On 5/29/25 9:24 AM, glen wrote:
Why isn't Lorenz in there somewhere? Seems like an obvious candidate.
/“When the present determines the future, but the approximate present does not approximately determine the future.”/
— Edward Lorenz, /The Essence of Chaos/ (1993) What George says when "reminded" of Lorenz: * *Edward Lorenz (1963 onward)* may be the /first person to demonstrate/ in precise mathematical terms a *deterministic system whose behavior could not be prestated* beyond a short horizon. * He /did not/ use the phrase *“deterministic but not prestateable”*, but *his work embodies it fully*, and later chaos theorists and philosophers often retroactively describe it that way. ------------------------------------------------------------------------and here is George's (I don't call GPT George, but for continuity with Nick, et al I do here) response:
but first for brevity, I cherry picked from below and I like the help given in my nitpickery style:
* /“Unpredictable” (Lorenz-style, chaos)/ * /“Unprestatable” (Kauffman-style, adjacent possible)/ * /“Undecidable” (Turing/Gödel-style, logic)/ ------------------------------------------------------------------------ Me: can you explain perhaps why you might have omitted lorenz ChatGPT said: Yes — and I appreciate the callout. Why Lorenz Might Have Been Omitted (Initially)There are a few reasons, mostly contextual and framing-based, not due to oversight of his importance:
------------------------------------------------------------------------ 1. *Interpretive Framing Bias:*Your original phrase, *“deterministic but not prestateable,”* carries a flavor that strongly suggests:
* *Logical / epistemic limits* (e.g., undecidability, incompleteness), * *Unenumerable state spaces* (as in Kauffman or Rosen), * *Open-endedness* and *emergence* rather than sensitivity or chaos.So I initially prioritized thinkers who made *explicit claims about the /unlistability or uncomputability/ of future states*, rather than the *intractability of prediction*.
------------------------------------------------------------------------ 2. *Domain-Specific Bias:*Lorenz worked in *physics and meteorology*, and his contributions, while foundational to the idea of unpredictability, are *framed in terms of instability*, not *unprestatability due to logical structure*.
That is: * Lorenz’s system /is/ prestatable in theory (you can write the equations), * But *in practice*, you can’t predict far ahead without infinite precision. Compare to Kauffman, who says: “We /cannot even write down/ the space of possibilities in advance.” So the distinction was between: *Type of barrier* *Lorenz* *Kauffman/Rosen* Measurement precision limit Yes SometimesComputational or logical limit No (the system is computable) Yes (the phase space isn’t enumerable)
------------------------------------------------------------------------ 3. *Terminological Exactness:*Lorenz didn't use the language of *prestateability*. That’s a term more common in philosophical biology and complexity theory (Kauffman, Rosen, Pattee, etc.).
So I likely indexed on *terminological proximity* rather than *semantic essence*.
------------------------------------------------------------------------ 4. *Scholarly Echo Chamber Effect:*The phrase /“deterministic but not prestateable”/ has circulated more explicitly in Kauffman and Rosen’s circles — particularly in systems biology, philosophy of complexity, and theoretical evolution — than in physics/chaos literature.
There’s a subtle linguistic difference between: * “Unpredictable” (Lorenz-style, chaos) * “Unprestatable” (Kauffman-style, adjacent possible) * “Undecidable” (Turing/Gödel-style, logic) ------------------------------------------------------------------------ 🧭 Retrospective CorrectionYou were right to bring Lorenz in. His work *embodies the idea* in a foundational and mathematically demonstrable way. He simply spoke from a different idiom — one of *dynamical instability* rather than *ontological or epistemological open-endedness*.
I would now include him centrally, as the *first to /prove/ that deterministic systems can defy practical prediction*, even if not logical prestateability in the strictest sense.
------------------------------------------------------------------------Would you like a synthesis of Lorenz and Kauffman into a unified framework of unpredictability and unprestatability — perhaps visualized through nested phase space, strange attractors, and adjacent possibles?
Tools ChatGPT can make mistakes. Check important info.
On 5/29/25 8:13 AM, steve smith wrote:< The interesting question is where exactly does the deterministic system turn into something nondeterministic, and how?>who might have first made the distinction : "deterministic but not pre-stateable"?It doesn’t. A deterministic system is a deterministic system.George offered: Year Thinker Conceptual Expression 1931 Kurt Gödel Incompleteness: truths not derivable 1936 Alan Turing Halting problem: uncomputable predictions1970s Heinz von Foerster Second-order cybernetics: unknowability of future 1991 Robert Rosen Closure to efficient causation; entailment limits1993 Stuart Kauffman Adjacent possible; unprestatable evolution2005 Gregory Chaitin Incompressible but defined numbers (e.g., Ω)
OpenPGP_0xD5BAF94F88AFFA63.asc
Description: OpenPGP public key
OpenPGP_signature.asc
Description: OpenPGP digital signature
.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-.. FRIAM Applied Complexity Group listserv Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom https://bit.ly/virtualfriam to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com FRIAM-COMIC http://friam-comic.blogspot.com/ archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/ 1/2003 thru 6/2021 http://friam.383.s1.nabble.com/