United States is approval from the union. Soviet Russia claims she
already has computer-run subways and even ships. The latter application
took place on the oil tanker _Engineer Pustoshkin_ plying the Caspian
Sea. The main complaint of the director of this research work, P.
Strumpe, is that ships are not yet designed for computer control and
will change for the better when their designers realize the error of
their ways.
[Illustration:
_Hughes Aircraft Company_
Mobot Mark II, carrying a Geiger counter in its “hands,” demonstrates
how it can substitute for men in dangerously radiated areas.
]
Minneapolis-Honeywell in this country is working toward the complete
automation of buildings, pointing out that they are as much machines as
structures. A 33-story skyscraper in Houston will use a central computer
to check 400 points automatically and continuously. Temperature and
humidity will be monitored, as well as doors and windows. Presence of
smoke and fire will be automatically detected, and all mechanical
equipment will be monitored and controlled. Equipped with cost figures,
the central computer will literally “run” the building for optimum
efficiency and economy. Harvard University has a central control for
seventy-six campus buildings, and in Denver work is being done toward a
central control for a number of large buildings. It is fitting that
automation of buildings be carried on, since historically it was in the
home that self-control of machines was pioneered with automatic control
of furnaces with thermostats.
[Illustration:
_Robodyne Division, U.S. Industries, Inc._
TransfeRobot assembly-line worker installs clockwork parts with speed
and precision.
]
In this country our traffic is crying for some kind of control, and New
York is already using punched-card programming to control part of the
city’s traffic. The Federal administration is studying a bold proposal
from RCA, Bendix, General Motors, and Westinghouse for an automatically
controlled highway. The reason? Traffic is getting to be too much for
the human brain to deal with. A better one has to be found, and the
computer is applying for the job.
The coming of automation has been likened to a tidal wave. It is useless
to shovel against it, and the job would seem to be to find suitable life
preservers to keep us afloat as it sweeps in over the world. One
approach is that of a nonprofit foundation to study the impact of
automation on workers. This group, a joint United States Industries,
Inc., and International Association of Machinists organization, has
already come up with a scheme for collecting “dues” from the machines,
in annual amounts of from $25 to $1,000, depending on the work output of
the machine.
A key project of the foundation is a study of effective retraining of
workers to fit them for jobs in the new, computerized factory. Such
studies may well have to be extended from the assembly line to the
white-collar worker and executive as well. The computer can wear many
different kinds of hats!
------------------------------------------------------------------------
Teaching Machine Age
Lilyn E. Carlton in _Saturday Review_
“_In the good old-fashioned school days,
Days of the golden rule,
Teacher said, ‘Good morning, class,’
And so she started school._
_Alas! How different things are now,
The school day can’t begin
Till someone finds the socket
And plugs the teacher in._”
10: The Academic Computer
It was inevitable that the computer invade, or perhaps “infiltrate”
is the better word, our education system. Mark I and ENIAC were
university-born and -bred, and early research work was done by many
institutions using computers. A logical development was to teach formal
courses in using the computer. While application of the machine in
mathematical and scientific work came first, its application to business
and to the training of executives for such use of the computer was soon
recognized. As an example, one of two computers installed by U.C.L.A. in
1957 was for use exclusively in training engineering executives as well
as undergraduates in engineering economy.
Early courses were aimed at those already in industry, in an attempt to
catch them up with the technology of computer-oriented systems in
business and science. As special courses, many of these carried a high
tuition fee. Next came the teaching of professors and deans of
engineering institutions in techniques of computer education for
undergraduates. Today the computer is being taught to many students in
many schools. New York University has a $3 million computer at its
Courant Institute of Mathematical Sciences, being used by students in
basic and applied research on projects ranging from the design of
bridges to the analysis of voting patterns in Congress.
M.I.T. recently added a digital computer to teach its students the
operation of electronic data-processing equipment. Another computer is
used in more sophisticated work including speech analysis, study of
bioelectrical signals, and the simulation of automata as in the “Hand”
project. At the computing center of the University of Michigan a second
generation of computers is being installed. Students in some one hundred
different courses use these computers, programming them with a language
developed at the University and called MAD, for Michigan Algorithm
Decoder. These are typical examples of perhaps two hundred schools using
computers.
That knowledge of computer techniques is essential for the engineering
graduate is evident in the fact that of a recent class of such students
at Purdue, 1,600 used the computer during the term. Less known is the
integration of computer courses in secondary education. The Royal McBee
Corporation teaches a special course on the computer to youngsters at
Staples High in Westport, Connecticut. At the end of the first four-week
session it was found that the students, fifteen to seventeen years old,
had learned faster than adults. At New York’s St. Vincent Ferrer
Catholic High School, 400 girls participated in a similar project
conducted by Royal McBee. Other high schools are following suit, and
computers are expected to appear in significant numbers in high schools
before the end of 1962. Textbooks on computers, written for high-school
students, are available. As an example of the ability of young people in
this field, David Malin of Walter Johnson High School in Rockville,
Maryland, read his own paper on the use of computers to simulate human
thought processes to science experts attending the 1961 Eastern Joint
Computer Conference held in Washington, D.C.
The use of the computer in the classroom encompasses not only colleges
and high schools, but extends even to prisons. Twenty inmates of a
Pennsylvania state institution attended a pilot program teaching
computer techniques with a UNIVAC machine.
[Illustration:
_Datamation_
Seventeen-year-old David Malin who presented a paper on computers at
the Eastern Joint Computer Conference in 1961.
]
The United States is not alone in placing importance on the computer in
schools. Our Department of Commerce has published details of Russian
work in this direction, noting that it began in 1955 and places high
priority on the training of specialists in computer research, machine
translation, automation, and so on. The Department of Commerce feels
that these courses, taught at the graduate, undergraduate, and even
high-school level, are of high quality.
_Teaching Machines_
Thus far we have talked of the computer only as a tool to be studied and
not as an aid to learning in itself. In just a few years, however, the
“teaching machine” has become familiar in the press and controversial
from a number of standpoints, including those of being a “dehumanizer”
of the process of teaching and a threat to the apple business!
Actually, the computer has functioned for some time outside the
classroom as a teaching machine. Early applications of analog computers
as flight simulators were true “teaching machines” although perhaps the
act was not as obvious as classroom use of a computer to teach the three
R’s. Even today, there are those who insist that such use of the
computer by the military or industry offers more potential than an
academic teaching machine. Assembly workers have been taught by
programmed audiovisual machines such as Hughes Aircraft’s Videosonic
trainer, and the government has taught many technicians by computer
techniques. A shrewd observer, however, noting that the computer is
called stupid, bluntly points out that any untaught student is in the
same category, and that perhaps it takes one to teach one.
A strong motivation for looking to the machine as a public teaching tool
is the desperation occasioned by the growing shortage of teachers. If
the teaching machine could take over even some of the more simple chores
of the classroom, early advocates said, it would be worth the effort.
Formal study of machine methods of teaching have a history of forty
years or more. In the 20’s, Sydney Pressey designed and built automatic
teaching—or more precisely, testing—machines at Ohio State University.
These were simply multiple-choice questions so mechanized as to be
answered by the push of a button rather than with a pencil mark. A right
answer advanced the machine to the next question, while an error
required the student to try again. Pressey wisely realized the value in
his machines; the student could proceed at his own pace, and his
learning was also stimulated by immediate recognition of achievement. To
further enforce this learning, some of the teaching machines dispensed
candy for a correct answer. Using this criterion, it would seem that
brighter students could be recognized by their weight.
Unfortunately, Pressey’s teaching machines did not make a very big
splash in the academic world, because of a combination of factors. The
machines themselves had limitations in that they did not present
material to be learned but were more of the nature of _a posteriori_
testing devices. Too, educators were loath to adopt the mechanized
teachers for a variety of reasons, including skepticism, inertia,
economics, and others. However, machine scoring of multiple-choice tests
marked with special current-conducting pencils became commonplace.
Another researcher, B. F. Skinner, commenced work on a different kind of
teaching machine thirty years ago at Harvard. Basically his method
consists of giving the subject small bits—not computer “bits,” but the
coincidence is interesting—of learning at a time, and reinforcing these
bits strongly and immediately. Skinner insists that actual “recall” of
information is more important than multiple-choice “recognition,” and he
asks for an answer rather than a choice. Called “operant reinforcement,”
the technique has been used not only on man, but on apes, monkeys, rats,
dogs, and surprisingly, pigeons.
During World War II, Dr. Skinner conducted “Project Pigeon” for the
military. In this unusual training program, the feathered students were
taught to peck at certain targets in return for which they received food
as a reward. This combination of apt pupils and advanced teaching
methods produced pigeons who could play ping-pong. This was in the early
days of missile guidance, and the pigeons next went into training as a
homing system for these new weapons! To make guidance more reliable, not
one but three pigeons were to be carried in the nose of the device.
Lenses in the missile projected an image before each pigeon, who
dutifully pecked at his “target.” If the target was in the center of the
cross hairs, the missile would continue on its course; if off to one
side, the pecking would actuate corrective maneuvers. As Project
“Orcon,” for Organic Control, this work was carried on for some time
after the end of the war. Fortunately for the birds, however, more
sophisticated, inorganic guidance systems were developed.
The implications of the pigeon studies in time led to a new teaching
method for human beings. Shortly after Skinner released a paper on his
work in operant reinforcement with the pigeons, many workers in the
teaching field began to move in this direction. For several years
Skinner and James Holland have been using machines of this type to teach
some sections of a course in human behavior to students at Radcliffe and
Harvard. Rheem Califone manufactures the DIDAK machine to Skinner’s
specifications.
To the reasons advanced by those who see teacher shortages looming,
Skinner adds the argument that a machine can often teach better. Too
much time, he feels, has been spent on details that are not basic to the
problem. Better salaries for teachers, more teachers, and more schools
do not in themselves improve the actual teaching. Operant reinforcement,
Skinner contends, _does_ get at the root of the problem and, in addition
to relieving the teacher of a heavy burden, the teaching machine
achieves better results in some phases of teaching. It also solves
another problem that plagues the educator today. It is well known that
not all of us can learn at the same rate. Since it is economically and
culturally impossible except in rare cases to teach children in groups
of equal ability, a compromise speed must be established. This is fine
for the “average” child, of whom there may actually be none in the
classroom; it penalizes the fast student, and the slow student perhaps
even more. The teaching machine, its proponents feel, takes care of this
difficulty and lets each proceed at his own rate. Since speed in itself
is no sure indicator of intelligence, the slow child, left to learn as
he can, may reach heights not before dreamed possible for him.
Many educators agree that automated teaching is past due. James D. Finn,
Professor of Education at the University of Southern California,
deplores the lack of modern technology in teaching. “Technology during
the period from 1900 to 1950 only washed lightly on the shores of
instruction,” he says. “The cake of custom proved to be too tough and
the mass production state, at least 100 years behind industry, was not
entered except here and there on little isolated islands.”
[Illustration:
_Educational Science Division,_
_U.S. Industries, Inc._
AutoTutor teaching machine has programs for teaching many subjects.
]
These little isolated islands are now getting bigger and closer
together. The Air Force has for some time trained technicians at Keesler
Field with U.S. Industries AutoTutor machines, and also uses them at the
Wright Air Development Center. The Post Office Department has purchased
fifty-five U.S. Industries’ Digiflex trainers. Following this lead,
public education is beginning to use teaching machines. San Francisco
has an electronic computer version that not only teaches, tests, and
coaches, but even sounds an alarm if the student tries to “goof off” on
any of the problems. The designers of the machine selected a sure-fire
intellectual acronym, PLATO, for Programmed Logic for Automatic Teaching
Operations. The System Development Corporation, the operations firm that
designed the SAGE computer, calls its computer-controlled classroom
teacher simply CLASS. This machine uses a Bendix G-15 computer to teach
twenty youngsters at a time.
To show the awareness of the publishers of texts and other educational
material, firms like Book of Knowledge, Encyclopedia Britannica Films,
and TMI Grolier are in the “teaching machine” business, and the
McGraw-Hill Book Company and Thompson Ramo Wooldridge, Inc., have teamed
to produce computerized teaching machines and the programs for them.
Other publishers using “programming” techniques in their books include
Harcourt-Brace with its 2600 series (for 2,600 programmed steps the
student must negotiate), Prentice-Hall, and D. C. Heath. Entirely new
firms like Learning, Incorporated, are now producing “programs” on many
subjects for teaching machines.
Subjects available in teaching machine form include algebra,
mathematics, trigonometry, slide rule fundamentals, electronics,
calculus, analytical geometry, plane geometry, probability theory,
electricity, Russian, German, Spanish, Hebrew, spelling, music
fundamentals, management science, and even Goren’s bridge for beginners.
While many of these teaching machines are simply textbooks programmed
for faster learning, the conversion of such material into
computer-handled presentation is merely one of economics. For example, a
Doubleday TutorText book costs only a few dollars; an automatic
AutoTutor Mark II costs $1,250 because of its complex searching facility
that requires several thousand branching responses. However, the
AutoTutor is faster and more effective and will operate twenty-four
hours a day if necessary. With sufficient demand the machine may be the
cheaper in the long run.
The System Development Corporation feels that its general concept of
automated group education will be feasible in the near future despite
the high cost of advanced electronic digital computers. It cites pilot
studies being conducted by the State of California on data-processing
for a number of schools through a central facility. Using this same
approach, a single central computer could serve several schools with
auxiliary lower-priced equipment. Even a moderately large computer used
in this way could teach a thousand or more students simultaneously and
_individually_, the Corporation feels. After school hours, the computer
can handle administrative tasks.
[Illustration:
_System Development Corp._
The CLASS facility incorporates an administrative area, hallway,
combined observation and counseling area, and a large classroom area
divided by a folding wall.
]
In the CLASS system developed by the System Development Corporation, the
“branching” concept is used. In a typical lesson program, if the student
immediately answers that America was discovered by Christopher Columbus,
he will be told he is correct and will then be branched to the next
item. If he answers Leif Ericson, the computer takes time out to
enlighten the pupil on that score. Next, it reinforces the correct date
in the student’s mind before asking another question. Although it would
seem that a lucky student could progress through the programmed lesson
on guesswork alone, the inexorable laws of probability rule this out. He
cannot complete the lesson until he has soaked up all the information it
is intended to impart. He can do this without an error, in a very short
time, or he can learn by the trial-and-error process, whichever is
better suited to his speed and mental ability.
Making up the program for the teaching machine is a difficult task and
requires the services of technical expert, psychologist, and programmer.
An English-like language is used in preparing a CLASS program for the
computer. Put on magnetic tape, the program goes into the memory of the
computer and is called out by proper responses from the student as he
progresses through the lesson.
[Illustration:
_System Development Corp._
Students in CLASS are learning French in a group mode of automated
instruction.
]
Complex as the programming is, entries from the student’s control are
processed into the computer in about one-tenth of a second, and an
answer is flashed back in about the same amount of time. Remember that
the CLASS computer is handling twenty students at a time, and that in
addition to teaching it is keeping a complete record of how the student
fared at each step of the lesson.
It is obvious that the binary or yes-no logic of the computer ties in
with the concept put forth by Skinner and others of presenting small
bits of information at a time. We can use the game of 20 Questions as a
good analogy. Even getting only simple yes-no answers, skilled players
can elicit an amazing amount of information in often far less than the
permitted number of questions. Thus even complex subjects can be broken
down into simple questions answerable by discrete choices from the
student.
The automated group education system of the System Development
Corporation is made up of the following components: a digital computer
to control and select the material presented and to analyze responses, a
magnetic tape storage unit, a typewriter for printing out data analysis,
a slide projector and screen for presenting educational materials, and
individual desks with keyboards for the students’ responses.
We have pointed out that even though it is possible to break down
educational material into multiple-choice or yes-no answers to which are
assigned intrinsic values, the ideal system permits answers on a linear
scale. In other words, instead of picking what he considers the most
nearly correct, a student writes his own answer. Some experts feel that
the advances being made in optical scanning, or “reading” techniques for
computers, will result in linear programming of the teaching machines
within the next ten years. Such a development will do much to alleviate
the complaint that the machine exerts a rigid mechanizing effect on the
teaching process.
While fear of displacement motivates some teachers to distrust the
machine, an honest belief that the human touch is necessary in the
schoolroom is also a large factor against acceptance. Yet these same
wary teachers generally use flash cards, flip charts, and other
mechanical aids with no qualms. The electronic computer is a logical
extension of audiovisual techniques, and in time the teacher will come
to accept it for what it is.
The human teacher will continue to be an indispensable element in
education, but he must recognize that as our technology becomes more
complex he will need more and more help. In 1960 there were about 44
million students in our classrooms, and about 135,000 too few teachers.
By 1965 it is estimated there will be 48 million students and 250,000
teachers fewer than we need. Parallel with this development is the
rapidly growing need for college graduates. One large industrial firm
which employs 150,000 hires only 300 college graduates a year at
present, but will need 7,000 when it automates its plants. The pressure
of need thus is forcing our educational system to make use of the most
efficient means of educating our students.
Beyond simply taking its place with other aids, however, the computer
will make great changes in our basic concepts of teaching, according to
Dr. Skinner. He asks the question “Are the students who learn in spite
of a confusing presentation of a subject better for the experience, or
were they better students at the outset?” He advances this argument to
say that perhaps “easy” learning is actually the best; that we would do
well to analyze the behavior called thinking and then produce it
according to these specifications. The traditional teacher finds the
prospect alarming and questions the soundness of minimizing failure and
maximizing success.
There is not yet definite agreement by other psychologists with
Skinner’s contention that recall rather than recognition is the desired
method. Neither is it sure that the negative reinforcement of a number
of incorrect choices may result in remembering wrong answers. And of
course the division between rote learning and creativity is an important
consideration. The answers may well lie in the computer, which when
properly programmed is about the most logical device we have available
to us. Thus the machine may determine the best teaching methods and then
use them to teach us. Regardless of these as yet unanswered questions,
however, the future of the teaching machine seems to be assured. One
authority has predicted that it will be a $100 million market by 1965.
An intriguing use of computer techniques in teaching is being
investigated by Corrigan Communications, which scores students answering
questions on telecourses. This work is being done with a course in
medicine, and with the rapid growth of educational television the
implications of combining it and teaching machine techniques are of
great importance.