As I wait for the murmuration to form, I make some deft amendments to the list of the authors, striking out a name here, a name there.
Feathers will fly, of course.
TEN BIRDS MIGHT not seem much but these birds are like precision instruments, guided with digital finesse. To begin with, we – I – restrict myself to only minor interventions.
I make the murmuration split into two distinct elements, then recombine. Suitably encouraged, I quarter it like a flag. I pull it apart into four rippling sheets of birds, with arcs of clear air between them. The edges are improbably straight, as if the birds are glassed-in, boxed by invisible planes. But that is the power of incredibly delicate control processes, of stimulus and feedback operations happening much too swiftly for human perception. If an edge starts losing coherence, the computer makes a tiny adjustment to one or more of the control starlings and the order is reestablished. This happens many times a second, at the speed of avian reactions.
They have always lived in a faster world than us. They live a hundred days in one of our hours. To them we are slow, lumbering, ogrelike beings, pinned to the ground by the stonelike mass of our bodies. We envy them; they pity us.
I push forward. I carve geometries out of the murmuration. I fold it into a torus, then a ribbon, then a Möbius strip. I do not need to know how to make these shapes, only to instruct the laptop in my desires. It works out the rest, and becomes more adept as it goes along.
I make the murmuration spell out letters, then I coax those letters into lumpy, smeared-out words. I spell my name in birds. They banner around me like the slogans towed by light aircraft. I laugh even as I feel that I have crossed some line, some invisible threshold between pristine science and sordid exploitation.
But I carry on anyway. I am starting to think about those domains, those hints of modular organisation.
How far could I push this, if I were so determined?
ANGRY EXCHANGES OF emails. Editor not happy with this latest change of direction. Much to-ing and fro-ing. Questions over the change in the listed authors – deemed most unorthodox. Accusations of unprofessionalism. If we were in a room together, the three of us, we might get somewhere. Or we might end up throwing textbooks.
Is this a travesty of the way science ought to be done, or is it science at its shining best – as loaded with passion and conviction as the any other human enterprise? No one would doubt that poets squabble, that a work of great literature might take some toll on its creator, that art forges enemies as readily as allies. Why do we hold science to a colder, more emotionless set of standards? If we care at all about the truth, should we not celebrate this anger, this clashing of viewpoints?
It means that something vital is at stake.
Hard in the spitting crucible of all this to remember that every one of us was drawn to this discipline because of a love of birds.
But that is science.
MY PROPOSITION IS simple. The domains are controllable. I can cause them to form, contain the shape and extent of their boundaries, determine the interaction of their vector groups with the surrounding elements. I can move the domains around with the flock. I can blend one domain into another, merging them like a pair of colliding galaxies. Depending on their vector properties, I can choose whether that act results in the destruction of both domains or the formation of a larger one.
I sense the possibility of being able to execute a kind of Boolean algebra. If the domains behave in a controllable and repeatable way, and I can determine their states – their aggregate vector sums – then I can treat them as inputs in a series of logic operations.
The thought thrills me. I cannot wait for the coming of dusk.
With the laptop reprogrammed, I quickly satisfy myself that the elements of my Boolean experiment are indeed workable. I create the simplest class of logic gate, an AND gate. I classify the input domain states as either being 0 or 1, and after some trials I achieve a reliable ‘truth table’ of outputs, with my gate only spitting out a ‘1’ if the two inputs share that value.
I push on. I create OR and NOT gates, a ‘not AND’ or NAND gate, a NOR gate, an XOR and XNOR gate. Each is trickier than the last, each requires defter control of the domains and vector states. To make things easier – at the burden of a high computational load on the computer and the ethernet network – I retrieve more birds from the snares, fitting them with additional digital hoods.
Now I can create finer domains, stringing them together like the modules in an electrical circuit.
I begin to ‘wire up’ the flock. I assign gates to perform logical operations, but also to store data. Again, I need only tell the computer what I want it to do – it takes care of the computational heavy-lifting. All I know is what my eyes tell me. The murmuration has grown knotted and clotted, dense with domain boundaries and threaded with the thick synapses of internal data corridors. It swoops and billows over me, a circuit of birds.
The astonishing thing is that on the level of individual starlings, they sense no strangeness – no inkling that they are participating in anything but a normal murmuration. The complexity is emergent, operating on a scale that the birds simply cannot sense, cannot share. They are cells in a larger organism.
I lash together a Perl script, a simple text to logic program on the laptop, enabling me to send natural language queries to the flock.
IS ONE AND ONE TWO?
There is a process of calculation. The circuit shuffles. I glean the flow of information along its processing channels – the physical movement of birds and their larger domain boundaries.
The answer returns. The laptop takes the Boolean configuration and converts it back into natural language.
>>YES.
I try another query.
IS ONE AND ZERO ZERO?
A swoop, a billow, a constant busy shuffling of birds.
>>NO.
I smile. Maybe a fluke.
IS ONE AND ZERO ONE?
>>YES.
I am elated.
Over the next thirty minutes, I run through question after question. The birds answer unfailingly. They are computing, and doing so with the utmost machinelike reliability.
>>YES YES NO YES YES NO NO NO.
I am doing algebra with starlings.
But as the gloom gathers, as the dusk deepens, something troubles me.
In all my interventions to date, one thing has remained true. The murmuration eventually dissipates. The roosting instinct overpowers the flocking instinct, and the birds cascade down into the trees. It happens very quickly, a kind of runaway escalation. Whenever I have witnessed it, I am always saddened, for it is the end of the show, but I am also amazed by what is another demonstration of marvellous collective action.
And then the skies are clear again, until the birds lift at dawn. This is what should happen.
But now the murmuration will not break up.
Some birds leave it, maybe a third, but a core remains. I hammer at the laptop – more puzzled than worried at first. I try to disrupt the logic flow, randomise the data, dismantle the knotty Boolean architecture. But the pattern remains obstinately present. The sky darkens, until only the cameras and rangefinders are able to track the birds, and then with difficulty.
But I can still hear them up there – a warm but unseen presence, like a clot of dark matter hovering over me.
I THINK IT’S time to recuse myself from refereeing this paper.