| Science is a swarm.
To the layperson, the stereotypical scientist is logical, clear-thinking, wellinformed but perhaps socially awkward, carefully planning his or her experiments and then analyzing the resulting data deliberately, with precision. The scientist works alone, emotion-free, searching only for truth, having been well advised about the pitfalls and temptations that lie along the path to discovery and the expansion of human knowledge.
Those who work in science understand how inaccurate this stereotype is. In reality, researchers’ daily routines follow a process better described as collective trial-and-error, nearly random at times. A most salient feature of scientific behavior is its collaborative nature. From applying for grants to seeking tenure, from literature reviews to peer review to conference presentations, every bit of the scientific enterprise is social, every step of the process is designed to make scientists aware of one another’s work, to force researchers to compare, to communicate, to study the work that others are doing, in order to push the paradigm forward - not as independent, isolated seekers-of-truth, but more like a swarm.
If we plotted a group of scientists as points on a space of dimensions of theories and methods, and ran the plot so we could see changes over time, we would see individuals colliding and crossing, escaping the group’s gravity field and returning, disintegrating but simultaneously cohering in some mysterious way and moving as a deliberate, purposeful bunch, across the space - constantly pushing toward a direction that improves the state of knowledge, sometimes stepping in the wrong direction, but relentlessly insisting toward an epistemological optimum.
The book you hold in your hand is a snapshot of the swarm that is the swarm paradigm, a flash photograph of work by researchers from all over the world, captured in mid-buzz as they search, using collective trial and error, for ways to take advantage of processes that are observed in nature and instantiated in computer programs. |
