A Necessary Being or Just a Collection?

In Chapter 7 of his Metaphysics (paperback, second edition), Peter Van Inwagen discusses the cosmological argument for the existence of a necessary being. He critiques traditional formulations which invoke the Principle of Sufficient Reason (PSR), and then presents some versions which rely on premises weaker than the traditional PSR. He thinks some of these versions have more merit (but none support a truly compelling argument in his view).

Below I look at one of the versions he discusses, which I found interesting and helpful for the purpose of exploring an argument I’ve been favoring. A key premise in both arguments has to do with whether the sum of everything that exists is actually an entity or being in its own right.

The first premise of the argument from Van Inwagen’s book relies on the following principle: “Every being has this feature: the fact that it exists has an explanation” (p.125). If we accept this principle then, Van Inwagen argues, it is plausible to suppose that here are no beings which are both independent and contingent. To be independent means there can be no explanation in terms of other beings: therefore, given that we demand some explanation for existence, the only answer can be that the being’s non-existence must be impossible. Its existence is necessary rather than contingent. It follows that we can construct what might be called a Monist or Pantheist cosmological argument (my labels):

Monist or Pantheist Cosmological Argument (adapted from Van Inwagen, p.125):

1. There are no independent and contingent beings (premise)
2. There is a being which is the totality of all beings: the world (premise)
3. The world is an independent being (follows from 2)
4. The world is not a contingent being (from 1 and 3)
5. The world is a necessary being (follows from 4)

I think there are good reasons to accept premise 1, but I want to discuss here the fact that even if one accepts it, many would reject the second premise. As Van Inwagen explains, a naturalist typically has no problem seeing the contingent objects in the world as mutually dependent, without any need to postulate a world-being. Traditional theists also reject premise 2, since they see the necessary being as being distinct from its contingent creation. When it comes to our actual world, I also would reject premise 2: I don’t see a good reason to consider the totality of things an additional fundamental thing (I have a post related to this topic here).

The form of this argument, however, helped bring out an implied premise in my own modal realist version of the cosmological argument (first sketched here). This is the premise that the sum of all metaphysical possibilities is a kind of totality-entity in a way that the sum of all actual things is not.

Let me lay out the argument as follows (note that while I will use the term “beings” to be consistent with the discussion above, this would not normally be my choice):

Modal Realist (or Panentheistic) Cosmological Argument:

1. There are no independent and contingent beings (premise)
2. Modal Realism is correct: all possible as well as actual beings exist (premise)
3. There is a being which is the totality of all beings: let’s call this the “megaverse” (premise)
4. The megaverse is an independent being (follows from 3)
5. The megaverse is not a contingent being (from 1 and 4)
6. The megaverse is a necessary being (follows from 5)

So the question I want to ponder is this: is premise 3 here more defensible than the similar premise 2 in the earlier argument? I think so, but I haven’t argued directly for this before.

My preliminary thoughts go something like this.

Certainly, the megaverse is metaphysically exhaustive in a way that the actual world is not, and so has a better claim to independence and necessary existence. But could it just be a “mere” collection of possible and actual beings rather than a totality entity/being?

Recall that in David Lewis’ model of modal realism, the set of possible worlds is indeed a mere collection. For Lewis, each world is causally distinct (the actual world is simply the one we find ourselves in). In my preferred model, on the other hand, there is an intimate connection between the actual and the possible at the level of each individual event. Each causal event is an actualization of a possibility. So there is a continual closeness or adjacency therefore between what we see as actual and what is possible. The actual “world” is a causal network of events embedded in a larger framework of possible events (more related discussion here). So the “megaverse” is not a collection of distinct worlds, but is better seen as this transcendent framework which our ever-evolving actual “world” fits into.

Still, this doesn’t seem to rule out that at the level of events (rather than “worlds”); we could view the megaverse as a “mere” collection.

I’ve been scratching my head on this, and have some more thoughts, but I’ll save them for a follow-up post.

Logos vs. Chaos, Part Two

This the second of two posts -- the first post is here.

So, according to Timothy O’Connor, there are two kinds of necessary being (NB, for short) which could provide the right kind of ultimate explanation for our contingent reality: a personal agent (Logos), or an impersonal primordial world generator (Chaos). In Chapter 4 of his book, Theism and Ultimate Explanation, he examines these options to see which provides the better explanation.

O’Connor considers several types of Chaos models. He distinguishes between single-stage (all creation at once) and multi-stage models. For single-stage models, he describes three versions: Immutable Chaos, Abundant Chaos, and Random Chaos.

In the Immutable Chaos model, the world is necessarily a product of the NB’s nature. On reflection, O’Connor finds this model hard to credit: it seems unlikely that our large, highly arbitrary and extensively structured world should need to follow from the NB’s nature. Also, the idea of a single world being a necessary product runs afoul of the need for the NB to provide a non-fully contrastive cause, as discussed earlier in the book.

In the Abundant Chaos model, many worlds are produced, including ours. Oddly, to me at least, O’Connor uses the same sort of objection he had to Immutable Chaos to argue against the plausibility of this model:

“For surely it is no explanation of why the nature of Chaos is ordered to this effect to say that it is because it is ordered to that entire range of effects, and this is one of them. Really, the situation is just made more problematic. How is it that a highly unified source could be causally ordered to just this effect, and also to just that one, and… (Emphasis original, p.94)”

This analysis seems wrong to me. If there are many worlds (perhaps an infinite number of them), then arguing from the highly particular nature of our world loses its force. It is, after all just our local neighborhood in a great expanse. In the extreme case, one might see the NB giving rise to (or, in the panentheistic mode, being constituted by) all metaphysically possible worlds. I’ll return to this point again below when I review O’Connor’s later invocation of the fine-tuning argument as a guide to choosing the correct NB model.

The third Chaos option is labeled Random Chaos. Here, the NB generates a world from its nature utilizing something like a random number generator. O’Connor sees this as having more merit than the other Chaos options. To me, however, this didn’t seem much different in a crucial sense from the Abundant Chaos model: myriad possibilities are viable and we find ourselves in one of them.

O’Connor also considers multi-stage versions of Random Chaos. For instance, perhaps at the termination point of a world (like Big Bang/Big Crunch points) a successor world is generated with random alterations from the previous entry. He says such models may have merit although the mechanisms are somewhat obscure to him (for a multiverse proposal with this feature, see physicist Lee Smolin’s first book). I didn’t find the single-stage/multi-stage distinction very important in all of this.

At this step in his discussion, O’Connor sees the Random Chaos model as the best alterative to the Logos (or agent) model. How might one choose between them? At this point, O’Connor invokes the fine-tuning argument. He believes that the fine-tuning argument fails as a stand-alone design argument for the existence of a NB, but given independent motivation of the NB from the cosmological argument from contingency, he thinks it can help choose between the options as to the NB’s nature.

I’ll skip reviewing some of O’Connor’s discussion of the fine-tuning argument, which would be familiar to those who have spent time on it (a nice summary is on these posts at Parableman). He concludes that it does succeed in elevating the Logos option over Chaos. He thinks the particular nature of our world argues for a NB which is an agent acting on purpose-driven intentions in creating our world. I disagree with this conclusion because I believe the (independently motivated) existence of a multiverse removes the force from the fine-tuning argument. O’Connor, on the other hand, concedes that the multiverse idea dilutes the force of fine-tuning, but he doesn’t think it eliminates it.

O’Connor does say that the strongest multiverse concept (for the purpose of countering fine-tuning) is one which invokes the existence of myriad metaphysically possible worlds as opposed to the cosmological models offered to-date by physicists (I agree). In evaluating whether the multiverse defeats the fine-tuning argument, he first offers the objection that positing the multiverse is less parsimonious then positing a NB designer -- however he concedes this is not obviously persuasive. Then he offers a second objection, which I found dubious, saying that a multiverse option would be but one of many “totality” possibilities, many of which may not contain intelligent life. But if the multiverse is the complete metaphysical manifold, then this objection is ill-founded. I conclude that O’Connor does not find a compelling objection to the metaphysical multiverse model as a defeater for the fine-tuning argument (discussion takes place on pages 107-108). Therefore, the fine-tuning argument fails to provide a basis for preferring Logos.

Further, we should recall the rationale we used for positing some kind of NB in the first place, which is to ground real modal truths of necessity and possibility. Given a robust modal realism (which O’Connor and I both endorse) I think we should conclude that the NB is the source for (and should perhaps even be identified with) the full manifold of metaphysical possibilities. (By the way, I don’t think it matters for this discussion whether the other possibilities are concretely realized or abstract.) So, I conclude that a “Chaos” model of the NB is the preferred model. I judge this on the grounds of parsimony, since the personalized Logos model needs to have the same metaphysical scope, but also adds extra features (purposes, intentions) which are unneeded.

For now, I will not be blogging about the last two chapters of O’Connor’s book. Chapter 5 deals with the matter of how many worlds a Logos NB would create and discusses how this applies to the problem of evil and other matters. Finally Chapter 6 is of less purely philosophical interest as O’ Connor discusses issues in Christian theology and the problem of reconciling the God of metaphysics to the God of the Bible.

So, this concludes my discussion of O’Connor’s though-provoking book. I will be following up, though, with some additional reflections on this topic of inferring the nature of a transcendent necessary entity from what we know about our world. Also, I understand that Peter Van Inwagen discussed this Chaos/Logos topic in his Metaphysics, so I may have some thoughts after reading the relevant passages (which I intend to do soon).

[UPDATE: 2 September 2008] In the paperback 2nd edition of Van Inwagen’s book (linked above), he concludes that the multiverse objection is decisive against the fine-tuning argument. He says: “It is the possibility of an interplay of chance and an observational selection effect that is the undoing of the teleological argument in the form in which we are considering it (p.158).” When subsequently discussing the options of Logos and Chaos, he therefore sees no reason to prefer one over the other:

“As far as our present knowledge goes (aside from any divine revelations various individuals and groups may be privy to), we have no reason to prefer either of the following two hypotheses to the other:
• This is the only cosmos, and some rational being has (or rational beings have) fine-tuned it in such a way that it is a suitable abode for life.
• This is only one among a vast number of cosmoi (some of which are -- a statistical certainty -– suitable abodes for life) (p.161).”

To be sure, he does not give any credence to my argument that the chaos option is preferable to logos on grounds of parsimony.

I might also mention some context for those who haven’t read the book: Van Inwagen, while being one of our most prominent philosophers who is also a theist, finds none of the traditional philosophical arguments for theism (ontological, cosmological and teleological) compelling.

Logos vs. Chaos, Part One

Timothy O’Connor has written an interesting book full of meaty metaphysics (plus some theology): Theism and Ultimate Explanation: The Necessary Shape of Contingency. In this post I will focus on the sections of the book which were of greatest interest to me, but I enjoyed reading the whole thing. The central goal of the book is to present an up-to-date cosmological argument from contingency for the existence of a transcendent necessary being.

Like O’Connor, I’m interested in pursuing the quest for ultimate metaphysical explanation, and I agree with him that it is the apparent contingent nature of our world which demands an answer (“why this?”).

Unlike O’Connor, however, I don’t come to this quest as a traditional theist. I’ve long thought that the purposes and activities of a personal god cannot provide an explanation without raising equivalently difficult questions. So, an intriguing part of the book is where O’Connor discusses and compares two models of a transcendent necessary being: the personal and the impersonal. Or as he labels them: Logos and Chaos.

Before getting to this point, O’Connor lays the groundwork. In the first two chapters of the book, he discusses the case for modal realism and sketches a theory of modal epistemology. I agree with him that our concepts of possibility and necessity are so fundamental to explanation and reasoning (as well as to everyday life) that they are grounded in real modal truths. With regard to his account of how we come to know such truths, I found the discussion interesting, but tentative, and will leave this aside for now.

In Chapter 3, we begin to look at where the quest for explanation will take us. If the contingency of the world is acknowledged, and the necessities we come to know (if imperfectly) are real, what explains this situation? After commenting on the tendency for many philosophers to dismiss this sort of question as unanswerable, O’Connor turns his attention to the options on offer (I would note that Graham Oppy, in his review, disagrees with O’Connor in that he counts “contingent world as brute fact” as one of the options).

First, O’Connor discusses the common naturalist option which asserts that there is an infinite causal chain of contingent things, and that no further explanation is needed. O’Connor says even if the infinite causal chain is the right model, we can still coherently ask for an explanation of why THIS chain of things, and not another. The rejoinder at this point is to point out that a fully contrastive explanation of why this chain is as it is, if it existed, would convert the contingent chain to a necessary one. But O’Connor responds by saying that we can still seek a complete explanation, even if it is a non-fully-contrastive one. He says the only way to do this will be to ground contingent reality in a transcendent-cause explanation that is non-contrastive.

But how can a necessary being (NB, for short) provide the right kind of explanation? The necessity of the cause would seem to “prove too much” and convert the contingent chain to necessary status. Indeed, this is a standard critique of the Leibnizian version of the cosmological argument, which features the Principle of Sufficient Reason (PSR). O’Connor concedes the effectiveness of this critique, and points out that his version will not invoke the PSR. But isn’t the PSR or something like it implied by the desire for an explanation of the contingent world?

O’Connor says he can avoid this as he isn’t offering a contrastive explanation (e.g. why this world instead of any other?) He says he can do this by invoking an explanation based on a model of the NB as an agent:

"In the context of an agent who exercises a capacity to freely act for a purpose, explanation is grounded in an internal similarity relation of the content of the prior purpose to that of the effective intention. To understand why an intention is freely generated, one need only identify its reasons-bearing content. This contrasts of course, with a mechanistic model of intentional action on which an agent’s purposes or desires or beliefs explain the choice, or formation of an intention, solely in terms of an external, causal relationship to it. But it is readily understandable in its own terms. (p.83)"

So the NB-agent’s creative activity of forming an intention, based on reasons (which is then satisfied by the creation of a contingent world), need not be seen as a necessary sequence of steps:

"And it would be a confusion to suppose that we needed a further explanation of the generation of the intention – for that is just the agent’s exercise of control over his state of intention and its product. (p.83, emphasis original)"

But still we want to ask: why intend contingent world C instead of C*? Well, we can’t ask this question, because an answer would provide for C being inevitable. But that doesn’t mean we haven’t given an explanation at all, according to O’Connor. The buck stops with the (contingent) intentions of the NB-agent.

To me, this seems a difficult stance. A naturalist might ask why we can’t stop the explanation at an earlier stage (before invoking the NB); O’Connor wants to stop the discussion at a later stage – but either way we eventually reach a “conversation-stopper”. Do the extra steps add enough value to prefer this NB-agent view, especially given all the (controversial) extra apparatus of intentions, purposes, activity, etc.?

Well, we need to recall the NB view’s main advantage, which is that it helps underwrite the reality of modal truths (there exists both contingent and necessary aspects to reality). But, do we really need the agent apparatus to provide this advantage?

To his credit, O’Connor sees the existence of another alternative (or set of alternatives). He starts by noting that, in addition to his agent model, there are many instances where we accept causal explanations even when they are not fully contrastive ones. Specifically, we do this in cases of “indeterministic mechanistic causal processes in the natural world, and the kinds of scientific explanations that may be given for them.” He gives the example of a disease which, when untreated, leads to a debilitating outcome 27% of the time. If the indeterminism here is fundamentally irreducible, we still have a good explanation for the outcome (the untreated disease) even if not a fully contrastive one (we can’t say why the outcome manifested itself this particular time).

So, the question becomes: could the necessary being be an impersonal “fount” of creative possibilities, which generates one or more contingent worlds?

{End of part one}

Note: I’m skipping over some sections along the way – for instance I’m leaving aside the discussion of the Spinozan view here (could our single world be the necessary being?). Also, O’Connor mentions in passing that it is hard to rule out a model of multiple NB’s -- although there seems to be no good argument for it either.

Conway and Kochen vs. Determinism

John Conway and Simon Kochen are out with a paper called "The Strong Free Will Theorem" (HT), updating and "strengthening" their earlier paper (discussed here). Recall that the theorem begins with axioms which, while idealized, flow from accepted aspects of quantum theory and relativity and then concludes that if humans are assumed to be free in setting up experiments, then particles have the same kind of freedom in selecting among experimental outcomes. The theorem also serves as another argument toward ruling out hidden-variable interpretations of QM.

This paper presents a "stronger" version of the theorem, by showing it still works if one of the axioms is loosened, but otherwise the thrust is unchanged. In keeping with the earlier paper, though, the authors add to the formal argument some provocative philosophical comments, which I enjoy. Here's how the paper concludes:

"Although...determinism may formally be shown to be consistent, there is no longer any evidence that supports it, in view of the fact that classical physics has been superceded by quantum mechanics, a non-deterministic theory. The import of the free will theorem is that it is not only current quantum theory, but the world itself that is non-deterministic, so that no future theory can return us to a clockwork universe."

3 Links

I'm currently reading Theism and Ultimate Explanation: The Necessary Shape of Contingency by Timothy O'Connor. I'm very interested in the cosmological argument from contingency, and this book is an up-to-date take on that and related metaphyical issues. I hope to have a post on this at some point but in the meantime here is a review from a naturalist's perspective by Graham Oppy (HT: sideblog at FQI).

I really enjoyed this insightful cartoon posted at Cosmic Variance along with the comments by Sean Carroll (the original source for the cartoon is here). No offense to cosmologists, but for purely philosophical reasons I think it is best to identify the actual world with the observable or causally connected universe (not that I think that's all there is, but because the regions we assume exist beyond the observable have a different ontological as well as epistemological status - see a related post here).

Finally, I want to post a friendly link to the discussion forum at Panendeism.org. Panendeism, as I understand it, is meant to be like Panentheism, but with the "deism" label stressing that this is a worldview arrived at through reason, without reliance on authority or revelation.

Sir John Templeton Dead at 95

Sir John was one of great investors and philanthropists (here is the press release from the John Templeton Foundation site). I had the opportunity to hear him speak a couple of times and I conversed personally with him for a few minutes once (about interest rates and currency markets) and he was extremely insightful.

The foundation’s work, a big part of which looks to bridge religion and science, is sometimes controversial (for more see this old post), but it has funded many worthy efforts, including for example the recent seeding of the Foundational Questions Institute. I hope they continue to fund this kind of scientific research. In terms of critique, my personal wish would be that they divert some funding from religion and theology to non-religiously motivated study of metaphysics and philosophy of science. In any case, I think our progress on the big questions should be better because of the existence of the foundation. So, thanks to Sir John and to those who work at implementing his vision.

Reduce Everything to Space-time?

[UPDATED 25 Sept.2009: Fixed Links]
I want to quickly comment on an interesting post by Justin at Panexperientialism. In it he reviews a book by Freya Mathews (called The Ecological Self) and also discusses a draft paper by Jonathan Schaffer (Spacetime the One Substance). Please check out his post, which discusses many aspects of Mathews’ ideas in particular beyond what I’m picking up on here (I have not read the book).

Both Mathews and Schaffer advocate a monistic metaphysical view where matter is effectively reduced to space-time.

I agree with these authors that the dual scheme of {space-time container plus material objects} must be rejected, but think they are slightly off-track in wanting to reduce the properties of matter fields to space-time (at least space-time anything like we currently think of it).

These brief comments focus on the relationship of this idea to the work of theoretical physicists. Mathews acknowledges that an early attempt to derive this reduction from general relativity failed (Wheeler’s Geometrodynamics), but still likes the metaphysical vision for philosophical reasons. In his paper Schaffer argues toward a similar goal, but along the way I think he overstates the degree to which GR and (especially) quantum field theory as we know them are congenial to this vision. QFT has matter fields housed in a separate space-time container. In GR the matter and space-time are dynamically intertwined, but the fact that you can model the geometry while leaving out matter shows that they remain distinct.

In some ways the quest for a theory of quantum gravity can (should?) be viewed as a quest for a monistic theory which is rid of the dual scheme. I continue to try to follow the different theories as a layperson to see how they come down on this issue.

String theory: originally an extension of QFT which retained the feature of having fields on a background space-time. Has evolved in many ways over the years and maybe can overcome this starting point (?).

Loop Quantum Gravity and Causal Dynamical Triangulations: these seek to formulate a quantum version of space-time with the promise of integrating matter into the picture later. I’m not sure if this promised integration would be more monistic than GR.

Causal Sets; Quantum Causal Histories/Geometrogenesis; Internal Relativity; Quantum Computing and Condensed Matter-based approaches: these programs seem best on this question as they try to specify a monistic underlying micro-theory from which space-time and matter fields as we know them may simultaneously emerge.

I would note that if the latter sort of approach works, it doesn’t support Schaffer’s advocacy of priority monism (see my previous post on this topic). The underlying network would not be a very coherent whole, but a fairly ill-behaved evolving pluralism of micro-events. Even though Schaffer wants to overcome the container/object scheme, his view of space-time as the holistic fundamental object still has a bit of a hangover from the container idea in my opinion.

CDT in Scientific American

Surprising and promising results have not come too frequently in quantum gravity research, but the Causal Dynamical Triangulations program led by Renate Loll, Jan Ambørn and Jerzy Jurkiewicz had an exciting moment in 2004. A computer simulation showed that a space-time model with the right dimensionality arose from a path integral superposition of fairly generic microscopic geometric building blocks. The team has kept up a steady stream of research investigating and seeking to extend this result, and now they have published a popular article in the latest Scientific American.

I recommend the article, if one has access to it. I also discussed (as best I could) the basics of the CDT approach in this earlier post, so I won’t repeat all that here. Also, I coincidentally had just read a recent paper by the team which showed how they have generated not just the right dimensionality, but also specifically find a de Sitter universe in a simulation.

The thing I find most interesting about CDT is that it may give some evidence that selecting asymmetric time and causality as fundamental features is important in quantum gravity (a caveat is that their simulation uses globally synchronized time, and I wonder if they can relax this assumption).

One reason to be cautious is that CDT at this point only deals with space-time, not matter. Like in Loop Quantum Gravity, there is an expectation that matter fields can be coupled to the theory later on. This is in contrast to research by Fotini Markopoulou and Olaf Dreyer (see posts here and here), who think that it is the matter fields which are to emerge from a micro-quantum substrate, and that space-time geometry is to be inferred from the matter. If this works, it seems conceptually more appealing, since you’ve dealt with both space-time and matter at once. (See also this recent FQXi article on Markopoulou and Dreyer).

One last interesting aspect which I hadn’t thought much about before was discussed in the recent paper. This is the fact that CDT (and I assume other “emergence-style” programs) need to be investigated by computer simulation, rather than by deriving a specific analytical result through mathematical formalism. This essentially means giving up on the traditional idea of a “theory of everything” which can be written down in a set of equations. The CDT team doesn’t see this as a weakness, and cites condensed matter theory (see also this post) for example as a field where emergent behaviors are profitably studied without the possibility of precise description at the micro-level. They also invoke the idea of “self-organizing” complex systems. Here’s a quote:

“Think of quantum gravity as a strongly coupled system of a very large number of microscopic constituents, which by its nature is largely inaccessible to analytic pen-and-paper methods. This is no reason for despair, but a common situation in many complex systems of theoretical interest in physics, biology and elsewhere, and merely calls for a dedicated set of technical tools and conceptual notions.”

Sensitivity to a Single Photon

Patrick Suppes and Jose Acacio de Barros wrote a paper called Quantum Mechanics and the Brain (HT: Clark’s sideblog). Obviously the title was irresistable to me. However, it turns out that the title is a bit misleading. The paper has a few paragraphs discussing previous proposals regarding the role QM might play in the brain and briefly gives the authors’ views on these -- frankly without saying anything new. The paper then shifts to what is its main focus, which is to highlight an interesting instance where there is a good indication that biological systems do exploit quantum level phenomena. This is the animal eye’s demonstrated ability to react to a very faint light signal: specifically the ability to detect and respond to a handful of photons, and probably a single photon.

The paper describes previous experimental results with insects and animals. Sensitivity to single photons was inferred using statistical methods in some of these tests (some of the older references in the paper were also discussed in this Usenet posting from 1996). The authors then outline a proposal for future experiments to confirm the results by utilizing the technological ability we now have to shoot single photons.

This result would demonstrate that organisms are sensitive to the quantum realm. While we remain far from understanding the role quantum effects might play in conscious experience, this is another step toward exploring the subject.

A Russell-Style Objection to Many-Worlds

In the standard formulation of Quantum Mechanics, there are two processes: in the absence of a measurement, a system described by a wave function evolves continuously and deterministically; when a measurement is made, the system instantaneously takes on a specific value according to the property being measured (often called a “collapse”). Everett’s Relative State formulation of Quantum Mechanics, the most famous version of which is called the Many-Worlds Interpretation, seeks to drop the collapse process from the theory. The interpretation then struggles to explain why we have the experience of measurement outcomes that we do (e.g. all outcomes are still happening but in many different worlds).

My common sense objection to the interpretation has always been based on the roots of the theory in scientific experimentation. QM was formulated to explain the phenomena we observe in the microscopic realm, including, obviously, the outcomes of measurements! It seems perverse to drop the process describing measurement.

Why then is the interpretation fairly popular? It is because the collapse process seems mysterious and “unphysical”, while the evolution of quantum systems in the absence of collapse is mathematically well behaved and, despite the many new complications and nuances, remains closer in spirit to the traditional dynamics of “matter in motion”.

I don’t know for certain whether Bertrand Russell ever wrote about Many-Worlds (it’s doubtful given the chronology – he died in 1970 and Everett’s 1957 work wasn’t widely discussed until the 70’s), but he made a careful philosophical study of physics (and all contemporary science) and I think his work helps clarify my objection to the interpretation.

Recall that Russell’s project was to analyze the data and methods of science and how they relate to human experience. He described how the realms of physical theory and experience can be brought closer together in an ontology of events and their causal relations.

Here are some quotes from Russell’s Human Knowledge (which I’m reading now, inspired by Carey Carlson’s references to it in his book):

Here’s the quote that first made me think of Many-Worlds: “Mathematical physics contains such a superstructure of theory that its basis in observation is obscured.” (p.41)

We need to remember: “…a datum for physics is something abstracted from a system of correlated psychological data.” (p.59)

And: “There is here a peculiarity: physics never mentions percepts [experiences] except when it speaks of empirical verification of laws; but if its laws are not concerned with percepts, how can percepts verify them?” (p.219)

The answer is that it doesn’t make sense to see the entities described in physics as something disjoint with the events of first-person experience. Science can be seen as describing a causal network which connects up in consistent ways with experiential events (think of how the human experiences of telescopic images provide the raw material and ongoing testing for astronomical and cosmological theories). Therefore, reality is best viewed as a web of events consisting of the directly experienced and the indirectly inferred, the latter of which are the usual target of physical theories.

In quantum physics, the wave function is derived as a description of reality which connects consistently with experiential observations. It has this in common with other theories of physics. Of course with quantum physics there is always a twist. The twist is that, unlike the entities of other physical theories, the wave function can itself only be viewed as a fully objective physical entity if we pretend that the measurement events don’t exist! Here we are unavoidably given a choice to elevate the reality of the directly experienced events (measurements) or the inferred physical entity (the wave function). The events must take priority since they are the part of reality which is not known through inference. In fact quantum measurement events are the best candidates to be the raw material for a consistent ontology of the concrete world.

Given all this, I think the best interpretation of quantum mechanics is a version of the relational or perspectivist approaches. In this interpretation, a relational network of measurement events (or interactions) constitutes the concrete world. All quantum systems interact (perform measurements upon each other) with no ontological distinction between the macroscopic and microscopic realms (or the human and non-human). The wave function in this interpretation describes a system’s propensities for interaction outcomes from the perspective of a particular measuring system.

Quantum Foundations Series (some highlights from posts with the Quantum Physics label):

The Duality at the Heart of Physics
I discuss how the two processes of QM are unavoidably interdependent.

Free Will All the Way Down
A paper describes the relationship between freedom at the human level and indeterminism at the micro-level.

Interpreting Quantum Probability
Is the probabilistic aspect of QM objectively real or epistemological?

The Limit of the Bayesian Interpretation
Why the Bayesian (subjective) interpretation of quantum probability is preferred but is still incomplete.

Wigner’s Friend and Perspectivist Quantum Theory
A superior interpretation of QM by Paul Merriam.

Merriam’s Quantum Relativity
How Merriam addresses the “preferred basis” problem for the perspectivist interpretation.