In still another post at PT (go here), I’m charged with committing a basic physics error in my book No Free Lunch, much to the delight of the gallery that comments there (based, by the way, on a deliberate misquote — see below). Too bad that Freeman Dyson agrees with me and not with them. Here, then, is the pearl: http://www.aleph.se/Trans/Global/Omega/dyson.txt. Go trample on it. And having trampled on it, go email Freeman and get him to distance himself from my views even though the section of NFL cited merely expands on his and Frank Tipler’s ideas.
The light from the distant galaxies will be strongly red-shifted. But the sky will never become empty and dark, if we can tune our eyes to longer and longer wavelengths as time goes on. –Freeman Dyson
In addition, the author of the PT post deliberately misquotes me, juxtaposing two passages from my work without any indication that several pages of text intervene between the passages. Here is the passage attributed to me at PT exactly as it appeared there (at the very least, there should have been an ellipsis before “Certainly quantum mechanics …” as well as an indication that his actually is not the start of a sentence):
WhatÃ¢â‚¬â„¢s more, the energy in quantum events is proportional to frequency or inversely proportional to wavelength. And since there is no upper limit to the wavelength of, for instance, electromagnetic radiation, there is no lower limit to the energy required to impart information. In the limit, a designer could therefore impart information into the universe without inputting any energy at all. Whether the designer works through quantum mechanical effects is not ultimately the issue here. Certainly quantum mechanics is much more hospitable to an information processing view of the universe than the older mechanical models. All thatÃ¢â‚¬â„¢s needed, however, is a universe whose constitution and dynamics are not reducible to deterministic natural laws. Such a universe will produce random events and thus have the possibility of producing events that exhibit specified complexity (i.e., events that stand out against the backdrop of randomness).
And now here is the full text with the two passages marked in bold. Note that the PT post simply kludges those passages together (you’ll have to scroll down quite a ways to see the connection). By the way, I’ve saved the page at PT just so that they don’t insert ellipses and say there never was a problem:
How much energy is required to impart information? We have sensors that can detect quantum events and amplify them to the macroscopic level. WhatÃ¢â‚¬â„¢s more, the energy in quantum events is proportional to frequency or inversely proportional to wavelength. And since there is no upper limit to the wavelength of, for instance, electromagnetic radiation, there is no lower limit to the energy required to impart information. In the limit, a designer could therefore impart information into the universe without inputting any energy at all.
Limits, however, are tricky things. To be sure, an embodied designer could impart information by employing arbitrarily small amounts of energy. But an arbitrarily small amount of energy is still a positive amount of energy, and any designer employing positive amounts of energy to impart information is still, in Paul DaviesÃ¢â‚¬â„¢s phrase, Ã¢â‚¬Å“moving the particles.Ã¢â‚¬Â [[In contrast to the PT post, the possibility of infinite wavelength, zero energy, and zero bandwidth therefore never arises. –WmAD]]. The question remains how can an unembodied designer influence the natural world without imparting any energy whatsoever. It is here that an indeterministic universe comes to the rescue. Although we can thank quantum mechanics for the widespread recognition that the universe is indeterministic, indeterminism has a long philosophical history, and appears in such diverse places as the atomism of Lucretius and the pragmatism of Charles Peirce and William James.
For now, however, quantum theory is probably the best place to locate indeterminism. True, there is a sense in which quantum mechanics is deterministic: The evolution of the state function by means of the Schroedinger equation is deterministic; that is, given the state function at a given time, the Schroedinger equation prescribes the exact state at some future time. Nonetheless, the state function itself characterizes a probability distribution, and all observation of quantum systems involves sampling from such probability distributions. An analogy may help here. Imagine an urn that always contains ten balls. On Monday there are two white balls and eight black balls in the urn, on Tuesday there are three white balls and seven black balls in the urn, …, and on Sunday there are eight white balls and two black balls in the urn. Day to day the number of balls in the urns is determined. But sampling from these urns any day of the week is probabilistic and therefore indeterministic.
I need here to add a word about quantum cosmology and the many-worlds interpretation of quantum mechanics. Many quantum cosmologists would cringe at my characterization of quantum mechanics. The emerging consensus among quantum cosmologists (and one now held by Murray Gell-Mann, Philip Anderson, Stephen Hawking, and Steven Weinberg) is that quantum mechanics is completely deterministic. Accordingly, the state function of quantum mechanics does not characterize a probability distributionÃ¢â‚¬â€we only interpret it as a probability distribution from our limited vantage. Instead, the state function describes an ensemble of universes. Thus, the emerging consensus among quantum cosmologists is a many-worlds view (see section 2.8).
Why has this view taken hold? In quantum cosmology, when trying to apply quantum mechanics to the universe as a whole, having the state function collapse, as it must within a probabilistic interpretation, leads to a break in the dynamics of the quantum equations. This is mathematically unappealing (and for many cosmologists also metaphysically unappealing since it gives up on full deterministic causalityÃ¢â‚¬â€this was EinsteinÃ¢â‚¬â„¢s worry about quantum mechanics). Thus, instead of allowing for state-function collapse, quantum cosmologists have come to prefer an expanded ontology in which all possible histories or worlds consistent with quantum mechanics get lived out.
The totality of physical reality for quantum cosmologists is therefore vastly bigger than the world we think we inhabit. We think we live in a world where Hitler lost World War IIÃ¢â‚¬â€and we are right as far as that goes. Yet from a many-worlds point of view, it would be more precise to say that we live within a world that is but one among a multiplicity of worlds each of which are as real as ours and whose union properly constitutes the whole of reality. Moreover, the role of quantum theory is to coordinate all those worlds. Thus within our world, Hitler lost World War II. But presumably there are quantum events that could have changed the course of that war, on account of which Hitler would then have won World War II, and whose consequences are being fully worked out in that alternate world. The totality of physical reality is thus no longer properly conceived as a universe in the traditional sense but as a multiverse consisting of multiple worlds, multiple histories, and multiple minds.
There is an old joke: ThereÃ¢â‚¬â„¢s speculation, wild speculation, and cosmology! When Alan Guth first began proposing his inflationary cosmology, Lenny Susskind remarked, Ã¢â‚¬Å“You know, the most amazing thing is that they pay us for this.Ã¢â‚¬Â Cicero likewise remarked, Ã¢â‚¬Å“There is nothing so ridiculous but that some philosopher has said itÃ¢â‚¬ÂÃ¢â‚¬â€no doubt he included the natural philosophers and cosmologists of his day. I would update CiceroÃ¢â‚¬â„¢s dictum as follows: Ã¢â‚¬Å“No cosmological idea is so crazy but that it becomes plausible and even compelling once it is given an elegant mathematical formulation and shown to underwrite physics as the ultimate science.Ã¢â‚¬Â My aim with these remarks is not ridicule but a reality check. Cosmologists are notorious for beginning with physics and ending in metaphysics. The problem is that if you are willing to monkey with metaphysics, you can get any result you like.
Many-worlds purchase complete determinism, but at a huge metaphysical cost. Many-worlds vastly inflate our ontology. In fact, inflated ontologies have become a dominant theme of recent cosmological speculation. The point to realize is that there is no reason to give these inflated ontologies, especially in the case of quantum many-worlds, any allegiance except as speculative hypotheses that are of interest because of the insights they generate. Why? Not merely because data underdetermine theories but because, in the case of inflated ontologies, data could never even in principle adjudicate among such theories (see section 2.8). David Lindley made this point beautifully in The End of Physics: The Myth of a Unified Theory. LindleyÃ¢â‚¬â„¢s choice of the word Ã¢â‚¬Å“mythÃ¢â‚¬Â was well-considered. The bloated ontologies of contemporary cosmological speculation, like the myths of old, bring unity to our understanding but at the cost of severing us from the data of actual experience.
I could go on with this sociological commentary on quantum cosmology and many-worlds, but let us consider the many-worlds interpretation on its own terms. First off, the many-worlds interpretation of quantum mechanics is an interpretation of quantum mechanics and not quantum mechanics itself. My minimalist probabilistic interpretation is likewise an interpretation of quantum mechanics. Because it is minimalist, it is compatible with all interpretations of quantum mechanics that allow for a fundamental indeterminism in the world. It is incompatible only with interpretations that view quantum mechanics as completely deterministic, as with the many-worlds interpretation. It is absolutely crucial here to understand that interpretations of quantum mechanics are empirically indistinguishable. I already quoted Anthony Sudbery to that effect in chapter 2, but it is worth repeating the quote: Ã¢â‚¬Å“An interpretation of quantum mechanics is essentially an answer to the question Ã¢â‚¬ËœWhat is the state vector?Ã¢â‚¬â„¢ Different interpretations cannot be distinguished on scientific groundsÃ¢â‚¬â€they do not have different experimental consequences; if they did they would constitute different theories.Ã¢â‚¬Â
How, then, do we decide between a minimalist probabilistic interpretation that stresses an indeterministic universe and a many-worlds interpretation that stresses a completely deterministic, albeit ontologically bloated, multiverse? Although empirics alone are not enough to distinguish the two, one consideration is, at least for me, decisive. That consideration centers on the priority of probabilities in quantum mechanics and on how we make sense of those probabilities. Historically, quantum mechanics did not begin with a many-worlds formulation and then derive probabilities. Historically, quantum mechanics began with trying to make sense of probabilistic phenomena. Then, because full deterministic causality had for centuries been elevated as a regulative ideal for physics, a way was found to interpret quantum mechanics nonprobabilistically via many-worlds. This historical priority of probabilities in the formulation of quantum mechanics suggests to me a conceptual and ontic priority: quantum mechanics is fundamentally a probabilistic theory describing an indeterministic world, and only with considerable finagling can it be interpreted as a completely deterministic theory. The minimalist probabilistic interpretation comes to terms with the probabilities arising out of quantum mechanics as such. The many-worlds interpretation begins with these same probabilities, must explain them away, but then must recover them for use in actual quantum mechanical experiments (it is the probabilities associated with measurements and not the many-worlds themselves that figure directly into quantum mechanical experiments).
Not only is the many-worlds interpretation parasitic on the minimal probabilistic interpretation; it is not even clear whether the many-worlds interpretation allows for a coherent recovery of probabilities. As Michael Dickson observes:
“Without a notion of identity across time of a world (or mind), it is unclear how probabilities can be made empirically manifest [within the many-worlds interpretation]; i.e., the connection between probabilities and relative frequencies (over time) is severed. Indeed, the very notion of performing an experiment (which inevitably takes time) is apparently unavailable without the prior notion of what constitutes the same world (or mind) over time.”
Within the many-worlds interpretation, worlds that constitute the multiverse are continually splitting in accord with the probabilities given by quantum theory. As a consequence, there is no experimental way to track those probabilities within a given world. To be sure, one can assign probabilities simply on the basis of quantum theory. But unless science is to become a purely rationalist enterprise, it is also necessary to ground those probabilities in experience. The many-worlds interpretation seems not to allow this. Indeed, experience can only take place within a world, not across worlds. There are other difficulties with the many-worlds interpretation. But my concern here is simply to address the growing sense that this interpretation is the only game in town. It is not. Yet even if it were, it would get around the problem of causal indeterminacy only to face a still deeper problem of contingencyÃ¢â‚¬â€why do we inhabit this world rather than another and why is our world chock-full of specified complexity while others are not?
Let us now return to the question of how an unembodied designer can impart information into the natural world without imparting any energy. Setting aside the many-worlds interpretation of quantum mechanics and the determinism it implies, let us grant that there is genuine indeterminism in the natural world. In that case, whether an unembodied designer works through quantum mechanical effects is not ultimately the issue. Certainly quantum mechanics is more hospitable to an information processing view of the universe than the older mechanical models. All that is needed, however, is a universe whose constitution and dynamics are not reducible to deterministic natural laws (indeed, given the imprecision inherent in our measurements, there is no way ever to establish determinism with finality). Such a universe will produce random events and thus have the possibility of producing events that exhibit specified complexity (i.e., events that stand out against the backdrop of randomness). As I have stressed throughout this book, specified complexity is a form of information, albeit a richer form than Shannon information, which trades purely in complexity (see chapter 3). WhatÃ¢â‚¬â„¢s more, I have argued that specified complexity is a reliable empirical marker of actual design. And indeed, our best empirical evidence confirms that we live in a nondeterministic universe that is open to novel information, that exhibits specified complexity, and that therefore offers convincing evidence of an unembodied designer who has imparted it with information.