83
Do Quanta Need a New Logic? John Stachel Frontiers of Fundamental Physics 14 Epistemology and Philosophy Marseille, 17 July 2014

Do Quanta Need a New Logic? John Stachel

  • Upload
    cheche

  • View
    59

  • Download
    2

Embed Size (px)

DESCRIPTION

Do Quanta Need a New Logic? John Stachel. Frontiers of Fundamental Physics 14 Epistemology and Philosophy Marseille, 17 July 2014 . Do Quanta Need a New Logic?. Three Aspects of a Logic. Syntax : A formalism, or set of symbols and rules for well-formed formulas using these symbols - PowerPoint PPT Presentation

Citation preview

Page 1: Do Quanta Need a New Logic? John  Stachel

Do Quanta Need a New Logic?

John Stachel

Frontiers of Fundamental Physics 14Epistemology and Philosophy

Marseille, 17 July 2014

Page 2: Do Quanta Need a New Logic? John  Stachel

Do Quanta Need a New Logic?

Page 3: Do Quanta Need a New Logic? John  Stachel

Three Aspects of a LogicSyntax: A formalism, or set of symbols and rules for well-formed formulas using these symbolsSemantics: An interpretation of the formalism: Its meaning as a description of some aspect of the worldPragmatics: Its application to that aspect of the world

Page 4: Do Quanta Need a New Logic? John  Stachel

Logic-Language-WorldThree steps:

Logic is about Language, Language is about The World.

Panlogism– The attempt to “short circuit” this process by

identifying the real object and the “concrete-in- thought” leads to the assertion:

Logic is about The World

Page 5: Do Quanta Need a New Logic? John  Stachel

Aron Gurwitsch

Page 6: Do Quanta Need a New Logic? John  Stachel

Leibniz: Philosophie des Panlogismus

“Things are realizations of concepts of reason. It is not sufficient to maintain that the logical and the ontological viewpoints can never be fully distinguished from each other, or that no separation, no abyss exists between reason and reality. One seems most faithful to the situation, if one speaks of an identity, or better of an equivalence of the logical and the ontological”

Page 7: Do Quanta Need a New Logic? John  Stachel

Panlogism Reborn! “ By panlogism I mean the philosophical

tendency to obliterate the distinction between logical and ontological principles” JS, “Do Quanta Need a New Logic?” (1986)

I have been combating this viewpoint for over 35 years:

”The ‘Logic’ of Quantum Logic” in PSA 1974 (Dordrecht: Reidel 1976), pp. 515-526.

Page 8: Do Quanta Need a New Logic? John  Stachel
Page 9: Do Quanta Need a New Logic? John  Stachel
Page 10: Do Quanta Need a New Logic? John  Stachel

Logic(s) For Quantum Mechanics

If we abandon panlogism, and the search for The Logic of Quantum Mechanics, we shall see that quantum logics (note the plural!) are just different ways of reformulating the same content of quantum theory.

Page 11: Do Quanta Need a New Logic? John  Stachel

The Danger:If we accept one formulation of The Logic of Quantum Mechanics as providing the answers, it prevents us from confronting the real questions about the nature of quantum mechanics.

Page 12: Do Quanta Need a New Logic? John  Stachel

Chris Isham

Page 13: Do Quanta Need a New Logic? John  Stachel

Is it True; or is it False; or Some-where In Between? The Logic of Quantum Theory

"A key feature of classical physics is that, at any given time, the system has a definite state, and this state determines-- and is uniquely determined by-- the values of all the physical quantities associated with the system.“ Realism is "the philosophical view that each physical quantity has a value for any given state of the system.“

Page 14: Do Quanta Need a New Logic? John  Stachel

In a Letter to Chris, I Raised Two Problems:

1) Conditional Properties:“each physical quantity has

a value” 2) The Primacy of Process:

“for any given state of the system”

Page 15: Do Quanta Need a New Logic? John  Stachel

First Problem:

1) Conditional Properties:“each physical quantity has

a value” 2) The Primacy of Process:

“for any given state of the system.”

Page 16: Do Quanta Need a New Logic? John  Stachel

1) Conditional PropertiesThis is just not true of conditional properties, as discussed in detail in Do Quanta Need a New Logic? The example I use concerns the properties "hardness h" and "viscosity v": Given a system defined by its chemical composition, the property "hardness" will only apply-- let alone have a numerical value on Moh's scale-- if the system is in its solid state; while "viscosity" will only apply if the system is in a fluid (liquid or gaseous) state.

Page 17: Do Quanta Need a New Logic? John  Stachel

Simple Classical ExampleHardness and Viscosity can be applied to any substance, but not simultaneously. If it is in solid state, hardness applies; if it is in a fluid state viscosity applies.

Page 18: Do Quanta Need a New Logic? John  Stachel

1) Conditional PropertiesWith the definition of the logical 'negation' operator, [logic] has already gotten as complicated as it gets. … [M]y article … discusses the difference between choice negation (-) and exclusion negation (∼ ) in general, and the inevitability of choice negation for a conditional predicate … if one wants to derive other predicates from it, and what follows from this choice even before getting to the special case of QM.

Page 19: Do Quanta Need a New Logic? John  Stachel

Non-Standard Logic NeededSemanticsIf we allow elementary propositions of the form:

“System S has hardness h,” “System S has viscosity v”

Then a non-standard logic is needed: negation (“not”), conjunction (“and”) and disjunction (“or”) cannot all follow the laws of classical logic

Page 20: Do Quanta Need a New Logic? John  Stachel

Which Negation?How shall we interpret the proposition:

“System S does not have hardness h1”It could mean:

“System S has some hardness h2≠ h1” or it could mean:

“System S is not in the solid state, so the concept of hardness does not apply”

Page 21: Do Quanta Need a New Logic? John  Stachel

Which Negation?The law of the excluded middle:

“p or not p is always true” (p∨∼p=I)If we choose

“System S has some hardness h2≠ h1”Then the law of the excluded middle Is not valid for this variant of intuitionistic logic. To keep the law, we must choose the disjunction:

“System S has some hardness h2≠ h1 or it is not in the solid state so the concept of hardness does not apply” This choice also leads to a non-standard logic!

Page 22: Do Quanta Need a New Logic? John  Stachel

Simplest ExampleOnly Two States, Solid and Liquid:Solid State, only two values h1, h2 of hardnessLiquid State, only two values v1, v2 of viscosityThen there are only four elementary propositions, symbolized by h1, h2, v1, v2

We also need symbols for:Negation (∼)The identically false proposition ϕ The dentically true proposition I

Page 23: Do Quanta Need a New Logic? John  Stachel

Simplest ExampleIf we plot all possible combinations of propositions p, starting from the identically false proposition ϕ and ending with the identically true proposition I, and use dotted lines to represent logical implication, then we get the following diagram of a propositional lattice

Page 24: Do Quanta Need a New Logic? John  Stachel

Propositional Lattice I (∼ϕ) . . . . . . . . . . . h1 (∼h2) v2 (∼v1) v1 (∼v2) h2(∼h1) . . . . . . . . . . . ϕ (∼I )

Page 25: Do Quanta Need a New Logic? John  Stachel

David Ritz Finkelstein

Page 26: Do Quanta Need a New Logic? John  Stachel

The Physics of Logic (1969)

Finkelstein uses the same diagram to describe the polarization of photons, and states:

“The system is the simplest quantum-like lattice and exhibits nondistributivity and coherence.”

Page 27: Do Quanta Need a New Logic? John  Stachel

Enter PragmaticsBut such propositions cannot be tested without

the additional specification of the conditions C under which system S is being observed (e.g., temperature T, pressure p).

So if we add these conditions to the form of the proposition:

“System S under conditions C has hardness h”

Page 28: Do Quanta Need a New Logic? John  Stachel

Classical Logic Returns!Then, either:

The proposition is not well-formed, if system S is not in the solid state under conditions C,

or: The proposition is well-formed, if system S is in the solid state under conditions C, and

Classical logic holds for all well-formed propositions!

Page 29: Do Quanta Need a New Logic? John  Stachel

ObjectionWell and good for this example, but there is no analogue of the superposition principle in it. We cannot superpose two states S1 and S2 of different hardness or a state of hardness and and a state of viscosity to get a new state SAnswer:Let’s look at another classical example:

Page 30: Do Quanta Need a New Logic? John  Stachel
Page 31: Do Quanta Need a New Logic? John  Stachel

What Are Colors?Colors are literally in the mind of the beholder: The human eye and brain combine to interpret all electromagnetic waves within a certain range of frequencies that impinge on the retina in terms of three dimensions, e.g.:

Brightness, hue and saturation orThree primary colors

So colors are conditional properties of an open system, forming a 3-dimensional vector space

Page 32: Do Quanta Need a New Logic? John  Stachel
Page 33: Do Quanta Need a New Logic? John  Stachel

Color LogicThree primary colors:

Red Green Blue

Page 34: Do Quanta Need a New Logic? John  Stachel

Additive Color

Page 35: Do Quanta Need a New Logic? John  Stachel
Page 36: Do Quanta Need a New Logic? John  Stachel

They can be mixed (color superposition) to get:

Magenta Yellow Cyan

Page 37: Do Quanta Need a New Logic? John  Stachel
Page 38: Do Quanta Need a New Logic? John  Stachel

ComplementationTwo colors are complementary if, when superposed (mixed) they produce white.

Red Cyan Green Magenta Blue Yellow

Page 39: Do Quanta Need a New Logic? John  Stachel
Page 40: Do Quanta Need a New Logic? John  Stachel

Enter LogicDefine elementary propositions:

“Object O has color c” Interpret negation (“not”) as color

complementationInterpret conjunction (“and”) as color

additionAnd you have a non-standard logic with superposition of colors!

Page 41: Do Quanta Need a New Logic? John  Stachel

What Is a Quantum System?

So neither conditional properties nor superposition are unique to quantum systems.Then what makes a quantum system? As we shall see, it is the role of h, the quantum of action.

Page 42: Do Quanta Need a New Logic? John  Stachel

What is Quantization? Quantization is just a way of accounting for

the effects of h, the quantum of action, on any process involving some system,– or rather on theoretical models of such a system-- “fundamental” or “composite”, in which the collective behavior of a set of more fundamental entities is quantized

Page 43: Do Quanta Need a New Logic? John  Stachel

“Atoms and Human Knowledge”--Niels Bohr 1957

“..an element of wholeness, so to speak, in the physical processes, a feature going far beyond the old doctrine of the restricted divisibility of matter. This element is called the universal quantum of action. It was discovered by Max Planck in the first year of this century and came to inaugurate a whole new epoch in physics and natural philosophy.

Page 44: Do Quanta Need a New Logic? John  Stachel

“Atoms and Human Knowledge”-- (cont’d)

We came to understand that the ordinary laws of physics, i.e., classical mechanics and electrodynamics, are idealizations that can only be applied in the analysis of phenomena in which the action involved at every stage is so large compared to the quantum that the latter can be completely disregarded.

Page 45: Do Quanta Need a New Logic? John  Stachel

The Second Problem1) Conditional Properties:

“each physical quantity has a value” 2) The Primacy of Process:

“for any given state of the system”

Page 46: Do Quanta Need a New Logic? John  Stachel

Lee Smolin

Page 47: Do Quanta Need a New Logic? John  Stachel

Three Roads to Quantum Gravity

“[R]elativity theory and quantum theory each ... tell us-- no, better, they scream at us-- that our world is a history of processes. Motion and change are primary. Nothing is, except in a very approximate and temporary sense. How something is, or what its state is, is an illusion.

Page 48: Do Quanta Need a New Logic? John  Stachel

Three Roads to Quantum Gravity

It may be a useful illusion for some purposes, but if we want to think fundamentally we must not lose sight of the essential fact that 'is' is an illusion. So to speak the language of the new physics we must learn a vocabulary in which process is more important than, and prior to, stasis.”

Page 49: Do Quanta Need a New Logic? John  Stachel

2) Primacy of ProcessPhrases such as "at any moment of time", "at any given time” may be applied in Newtonian-Galileian physics, which is based on a global absolute time. But from SR on to GR, this phrase involves a convention defining a global time., and nothing physical can depend on the choice of convention!

Page 50: Do Quanta Need a New Logic? John  Stachel

2) Primacy of ProcessThe only convention-invariant things are processes, each involving a space-time region. This suggests-- as do many other considerations-- that the fundamental entities in quantum theory are the transition amplitudes, and that states should be taken in the c.g.s. system (cum grano salis).

Page 51: Do Quanta Need a New Logic? John  Stachel

2) Primacy of ProcessAnd this is true of our measurements as well: any measurement involves a finite time interval and a finite 3-dimensional spatial region. Sometimes, we can get away with neglecting this, and talking, for example in NR QM, about ideal instantaneous measurements.

Page 52: Do Quanta Need a New Logic? John  Stachel

2) Primacy of ProcessBut sometimes we most definitely cannot, as Bohr and Rosenfeld demonstrated for E-M QFT, where the basic quantities defined by the theory (and therefore measurable-- I am not an operationalist!) are space-time averages. Their critique of Heisenberg shows what happens if you forget this!

Page 53: Do Quanta Need a New Logic? John  Stachel

Closed versus Open Systems System Key Concept Closed Determinism Open Causality

Determinism means fatalism: nothing can change what happensCausality means control: by manipulating the causes, one can change the outcome

“Determinism is really an article of philosophical faith, not a scientific result” (JS 1968).

Page 54: Do Quanta Need a New Logic? John  Stachel

The Dogma of ClosureWhen classical physics treated open systems, it was tacitly assumed (as an article of faith) that, by suitable enlargement of the system, it could always be included in closed system of a deterministic type. … The contrast between open and closed should not be taken as identical with the contrast between ‘phenomenological’ and ‘fundamental’ …(JS: “Comments on ‘Causality Requirements and the Theory of Relativity,” 1968)

Page 55: Do Quanta Need a New Logic? John  Stachel

Do We Really Want Global?The systems we actually model are finite processes, and all finite processes are open. A finite process is a bounded region in space- time: Its boundary is where new data (information) can be fed into the system and the resulting data can be extracted from it. Example: Asymptotically free in- and out-states in a scattering process.

Page 56: Do Quanta Need a New Logic? John  Stachel

Properties: Intrinsic and Extrinsic

Intrinsic properties: Define the nature of the systemExample: The mass, charge, and spin of an electron, proton, neutron, etc.Extrinsic properties: Depend on the relation of the system to its surroundingsExample: The position, momentum, angular momentum of a system relative to some inertial frame of reference.

Page 57: Do Quanta Need a New Logic? John  Stachel

Heisenberg Physics and PhilosophyIntroduction by Paul Davies

One is used to uncertainty in many physical processes – for example, in the stock market or in thermodynamics – but in these cases the uncertainty is due to missing information rather than to any fundamental limitation in what may be known about these systems.

Page 58: Do Quanta Need a New Logic? John  Stachel

ProcessesAn open system can undergo a process:

Preparation of the system, Interaction with its

environ- ment, Registration of some

result.

Page 59: Do Quanta Need a New Logic? John  Stachel

Experiment vs ObservationExperiment: The preparation result is fixed, predict the result of the registration (laboratory).Observation: The registration result is fixed, retrodict the result of preparation that led to it (astronomy, cosmology).

Page 60: Do Quanta Need a New Logic? John  Stachel

Classical System:Complete information allows us

to predict/retrodiction with certainty.

Incomplete information leads to compute the probability of a prediction/retrodiction.

Page 61: Do Quanta Need a New Logic? John  Stachel

What is ProbabilityEnsemble interpretation:

Large number of copies of the process.Probability is the ratio of number of

copies giving the predicted outcome to the total numberPropensity interpretation:

Probability is the propensity of the process to give the predicted outcome

Page 62: Do Quanta Need a New Logic? John  Stachel

Quantum Logic (J.S.)The conditional probability for a system initially prepared with position qi at time ti to be found in an interval dq around q at time t is given by:

P(qi , ti; q, t) dq,where the probability density P(qi , ti; q, t) is proportional to the Van Vleck determinant of Hamilton’s principal function. This probability may be given a propensity interpretation for a virtual ensemble associated with a single system, or a frequency interpretation for an ensemble of identically prepared systems.

Page 63: Do Quanta Need a New Logic? John  Stachel

Heisenberg Physics and PhilosophyIntroduction by Paul Davies

It is essential to appreciate that this uncertainty is inherent in nature and not merely the result of technological limitations in measurement. It is not that the experimenter is merely too clumsy to measure position and momentum simultaneously. The particle simply does not possess simultaneously precise values of these two attributes.

Page 64: Do Quanta Need a New Logic? John  Stachel

Bohr: Atomic Physics and Human Knowledge

On the lines of objective description, it is indeed more appropriate to use the word phenomenon to refer only to observations obtained under circumstances whose description includes an account of the whole experimental arrangement. In such terminology, the observational problem in quantum physics is deprived of any special intricacy,

Page 65: Do Quanta Need a New Logic? John  Stachel

Atomic Physics and Human Knowledge

and we are, moreover, directly reminded that every atomic phenomenon is closed in the sense that its observation is based on registrations obtained by means of suitable amplification devices with irreversible functioning such as, for example, permanent marks on a photographic plate, caused by the penetration of electrons into the emulsion.

Page 66: Do Quanta Need a New Logic? John  Stachel

What is a Quantum Process?

A quantum process involves three stages: preparation, interaction, registration.

Big question: How does h figure in the preparation and registration procedures?

Page 67: Do Quanta Need a New Logic? John  Stachel

Quantum System:The effects of the quantum of action h on the process cannot be neglected. In general, only probabilities 0 p1 for such processes can be calculated (this does not exclude the occasional p = 0 or 1)

Page 68: Do Quanta Need a New Logic? John  Stachel

The Big Difference:If in some process, different interaction paths can lead from preparation to registration; then to get the total probability for that process, if the paths are:Distinguishable: add the probabilities for each path (classical)Indistinguishable: add the probability amplitudes for each path (quantum).

Page 69: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

Much of the debate over quantum logic(s) hinges (implicitly if not explicitly) on the question of the quantum-mechanical analogue of the classical-mechanical propositions. Propositions about a classical system can refer to only the properties of the system itself, considered as closed-- that is, not interacting with anything outside the system.

Page 70: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

But, as Bohr emphasized, the existence of the quantum of action h prevents such a complete separation between a quantum-mechanical system and its macroscopic surroundings. Quantum Mechanics can only deal with open systems. Two major consequences are:

Page 71: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

(i) A full description of a quantum-mechanical phenomenon (see Bohr 1958) or process (see Feynman 1968– the word “process” will be used hereafter) must include a specification of the result of an initial preparation of the system, an account of the type of interactions it undergoes subsequently, and of the result of some act of registration (“measurement”) to which the system is finally subjected.

Page 72: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

(ii) A maximal quantum-mechanical preparation or registration only specifies “half” the data about a system that would be specifiable classically. For example, while one could in principle prepare or register a classical-mechanical system with a determinate position and momentum, one can only prepare or register a quantum-mechanical system with either a determinate position or momentum.

Page 73: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

A typical proposition about a process involving an electron might read:

“At time t1 the electron was prepared with momentum p0, subsequently passed through a certain electric field E, and at a later time t2 was registered at position q0.” Quantum mechanics assigns a probability to such a proposition.

Page 74: Do Quanta Need a New Logic? John  Stachel

Conditional ProbabilitiesNote that, as in the classical case, all the quantum mechanical probabilities for a process are conditional; one does not ask for the probability of a final value b tout court, but its probability given an initial prepared value a. Once computed, these conditional probabilities obey the laws of the classical probability calculus based on classical logic.

Page 75: Do Quanta Need a New Logic? John  Stachel

Conditional ProbabilitiesIn contrast to the classical case, quantum-mechanical conditional probabilities cannot be attributed to ignorance but are fundamental. Without altering the physical conditions defining the process considered (thus producing a different process), it is impossible to further subdivide a quantum-mechanical ensemble into sub-ensembles– let alone into individual trajectories–

Page 76: Do Quanta Need a New Logic? John  Stachel

Conditional Probabilitiesfor which probabilities (as opposed to probability amplitudes) can be defined. For example, a sample of undecayed radioactive nuclei all with the same average half-life cannot be subdivided it into sub-ensembles each with different predicted average life-times. (Of course, retroactively, after some or all have decayed, it is easy to do so.).

Page 77: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

Classical propositional logic is all that is needed to handle such propositions, describing a complete quantum-mechanical process. Logical problems begin when such propositions are truncated by omission of reference to the preparation and/or registration.

Page 78: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

Attention is focused on “the state of the system at time t,” to which one tries to attach a significance similar to that of the state in the classical case. The maximal goal of such an approach is to attribute a complete set of classical properties, e.g., both position and momentum, to a quantum system in a given state.

Page 79: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

These elementary propositions are then compounded by suitably-defined operations of conjunction (“and”) and disjunction (“or”) in an attempt to give meaning to the resulting compound propositions, even when the properties referred to in such compound propositions are incompatible quantum mechanically.

Page 80: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

I shall not enter into further detail about the various possible non-standard logics that have been proposed for quantum mechanics (see my article Quantum Logic), but end with two reminders:

Page 81: Do Quanta Need a New Logic? John  Stachel

What is a Quantum-Mechanical Proposition?

1) If you accept panlogism (“The Logic of …”), this philosophical choice has nothing to do with quantum theory.2) The choice of a logic (“Logics for … “) is just a matter of selecting one of various different ways to express the same content.

Page 82: Do Quanta Need a New Logic? John  Stachel

Final Word on Quantum Logic (Niels Bohr, 1939)

The question of the logical forms which are best adapted to quantum theory is in fact a practical problem, concerned with the most convenient manner in which to express the new situation that arises in this domain.

Page 83: Do Quanta Need a New Logic? John  Stachel