Preface

This book, which focuses on the domain of food science, is an excellent occasion to consider various issues related to optimization. Optimality, in any domain, is an open question, raising various complex issues. Questioning the purpose of optimization, its ability to answer important real-life questions, is in essence an intellectual exercise: are we able to address the appropriate issues with the help of modern computational tools? Do we believe too much in computation? Are we able to address the right issues with the right tools?

These questions have been considered with the help of a philosopher, Nathalie Godefroid1, and this preface is the result of our conversations on this vast subject.

The sources

The idea of optimization has its roots in what has been called "modernity" since 16th and 17th Centuries, based on a fundamental change in the perception of man within nature. During Antiquity and the Middle Ages nature was considered a "cosmos", that is a big whole, symbolic, sacred, and respected hierarchy (each creature has its own position). "Modernity", however, initiated a neutral standpoint, from which symbolics are progressively removed. The universe is infinite, without purpose, and nature is just a set of physical laws that can be understood and controlled, and therefore submitted to human needs and desires. Descartes' project is to become "owner and master" of nature (Discours de la méthode, [DES 37]): managing and predicting natural phenomena becomes an attractive and reachable challenge. This control of nature is based on mathematics: "nature is a book written in mathematical language" (Galilée, [MAR 02]), the aim is to depart from mistery and contingency using rationality and mathematics.

This humanist project (knowledge and progress must benefit human beings, their freedom and their happiness), however, is, the source of other troubles, as highlighted by Heidegger. Technique progressively becomes unlimited, mandatory and above human beings, their projects and activities. Everything then becomes a product, a consumer good, including humans: everything behaves in a way that is "computable". Heidegger shows that technique is no longer an instrument at the service of humans, but an end in itself. The search for performance and optimization is a technical ideal. Rationality in technique is relying on a value system (an evaluation function) in connection with economic interests. The technical means thus dictate a value to the user: efficacy. Efficacy has gained supremacy everywhere: economy, pedagogy, sport, research, social organization, politics, sex, everyday life, etc. "The technical phenomenon is the concern of the immense majority of men today, that is to search in everything for the most efficient method" [ELL 77].

Technique, power and language

With technique, men control nature using a complex and evolving set of means. "Not merely its application, but technique itself is domination - over nature and over men: methodical, clairvoyant domination. The aims and interests of domination are not additional or dictated to technique from above - they enter into the construction of the technical apparatus itself. For example, technique is a social and historical project: into it is projected what a society and its ruling interests decide to make of man and things. The aims of domination are substantive, and belong to the form of technical reason itself" [MAR 64].

Another hazard caused by technique, a major hazard according to Heidegger, is the modification of language: technique triggers the ideal of a non-ambiguous communication language. This ideal language dedicated to information encoding is non-hermeneutic2, in contrast to natural language, which predates and is external to technique, like the poetic language. This impact of technique on language, with its pure utilitarian approach, is a threat to human essence, as it discards philosophical and religious thoughts, meditations and contemplations, which are typical disinterested, non-measurable activities.

According to the sociologist Philippe Breton (L'Utopie de la communication, [BRE 92]), a true social utopia has been built since the Second World War with cybernetics and the work of Norbert Wiener (an American mathematician and philosopher, deemed to be the originator of cybernetics). Considering that everything is information and information sharing, living organisms and machines are on one level: the brain is a computer, thinking is computing. Even if this viewpoint forms the groundwork for artificial intelligence, this posttraumatic utopia emerged after the Second World War, with the intention of discarding such horrors forever. The main values are transparency, consensus and information circulation, as opposed to entropy and chaos. Machines would be more efficient and rational than human beings in making decisions, particularly in politics.

Technical developments thus answer the desire for full control in an uncertain and complex world. But complexities are the essence of life and the human brain's creativity. Randomness and unpredictability are a major characteristic of many systems, including living organisms, populations and ecosystems. Creativity of life may remain outside of the scope of mathematical modeling. In Ancient Greek philosophy, an "opportunity" is a recurrent topic: man is the one who knows or should know how to exploit opportunity in a world where nothing is perfectly predictable. Intuition and improvisation are "human" capabilities (in particular in the musical domain, as highlighted by Jankélévitch). Even if computer science and artificial intelligence have made huge progress, the question of the respective roles of man and machine remains.

The human factor in computer science

The human body is synonym of imperfection (Plátõn), flesh is a source of corruption: emotions, illness, death. Medical and technical progress aim at repairing, improving and augmenting the human body. But in the scientific domain, the body is often considered a neutral material, a source of information, of unpredictable data or signals, emptied of its symbolic meaning. Embodiment, the humility of the human condition, and finally the fear of death are at the source of the modern fantasy aiming at abolishing the body. This idea actually also comes from modernity, from Descartes and the first anatomists: humanity is thought; the body, a hindrance. The sociologist David Le Breton ([LE 99]) draws a parallel with the fact that we now use our body less and less in everyday life (cars, lifts, sitting position for working, the Internet, virtual world, etc.): the body has atrophied. This restriction of physical and sensorial activities changes our perception of the world, limits our impact on reality and weakens our actual knowledge of things.

Norbert Wiener was one of the first to blur the line between machine and life. The brain is an intelligent machine that can be mimicked with a computer. The body is inessential, and we may dream of downloading a spirit into a computer, as in some science fiction film ..

But according to Hubert Dreyfus, artificial intelligence underlies some erroneous metaphysics3 [DRE 79]:

• - a biological assumption: "The brain processes information in discrete operations by way of some biological equivalent of on/off switches";
• - a psychological assumption: "The mind can be viewed as a device operating on bits of information according to formal rules";
• - an epistemological assumption: "All knowledge can be formalized";
• - an ontological assumption: "The world consists of independent facts that can be represented by independent symbols".

It has been found that the current understanding of the human mind was based on engineering principles and problem-solving techniques related to management science. Modern artificial intelligence research is now more open to issues that have become important to modern European philosophy and psychology, such as situatedness, embodiment, perception and gestalt.

For the moment, we can still claim that there are fundamental differences between man and machine: the computer has no "marginal consciousness", making man sensitive to various and multiple facts of his environment. The computer is not able to use a context and bring ambiguous words or situations into perspective, making them thus intelligible; it does not distinguish what is essential and what is auxiliary using intuition. The computer is designed for precise works; it is not as versatile and adaptive, as the human brain is. And finally the computer has no body (except maybe if we consider robots).

This question of the role of the body in the implementation of intelligence is a major question. Humans are not facing a world made up of parameters to be recorded and processed. Understanding relies on a symbolic system: language and body. Language is not a code made up of unambiguous signs. It is based on culture and history (except maybe for a common core that can be found in all languages, see Noam Chomsky's theories). Words always impart more than their definition; they have an evocative power (now and then magic and religious). The body is a measurement of the world: through his/her body, an individual interprets his/her environment and acts on it according to some influences related to his/her habits or education.

Perception is at the...