Chapter 1: Introduction to the Many Worlds Debate

The theoretical physicist and science communicator Sabine Hossenfelder has produced several engaging videos discussing the Many Worlds Interpretation (MWI) of Quantum Mechanics. While her content is captivating, I believe Hossenfelder makes several critical errors in her understanding of MWI (and science more broadly). This document serves as a critique of her perspective.

I will not provide an overview of either MWI, the alternative Copenhagen Interpretation (CI), or Quantum Mechanics in general, so prior knowledge is assumed.

Let’s dive right into the critiques I wish to present. In her video titled "Does the Many Worlds Interpretation make sense?", Hossenfelder asserts:

"The primary issue with many worlds is explaining why an observer doesn't branch with those worlds, hence witnessing all outcomes, yet only experiences one at random. I’ve yet to find a satisfactory explanation for this."

This assertion is not a significant challenge for MWI; in fact, it isn’t a challenge at all, as observers do branch along with these worlds. Earlier in the video, she provides the example of a beam splitter, which allows a photon to take two distinct paths in a superposition of possibilities:

Illustration inspired by the original image from Sabine Hossenfelder

Both pathways present a 50% probability. However, once an observer—let's name him Bob—measures the photon on one of these paths, he is certain of its position with 100% probability. In the CI, this corresponds to the wave function collapse, resulting in one Bob witnessing a single outcome. Conversely, in MWI, both paths are realized in parallel universes, each containing its own Bob. For simplicity, let's consider two universes: one where the photon moves to the right and the other where it moves downward. Each universe has a Bob, and upon observation, they see the photon taking different paths. Thus, Bob does indeed "branch with those worlds" and collectively sees all outcomes.

Regarding this experiment, Hossenfelder states:

"However, once you measure the particle on the right side, you know it’s not on the left side. This means you have to update the wave function."

(This assertion is incorrect, but I will elaborate on that later.)

This update of the wave function, often referred to as the collapse or reduction of the wave function, is a fundamental aspect of quantum mechanics. Neglecting to update the wave function leads to incorrect probabilities.

(Again, this is not accurate, but further explanation will follow.)

Hossenfelder later critiques supporters of MWI by claiming:

"Proponents of many worlds often assert that their interpretation is simpler because it eliminates the collapse postulate. However, as discussed earlier, you require the collapse postulate to calculate probabilities."

This is misleading! The probabilities are derived from the Born rule. In MWI, these probabilities remain consistent before and after an observation, representing the fraction of universes corresponding to each specific outcome. There is no necessity for any collapse in this framework. Uncertainty exists solely in the perception of individual observers!

To illustrate, let’s revisit the photon experiment. From the MWI viewpoint, we can frame it as follows: before any measurement occurs, there are two parallel universes (again, for simplification). In one, after interacting with the beam splitter, the photon moves right; in the other, it travels downward. Both parallel universes include a Bob. Upon measurement, Bob 1 sees the photon on the right, while Bob 2 sees it moving downwards.

This demonstrates that the probabilities do not alter post-measurement: two parallel universes existed initially, and they persist afterward. The transformation lies in the fact that, prior to measurement, Bob 1 and Bob 2 were the same person: Bob. Initially, Bob was uncertain about the photon’s path, but Bob 1 and Bob 2 now possess knowledge regarding their respective paths. Bob believed there was a 50% likelihood of the photon traveling right, while Bob 1 now perceives this probability as 100%. This reflects a change in probability, but only in Bob and Bob 1’s understanding of the circumstances: it shifted because Bob 1 has discovered the universe in which he exists. The actual probabilities—the ratios of parallel universes with specific outcomes—remain unchanged.

Therefore, MWI does not require a collapse postulate. CI does, which adds complexity (in terms of Kolmogorov Complexity) compared to MWI. Superposition occurs when the Schrödinger Equation has multiple solutions, which merely describes multiple (parallel) universes. CI attempts to predict a single universe by incorporating the collapse postulate, but without supportive evidence, this introduces unnecessary complexity. Occam’s Razor favors MWI!

Next, we turn to Hossenfelder’s video titled "The Multiverse: Science, Religion or Pseudoscience?" In this video, she asserts (note that she references various multiverses, including that of MWI):

"The problem with these different multiverses is that they propose the existence of something unobservable: those other universes. Not only can you not see them, but you also cannot interact with them in any way. They are completely detached from ours. There is no observation you could make to suggest their existence, not even in principle. For this reason, suggesting that other universes exist is unnecessary to explain our observations, and therefore is something a scientist should avoid. Making an unfounded assumption is akin to proposing the existence of an unobservable deity, a flying spaghetti monster, or an omniscient dwarf residing in your wardrobe."

By implying (and later explicitly stating) that multiverses lack scientific merit, Hossenfelder’s statement contains a critical flaw: MWI does not posit the existence of parallel universes; rather, it is the Schrödinger Equation that predicts them. The Schrödinger Equation indeed produces other predictions that can be tested (and have been). As previously stated, MWI discards a postulate (the collapse postulate). In my view, MWI elucidates observed phenomena better than CI since MWI remains deterministic and does not suggest non-locality.

Believers in these other universes maintain that they exist because they emerge from their mathematical formulations. The mathematics describe observable phenomena, leading to the inference that everything else their mathematics depicts must also exist. This conflates mathematical models with reality.

It remains somewhat ambiguous whether this statement also pertains to the Schrödinger Equation predicting parallel universes, but I will address it nonetheless. Science constructs theories, and if the most robust theory available to explain an observed event also predicts something unobserved, that prediction warrants serious consideration. It does not imply that the unobserved entity—such as a parallel universe—must exist, but it does increase its likelihood. I recognize that this is somewhat vague, yet Hossenfelder asserts that science makes no claims regarding the existence of parallel universes, when, in fact, a theory can indicate their existence.

Hossenfelder later reiterates her earlier point about the inability to observe other universes:

"The second objection I encounter is that we can only observe a segment of our universe since light requires time to travel, and it has only traveled so far since the Big Bang. However, no one would argue that this means the universe ceases to exist outside our observable region. Certainly not. No one claims that if something is unobservable, it doesn’t exist. The critical point is that if you can’t observe it, science has nothing to say about its existence."

So, does science not address whether anything exists beyond the observable universe? This quote seems to imply that notion directly, and I find it challenging to believe that this is Hossenfelder's actual stance. Regardless, the larger issue is that this statement misrepresents the essence of science. Observation is undeniably crucial to scientific inquiry, but its role is to potentially falsify existing theories rather than serve as a definitive measure of whether something is scientific or not. In essence, to reiterate my earlier argument, a theory is not unscientific simply because it cannot be observed in principle.

In conclusion, I appreciate your reading this critique!