The Critical Rationalist                       Vol. 02  No. 03
ISSN: 1393-3809                                    15-Sep-1997

3 Reductionist NightmaresGame Trees and Ant Country

1 Introduction

2 Components of the Problem

(2) Ian Stewart and I raised these questions in different contexts in three papers and a popular-science book (Cohen & Stewart 1991a, Cohen & Stewart 1991b, Cohen & Stewart 1994, Stewart & Cohen 1994); the enthusiastic reception of the book, The Collapse of Chaos, gives me some authority to raise these issues here. A more sophisticated version of the argument is in our next book (Figments of Reality, 1997).

(3) There are, I think, four components of the problem:

The adequacy of any theories at all, an issue addressed well by Popper (e.g. 1963b--but not as naively as I wish to). Stewart and I are impressed that most of the Science we "know" is revealed as "Lies-to-Children" (Stewart & Cohen 1994, Stewart & Cohen 1997) as soon as it is examined closely (by this we meant that these matters are simplified to the point where they're not in any sense "true", but are the standard myth for educational purposes--they fit with other myths, perhaps). Examples: rainbows produced by raindrops acting as tiny prisms (all exactly pointing in the right direction for your eyes? And for your friend 20 metres to your left?); the Bernoulli story of lower pressure over the curved upper surface of airplane wings "pulling upwards" (so that if airplanes try to fly upside down they dive into the ground?); the anatomy of female genital tracts as portrayed in diagrams (always shown as having large cavities in uterus and vagina - the cavities are nominal, 10 microlitres in uterus, perhaps 10x that in vagina! See Cohen 1996). The intellectual context of testing any theory must be suspect, therefore, if it is tested against these educational constructs rather than against deeper appreciations. As Editor-in-Chief of Speculations in Science and Technology I receive many submissions which have precisely this structure: a speculation among fairy-tale (text-book) suppositions. Then I find that many of my own speculations, as well as those I find in Nature and Science, also have this naive setting. Rejecting most initial hypotheses because a disproof is "obvious" sounds persuasive, but not if the touchstone is itself bogus.
Is there one physics? This relates to post-Modern stances, to questions about the "Theory of Everything" and its utility if invented (found?). If there are many equally-congruent webs of theory at the physical "level" (perhaps even non-overlapping) then our parochial starting-with-the-electron one, with its successively deeper levels of fundamental particles, is digging its roots down into a progressively exclusive view, not generalizing but limiting our ability to extend intellectual frontiers. As we extend the zoo of particles, our commitment to this classification is reinforced (and the balancing of mathematical equations, rather than disproof, is the task at hand). This is a biological stance, perhaps inappropriate to the mathematically-tied theories of physics (what Stewart and I called the "Sherlock Holmes Stories", each self-consistent but not universal in application).
Tactics of disproof commonly have a reductionist basis: internal-to-the-theory predictions are set up for test. But Stewart and I showed that much, perhaps most, of apparently reductionist scientific theory depends for its argument on contextual elements (for example, the Gas Laws depend on Volume, Pressure, Temperature, all contextually determined; gas molecules don't know what volume they're in). Ceteris paribus arguments, on which most disproofs depend, specifically factor out context. Arguments in molecular genetics, as well as most classical physics and chemistry, are susceptible to this criticism. Results can only be interpreted within the ceteris paribus experimental design (this is close to, but more parochial than, Kuhn's paradigm argument and dangerously close to Feyerabend's criticism of scientific argument in general). "Normal science" must be blind to context, but context is vital for locating the problems under consideration, and for describing the phase space of possibilities in which an understanding of the problem is sought. A theoretical physicist can only find (invent) a new particle, to match the equations with new particle-collision results, if (s)he knows the limits of possible properties of particles within present paradigm(s); an evolutionary biologist can only invent (discover) a palaeontological phylogeny if (s)he has in mind all the ways in which the group might have evolved. I can only claim to understand the evolution of life on Earth if I can begin to answer "What might, and might not, happen if we ran the system again on Earth--or what might we find on another aqueous planet?" (Cohen 1991, Cohen 1993, Stewart & Cohen 1997).
The converse of 3 above: emergent properties frequently show independence of antecedent causes or more "basic" components. Stewart and I borrowed the word "fungibility" from the legal lexicon to describe such cases: a bridge can be made from rope, from steel girders, or from concrete and can exhibit the same special properties, so the components are fungible. And the reductionist proposition that the properties of the bridge can be argued from--depend on--the properties of its components and their interactions simply does not apply. The emergent properties are contextual, the rest of the Universe sees the bridge as a link between the island and the mainland whatever it's made of (even a tunnel, "made of nothing", might be equivalent). Popper got involved directly in this problem. Natural selection as a basis for organic evolution had a strange, perhaps un-disprovable status for him, at least partly because it could not be linked inexorably to nice reductionist DNA molecular biology. Predictions, particularly disprovable within the same theories, could not be made because whatever the genetic substrate, evolution would still occur in much the same way: on Antares 3, aquatic carnivores would be streamlined like pike, dolphins or ichthyosaurs; terrestrial carnivores would have their eyes in front and herbivores would have them at the sides (cf. Popper 1978, though he did not put it in this way)

3 Reductionist NightmaresGame Trees and Ant Country

1 Introduction

The Critical Rationalist                       Vol. 02  No. 03
ISSN: 1393-3809                                    15-Sep-1997

tcr-editors@www.eeng.dcu.ie