A Pundit Postscript on Prigogine for Preserving Palatability in Parts of a Previous Post.

"Walking before you can run works well as a strategy for getting somewhere, because it's easier to identify a mis-step while walking than running. However, once you can walk, running can be pretty useful for getting somewhere else, but it's not the only way of getting there. Neither is walking." Pundit's Parable about the Reliability of Reductionism.

Reductionism, The Reductionistic Fallacy, and The End of Strict Determinism.

In a the previous post, Pundit discussed a rather somber book, and remarks in it by Peter R Wills about reductionism in biology, and in his own hopefully light-hearted posting, made a passing reference to a second book about some very interesting parts of modern Physics.

Book number two, illustrated above, is The End of Certainty:Time, Chaos and the New Laws of Nature, by Nobel Laureate, Ilya Prigogine.

Pundit has transferred his comments about The End of Certainty here, to help keep the previous post more light-hearted and enjoyable, and being diligent, has reread The End of Certainty, which has triggered several ideas for a new post.

But he will have to ask readers here to be patient. While it's possible, but not easy, to switch from reading molecular genetics and biology to books on modern physics without getting indigestion when the topic is extremely well written (as a Prigogine, or a Richard Feynman can do), the Pundit can't easily do this off the cuff, and also ensure that the casual reader can both understand and enjoy Pundit prose.

It's a big jump.

Pundit needs to do some extra thinking, drink several strong espresso caffes, and work through the re-drafts. So readers will see this post evolve as the caffeine takes effect.

The start with, here's a few explanatory remarks about The End of Certainty, to add to the of the beautiful image of its cover shown at the top.

A summary of Prigogines book:

Simplifying thinking about matter to the level of single components, such as molecules are other fundamental units, has been enormously successful in increasing our understanding of the universe. But it is not the only way physics proceeds, and its certainly not the only approach that biologists use.

For instance, when you attempt to predict the behaviour of large numbers of molecules, instead of one molecule, especially in realistic situations subject to energy flow from outside, away from conditions of chemical equilibrium- that is, predicting real life on the earth - completely new ideas are needed, and new behaviours of matter emerge that don't mesh easily with what we know about behaviour of single isolated components.

Phase transitions of matter, such as melting of ice, are one instance of this emergence of new properties. A single water molecule doesn't melt, but ice does.

And to any biologist who has taken a course in physiology of multi-cellular organisms, the idea that all biology is or could be studied only by moving progressively downward to cells and moleculular components, and only in this downward direction to smaller components in isolation from the companion investigation of the actual behaviour of functioning organisms, and indeed, populations of organisms, is simply a ludicrous proposition.

Reductionism in the sense of reducing models to simplified components such as single molecules is not even the only way physics proceeds, let alone biology, and this truism has been obvious the physicists for nearly a century, if not longer.

For instance, analysing matter using statistical concepts applied to large numbers of components is used in the discipline of statistical mechanics, which involves predictions about properties of bulk matter that emerge from the modeled or theoretical statistical properties of large numbers or ensembles of its component molecules. This is the logical opposite path of reasoning to that involved in reducing matter to its component parts - it is integrating the behaviour of many components to understand the emerging properties of matter in bulk.

(Readers totally unfamiliar with statistical mechanics may wish to consult the Stanford Internet Encyclopedia of Philosopy which has the to say:

Statistical mechanics was the first foundational physical theory in which probabilistic concepts and probabilistic explanation played a fundamental role. For the philosopher it provides a crucial test case in which to compare the philosophers' ideas about the meaning of probabilistic assertions and the role of probability in explanation with what actually goes on when probability enters a foundational physical theory. The account offered by statistical mechanics of the asymmetry in time of physical processes also plays an important role in the philosopher's attempt to understand the alleged asymmetries of causation and of time itself.

Sklar, Lawrence, "Philosophy of Statistical Mechanics", The Stanford Encyclopedia of Philosophy (Summer 2004 Edition), Edward N. Zalta (ed.), URL = http://plato.stanford.edu/archives/sum2004/entries/statphys-statmech/)

This pathway - the physics of complexity- of proceeding from simplified components to emerging complex behaviours, especially irreversible behaviours, and to modelling reality in detail, has been taken several steps further in degree of sophistication in recent years by Nobel prize-winner Ilya Prigogine and his colleagues.

His ideas in this direction have helped improve fundamental understanding of the meaning of time itself. Most remarkably, they imply that strict determinism - namely thinking that the future is precisely predictable from the present - an assumtion which has bothered lots of philosophers and lots ordinary people too, is finished as a credible scientific concept. That's good news. Laplace was wrong.


One quotation from The End of Certainty will be given to illustrate the relevance of Prigogine's book to Wills' reductionism argument that was the lead in to this post.

In essence, Wills' argument takes reductionism to mean" seeing the world as being explained only by its simplest smaller components such as atoms". But the quote from Prigogine that follows shows that reductionism in this sense is only part of the full story of modern physical sciences, just as it is clearly only part of biology, including molecular biology. The Reductionistic Fallacy can also be made during attempts to apply holistic approaches to the real world, as dicussed below.

But first, consider page 54, of Prigogines's book, after reminding ourselves first that physical phase transitions of matter (such as melting of ice into water) are an emergent property of large numbers of molecules, we find the following comment:

As long as we consider merely a few particles, we cannot say whether they form a liquid or a gas. States of matter are ultimately as well as phase transitions are ultimately defined by [whether there are large numbers N of molecules and the volume V becomes large]. The existence of phase transitions shows that we have to be careful when we adopt a reductionist attitude. Phase transitions correspond to emerging properties. They are meaningful only at the level of populations, and not of single particles.

One could easily argue that most of physiology (or to use today's buzzword - Systems Biology) is a study of the emergent properties of cells, and thus physiology (and Systems Biology) itself is the essence of anti-reductionism.

Amazingly, we now know even the simplest organisms display phase transitions when they get together in larger populations. That is, they radically transform into new types of cell populations – e.g. by becoming luminescent, or killing their host quickly if they are parasites. The study of such interesting group phenomena displayed by bacteria is currently a hot topic in microbiological research, and is called “Quorum Sensing” in the jargon of the experts.

The Non-Reductionist activities of Molecular Biologists is an extremely broad topic and a very active current research field. To illustrate this, here is a snippet from the Pundit's Email, at the time this post was first drafted:

Systems Biology
From Single Cells to Environments
15 - 26 May 2006
Venue: Plant Biotechnology Centre, Primary Industries Research Victoria, Department of Primary
Industries, Victorian AgriBiosciences Centre, 1 Park Drive, La Trobe R&D Park, Bundoora, Victoria

Monday 15th May 2006

09.30 – 10.30 Plant and animal whole genome shotgun sequencing
Huanming Yang, Beijing Genomics Institute, China
11.00 – 12.00 A scan for positive selection in the genomes of humans and
chimpanzees
Melissa Hubisz, University of Chicago, USA

Tuesday 16th May 2006
09.30 – 10.30 Discoveries through integrated functional genomics studies of the
model legume Medicago truncatula
Lloyd Sumner, Noble Foundation, USA
11.00 – 12.00 From microbial genomics and metagenomics to functional
genomics: insights into microbial physiology and evolution
Ian Paulsen, The Institute for Genomic Research, USA
Wednesday 17th May 2006
09.30 – 10.30 Application of microarrays in agricultural genomics and comparative
genomics
Juan Loor, University of Illinois, USA
11.00 – 12.00 Flu typing and biosecurity applications using microelectrochemistry
oligoarrays
Andy McShea, CombiMatrix, USA
Thursday 18th May 2006
09.30 – 10.30 Tissue-specific molecular profiling of transcripts, proteins and
metabolites
Julia Kehr, Max-Planck-Institute for Molecular Plant Physiology, Germany
11.00 – 12.00 Metabolism in specialised cell types
Glenn Turner, Washington State University, USA
Friday 19th May 2006
09.30 – 10.30 Transplastomics: opportunities and challenges
Ulrich Koop, University of Munich, Germany
11.00 – 12.00 Phenomics: In vivo functional analysis of genomes including
systems biology of early embryogenesis in Caenorhabditis
Fabio Piano, New York University, USA
Monday 22nd May 2006
09.30 – 10.30 Genomic technologies for systems approaches to energy
alternatives
Paul Richardson, DOE Joint Genome Institute, USA
11.00 – 12.00 Genome-wide non-mendelian inheritance of extra-genomic
information in Arabidopsis
Robert Pruitt, Purdue University, USA
Tuesday 23rd May 2006
09.30 – 10.30 Computational systems biology and semantics of genomes
Eric Werner, Cellnomica Inc., USA
11.00 – 12.00 Eucaryotic regulatory networks
Anthony Borneman, Yale University, USA
Wednesday 24th May 2006
09.30 – 10.30 Epigenetics and gene silencing in plants
Robert Martienssen, Cold Spring Harbor Lab, USA
11.00 – 12.00 Systems approach to understand cellular networks
Steve Wiley, Biomolecular Systems Initiative, USA

The Reductionist(ic) Fallacy.

REDUCTIONISTIC FALLACY--This fallacy is to reduce some multifaceted or complex concept or entity to only one of its elements or qualities. "A flag’s only piece of cloth." "God is only a father figure." And you are only a bunch of molecules, so how can I possibly pay any attention to your opinions concerning flags and God? Oops. That was illogical . However, it was highly illustrative, you have to admit.

From an Irrational Fallacies web-site.

But Peter Wills' view, that reductionism is to do with too much focus on sub-components such as molecules to explain the biological world, misses out on capturing the full extent of a logical fallacy, known surprisingly enough as the Reductionistic Fallacy.

In fact, so-called holistic approaches can easily suffer from a Reductionist Fallacy too, if the concepts or assumptions used to understand the whole behaviour of a system are too simplistic. In other words, a total system approach can be in error if it reduces the system with too much simplification- i.e. holism itself is necessarily reductionist, and this reductionism too, can be wrong. What's more, empirically testing whether an holistic approach is false is inherently more difficult than testing the validity of theories about its simpler components.

This is a simple fact of reality which underlies the success of most modern science. Walking before you can run works well as a strategy for getting somewhere, because it's easier to identify a mis-step while walking than while running. However, once you can walk, running can be pretty useful for getting somewhere else, but it is not the only way of getting there.

Jon Richfield richfield-AT--telkomsa.net had the following perspective perspective to contribute (provided in a slightly different context) in a recent response to an essay by Freeman Dyson that seems very relevant:

It this he points out that the conceptual simplification of some form of reductionism is part of every approach to understand the world, including holistic approaches, and that a "reductionistic fallacy" can easily worm its way into holistic approaches too.

There are logical pitfalls everywhere.

The weary and wearying harping on reductionism in science in general and the soft sciences in particular, tediously conflates, or at least fails to distinguish clearly between, two different things. These are reductionist fallacy and reductionism.

Reductionism as such is a valid and vital tool and strategy for investigation into mechanisms and for distinction between rival hypotheses. Reductionism as a tool is what we all do when we try to sort out how things probably happen, will happen, or might have happened in the past. It is reduction of the complexity of the aspects of the system under investigation. We attempt thereby to isolate the essentials that we suspect of being the independent variables of the mechanism under consideration. To the extent that we succeed, this puts us in a position to vary the parameters and predict the effects. To the extent that our predictions are successful, they offer us grounds to defend our support for any one of the possible hypotheses that anyone might have proposed in the absence of our reductionistic investigations.
...
On the other hand, if anyone thinks that it follows that non-reductionistic research is easier to perform validly than reductionistic, please do not disturb me in my innocence to tell me about it before you are sure how to answer some inconvenient questions.

For example, how is one to justify confidence in the results? How is one to justify the cost of research in which we do not apply reductionistic techniques to avoid combinatorially large ranges of unconstrained possibilities? What form might so-called holistic research take, that is not logically equivalent to standard research techniques? What would the special value of the outcomes of such research be? And so on.
...
One way or another, I trust it is clear in this context, that to speak of "the obsolescence of reductionist biology as it has been practiced for the last hundred years" makes as much sense as it might to speak of "the obsolescence of reductionist biochemistry as it has been practiced for the last hundred years ". Even if what matters is not the reductionistic parts, nor the organism, nor the community, but the Gaian planetary whole, we cannot make a single valid holistic statement if even by implication, it contradicts the basic facts of physiology, of biochemistry, of physics, of arithmetic, or of basic logic.

This is where we encounter the reductionistic fallacy, as opposed to reductionistic strategies. The fallacy is when one omits, or fails to recognise or allow for, relevant variables in a complex situation. The classical form of the fallacy is to say that X is "nothing but" Y. For instance, one might imply that fluorine and iodine are nothing but halogens, or deuterium nothing but hydrogen, or that Homo, Hirudo, and Hirundo are nothing but animals. If such statements were valid and relevant, then the respective items would be identical to the other elements of their classes, which is not in general the case.

Overlooking the distinctions between entities is one form of the fallacy. And certainly a lot of research has been rendered invalid by failure to recognise invalid assumptions.

To speak of "the need for a new biology based on communities and ecosystems rather than on genes and molecules" is nonsense. You cannot have any meaningful biology at all without genes and molecules, nor indeed without electrons and radicals. The research must be "based on", take account of, not only the existence, but the nature of its components. Nor can one have communities and ecosystems without individuals and organs as well as molecules and particles. To ignore this, even to under-emphasise this, would exemplify the reductionistic fallacy in its crudest and most intense form.

In fact, it goes beyond that. For anyone even to speak of complex systems such as communities and ecosystems without criteria for defining them as entities is conceptually questionable, to put it politely. To speak of them without retaining recognition of their "reductionistically" recognisable components and their component functions would be flatly incompetent. It would be about as silly, and as fallaciously reductionistic, as speaking of cells as "simply cells", without bearing in mind the nature of their membranes, their organelles, their endosymbionts, all of which can vary relevantly even within a single organism.
...
In spite of the simplistic partisan views of certain classes of computer science zealots, top-down is not intrinsically better than bottom-up; context is crucial at all times. The better one understands the infrastructural components the more likely the top-down is to be appropriate, but until then we cannot get far without the requisite bottom-up spadework. In science, including the science of "communities and ecosystems", we are far from the point where we can safely skip that spadework.

It does not follow that we must defer everything but low-level work, but to suggest that "holistic" approaches are any more important, sounder, more overdue, or less vulnerable to reductionistic fallacy is ridiculous. It is equally ridiculous to disparage the volumes of ecological biology that already have been done, or to characterise other studies pejoratively and unsoundly as "reductionistic".

See also

Hoefer, Carl, "Causal Determinism", The Stanford Encyclopedia of Philosophy (Summer 2005 Edition), Edward N. Zalta (ed.), URL = http://plato.stanford.edu/archives/sum2005/entries/determinism-causal/

Groarke, Leo "Informal Logic", The Stanford Encyclopedia of Philosophy (Summer 2006 Edition), Edward N. Zalta (ed.), forthcoming URL = http://plato.stanford.edu/archives/sum2006/entries/logic-informal/