Skip to content

Tim Maudlin on Two Answers to the Big Question

November 6, 2012

Tim Maudlin has recently weighed in on two books which explore the question of why there is something rather than nothing.  The first, ‘A Universe From Nothing’ by physicist Laurence Krauss, has been controversially panned by our group’s researcher David Albert; the second, ‘Why Does the World Exist’ by Jim Holt, has been controversially panned by physicist  Freeman Dyson.

Maudlin’s take is  somewhat less critical than either of these previous reviews, but points  out that the two authors seem to be answering (or trying to answer) different questions.  Maudlin takes Krauss to task for being sloppy in his argument that the vacuum state is empty;  he also discusses why physicists see some states, such as our current state, as requiring explanation, but others, such as the ‘false vacuum’ of inflationary theory, as natural or preferred.

Tim has given us permission to post the paper here to be subject to comments.  Maybe he will reply to some.  As usual, comments will be moderated.

10 Comments leave one →
  1. Andre Mirabelli permalink
    November 21, 2012 11:17 am


    1. Near the end of your second page when addressing Krauss’s question, you say:
    Suppose there is a certain physical (sic) state, not of the whole of existence but
    rather of just a part, that deserves the name “vacuum state”.

    Though it is not explicit, I believe that here “part” is meant to imply temporal partiality but special totality, i.e. some relativistic time-slice of some, probably non-zero, thickness. Is that correct? If that is not your intent, then the rest of the paragraph confuses me.

    2. Regarding your question, in the middle of page 3, about the Big Bang,
    Does the very temporal notion of “before” break down, …

    I would suggest consideration of approaching this question through a theoretical operationalism based on thought experiments:
    Time is what is defined as that that is measured in principle by an ideal clock. Such ideal clocks are recurrent processes. If our best theory says that no recurrent processes predated a certain state of the universe, then such a state has no prior time. That is to say that if our theory says that there does not exist two states of the universe that are actual but are not subsequent to the Big Bang in which some property can be defined that is actually, but not necessarily, the same in both states, then there was no time before the Big Bang.

    • November 27, 2012 1:42 pm

      Dear Andre,

      Regarding the way to delimit only part of the whole universe as in the vacuum state, it is not incumbent on me to say exactly how to do this, but incumbent on Krauss. The point is that if one wants to say that “something came from nothing” then the “nothing” (i.e. the vacuum state) cannot be the state of the entire universe, else there would not be “something”. So the vacuum state has to be a state of only a part. Whether the part could be demarcated by something like a spacelike hypersurface, or have itself any temporal thickness, or not be characterizable by any such property, is not a question I can answer. But without some sort of answer, the idea of something coming from nothing cannot be maintained.

      With regard to your comment about time, I think you have the definitional dependences backwards. It is not that time is defined by reference to ideal clocks, it is that that the notion of an ideal clock is defined by reference to an independently existing temporal magnitude. A material system functions as a ideal clock exactly if its dynamics are such that some observable feature of the clock can be used to mark off equal temporal intervals. But for this definition of an ideal clock to make sense, there must already be a temporal structure with a metric to define “equal intervals”.

      The temporal structure is used in specifying the fundamental dynamical laws. Using these laws, one can then analyze the behavior of a material system to see if it approximates an ideal clock, and to what degree, and in what circumstances.

      Note that the notion of a recurrent process nowhere appears in this definition, and indeed a clock need not involve any recurrent process at all. For example, suppose you have a large chunk of some radioactive material. Then you can use it as a clock by employing the half-life: when only half of the original chunk remains that counts as the first “tick”, when half of the remainder remains that is the second “tick” etc. Given a large enough sample, such a clock can be made arbitrarily likely to be arbitrarily accurate, short of 100% likely to be perfectly accurate. And nothing recurs. What is essential is how the behavior of the chunk depends on the temporal magnitude, in this case via the exponential decay. This dependence produces a material system that can accurately indicate the passage of time.



      • April 17, 2013 9:01 am

        Hi Tim,
        Why can’t we define certain uniform motion as primitive, and ideal clock is any system with parts changing uniformly?
        Then time is simply whatever the ideal clock shows (not what it measures, as it does not measure anything). Time parameter in dynamical laws results from comparison of object motion and the motion of this ideal clock.
        Regarding your example of half-life, I think that it can function as a clock is because there is a theory that says it decays away half of its sample in same time intervals, as shown by some ideal clock. In QM we still need (classical) clocks to function as time parameter.

  2. Andre Mirabelli permalink
    December 3, 2012 1:44 pm


    I would say that in your clock, the property of ‘being half the amount as at the previous demarcation’ is what is recurrent. The property that is recurrent can be defined however one wishes (first, second or … order predicate) as long as there is some clearly defined recurrent property.
    As to the order of dependencies: I think that while the order of being might be different than the order of knowing, the knowing has to make sense to us in order for us to be able to address the being. Ideal clocks (or, in general, operational thought experiments) are what ensure us that we know what we are taking about when we point to the property of time (or, in general, any property) within a theory.
    I do not believe that realism is up for grabs. (The universe with or without a property of Time is what it is.) I believe that what is at stake is the clarity of our understanding of what we are speaking about. (Do we know what we mean to refer to by the property of time modeled by/within our theories?)
    If we do not understand what we are referring to by ‘time’ then the question of what came ‘before’ the big bang is meaningless to us, no matter what the ontology and truth of the universe is.


    • December 4, 2012 9:54 am


      Maybe I should have put the point another way. We write dynamical laws in terms of a time variable: it is that variable that an “ideal clock” would accurately allow us to observationally determine. (This might be the proper time along a path in Relativity). In the case of the exponential decay law, it is trivial to give the elapsed time as a function of the percentage of material remaining, so on the assumption that the percentage is something observable we have a clock, and it can be made as ideal as one likes to as high probability one likes. The idea of “ticking” is a completely artificial add-on: one doesn’t need it at all. And nothing in the physics is recurrent or cyclic.

      The point about definitions is about how one specifies what it means to be an “ideal clock”. I just did it, but by reference to an time parameter which is not itself defined by reference to clocks. It is just a postulated physical-geometrical magnitude that appears in the laws of physics. It is not directly observable, but one can analyze systems, like the radioactive material, to see if (given the physics) their observable properties can be used to indicate elapsed time, with high probability. To reverse the procedure, one would have to be able to define “ideal clock” without reference to the physical magnitude that it is supposed to accurately measure. I have absolutely no idea how one could do this, and in any case it is not at all what has been done. Of course, clocks that have recurrent physical cycles have some practical advantages, but that has nothing to do with definitions.

      If you think you can do it some other way, then try to define an “ideal clock” without reference to the antecedently existing temporal magnitude. I don’t know where you would begin.


  3. Andre Mirabelli permalink
    December 6, 2012 2:35 am


    Let us say that one wanted to test the effects of electricity at the nuclear level. One would see that the (semi-classical) Hamiltonian for the atom from which one would calculate decay rates has a term dependent on the electrical potential (even though it is swamped by other ‘forces’). So one might put your decaying material in a charged capacitor. Taking your view, we would need to assume the regularity of some other clock to test the temporal regularity of the decay rate. But which other periodicity? If the question is the effect of electricity on atomic processes, then atomic (or electro-magnetic) clocks are no longer a reliable standard, due to the ubiquity of (possibly minute) electrical effects. That is why I believe that it is the statistical mutual periodicity, which is independent of any other specific physical theory, that is being intellectually referred to by the “time variable” in dynamical laws and not to any particular (type of) physical instance (such as nuclear decay) that may be pointed to.

    I see this as a question as to whether the circularity of physical theory can be ameliorated by the separation of responsibility among aspects of a total theory, or whether the circularity must be holistic in every aspect and throughout. My proclivity is to encourage the falsifiability of theories by disengaging, to as much extent as possible, the thought experiments defining the measurement instruments from that of the thought experiments of applying them to the tests of the theories themselves. This allows us to better understand the claims of a theory rather then saying things like “By time I mean the measure of ‘t’ that makes my theory (of the moment) true.” Especially since no historical moment has ever seemed to have a defendable TOE, so no ‘t’ or anything else would ever be defined.
    To me, this is not about experiments or even about the truth of any particular theory. It is about understanding what exactly it is (factually true or false) that a particular theory is claiming. Thus, my repeated reference to ‘thought experiments.” The falsifiability (not the ‘falsity’) of the claims of a theory about the relationship between two processes are clearer if the two process are not dependently defined.

    As to your comment that “nothing in physics is recurrent”: My breathing is recurrent. With the freedom of language, there is no limit to the properties that I could define as “recurrent.” I presume you mean “periodic” instead of “recurrent.” But periodicity is exactly what is at stake: Is the duration of this recurrence the same as the duration of some past, future or elsewhere recurrence? If your point here is about the lack of precision in actual experiments, that, as I said, is not my concern here. However, the “mutuality” of periodicity does require a nesting process for precision even in the conceptual definition of “statistical mutual periodicity”. But that can be supplied.

    As to your point that no theory includes instructions as to how to identify its most basic concepts in actuality, I say touché. The “ostentatious” (sic) concepts of “here”, “now”, “there” and “when” are buried too deep in natural language for me to begin to deconstruct. The process of conceptual clarification has its limits. I would just rather not have these limits lie within the presentation of the formal theories themselves.


  4. December 10, 2012 11:06 am

    Dear Andre,

    It would help matters to be accurate. I never wrote “nothing in physics is recurrent”. I wrote “nothing in the physics”, i.e. in the physics of the decaying radioactive material, is recurrent. And, as I said, the notion of an ideal clock is refined by reference to a temporal structure, not the other way around. It is by physical analysis, in light of the accepted theory, that one can make corrections and interpret data, not by have to find some other recurrent process. You just can’t write down dynamical equations without a dynamical variable.


    • Andre Mirabelli permalink
      December 16, 2012 1:31 am


      I take our basic difference to be the extent to which it is desirable/helpful to view a theory holistically versus, to the extent possible, to have independent theories that can be, perfectly consistently, joined together at more inclusive levels.
      I agree with realism, and thus that a temporal structure is what clocks are intended to measure. I am definitely not proposing an operationalism that reverses that dependency. Our difference seems to be about whether it helps, to the extent possible, to understand the intent of a theory to be able to understand some of its elements separate from its structure as a whole.
      Re: “It is by physical analysis, in light of the accepted theory, that one can make corrections and interpret data, not by have to find some other recurrent process.”
      The terms of an “accepted theory” need to be interpreted. There is historical precedent, but that rubs with the possibility that the theory might differ on the intent of any particular term. So, I look at the time variable in QFT and ask what regularity does it imply for testing the rest of the theory. If I take the theory completely holistically then any discrepancy can be immediately set at the foot of the clock, and one would be off pursuing various possibilities as to what is regular so as to dispel the problem. This is not what happens. Nor is the truth of conflicting theories when applied to clocks required. Rather clocks are taken as separately provided by testing periodicity independently. The separate criteria can be informed by other theories, but its purpose is to find the commonly definable mutual periodicity of many processes that can be used to adjudicate conflicting theories. The process of confirming a broad theory is not to throw everything open to consideration at once. It is not to find a set of clocks, rods and other perimeters in the large (mathematical) space of joint possibilities of them all together in order to find one particular combination that makes the theory true. Testing that set of infinite possibilities truly seems a fools errand.
      Thus, clocks acceptable to our theories seem to have been and continue to be definable independent of our theories. This seems to me to be a good thing for our understanding of the claims of those theories.
      That we intend our theories to ultimately consistently explain everything, including each particular clock, is not denied. What is denied is that it is more informative to have one large explanatory circle as opposed to connected areas with their separate theories and independent empirical dependencies.
      Thus, I see it as a good thing that one can define an ideal clock and rod and other elements, possibly such as local inertial frames and internal properties, and then state a theory that implies a structure that includes them. You seem to be arguing that it would be better to just state the theory and then claim that our understanding of each and every part of it depends on its complete validity. I think that history favors my view of a world open to layered understanding: we are able to understand how to measure the dynamical variable of time at reasonable degrees of precision before we have a satisfactory theory of everything in which it might (or might come not to) play a fundamental part.
      That the same thought experiment could define a clock that can measure the time employed in successive theories is a positive thing for our understanding of what the theories intend to claim the world is like. It would be very confounding if the time variable used by successive theories had different meanings and, thus, different conceptual connections to everything else. It would make it extremely difficult for us to know what the theory intended us to understand about our world.
      I would deny that the overthrow of classical physics required that its rational reconstruction is based on an incorrect thought experiment for defining the clocks that measured its time variable. Rather, the time measured by independently conceived clocks could play a similar role in various incompatible theories. The incompatibility of various theories does not require that all the elements in each theory have a different theoretical meaning in each theory, as complete holism would require. If we come to propose a theory where a particular time measure is not appropriate, then some other variables that we could conceptually imagine measuring separately from the theory as a whole will be required for us to get a grip on what the broader theory is proposing the world is like. As of now, all our successive theories have had a time property that each would agree is properly measured by ‘the recurrent process that makes the most other recurrent processes periodic.’ (Accidents aside, it will win out that the recurrence in tune with the inherent regularity of the true, the heartbeat of the toe, is what will be found to be most common.) This theoretical understanding of the time measure allows us to engage newly proposed theories, instead of getting lost with no bearings at all in the infinite space of their logically possible interpretations.
      To reprise, an uncompromising holism is a logical sea offering no direction as to its interpretation. Introducing independent theories that can be separately understood but subsequently employed in broader theories provides anchors for our conceptual entry into a new theory. That is what a statistically based thought experiment for defining an ideal clock has offered.


      • December 16, 2012 1:37 pm


        You are attributing to me a lot of things I never said and frankly do not even understand. I said nothing at all about holism. If you have QFT, for example, then you can use it to try to find systems that will observably indicate the time variable with high probability. As I have repeatedly said, these may or may not be recurrent…it makes no difference. As for the rest, you seem to have some very complicated theory in mind. But it has little to do, as far as I can tell with anything I said in the review of elsewhere. This does not seem an appropriate venue to hash out your theory.


  5. Larry Hitterdale permalink
    March 20, 2013 9:36 pm

    I agree that (many) physicists do regard certain symmetries as making a physical state “to be expected” or “not in need of further explanation”. I also agree that neither Krauss nor Holt addresses this issue in a satisfactory way. If we do examine the matter in more detail, we can see that the assumption is unwarranted. Neither the content nor practice of current physics indicates that any particular kind of state needs no further explanation. Every symmetry which some physicists might suppose to need no further explanation, other physicists are trying to explain. So, the basis for the assumption must come from outside physics. Symmetry as a selector looks like a version of simplicity. But, if so, symmetry is not simple enough. If simplicity were the selector for existence, then one might have expected all reality to be one non-composite particle characterized by one bland property. In any case, simplicity is implausible as a guess as to what the selector might be. What I mean is this: Suppose we know only that something exists, but we are ignorant about what exists. Logically, we should guess that existence is a generic or run-of-the-mill selection from possibilities. If existence instantiated a generic possibility, then existence would be much less orderly than it is. There are many more possible ways to be chaotic than to be organized. So, the fact that the world seems built on symmetries, far from needing no explanation, is especially in need of explanation. On this point, the view of many physicists is the reverse of the truth. Of course, for various reasons, we might reject talk about generic possibilities, comparing possibilities, and so on. Then the ideas of “to be expected” or “not needing an explanation” would be wrong, but for a different reason, because there would be no principles on the basis of which they could be either affirmed or denied. We could only say that, at the present time in the history of inquiry, explanation has reached to some principles or states, but no further. What, if anything, lies beyond would be completely unknown.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: