Complexity of science or science of complexity

 kw: book reviews, nonfiction, science, scientists, scientific method, complex systems

The opening salvo in Dr Giorgio Parisi's book In a Flight of Starlings: The Wonders of Complex Systems is the story of how he and his colleagues studied starling murmurations.

When I lived in Oklahoma I would see these amazing formations in the sky in late summer and fall. A flock like this can have tens of thousands of birds, that swoop about like parts of a sky-borne animal. The author's team used multiple synchronized cameras shooting frequent images, which allowed them to identify individual birds and their relationships in three dimensions. As time passed and cameras became better and faster, their work became more and more precise.

Starlings don't flock like other birds such as sparrows, as shown here, nor do they use "V" formations like geese. Analyzing the starlings' flights in detail showed that each bird keeps track of only a few nearest neighbors, with two surprising characteristics: Firstly, the distance between near neighbors fore-and-aft is quite a bit greater than side-to-side, and secondly, a section of the flock is denser near the boundary than in the middle. These attributes seem to be driven by the tension between maintaining sufficient confusion to avoid predators such as falcons, and the need to avoid bumping into one another. Changes of speed and direction seem to ripple through parts of a murmuration like a "wave" through a football stadium, although with less overall coordination.

The author's work is in the general realm of the mathematics of theoretical physics as applied to complex systems. In the book many mathematical concepts and operations are discussed, somehow without the need to print a single equation. His aim is to show aspects of how science is done via stories about his own work, including the work that earned him a Nobel Prize in 2021. Two chapters of the book focus on phase transitions and transitions between order and disorder. If you think about it, these are two ways to understand the same thing. The phase transitions familiar to us, such as water either freezing or boiling, are brought about as, in the case of freezing, disorder gives way to order, and in the case of boiling, one style of disorder gives way to another. A less-recognized but common transition is sublimation, in which ice evaporates directly without passing through a liquid phase. But there are also transitions between one kind of order and another, such as the conversion of graphite to diamond, or the arcane transitions found by laboratory studies of at least ten crystal structures of ice, each of which is stable in a particular range of pressures and temperatures.

While the scientific stories are fascinating, I confess that, though I am a scientist also, and familiar with mathematical physics, Dr. Parisi lives on a very different level, and I struggled to understand the mathematical concepts he discusses. The chapter on spin glasses (a concept relating to the orientation of electrons in a solid material) was a complete mystery to me. However, I could just relax and take in whatever was meaningful, and that was satisfactory. He follows that chapter with one on metaphor and its limits, which is quite comprehensible because we live lives of metaphor. Metaphors and other similitudes let us model a phenomenon, which allows us to gradually build an understanding of what is going on.

The author's aim is to foster a greater understanding of science and how it operates, and to increase public trust in science, which has been so badly eroded in recent years. I hope he understands that the problem is not with science itself, but with dishonest scientists such as the very public figure who proclaimed, "I am Science!", and thus deeply offended the great majority of Americans who could tell he had been lying to us for years. I must also point out that anyone who says something is "settled science" displays a fundamental and deep misunderstanding of science. Perhaps you've heard the saying: "Half of all medical 'knowledge' is shown to be false about every ten years. The trouble is, we never can tell which half beforehand."

This is true. I have a small journal of medical mistakes I have suffered, one of which had me at death's door until I wrested control of my medical care from a mistaken doctor and "jumped over some fences" to get appropriate treatment. Medical science is admittedly at a level of complexity far beyond the "complex systems" studied by a physicist, but the principle is similar. In the next-to-last chapter in particular we read how an incorrect metaphor often leads a scientist to revise and develop a better metaphor, over and over, until the model and the system correspond, and you can say, "Now I understand this system." The stuff that is interesting in science is where we don't yet have full understanding, and still find things that make us say, "That's funny…"

Science is a human enterprise. It is saved from disaster and banality by the self-correcting nature of the process. The backing-and-forthing that goes on can be frustrating, but we haven't produced anything better. It is extremely unlikely that a better process is possible, no matter now much smarter we get, or how smart our so-called-AI (I prefer to say SI for "simulated intelligence") systems become.