Link Search Menu Expand Document

Reading Notes

PHIL 560

Table of contents
  1. “Two Concepts of Intertheoretic Reduction”, Thomas Nickles
  2. “The Reduction of Theories” from The Structure of Science, Earnest Nagel
  3. “Special Sciences (or: The Disunity of Science as a Working Hypothesis)”, Jerry Fodor
  4. “Multiple Realizations”, Lawrence Shapiro
  5. “Making Sense of Emergence”, Jaegwon Kim
  6. “Kim on Emergence”, Sydney Shoemaker
  7. Metaphysical Emergence, Jessica M. Wilson
    1. Chapter 1: Key issues and questions
    2. Chapter 2: Two Schemas for Metaphysical Emergence
    3. Chapter 3: The viability of Weak emergence
    4. Chapter 4: The viability of Strong emergence
  8. “The Metaphysics of Realization, Multiple Realizability, and the Special Sciences”, Carl Gillett
  9. “Thinking About Mechanisms”, Peter Machamer, Lindley Darden, and Carl Craver
  10. “Aggregativity: Reductive Heuristics for Finding Emergence”, William C. Wimsatt
  11. “Computability Theory and Ontological Emergence”, Jon Cogburn and Mark Silcov
  12. “Determinism, predictability and open-ended evolution: lessons from computational emergence”, Philippe Huneman
  13. “The Architecture of Complexity”, Herbert A. Simon
  14. Approaches to Reduction, Kenneth F. Schaffner
  15. “Is Anything Ever New? Considering Emergence”, James P. Crutchfield
  16. “Alternative Views of Complexity”, Herbert Simon
  17. “More is Different”, P.W. Anderson
  18. “The Status of Emergence”, Paul Henly
  19. “Emergence, Closure and Inter-level Causation in Biological Systems”, Mossio et al.

“Two Concepts of Intertheoretic Reduction”, Thomas Nickles

  • Two different kinds of reduction: Nagel’s (explanation of one theory by another) and a new proposed one (a “varied collection of intertheoretic relations”)

Two Concepts of Reduction

  • Does classical mechanics (CM) reduce to special theory of relativity (STR) or vice versa?
  • It is in fact not clear.
  • The more general STR reduces to the less general CM in the limit of low velocities
  • So what exactly does ‘reduce to’ mean?
  • Reduction 1: “an increase in the overall efficiency of a conceptual scheme” – “elimination, trimming down, consolidation”
  • But what about ‘historical reduction’ and predecessors?
  • “Reduction 2”: taking a limit on something, extracting a property
  • Consolidating logically superfluous postulates is part of Reduction 1, derivational reduction. The reduced theory must be a logical consequence of the reducing theory. (!!!)
  • For Reduction 2, not all reduction is explanation. Guided by special cases or relations
  • Reduction 2 paradigms are ‘domain-preserving’ reductions of successor theories to predecessors
  • Reduction may provide confirmation of a new theory and establish its importance
  • There can many ways to reduce-2, whereas only one to reduce-1
  • For reduction-2, the reduced theory does not need to be logically compatible with the reducing theory
  • Reduction-2 is not so sentiive to meaning change, overcomes Nagel’s objections

Nagel on Reduction-1

  • Distinction of homogenous and heterogenous reduction
    • Homogenous: descriptive vocab of a reduced theory is a subset of the descriptive vocab of the reducing theory
    • Heterogenous: reduced theory contains descriptive terms not in the reducing theory
    • Nagel wants to turn heterogenous reduction into homogenous reduction
    • T1 reduces T2 when T2 is derivable from T1 (T2 is reduced)
  • Domain-combining reductions are heterogenous
  • Historically distinct domains involve different descriptive vocabs
  • “Meaning change” objection: passing from one theory system to another involves an extensive conceptual shift – e.g. ‘mass’ in CM vs. STR
    • Fallacy of equivocation?
  • Can we apply bridge laws? Would turn homogenous reductions into heterogenous ones
  • Feyerabend: reduced and reducing theoreis are almost always logically incompatible; reductions are only approximate
  • Partial reduction? Derivational reduction?
  • Nagel’s critics forget that philosophical models are deliberately idealized
  • Many claimed reduction-1s fail because of logical incompatibilities.
  • Nagel;’s condition are not necessar fro all types of intertheoretic reduction
  • However, correlation is not enough for reduction

Identificatory Reduction

  • Sklar and Schnaffer: importance of theoretical identifiation in reduction
  • Nagel’s corollary laws are too weak

“The Reduction of Theories” from The Structure of Science, Earnest Nagel

  • Classical mechanics is no longer regarded as the universal / fundamental science of nature
  • Many physicists are skeptical if you can have a theory which integrates all natural science into a common set of principles.
  • Recurrent feature of modern science: absorbption by or reduction to a more inclusive theory
  • Findings of physics are often incompatible with ‘common sense’
  • The ‘bankruptcy’ of classical physics
  • What are the conditions for reduction?

I. The Reduction of Thermodynamics to Statistical Mechanics

  • Reduction – explanation o a theory established in one area of inquiry by another theory
  • “Secondary science” – reducing theory
  • “Primary science” reduced theory
  • A theory applying to a specific set of things may expand its domain.
    • Mechanics moved from point-masses to deformable bodies
    • Extension from Galileo’s laws to Newtonian mechanics and gravitational theory – no new concepts
  • Reductions of this type establish deductive relaitons between homogenous vocabularies – these are accepted as normal phases
  • A second kind of reduction: assimilation into a distinct theory
    • Primary science wipes out familiar distinctions as spurious, maintains that some prima facie different traits are really identical, etc.
    • Often happens when the secondary science is macropscopic and the primary science is a microscopic constitution for macroscopic processes.
    • Change in knowledge of what ‘temperature’ is – from a feeling to the temperature to the mean kinetic energy of molecules
    • What is the genuine reality?
  • Is temperature an ‘emergent’ trait, manifesting at higher levels of nature but not at lower levels?
  • “Heterogenous” reductions
  • Example of reduction with thermodynamics with mechanics (statistical mechanics)
    1. Thermal phenomena goes back to modern times to Galileo
    2. Laws have systematic interrelations
    3. Thermodynamics: uses concepts, distinctions, general laws also employed in mechanics, but also adds notions like temperature, heat, and entropy
    4. How to find a more intimate connection between thermal and mechanical phenomena?
    5. Introduction of a statistical assumption

II. Formal Conditions for Reduction

  • Axioms, special hypotheses, and experimental laws must be available as explicitly formulated statements
  • Four classes of statements under a science \(S\):
    1. Theoretical postulates of \(S\) and the theorems derivable from them, as well as coordinating definitions
    2. Experimental laws
    3. Observation statements
    4. Borrowed laws
  • Every statement of a science can be understood as a linguistic structure
    • There will always be a set of primitive expressions
    • Observable expressions / primitives
    • Theoretical expressions / primitives
  • Primary and secondary sciences often have expressions with the same meanings
  • Reduction happens when the experimental laws of a secondary science are logical consequences of theoretical assumptions of a primary science.
  • Two formal conditions for reduction of a primary science to a primary science
    1. Assumptions must be introduced into the primary science to explain what “A” (from the secondary science) means
    2. All laws of the secondary science (including ‘A”) must be logically derivable from the primary theory
  • Three possibilities for linkages:
    1. Logical connections – ‘A’ in secondary science is entailed by ‘B’ in primary science
    2. Linkages are conventions, experimental significance is assigned
    3. Linkages are factual or material
  • To reduce thermodynamics to mechanics, you have to connect temperature to mean kinetic energy of gas molecules
    • To reduce thermodynamics to mechanics, you have to assert that it is proportional to the mean kinetic energy of gas molecules
    • “Similar syntactical structural” derivable, but there is a different form

III. Nonformal Conditions for Reduction

  • Two formal conditions do not distinguish trivial from notewrothy scientific achievements
    • The sole requirement for reduction being logical deducation can be satisfied very easily
    • Extrapolation of theory required
  • The reduction of thermodynamics to the kinetic theory of gases gives us much more than the deduction of Boyle-Charles’ law – the reduction is a significant scientific accomplishment
    • We can deduce from the primary theory more complex assumptions
  • It is not enough for the previously established laws of a secondary science to be represented in the primary discipline – the theory must also give useful suggestions for developing the secondary science
  • A number of experimental laws seem to have an invariant – so the reduction of thermodynamics to kinetic theory also integrates laws so that directly relevant evidence for one law can be indirect evidence for others
  • Perhaps little to no new knowledge or increased power is gained by reducing one science to another – it in fact might be counterproductive
  • Sciences have a history
  • Reducability of one science to another is contingent upon the specific theory employed by a discipline at some time.
  • Psychology an “autonomous discipline” because “a headache is not a rearrangemen tof particles in the cranium”, or the mind and physical processes.
  • Mistaken notion: the question of if a science is reducible to another is based on inspecting properites or natures of things rather than logical consequences
    • Deducing certain properties from other ‘natures’ – turns a logical / empricial question into a speculative and irresolvable one. (365)
  • “Laplace in error” – the Divine Intelligence
  • Reduction does not wipeout distinctions recognized by the secondary science.
    • Reduction is not a process of elimination
    • Headaches explained by physical, chemical, physiological headaches.
    • States conditions under which a determinate psychological phenomenon takes place.

IV. The Doctrine of Emergence

  • Doctrine of ‘emergent evolution’ or ‘holism’
  • Emergence of properties at ‘higher’ levels of organization which are not predictable from properties at lower levels
  • The interaction of elements may enter into relations with each other which are different from existing relations.
  • “Complete knowledge” on the lements of \(O\): we know all the propertie sht elements possess in isolation, and all the properties exhibited by complexes other than \(O\).
  • Two cases
    1. Possible to predict if elements will occur in a relation, and what objects will be formed
    2. There is a property such that it is impossible to predict if the element stand in known relation, if an object will be formed. This object is an ‘emergent object’ and the property is the ‘emergent property’
  • “most of the chemical and physical properties of water have little known connection, either quanitative or qualitative, with those of oxygen and hydrogen”
  • Howwever, Nagel thinks this is not true. It is impossible to deduce properties of water from properties of hydrogen alone or of oxygen alone – but it is also impossible to deduce the behavior of a clock merely from the organization of its constituent parts
    • It is not properties, but statements which can be deduced
    • Statemetns about complex wholes can be deduced from statements about constitutents if premises contain a suitable theory – you have to give the terms to be able ot articulate something
  • A given property is clearly emergent if it is possessed relative to one theory but not in another theory.
  • Emergence: stating certain logical facts about formal relations between statements
  • We cannot appeal to ‘inherent natures’
  • Atomic theory of matter revived by Dalton to account systematiclaly for a range of chemical facts
    • Dalton posited few properties of atoms and could not explain many chemical transformations, but was modified to account for them.
    • Development of a notion of the ‘intrinsic’ nature of atoms – there is a shifting notion of the ‘nature’ of the atom in each case, from Democritus to Dalton to pysiochemcial theory
  • We should not think that different atomic theories represent progress towards a fixed set of submicroscopic objects – the notion that atoms have ‘inehrent natures’ independent of any particular theory is false
  • It is not possible to assert what can be predicted from the ‘natures’ of atomic particles independent of any particular atomic theory
  • Properties of compounds can be predicted from current electronic theory of composition of atoms
  • Propertie smay be emergent wrt some theory but not another.
  • It is also wrong to maintain that we are “baptizing our ignorance” by characterizing a trait as emergent. Our ignorance or incomplete knowledge of the natures of atoms is irrelevant to the issue at stake.
  • The doctrine of emergence is correct as a logical relation between statements.
  • Opposition to mechanical reductionists: if some properties are emergent, their occurrence is inexplicable in mechanical terms. Emergence sets limits to the science of mechanics
  • “Predicted from” – deductive or inductive?
  • Qualities, structures, modes of behavior come into existence from time to time
  • What is a ‘different’ property or arrangement?
  • It is not just that our knowledge of the structures of events change, but that the structures change themselves
  • Creation of new substances by chemists which have never existed before: is this a ‘realization of potentialities’ or something more.

V. Wholes, Sums, and Organic Unities

  • “Organic unities” – not merely aggregates but “the whole is more than the sum of its parts”
  • What is the cognitive worth of these notions of ‘whole’, ‘sum’, etc.?
  • ‘Whole’ – spatial extension, ‘part’ – spatially included in it
  • ‘Whole”: temporal period
  • ‘Whole’: – aggregate of elements
  • ‘Whole’: patterns of relations
  • Sums are not so easily understood
  • “The different actions of a chemical compound will never be found to be the sums of actions of its separate parts”
  • Relative to classical kinetic theory the thermal properties of solids are not sums of the properties of their parts, but relative to quantum theory properties are their sums
  • “Organic / functional” wholes: the behavior is not determined by individual elements, but where part-processes are determined by the intrinsic nature of the whole. Parts do not act independently of each other, but are all related.
    • e.g. the distribution of charges on a conductor
  • Piecewise construction vs analysis
  • Additive and nonadditive analysis – particle physics of classical mechanics vs field approach of electrodynamics.
  • Some wholes can be analyzed in this manner, for others it still has to be worked out.

“Special Sciences (or: The Disunity of Science as a Working Hypothesis)”, Jerry Fodor

  • Positivistic philosophy of science: all true theories in the special sciences should reduce to physical theories in the long run
  • Many philosophers who accept reductivism do so to endorse the generality of physics
  • Conflation: “physics is basic science” and “theories in the special science reduce to physical theories”
  • The “unity of science” is much less plausible than the generality of physics
  • “the more the special sciences succeed, the more they ought to disappear”
  • Pyschology and physics?

Part I

  • Reductivism: the special sciences reduce to physics
  • Let \(S_1 x \to S_2 x\) be a law of the special sciences, suggesting that all \(S_1\) situations bring about \(S_2\) situations
  • If we show \(S_1 x \iff P_1 x\) and \(S_2 x \iff P_2 x\), then we can show \(P_1 x \to P_2 x\)
  • The connective \(\to\) must be transitive; psychology can reduce to physics through neurology, biochemistry, etc.
  • If bridge laws are not identity statements, then the new claim that can be made is at most one of causal correlation.
  • The truth of reductivism does not guarantee the generality of physics, since there are some events which are not physical.
  • If bridge laws express event identities, we get “token physicalism”: all events subject to the sciences are physical events.
    • Weaker than ‘materialism’, which says physicalism is true and every event falls under the laws of science. Why bother being a token physicalist without being a materialist?
    • Weaker than ‘type physicalism’: every property in the laws of science is a physical property.
    • Weaker than ‘reductivism’: the conjunction of token physicalism with the assumption of total correspondence between special sciences and physics. Token physicalism does not imply reductivism. Reductivism is a sufficient condition fro token physicalism.
  • Reduction of psychology to neurology entails any event of psychology is identical with an event of neurology.
  • Reductivism is too strong, but token physicalism works

Section II

  • Every science sets forth a taxonomy of events
  • Descriptive vocabulary of theoretical and observation predicates
  • If reductivism is true, then every natural kind is a physical natural kind.
  • The question of if reductivism is too strong is finally an empirical question.
    • Is it really?
    • What about the “science of phenomenology”?
  • It is unlikely that every natural kind corresponds to a physical natural kind because
    • interesting generalizations can be made about events whose physical descriptions have nothing in common
    • special sciences are involved in making these generalizations
  • What is the disjunctive predicate which corresponds to “\(x\) is a monetary exchange \(\iff ...\)”?
    • Why not through a neural model? If we reduce the human mind to physicalism then does not everything which the human mind constitute become reducible?
    • “Psychobiology” or “psychology and brain science” departments are institutionalized gambles that lawful co-extensions can be found
  • Confusing token physicalism with reductivism?
  • There seem to many nomoloical possible systems which satisfy natural kind predicates in psychology but not in neurologial prediates.
  • Psychological reductivism is not equivalent to token physicalism
  • To confirm token physicalism, we do not need to show that neurological countertypes of type-identical psychological events are type-identical, just that theya re identical with respec tto properties determining the kind of psychological event.
  • The classical understanding of the unity of science has misconstrued the goal of scientiifc reduction: not to find natural kind prediates of physics for everything, but rather to explicate the physical mechanisms by which events conform to laws of the special sciences.

Section III

  • Reductivism is too strong for the unity of science; incompatible with special sciences results and more than we need to assume to be token physcialists
  • “the problem all along has been that there is an open empirical possibility that what corresponds to the natural kind predicates of a reduced science may be a heterogenous and unsystematic disjunction of predicates in the reducing science, and we do not want the unity of science to be prejudiced by this possibility”
  • We can’t have “unnatural disjunctions”
  • Laws of the special sciences can have exceptions, but the basic sciences cannot
  • There are special sciences not because of our epistemic relation to the world (this is what the reductivists say), but because of the way the world is put together: not all natural kinds (classes which are important) correpsond to physical natural kinds
  • Special sciences are in the business of stating some true counterfactual generalizations
  • If physics is basic, each of those things should be physical; but the taxonomies of the speical sciences don’t need to reduce to the taxonomy of physics.

“Multiple Realizations”, Lawrence Shapiro

  • Multiple realizability of hihger-level properties in lower-level properties stops theoretical reduction.
  • Putnam introduces multiple realizability to “demolish the hypothesis that pyschological states are reducible to brain states”
  • Jerry Fodor generalizes Putnam, uses the multiple realizability thesis against any attempt to reduce any special science
  • The task of specifying exactly what is meant by MRT

I. An A Priori Argument for Multiple Realizability

  • Putnam: “two systems are fundamentally isomorphic if there is a correspondence between the states of one and the states of the other that preserves functional relations”
    • Functional isomorphism implies MRT
  • MRT: it is physically possible for some undefined class of kinds to have multiple realizations
  • So perhaps MRT must be judged on an empirical basis. But it is articulated a priori.
  • Functional isomorphism doesn’t really imply anything about the truth of MRT

II. A Likelihood Argument for Multiple Realizability

  • Two hypotheses:
    • (TI) Psychology states are type identical to physical-chemical states of the brain
    • (MR): Psychological states are multiply realizable
  • If (TI) is true, we would be surprised if mammals, reptiles, etc. share the same psychological state b/c that means they share the same physical-chemical state.
  • MR makes identity of psychological states across life forms more likely (??? how is this compelling??)

III. Empirical Evidence for Multiple Realizability

  • The argument against physicalism rests upon the empirical likelihood that creatures of different composition and structures, which are in no interesting sense in identical physical states, can nevertheless be in identical psychological states” – Putnam
  • The possibility of machine minds suffices to establish MRT, supposedly
  • Does this all establish MRT?

IV. Evaluating the Empirical Evidence for Multiple Realizability

  • Philosophers tend to associate MRT with mind-brain reduction
  • Multiple realizations count truly as multiple realizations when they differ in causally reelvant properties
  • Showing a kind is multiply realizable takes work
  • No reason to count the brain and replacing each of its neurons with a silicon chip that performs the exact same causally relevant function as multiple realizations

V. A Dilemma for Multiple Realizability

  • Two realizations of a kind \(T\) are in fact different realizations of \(T\) if and only if they differ in some causally relevant property which individuates \(T\)
  • Why is MRT a coherent thesis at all?
  • Why do we call camera eye and the compound eye, eyes?
  • What unifies different realizations of a kind?
  • Circular dilemma: if we say that the realizations are the same because they are the same kind, then we are begging the question. If we say they are the same kind because they are the same realizations, then we are being circular.
  • Realizations which satisfy the difference in causal capacity cirterion may not be of the same kind if there are no interesting laws that unify them

VI. Multiple Realizability and the Special Sciences

  • What is the value of categorizing something as \(P\) if there are no empirical laws involving \(P\)?
  • Let’s re-examine the goals of the special sciences
  • SPeical sciences focus on mechanism: we need to know about optical principles to understand hwy aquatic camera principles need graded spherical lenses
  • We can say nothing more about eyes in general than that they have the function to see
  • MRT is not an obvious thesis
  • Genuinely different realizations cannot have anything in common except theri function
  • There are no laws true of all realizers qua realizers of a functional kind except analytically true laws
  • We should abandon characterizing speical sciences as a search for laws over functional kinds, rather we understand how functional kinds produce capacities which make them interesting
  • Natural kinds – the items which you can make generalizations over
  • Is Fodor limiting himself to the current state of the sciences but trying to make claims about ideal state of science?
  • Science in relation to our epistemic limitations as beings

“Making Sense of Emergence”, Jaegwon Kim

Section I

  • Emergentism: “as systems acquire increasingly higher degrees of organizational complexity they begin to exhibit novel properties that in some sense transcend the properties of their constituent parts, and behave in ways that cannot be predicted on the basis of the laws governing simpler systems”
  • Carl Hempel and Ernest Nagel – confused and incoherent, neovitalism, trivial
  • Emergence refuses to die as an idea
  • Emergentism part of “nonreductive materialism”
  • Has emergentism returned? Yes
  • The core ontology amounts tol old resurgence
  • Two groups of ideas
    • Emergent properties are novel / unpredictable b/c not explainable in terms of the properties of the constituents
    • Causal powers: emergents bring into the world new causal powers on their own

Section II

  • Supposedly, emergent phenomena / properties are not explainable / reductivlely explainable
  • Assume every material object has a unique complete microstructural description
  • Mereological supervenience: systems with identical total microstructural property have other properties in common. ALl properties of a physical system supervene on / determined by its total microstructural prooperty.
  • Supposedly, some supervenient properties are emergent (unpredictable), while the rest are resultant (predictable from a system’s total microstructural property)
    • Is this merely an epistemic distinction?
  • ‘New’, epistemic, or ‘new’, metaphysical (new causal powers)
  • Inductive predictability vs theoretical predictability
    • Even emergent properties are inductiely predictable
    • What is denied is theoretical predictability of \(E\) on the basis of \(M\), which may result from not even having the concept of \(E\)
    • e.g. \(E\) may be a phenomenal property of experiences
    • Where do we begin with designing phenomenal properties as opposed to merely mental properties? Best idea: create a duplicate of our brain

Section III

  • Another idea: unpredictability of emergents = emergence cannot be explained in terms of the properties of the constituents
  • Emergent properties are not reductively explainable in terms of the properties of the constituents
  • Let \(B\) be a domain of properties. Reduction of a property \(E\) to \(B\) requires:
    1. \(E\) must be functionalized (construed as a property defined by causal or nomic relations to other properties in \(B\)): that is, \(E\) means having some property \(P\) in \(B\) such that various causes cause \(P\) to be instantiated, and \(P\) causes various other effects to be instantiated. Multiple realizers for \(E\) are allowed.
    2. Find all realizers of \(E\) in \(B\).
    3. Find a theory that explains how realizers of \(E\) perform the causal task that is constitutive of \(E\).
  • Differs from Nagel’s understanding of intertheoretic reduction, nod iscussion of bridge laws or derivation of laws
  • The “mystery” of consciousness is not dispelled by any Nagelian reduction when taking bridge laws as unexplained primitives
  • Nagelian reduction is consistent with and even entailed by many dualist mind-body theories
  • Putnam and Fodor’s multiple realization arguments are irrelevant to the real issue of mind-body reduction

Section IV

  • \(E\) is a proeprty targeted for reduction
  • \(E\) has been functionalized as the property of having some property \(P\) meeting causal specification \(C\)
  • Why does a system exhibit \(E\) at \(t\)? Having \(E\) is having a property with casual role \(C\) and the system has some property \(Q\) which fills casual role \(C\); \(Q\) fills \(C\)
  • What enables ascent from the reduction base to higher properties is conceptual connections generated by the functionalization of higher properties.
  • Why can we not design new physical devices that exhibit phenomenal consciousness? Brute bridge laws are all we can get to connect phenomenal properties with physical properties, whereas we need to make theoretical predictions (i.e. by functionalizing phenomenal experience).
  • How is a functional model a model of reduction? Reduction is not necessarily elimination. We can have ‘conservative reduction’ vs. eliminative reduction. Conservative reduction requires identities.
  • How to deal with the question of multiple realizability?
    • \(E\) is a legitimate higher-level property irreducible to its realizers \(Q\)
    • \(E\) is the disjunction of its realizers
  • If multiplicity of realizers means anything, they must be casually diverse. Unelss two realizers show causal diversity, there should be no reason to count them as two
  • \(E\) can even be reduced eliminatively as a significant scientific property
  • Functionalization of a property is necessary and sufficient for reduction
    • Explains why reducible properties are predictable and explainable
    • Explains why irreducible properties are neither predictable nor explainable
  • Are there emergent properties?
  • Phenomenal properties of consciousness like qualia are the most promising candidates

Section V

  • Emergence associated with “downward causation”
  • Emergent properties exercise causal powers downward wrt processes at lower levels (from which they emerge)
  • Emergence vs resultance
  • Emergent properties are unpredictable, unexplainable, irreducible
  • What can emergent properties do after they have emrged?
  • New doctrine of emergentism: emergent properties have causal powers on their own, irreducible to the causal powers of basal constituents.
  • Upwards causation and same-level causation seem to presuppose downwards causation
  • To cause anyh property to be instantiated, you must cause the basal conditions from which it arises. (principle of downward causation)

Section VI

  • Is downward causation even a coherent notion?
  • How can higher-level properties causally influence the conditions from which they arise?
  • Downward causation doesn’t have to be strange – complex systems cause change at lower microlevels by virtue of macrolevel properties
  • Some activity or event can be the cause of events involving its own microconstituents, reflexive downward causation
  • Reflexive downward causation is combined with upwards determination

Section VII

  • Downward casuation presents us with no special problems
  • Sperry wants reflexive downward causation
  • Is this an absurdity? Is it coherent at all?
  • A temporal analysis shows you cannot have synchronous reflexive downward causation
  • If it’s diachronic, it doesn’t seem to be important

Section VIII

  • Mental-to-physical causation – the mental occupying a higher emergent level relative to the physical level
  • If an emergent \(M\) emerges from a basal condition \(P\), why can’t \(P\) displace \(M\) as a cause of any putative effect of \(M\)?
  • If emergent properties exist, they are largely useless for the purpose of causal / explanatory theories

“Kim on Emergence”, Sydney Shoemaker

  • Emergence: properties of silver chloride relative to chemical elements silver and chlorine.
  • Elements have latent causal powers
  • Component entities have powers that collectively determine the instantiation of the emergent property
  • If emergentism is false, micro-entities only have causal powers, and macro-entities are just collections of micro-entities. Properties of macro-entities are predictable on the basis of the properties of micro-entities.
  • If emergentism is true, properties of macro-entities are not predictable on the basis of micro-facts. But they are realized in the micro-facts
  • We know micro-facts fix the macro-facts, but emergentists say the micro-facts include the instantiation of micro-latent powers.

Metaphysical Emergence, Jessica M. Wilson

Chapter 1: Key issues and questions

Metaphysical emergence: dependence with autonomy

  • Entities treated by the special sciences: cells, orangs, trees, birgs, lakes, hurricans, humans, etc.
  • Supposedly macro-entities are dependent on configruations of fundamental physical entities
  • Macro-entities inherit matter from micro-configurations
  • Do the speical sciences have ontological and causal autonomy?
  • Emergence: cotemporal material dependence, and ontological and causal autonomy
  • Metaphysical emergence
  • Why can special scientific and macro entities be materially dependent on micro-configurations of fundamental physical entities
  • Broadly empirical motivations.
  • Features of macro-entities are in part functions of features of micro-configruations
  • Basic examples of constitution: everything is made of atoms, quarks and leptons form everything, etc., quantum mechanics.
  • Why should we think that macro-entities are autonomous from micro-configruations?
    • Distinctive special-scientific taxonomy. Provides a prima facie reason
    • Distinctive special-science features. Special entities have distinctive features. Entities with distinct features.
    • Distinctive special-science laws. Special science entities are governed by special science laws. Special science laws seemingly distinctive, causal.
  • Two more reasons for thinking that speical science entities are autonomous
    • Universal properties and behaviors. Many special science entities exhibit features functionally independent from features of underlying micro-configurations
    • Elimination of microphysical degrees of freedom. States of speical science entities are specifiable by stricter fewer degrees of freedom than the state of micro-configurations. (Information)
  • Additional prima facie motivations
    • Perceptual unity of macro-entities
    • Compositional flexibility
    • Proper names and definite descriptions
    • Truth and meaning
  • Final motivation: seemingly free will, appears to be different from the micro-configurations
  • There are many reasons why we should think natural entities are cotemporally materially dependent on micro-configurations but also ontologically and causally autonomous
  • There is metaphysical emergence

Two key questions

  • Two questions
    • What is metaphysical emergence? Is there more than one form of it?
    • Is there actually metaphysical emergence? Are there any problems which minimize the significance of metaphysical emergence?
  • What is metaphysical emergence?
    • Various notions of dependency and autonomy?
    • How do different accounts typically agree?
      • The only substance is material or physical. Target matter for reduction is physics.
      • Some notion of combination of material dependence and autonomy. Emergents are distinct from their dependence bases and distinctively efficacious wrt bases.
      • Emergence of entities can be investigated by attention to emergence of features of entities at issue. If an entity is emergent, it is b/c there is some characteristic feature
        • Features reasoned about via entities
        • Type vs token matters
      • Emergence has correlational or modal implications
        • “Emergent features supervene on base features” with nomological necessity”
      • Natural reality exhibits a layered structure. Emergentism attempts to make sense of this in a way which is compatible with the experience of cotemporal material dependence with autonomy
    • There is a wide variety in terms of accounts of emergence – there are often incompatible interpretations of core notions
    • Emergence has no “settled meaning”, irreconcilable, etc.
  • Is there actually metaphysical emergence? It’s not clear
  • The aim: provide clear, compelling, systematic answers to key questions
  • There are two schemas for metaphysical emergence, physicalism-compatible and not
  • Each form of metaphyiscal emergence is viable

Operative Notions

  • The physical: it’s assumed that all things are ultimately physical. But what does this mean? Characterization of physical entities is now determined a posteriori. Neither current nor future physics can characterize the domain of the physical. But also, physicalists have not totally given reign to physics to determine what is physical. Materialism: mentality is not over the mateiral goings on. “No fundamental mentality.” Physical entities can be treated as basically compositionally basic
  • The individuation of levels: natural to think of emergent entities as ‘higher-level’ wrt ‘lower level’ going-ons. Do levels hav eontologically singificance or is reality single-layered? What combinations of features should be taken to exist at some level of scientific reality? How should levels be individuated?
    • Lightweight combination – mereological sums, uncontroversially aggregative operations. These are ontologically “lightweight” operations
    • Law-consequence approach – appeals to scientific laws. Laws are metaphysical, encode rules about features. Laws governing entities characteristic of a level can explain the level. Advantage: differentiation is not defined but follows from the laws. Causal conseuqences of \(L\)-level laws are placed at \(L\). Does not rule out weak emergence. Laws require certain kinds of information to operate.
  • The fundamental – what does it mean for things to be fundamental?
    • Non-primitivist: fundamentality defined in terms of dependence. Fundamental things are independent metaphysically.
    • Primitivist: fundametnal things are not metaphysically analyzable in other terms, positive or negative
    • We should reject the idea that fundametnal reality needs a relational underpinning. Fundamental goings-on play a role as axioms in a theory. The primitives are the things which make the goings-on which are fundamental.

Causes and Powers

  • “Power” – causal contributions a given feature makes to an entity’s bringing about of an effect. What entities do depends on what they are. Everyone can accept this notion of powers. We do not need to commit to a lot of things.
  • Physical equations are expressed mathematically rather than causally – scientific focus on the prediction of quantities
    • Is “=” casual or not?
    • Is physics incompatible with causation? Probably not.


  • Cartesian caricature of metaphysicians
  • Contemporary metaphysicians endorse an abductive methodology, “inference tot he best explanation” – attention towards various theroetical desiderata
  • Advance in metaphysical understanding indicated by the acceptance of abductive reasoning
  • We shoudl aim to accomodate the appearance of metaphysical emergence
  • Criterion of appropriate accomodation: an account of metaphysical emergence should make natural and realistic sense of appearances of metaphysical emergence
    • show the emergentist how to stay on their own horse
  • Criterion of illuminating accomodation: an adequate account of metaphysical emergence should provide an illuminating basis for accomodating the appearances of metaphysical emergence in a natural fashion.

Chapter 2: Two Schemas for Metaphysical Emergence

  • Problem of higher-level causation challenges the appearances of an emergent structure: higher-level features causally overdetermine effects already brought about by base features
  • We can lob two responses
  • We have reason to think satisfaction of conditions for weak and strong emergence are core and crucial to metaphysical emergence

2.1. The problem of higher-level causation

  • Entities are efficacious in having efficacious features
  • Causation is a relation between spatiotemporally located things going on
  • Kim’s argument: motivates rejection of the premise that special-science features are distinct from base features, motivating reductionism
  • You can avoid overdetermination by denying each of four premises (dependence, reality, efficacy, distinctness) but not in a way which preserves metaphysical emergence
  • So how can you preserve metaphysical emergence?
  • Deny physical causal closure: strong emergentism. Deny that every lower-level physical effect has a sufficeint purely lower-level physical cause.
  • Deny non-overdetermination: weak emergentism. Allow that effects of \(S\) are overdetermined by \(P\)

2.2. Strong emergentism and the New Power Condition

  • Strong emergentists: some special-science features are real, and deny physical causal closure (incompatible with physicalism)
  • British Emergentism: fudnamentally novel power
  • Mill – emergence is ‘heteropathic’ effects of joint causes
  • Novel powers are fundamentally novel
  • Emergent features have fundamentally novel powers, not b base features or associated micro-configurations, irreducible to capacities of features of parts in relation
  • “radical emergence” against physicalism
    • threaten the physical world as a closed causal system
  • New Power Condition: Token feature \(S\) has, on a given occasion, at least one token power not identical with any token power of the token feature \(P\) upon which \(S\) cotemporally materially depends, on that occasion.
  • Strong emergentist position associated with a strategy of response by rejecting closure
  • Only by rejecting closure does the new power condition provide a basis for avoiding overdetermination of higher-level and lower-level effects.
  • If \(S\) satisfies the New Power Condition, then it has the fundamentally novel power to bring about \(P^*\). But \(P^*\) does not have a sufficeint purely lower-level physical cause. \(P\) is not at all a cause of \(P^*\).
  • If \(S\) causes another speical-science feature \(S^*\), \(P\) does not have the power to bring about \(S^*\). We still require physical causal closure being false. If closure is held, \(P\) would have a non-derivative power to cause \(S^*\) by being a sufficient purely lower-level physical cause
  • Strong emergentist responses require “downward” causation”
  • Strong Emergence schema for metaphysical emergence: WHat it is for token feature \(S\) to be Strongly metaphysically emergent from token feature \(P\) on a given occasion is for it to be the case, on that occasion, (i) that \(S\) cotemporally materially depends on \(P\), and (ii) that \(S\) has at least one token power not identical with any token power of \(P\).

2.3. Nonreductive physicalism and the Proper Subset of Powers Condition

  • As physicalists, they cannot reject physical causal closure; instead, they reject non-overdetermination; distinct speical-science and base features can be a sufficient cause of a single effect.
  • “Realization” relation between tokens and types, e.g. Functional realization, constitutive mechanism, mereological realization
  • Overdetermination is unproblematic because the feature is efficacious qua the type of feature it is (the realized feature)
  • Physicalists are committed to every SSF satisfying the Token Identity of Powers Condition: every token power of token feature \(S\) is identical with a token power of teh token feature \(P\) on which \(S\) cotemporally amterially depends
  • Stronger condition, Proper Subset of Powers Condition: token feature \(S\) has a non-empty proper subset of the token powers of the token feature \(P\) on which \(S\) cotemporally materially depends.
  • For functionalists, a causal role is just a collection of powers. Functionally relaized features only have a proper subset of the powers of its realizing features in cases of multipel realizability.
  • A token feature \(S\) satisfying PSPC will be distinct from \(P\). Might \(S\) also be causally autonomous wrt \(P\)? \(S\)’s causal autonomy does not require \(S\) have distinctive power, it can have a plurality of powers.
  • Multiple realizability is a good indicator of when a comparatively abstract ontological and causal joint is in place, implies that the joint does not hinge on multiple realizability
  • There are two ways for a higher-level feature to be distinctively efficacious w.r.t supervening lower-level features. Kim: higher-level feature associated with new power to produce the effect (distinctive efficacy). Or, associated with a distinctive subset of powers relevantly proportional to the effect. Power profile.

2.4. Schemas as core and crucial to metaphysical emergence

  • Attention to the problem of higher-level causation allows us to posit two strategies of response – Strong emergentism and nonreductive physicalism
  • Each provides a condition on the token powers of higher-level features w.r.t. token powers of lower-level features
  • There are limited ways in which cotemporally materially dependent higher-level features can be causally autonomous wrt base features.
    • Feature may have more or less powers
    • Feature may be distinctively efficaious in contributing to more or fewer effects
  • Exhausting all options, since token powers being identical isn’t distinctive.
  • Conditions in the schemas are necessary and sufficient for metaphysical emergence

Chapter 3: The viability of Weak emergence


  • Is PSP a mind-dependent abstraction?
  • Picture theory: represent corresponds to reality; when a predicate applies truly to an object, it designates a property possessed by that object and by every object to which the predicate applies
  • Motivations for positing metaphysically emergent entities are better interpreted as tracking inexact similarities between lower-level entities or features
  • Anti-realism motivated by Kim-concern that higher-level properties are causally excluded by lower-level realizers
  • We can’t deny when the realizer of a Weakly emergent feature might be the cause of a given effect
  • The higher-level features must be distinctively, not necessarily uniquely efficacious
  • Power vs. power profile
  • There can be different power profile tracking differences: laws, causal joints
  • We have not been good reason to resist taking prima facie appearances of higher-level reality seriously

Compatibility with reductionism

  • Satisfaction of weak emergence is compatible with ontological reduction to an other lower physicallya cceptable feature

Compatibility with physical unacceptability

  • A feature satisfying the conditions in the schema might be “over and above” its dependence base feature, rendering \(S\) physically unacceptable


Concluding Remarks

  • Range of objections to Weak emergence: condtions compatible with anti-realism,r eductionism, phsical unacceptability, not necessary. But this admits responses for proponetns of the schema.

Chapter 4: The viability of Strong emergence

Incompatibility with scientific theory of practice

  • The positing of fundamental configruational forces is compatible with scientific practice


  • Strongly emergent features should collapse into the lower-level base features upon which thye depend
  • Each objection admits 1+ responses available to proponents of diverse implementations of the schema
  • Maybe some of the conditions in Strong Emergence need to be tweaked

“The Metaphysics of Realization, Multiple Realizability, and the Special Sciences”, Carl Gillett

  • “Received view” of multiple realizability
  • Shapiro: the mammalian and octopus eye are not examples of multiple realization
  • What is realization?

I. Two Views of the Metaphysics of Realization

  • “Flat” account: defended by Jaegwon Kim, Sydney Shoemaker.
    • A property instance \(X\) realizes a property instance \(Y\) only if \(X, YY\) are instantiated in the same individual.
    • A property instance \(X\) realizes a property instance \(Y\) only if the causal powers individuative of the instance of \(Y\) match causal powers contributed by the instance of \(X\).
  • Realized and realizer properties share the individual in which they are instantiated
  • Dimensional account: realizer and realized properties may be instantiated in same or different individuals.
  • Hilary Putnam does not take the mammaliana nd octopus eye to be a case of multiple realization, but Fodor and Block do.
    • Flat metaphysics: the only properties relevant to realizing a property in some individual are the other properties instantiated in that individual. The individuals instantiating the property of being an eye are virtually identical w.r.t. those properties resulting in instantiating the property of being an eye.
    • Dimensional metaphysics:properties / relations of the constituents of an individual are potential realizers of this individuals’ properties.

II. Understanding the relevance criterion

  • Two corkscrews that differ only in color
  • Such a case does not constitute multiple realization of the property of removing corks.
  • Relevance criterion: instances are only distinct realizations of a property if the difference between properties are those which are individuative of the property being realized.
  • Shapiro assumes a flat metaphysics
  • We do not reach Shapiro’s negative conclusions if we assume a dimensional metaphysics
  • Dimensioned view: properties and relations of constituents can be constituents.
  • There is a wider view of causal role playing and a realizer property
  • Do all proponents of dimensioned metaphysics take all differences of composition to be multiple realization? No – things dont have to contribute to the individuative powers of the property
  • Shapiro is not necessarily a critic of Block and Fodor, unless you assume an account of realization

III. Realization and the Nature of the Special Sciences

  • Shapiro: why is the MRT coherent at all?
  • Commitment to flat metaphysics is necessary for the dilemma to arise
  • Under the dimensioned view, realizer properties do not have to contribute a common set of powers.
  • We should maybe reject the flat view in favor of the dimensioned view

“Thinking About Mechanisms”, Peter Machamer, Lindley Darden, and Carl Craver


  • In many areas of science, a satisfactorye xplanation requires providing a description of a mechanism
  • The practice of science can be the discovery and descritpion of mechanisms


  • Mechanisms explain how a phenomenon comes about
  • Mehcanisms are entities and activities organized such that they are productive of regular changes from start or set-up to finish or termination conditions
  • Descriptions of mechanisms show how termination conditionsa re produced by initialization adn intermediate stages
  • Mechanisms are composed of entities (things that engage in activities) and activities (producers of change)
  • Mechanisms are regular – productive continuity
  • Emphasis on activities in mechanisms – producers of change

Ontic Status of Mechanisms

  • Dualist ontics of mechanism
  • Activities can be identified and individiauted like entities
  • Functions are properties “had by” entities – or maybe they are activities
  • Entities and properties determine activities they can engage in
  • Entities and activities are correlatives, interdependent
  • Activities are types of causes
  • There are few universal laws in neurobiology or molecular biology.
  • Mechanism: series of activities of entities that bring about finish in a regular way.
  • Synapses: chemical transmission
  • Termination conditions
  • These states are privileged (graphs!)
  • Intermediate activities
  • Mechanisms can be hierarchical, nested, etc.

“Aggregativity: Reductive Heuristics for Finding Emergence”, William C. Wimsatt

Reduction and Emergence

  • Traditional reductionists posit that emergence claimsa re nothing more than temporal confessions of ignorance.
  • Giving reductionistic explanations does not deny their importance or make them less emergent; it explains how they are important, ineliminably so
  • Multiple realizability is supposedly a criterion for emergence, but it follows from the existence of compositional levels of organization, is very wide, and not very mysterious.
  • An emergent property is a system property which is dependent upon the mode of organization of the system’s parts.
    • Compatible with reductionism
    • Common
    • Weak for antireductionists, but it’s still powerful.
    • Why might we prefer some decompositionso ver others?
  • Emergence can indicate context-sensitivity of relational parts’ properties to intra-systemic conditions
  • Level and scope switching: can be useful to remove biases because of our perceptual biases
  • “the whole is more than the sum of the parts” – e.g. Huygen, pendulum clocks. There is a “virtual metronome” not present at any of the clocks. This is not antireductionistic: oscillation will center around a true mean. It is also an intuitive case of emergence.
  • We need an analysis of emergence which is consistence with reductionism – hemoglobin molecules, traffic jams, critical mass for fissionable materials, etc.

Conditions of Aggregativity

  • Emergence involves an interdependence of diverse parts, but there are many possible forms of such interaction
  • It is easier to discuss what it means for a system property not to be emergent – a “mere aggregate” of its parts’ properties.
  • Four conditions for aggregativity
    • Intersubstitution or rearrangement of parts – invariance of the system property under operations rearranging parts in the system, e.g. commutativity of composition function
    • Addition or subtraction of parts – quantiative similarity of the system property under addition or subtraction of parts, e.g. inductive definitions
    • Decomposition and reaggregation of parts – invariance of the system under operations involving decomposition and reaggregation, e.g. associativity of composition function
    • A linearity condition – no cooperative or inhibitory interactions among parts of the system
  • For a system property to be truly aggregative, it must meet these conditions fro all possible decompositions of the systems into parts – but this is rarely met. More interesting are when conditions are met for some decompositions but not others.
  • Testing conditions against different ways of decomposing the system is revealing of its organization.

Heuristic and Other Uses of Limited Aggregativity

  • Blindness to assumed constraints
  • Conservation laws of physics, e.g. mass, energy, momentum, net charge are aggregative under all decompositions
  • Maybe anything which is invariant across many transformations must have a major conservation law associated with it
  • Volume is not aggregative if you consdier solvent-solute interactions in chemistry.

Aggregativity, Vulgar Reductionism, and Detecting Organizational Properties

  • Few system properties are aggregative. Emergence therefore is very common.
  • We go get more complex (from aggregative models)
  • Aggregativity is a natural criteria for choosing among decompositiosn: aggregative decompositions are natural compositions, joints of nature.
  • Better grounding for vulgar reductionisms

“Computability Theory and Ontological Emergence”, Jon Cogburn and Mark Silcov


  • Often helpful in metaphysics to reflect on how those principles reflect on claims in logic and mathematics.
  • Induction: define a base class and then inclusion by reference to the base class
  • Conditions can be deflationary or inflationary
  • There is an analog in metaphysics
  • Equations reduce at the limit
  • Why has the inflationary strategy been less successful?
    • Humans don’t know what the relevant base class of properties is
  • Questions concerning ultimate properties can be distinguished from questions concerning emergence. Why should a doctrine of emergence say anything interesting about the base class?
  • Distinguish ontologically emergent properties from those which are merely epistemically emergent by simply assuming that the base class of properties is well-defined and asking which properties are emergent relative to them
  • Should we focus exclusively on the inductive component of the inflationary strategy?


  • “Emergent properties”: properties of objects that are not properties of the individual parts of those objects.
  • “Ontological emergence”: emergent properties that are part of the fundamental furniture of the universe, “the really real ones”.
  • “Epistemically emergent” properties: not really part of the fundamental furniture of the universe, but serve as heuristic devices.
  • Three wasy of distinguishing between epistemic and ontological emergence


  • Barry Smith: epistemically emergent properties are all and only secondary properties, ontologically emergent properties are recognition-independent
  • Secondary properties seem to be dependent on us; primary properties might exist in a world which does not contain thinkers
  • Looking pretty, tasting salty: both dependent upon us
  • What about being morally praiseworthy? What about color?
  • Some secondary properties are also ontologically emergent


  • Ontologically emergent properties have genuine causal powers that are irreducible to the instances of the base properties from which they emergence
  • Popular approach to metaphysics of emergence, e.g. Jaegwon Kim
  • Doesn’t help answering our questions: causality not better understood than emergence
  • Kim’s definition of ontological emergence seems to suggest that all emergent properties are ontologically emergent
  • Deductive-nomological view of emergence faces Nickles’ dilema: if you characterize the difference between epistemic and ontological emergence in terms of irreducibility, then with your formal model of reduction to which ontologically emergent activities fail to conform, the theory vacuously determines all emergent properties to be vacuously emergent.
  • Too much focus on reducibility?


  • Kim’s functional view of emergence
  • Epistemically emergent properties are characterizable inpurely functional terms, whereas ontologically emergent properties fail
  • What is function? – ask philosophy of biologist people
  • Kim does not consider properties that satisfy functional specificability but not relevant physical causal roles (?)
  • Does not differentiate between wildly disjunctive physical realizers of functional states and those which are not


  • Emergent properties denote properties of mereological sums that are not instantiated by their parts
  • An ontolgocially emergent property is an emergent property that is really real
  • Borrowing from computability theory: a property is epistemically emergent iff there is an algorithmic procedure that takes in input observed instantiations of properties of a mereological sum’s parts, and gives an output “1” if the emergent property is instantiated and “0” if not. Ontologically emergent if there is no such procedure.
  • Godel, Turing, Church: the set of computationally specified numerical functions
  • There is no method for associating perceptually distinct colors with specific unique and distinct wavelengths
  • The color of an object is ontologically emergnet upon the physical properties of the boject
  • Daniel Dennet: color properties depend on utterly contingent biological features of organisms
  • Non-trivial explanation of why some properties are best classified as being merely heuristic

“Determinism, predictability and open-ended evolution: lessons from computational emergence”, Philippe Huneman


  • Emergence in terms of computational incompressibility
  • Computational view has the advantage of being “objective”
  • Practice of computer simulations?
  • Are too many things made emergent?

Emergence: the concept and its instances

  • Emergence: novelty, irreducibility, unpredictability, downward causation
  • “Unpredictability” – some kind of derivation of a state of affairs
  • Emergence as irreducibility
  • …but dependent on the explanatory framework of the cognitive subjects
  • Might emergence be an epistemological property and not ontological?
  • Genetic algorithms, cellular automata, etc. are ways to rigorously discuss emergence: “computational emergnece”
  • A computational simulation process is emergent iff its result cannot be reached except by running the whole simulation; one cannot compress the simulation. Supposedly this definition does not rely on our epistemic capacities.
  • Many unpredictable features are not at all interesting
  • Computationally incompressible processes
  • Automata display global regularities not implemented in first design: emergent features exhibit a kind of order irreducible to properties of the parts
  • Weak emergent is obht dependent and autonomous
  • Emergence does not oppose reductionism: the processes take place in the building blocks or the cells completely determine the behavior of the automaton, but a more explanatory set of counterfactual dependencies is produced through the emergent, incomputable process.

Emergence, predictability, and determinism

  • Disentangles determinism and predictability
  • Cellular automata are purely deterministic, yet because processes might be incompressible, they yield unpredictable results.
  • What is predictability?
  • Individual predictions, the trajectory of a micro-element in the system, is not predictable. But macor-predictions can emerge.
  • Trajectories of some agents are not predictable, but you can run several rounds of simulations and show that certain ranges of state parameters always yield the same pattern of behavior.
  • Robust emergence implies the possibility of predictions based on counterfactual dependnecies between global-level entities
  • General classes can be related to ranges of variables.
  • Computational emergence: changing the values of initial states of cells precludes prediction of final states of agents
  • The study of complex systems can be done through design of simulations and comparisons

Robustness analysis, relevant causal factors, and the instantiations of the formal computational concept of emergence

  • Computational emergence might fit simulations, but what about the real world?
  • The system modeled is captured in its causal structure by the model; so when there is emergence in the model, reality itself displays emergent processes/properties.
  • Can you read causal factors off of scientific models?
  • Downward causation and overdetermination doesn’t apply: it’s not two levels

Novelty: a case study about artificial life and open-ended evolution

“The Architecture of Complexity”, Herbert A. Simon

  • Should we try to find common properties among different kinds of complex systems?
  • Given the properteis of parts and their interaction, it is not trivial to infer properties of the whole
  • Hierarchy: one of the central structural schemes

Hierarchic Systems

  • A system composed of interrelated sub-systems
  • “Elementary” units are bsaically pragmatic
  • Examples: business firms, governments, universities, families, tribes; th ecell, organs, systems,atoms & protons, neutrons, electrons; books, episodes, etc.

The Evolution of Complex Systems

  • “Decrease in watch-building time”
  • We can think about these as “stable states”
  • Not all large systems appear hierarchical, e.g. linear polymers
  • Problem solving as natural selection. Problem solving is selective trial and erroir.
  • Notion of heuristics
  • How do we make selections?
    • Stable configurations
    • Previous experience
  • On empires and empire-building: on Alexander’s death, his empire did not crumble to dust, but fragmented into the major subsystems.
  • Complex systems evolve from simple systems much more rapidly if there are stable intermediate forms
  • “Among possible complex forms, hierarchies are the ones that have the time to evolve”

Nearly Decomposable Systems

  • In hierarchic systems, we can distinguish between interactions among subsystems
  • A system is decomposable into subparts; roughly independent behavior
  • Intra-component linkages are stronger than inter-component linkages

The Description of Complexity

  • If you ask a person to draw a complex object, they will drawn in a hierarchic fashion.
  • Comparatively little information is lost by representing systems as hierarchies
  • You can have very simple descriptions of complex systems.
  • Exploiting redundancy in the original structure
  • If a complex structure is completely unredundant, then it is its own simplest description
  • State descriptions vs process descriptions (DNA)
  • Recursion? self-reproducing systems
  • Ontogeny recapitulates phylogeny: solve a complex problem by reducing it to a problem previously solved
  • Hierarchy as central to a nontrivial theory of complex systems
  • Hierarchies have near-decomposability which greatly simplifies their behavior

Approaches to Reduction, Kenneth F. Schaffner


  • What happens when one scientific theory is explained by a theroy from a different branch of science?
  • Theory reduction: the process of explaining one theory by another

Four Reduction Paradigms

  • Nagel-Woodger-Quine (NWQ): direct reduction in which basic terms of one theory are related to basic terms of another; laws of the reduced theory are derivable form the reducing theory
    • e.g. Nagel: reduction of thermodynamics to statistical mechanics
  • Kemeny-Oppenheim (KO): indirect reduction, you obtain identical observable predictions from both theories
    • Lavoisier’s oxidation theory of phlogiston
  • Popper-Feyerabend-Kuhn (PFK): relation of later scientific theories to earlier ones, a later theory is able to explain why the previous one “worked” or to “correct” it.
    • Reduction of Galilean law of free fall to Newtonian mechanics
  • Suppses: find isomorphisms in structures

Formal Presentation of the Paradigms

Scientific Interlude

  • There are many examples of reduction in the literature that we can appeal to
  • In practice, reduction conforms best to PFK
  • Some reductions are only approximate, and some are even impossible; others are entirely oerturned rather than reduced – careful correction
  • Historical evolution FTW
  • A reduction can give us new information about the reduced sicence

The General Reduction Paradigm

  • Reduction occurs iff
    1. All the primitive terms in the corrected secondary theory appear in the primary theory or are associated with them such that
      • you can have a bijection of individuals in the theories
      • all effective predicates of the secondary theory are associated with a sentence in the primary theory
      • all reduction functions ahve empirical support
    2. The secondary theory is derivable from the primary theory when conjoined with reduction functions
    3. The corrected secondary theory provides more accurate experimentally verifiable predictions than the original secondary theory
    4. The secondary theory should be explicable by the primary theory
    5. Relations between the corrected secondary theory and the primary theory should be one of “strong analogy”
  • It is possible to onstruct a Suppes reduction if it is possible to construct a NWQ reduction.
  • Reduction is a scientific fact, not simple, but also not very recalcitrant that you cannot have a general logic of reduction.

“Is Anything Ever New? Considering Emergence”, James P. Crutchfield

  • “Some of the most engaging and perplexing natural phenomena are those in which highly-structured collective behavior emerges over time from the interaction of simple subsystems”
  • “Emergence is generally understood to be a process that leads to the appearance of structure not directly described by the defining constraints and instantaneous forces that control a system”
  • “Over time ‘‘something new’’ appears at scales not directly specified by the equations of motion”
  • WHat is “something”, and what does it mean to be “new”?
  • Deterministic chaos: where in the determinism did the randomness come from? – emergence of the disorder is a result of the complicated behavior of nonlinear dynamic systems and the limitations of the observer
  • Unpredictability and self-similarity: the emergent feature stands in opposition to the system’s defining character
  • “The newness in both cases is in the eye of an observer: the observer whose predictions fail or the analyst who notes that the feature of statistical self-similarity captures a commonality across length scales”
  • “Briefly stated, in the realm of pattern formation ‘‘patterns’’ are guessed and then verified.”
  • INtrinsic emergence: pattern formation is insufficient to capture the essential aspect
  • Newness is always referred to outside of the system
  • “The problem is that the ‘‘newness’’ in the emergence of pattern is always referred outside the system to some observer that anticipates the structures via a fixed palette of possible regularities”
  • When a new state of matter emerges from a phase transition, for example, initially no one knows the governing ‘‘order parameter.’’
  • What is distinctive about intrinsic emergence is that the patterns formed confer additional functionality which supports global information processing. Recently, examples of this sort have fallen under the rubric of ‘‘emergent computation.
  • In summary, three notions will be distinguished:
    1. The intuitive definition of emergence: ‘‘something new appears’’;
    2. Pattern formation: an observer identifies ‘‘organization’’ in a dynamical system; and
    3. Intrinsic emergence: the system itself capitalizes on patterns that appear.
  • “In pattern formation, the observer is the scientist that uses prior concepts—e.g., ‘‘spiral’’ or ‘‘vortex’’—to detect structure in experimental data and so to verify or falsify their applicability to the phenomenon”
  • What’s in a model?
    • Essential novelty involved has to be referred to an evaluating entity
    • The observer in this view is a subprocess of the entire system
  • A fundamental point is that any act of modeling makes a distinction between data that is accounted for—the ordered part—and data that is not described—the apparently random part. This distinction might be a null one: for example, for either completely predictable or ideally random (unstructured) sources the data is explained by one descriptive extreme or the other. Nature is seldom so simple. It appears that natural processes are an amalgam of randomness and order. In our view it is the organization of the interplay between order and randomness that makes nature ‘‘complex.’
  • Then the size of the ‘‘best’’ model is a measure of the data’s intrinsic structure. If we believe the data is a faithful representation of the raw behavior of the underlying process, this then translates into a measure of structure in the natural phenomenon originally studied
  • “It is helpful to draw a distinction between discovery and emergence. The level of pattern formation and the modeling framework of computational mechanics concern discovery. Above it was suggested that innovation based on hierarchical machine reconstruction is one type of discovery, in the sense that new regularities across increasingly-accurate models are detected and then taken as a new basis for representation. Discovery, though, is not the same thing as emergence, which at a minimum is dynamical: over time, or over generations in an evolutionary system, something new appears. Discovery, in this sense, is atemporal: the change in state and increased knowledge of the observer are not the focus of the analysis activity; the products of model fitting and statistical parameter estimation are.”
  • “In contrast, emergence concerns the process of discovery. Moreover, intrinsic emergence puts the subjective aspects of discovery into the system under study. In short, emergence pushes the semantic stack down one level. In this view analyzing emergence is more objective than analyzing pattern formation in that detecting emergence requires modeling the dynamics of discovery, not just implementing a discovery procedure.
  • ‘‘Emergence’’ is meaningless unless it is defined within the context of processes themselves; the only well-defined notion of emergence would seem to be intrinsic emergence. Why? Simply because emergence defined without this closure leads to an infinite regress of observers detecting patterns of observers detecting patterns. . . . This is not a satisfactory definition, since it is not finite. The regress must be folded into the system, it must be immanent in the dynamics”
  • Modern evolutionary theory continues to be governed by Darwin’s view of the natural selection of individuals that reproduce with variation.
    • Selectionists hold that structure in the biological world is due primarily to the fitness-based selection of individuals in populations whose diversity is maintained by genetic variation. (In a sense, genetic variation is a destabilizing mechanism that provides the raw diversity of structure. Natural selection then is a stabilizing dynamic that acts on the expression of that variation.)
    • The second, anarchistic camp consists of the Historicists who hold fast to the Darwinian mechanisms of selection and variation, but emphasize the accidental determinants of biological form (What distinguishes this position from the Selectionists is the claim that major changes in structure can be and have been nonadaptive. While these changes have had the largest effect on the forms of present day life, at the time they occurred they conferred no survival advantage.)
    • there are the Structuralists whose goal is to elucidate the ‘‘principles of organization’’ that guide the appearance of biological structure. They contend that energetic, mechanical, biomolecular, and morphogenetic constraints limit the infinite range of possible biological form
  • Impressions
    • Natural selection’s culling of genetic variation provides the Selectionists with a theory of transformation. But the approach does not provide a theory of structure
    • The Historicists also have a theory of transformation, but they offer neither a theory of structure nor, apparently, a justification for a high fraction of functionality over the space of structures
    • the Structuralists do not offer a theory of transformation. Neither do they, despite claims for the primacy of organization in evolutionary processes, provide a theory of structure itself
  • Mechanisms
    • Deterministic mechanisms, statistical mechanics, computational mechanics, evolutionary mechanics
  • So is anything ever new? I would answer ‘‘most definitely.’’ With careful attention to the location of the observer and the system-under-study, with detailed accounting of intrinsic computation, with quantitative measures of complexity, we can analyze the patterns, structures, and novel information processing architectures that emerge in nonlinear processes. In this way, we demonstrate that something new has appeared.

“Alternative Views of Complexity”, Herbert Simon

  • Interest in complexity and complex systems
  • WWI: holism, Gestalts, creative evolution.
  • WWII: information, feedback, cybernetics, general systems
  • Now: chaos, adaptive systems, genetic algorithms, cellular automata.

Holism and Reductionism

  • Natural objects are wholes
  • Strong: Applied to complex systems in general, it postulates new system properties and relations among subsystems that had no place in the system components; hence it calls for emergence, a ‘‘creative’’ principle. Mechanistic explanations of emergence are rejected
  • Weak: emergence simply means that the parts of a complex system have mutual relations that do not exist for the parts in isolation

Cybernetics and General Systems

  • Norbert Wiener: cybernetics, combination of servomechanism theory, information theory, computers
  • Information theory: organized complexity explained in terms of reduction of entropy / disorder
  • In information theory, energy, information, and pattern all correspond to negative entropy.
  • Feedback control shows how a system can work toward goals and adapt to a changing environment,3 thereby removing the mystery from teleology
  • Although looking at complex systems generally might be difficult, we can look at classes of complex systems.

Current Interest in Complexity

  • Understanding the environment, society, and organisms
  • Relevant mathematics and computational algorithms emerging
  • Age of anxiety: “chaos”, “catastrophe”

Catastrophe Theory

  • Classification of nonlinear dynamic systems according to their modes of behavior
  • Such a system can assume two (or more) distinct steady states (static equilibria, for example, or periodic cycles); but when the system is in one of these states, a moderate change in a system parameter may cause it to shift suddenly to the other—or into an unstable state in which variables increase without limit.
  • It is not hard to think of natural systems that exhibit behavior of this kind—stable behavior followed by a sudden shift to disequilibrium or to another, quite different, equilibrium.

Complexity and Chaos

  • Theory of chaos from Poincare
  • Chaotic systems are deterministic dynamic systems that, if their initial conditions are disturbed even infinitesimally, may alter their paths radically
  • Nonlinear differential equations
  • Deep understanding has now been achieved of many aspects of chaos, but to say that we ‘‘understand’’ does not imply that we can predict.
  • Chaos led to the recognition of a new, generalized, notion of equilibrium—the so-called ‘‘strange attractor.’’

Rationality in a Catastrophic or Chaotic World

  • On the basis of the evidence, we should suppose neither that all of the complex systems we encounter in the world are chaotic, nor that few of them are. Moreover, as the airplane example shows, the ominous term ‘‘chaotic’’ should not be read as ‘‘unmanageable.’’

Complexity and Evolution

  • Much current research on complex systems focuses upon the emergence of complexity, that is, system evolution


  • What is new about the present activity is not the study of particular complex systems but the study of the phenomenon of complexity in its own right.

“More is Different”, P.W. Anderson

  • The reductionist hypothesis may still be a topic for controversy among philosophers, but among the great majority of active scientists I think it is accepted without question
  • The main fallacy in this kind of thinking is that the reductionist hypothesis does not by any means imply a ‘‘constructionist’’ one: The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe.
  • The constructionist hypothesis breaks down when confronted with the twin difficulties of scale and complexity. The behavior of large and complex aggregates of elementary particles, it turns out, is not to be understood in terms of a simple extrapolation of the properties of a few particles
  • At each stage entirely new laws, concepts, and generalizations are necessary, requiring inspiration and creativity to just as great a degree as in the previous one.
  • In my own field of many-body physics, we are, perhaps, closer to our fundamental, intensive underpinnings than in any other science in which non-trivial complexities occur, and as a result we have begun to formulate a general theory of just how this shift from quantitative to qualitative differentiation takes place. This formulation, called the theory of ‘‘broken symmetry,’’ may be of help in making more generally clear the breakdown of the constructionist converse of reductionism.
  • At this point we must forget about the possibility of inversion and ignore the parity symmetry: the symmetry laws have been, not repealed, but broken.
  • three inferences
    1. One is that symmetry is of great importance in physics. By symmetry we mean the existence of different viewpoints from which the system appears the same. It is only slightly overstating the case to say that physics is the study of symmetry
    2. the internal structural of a piece of matter need not be symmetrical even if the total state of it is, e.g. limits to infinity
    3. s that the state of a really big system does not at all have to have the symmetry of the laws which govern it; in fact, it usually has less symmetry
      • e.g. a crystal; built from a substrate of atoms, express perfect homogeneity of space; all of a sudden displays a new symmetry
  • The essential idea is that in the so-called N ! y limit of large systems (on our own, macroscopic scale) it is not only convenient but essential to realize that matter will undergo mathematically sharp, singular ‘‘phase transitions’’ to states in which the microscopic symmetries, and even the microscopic equations of motion, are in a sense violated

“The Status of Emergence”, Paul Henly

  • THE doctrine of emergence, having passed through the stage of acute philosophic controversy, has become to a large extent a generally accepted view. Unfortunately it has not become an un- ambiguous view. In one sense it seems to amount to no more than the insistence that there are novelties in the world
  • If belief in novelty were confined to this subjective aspect of experience, then there could be no question of its existene
  • Novelties, moreover, are assumed to arise not everywhere, but merely at certain stages in the evolution of the world, and are des- ignated as emergents, in contrast to resultants which involve no novelty
  • Novelty, then, pertains primarily to the qualities of things and a quality may be said to be new if it was never before embodied.
  • The crux of these exceptions seems to be that emergent evo- lution has to do with a causal situation. If one claims that life is a quality emergent from the inorganic, this is to claim that living organisms are brought into being by inorganic processes, though they are not explained by these processes
  • It is because of the failure of one group of qualities to explain another that unpredictability in advance of the fact can be used as the criterion of emergence. If there is an adequate basis in one group of characteristics for the explanation of the second, the latter could be predicted in advance of occurrence
  • With the exception just noted, however, unpredietability has been taken as the criterion of emergence, and here discussions of emergent evolution have been at their weakest, for unpredietabil- ity has been treated as a simple quality of events
  • predictability is seen once more to be relative to knowledge
  • Once the dependence of predictability upon evidence is ad- mitted, it becomes clear that unpredietability is not peculiarly the mark of the novel, but merely of what may be loosely called the logically unrelated

“Emergence, Closure and Inter-level Causation in Biological Systems”, Mossio et al.