Читать книгу Cyber-physical Systems - Pedro H. J. Nardelli - Страница 15

Science is a special type of formal discourse that claims to hold objective true knowledge of well‐determined objects. Different sciences have different objects, requiring different methods to state the truth value of different statements. A given science is presented as a theory (i.e. a systematic, consistent discourse) that articulates different concepts through a chain of determinations (e.g. causal or structural relations) that are independent of any agent (subject) involved in the production of scientific knowledge. This, however, does not preclude the importance of scientists: they are the necessary agents of the scientific practice. Scientific practice can then be thought as the way to produce new knowledge about a given object, where scientists work on theoretical raw material (e.g. commonsense knowledge, know‐how knowledge, empirical facts, established scientific knowledge) following historically established norms and methods in a specific scientific field to produce new scientific knowledge. In other words, scientific practice is the historically defined production process of objective true knowledge. Note that these norms, despite not being fixed, have a relatively stable structure since the object itself constrains which are the valid methods eligible to produce the knowledge effect.

Moreover, scientific knowledge poses general statements about its object. Such a generality comes with abstraction, moving from particular (narrow) abstractions of real‐world, concrete objects to abstract, symbolic ones. Particular variations of a class of concrete objects can be used as the raw material by scientists to build a general theory that is capable of covering all, known and unknown, concrete variations of that class of objects. This general theory is built upon abstract objects that provide knowledge of concrete objects. However, this differentiation is of key importance since a one‐to‐one map between the concrete and abstract realities may not exist. Abstract (symbolic) objects as part of scientific theories produce a knowledge effect on concrete objects, understood as realizations of the theory, not as a reduction or special case. At any rate, despite the apparent preponderance of abstractions, the concrete reality is what determines in the last instance the validity of the theory (even in the “concrete” symbolic reality of pure mathematics, concreteness is defined by the foundational axioms and valid operations).

To illustrate this position, let us think about dogs. Although the concept of dog cannot bark, dogs do bark. Clearly, in the symbolic reality in which the concept of dog exists, it has the ability of barking. The concept, though, cannot transcend this domain so we cannot hear in the real world the barking sound of the abstracted dog. Conversely, we all hear real dogs barking, and therefore, any abstraction of dogs that assumes that they cannot bark shall not be considered scientific at all. This seems trivial when presented with this naive example, but we will see throughout this book the implications of unsound abstractions in different, more elusive domains. This is even more critical when incorrect abstractions are accompanied by heavily mathematized (therefore consistent) models. For instance, the fact that some statement is a true knowledge in mathematical sciences does not imply it is true in economics. Always remember: a mathematically consistent model is not synonymous with a scientific theory.

Philosophy, like science, is also a theoretical discourse but with a very important difference: it works by demarcating positions as correct or incorrect based on its own philosophical system that defines categories and their relations [8]. Unlike scientific proofs, philosophy works through rational argumentation to defend positions (i.e. theses), usually trying to answer universal and timeless questions about, for example, existence of freedom. In this case, philosophy has no specific (concrete) object as sciences do; consequently, it is not a science in the way we just defined. Philosophy then becomes its own practice: rational argumentation based on a totalizing system of categories defining positions about everything that exists or not. Following this line of thought, philosophy is not a science of sciences; it can neither judge the truth value of propositions internally established by the different sciences nor state de jure conditions for scientific knowledge from the outside.

Besides, scientific and philosophical practices exist among several other social practices. They are part of an articulated historical social whole, where different practices coexist and interfere with each other at certain degrees and levels of effectivity. As previously discussed, scientific practice produces general objective true knowledge about abstract objects, which very usually contradicts the commonsense ideas that are usually related to immediate representations arising from other social practices of our daily lives. This clearly leads to obstacles to scientists, who are both agents of the scientific practice and individuals living in society. The totalizing tendency of philosophy also plays a role: it either distorts scientific theories and concepts to fit in universal systems of philosophical categories or judges their truth value based on universal methodological assumptions. This directly or indirectly affects the self‐understanding of the relation that the scientists have with their own practice, creating new obstacles to the science development [9, 10].

Example 1.2 Differences between daily language, philosophical categories, and scientific concepts. One word that exemplifies very well the difference is time. We use word time in many ways in our daily lives: to discuss about our activities, plans, routine, and the like. However, in philosophy, the category Time has different roles depending on the philosophical system to be considered – this usually comes with the relation between other categories like Causality, Origin, and End. In sciences, time is also a concept in different disciplines. In physics, time is a very precise concept that has been changing throughout its history, changing (not without pain) from the classical definition that time is an absolute measure (i.e. the same everywhere) to today's relativity theory where time is relative (and the speed of light is absolute). Such a scientific definition of the concept of time is not intuitive at all, and goes against most of our immediate use of the word. In this sense, scientists may find it difficult to operate with the scientific concept of time in relativity theory because of the other more usual meanings of the word. Besides, such a confusion between the scientific and the nonscientific may open philosophical questions and nonscientific interpretations of the scientific results.

In addition to this unavoidable challenge, the rationalization required by scientific theories appears in different forms. In this case, philosophical practice can help scientific practice by classifying the different types of rationality depending on the object under consideration. Motivated by Lepskiy [11] (but understood here in a different manner) and Althusser [8], we propose the following division.

 Classical scientific rationality: Direct observations and empirical falsification are possible for all elements of the theory, i.e. there is a one‐to‐one map between the physical and abstract realities.

 Nonclassical scientific rationality: Observations are not directly possible, i.e. the process of abstraction leads to nonobservable steps, resulting in a relatively autonomous theoretical domain.

 Interventionist scientific rationality: Active elements with internal awareness with objectives and goals exist, leading to a theory of the fact to be accomplished in contrast to theories of the accomplished facts.

By acknowledging the differences between these forms of rationality, sciences and scientific knowledge can be internalized as a social practice within the existing mode of production. Different from positivist and existentialist traditions in philosophy, this practice of philosophy attempts to articulate the scientific practice within the historical social whole, critically building demarcations of the correctness of the reach of scientific knowledge by rational argumentation [10].

Example 1.3 Scientific efforts related to COVID‐19. In 2020, an unprecedented channeling of research activities was directed to combat the COVID‐19 pandemics. These activities can be classified following the three aforementioned scientific rationalities. The classical rationality can be exemplified by the process to test the effectiveness of vaccines following the historically defined norms. The nonclassical rationality incorporates the mathematical models for epidemics based on nonlinear dynamical processes over graphs where not all variables are observable. The interventionist rationality considers lockdown policies to control the virus propagation as a fact to be accomplished. A critical philosophical practice demarcates the reach of the three different scientific activities, determining both their interrelations and the articulation with other social practices. For example, a vaccine that is proved to work can be modeled by a mathematical model, which can be used to change the lockdown policy. However, within the capitalist mode of production, these activities are directly or indirectly determined by the economical reality – from the funds available to develop the vaccine and its respective property rights to the economic impact of lockdown policies and its justifications based on a wide range of epidemiological models. A critical philosophical practice acknowledges the autonomy of the results obtained through the scientific practice with respect to its object while it internalizes such a practice in the articulated social whole.

This philosophical practice goes hand in hand with the scientific practice by helping scientists to avoid overreaching tendencies related to their own theoretical findings. It also indicates critical points where other practices might be interfering in the scientific activity and vice versa. Although a deep discussion of the complex relations between scientific practice and other practices are far beyond our aim here, we will throughout this book deal with one specific relation: how scientific practice is related to the technological development. We have seen so far that the practical development of techniques does not require the intervention of (abstract) scientific rationalization. On the other hand, the knowledge produced by the sciences has a lot to offer to practical techniques. The existence of the term technology, referring to techniques developed or rectified by the sciences, indicates such a relation. More than what this definition might suggest, technology cannot be simply reduced to a mere application of scientific knowledge; it can indeed create new domains and objects subject to a new scientific discourse.

The aforementioned control and information theories perfectly exemplify this. New technological artifacts had been constructed using the up‐to‐date knowledge of physical laws to solve specific concrete problems, almost in a trial‐and‐error basis to create know‐how‐type of knowledge pushed by the needs of the industrial revolution. At some point, these concrete artifacts were conceptualized as abstract objects toward a scientific theory with its own methods, proofs, and research questions, constituting a relatively autonomous science of specific technological objects. The new established science not only indicates how to improve the efficiency of existing techniques and/or artifacts but also (and very importantly) defines their fundamental characteristics, conditions, and limits.

Example 1.4 Information theory. Although large‐scale communications systems had already been deployed for some decades before the 1940s, the engineers considered that errors in transmission were somehow inevitable. This commonsense practical knowledge was scientifically proven false when Claude Shannon published A Mathematical Theory of Communication [12] formulating the concept of information entropy and mutual information. Using these concepts, Shannon mathematically proved the existence of a code that leads to error‐free communication if, and only if, the coding rate is below the channel capacity. This theory proposed in 1948 opened up a new field of theoretical research and also oriented practical deployments by giving an absolute indication of how far from the fundamental limit specific technologies are. It is noteworthy that, although Shannon had mathematically proven the existence of capacity‐achieving codes, he has not indicated how to practically design them. For many years, researchers and engineers have pushed the technological boundaries and have developed different coding schemes. Only with the new millennium, feasible solutions have been proposed (or rediscovered) and, currently, the turbo codes and low‐density parity‐check (LDPC) codes are feasible options to reach a performance close to Shannon's limit. These high‐performance techniques are used for example in cellular networks and satellite communications. The fundamental limit proposed by Shannon, though, cannot be surpassed by any existing or future technologies. A similar development happened in physics when the fundamental laws and limits of thermodynamics; firstly motivated by the development of thermal engines, the thermodynamic laws imposed fundamental limits of all existing or future engines [10].

An important remark is that sciences as theoretical discourses are historical and objective, holding a truth value relative to what is scientifically known at that time considering limitations in both theoretical and experimental domains. In this sense, scientific practice is an open‐ending activity constituted by historically established norms. These norms, which are not the same for the different sciences and are internally defined through the scientific practice, determine the valid methodologies to produce scientific knowledge. Once established, this knowledge can then be used as raw material not only for the scientific practice from where it originates but also it can be (directly or indirectly) employed by other practices. As demonstrated in, for example, Noble [13], Feenberg [14], the scientific and technical development as a historical phenomenon cannot be studied isolated from the society and its articulation with the social whole becomes necessary.

From this perspective, this book will pedagogically construct a scientific foundation for CPSs based on existing scientific concepts and theories without distorting and displacing their specific objects. The resulting general theory will then be used to explain and explore different particular existing realizations of the well‐defined abstract scientific object called CPS following the three proposed scientific rationalities. We are now ready to discuss the book structure and its rationale to then start our theoretical tour.