Re: Rosenean theory, Nature article

[Dan Fiscus <***>]

Tim,

Good start on a letter to Nature. Some comments/excerpts:

Tim Gwinn wrote:

> The question that arises here is that - if this is true - then what is
> required of the models for such a system? Clearly, at the least, what is
> required is that the models can reflect this nonfractionability.
>
> However, state-based formalisms (and thus, models expressed in those
> formalisms) do not have within them the capacity to express
> nonfractionability. If material systems are reducible to some canonical
> expression in a state-based formalism, then the imposition of additional
> constraints (such as "power-law" relationships) to a state-based description
> do not help the situation at all. Another way to state this is to say that
> there is nothing in a state-based description that can entail
> nonfractionable features or properties.(see LI 209-210)

These make sense to me, but I think we'll need more friendly
words, clear definitions in easier words, or good examples to
use for 1) state-based formalisms, 2) canonical expressions,
3) constraints, and 4) entail. Given the level knowledge of
that Ottini I think we need to communicate to a general
audience rather than experts.

> This consideration informs us that other formalisms are required for
> modelling complex systems, and that these formalisms do not take the form of
> additional constraints on existing state-based formalisms. Instead, the
> requirement is to invoke formalisms which inherently have the capacity to
> express the nonfractionable properties, and which are thus incommensurable
> with a state-based description. What this says is not that the 'laws of
> physics' are wrong per se as currently expressed, but rather that the
> assumption that the current laws of physics are exhaustive is wrong, and
> that the quest for a single canonical formal description of all material
> systems is not well-founded.

Other terms to deal with 1) complex, 2) incommensurable,
3) exhaustive, 4) single canonical formal description, 5) material
system, 6) well-founded. I am just flagging these for myself as
a note on what terms/concepts are most important to get
across clearly.

> On another note, I would be highly skeptical of calling the web and
> power-grid systems "complex", insofar as they do not - of their own accord -
> grow and adapt themselves. Without human direction, they are quite inert.

I can see your point, but one potential advantage such systems
provide for revealing/developing/explaining Rosen theory is the
combo of 1) existing science does not seem capable of predicting
their behavior and guiding their management as in for robust and
fault tolerant performance, and 2) to the degree that such
understanding and management issues are involved with network
topology and/or relational systems they ought to be great test
cases for application of Rosen theory.

I don't have a specific proposal or great details on these points, but
I offer that they might be worth exploring. One quick idea that
comes to mind is to compare network topologies between human
(power grid, internet) and natural (ecosystems, neurons, etc.)
systems to see if there are general principles able to show major
similarities or differences.

I'll take your starter points above and try to draft a short letter.
I think we could start with a very short letter just to flag the
issue that based on this Ottini article and its foray into complexity
Nautre and anyone that wants to really work on complexity ought
to know about Rosen and his work/ideas. Then we can present a
few examples such as the ones you treat above. We might even
mention efforts such as current researchers, the lists, websites, etc.
Less of a science publication and more an informal letter to the
editor. Sound OK? (After that I think we should do a more formal
publication.)

So far it is you, me and Judith for authors...

Dan