into an engineering discipline. The latter holds the promise of making biology an even more fundamental science, one that, along with physics, probes and defines the nature of reality.1
Powerful and provocative words indeed. But in a sharp critique of holism, the Nobel biologist, Sydney Brenner, recently wrote: ‘The new science of systems biology claims to be able to solve the problem but I contend that this approach will fail because deducing models of function from the behaviour of a complex system is an inverse problem that is impossible to solve.’23
Despite that treacherously uncertain backdrop, let us now briefly venture into this philosophic lion’s den. I will offer some thoughts on this troublesome philosophic divide and how it impacts on our goal of better understanding living systems. At least in the context of life, I propose that the reductionist–holistic divide is more semantic than substantive, and that holism, when probed more deeply, can be thought of as just a more elaborate form of reduction.
At the risk of gross oversimplification we may state that the most useful application of reductionist philosophy, when viewed as a scientific methodology, is the one termed ‘hierarchical reduction’, the idea being that phenomena at one hierarchical level can be explained using concepts taken from a lower hierarchical level. Steven Weinberg recently expressed the idea succinctly: ‘explanatory arrows always point downward’.24 Thus, to illustrate, one attempts to explain social behaviour based on individual organismic behaviour, organismic behaviour in terms of cellular behaviour, cellular behaviour based on biochemical cycles, and biochemical cycles rest upon more basic physical and chemical concepts of molecular structure and reactivity, and so on, continuing down to fundamental subatomic particles. Hierarchical reduction seeks to provide understanding level by level, with phenomena at each level being explained by the conceptual framework associated with the level immediately below. Much of the spectacular advance witnessed in the physical sciences since the scientific revolution of the seventeenth century can be directly attributed to the successful implementation of that methodology. Within the biological sciences, the reductionist harvest has been particularly abundant. The enormous advances in our understanding of biological processes, such as DNA replication, protein synthesis, metabolic cycles, etc., all derive from the reductionist methodology. Without question molecular biology has revealed many of the wonders of cell function at the molecular level—reduction par excellence.
But, as noted in chapter 1, the enormous complexity of biological systems often makes the reductionist methodology difficult to implement, and it is that difficulty that has been responsible for the burgeoning anti-reductionist, holistic approach to biological systems of recent decades. The holistic view derives its persuasive influence from the systems theory school of thought that builds on the idea that within complex systems, systemic relations arise that produce novel and quite unpredictable characteristics. So, in recalling Weinberg’s reductionist comment ‘explanatory arrows always point downward’, together with June Goodfield’s despairing commentary, how are we to respond to the two opposing viewpoints? And what are the implications of this apparently fundamental disagreement with respect to our attempts to understand life?
To a large extent criticism of the reductionist approach derives from extreme expressions of reduction, such as the one offered by Francis Crick,25 who claimed that ‘the ultimate aim of the modern movement in biology is to explain all biology in terms of physics and chemistry’. Such claims appear unrealistic for the foreseeable future, and remain an ultimate aim in just the same way that the ‘ultimate aim’ of chemistry is to predict all chemical phenomena through solving Schrödinger’s famous wave equation. In that sense the broadly based critique of reduction, given its inherent limitations, is on solid ground. But the idea that a more measured reductionist approach is unable to deal with emergent properties at all is clearly incorrect; emergent properties are regularly addressed and understood through reduction.
To take a simple example, consider the physical properties of condensed states (that’s just the term for solids and liquids) that we discussed earlier. Condensed states exhibit a variety of emergent properties that are totally absent at the single molecule level. The condensed state may be solid or liquid, it may be conducting or insulating, shiny or dull. A single molecule does not possess any of those condensed state properties. A single molecule is neither solid, nor liquid, neither shiny nor dull. Nonetheless, despite the absence of these collective properties at the molecular level, these condensed state properties are well understood based on the electronic characteristics of the individual molecules. So we understand why at room temperature molecular hydrogen is a gas, water is a liquid, and regular table salt is a solid, based solely on properties of the individual molecules (molecular weight, charge character, etc.) and the corresponding intermolecular forces that would be expected in those materials. Similarly we may usefully predict the solid state conductivity of a material by carrying out a particular kind of theoretical analysis on the individual isolated molecule.
My point is that physics and chemistry are replete with such reductionist analyses that offer insight into the underlying reasons for a wide range of emergent properties. The oft cited claim that some properties cannot be explained by reduction because they are emergent is simply incorrect, though, of course, this does not mean that all emergent properties can be explained by reduction. Reduction as a methodology does have its limitations, as does any methodology. Complex systems cannot always be readily reduced to their component parts. Unexpected emergent properties can and do appear and in those cases, it could be argued, a holistic approach may be necessary. But a deeper appraisal of the holistic view suggests that its anti-reductionist claim is misstated to a degree. The problem lies primarily with the meaning that the term ‘holistic’ conveys. If ‘holistic’ is intended to convey the impression that the entire system is treated as a whole entity, that reduction into components is avoided, then that is certainly not the case. The systems approach dissects the complex whole into component parts as does the reductionist approach, but addresses the complex nature of interactions within the system in a more realistic fashion. The holistic view recognizes that in addition to ‘upward causation’ from lower-level hierarchies to higher ones, one must also consider the possibility of ‘downward causation’ where higher-level phenomena influence actions at lower levels.
These kinds of feedback effects can lead to quite unexpected emergent properties that cannot be easily foreseen and are not readily amenable to a simple reductionist analysis. Nonetheless a moment’s thought reveals
that a reductionist philosophy is at the heart of holism as well. The holistic systems approach to understanding the complexity of a biological system continues to reduce the complex system into simpler elements, though placing greater emphasis on the complex nature of the interactions between those elements. In other words the holistic approach merely preaches a more elaborate form of reduction, one that recognizes that causal relations within a system can be more complex than those implied by a simple bottom-up causal chain. To quote Athel Cornish-Bowden, the British biologist:
the classical reductionist approach to science can be understood as a way of understanding the functioning of a whole system in terms of the properties of its parts, but now we must learn to understand the parts in terms of the whole.26
Reduction as an explanatory tool in science is difficult to circumvent because reduction is a key means of obtaining scientific understanding. Despite several decades of groping expectantly toward some kind of non-reductionist or even anti-reductionist methodology, that activity does not seem as yet to have born edible fruit. Holism, despite its name, can be thought of as just a reductionist elaboration, a potentially valuable elaboration for sure, but an elaboration nonetheless. Reduction in its various forms and sub-forms, was, is, and will likely remain the central conceptual tool in scientific endeavour. To the extent that the ‘what is life’ question can be satisfactorily resolved, I believe it can only be through a fundamentally reductionist approach—by seeking the underlying connections between chemistry and biology, by identifying the process responsible for biological complexification. Ultimately the difference between animate and inanimate must be reduced to differences in the nature of the materials within the two worlds and, in particular, in the way those materials interact and react.