"How is it possible that the brain, apparently so solid, is the support for things as impalpable as thoughts?" The question that Marvin Minsky poses at the beginning of his celebrated book The Society of Mind is, in a sense, the question that, besides igniting a very heated debate on a whole series of issues spanning philosophy, psychology and linguistics, "certified" in 1985, the year the book appeared, thirty years of research into artificial intelligence and into the attempt to reproduce a typically human faculty: thought.
Minsky's recent passing, in January 2016 at the age of 88, did not go unnoticed. Minsky was part of that small band of figures, truly extraordinary and also a little "mad", who marked the history of contemporary technology and who bring back to mind everything that was done, back then, and which today we hardly notice when we use a telephone, a tablet, a computer, or when a doctor performs robotic surgery or a pilot trains with a flight simulator.
Let us proceed in order: let us go, for a moment, to an extraordinary period in the history of technology: the year is 1955. The term "informatics" has not yet even been conceived, even though the most recent period had been very intense as regards technological development.
The preceding years had indeed been dense with spectacular achievements: Alan Turing, after formalizing the concept of the "algorithm", fundamental to the science of computers, and after devising the test that bears his name, indispensable for the future studies on artificial intelligence, died tragically in 1954.
John von Neumann, the so-called evil genius[1], reworking Turing's 1936 ideas on the construction of a so-called "universal machine", conceived the EDVAC (Electronic Discrete Variables Automatic Computer), the first programmable digital machine (thus with a stored program), based on what would later be called the famous "von Neumann architecture", an important logical and functional scheme that still today guides the construction of modern computers, even though at the time the size of its memory was a mere 5.5 Kilobytes.
As we were saying, the year is 1955. The figures in question are John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon. The document drawn up on 31 August of that year by the four, the so-called "Dartmouth proposal", announces what the organizers consider the main themes of the field of research, among them neural networks, computability theory, creativity, and the processing and recognition of natural language, and proposes to meet the following summer. The original words of the four:

“We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.[2]“

McCarthy (1927-2011), the inventor of the term "artificial intelligence", was an electronic engineer and, among other things, the author of the celebrated programming language Lisp, the oldest high-level programming language still in use today, after Fortran. In the early 1950s, when he was at MIT, McCarthy had devised time-sharing, a procedure, adopted from that moment on by most operating systems, through which many users can work on a single computer at the same time and without which it would have been hard to imagine today's Internet. From this idea there later sprang the concept of multitasking, widely applied in the operating systems of recent generations, according to which the operating system can run several programs at the same time.

Minsky (1927-2016), for his part, graduated in mathematics at Harvard with top marks in 1950, but also studied physics, neurophysiology and psychology, and after obtaining a doctoral fellowship at Princeton University, where he built the first neural network simulator, he moved to the Massachusetts Institute of Technology, where he would remain for the rest of his life. At MIT he would later found the Artificial Intelligence Lab and, together with Seymour Papert, would write important pages for the future studies on the logical omnipotence of computers and, again together with Papert, would create the Logo language, the first programming language with educational aims.
Besides being one of the founders of the so-called free software movement, forerunner of open source, Rochester (1919-2001) too was an electronic engineer at MIT.

In the early 1950s he designed the IBM 701, the first commercially produced computer in history; he also designed the first symbolic assembly language and would later lend McCarthy a hand in fine-tuning the LISP language.

Shannon (1916-2001), for his part, was a mathematician who became famous for laying the foundations of information theory. An expert in cryptography for the government during the Second World War, Shannon applied the numbers 0 and 1 of Boolean logic to electrical circuits (thus introducing the bit, binary digit, the minimal unit of information), and so laid the theoretical foundations of the systems for encoding, processing and digitally transmitting information. Without Shannon's theory there would be no CDs, no concepts of input and output, and not even lossless file compression (such as .zip files), lossy encoding (as in .mp3 files) and the digital modulations used in ethernet transmissions, along with many other fundamental features of today's technology. In the photo at bottom left, Shannon is with one of his automata, the electromechanical mouse Theseus, able to "learn" to find the way out of a maze, one of the first experiments in "learning by experience" (1952).

Our four are also rather eclectic and creative: while Shannon dabbled in juggling, Minsky invented the confocal scanning microscope, predecessor of the laser scanning confocal microscope so widespread today; he played the piano, and in 1968 he would serve as scientific adviser on the film production of 2001: A Space Odyssey.
In 1956 the conference takes place. The venue is Dartmouth College. The event, under the name "Dartmouth Summer Research Project on Artificial Intelligence", officially marks the starting date of studies on the subject. From that moment, one might say with a common phrase, nothing would ever be the same again.
For the occasion, the four of the first hour are joined by Ray Solomonoff, Oliver Selfridge, Trenchard More, Arthur Samuel, Allen Newell and Herbert Simon. Without taking anything away from the others, Newell and Simon deserve a place of note: the former a mathematician, psychologist and computer scientist, author of the celebrated General Problem Solver; the latter a future Nobel laureate in economics for his studies on business organization. Newell and Simon shared the Turing Award in 1975[3].
During the conference, Newell and Simon presented the Logic Theorist, the first program explicitly designed to imitate the problem-solving abilities of human beings.
The spring, or in a certain sense the provocation, driving the group is precisely the attempt to demonstrate that every aspect of human learning, indeed every feature of human intelligence, can be simulated by a machine and, given that the machine par excellence at that moment is the calculating machine, by a computer. The "cognitivist" turn in psychology, which as is well known privileged the logical-symbolic, rational, computational aspects of the human mind and identified in the computer a reference model for attempting to explain many of the processes through which human beings know and reason, provides the frame for the proposal.
In this field, the questions multiply: can we speak of knowledge through machines, as when they are used to dispense information or mediate human communication, or even of knowledge belonging to machines? What is intelligence? Is it merely the manipulation of symbols, is it only a computational process, in a word is it "calculation", as many AI theorists have always maintained, or is it something more? And again, is there an analogy between brain and computer? Or, at least, which aspects of the human mind can be reproduced by a computer? The machine carries out, infallibly and deterministically, a defined series of instructions dictated by man, and could never take the initiative; and yet anyone who has had occasion to work with a computer knows that it sometimes astonishes us with the surprising speed at which it draws correlations among complex data. And finally, can a computer that wins a game against the world chess champion be considered "intelligent"? Can an artificial mind be produced, just as a car or a household appliance is produced[4]?
Minsky and Papert sought to explain that within the human brain there is organized a myriad of "little agents", each specialized in carrying out a single operation (keeping one's balance, understanding language, etc.), but which in themselves are not intelligent. Intelligence derives, instead, from the cooperation and interaction of these agents. This vision was echoed by Hilary Putnam's celebrated functionalist dualism, according to which the mind is to the brain as the program (software) is to the machine (hardware): a vision with a strong behaviorist influence, according to which attention was to be placed not on what constitutes a system, but on how it functions: the differences between man and machine could be identified only by their behavior.
In the end, however, the debate revealed positions that were unquestionably ideological: on one side, a materialistic vision of the human mind, with a sort of mechanism of mental states; on the other, a conception that tries to highlight the specific and ineliminable traits of the human being with respect to any device.
Perhaps we need to return to the distinction, introduced by John Searle, between weak artificial intelligence and strong artificial intelligence. While the former holds that the computer's main contribution to the study of the mind is to give us a very powerful and precise tool for formulating and examining hypotheses, the latter maintains that the brain is nothing but a digital computer and the mind nothing but a computer program: the consequence is that the brain is one among an indefinitely large number of types of hardware capable of supporting the programs that constitute human intelligence. It is precisely the idea of strong artificial intelligence that unleashed an uproar in the contemporary debate, turning a few working hypotheses into a very heated philosophical dispute.
Today we know that, after all, a machine that defeats the world chess champion cannot be considered intelligent, only extremely fast at manipulating symbols, and the ambitious project of building a universal translator failed precisely because the machine lacks the notion of "context", indispensable for carrying a text from one language into another by choosing among the many terms with which a word can be rendered.
So the ability to calculate rapidly and the manipulation of symbols do not seem to be the only characteristics of human thought; perhaps the four of Dartmouth and their successors gave too much weight to those aspects of our intelligence that lend themselves best to being replaced by the machine, such as memory, the speed of connections, formal logical reasoning, etc., but there are other aspects of our being human that cannot be reduced to pure representation.
Perhaps the enormous debate that arose was not even among the aims of Dartmouth, but it has provided indispensable help in understanding what we can ask of a computer and what remains an exclusively human prerogative.

Notes

[1] In reference to his involvement in the project to build the atomic bomb and to his hatred of the Nazis, the Japanese and the Soviets, which would culminate in his proposal for a preventive bombing of the Soviet Union.

[2] McCarthy, J., Minsky, M.L., Rochester, N., Shannon, C.E., A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence – August 31, 1955, “AI Magazine”, 2006, vol. XXVII, (4) – aaai.org

[3] The A.M. Turing Award is a prize, presented annually by the Association for Computing Machinery (ACM), to an individual who stands out for technical contributions offered to the computing community, in particular for advances that are lasting and of high technical importance. The award is often also called the "Nobel Prize of computing" (it.wikipedia.org).

[4] Cf. Capponi, M., Il computer come ambiente di apprendimento, Perugia, Morlacchi, 2003, pp. 53-54.

Massimo Capponi is a tenured university researcher at the University of Perugia, where he teaches Computer Science Applied to Education in the degree program in Education Sciences.