Seymour Papert (1928-2016)

October 24, 2016

When my son was young, and personal computers even younger, I wrote a computer aided instruction (CAI) program for him for the elementary mathematics operations of addition, subtraction, multiplication, and division. At that time, we had an Apple II clone, a Franklin computer, for our children to use, so I wrote the program in BASIC. The program was an electronic version of a school workbook that would present practice problems and wait for answers to be entered.

A portion of page 51 of S. A. Felter and Erskine Hewitt, "The first lessons in numbers : an illustrated table book, designed for elementary instruction," Scribner, Armstrong, and Co. (New York, 1868), pp. 108, via archive.org.

Answers were checked, and there was a time limit on accepting an answer before the next problem was presented. In that way, I was able to include logic to give more emphasis on the type of problem missed, reverting back to easier examples. As I remember, my son disliked using this program, but he was forced to used it nonetheless. Later, he became a computer scientist. This program was text-based, but in later years I wrote a graphical CAI program in Visual Basic that gave instruction on Latin noun declension, this time for my daughter.

These were not my first forays into CAI. As a graduate student, I was paid to write a CAI program on a simple statistical concept for the students of a psychology professor. This program was hosted on an IBM 360 mainframe computer, and it was written in the APL programming language. I didn't get any user feedback, but his students likely hated using this program as much as my son hated his math program.

Some programming topics have been with us from the origins of useful computer hardware. There's bookkeeping and payroll, language translation, and computer-aided instruction. The first general-purpose CAI system was PLATO, an acronym for "Programmed Logic for Automated Teaching Operations." PLATO was conceived and programmed by physicists, who routinely wander into areas not associated with physics.[1]

In the late 1950s, Chalmers W. Sherwin, a physicist at the University of Illinois (Urbana, Illinois), proposed a CAI system to the dean of engineering, the engineering school being the haven for a university computer and programming talent. Donald Bitzer, an assistant to physics professor, Daniel Alpert, built a demonstration system. Bitzer, who realized that a graphical user interface was important to a CAI system, is considered to be the Father of PLATO.

As they say, "A picture is worth a thousand words." This PLATO V terminal, with an orange plasma display, is dated 1981. (Via Wikimedia Commons).

As the project progressed, the National Science Foundation provided funding for PLATO and a CAI laboratory at the University of Illinois. Several years thereafter, PLATO was being implemented on a variety of mainframe computers worldwide, and it was eventually licensed by Control Data Corporation (CDC). Unfortunately, the expense of running such a system resulted in few customers, and PLATO was discontinued at CDC by the mid-1980s.

In the decade after PLATO, a type of CAI specific to computer programming was developed at the MIT spin-off, Bolt, Beranek and Newman (BBN), by a research team that included Seymour Papert. Papert is known, also, for co-authoring the book, Perceptrons with Marvin Minsky.[2] The BBN CAI instructor was Logo, a programming language that used a pen-wielding "turtle" as an output device. Seymour Papert died on July 31, 2016, at age 88.[3]

Seymour Papert (1928-2016). Papert is shown with a Logo robot turtle. Modern versions of Logo create turtle graphics on computer screens, not on paper. However, current student interest in robotics may bring back the mechanical turtle. (Matematicamente.it photo, via Wikimedia Commons.)

The Logo language has simple commands for turtle movement, and it introduces the concept of loops, as shown by this short example for drawing a square 50 units on a side.

REPEAT 4 [FD 50 RT 90]

In this example, the command FD 50 has the turtle move forward 50 units, the command RT 90 calls for a 90-degree right turn, and REPEAT 4 has everything within the brackets repeat four times to give the four sides of the square. As an exercise, a computer science student could easily knock-out a Logo interpreter in one of today's high-level languages as a class project. Some complex images can be made using Logo, as the following figure demonstrates.

One level of a Hilbert curve drawn using Logo. German mathematician, David Hilbert, invented this fractal curve in 1891. If the curve continued, it would completely fill the plane. (Via twentygototen.org.)[4]

Seymour Papert was born on February 29, 1928, in Pretoria, South Africa. Papert, whose father was an entomologist, was awarded a Ph.D. in mathematics from the University of the Witwatersrand (Johannesburg, South Africa) in 1952.[4] He received a second Ph.D. in mathematics, this time from Cambridge University (Cambridge, UK), in 1959.[4]

In a surprising shift for someone with two mathematics Ph.D.s, Papert went to the University of Geneva to work with the Swiss psychologist, Jean Piaget, from 1959-1963. It was there that he developed his theory of education that children learn from doing more than from studying.[4]

Papert came to the Massachusetts Institute of Technology (MIT, Cambridge, Massachusetts) in 1963 to work on artificial intelligence with Marvin Minsky. In 1967, he was named a professor of applied mathematics, and he served as co-director with Minsky of the MIT Artificial Intelligence Laboratory.

Papert was one of the founding faculty of the MIT Media Laboratory in 1985, and he participated in its One Laptop per Child (OLPC) project to manufacture inexpensive laptop computers to be used worldwide. While it's estimated that more than 3 million such laptops were shipped, the advent of inexpensive cellphones and tablet computers has offered an inexpensive alternative. Papert was also involved in the creation of Lego Mindstorms.

Papert's philosophy of education is summarized in Papert's principle, that

"Some of the most crucial steps in mental growth are based not simply on acquiring new skills, but on acquiring new administrative ways to use what one already knows."[5]

References:

1. In the 1970s, there was a joke that a physicist is the type of programmer who writes a database application in Fortran. That's because Fortran is the staple language of physics. There were better database languages, even in those dim days.
2. Marvin Minsky and Seymour A. Papert, "Perceptrons: An Introduction to Computational Geometry, Expanded Edition," The MIT Press, December 28, 1987, Paperback, 308 pp., ISBN-13: 978-0262631112.
3. Gary S. Stager, "Seymour Papert (1928–2016)," Nature, vol. 537, no. 7620 (September 15, 2016), doi:10.1038/537308a.
4. Examples: koch snowflake, hilbert curve, spiral, twentygototen.org. "20 GOTO 10" specifies an infinite loop in several languages.
5. Marvin Minsky,"Chapter 10: Papert's Principle," from The Society of Mind, Simon & (New York, 1988), p. 102 (via Amazon).