The main theme of August's FQXi conference was centered around the physics of the observer and so, in wrapping up our discussion of the conference, it remains to be asked if any progress was made toward a better understanding of the concept. As with just about any conference or meeting of researchers from such diverse backgrounds, it is natural to expect that each of us came in to the conference with out own opinions, ideas, and, yes, biases. It's difficult to say if anyone actually changed their mind about anything or was swayed by a reasonable but opposing argument, as a result of this conference.
Of course, science is not based on opinion. The final arbiter of science must always be carefully constructed experiment. Certainly there are some, including a few attendees of this conference, who have suggested that science, and most notably physics, has moved into the post-experiment era in which an elegant mathematical description is all that is required to prove a physical "truth." As someone who wrote his doctoral thesis on one failed attempt to reduce physics to a purely deductive exercise, I can say that this argument is actually nothing new. Physicists have occasionally entertained this idea at one time or another throughout the past four centuries. In some sense, Hilbert formally challenged physicists to do exactly this---axiomatize physics---in his famous Sixth Problem (though, it should be noted, that his actual statement of the problem can be interpreted as having a narrower focus). Experiment, however, persists as the final arbiter of physical "truth" precisely because it has the most direct connection to our senses which remains the only way in which we directly interact with the world. In other words, we expect science to interpret the world of our experience.
That being said, modern physicists know well that direct experience can often be deceiving. Both relativity and quantum mechanics---the two foundational pillars of modern physics---raise direct issues concerning the role of the observer in our understanding of the world. Indeed, some of these issues have been at the forefront of physics for four centuries. After all, it was Galileo (and not Einstein as many incorrectly believe) who developed the principle of relativity which states that the laws of physics should be the same in all inertial reference frames. Galileo's thought experiment involved someone in a windowless cabin on a ship in calm waters---would that person be able to determine with absolute certainty if they were moving or not? The answer, of course, is no.
The key here is that this says something profound about the nature of the observer---all inertial observers must agree on the laws of physics, though not necessarily on the specific outcomes of individual experiments. In other words, the universe must be self-consistent, i.e. it must operate in the same manner for all inertial observers.
In the quantum realm, however, we learn that the observer can affect the outcome of experiments. Is this a violation of the principle of relativity, then? Not necessarily. The principle simply says that the same rules must apply to all inertial observers. Variation in the outcomes of individual experiments is allowed as long as the rules that led to those results are the same in each case.
And that, right there, is the real key to this entire discussion. Science has very carefully built a structure and methodology for addressing these issues over the course of those same four centuries that has been a consistently reliable predictor of future outcomes. Arguable this methodology---which defines modern science---is the greatest achievement in the history of humanity. One of its hallmarks, particularly in physics, is its emphasis on rigor and clarity (perhaps due to the close ties between physics and mathematics). We need to know what it is that we're talking about otherwise we end up just talking past one another, sometimes without realizing. John Wheeler's argument that we need to move beyond defining things sounds almost Aristotelean in contrast. Indeed, one way in which this grand enterprise we call science has progressed has been by agreeing on definitions and then testing those definitions. In theory, this should lead to further refinement of those definitions. Unfortunately, science, including physics itself, has become fractured enough that universal consensus on some definitions---including "the observer"---is lacking.
A classic example of this that recalls several conversations and exchanges made at the last few FQXi conferences, is the concept of entropy. Despite the fact that Boltzmann and Gibbs clarified the definition of entropy in the 19th century (which Jaynes so elegantly further clarified in the mid-20th century), there are those that persist in defining entropy via a Clausius relation. While such a relation is certainly a valid manner in which to describe certain types of entropy, it is abundantly clear from even a cursory reading of Boltzmann, Gibbs, and Jaynes that such a relation does not work as a universal definition. It is simply a way in which entropy behaves under certain, limiting conditions and tells us nothing about its actual nature.
At any rate, this brings us back to the question at hand. What is an observer? For that matter, what is an event? Does the latter require the former? These were both questions that ostensibly were to be addressed at the conference. In fact an entire session was dedicated solely to the observer. During the Q&A session of the associated panel, Jeremy Butterfield pointed out that modern philosophy, via Frege, Russell, et. al., had established set definitions for many of these "truths," relations, etc. that are generally free from the types of disciplinary bias you get in science. For example, those who prefer to define entropy via a Clausius relation are typically those who work in areas reliant on classical thermodynamics. This comment by Butterfield raises a few interesting points. Notably, it sets a definitive role for philosophy in the advancement of science. Scientists are often dismissive of the role of philosophy, but, just like science, philosophy is not monolithic. While it certainly contains its share of post-modernist rubbish, it also includes some very important and rigorous work including, as Butterfield pointed out, in defining certain important terms used by science. Having these independent, broadly developed ideas and definitions can help to relieve some of the tensions surrounding some of these definitions.
All of this is to say that we still have no consistent definition of an observer, per se. Even Butterfield did not offer one. The session and the subsequent panel did little to clear the fog on this issue, Butterfield's comments notwithstanding. That's not to say that the others didn't offer interesting and cogent opinions. David Wolpert, for instance, suggested that an observer, whatever one is, must be in a non-equilibrium state. While this is an intriguing idea, the other panelists showed no inclination to jump on that bandwagon. (It bears mentioning that, if observers automatically come equipped with a reference frame and since reference frames naturally break some symmetry when they are introduced to a problem, then Wolpert's idea might be the germ of something more generally valid and useful.)
At any rate, Jim Hartle offered up an amusing anecdote at one point that could be a kind of metaphor for the session. He mentioned that Murray Gell-Mann once compared something to "sticking a pin in the I Ching, but no one understood what the hell he meant." This may sound like a harsh indictment of the session, but it shouldn't be taken in that way. In fact, the session had the essence of one of those good, working sessions from legendary physics conferences of old like the Shelter Island Conference in 1947 (which Oppenheimer felt was the best conference he had ever attended). In that sense, the session lent a feeling to the proceedings that this---the FQXi conference---was, first and foremost, a working conference. And that is what makes these conferences so unique. Some of the brightest minds in physics, philosophy, neuroscience, mathematics, biology, computer science, and elsewhere communicate with one another, poking and prodding at the heart of difficult questions, nudging the scientific process along. This is where the real work gets done.