Computation in Natural Systems


In collaboration with Keith Mott, a plant biologist at USU, and several graduate and undergraduate students, I am developing methods to quantitatively identify and characterize how plants manage to solve sophisticated problems.  One of our typical reflective moments is shown below. 




If you are not a plant biologist (and perhaps even if you are), you may be surprised to learn that plants are capable of solving subtle problems.  But it’s true.  Each day, a plant is faced with the dilemma of trying to acquire as much CO2 as possible from the atmosphere (which it then uses to store energy via photosynthesis), while at the same time trying to mitigate the possibly deleterious effects of excessive water loss (via evaporation).


The hardware a plant uses to regulate the exchange of gases between the air and its interior are variable aperture pores – stomata – that cover the surfaces of its leaves.  Bright light causes a plant’s stomata to open, and this, in turn, permits CO2 to enter the leaf and water vapor to leave. 


Because water evaporation increases as temperature increases, a straightforward strategy for maximizing CO2 uptake while minimizing water loss would seem to be: open stomata wide when it’s cool, close them shut when it’s hot.  This is where the subtlety comes in.  Even when a pore is wide open, CO2 can only enter a leaf at a limited rate because CO2 molecules have to diffuse through air.  Thus, to maximize CO2 uptake over the course of a day requires that pores be open to some degree all of the time.  So, what is the correct openness for each temperature, humidity, and light level condition? 


Keith and I believe that plants may solve this complicated problem by a process called “emergent computation.”  Much more about how stomata work, what emergent computation is, and what the evidence is for plants performing emergent computation can be found at our Complexity and Stomatal Behavior web site.


The work I have done on uncovering possible computation in plants is a part of a larger interest centered on the question, are the behaviors of natural systems the result of formal computations?  Among other things, computation seems to involve the deterministic manipulation of information to perform a purposeful task.  So, then, is a falling rock a computation?  When a rock falls, its trajectory is determined by some initial data (the rock’s initial position and velocity) and by the requirement that a quantity called the “action” be a minimum.  In other words, the trajectory of a falling rock is a set of output data that “solves the problem” of minimizing action.  Sounds kind of like a computation.  Still, a falling rock doesn’t seem very subtle or sophisticated.  Isn’t falling “just the way it is?” 


Trying to figure out how to tell when a phenomenon is a sophisticated computation and when it is “just the way it is,” is my current major research interest.



Relevant Publications:


D. Peak, J.D. West+, S.M. Messinger+, and K.A. Mott, “Evidence for complex, collective dynamics and emergent, distributed computation in plants,” Proceedings of the National Academy of Sciences 101, 918-922 (2004).


J.D. West+, D. Peak, J.Q. Peterson, and K.A. Mott, “Dynamics of stomatal patches for a single surface of Xanthium strumarium L. leaves observed with fluorescence and thermal images,” Plant, Cell & Environment 28, 633-641(2005).


K.A. Mott and D. Peak, “Stomatal patchiness and task-performing networks,” Annals of Botany, doi: 10.1093/aob/mcl234, 1-8 (2006)


S.M. Messinger+, K.A. Mott, and D. Peak, “Task-performing dynamics in irregular, biomimetic networks,” Complexity 12, 14-21 (2007)