#JuarreroBook Chapter 6 Part 1 There's a lot in this chapter, and some of it I find hard to understand. So I'd like to split things up. We are now on context dependent constraints, the nature of which is to "take conditions away from independence" The...
-
#JuarreroBook Chapter 6 Part 1 There's a lot in this chapter, and some of it I find hard to understand. So I'd like to split things up. We are now on context dependent constraints, the nature of which is to "take conditions away from independence"
The ch. outlines "three examples of the emergence of long-range correlations generated in virtue of context-dependent constraints. The first serves as a metaphor of phase transitions. The second illustrates inter-dependent dynamics among oscillators. The third is the textbook case of self- organizing, nonlinear, and far from equilibrium processes in the natural world. All three show how context-dependent constraints, operating against a backdrop established by context-independent constraints, weave global forms of order".
The examples are: 1. the phase transition of a random graph with sufficiently many links that it moves to connectedness
-
synchronising pendulum clocks on a shelf
-
convection patterns such as Bernard cells
What do people make of these examples? do they involve "transitions to a new possibility space" (pg. 70)? Do the constraints seem 'real' (metaphysically)? Is 'constraint satisfaction' as seen in these examples "an important form of "causality" that has been systematically ignored by modern science and philosophy" (pg. 72) ?
(comment on Context Changes Everything, p. 72)
-
-
-
@uh @UlrikeHahn some more info on those examples themselves. I think the text on the buttons example in the book is garbled. Kauffmann just refers to the standard phase transition in Erdos Renyi random graphs as you add more and more links - pg 30 here https://avalonlibrary.net/ebooks/Stuart%20Kauffman%20-%20At%20Home%20in%20the%20Universe%20The%20Search%20for%20the%20Laws%20of%20Self%20Organization%20and%20Complexity.pdf
so that’s what I’m taking the example to be….
-
@uh a good description (and figure) of the way network structure changes as a function of the average degree (number of links each node has), can be found in the Barabasi book, pg 16, https://barabasi.com/f/624.pdf
-
@uh and on the pendulum clocks example I found this nice article https://physicsworld.com/a/the-secret-of-the-synchronized-pendulums/
-
-
@uh @[email protected] @[email protected]
Indeed this is a much longer chapter! I stopped at p. 73 for now.
I share your concern about the ontology here: Juarrero talks constantly about constraints 'doing' things, and even being a form of causality (though she doesn't say how). But her examples suggest that what she calls constraints are just ways of describing patterns that appear when certain entities interact with each other in specific organised ways. This impression is reinforced by the apparently circular treatment of context-dependent constraints on p.70: they are characterised by appeal to constrained interactions...
But then, rather than being something ontologically additional that does things, constraints are just ways of talking about features of such patterns, which are in their turn constituted by the familiar kinds of causal interactions between entities. So, nothing ontologically new, just, at most, new-ish alternative explanatory tools.
(This connects, I think, to the Deacon vs Heil discussion we had last week. Hat tip: @[email protected], @[email protected]) (Also connected, perhaps, to the traditional criticisms of Dynamical Systems Theory, when applied to CogSci, in terms of it offering just redescriptions, rather than explanations)
Though I promised not to complain any more about the general obscurity and lack of rigour, I can't help myself. Juarrero talks about enabling constraints 'locking in' (??) information and energy flow to 'real-world traits/characteristics' (what others are there? what do such constraints do that has to do with real-world traits rather than... what exactly?). Then on p.71 she seems to give 3 non-equivalent characterisations of enabling constraints in the space of two paragraphs...
-
@[email protected]
Intrigued! Do you have a good ref or pointer on the redescription complaint (of dyn sys in cog sci)? -
@[email protected] @NicoleCRust @[email protected] @awaisaftab @uh I’m more unsure than you about the new examples, Dimitri, and I’m still confused by the first one at the moment. It feels natural(ish?) to think of a real world group of objects and a process establishing connectivity between them and say something like ‘they became fully connected because the number of links exceeded ln(N), and that that feels not just like a description but like an explanation as something bound to happen. 1/n
-
@[email protected] @NicoleCRust @[email protected] @awaisaftab @uh in that way it feels constraint like? But what is it? I was thinking about logical constraints in the chapter on context independent constraints and that had me confused too. If one is thinking in terms of possibility spaces and what narrows them down (‘constrains’ them) then both logical constraints and the ln(N) threshold (whatever it is) feel as real as the notion of a possibility space itself (but what is the latter?)
-
@[email protected] @NicoleCRust @[email protected] @awaisaftab @uh 3/n but I’m also confused by the phase transition threshold in other ways: how is it like or unlike saying ‘if I have 2 objects and add one I now have >2’ because the transition threshold is neither strict (network doesn’t have to be conn.) or exact (it wouldn’t be wrong, in some sense, to say the network was connected because it had 1.5 ln(N) links…), and am I confusing a description with the constraint itself?….am lost in the weeds
-
@[email protected] @NicoleCRust @[email protected] @dsmith @awaisaftab @uh 4/n and very last thought: it’s a very long time ago that I read the Kauffmann book, but I remember the point of it as being that evolution alone is insufficient to explain life and life forms as we see them, and that further constraints on generating structure are required to render it as anything other than wildly improbable - cue emergence and complexity. So he presumably thought example shows meaningful constraints in action?
-
@[email protected] @NicoleCRust @[email protected] @dsmith @uh ..5/5 which links to Dylan’s point?
-
@[email protected] @UlrikeHahn @NicoleCRust @[email protected] @awaisaftab @uh
I like the term “constraint satisfaction” as it captures how conditions/experience can move an individual toward a regulated state (not unlike the synchronizing clocks).
cf: Blair describes causal role of SES, cortisone, etc., in emotional reactivity, school readiness.
Blair, C. (2002). School readiness: Integrating cognition and emotion in a neurobio conceptualization of children's functioning at school entry. The Amer Psych, 57, 111-127.
-
@dsmith @[email protected] @NicoleCRust @[email protected] @awaisaftab @uh Dylan, I like the notion of constraint satisfaction in many contexts, but I think my specific problem here with these three examples is identifying what exactly *is* the constraint and what is its nature.
-
@UlrikeHahn @[email protected] @NicoleCRust @[email protected] @awaisaftab @uh
I think I understand this difficulty. Not easy to describe a constraint... What caused your window to blow in, a specific wind or the entire storm? Any attempt to explain needs to account the big-picture storm, but then the explanation smacks of description.
Solution?.. I think the value of group discussion is that each of you brings diverse experience that can help all to "pan out" via concrete examples.
The Blair article is brilliant, btw.
-
@UlrikeHahn @[email protected] @NicoleCRust @[email protected] @awaisaftab @uh
Back to the clocks... Any one of the clocks cannot continue to do just what it was doing because the other clocks create conditions (constraints) that keep it from doing so. That is, the other clocks are "canalizing" the behaviour of the clock in question.
-
@dsmith @[email protected] @NicoleCRust @[email protected] @awaisaftab @uh
or is the resonator (the shelf) the constraint?
or is the fact that you cannot maintain the interfering vibrations the constraint?
and isn't the presence of the resonator also a context?
or is the context the thing that provides the physical manifestation of the constraint?
?????
-
@NicoleCRust @[email protected] @UlrikeHahn @awaisaftab @uh This recent overview paper summarises a bit of that discussion and gives some of the key references on the first paragraph or so of section 3
https://compass.onlinelibrary.wiley.com/doi/10.1111/phc3.12695
-
@UlrikeHahn @dsmith @[email protected] @NicoleCRust @awaisaftab @uh @dsmith @[email protected] @UlrikeHahn @NicoleCRust @awaisaftab @uh The worry raised by critics of DST in CogSci is that talk of constraints, etc., is at best a shorthand for referring to complex causal relations, and only the latter do the real causal work.
To be clear, I don't endorse that view, at least not in general, as I think that pluralism about forms of explanation is the way to go.
-
@UlrikeHahn @dsmith @[email protected] @NicoleCRust @awaisaftab @uh However, pluralism does not mean anything goes, and we need to try and figure out which explanations are appropriate for what and for what purposes.
DST explanations may be mere redescriptions in some cases, but may be good explanations in others. They need not exclude mechanistic explanations. On the contrary, they may complement each other, with the latter providing an explanation for why those dynamics arise, etc.