Stefano Franchi

The book brings into relief the variety of approaches and disciplines that have informed the quest for a theory of cognition. The center of interest are the historical, geographical, and theoretical peripheries of classic AI's mainstream research program. The twelve chapters bring back into focus the variety of strategies and theoretical questions that researchers explored while working toward a scientific theory of cognition and pre-cognition. The volume is organized in four parts, each one including three essays. The first one deals with cybernetics, the approach that may be considered as the most important periphery of classic AI research. The second part focuses on the geographical periphery of AI research. It examines how the theories and techniques developed on AI's home ground were translated into countries with different cultures and traditions: Italy, France, and the Soviet Union. The third part focuses on AI's periphery understood in the cultural and historical meaning of the term. It contains essays that locate some of the central concepts of AI, like representation and computability, within a broader philosophical (Descartes, Aristotle, Leibniz) and technical background (programming theory and practice). The fourth and final part of the volume is focused directly on the limitation of Turing's classic computability theory and its possible alternatives, some of which were studied in the early years of AI's research (e.g. Ashby's re-entrant information model), while others have been intensely studied in recent times (quantum automata).

Research in Artificial Intelligence has always had a very strong relationship with games and game-playing, and especially with chess. Workers in AI have always denied that this interest was more than purely accidental. Parlor games, they claimed, became a favorite topic of interest because they provided the ideal test case for any simulation of intelligence. Chess is the Drosophila of AI, it was said, with reference to the fruit-fly whose fast reproductive cycle made it into a favorite test bed for genetic theories for almost a century. In this paper I will try to show Artificial Intelligence’s relationship to games is quite different from what this analogy suggests. In fact, I will argue that AI is, at core, a theory of games and a theory of subjectivity as game-playing.

Open peer commentary on the target article “From Objects to Processes: A Proposal to Rewrite Radical Constructivism” by Siegfried J. Schmidt. Upshot: I focus my commentary on the fundamental metaphysical issue that Siegfried J. Schmidt’s very stimulating paper addresses in §45 and particularly upon the relationship between the ontological status of the processes from which worlds emerge and the temporality of the objects to be found therein. I argue that Schmidt’s emphasis on world-forming processes raises many questions concerning the temporal stability of objects and the relationship between objects and actors belonging to different worlds. I suggest that some classic as well as contemporary thinkers (e.g., Fichte, Hegel, Heidegger, Gadamer, Foucault, and S. J. Gould) have faced similar problems and discuss how their answers could be integrated within Schmidt’s revised radical constructivism

Context: W. R. Ashby’s work on homeostasis as the basic mechanism underlying all kinds of physiological as well as cognitive functions has aroused renewed interest in cognitive science and related disciplines. Researchers have successfully incorporated some of Ashby’s technical results, such as ultrastability, into modern frameworks (e.g., CTRNN networks). Problem: The recovery of Ashby’s technical contributions has left in the background Ashby’s far more controversial non-technical views, according to which homeostatic adaptation to the environment governs all aspects of all forms of life. This thesis entails that life is fundamentally “heteronomous” and it is conceptually at odds with the autopoiesis framework adopted by Ashby’s recent defenders as well as with the primacy of autonomy in human life that most of the Western philosophical tradition upholds. The paper argues that the use of computer simulations focused on the more conceptual aspects of Ashby’s thought may help us recover, extend and consequently assess an overall view of life as heteronomy. Method: The paper discusses some computer simulations of Ashby’s original electro-mechanical device (the homeostat) that implement his techniques (double-feedback loops and random parameter-switching). Results: First simulation results show that even though Ashby’s claims about homeostatic adaptivity need to be slightly weakened, his overall results are confirmed, thereby suggesting that an extension to virtual robots engaged in minimal cognitive tasks may be successful. Implications: The paper shows that a fuller incorporation of Ashby’s original results into recent cognitive science research may trigger a philosophical and technical reevaluation of the traditional distinction between heteronomous and autonomous behavior. Constructivist content: The research outlined in the paper supports an extended constructionist perspective in which agency as autonomy plays a more limited role

Open peer commentary on the article “Constructivism and Computation: Can Computer-Based Modeling Add to the Case for Constructivism?” by Manfred Füllsack. Upshot: While I do agree with Füllsack’s positive assessment of the use of computer simulations in advancing constructivism’s program, I am less convinced by the alleged opposition between computers and constructivism he builds up. In my opinion, his depiction of computers and computation is inaccurate in several respects. As a result, the alleged incompatibility with constructivism Füllsack objects to disappears, and his whole essay could then be construed as a classic straw man fallacy. This is unfortunate, because it could lead to an unwarranted dismissal of his positive contribution

Open peer commentary on the article “A Cybernetic Computational Model for Learning and Skill Acquisition” by Bernard Scott & Abhinav Bansal. Upshot: Scott and Bansal’s assessment of the limitations of their work relies on a concept of simulation that I find problematic. It assumes that the ultimate goal of a model is a replication of the phenomena it applies, whereas a limited model produces only simulations. I argue that this position leads to unfortunate epistemological results, and it ends up assigning an unduly exclusive role to the study of the biochemical substrate of cognition

Herbert Simon’s work presents a curious anomaly to the historian and philosopher trying to understand the development of classic Artificial Intelligence (AI). Simon was one of most influential figures in AI since its birth, and yet it is always with some difficulties that his work can be made to fit within the received canon of AI’s development and goals. In fact, he differed from every other figure in early AI on most counts: in terms of the recognized intellectual heritage of AI, of his own background and training, of the goals he set for his own AI work and the assessment criteria he accepted. I will argue that these differences provide important clues toward a reevaluation of the relationship between Artificial Intelligence and Herbert Simon’s work that may change our current understanding of both. On the one hand, classic Artificial Intelligence (or Complex Information Processing, as Simon preferred to call it for a number of years), provided the tool he needed to pursue a much broader research agenda that strove to encompass human beings in their cognitive, emotional, social, and political dimensions. On the other hand, AI’s curious status as the discipline that straggles the boundaries between engineering, science, and philosophy can be recast as the tool that allowed researchers to pursue philosophy’s old goals with an alternative methodology. From this perspective, Simon’s version of Artificial Intelligence becomes a full-fledged form of ”anti-philosophy” as ambitious and broad-ranging as old-fashioned metaphysics and as revolutionary as the latter in the radical refashioning of its methodology. It follows that a philosophical assessment of AI, at in least in its Simonian incarnation, must be more farreaching than it is usually thought. At the methodological level, it must discuss whether AI’s invention of computer simulation as the tool that overcomes the a-priori/a-posteriori distinction by actually producing the behavior it wants to explain is really adequate to the job at hand..

Upshot: The target article suggested that Ashby’s device, the homeostat, embodies and illustrates a conception of life as a passive-contingent phenomenon. It advocated renewed experiments with updated and extended versions of his device that would allow us to understand better what passive-contingent life “would be like.” In assessing the proposal, we should be particularly careful when dealing with the concept of “passivity,” and we should not mistake the proposed theoretical exploration for a substantial metaphysical thesis about life in general

This dissertation examines the contention that philosophy might have recently come to an end and that a decisive use of the concepts of game and play may prove crucial to a successful overcoming or a definitive replacement of Western metaphysics. I show that this double contention is marred by two difficult problems that must be clarified: the possibility of the "end of philosophy" pervades the very definition of philosophy at work in the Western canon, and the concept of play, or a family of concepts related to play, have classically been used in conjunction with the project of "ending philosophy." ;A close reading of the Lectures on the History of Philosophy and of the Phenomenology of Spirit provides ample evidence for the general validity of the Hegelian argument: philosophy must reach an end, in all the variety of meanings of this term , if the discipline is to have any hope to be transformed from a mere love of knowledge into an accomplished Wissenschaft. The second point is established through a broad analysis of both Hegelian philosophy and other traditions that see themselves as non-philosophy, i.e. as replacements of metaphysics : in all cases, a wide recourse to the concept of play is crucial, but ultimately unsuccessful, to either the accomplishment or the overcoming of metaphysics

Open peer commentary on the article “What Can the Global Observer Know?” by Diana Gasparyan. Upshot: While I agree with Gasparyan’s incisive critique of the concept of the “general observer,” her use of the concept of “event” is somewhat ambiguous. On the one hand, she equates “events” to Wittgenstein’s and “configurations of objects” or “states of affairs” and she consider the world as a collection of such states of affairs. On the other hand, she cites Badiou’s work in support of her criticism of the “general observer.” Yet Badiou’s conception of events is very different from Wittgenstein’s: they are locally produced by historically situated agents in their concrete inter-subjective activities. I argue that a resolution of the target article’s ambiguity in favour of Badiou’s conception of events would allow constructivism to escape a narrow epistemological reading of its basic thesis in favor of a broader interpretation

In lieu of an abstract, here is a brief excerpt of the content:Hunters, Cooks, and Nooks:Two Interpretations of the Tangled Relationship Between Philosophy and ScienceStefano Franchi (bio)Knowledge is the measure of all things.—Plato, Prot. 361b1Preliminaries: Double QuestioningWhen philosophers ask questions about science, they usually do so in the context of one specific discipline whose latest results or whose historical development seem to pose genuinely philosophical problems: for instance, the nature of space/time, the nature of intelligence, the nature/nurture debate. It is rarer to hear a philosopher asking questions about the general relationship between philosophy and science, even when, as in the cases above, the answer to this question provides the necessary ground for any specific investigation of the allegedly genuine philosophical problems arising in physics, biology, or cognitive science.The main difficulty lies in the intrinsic ambiguity of the term "science." On the one hand, "science" can be considered as a descriptive term: it is the genus of the empirically given, scientific disciplines (the "sciences," in the plural). On the other hand, "science" (episteme, scientia, Wissenschaft) (always in the singular) is a theoretical term that denotes a form of knowledge. More precisely, it refers, as philosophers say, to a "conceptually unified body of objective knowledge." Traditionally, the latter meaning includes philosophy itself within its denotation. Thus, when both meanings of the term "science" are considered, to interrogate the relationship between philosophy and science is always, and preliminarily, to question the nature of philosophy itself, even when the question as such is not explicitly raised. In all these cases, a characterization of the nature of philosophy is presupposed by a philosophical analysis of scientific results.However, unless the specificity of concrete scientific investigations is denied from the outset, the questioning, even when conducted from a circumscribed philosophical standpoint, cannot be reduced to a single gesture that interrogates the scientificity of philosophy. On the contrary, it must also concurrently question the possibility of a specific relation between a determinate philosophical endeavor and specific scientific pursuits. Any questioning of the relationship between philosophy and science will always imply a double gesture: on the one hand, a question about what makes or does not make philosophy a science; on the other, what brings philosophy closer, or moves [End Page 98] it farther apart, from the sciences. Most often, philosophers focus their attention on just one aspect of this double question. In this paper, I would like to address both sides of the issue. In particular, I would like to examine the two very different answers to this general problem given by Plato and by Michel Foucault. Although their answers address fundamentally the same issue, they are so diametrically opposed, as I will show, that they may be considered as the external boundaries of a very broad range of possible responses.2SignpostsLet me begin with Plato. In a well-known passage from the Euthydemus, the main characters are discussing which kind of knowledge is required by wisdom. In this context, to the delighted although belated surprise of Crito, the young Cleinias introduces a crucial distinction between hunters and cooks: the former know how to find the prey, but only the latter know how to use it:No art of actual hunting, [Cleinias] said, extends any further than pursuing and capturing; whenever the hunters catch what they are pursuing they are incapable of using it, but they and the fishermen hand over their prey to the cooks. And again, geometers and astronomers and calculators (who are hunters too, in a way, for none of these make their diagrams; they simply discover those which already exist), since they themselves have no idea of how to use their prey but only how to hunt it, hand over the task of using their discoveries to the dialecticians—at least, those of them who are not completely useless [anoetoi].[290b–c]1Cleinias's account acknowledges the reality and value of the knowledge gained by the sciences, but denies that scientists have full awareness of what they are doing and of the implications of their own work. Scientific knowledge lacks both self-reflexivity and transparency: it does not have the ability to bring to full evidence the principles it rests upon...

Believing that the enterprise of constructing "artificial intelligence" transcends the bounds of any one discipline, the editors of Mechanical Bodies, Computational Minds have brought together researchers in AI and scholars in the humanities to reexamine the fundamental assumptions of both areas. The AI community, for example, could benefit from explorations of human intelligence and creativity by philosophers, psychologists, anthropologists, literary critics, and others, while analysis of AI's theoretical struggles and technical advances could yield insights into such traditional humanist concerns as the nature of rationality and the mind-body dichotomy.The contributions include a continuation of the famous Hubert Dreyfus-Daniel Dennett debate over Kasparov's defeat by IBM's Deep Blue; Philip Agre's tracing of difficulties in AI research to the inherited tensions of Cartesian dualism; Evelyn Fox Keller's examination of the development of computer technology in relation to biology; Douglas Hofstadter's argument that thinking is more than the theorem-solving activities of AI; and Alison Adam's discussion of the implicitly male universal subject used in AI