As opposed to data, information cannot exist on its own. It arises dynamically from the interaction of data with that which makes use of it. It is a disconcerting feeling for a computer scientist to realize that information, the very object of his profession is not perfect, but rather something that will be different to every single user that observes it.

To emphasize this imperfectness, let us consider the following statement:

  • Example 1
  • “The man is 35 years old.”

It is a well formed statement that conveys a very valid piece of information about the person in discussion, but to most of us, this sentence contains no real information. In the best case it is a piece of data torn out of a context that we are not aware of. In order for this to represent actual information one requires considerable knowledge in what this specific context is concerned: “Who is the man?”, “When was this stated?”, are just two of the questions that can immediately be asked by somebody who’s reading this single sentence in a transcript of the underlying conversation, not having a context to place it in.

  • Example 2
  • “The man who lives on Elm Street 99999, apartment X, Aukland, New Zeeland is 35 years old.”
  • “The man who lives on Elm Street 99999, apartment X, Aukland, New Zeeland was 35 years old in year 2013.”

The somewhat more complete sentence in example: 2 can quickly respond to these questions, but again the statements imply a great deal of assumptions about who is going to read them: “What is an Aukland?”, “What is a New Zeeland?”, “What does year 2013 actually mean?”.

The fact of the matter is, that regardless of how much we describe the scenario, there will still be questions that are unanswered and assumptions that have to be made regarding a-priori knowledge possessed by the data consumer, regarding the context in which the information resides.

This effect is not limited to information conveyed via spoken or written language. Any object is potentially describable by an infinite or unreasonably large amount of attributes, some of which may even be inaccessible, and as such any information drawn from that subject is inherently incomplete.

Failure to recognize this aspects could have grave consequences with regards to information processing. If one does not recognize and accept incompleteness of information one would be tempted to analyze an object ad infinitum, trying to grasp all the details and characteristics of it. The process itself would likely generate an information overload. In the world of humans though, this is not the case. The human brain is very well equipped for these particularities. Most of the time it will only draw just as much information as we actually need to identify important aspects about the object: categorize the object, identify whether it is dangerous, useful, etc. Irrelevant facts, even if identified, are quickly forgotten making room in the memory for the next thing.

This selective observation (extraction of information) takes us directly to the other important aspect of information: subjectivity. All subjects draw information using their own particular sensors and interpreters and will filter it through their own preexisting knowledge and angle of interest. As such they will identify particular aspects of the object that are unique to them, thus giving information a highly subjective character.

It will become clear later on in this chapter, why these two seemingly evident characteristics are so important within the context of “knowledge representation” and how present representation systems fail to properly account for them.