Database, Information and Knowledge (DIK)

Have decided to swiftly read through a copy of Luciano Floridi’s Philosophy and Computing, which I have had borrowed out from the library for a while. It was published in 1999 and seems to be a precursor of the establishment of the Philosophy of Information.

One part I found particularly interesting, titled ‘Data, information and knowledge: an erotetic approach’, is a short subsection of the fourth chapter, ‘The digital domain: Infosphere, databases and hypertexts’.

Floridi employs the conceptual framework provided by database technology to get working definitions of “data”, “information” and “knowledge”, using an erotetic framework (erotetic logic is the logic of questions and answers).


A datum is anything that makes a difference: a light in the dark, a black dot on a white page, a 1 opposed to a 0, a sound in silence. As such, it is a semantic atom: there can be no smaller extension of meaning, because a non-datum is a not-given nothingness, in which not even the presence of nothing is significant. From the point of view of our erotetic model, a datum can then be defined as an answer without a question: 12 is a sign that makes a difference, but it is not yet informative, for it could be the number of the astrological signs, the size of a pair of shoes or the name of a bus route in London, we do not know which.

Secondly, whilst the standard definition of information describes it as “data + meaning” or “meaningful data”, Floridi defines it in the following way relative to the erotetic framework: ‘A piece of information can be defined as a semantic molecule consisting of a data and a relevant question: 12 becomes informative once we know that it is the answer to the question “how many apostles were there?”‘.

Interestingly he then writes ‘note that a datum does not need to be the correct answer to function as a proper compound of a semantic molecule: disinformation and propaganda are still forms of information, no matter whether wrong or incorrect.’ This claim is in contrast to his subsequently developed veridicality thesis for information, according to which semantic information encapsulates truth.

Finally, in this erotetic framework ‘information becomes knowledge only if it is associated with the relevant explanation of the reasons why things are the way they are said to be. Knowledge involves understanding, not merely the contingent possession of a correct justification, and therefore insight’.

So knowledge can be defined as information + understanding or as a body of information + relevant explanation.

We thus get a linear hierarchy: data > information > knowledge. Given these definitions, Floridi goes on to write:

there are many more data than corresponding items of information, and the greatest parts of our epistemic life is based on true information, not on knowledge, since understanding is a rare and often unnecessary phenomenon. Strictly speaking, we do not know our names, our age or that “this is my hand”; we are simply informed about these facts. We may not know where we live and who our best friends are, but we are able to answer both questions correctly; most of us are only informed that water is H20 and that (a + b)2 = (a2 + 2ab + b2). Moreover, many things are just the way they are and there may be no particular explanations to give, or none that we can ever discover. “The person who will get the job owns a Ford”, if true, is a piece of information that may be intrinsically unexplainable, for there may be no specific reason why that is the case, and hence no way of raising it to the status of knowledge


So information is a common phenomenon, whereas knowledge is a rare exception? Whilst I am happy with this particular definition of knowledge as information + understanding, I think that application of the concept of understanding here is too heavy and results in an unduly small set of knowledge instances.

Leave a Reply

Your email address will not be published. Required fields are marked *