The Precepts

  • Michael K. Bergman


Truth, though fallible, exists. Knowledge thus should express a coherent reality, to reflect a logical consistency and structure that comport with our observations about the world. How we represent that reality has syntactic variation and ambiguities of a semantic nature that can only be resolved by context. To deal in the realm of knowledge and belief, the purpose of our KR systems, we must be able to ingest and process electronic data in any form. This information includes any managed by databases, but it can also include messages, documents, Web pages, or data tables. A hub-and-spoke design with a canonical data model is a superior way to organize, manipulate, and manage input information. Our source information embodies conceptual differences and value and attributes differences in raw data. By understanding the sources of semantic heterogeneity, we set the basis for extracting meaning and resolving ambiguities. Once we resolve the source information, we need to organize it into ‘natural’ classes and relate those classes coherently and consistently to one another. This organization takes the form of a knowledge graph. Semantic technologies have the flexibility and openness required for this task. Traditional relational databases do not; they are inflexible and fragile when the nature (schema) of the world changes, and require expensive re-architecting in the face of new knowledge or new relationships. In comparison, knowledge graphs provide natural groupings of concepts and entity types to characterize the domain at hand, situated to one another with testable relations.


Data model Fallibility Semantic heterogeneity Knowledge graph 


  1. 1.
    C. Anderson, Leveraging data to drive innovation, Wall Street Journal,
  2. 2.
    P. Schäuble, SPIDER: A multiuser information retrieval system for semistructured and dynamic data, in Proceedings of the 16th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (ACM, 1993), pp. 318–327Google Scholar
  3. 3.
    M. Magnani, D. Montesi, A Unified Approach to Structured, Semistructured and Unstructured Data, TECHREPORT in Education, Information Processing and Management 2, 263–275 (2004)Google Scholar
  4. 4.
    J. Madhavan, S.R. Jeffery, S. Cohen, X. Dong, D. Ko, C. Yu, A. Halevy, Web-scale data integration: You can only afford to pay as you go, in CIDR (2007)Google Scholar
  5. 5.
    A. Singhal, Introducing the knowledge graph: Things, not strings, Official Google Blog (May 2012)Google Scholar
  6. 6.
    C. Pluempitiwiriyawej, J. Hammer, A Classification Scheme for Semantic and Schematic Heterogeneities in XML Data Sources, vol 36 (University of Florida, Gainesville, FL, 2000)Google Scholar
  7. 7.
    P. Steiner, CS Peirce and artificial intelligence: Historical heritage and (new) theoretical stakes, in Philosophy and Theory of Artificial Intelligence, ed. by V. C. Müller (Springer, Heidelberg, 2013), pp. 265–276Google Scholar
  8. 8.
    M.K. Bergman, Tutorial: Internet Languages, Character Sets and Encodings, BrightPlanet Corporation Technical Documentation (Mar. 2006)Google Scholar
  9. 9.
    M.R. Ayers, Locke versus Aristotle on natural kinds. J. Philos. 78, 247–272 (1981)CrossRefGoogle Scholar
  10. 10.
    M. Hulswit, Peirce’s teleological approach to natural classes. Trans. Charles S. Peirce Soc. 33, 722–772 (1997)Google Scholar
  11. 11.
    M.K. Bergman, Shaping wikipedia into a computable knowledge base, in AI3:::Adaptive Information (Mar. 2015)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Michael K. Bergman
    • 1
  1. 1.Cognonto CorporationCoralvilleUSA

Personalised recommendations