Zachary > Recent Status Updates

A prioristic views of method are incompatible with the values of science. Investigations should be performed in the real world, not in the idealized conceptual worlds of our own construction. Models that limit an agent to too few degrees of freedom are often misleading. Logically necessary consequences from such models are irrelevant. Models that hold in principle but not in practice should not hold our attention.

Overly strict normative idealizations are compromising to human rationality & scientific practice. Reasonable goal setting has been shown to result in better performance. Yet our implicit models are more demanding than plausible. Idealizing human rationality ignores the constraints of our biology. What we really do is make mistakes & learn by studying how things break down. Why not develop our heuristics around this?

The dominant reductionist science is flawed by idealizations about rational decision which requires unrealistic degrees of knowledge and inferencial power. In reality we craft heuristic tools that are neither axiomatic nor algorithmic in origin but refined with experience. Rather than formalisms and foundations, science should rely on contextual adaptability that is tolerant of and engineered to learn from error.

Classical deterministic science posits that every event has a cause and the same cause leads to the same effect. However, many effects, especially in biology, seem to result from causal indeterminacy (randomness) or even no cause at all. Determinacy and indeterminacy are often detectable in the same system although at different hierarchical levels.

We'd expect natural selection to drive biological systems from chaotic to more orderly states while genetic drift, being entirely random, would do the opposite. However, drift typically results in genetic homozygotization which is a decease in genetic chaos on the population level while natural selection results in heterozygotization which is an increases in genetic chaos.

There are three types of randomness. Two of them, quantum randomness and molecular stochasticity, are well known for all forms of matter. The third, randomness of allele transmission, is limited to eukaryotic species that use a meiotic process and is a strong source of variability.

Deterministic descriptions of gene expression are not possible. Different gene expressions can be observed in genetically identical cells exposed to the same environmental conditions. This "noise" is related to aspects of transcription and translation. Epigenetics being dynamic and stochastic are also a major source of genetic noise. Gene networks, however, have the ability to contain excessive noise.

The ability of recombination to generate abundant random genetic combinations allows for populations to explore a wide range of variation and meet adaptive challenges. Irregularities in recombination are often also the cause of reproductive isolation and speciation.

There is comprehensive data explaining HOW spontaneous mutations occur but WHY they occur is unclear. In biological processes the line between the deterministic and the stochastic depends on the size of the data set. Deterministic predictions are possible for short spans of time but the long term shows too much stochasticity for outcomes to be accurately predicted.

Some great achievements of inquiry have established that there are limits to the "knowable" in sciences, mathematics, and logic. Proper realization of these important advancements is an indication of a mature science and can be particularly useful in the rapidly developing areas of genetics.

Phenotypic plasticity operates on all levels of biological organization. The same stablized genotype can have multiple phenotypes meaning that plasticity can be genetically derived. Newly stabilized plastic phenotypes can also be "assimilated" into newly mutated genotypes meaning that genotypes can be phenotypically derived. Thus, the "genotype-first"/"phenotype-first" debate is a false dichotomy.

Exposure to slowly changing environments can allow for innovative adaptations while maintaining easy access to previous adaptations sometimes even without the need for reverted genotypic change. Developmental, physiochemical, selective, and genetic constraints result from processes that form phenotypes from genotypes and robustness in the relevant genotype spaces.

Both gene duplication which increases genetic robustness and recombination which can result in large jumps through genotype space have the potential to facilitate evolutionary innovation.

Neutral genes that neighbour a fitness relevant genotype network can become non-neutral with change in environmental context. These "molecular exaptations" show that neutralism and selectionism are not incompatible but equally important aspects of evolution.

Metabolic processes are not linear but highly reticulated networks that can share up to 30% of their reactions with multiple unrelated metabolic processes. Since a huge number of variant gene networks can have the same phenotypic expression, innovative metabolic processes can emerge through a variety of mutations even while preserving the origins process.

Not sure. Acknowledges population genetics yet seems to ignore its implications. Often unclear when the author is speaking of experimental findings or just theoretical science.