Hemingway, User690, and you
October 2, 2009
Our recent paper on human correspondence patterns has garnered a significant amount of attention in the press this past week, the full list of which can be found here. At Mike's behest, I thought I would take a few minutes to offer some personal commentary on the paper and remark on a few interesting consequences of our work.
The main finding of our paper is that, regardless of whether it is letter correspondence from the 1600s or modern day e-mail correspondence, human correspondence can be described by the same mechanistic model. The thing that distinguishes you from Ernest Hemingway or anyone else for that matter, is the parameterization of this model (i.e. how many correspondants you write in a single session). Similar to how the same equation describes the fluid mechanics of oil and water despite their different mechanical properties (density, viscosity, etc.), human correspondence can be described by the same model despite our obvious differences.
This finding is significant for many reasons, but I would like to highlight one of them here: just as we can compare nearly all fluids by their mechanical properties, we can compare humans by the parameterization of this model. From a practical perspective, we may be able to quantify behavior, to detect anomolous activity, or even to predict outcomes like productivity, success, or consumer loyalty based on the parameterization of our model.
Of course, the utility of using our model in areas ranging from the quantification of behavior to measures of productivity heavily depends on its applicability to human activities other than human correspondence. The fact is that our model is not correspondence-specific; our model only accounts for the weekly work cycle, the circadian cycle, and task repetition. Since these behaviors are certainly important for a wide range of human activities, it seems plausible—even likely—that our approach can be used to quantify human activity within the same, universal framework.