1/N I've never used Twitter to present a new scientific view. But these ideas span evolutionary & developmental biology and philosophy, so it won't be published in any widely read journal, and even if it is, it will take several years to happen, and few will see it. So here goes.

2/N 1. Cells are autonomous agents. They strive to maintain their lives and have dynamical and conditional interactions with their environments, from which they draw nutrition and energy. They reproduce and flourish in niches that are constructed in part by their own activities.

3/N 2. Although it seems that there are no processes inside cells that violate laws of chemistry and physics, their internal organization is not understood in conventional terms. The evolutionary steps that led to their emergence as autonomous agents are thus very obscure.

4/N 3. Aggregates of cells, however, are not autonomous agents. The evolution of multicellular organisms like animals and complex plants may have begun as associations of cells, but simple aggregates are not autonomous agents. Multilevel selection theory (MST) is a Darwinian

5/N to understanding how associations of cells can become agents, but it is often only a collection of Just-So stories.
4. There are physical steps that promote autonomy: the emergence of the egg-stage of development in proto-animals ensured that cells in a cluster were clonal

6/N that is, genetically identical. Also, the evolution of an egg container (shell, zona pellucida) kept even cells that resulted from non-cleavage division stayed together, and were thereby genetically uniform.
5. But clusters of genetically equivalent cells do not have agency

that is different in quality from the cells that compose them. For this to happen they have to develop novel feature that permit them to interact with the environment in new ways, and create new niches (Gibson's affordances). There are Darwinian scenarios for this based on

[stopping tweet numbering] competition, relative fitness, and so forth. But these are not needed either. There are self-organizing processes that shape and pattern cell masses based on physics of the middle scale - liquid and liquid crystalline properties, phase separation,

biochemical oscillation and synchronization, Turing processes, that lead to many different kinds of morphological novelty - multiple tissue layers, body cavities, segments and other repeated structures such as appendages and digits, that arise from inherent material properties

and (other than whether they persist in a given population of not once they arise) do not need natural selection to explain their occurrence.
6. But multicellular organisms with novel morphological attributes and affordances are not yet autonomous agents like the cells that

constitute them. In order for them to become autonomous more evolution is needed, so that they become biologically distinct from other such associations with a similar origin
7. These further rounds of evolution need not change the outward appearance of the organisms,

but mainly reinforce the reliability and reproducibility of their developmental routines, making them less "plastic," i.e., subject to external influences, and more "programmatic." The result of this "developmental system drift" is morphological stasis and enhanced stability

of generative processes. The multicellular organisms evolve into "natural kinds," fulfilling criteria laid out by philosophers for being such, though in these cases the "natural kindness" is an evolved property (which it is for the chemical elements as well. Ultimately,

such organisms become fully autonomous agents (the animals and plants that co-inhabit the Earth with us) and come to fulfill Kant's definition of "organized beings" exhibiting "natural purposes," entities that are both the causes and effects of themselves. [The end, for now.]