Even if I’m just a newbie in Semantic Web and Semantic Technologies, in the past months I had the chance to give some thought to this idea. The way I see it, the Semantic Web targets to make content processable and understandable by machines, and not necessarily targets, at this stage, to evolve the Web to a space of “machine reasoning”, in which machines can replace human reasoning with comparable results. This state of art won’t be achieved also in the next foreseen stage, named Noosphere by a few Web theoreticians, for example [1].
The Noosphere, formed from nous (mind) and sphere (space or circle), can be regarded as a “space of human thought” supposed to reflect in real time the dynamics of collective intelligence, the role of visualization (reflection) and aggregation tools being essential. In time, I suppose that machines will grow (in) intelligence, being more and more capable to handle various tasks more like humans. When this will happen?! Who knows… Along the time the world’s theoretical models barriers were pushed beyond previous existing limits, so everything is possible, even braking the barriers of Goedel’s incompleteness theorems.
The Semantic Web is just a stage in the evolution of the Web, same as Web’s versions, it reflects a new way of thinking about Web, its role, expectations and tools supposed to fulfill them. Each person or community can have its own expectations and way of approaching the Semantic Web, lot of effort being spent in different directions, reinventing the wheel, technologies that die soon after they were born. Most of the researchers consider ontologies as the backbone on which the Semantic Web has to be built, many technologies focusing on this perspective. There are also scientists who question the achievability of a Semantic Web or the role of ontologies in this picture. C. Shirky’s [2], supported also by P. Gaendenfors [3], sustains that “the Semantic Web is a machine for creating syllogisms” and therefore it will improve only the areas that uses syllogistic reasoning.
Conversely, ontologies are just islands of knowledge not anchored in reality, they offer only a view/map of the world, and even if they reflect the commitment to common agreement, they are not a commitment to completeness. Ontologies are supposed to be created mainly by experts, involve high costs, considerable effort and coordination, and it seems that they follow the fallacies of OOP programming, breaking apart in their own complexity and require redesign when new facts are brought into the picture or the scope changes. As new knowledge is acquired or the requirements changes, the work on ontologies never ends, ontologies matching and integration involving other type of issues. Even more, to make things even fuzzier, M.K. Gergman [4] mentions more than 40 information structures that have been labeled in one way or another as ontology – tag cloud, controlled vocabulary, topic map, concept map, etc. Another important aspect neglected by ontologies seems to be the fuzzy nature of truth, while other issues derive from the information representational structure problems: symbol grounding problem, frame problem [5] and contextual emergence [6].
The goal of the Semantic Web is to “get people to use more meta-data” [2], and why not to create metadata, of harnessing the Collective Intelligence, as [7] formulates it. It has started with wikis and folksonomies, and might continue with more complex annotations, for example Knowledge Maps. It is created thus a layer of connectivity on top of physical structure of the Web Graph.
From my point of view Web theoreticians focus on high level goals and ignore the immediate needs of the users, which are often excluded from the Semantic Web equation. Models and technologies that target only the scientific world (e.g. ontologies) have low chances to make a difference in the Web space. The Web users need (free) tools that can be used for metadata creation, collaboration, information processing, knowledge mapping and diffusion. At least in the near future machines won’t achieve the thinking performances of humans about the world, though maybe once the Web riches the state of a Semantic Web, things would be much simpler.
References:
[1] Levy, P. (2005). From Cyberspace to Noosphere. [Online] Available from:
http://www.minervaeurope.org/events/parma/papers/levy_ppt.ppt (Accessed: 26 January 2009)
[2] Shirky, C. (2003). The Semantic Web, Syllogism, and Worldview. [Online] Available from:
http://www.shirky.com/writings/semantic_syllogism.html (Accessed: 7 February 2009)
[3] Gaerdenfors, P. (2004 B). Conceptual Spaces as a Framework for Knowledge Representation. [Online] Available from:
http://www.mindmatter.de/mmpdf/gaerdenfors.pdf (Accessed: 9 January 2008)
[4] M.K. Gergman. 2009. ‘Structs’: Naïve Data Formats and the ABox.[Online] Available from:
http://www.mkbergman.com/?p=471(Accessed: 17 April 2008)
[5] Duch, W. (1995) From cognitive models to neurofuzzy systems - the mind space approach. [Online] Available from:
http://www.fizyka.umk.pl/publications/kmk/95sams.pdf (Accessed: 13 January 2009)
[6] Atmanspacher, H., Foundation, P., beim Graben, P. (2005). Contextual Emergence of Mental States from Neurodynamics. [Online] Available from:
http://www.igpp.de/english/tda/pdf/potsdama12.pdf (Accessed: 13 January 2009)
[7] O’Reilly, T. (2005). What Is Web 2.0: Design Patterns and Business Models for the Next Generation of Software. [Online] Available from:
http://www.oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20 (Accessed: 18 April 2009)
