Evidence for Synergies section: The discussion of the hallmark quality of synergies needs some work. Once assembled, it is a relatively autonomous organization of degrees of freedom that take care of themselves. Components adjust to (a) mutual fluctuations and (b) fluctuations in the external force field (more generically, the synergy's environment) so as to sustain integrity of function. Retaining stability is, for a synergy, the retaining of functional integrity.
The speech data are not doing enough conceptual work. That the same parts are involved in different functions in different ways needs emphasizing. Likewise, that the adjustments are task specific. Even better: if true, then it is clear that the assembling and disassembling of synergies cannot be hard suggesting that the stabilities in question are marginal in a peculiar way. They are simultaneously stable and unstable, ready to become something else at an instant. In Iberall's terms, synergic states are equienergetic. It's equally easy to go from anyone to any other. In Kelso's preferred terminology, synergic dynamics are metastable.
I think the Bernstein quotation should include a paraphrase of: "not chains of details but structures that are differentiated into details."
Synergies learned or innate section: I wonder if synergies as such are things that can be developed or learned. Is ever a self-organized thing learned? Lessons from Thelen's work are that "whatever the course of brain development, behavioral expression is entirely context dependent” “…[t]he organic components and the context are equally causal and privileged” (Thelen & Smith, 1994). Similar lessons can be distilled from Adolph's work. How to assemble function-specific organizations might be developed, might be learned.