Tuesday, December 20, 2016

Is genetics still metaphysical? Part IV: The 'specifisticity' of life

I had not intended a 4th post in the series (part 1 is here) about whether genetics is 'metaphysical', and what that might mean in our search to understand biological causation.  However, I just listened to a very good** discussion of the Copernican 'revolution' in understanding the movement we observe in the skies, and its relevance to scientific inference generally.  That led me to write this follow-up.

Copernicus showed that viewing the Sun rather than the Earth as the center of the known universe seemed more natural and in some ways easier than the prior Ptolemaic system.  Of course, the issues were deeper because they were relevant to theological explanations of existence, in which the Earth was the center of God's creation.

The Ptolemaic system that dated back to the classical era had led to the discovery that planets did not orbit the earth in simple circles; instead, to predict their position one had to invoke 'epicycles', short occasional small circular detours in planetary paths.  By placing the Sun at the center of celestial motions, the Copernican system was somewhat simpler, though it had its own equivalent of epicycles and wasn't entirely more obvious.  The important follow-up that did simplify things was Kepler's showing that the orbits were ellipses rather than circles.  This was much easier and more natural, even if also imperfect**, and indeed Tycho Brahe had shown ways to make Earth-centered planetary predictions that basically matched those of Copernicus, by having the moon and planets circle the Earth but the Sun circle the Earth, in a sense salvaging a geocentric cosmos.

The BBC program's discussion concerned the fact that at some point, truth becomes as much philosophical  as it is scientific. Matters of computational convenience were not necessarily about what view is 'true'.  Indeed, we may never be able to know absolute truth, and indeed that concept may be inherently philosophical, as pointed out on the program by the philosopher of science Massimo Pigliucci.  Thus, none of the competing planetary computational approaches necessarily need be 'true': each had its own level of complexity and limits of accuracy.  These days science takes the heliocentric system as obviously 'true', and in particular, that theological assertions that the Earth is the center of the universe is wrong.

More relevant to this post, however, I think that at some point every science becomes an axiomatic system, built upon terms and relationships that are defined but not examined or examinable in any further depth.  In classical geometry, for example, such terms include 'point' or 'line', and I think that currently, the 'electron' is like that: it is not clearly a 'thing' nor a 'wave' and if it is 'energy' that, too, is something whose effect can be defined observationally but whose essence is not further explored.  In a sense, the ultimate nature of these fundamentals is 'metaphysical' that is, is above the physical, something whose 'true' reality or essence we cannot see, at least in our current stage of a science.

Mathematics is an axiomatic system, based on entities like numbers and relationships like equality, addition, and so on.  We deduce things from these primary entities or prove their relative properties, but we can go a very long way in physics and cosmology using mathematics and the various principles and assumptions that we currently make.   We don't need to ask what an electron or electromagnetic wave 'is' in order to make precise use of it in building a model of physical existence, whether or not some day we will be able to probe such things more deeply.

If this is the basis of science, isn't the same true of the biological equivalent of what are currently considered its primary 'things'--that is, 'genes'?  I think not.

Genetics and metaphysics in a comparative context
In the previous posts in this series, I asked whether genetics (and by extension, evolution) was still essentially metaphysical.  Since the term 'gene' (and its historical antecedents) was defined by observable facts, such as patterns of inheritance, it was assumed to be a real entity, even if nothing was known or, at the time, knowable about its essence.  That was in the realm of speculation but, like electrons or geometrical points perhaps, it was at least assumed to be a kind of real 'thing' because it seemed to behave is if it were.  But what kind of thing was purely speculative and at best indirectly supported by evidence of its putative causal result, the directly observable traits of organisms.

As we tried to explain in this series, the idea of a 'gene' historically grew out of Mendel's work with carefully selected traits in peas, which he chose specifically as being useful for plant improvement. The resulting metaphoric or metaphysical notion of life's primary causal element (at the time, perhaps, literally comparable to atomic 'elements'--indicated by Mendel's use of that word) led to a perhaps unprecedentedly productive research strategy, which yielded the discovery of RNA and DNA itself, that particular regions of DNA code for the structure of proteins.

However, to a much under-appreciated extent, that very success itself led us to discover that no 'gene'--no bit of DNA--acts on its own, that only sometimes is a given coding stretch used to produce a given protein because of context-specific variable exon usage; that the code only works when other types of DNA-based codes are used to control the expression of the gene; that the code is sometimes altered after transcription; and much, much more. Indeed, it seems very possible that important or even fundamental aspects of what DNA does remain unknown.

We've dealt with some of these issues in many other posts.  In particular, the 'gene' is currently not a fundamental concept comparable to 'point', 'electron', 'square root', and so on.  It is not something that is a fundamental, irreducible causal element whose internal nature or identity cannot be probed more deeply. Unlike points and electrons, not all genes are identical; indeed no two genes are. A gene is not a primary causal unit in the same sense.  Earlier in this series we quoted a new suggested definition of 'gene' that makes this point by inadvertently being so useless that it might as well not have been suggested.

Today, except for some restricted, usually vague and often conveniently self-serving situations, we do not have a good concept of what a gene 'is'--or even if life is based on some such concept.  First, we define genes in terms of biological 'functions', that is, some purportedly causative outcomes that we like to measure, like the production of skin or eyes, or intelligence, or disease.  One thing that is relevant and does seem very clear is that aspects of DNA have functions that are fundamentally due to interaction, or even that interaction is all that DNA function is about.  The word ('gene') no longer unambiguously refers to a clear kind of basic element: its referents have to be defined ad hoc. The same purported unit has different causative aspects in different experimental and natural contexts.  It is also not a proper fundamental unit because in today's usages a 'gene' has internal components (in DNA sequence, modification by other chemicals etc.).  In some selective situations the word has utility (e.g., referring to the BRCA1 gene in a causative context related to breast cancer), but even that is typically limited, and worse, limited to a typically unknown and/or variable extent. These statements reflect the success of the science to date, but also show how deep our need for conceptual reform really is.

Life is specifisitic 
It was probably understatement on our part to have ended the 3rd element of this series by saying that the 'gene' is still, and perhaps essentially, a metaphysical concept.  That's because it's not really clear, yet, whether it even is a coherent concept, much less whether it, or any fundamental unit of causation applies to life in the way such concepts (may) apply to physics and chemistry.  In a sense, perhaps similar to the views of Einstein and Ernst Mach that we touched on in part I of this series, the fundamental units of life are relationships, not things.  This may be similar to the issue in physics about when or whether or how reality is made of waves or things.  But there would be more, because of the fact that unlike physics, the fundamental units are not replicable the way electrons are.

Perhaps a different unique-context-centered causative concept is needed for understanding the essential nature of life: a  fundamental 'specifismology'.  At this stage of our knowledge, in particular in relation to prediction, and also in the political economy of contemporary science, we are far from that level of of understanding. But we've said that many times before!

It is in the nature of science that how and when we'll get a break of deeper insight, no one can say.

**You would probably enjoy this podcast or online-stream.  It is from the BBC Radio 4 science program series, Discovery.  The discussion of Copernicus and his work is very interesting, but as a science program, the Beeb committed at least a minor error in a kind of de facto assumption that there is some sort of underlying truth in this aspect of cosmology and the history concerns which measurement approach is most accurate or easiest.  That is essentially a Newtonian view, of space as having an absolute reference frame for which one tries to find the easiest computational system. That's what 'the Solar System' means: we place the Sun at the origin of 3-dimensional linear coordinates.  However, in post-Einstein relativistic times we now accept that there is no reference frame from which to decide which view is 'true'; heliocentric models are simply more practicably useful.  Whether the relativistic nature of reference frames applies to biology in a seriously relevant way is a separate, but interesting question.


Chris said...

Dear Ken,

I've found this series of posts to be quite thought-provoking. To return to the topic of Part I, I was wondering what you think a theory of biology should (realistically) aspire to predict?


Ken Weiss said...

I will think about your ver important question--'the' question that should be asked--and write another post because (1) I can't make a quick simple answer in a comment-reply, and (2) it can be somewhat longer and give me a chance to think more about it. But (3) if I had the answer I'd say what it was! I think this is a key question, for any science: how can we best know if we are at a stage where some serious ('revolutionary') change in thinking is appropriate or needed, and how on earth can we stimulate the research environment to make it more likely to happen sooner? I think there are no 'answers' but that thinking about the question might help someone(s) find it/them. I make no grandiose claims for my own insights about this, of course!

I'll post my thinking in a part V of the series in day or so, and thanks for your comment that provokes it.