All non-black objects are non-ravens
The Economist writes about a new way of securing data in databases based on philosophy:
In the 1940s a philosopher called Carl Hempel showed that by manipulating the logical statement all ravens are black you could derive the equivalent all non-black objects are non-ravens. Such topsyturvy transformations might seem reason enough to keep philosophers locked up safely on university campuses, where they cannot do too much damage. (The Economist 2006 8493(380) p. 72)
The underlying principle is similar to the immune system, where lymphocytes (e.g. T-cells) do not have information about any potentially harmful intruders, i.e. all their information refer to body cells. By negating their information base, everything they do not know is a potential threat and thus will be attacked. Hence, lymphocytes do not need information about virusses, bacteria, etc.
Transferring this into the world of computers results in negative databases, that store everything but the actual information. Such negative databases are suitable for protecting privacy, e.g. by just storing the negative information of credit card information. For instance it is possible to query the database for the information about one entry (”Give me the credit card information of person X”). However, it is not possible to process Select *-statements (”Give me all credit card information of your German customers”) (Esponda et al. 2004).
Practically, this means that for instance surveys can use the information stored in such negative databases and derive statistical information of it and do not conflict with privacy concerns of the individuals, whose date is stored in the database (Esponda 2005).
Another application could be data exchange through web services. The big security and privacy problem of web services is that I have to hand out date to some probably unknown entity. So basically web services rely on the hope, that my business partner does not publish my information to the world or sell it to competitors (There is an interesting interview with Withfield Diffie on this topic here). Wouldn’t it be possible to design negative statements of information that I want to be processed by somebody else, that allows the external entity to process it and return a meaningful result without actually knowing the exact nature of the result? Obviously it would not work on any data, e.g. you definitely need to present shipping information to UPS to get the package delivered.
The analogy with the human immune system results in adopting its weaknesses. Failures in knowing what self is can do great damages. Allergic reactions then seem to be the result of the wrong negative information.
References
(Esponda et al 2004) F. Esponda, E.S. Ackley, S. Forrest and P. Helman. On-line Negative Databases. Third International Conference on Artificial Immune Systems (ICARIS 2004) Proceedings, pp.175-188, September 2004.
[Esponda 2005) Esponda, F. Negative representations of information. University of New Mexico Dissertation, 2005.
