Contenu du sommaire : La communication au cœur des connaissances

Algorithms are integral to our environment and have been incorporated into many decision-making techniques. Computation and analysis – which are often linked to collections of data produced by consumers/citizens/users themselves – are powerful sources of genuine progress : better resource allocation, improvements in how products and services meet needs, and so on. However, algorithms remain human constructions and should be considered as such. Their biases must be understood, and their power and roles have to be assessed from a critical perspective. They need to be demystified.

Algorithms have become ubiquitous. Some are procedural, but there is another type of algorithm that is experiencing extremely rapid growth, and which we are not able to control. We do, however, have control of the data. In all areas of decision-making, it is thus crucial to define carefully what will be delegated to algorithms. This raises three questions : 1) what responsibility do algorithms have ? ; 2) what authority should we concede to them ? ; 3) what sort of governance is needed ? (this should not be conflated with questions around net neutrality).

We are currently witnessing a genuine revolution in the field of machine translation. The spectacular progress that has been made in the last two decades is even more valuable because our world is rapidly becoming “rebabelized.” However, the idea that machines will supersede “human” translators and interpreters belongs more to science fiction than to fact, and is based on a reductionist vision of “human” language and translation, which no machine will ever emulate.

Machine translation has made sudden and spectacular progress of late, and computers are (finally) capable of producing translations that are useful for a wide range of situations. Should translators be concerned about this new form of competition ? In this article, a study of “translationese,” the language of translation, explains why fully automated machine translation remains a very distant goal. In the short term, however, advances in the field are allowing humans and machines to work together more effectively.

Interdisciplinarity is the connecting up of fields of study that have been separated artificially by disciplinary barriers. In France, one of the key moments of its prehistory was the publication of the Encyclopédie, edited by Diderot and d'Alembert ; its history itself began with Henri Berr's creation of the Centre international de synthèse in 1925. During the last three decades of the twentieth century, interdisciplinarity gave rise to a number of scientific research programs. It has also been provided with different yet parallel institutional frameworks at French grandes écoles, universities, and the Centre national de la recherche scientifique, France's largest governmental research organization. As a mindset associated with a commitment to openness, it involves a form of mental conversion that has only, at present, been partially achieved.

This article, which is based on an educational research project, presents three illustrations of good interdisciplinarity : a kind of thinking that is both action-oriented and conceptual. It shows how such thinking has the potential to advance specific disciplines, and also highlights the conditions under which interdisciplinarity can be implemented, in terms of the deepening of epistemology, the organization of communication, and institutional structure. Finally, it suggests that interdisciplinary affinities are worth cultivating.

Political and scientific communication are two distinct registers. However, and despite what a Weberian analysis would suggest, they are not opposed to one another. When researchers intervene in the public sphere in order to show how research – whether it be their own or from their field – can shed light on issues of general concern such as climate change, they engage in a form of public science in which scientific and political communication converge in a single objective : to preserve our democracies through the search for individual and collective emancipation. This does not involve “manufacturing consent” ; instead, it constructs productive dissent.

Scientists play an essential role in helping us to understand the complexities of the dynamics of societies and their environments at a time of global change. A sustainable world will involve transmitting acquired knowledge and producing new knowledge to enlighten decision makers, managers, technicians, and other social actors. For this reason, this article uses examples related to climate change and the sustainable development goals (SDGs) in order to explore how scientists communicate and politicians appropriate both acquired and emerging knowledge.

Europe has been weakened of late, in both its international institutions and its very concept. The emergence of populist currents raises questions. In this time of European turbulence, the present text ventures to hope that another modernity (or post-modernity) is possible for Europe. This could be grounded on three points that have, at present, broken down : 1) a rootedness in common European culture in place of insular national identities ; 2) a renewed spirit of compromise rather than limp consensus ; and 3) an open cosmopolitanism that respects the “soul of the nation,” instead of an abstract universalism arising from globalization.

In 2015-2016, Europe faced an unprecedented influx of refugees. This revealed both the existence of a European incommunication in relation to EU member states and a crisis of confidence aggravated by a lack of solidarity among Eastern and Western European countries. The headlong rush to try to establish a dialogue with countries of the Mediterranean south was fraught with misunderstanding, since the immigration policies of Europe and its member states were more concerned with appeasing public sentiment than with providing a rational response to the immigration issue.

One practice at the heart of the scientific understanding of real-world phenomena involves reducing the latter to one or more curves on a graph. This takes a very specific form in the social and life sciences. Curves enable the researcher to become aware of certain aspects of the phenomena that are being studied, phenomena that would otherwise remain vague and only partially intelligible. This is the case for the major products of human thought, among which artificial intelligence (AI) is no exception. This article shows how AI follows the trajectory of natural logic, which is presided over by the dynamics of the perfect curve of exponential growth.