Perception element; that which is given through perception
Introduction to the notion
Etymology: plural of datum, from the Latin datum = (thing) given.
This notion, key in IT, has a long history and its rich background is not something that generally enters our mind.
Before data, the given
“That which immediately enters the mind before the latter has applied its mental processing to it.”, stated André Lalande, in his Vocabulaire technique et critique de la philosophie (1926). His colleague, Jules Lachelier, added: “It is likely that the first meaning of this word was that of given quantities…” (ibid.). In French, in philosophical vocabulary, the word sometimes appears in its masculine form and sometimes in its feminine form, as it the title “Essai sur les données immédiates de la conscience », by Henri Bergson (1889).
In this original meaning, the given is comparable to the percept “object of perception (without an ontological reference to any thing-in-itself)” (Le Grand Robert). The subtle difference between the notion of data and that of information reflects that opposition between percept and concept.
In systematic data processing – the sheer quantity of data which weighs us down from now on –, the difficulty lies in the transition from data to information, that is to say in the interpretation. The gap between the two is filled by “semantization”, if we establish a connection between data (in its raw state) and form. This connection has to be prepared by semantic modeling which builds the conceptual structure into which the perceptions slip. Access to meaning is the result of both of these operations: first, have the network of concepts ready; then, relate the percept to the concept. This complex process takes place instantly in the human mind, presupposing a prior investment that is none other than education. In the enterprise, this same process also requires an investment: drawing up the semantic model.
These thoughts, as old as philosophy, find a new application with the techniques associated with big data.
Related terms: software data, data policy
