The Intelligent MDM Blog

Even for Data, "More is Different"

In 1972, Philip Warren Anderson, American Physicist and Nobel Laureate, published in "Science" an article that I enjoy reading regularly, as it is in my mind an importante milestone in modern science.

The title is "More is Different - Broken symmetry and the nature of the hierarchical structure of science".
There are a few frightening scientific words and concepts in the article, but Anderson makes it simple for us to understand his point. And if you can remember a few terms or quotes, you'll be a star at your next dinner!

In this article, Anderson breaks down the reductionist hypothesis. This hypothesis assumes that a complex system can be understood through the understanding of the smaller parts of the system. Based on this hypothesis, Anderson explains that we can build a linear hiearchy of sciences where "the elementary entities of science X obey the laws of science Y". For example, entites in Social Sciences (humans) can be understood through the laws of Psychology, which can be understood through the laws of Physiology, and so on and so forth, until we arrive to Chemistry, and then to Particle Physics. So by mastering the laws at a granular level (particles) we would be able to extrapolate the behavior at the larger scales.

Unfortunately, things do not work as simply this way in the real world. In "More is Different", Anderson gives an example at the molecular level: A peculiar broken symmetry appears in larger scale molecules, apparently against a law defined at the smaller scale. This broken symmetry is a new effect that appears when the scale changes. To quote Anderson: "we can see how the whole becomes not only more than but very different from the sum of the parts.".

In three words: "More is Different".

Today, after seeing data grow dramatically in volume, velocity and variety, we start building the capabilities to use, process and govern this data. As data is reaching new stages in terms of scale and complexity (becoming "More"), to take the words of Anderson, we must expect in data management and data science "to encounter fascinating, and I believe, very fundamental questions at each stage in fitting together less complicated pieces into the more complicated system and understand the basically new types of behavior which can result." (something "Different")

In addition, as data becomes more and more valuable and pervasive, exploring new data opportunities mandates that its management no longer remains the sole prerogative of a minority of IT experts. A better approach should involve a diversity of people, skills, mindsets and backgrounds coming from outside the "data and information science valley". 40 years later, the words of Anderson sound again very true: "So it is not true [..] that we each should 'cultivate our own valley, and not attempt to build roads over the mountain ranges ... between the sciences.' Rather, we should recognize that such roads, while often the quickest shortcut to another part of our own science, are not visible from the viewpoint of one science alone."

Today, as in 1972, Anderson's intuition is still true: "More is Different".