Quantum Field Theories as Statistical Field Theories ?· Quantum Field Theories as Statistical Field…

  • View
    212

  • Download
    0

Embed Size (px)

Transcript

  • Quantum Field Theoriesas

    Statistical Field Theories

    Antoine TilloyMax Planck Institute of Quantum Optics, Germany

    Oberseminar Mathematische Physik LMU, Munich, GermanyFebruary 1, 2017

  • First problem:

    Quantum field theory is not about fields, there are no fields, itsabout correlation functions of ... macroscopic stuff made of ...

    (x1)(x2)(x3)(x4) =

    (x1)(x2)(x3)(x4) =

    what we want what we get

  • First problem:

    Quantum field theory is not about fields, there are no fields, itsabout correlation functions of ... macroscopic stuff made of ...

    (x1)(x2)(x3)(x4) =

    (x1)(x2)(x3)(x4) =

    what we want what we get

  • Second problem:

    Even as operational theories, physical interacting quantum fieldtheories are ill defined.

    (x1)(x2)(x3)(x4) =+k=0

    ()k gk

    (x1)(x2)(x3)(x4) =+k=0

    ()k gk

    These are actually the two classes of difficulties noted by Paul Dirac in 1963

  • Second problem:

    Even as operational theories, physical interacting quantum fieldtheories are ill defined.

    (x1)(x2)(x3)(x4) =+k=0

    ()k gk

    (x1)(x2)(x3)(x4) =+k=0

    ()k gk

    These are actually the two classes of difficulties noted by Paul Dirac in 1963

  • Reformulations of quantum theory are needed

  • The Bohmian way

    dQk(t)dt = v

    k (Q1(t), ,Qn(t))

    i ddt = HL. de Broglie 1927

    D. Bohm 1952

  • The Bohmian way

    dQk(t)dt = v

    k (Q1(t), ,Qn(t))

    i ddt = HL. de Broglie 1927

    D. Bohm 1952

  • Subtleties

    Lorentz invariance

    Either fixed foliation of foliation determined by the wave function the letter but not the spirit

    Particle creation-annihilation 2 solutions

    Stochastic creation-annihilation events Resurrect the Dirac sea

  • Subtleties

    Lorentz invariance

    Either fixed foliation of foliation determined by the wave function the letter but not the spirit

    Particle creation-annihilation 2 solutions

    Stochastic creation-annihilation events Resurrect the Dirac sea

  • Subtleties

    Lorentz invariance

    Either fixed foliation of foliation determined by the wave function the letter but not the spirit

    Particle creation-annihilation

    2 solutions

    Stochastic creation-annihilation events Resurrect the Dirac sea

  • Subtleties

    Lorentz invariance

    Either fixed foliation of foliation determined by the wave function the letter but not the spirit

    Particle creation-annihilation 2 solutions

    Stochastic creation-annihilation events Resurrect the Dirac sea

  • The Collapse way

    A modified Schrdinger equation

    t |w = iH |w+ tiny(w, w)

    A primitive ontology

    fields, particles, flashes

  • The Collapse way

    A modified Schrdinger equation

    t |w = iH |w+ tiny(w, w)

    A primitive ontology

    fields, particles, flashes

  • What is known:

    A Lorentz invariant GRW (Tumulka, 2006)

    but no interactions and a wave function formalism

    A Lorentz invariant CSL (Bedingham, 2011)

    but a non-linearity in the smearing function making the statisticalinterpretation unclear

    In both cases anyway, predictions = QFT.

  • What is known:

    A Lorentz invariant GRW (Tumulka, 2006)

    but no interactions and a wave function formalism

    A Lorentz invariant CSL (Bedingham, 2011)

    but a non-linearity in the smearing function making the statisticalinterpretation unclear

    In both cases anyway, predictions = QFT.

  • What is known:

    A Lorentz invariant GRW (Tumulka, 2006)

    but no interactions and a wave function formalism

    A Lorentz invariant CSL (Bedingham, 2011)

    but a non-linearity in the smearing function making the statisticalinterpretation unclear

    In both cases anyway, predictions = QFT.

  • What is known:

    A Lorentz invariant GRW (Tumulka, 2006)

    but no interactions and a wave function formalism

    A Lorentz invariant CSL (Bedingham, 2011)

    but a non-linearity in the smearing function making the statisticalinterpretation unclear

    In both cases anyway, predictions = QFT.

  • Objective:

    Construct a relativistic collapse model, naturally written as aLorentz invariant statistical field theory, that can be embeddedwithin an orthodox interacting quantum field theory.

    In provocative form

    1. Collapse models can be made equivalent to quantum theory2. Quantum field theories can be written as statistical field

    theories3. The two previous points are equivalent

  • Objective:

    Construct a relativistic collapse model, naturally written as aLorentz invariant statistical field theory, that can be embeddedwithin an orthodox interacting quantum field theory.

    In provocative form

    1. Collapse models can be made equivalent to quantum theory2. Quantum field theories can be written as statistical field

    theories3. The two previous points are equivalent

  • Objective:

    Construct a relativistic collapse model, naturally written as aLorentz invariant statistical field theory, that can be embeddedwithin an orthodox interacting quantum field theory.

    In provocative form

    1. Collapse models can be made equivalent to quantum theory

    2. Quantum field theories can be written as statistical fieldtheories

    3. The two previous points are equivalent

  • Objective:

    Construct a relativistic collapse model, naturally written as aLorentz invariant statistical field theory, that can be embeddedwithin an orthodox interacting quantum field theory.

    In provocative form

    1. Collapse models can be made equivalent to quantum theory2. Quantum field theories can be written as statistical field

    theories

    3. The two previous points are equivalent

  • Objective:

    Construct a relativistic collapse model, naturally written as aLorentz invariant statistical field theory, that can be embeddedwithin an orthodox interacting quantum field theory.

    In provocative form

    1. Collapse models can be made equivalent to quantum theory2. Quantum field theories can be written as statistical field

    theories3. The two previous points are equivalent

  • dynamical reduction models

    A modified Schrdinger equation

    t |w = iH |w+ tiny(w, w)

    A primitive ontology

    fields, particles, flashes

  • an instantiation: the csl model

    Linear collapse equation:

    ddt |w(t) =

    {iH0 +

    R3

    d3x M(x)wt(x)white

    2 M2(x)

    }|w(t) ,

    with M(x) =1

    (2)32

    Nk=1

    R3

    dyk mk e |xyk|

    2

    22 |yk yk| ,

    Master equation for t = E[|w(t) w(t)|

    ]ddt (t) = i [H0, (t)]

    2

    R3

    d3x[M(x), [M(x), (t)]

    ].

    More generally t = t 0with linear Completely Positive Trace Preserving

    needs to be an unraveling of a nice open-system evolution

    Gisin Diosi

  • an instantiation: the csl model

    Linear collapse equation:

    ddt |w(t) =

    {iH0 +

    R3

    d3x M(x)wt(x)white

    2 M2(x)

    }|w(t) ,

    with M(x) =1

    (2)32

    Nk=1

    R3

    dyk mk e |xyk|

    2

    22 |yk yk| ,

    Master equation for t = E[|w(t) w(t)|

    ]ddt (t) = i [H0, (t)]

    2

    R3

    d3x[M(x), [M(x), (t)]

    ].

    More generally t = t 0with linear Completely Positive Trace Preserving

    needs to be an unraveling of a nice open-system evolution

    Gisin Diosi

  • an instantiation: the csl model

    Linear collapse equation:

    ddt |w(t) =

    {iH0 +

    R3

    d3x M(x)wt(x)white

    2 M2(x)

    }|w(t) ,

    with M(x) =1

    (2)32

    Nk=1

    R3

    dyk mk e |xyk|

    2

    22 |yk yk| ,

    Master equation for t = E[|w(t) w(t)|

    ]ddt (t) = i [H0, (t)]

    2

    R3

    d3x[M(x), [M(x), (t)]

    ].

    More generally t = t 0with linear Completely Positive Trace Preserving

    needs to be an unraveling of a nice open-system evolution

    Gisin Diosi

  • an instantiation: the csl model

    Linear collapse equation:

    ddt |w(t) =

    {iH0 +

    R3

    d3x M(x)wt(x)

    2 M2(x)

    }|w(t) ,

    Normalization and cooking:

    w(t) = |w(t)w(t)|w(t)dt(w) = w(t)|w(t) d0(w)

    wt(x) = 2 M(x)+ bt(x)

    white.

  • an instantiation: the csl model

    Linear collapse equation:

    ddt |w(t) =

    {iH0 +

    R3

    d3x M(x)wt(x)

    2 M2(x)

    }|w(t) ,

    Normalization and cooking:

    w(t) = |w(t)w(t)|w(t)dt(w) = w(t)|w(t) d0(w)

    wt(x) = 2 M(x)+ bt(x)

    white.

  • an instantiation: the csl model

    The good choice of primitive ontology is w, not M(x):

    Is the natural object appearing in the linear eq. Does projects down to R3 the localization of the state:

    wt(x) = 2 M(x)+ bt(x)

    white

    Allows to reconstruct the state exactly: w |w Allows a time symmetric formulation of collapse

    Bedingham & Maroney (2015)

    Is the natural Lorentz invariant object in generalizationsTumulka (2006), Bedingham (2011)

    Allows quantum classical couplingAT & Disi (2016)

  • an instantiation: the csl model

    The good choice of primitive ontology is w, not M(x):

    Is the natural object appearing in the linear eq.

    Does projects down to R3 the localization of the state:wt(x) = 2

    M(x)+ bt(x)

    white

    Allows to reconstruct the state exactly: w |w Allows a time symmetric formulation of collapse

    Bedingham & Maroney (2015)

    Is the natural Lorentz invariant object in generalizationsTumulka (2006), Bedingham (2011)

    Allows quantum classical couplingAT