30
Emergence Explained What’s right and what’s wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science The Aerospace Corp. Computer Systems Division Computer Science & Technology Subdivision Information Technology Department That’s setting the expectation level pretty high. Can he really pull it off? That’s setting the expectation level pretty high. Can he really pull it off?

Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

Embed Size (px)

Citation preview

Page 1: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

Emergence ExplainedWhat’s rig

ht and what’s wrong about reductio

nism

Russ Abbott

California State University, Los AngelesDepartment of Computer Science

The Aerospace Corp.Computer Systems DivisionComputer Science & Technology SubdivisionInformation Technology Department

That’s setting the expectation level pretty high. Can

he really pull it off?

That’s setting the expectation level pretty high. Can

he really pull it off?

Page 2: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

2

Emergence the holy grail of complex system computing

How macroscopic behavior arises from microscopic behavior.

Yaneer Bar-Yam http://necsi.org/guide/concepts/emergence.html

Emergent entities (properties or substances) ‘arise’ out of more fundamental entities and yet are ‘novel’ or ‘irreducible’ with respect to them.Stanford Encyclopedia of Philosophy http://plato.stanford.edu/entries/properties-emergent/

Macro from micro. Multiscale phenomena.

The ‘scare’ quotes identify

problematic areas.

Page 3: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

3

Emergence the holy grail of complex system computing

John Holland, Emergence: From Chaos to Order

It is unlikely that a topic as complicated as emergence will submit meekly to a concise definition, and I have no such definition to offer.

The father of genetic algorithms. One of the founders of the Santa Fe Institute.

Page 4: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

4

Cosma Shalizihttp://cscs.umich.edu/~crshalizi/reviews/holland-on-emergence/

Someplace … where quantum field theory meets general relativity and atoms and void merge into one another, we may take “the rules of the game” to be given.

But the rest of the observable, exploitable order in the universe

benzene molecules, PV = nRT, snowflakes, cyclonic storms, kittens, cats, young love, middle-aged remorse, financial euphoria accompanied with acute gullibility, prevaricating candidates for public office, tapeworms, jet-lag, and unfolding cherry blossoms

Where do all these regularities come from?

Call this emergence if you like.

It’s a fine-sounding word, and brings to mind southwestern creation myths in an oddly apt way.

Physicist (& poet)

Page 5: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

5

Steven Weinberg

Reductionism may or may not be a good guide for a program of weather forecasting, but it provides the necessary insight that there are no autonomous laws of weather that are logically independent of the principles of physics.

[T]he reductionist view emphasizes that the weather behaves the way it does because of the general principles of aerodynamics, radiation flow, and so on (as well as historical accidents like the size and orbit of the earth), but in order to predict the weather tomorrow it may be more useful to think about cold fronts and thunderstorms. “Reductionism Redux,” in Cornwell, J. (ed), Nature's Imagination: The Frontiers of Scientific Vision, Oxford University Press, 1995

Weinberg does not deny others the right to drink the orange juice of emergent phenomena — the idea that a system with many mutually interacting parts can lead to novel macroscopic behaviour — but he asserts vigorously his right to drink the gin of reductionism. http://physicsweb.org/articles/review/15/4/1/1 review of Facing Up

Real or just conceptual

conveniences?

No autonomous forces; but perhaps autonomous laws.

The ultimate reductionist.

Page 6: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

6

An originator of and outspoken defender of functionalism

“Special Sciences; Still Autonomous after All These Years,” Philosophical Perspectives, 1998.

Well, I admit that I don’t know why. I don’t even know how to think about why. I expect to figure out why there is anything except physics the day before I figure out why there is anything at all.

Damn near everything we know about the world suggests that unimaginably complicated to-ings and fro-ings of bits and pieces at the extreme micro-level manage somehow to converge on stable macro-level properties.

[T]he somehow really is entirely mysterious. Why is there anything except physics?

Jerrold FodorMountains are made of all sorts of stuff. [Yet] generalizations about mountains-as-such … continue to serve geology in good stead.

Autonomous laws of mountains?

Page 7: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

7

Three types of emergence Static (petty reductionism)

a house, cloth, hardness, e.g., of a diamond, pressure, temperature. Dynamic (grand reductionism): most agent-based models,

market phenomena, (un)intended consequences. Entity-environment interactions; stigmergic effects.

Strong: new forces of nature, e.g., vitalism: “life” from “lifeless” chemicals. Magic; non-reductionist; what Weinberg doesn’t like.

Quantum entanglement violates whole/part (petty) reductionism.

Reductionism doesn’t outlaw new fundamental forces, e.g., dark energy.

But don’t do this too often.Tangent

Page 8: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

8

Reductionism vs. strong emergenceForce: any influence [that] tends to change [the motion of] an object.http://hyperphysics.phy-astr.gsu.edu/hbase/hframe.html

Reductionism: the only forces in the universe are the n fundamental forces, for some small n.

An absolutely stark choice.

Strong emergence: new forces of nature may appear at many levels of emergence.

What are the forces that make things happen?

Link structure to time

Page 9: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

9

Surprise is not the pointTangent

A resident of a primitive society would be quite surprised to see someone assemble a car from its components and then get in and drive off.

Page 10: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

10

Is emergence (just) epiphenomenal? Epiphenomenon: a secondary phenomenon

that is a by-product of another phenomenon. http://wordnet.princeton.edu/

“[Species diversification] is an epiphenomenon of the basic components of replicating DNA, mutations, geography, and limiting resources.” http://lifesci.rutgers.edu/~heylab/sconcept/conclusions.html

The elliptical shape of the earth’s orbit is an epiphenomenon of the force of gravity acting on a body in motion.

Imagine the earth as an “agent” that follows the (local) rules of inertia and gravitational attraction.

A very strange idea.

Just one last philosophical

concept.

Page 11: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

11

GameOfLife.jar

The Game of LifeTry to take it seriously as a (very) simple agent-based model.

Built on a rectangular grid. A totalistic two-dimensional cellular automaton.

An agent (cell) is either alive or dead. (Can’t move.) Rules [analogous to the basic forces of nature;

Fredkin & Zuse, Wolfram, http://www.math.usf.edu/~eclark/ANKOS_zuse_fredkin_thesis.html]

The 8 surrounding agents are an agent’s neighbors. A live agent with two or three live neighbors stays

alive; otherwise it dies. A dead agent with exactly three live neighbors is

(miraculously) (re)born and becomes alive. Article: http://www.math.com/students/wonders/life/life.html (bad applet?) Applet: http://www.ibiblio.org/lifepatterns/

• Agent-based model.• Universe as CA.• (Later) programming

platform.

Page 12: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

12

Gliders (waves of births and deaths? epidemics?) are (amazing) epiphenomena of the Game of Life rules—whose only(!) consequences are to switch agents/cells on and off.

Gliders (and other epiphenomena) are causally powerless. A glider does not change how the rules operate or which cells will be

switched on and off. A glider doesn’t “go to an agent and turn it on.” A Game of Life run will proceed in exactly the same way whether one

notices the gliders or not. A very reductionist stance. Agents don’t “notice” gliders—any more than gliders “notice” agents.

Gliders exemplify dynamic emergence. Gliders are not generated explicitly. There is no glider algorithm. Gliders are not visible in the rules. Gliders are generated stigmergically.

Epiphenomenal gliders

All software is stigmergically epiphenomenal over the

instruction execution cycle, which is stigmergically epiphenomenal over

electron flows.

The rules are the only forces!

Page 13: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

13

Amazing as they are, gliders are also trivial. Once we know how to produce a glider, it’s

simple to make them. Can build a library of Game of Life patterns.

By suitably arranging these patterns, one can simulate a Turing Machine. Paul Rendell. http://rendell.server.org.uk/gol/tmdetails.htm

Game of Life Programming Platform

and their interaction APIs.

You don't do

this with your

models.

A second level of emergence. Again, no algorithm; just stigmergy.

Also stigmergic.Very fragile.

Page 14: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

14

It’s the design that matters

To prove that a Game-of-Life simulation of a Turing Machine works, must reason about epiphenomenal interactions among epiphenomenal patterns. Must show that the design: Simulates a Turing Machine. (Reify the design; treat it as real.) Can be implemented on a Game-of-Life platform.

What does it mean for epiphenomenal gliders and other epiphenomenal patterns to simulate a Turing Machine?

Reductionism: the patterns don’t really interact! Functionalism: it’s (only!) the

design that matters. (Set it free!) Multiple possible implementations.

Disengage the model from the implementation.

Levels of abstraction? Layered hierarchies?

Very useful—but not real.

Page 15: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

15

A Game of Life anthropologist Find a lost tribe of Game of Life runs “in the wild.” Get a grant to study them. Figure out the Game of Life rules.

Model even explains gliders as emergents! Publish results.

But the rules do not explain the functionality of a Turing Machine simulation — which is logically independent of the rules.

Let’s pretend.

Recall Weinberg: no autonomous weather laws.

Lots of autonomous Turing Machines laws!

Page 16: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

16

Emergence: non-reductive regularity A regularity: Game-of-Life run simulating a Turing Machine. Can explain every step by appeal to the Game-of-Life rules.

The rules are the only forces at work. Yet those rules don’t explain:

What the system is doing functionally. How its design is accomplishing it.

Both Reductionism: only fundamental forces exist. Emergence (the-whole-is-more-than-just-the-sum-of-its-parts):

constructive/creative functionality. Not reducible to rules. Emergence is any non-reductive functionality/regularity.

Getting epiphenomena to do real work.

Recall Shalizi’s notion of emergence as all the order in the universe beyond quarks/strings, etc.

Page 17: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

17

Functionality for an EnvironmentToo artificial? Too designer-oriented?

(General) evolution’s “blind watchmaker”

Something survives if its design/functionality works in its environment.

Functionality understood on its own level.

Intelligent designer/architect Theoretical considerations for a

Turing Machine. Real considerations about a

real environment for an embedded system. (Systems engineering/architecting.)

In both cases concerned about interaction with an environment. The functionality is “outward looking” toward the environment. The “downward looking” implementation isn’t the point.

Two sources of design.

Page 18: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

18

A hurricane as a far-from-equilibrium

entity Generates heat internally — by condensation rather than combustion.

“Consumes fuel” — different from Prigogine’s dissipative structures.

Energy produced powers its self-perpetuating processes.

Design: a hurricane has a design; one can talk about how it works.

Fitness: persists (self-perpetuating) as long as the environment within

which it finds itself provides adequate resources given its design.

Moist warm surface air — what it “eats.”

Cool dryer condensation area in upper atmosphere — “waste product

removal.”

http://earthobservatory.nasa.gov/Library/Hurricanes/

Self-perpetuating far-from-equilibrium processes define regions of reduced entropy: you, me, Theseus’s ship, …What’s real?

Page 19: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

19

Dissipative structures vs. self-perpetuating entities

Dissipative structuresSelf-perpetuating

entities

Pure epiphenomena, e.g., 2-chamber example

Has functional “design,” e.g., Turing Machine

Artificial boundaries Self-created boundaries

Externally maintained energy gradient

Imports, stores, distributes, and selectively uses energy

Tangent

What is the simplest example of a (proto-)biological system

with these qualities?

Page 20: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

20

Why are there regularities at all? Why do un-designed regularities (functionality fragments) such

as gliders occur? Gliders are not designed. They have no environment to satisfy.

Why is there number theory, e.g., Fermat’s last theorem?

Why is the earth’s orbit a simple ellipse? “How can it be that mathematics, being after all a product of

human thought independent of experience, is so admirably adapted to the objects of reality?” A. Einstein.

“The Unreasonable Effectiveness of Mathematics in the Natural Sciences,” E. Wigner.Communications in Pure and Applied Mathematics, vol. 13, No. I (February 1960). John Wiley & Sons, http://www.dartmouth.edu/~matc/MathDrama/reading/Wigner.html

Tangent

Page 21: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

21

Modeling problems:the difficulty of looking downward

Strict reductionism implies that it is impossible to find a non-arbitrary base level for models. What are we leaving out that might matter?

Use Morse code to transmit messages on encrypted lines.

No good models of biological arms races. Combatants exploit and/or disrupt or otherwise foil

each other’s processes. Insects vs. plants: bark, bark boring, toxin, anti-toxin, … .

Geckos use the Van der Waals “force” to climb.

Models of computer security or terrorism will always be incomplete.

Can only model unimaginative enemies.

Universe is not segmented into disjoint layers.

epiphenomenalABM w/GP

Page 22: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

22

Don’t know how to build models that can notice emergent processes and characterize their interactions. We don’t know what we aren’t noticing. We/they can use our commercial airline system to deliver

mail/bombs. (Food distribution → Botulism toxin.)

Model gravity as an agent-based system. Ask system to find equation of earth’s orbit. Once told what to look for, system can find ellipse. But it won’t notice the yearly cycle of the seasons — even

though it is similarly emergent.

Modeling problems:the difficulty of looking upward

Models of computer security or terrorism will always be incomplete.

Can only model unimaginative enemies.

Exploit an existing process.

Page 23: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

23

The emergence of complexity

Anderson: The ability to reduce everything to

simple fundamental laws [does not imply] the

ability to start from those laws and reconstruct

the universe. “More is Different,” Science, 1972.

A non-reductive, creative, contingent, stigmergic, historical process. What existing

processes can I exploit to accomplish my goals?

Image fetching → hit count,Google map hacks, GPS. Real market for virtual assets.James Burke’s Connections.

Early member of the Santa Fe Institute.

The universe as bricolage

What designers do for a living.

Designs must be implementable not reducible.

Condensed matter

physics.

Page 24: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

24

Contact and paper [email protected]

Emergence Explained (.pdf): http://abbott.calstatela.edu/PapersAndTalks/Emergence Explained.pdf

Go to http://abbott.calstatela.edu/ and click Emergence Explained.

Search Google for “Emergence Explained”

Page 25: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

25

FoundationsSummary slides

Page 26: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

26

“Downward causation” Using epiphenomena it’s easy to compute:

the trajectory of a nail on the rim of a rolling wheel (Sperry). the force applied by one billiard ball to another. whether a cell in a Game-of-Life grid will ever be turned on by a glider. the position of the earth in elliptical orbit around the sun.

All are far more complex to compute in terms of fundamental forces. There is no fundamental force corresponding to kinetic energy. The gravitational n-body problem has no closed form. For the glider it’s the difference between partial and total decidability.

When a Game-of-Life Turing Machine turns on a tape cell (or a software animation turns on a phosphor), it is not turning on a grid cell. It is turning on a tape cell that is implemented as a grid cell. Although it may seem like downward causation, it isn’t. The Game-of-Life rules control everything!

Tangent

Page 27: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

27

In Computer Science we assume that one can specify a Turing Machine, a Finite State Automaton, or a piece of software, and it will do its thing — for free.

In the real world one needs energy to drive processes. To run real software in the real world requires a real computer.

Thermodynamic computation:nihil ex nihilo

Page 28: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

28

Categories of emergent entitiesUses energy? Intrinsically bound Extrinsically bound

At or moving toward an equilibrium

Mass-basedatomic nucleus

Attractor-baseda river

Far from equilibrium

Process-basedbacteria, me, hurricane,

Aerospace Corp.

Maintenance-basedTheseus’s ship,

satellite systems

In our examples, gliders are epiphenomena. In the real world, entities, which represent persistent

regions of reduced entropy, are still emergent but real. Have properties that do not apply to their components and

that depend on how they are held together.

The most interesting

entities.

Page 29: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

29

Who/what “Change agent” Environment

Pheromone-driven systems Ants in ant colony Environment

Agent-based modeling Agents Other agents, etc.

Varela & Maturana; autopoiesis Biological entity Environment

Conrad Game of Life rules Game of Life grid

Turing TM head (an FSA) TM tape

Von Neumann CPU instruction cycle Memory/software

System engineering; use case analysis Embedded system Larger context

Wegner; interactive computing Computer Users/environment

Fredkin-Zuse, Wolfram; nature as CAForces of nature—as outside the objects, e.g., virtual particles

Initial (or current) conditions

Stigmergy: controlling a process by changing its environment.

Computing/Nature as weak emergence

Page 30: Emergence Explained Whats right and whats wrong about reductionism Russ Abbott California State University, Los Angeles Department of Computer Science

30

Emergence is non-reductive (creative) regularity/functionality. Designers produce (usually nominal) emergence all the time. Weak emergence typically involves indirect interaction through an

environment: stigmergy. Designs must be implementable but not reducible.

Higher level entities (regions of reduced entropy ) are real — not epiphenomenal and not just conceptual conveniences.

Interactions are virtual — and are vulnerable at every level. Strictly reductionist. It's important that the interactions accomplish what they are supposed to

accomplish within the design. But ultimately, they are epiphenomenal. The only real action is at the lowest level.

Leads to two fundamental modeling problems: difficulty in looking upward and downward.

Summary