[Open Systems Science]

The basic methodology of science has been to define a domain of a problem, abstract the problem involved so the true nature of the problem becomes clear, then discover the underlying principles of the domain. This approach was proposed by Descartes in the 17th century as the basic methodology of science. When a domain of a problem is too unwieldy and large to easily abstract the problem, it is broken up into smaller elements that are subjected to the process of abstraction to discover their basic principles, and hence it is called reductionism. This methodology contributed enormously to scientific advances from the 18th century, technological progress from the 19th century, and can be largely credited with the industrial prosperity and economic development the world has achieved today. Approaching the end of the 20th century, certainly many problems have been resolved, but there are still plenty of stubborn issues that are not susceptible to easy resolution that must be resolved. Issues pertaining to life and human health are examples, as well as problems relating to the brain and the mind. There are a host of intractable challenges relating to society, the economy, the global environment, energy, food, and other matters. The reliability and dependability of social infrastructure interconnected by the Internet is yet another difficult issue. What all of these issues have in common is that they are enormous complex systems made up of many constituent elements which are closely interrelated to each other.

It has been for more than 15 years or so that I have begun to really question how we can unravel these immensely complicated systematic problems, and whether the time-honored scientific methodologies of the past can continue to serve us well as they did in the past. I held extensive wide-ranging talks with all the scientists and researchers around me, but coming up with a solution to this problem has been immensely difficult and taken a very long time to finally glimpse the shape of a new methodology. At long last, a new methodology has emerged in the form of open system science. Here I will say a few words about the nature of this open system science.

The type of issues to be addressed by this new methodology of open system science is that of immense complicated systems that we touched on earlier. These systems are made of various elements that interact and affect one another in highly complex ways. This means that even if one wanted to break apart the constituent elements, this would not only be a very difficult proposition but some vital aspects of the whole in the process of breaking it up. Moreover, these issues have to be resolved while the underlying systems are alive and operating. Dealing with life or brain related problems means fixing the person or the animal that is afflicted. Social, economic and environmental issues cannot be resolved by simply resetting or restarting the system like a computer. And problems affecting Internet-connected social infrastructure must be addressed even while services remain up and running.

Considered in this light, it will be apparent that the new science methodology carrying us into the future will require a few new perspectives in addition to the reductionism that has served us up to now. Indeed, there has already been a partial revision to prevailing scientific methodology. That is the addition of a synthetic perspective, that is the basis in the creation of things combining fundamental principles. Since reductionism is an analytical perspective, this signifies a convergence of analytic and synthetic perspectives. These perspectives are already interacting with each other in the sense that their analysis is needed for engineering and the new analytical results are used in engineering, and the results can be seen in the achievements of current science and technology.

But this is not enough to resolve the enormous complex problems alluded to earlier, and the fact that an additional new perspective is required is where the open system science methodology comes in. The other new perspective that I believe must be added is that of management. We have believed that the notion of management inhabits a totally different sphere than science and technology, but when you think about it, what is life science if not the management of life? The creatures always pursuing next-best solutions when total solutions are unobtainable, somehow doing what's necessary to extend life, doing what is necessary to ensure the survival of descendants. This involves a time line, and events to be managed. Problems involving the brain and mind are essentially the same. Most of those problems are due to the way of effective management. Similarly in the case of social, economic, global environmental, energy, and food related issues, it is necessary to continue chipping away at these problems without let up, and in this sense a management perspective is critically important. Efforts are also needed to counter service outages and deliberate attacks on the immense Internet-connected social infrastructure, and future upgrades and modifications must be taken into consideration at the initial design stage. Here again, a management perspective is essential. So in order to pursue solutions to the enormous and complex issues we face even while these issues are live and active, it requires a three-perspective approach in which a management perspective is incorporated from the outset in addition to analytic and synthetic perspectives. This three-perspective approach is the essence of the open system science methodology.

If this abstract discussion has left some of my readers baffled and scratching their heads, here I will mention some of the specific projects carried out here at Sony CSL that well illustrate what I am talking about. In his pioneering work on systems biology, Hiroaki Kitano has shed light on the essence of life by defining it in terms of managing a functional network, and this approach has proved immensely useful in the discovery of new treatments and drugs for cancer, diabetes, and immunological diseases. And approaching systems biology from the standpoint of epigenetics, Kazuhiro Sakurada has formulated new theories throwing further light on life by incorporating epigenetics (acquired traits) management perspective into genetics (inherited traits). In his work on system neuroscience, Ken Mogi has made headway unraveling qualia and other issues relating to the brain and mind by employing a contingency perspective. Luc Steels' work on semiotic dynamics elucidates the dynamics of natural language development and evolution over time. An open system science approach is adopted in the work of other CSL researchers as well including François Pachet's approach to creativity through reflexive interaction, Jun Rekimoto's proposed cybernetic earth as a globe-girding information system, Hideki Takayasu's analysis of economic activity by statistical dynamics, and Frank Nielsen's work in the area of computational information geometry. What these researchers have in common is that they are forging new fields of research by treating their areas of concern as open systems, and inevitably the process involves an open system science approach.

To commemorate the 20th anniversary since Sony CSL was established, we have compiled these recent research developments and published them in a book. This work would not have been possible without our loyal supporters, so we express our great appreciation for support in the past and years ahead.

Mario Tokoro, President and CEO, Sony Computer Science Laboratories, Inc.

(February 9, 2009)