Often, our fascination with technology leads us to focus on the details of our work rather than its essence; we risk seeing technology not as a means to an end but as the goal, and this is especially true when we talk about data integration.
To better appreciate the purpose of data integration, we must remind ourselves that the goal of doing so is the digital representation of phenomena in the physical world, a world that we must observe in order to understand, because without understanding, any decisions we make will be intrinsically uncertain.
Representation is also a lever for sharing, because every individual representation is always a choral action, in that each individual’s private vision must harmonize with the corporate vision, and each organization, though it serves a set of common goals, must necessarily consider the different points of view of its individual members.
A Living, Shared Representation
If we accept this point of view, it follows that the major effort will be to create a universally shared representation, one that can give concreteness to the conceptualization of what we observe and which, moreover, must be “alive” in the sense of linking the intensional component to the extensional one, because a representation makes sense only if it facilitates the passage from the meaning of the data to the individual instances of the data and, vice versa, from what is detected and measured to a model that gives it overall semantic coherence.
However, we must not underestimate the complexity of this task. We must not think of meaning as something taken for granted, or assume that the private visions of those who perform this arduous task is equivalent to the collective one, which is more multifaceted, articulated, and fluid, and we must not trust that what is normally referred to as “semantic by design” is something that, once put into being, remains immutable over time. Instead, we must treat it as something that is profoundly adaptable to a world that, by its very nature, is changeable.
A semantic model is, in fact, a living thing, capable of adapting to context, and it could not be otherwise. A semantic model has the capacity to grasp different points of view and harmonize them, to constitute not only a coherent and accurate representation, but one that, above all, can be the basis for sharing and communication between different people who, thanks to a common, shared model, will be able to give value to what they have to do, with certainty that the deep texture of the model enables them to connect individual visions with global ones.
Experience and Experimentation for Data Integration
When we form ideas about phenomena, we act in two different ways: When the phenomena is familiar to us, we define the concepts and map them with observations; if the phenomena is unknown, we let ourselves be guided by our observations, and we try to provide a coherent vision of what we observe. Both ways have equal dignity: the first is based on experience (semantics by design); the second on experimentation (semantics by experimentation), and discovery and history teaches us that progress is always a combination of the two, which implies that the construction of a representational model must necessarily allow and support both approaches.
Whether one proceeds by experience or by experimentation, we will still have to face the question – or the problem – of who is the ultimate judge, assuming that it exists, of the representation that we want to construct. It is common experience, in fact, that the way in which each of us observes and conceptualizes our world is based on two fundamental elements: (1) being part of a community that shares those elements that enable mutual interaction, and (2) each of us have our own individuality, which leads us to see and interpret what we observe according to the private component linked to the experience that we accumulate over time.
If the first component gives us direction, the second traces its path, and one does not live without the other, because the hope of a common vision that cancels any difference of views is a hope, but unfortunately in vain.
An Adventure in Semantics, to Guide Decision-Making
To unleash the full potential of data, we must integrate it, but we must recognize that although this is a technological process, it is also a semantic adventure, one in which the meaning of different data, created by different subjects, in different eras, must be traced back to a single model that represents them, harmonizing the different points of view and connecting them to each other, linking the data to its observations which, over time, have been made, because, if the meaning is the use (as Wittgenstein tells us in Philosophische Untersuchungen, 1953) and use is always here and now, it is also true that use without representation and understanding is ephemeral, vain, leading nowhere, especially when the use is aimed at guiding our decision-making.
He is passionate about philosophy, which he considers an indispensable discipline for those wishing to deal with Artificial Intelligence. He also deals with non-verbal communication, offering specialized seminars and advice (https://www.decorporisvoce.com/) and, in his spare time, is a Food Blogger (https://www.trapignatteesgommarelli.com/).
- Data Integration: It’s not a Technological Challenge, but a Semantic Adventure - February 15, 2024
- Data Ecology - December 22, 2022
- Data Virtualization: Be the Driver, not the Mechanic - January 5, 2022