Crystals are among the most beautiful objects in the natural word. They are well understood, ubiquitously used and much admired.
And yet the way scientists define them is spectacularly dull. The International Union of Crystallography defines crystals as structures that produce a diffraction pattern with discrete points.
In other words, these objects are defined by the single process used to measure them. If it doesn’t produce the required diffraction pattern, it ain’t a crystal.
Today, Julyan Cartwright at the University of Granada in Spain and Alan Mackay at the University of London in the UK argue that this definition is shortsighted and unnecessarily limiting.
They point out that the convergence of crystallography, materials science and biology is opening up a new approach to the study of structure, form and function. This new science is concerned not with static shapes in stable equilibria but metastable forms that are at the mercy of the energy landscape in which they exist and the flow of information to and from the environment.
Cartwright and Mackay give as an example the structure of mother of pearl, the beautiful irridescent biomineral that certain molluscs produce as an inner lining for their shells.
This is certainly an ordered, crystal-like structure but not one that produces the necessary diffraction pattern to be classed as a crystal.
That’s the result of its complexity. Mother of pearl is formed from layers of hexagonal plates of calcium carbonate in a ‘brickwork’ arrangement. These layers are separated by sheets of biopolymers, such as chitin.
This combination produces useful properties. The organic sheets prevent the spread of cracks while the platelets provide strength. So mother of pearl is strong, hard wearing and beautiful.
But how should such a material be described and analysed? Cartwright and Mackay say that an important consideration is the information a mollusc uses to make mother of pearl; and that is determined by its genome, proteome and so on, which together they call a conchome.
Somehow, from all this complexity and self organisation, mother of pearl emerges. Nobody is quite sure how.
Nevertheless, their key point is that this structure a phenomenon of information. And that this information is a kind of algorithm or formula for producing mother of pearl, analogous to an algorithm that produces the digits of pi.
Only a science that takes this information into account will be capable of a full description of mother of pearl and other materials like it, say Cartwright and Mackay.
That’s an interesting and ambitious approach that has the potential to profoundly change the way materials scientists and biologists think about form and structure.
What’s interesting is that a similar change in thinking about form and function is also emerging in the entirely different field of robotics and artificial intelligence.
For many years, roboticists attempted to copy human abilities like walking and running and so by building devices with a central processor that controlled every aspect of movement.
This required robots with sensors on every joint that sent back signals about the state of each limb at all times. The central processor then decided on a movement strategy, calculated a trajectory for the limb and then moved it accordingly. That’s exactly how humans do it. Or so they assumed.
But that approach fails spectacularly because the problem of co-ordinating all these joints becomes computationally hard when conditions are changing, when walking outside or upstairs or breaking into a jog for example.
So roboticists have had to embrace a new approach. It turns out that humans perform many actions that are so quick that the human brain cannot possibly be involved. The tensioning, acceleration and deceleration of muscles, tendons and ligaments when you jump off a wall, for example.
All these changes in material properties occur in the blink of an eye without any involvement of the the brain. Instead, the structure, form and properties of the materials themselves carry out this task–the intelligence is built in.
In a sense, the brain outsources the control of this movement to materials themselves.
In fact, roboticists have begun to think about this kind of movement as a computation since it can be roughly equated to the amount of computational horsepower that a central processor would need to carry out a similar task. And they’ve begin to design robots based on this principle of so-called morphological computing.
That’s now beginning to revolutionise robotics. Instead of centrally controlled robots, engineers are building bots in which the intelligence is built into the shape and form of the structure. These can carry out seemingly complex tasks like walking, running and swimming with little if any computational oversight.
A key insight in all this has been a better understanding of the role of environment. Put a walking robot in a swimming pool and it is helpless. So the the shape and form alone does not confer intelligence–it is the interaction between shape and form and a particular environment that is crucial.
The way information can be extracted from the environment is sometimes be spectacular. One example is a seemingly intelligent blob that can solve a maze. But of course every maze encodes its solution in its structure. The trick is designing a simple system that extracts this information.
This element–the crucial role of the environment–does not yet feature strongly in Cartwright and Mackay’s ideas. They would be the first to acknowledge that the environment plays a crucial role in the formation of any crystal or biological structure.
But there is a sense in which the processes of crystallisation and self organisation are like the maze-solving blob: the pattern or structure is clearly result of some kind of information extraction or exchange. Such an approach may also throw light on the origin of life.
The key will be to understand and characterise the relationship between the environment, the structures that form in it and the flow of information that makes this possible.
And if crystallographers, materials scientists and biologists want to solve it, they might do well to team up with the roboticists, engineers and evolutionary biologists who are toying with very similar ideas.
Ref: arxiv.org/abs/1207.3997: Beyond Crystals: The Dialectic Of Materials And Information
DeepMind’s cofounder: Generative AI is just a phase. What’s next is interactive AI.
“This is a profound moment in the history of technology,” says Mustafa Suleyman.
What to know about this autumn’s covid vaccines
New variants will pose a challenge, but early signs suggest the shots will still boost antibody responses.
Human-plus-AI solutions mitigate security threats
With the right human oversight, emerging technologies like artificial intelligence can help keep business and customer data secure
Next slide, please: A brief history of the corporate presentation
From million-dollar slide shows to Steve Jobs’s introduction of the iPhone, a bit of show business never hurt plain old business.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.