8th Conference of the European Society for Fuzzy Logic and Technology (EUSFLAT 2013)

The webbed emergence of fuzzy sets and computer

science education from electrical engineering
Rudolf Seising1 Marco Elio Tabacchi2
1
European Center for Soft Computing, Mieres, Asturias (Spain)
2
DMI Università degli Studi di Palermo and Istituto Nazionale di Ricerche Demopolis (Italy)

Abstract In the present contribution our aim is to show that we

are dealing with fuzzy concepts in the area of science
Historically, Computer science emerged from electrical and technology. The example employed in this paper is
engineering and from mathematics in the 1960s. From the scientific discipline of computer science (CS) that
the content of some unpublished documents and also emerged from Electrical engineering (EE) in the 1960s.
some rather less-well-known papers by Lotfi A. Zadeh In section II we will sketch some aspects of that histori-
it is argued that the emergences of Computer science cal episode from the viewpoint of the electrical engi-
and Fuzzy Set Theory have been interlinked. Zadeh’s neer Lotfi A. Zadeh who was on one hand a decisive
task as Chair of the Electrical Engineering Department protagonist of this development at the University of
in Berkeley in the 1960s, his activities in Education of California, Berkeley and who founded on the other
Engineering and his creation of the theory of Fuzzy sets hand the theory of Fuzzy Sets (FS) during that period.
generated his view on the scientific discipline of Com- From the content in non-published or not well known
puter science as a fuzzy set. This view could establish a documents, found by one of the authors in Zadeh’s pri-
new approach to history and philosophy of science. vate archives in his office and at his home follows the
claim that the two developments in science, the birth of
Keywords: Electrical engineering, Fuzzy sets CS and FS, resp. have been interlinked, viz.:

1. Introduction • The theory of FS emerged from scientific studies in

Circuit theory, Network theory, System theory, and
Science is the general analysis of nature and that is Information theory as parts of EE, as well as from
what in ancient times was called natural philosophy. mathematical theories and logic, after computers
Later on, natural philosophy differentiated into the dis- entered the field of technology.
ciplines of physics (ancient Greek: φύσις (physis) “na- • We can treat the stemming of CS from EE and oth-
ture”) (including astronomy), chemistry, biology and er scientific theories and methodologies as a fuzzy
some parts of mathematics. However, there are no strict relationship because the subjects of CS – some of
boundaries between these disciplines and there are them are also subjects of EE or mathematics or log-
overlaps: e.g., biophysics, quantum chemistry etc. ic etc. and some of them newly created – became
Furthermore today the main branches of physics part of the new scientific discipline CS only to
are much more determined and differentiated e.g., as some degree, as such that CS can be considered it-
(Classical) Mechanics, Electromagnetism, Thermody- self as a fuzzy set!
namics and Statistical mechanics, Relativity theory and
Quantum mechanics. However, again we must notice
Section III deals with this fuzziness of CS in the 1960s.
that we cannot sharply separate these elaborated parts
In Section IV we give a conclusion of this historical re-
of physics from each other. Sometimes, they have con-
search and in Section V a short outlook considers the
cepts in common, e.g., particles, mass points, waves, or
fuzziness of scientific disciplines in general.
they use the same rules and methods, some theories are
specialization of more general theories, e.g., Hook’s
mechanics and Newton’s mechanics etc. 2. Electrical engineering and the computer era
In addition, during the 19th century, scientist and
Computers have been the most famous technical prod-
engineers established relationships between physics and
uct of Second World War research ̶ along with the
engineering. Before that time, electricity was consid-
ered as part of physics but then, e.g., in 1882, the first atomic bomb. Computers became popular as “electronic
chair of Electrical engineering (EE) was founded at the brains” or “thinking machines”. Even if the “era of
University of Technology (TU) in Darmstadt (Germa- computers” was already started by the analogue MIT
ny) and in the same year the MIT in Boston, USA, of- Differential Analyzer of Vannevar Bush (1890-1974),
the technological development of digital computers was
fered he first option of EE within a physics department.
One year later TU Darmstadt and also Cornell Universi- an eminent push with ENIAC (Electronic Numerical
ty introduced the world's first courses of study in EE Integrator and Computer) and EDVAC (Electronic
and in 1885 the University College London founded the Discrete Variable Computer), both designed by John
first chair of electrical engineering in the UK. One year Presper Eckert (1919-1995) and John William Mauchly
later the University of Missouri established the first de- (1907-1980).
partment of EE in the USA.

© 2013. The authors - Published by Atlantis Press 432

2.1. Computers – a new field in electrical engineer- relationship of the system states that the output is just a
ing function of the input.
Zadeh believed that it was only a matter of time be-
As early as 1950 the British mathematician Alan fore system theory would attain acceptance. It turns out
Mathison Turing (1912-1954) asked in his famous that he was right: Eight years later, he wrote the article
Mind-article “Computing Machinery and Intelligence” “From Circuit Theory to System Theory” [4] for the
[1] “Can machines think?”. Instead of an answer he anniversary edition of the Proceedings of the IRE ap-
proposed a variation of the then well-known imitation peared in May 1962 to mark the 50th year of the Insti-
game, lately renamed the “Turing test”, that highlighted tute of Radio Engineers (IRE). He could describe prob-
a philosophical interest in the problem of deciding lems and applications of system theory and its relations
whether a computer or a program could think like a to network theory, control theory, and information theo-
human being or not. In those days Lotfi Aliasker Zadeh ry: “It has been brought about, largely within the past
(born 1921) was a young electrical engineer with deep two decades, by the great progress in our understanding
interest in the new-by-then computing machines. He of the behavior of both inanimate and animate systems-
wrote a paper “Thinking Machines. A New Field in progress which resulted on the one hand from a vast
Electrical Engineering”, which appeared in The Colum- expansion in the scientific and technological activities
bia Engineering Quarterly in the same year [2]. Here, directed toward the development of highly complex
Zadeh put for discussion questions “How will «elec- systems for such purposes as automatic control, pattern
tronic brains» or «thinking machines» affect our way of recognition, data-processing, communication, and ma-
living?” and “What is the role played by electrical engi- chine computation, and, on the other hand, by attempts
neers in the design of these devices?” ([2], p. 12.) He at quantitative analyses of the extremely complex ani-
was looking for “the principles and organization of ma- mate and man-machine systems which are encountered
chines which behave like a human brain. Such ma- in biology, neurophysiology, econometrics, operations
chines were then variously referred to as “thinking ma- research and other fields” ([4], p. 856f.).
chines”, “electronic brains”, “thinking robots”, and oth- After 1959, when Zadeh became a professor of elec-
er similar names. trical engineering at the University of California,
Berkeley, he published papers on system theory and
two well-known books with colleagues at his EE de-
partment [5, 6].

2.3. Electrical filtering

Already during his time in New York, when he joined

the faculty of Columbia University as an instructor and
since 1950 he was appointed assistant professor, he was
excited by Cybernetics [7] of Norbert Wiener (1894-
Figure 1: Turing, Zadeh, Ragazzini.
1964) and he was interested in the theory of ideal and
optimal filtering. Together with his supervisor John
The fundamental principles of thinking machines, Ralph Ragazzini (1912-1988) he published in 1950 “An
Zadeh stressed, were developed by mathematicians, but Extension of Wiener’s Theory of Prediction” [8], an
the physical realization, the construction of the thinking important milestone in the development of network syn-
machines, was the task of electrical engineers, who de- thesis.
sign and build the memory chips, processors, The mathematical techniques of the theory of predic-
“computors”, decision makers, etc. Until now, if elec- tion and filtering have been commonly employed in
trical engineers had come into contact at all with such mathematical physics, particularly in quantum mechan-
heretofore far-removed subjects as Boolean algebra, ics, e.g., multi-dimensional Euclidean spaces and Hil-
polyvalent logic, etc., it was through so far off when it bert space representation. In his many papers on predic-
would be just as important for post-graduate electrical tion and filtering, linear and nonlinear systems, time-
engineers to take classes in mathematical logic as clas- variying networks, etc. in the first 1950s, Zadeh used
ses on complex variables: “Time marches on.” ([2], p. the mathematical calculus of functionals and operators.
31). The analogy between projection in a function space and
filtration with an ideal filter led Zadeh in the early
1950s to a functional symbolism of filters [9]. Follow-
2.2. From circuits to systems ing Wiener’s work and continuing his collaboration
with his supervisor Ragazzini Zadeh used the concept
In 1954 Zadeh published, again in the Columbia Engi- of optimal filters instead of ideal filters. Ideal filters are
neering Quarterly, the article “System Theory” [3] defined as filters which achieve a perfect separation of
where he characterized systems as a “black boxes” with signal and noise, but in reality no such filter exists. He
inputs and outputs, and asserts that if these inputs and knew from experience that characteristics of electrical
outputs are describable as time dependent functions filters don’t show an exact step at the limiting frequen-
then the dynamic behaviour of the system can be stud- cy but smooth functions. Zadeh discussed optimal fil-
ied mathematically, and as such the input-output- ters that give the “best approximation” of a signal and

433
he noticed that “best approximations” depend on rea- though it was rare for requests by scientists who were
sonable criteria. In that time he formulated these criteria neither mathematicians nor theoretical physicists or his-
in statistical terms, but during the course of his work on torians to receive a positive response [12]. Zadeh ini-
optimum filters in the mid-50s he turned away from sta- tially took a half-year sabbatical from Columbia Uni-
tistical methods, and recognized that a more promising versity in 1956. He wanted to learn more about logic, an
approach was that of finding an optimum filter relative interest he had cultivated since 1950, when he predicted
to a distance to be minimized in the function space of that logic, and particularly multi-valued logic, would
the signals. become increasingly more important to the problems of
At the time Zadeh was working constructively to electrical engineering in the future. [19]
bridge the gap between theory and practice; he was
however forced to recognize that these attempts would
not be successful: “As a mathematically oriented sys-
tem theorist, I had been conditioned to believe that the
analytical tools based on set theory and two-valued log-
ic are all that is needed to build a framework for a pre-
cise, rigorous and effective body of techniques for the
analysis of almost any kind of man-made or natural sys-
tem. Then, in 1961-1963, in the course of writing a
book on system theory (with C. A. Desoer), I began to
feel that complex systems cannot be dealt with effec-
tively by the use of conventional approaches largely be- Figure 3: Montgomery, Kleene, Robbins.
cause the description languages based on classical
mathematics are not sufficiently expressive to serve as a The Princeton ambiance quickly inspired Zadeh who
means of characterization of input-output relations in an remarked it being the “Mecca for mathematicians” [12].
environment of imprecision, uncertainty and incom- He attended lectures by Stephen Kleene (1909-1994)
pleteness of information.” [10] (Figure 3, mid), who had also continued developing the
multi-valued logic devised by the Polish school of log-
ic. Kleene became Zadeh’s friend and mentor at Prince-
ton: “Steven Kleene was my teacher in logic. Yes, I
learned logic from Steven Kleene!” [10].
This residency had revealed to Zadeh some com-
pletely new perspectives of scientific life and work.
New ways of thinking had come from the mathematics
philosophers in Princeton, and thanks to them he had
learned new mathematical methods from statistics,
game and decision theory. He also experienced new
views of system theory and the newly established au-
Figure 2:Shannon, Desoer, Wiener. tomata theory. He had apparently become familiar with
Automata Studies, published during this period by
There were two ways of overcoming this situation. In Claude Elwood Shannon (1916-2001) and John McCar-
order to describe the actual systems appropriately, he thy (1927-2011). All of the knowledge, impetus and
could try to increase the mathematical precision even impressions he found at the IAS would have a lasting
further, but this course of action led nowhere. The other effect on Zadeh’s future endeavors!
way presented itself in the year 1964, when Zadeh dis- Zadeh found multi-valued logic to be a natural gen-
covered how he could describe real systems as they ap- eralization of the conventional logic of just two values
peared to people. “I’m always sort of gravitated toward into n values, similar to the leap from two-
something that would be closer to the real world” [11]. dimensionality to n-dimensionality in mathematics. [12]
The “closer to the real world” thing was the Theory of He was now also toying with the idea of introducing
Fuzzy Sets (FS)! multi-valued logic into automata theory and implement-
ing it in electric circuits, and once he had returned to
2.4. Mathematics and logics Columbia University in New York he assigned two dis-
sertations that dealt with the subjects of multi-valued
Zadeh’s resignation with usual mathematics would soon logic in the design of transistor circuits and with multi-
lead to capitulation. He was nearing a crossroads. Her- valued coding:
bert E. Robbins (1915-2001) (Figure 3, right) was the
chairman of Columbia University’s department of • Oscar Lowenschuss wrote the dissertation
mathematical statistics at the time. He was a good Multi-Valued Logic and Sequential Machines
friend of Zadeh as well as of Deane Montgomery or Non-Binary Switching Theory the following
(1909-2002) (Figure 3, left), a member of the Institute year [13]. Parts of this paper had been pub-
for Advanced Study (IAS) in Princeton. Robbins and lished previously [14]. See also the later publi-
Montgomery campaigned for the approval of the IAS cation [15];
guest residency for which Zadeh had applied, even

434
 • Werner Ulrich managed to finish his disserta- Institutes or the Institute for Automatics and
tion Nonbinary Error Correction Codes in Telemechanics in the Soviet Union did not necessarily
1957 [16]. have to serve as models for the institute he hoped to es-
tablish. Rather it would be designed to meet the specific
“That’s why I wanted to know about logics!” Zadeh re- needs and interests of workers in the fields of infor-
called when interview by Seising in the year 2000 [10]. mation theory, communication theory, system theory,
control theory, automata, biological systems, computa-
tion, machine translation of languages and related
fields. It would be concerned with both theoretical and
experimental research in this area [17].

Figure 5: Zadeh’s editorial in 1963, [17], excerpt.

After Zadeh had also joined the editorial board of the

IRE Transactions on Automatic Control in 1962, he
wrote “A Critical View of Our Research in Automatic
Control” for the April edition, an article in which he
repeated his call for the founding of such an institute.
Here he suggested establishing an Institute for Control
Science and Engineering, then he added: “(or, more
broadly, an Institute for Research in Information Sci-
ences) which would serve as a focal point on the na-
tional level for research in control theory and its appli-
cations as well as in such related fields as system theo-
ry, information and communication theories, circuit
Figure 4: First pages of the Ph D dissertation thesis theory, machine computation, automata theory, bionics,
by Ulrich [16] and Lowenschuss [13] in 1957 and 1958. etc.” ([18], p. 74).
The many new impressions, encounters, discussions
2.5. Information and communication and insights in Princeton had steered Zadeh’s future
scientific work in new directions.
The value to that scientific community which Zadeh
recognized in an institute like IAS in Princeton can also 2.6. Fuzzy sets
be seen in his dedication to establishing a similar insti-
tute for his own scientific community. Since the late As we have seen before, Zadeh had extensively criti-
1940s he was enthused over Shannon’s “Mathematical cized the relationship between mathematics and his own
Theory of Communication” (later called for short In- scientific-technical EE discipline. The tool offered by
formation theory) [13], and when he once again com- mathematics was not appropriate to the problems that
posed an editorial for the IRE Transactions on Infor- needed to be handled in the engineering sciences. In-
mation Theory in 1960, it was entitled “Toward an In- formation and communications technology had led to
stitute for Research in Communication Science” [17] the construction and design of systems that were so
(Figure 5). He had in the meantime become a professor complex that it took much more effort to measure and
at Berkeley and he apparently sometimes longed for the analyze these systems than had been the case just a few
freedom he had enjoyed as a guest scientist in Prince- years before. Much more exact methods were now re-
ton. He called for an Institute for Communication Sci- quired to identify, classify or characterize such systems
ence to be founded, where scientists could spend a year or to evaluate and compare them in terms of their per-
or two concentrating exclusively on their research with- formance or adaptivity.
out being distracted by teaching and administrative du- In order to provide a mathematically exact expression
ties, contract negotiations and doctoral advising. This of experimental research with real systems, it was nec-
was the only way to guarantee a free choice of research essary to employ meticulous case differentiations, dif-
topics and scientifically communicative exchange with ferentiated terminology and definitions that were
no outside pressure! adapted to the actual circumstances, things for which
A number of well-known institutions both in the the language normally used in mathematics could not
United States and abroad have these characteristics, but account. The circumstances observed in reality could no
they are embodied perhaps in their purist form in the longer simply be described using the available mathe-
IAS, which since its inception in 1930 had played a matical means.
very significant role in the development of mathematics In the summer of 1964 Zadeh was thinking about pat-
in their country [17]. The IAS, Germany’s Max-Planck- tern recognition problems and grades of membership of

435
an object to be an element of a class; almost 50 years 2.7. From the department of electrical engineering
later, he returns with the mind to these times as such: (EE) to the department of electrical engineer-
“While I was serving as Chair, I continued to do a lot of ing and computer sciences (EECS)
thinking about basic issues in systems analysis, espe-
cially the issue of unsharpness of class boundaries. In
July 1964, I was attending a conference in New York In one of Seising’s interviews Zadeh recalled: “System
and was staying at the home of my parents. They were Theory came grown up but then computers and com-
away. I had a dinner engagement but it had to be can- puters then took over. In other words: the center of at-
celed. I was alone in the apartment. My thoughts turned tention shifted. ... So, before that, there were some uni-
to the unsharpness of class boundaries. It was at that versities that started departments of system sciences,
point that the simple concept of a fuzzy set occurred to departments of system engineering, something like that,
me. It did not take me long to put my thoughts together but then they all went down. They all went down be-
and write a paper on the subject. This was the genesis cause computer science took over.” [11]
of fuzzy set theory.” ([19], p. 7).1
Zadeh submitted his first article “Fuzzy Sets” to the
editors of Information and Control in November 1964
and it appeared in this journal in the following June
[20]. He introduced new mathematical entities as clas-
ses or sets that ‘‘are not classes or sets in the usual
sense of these terms, since they do not dichotomize all
objects into those that belong to the class and those that
do not.’’ He introduced ‘‘the concept of a fuzzy set, that
is a class in which there may be a continuous infinity of
grades of membership, with the grade of membership of
an object x in a fuzzy set A represented by a number
fA(x) in the interval [0,1].’’ [21]
The question was how to generalize various con-
cepts, union of sets, intersection of sets, and so forth.
Zadeh defined equality, containment, complementation,
intersection and union (Figure 6) relating to fuzzy sets
A, B in any universe of discourse X as follows (for all x
∈ X):

• A = B if and only if µA(x) = µB(x), Figure 7: Zadeh became chair in 1963, (excerpt).
• A ⊆B if and only if µA(x) ≤ µB(x),
• ¬A is the complement of A, When Zadeh become chairman of the department of
if and only if µ¬A(x) = 1- µA(x), electrical engineering (EE) at Berkeley in 1963 (Figure
• A ∪ B if and only if µA∪B(x) = max (µA(x), 7), he experienced such shifts very intensively, for it
µB(x)), was during his five-year tenure in this position that his
department was renamed the Department of Electrical
• A ∩ B if and only if µA∩B(x) = min (µA(x),
Engineering and Computer Science (EECS) [11]. In
µB(x)).
November 1992 Zadeh was asked to give an after din-
ner talk on the history of CS at Berkeley. In the manu-
script for this talk he wrote: “I joined the EE Depart-
ment in 1959. At that time, the EE Department was best
known for its work in electromagnetism and micro-
waves. A decision was made to build up the area of cir-
cuits, systems and control. Don Pederson, Ernie Kuh
and Charles Desoer came from Bell Labs. Eli Jury, Art
Bergen and I came from Columbia University. […]
There wasn’t much activity in the computer field at that
time, but there was was significant. […] There was a
Computer Centre in Cory Hall that was run by the EE
Figure 6: Zadeh’s Illustration of fuzzy sets in R1: “The membership Department. The principal and only figures in computer
function of the union is comprised of curve segments 1 and 2; that of
the intersection is comprised of segments 3 and 4 (heavy lines).” science and engineering in EE at that time were Paul
([21], p. 342). Morton and Harry Huskey. They can be rightly regard-
ed as the progenitors of Computer science and engi-
neering at Berkeley.” […] When I was appointed as
Chairman in 1963, I was not a computer person and I
am not a computer user to this day, I regret to say. But I
1
was always a very strong believer in the importance of
A detailed presentation of the history of the theory of computers and digital technology. My first action as
FS is given in the one of the authors’ book [20].

436
Chairman was to send a memo to the faculty in which I Berkeley’s “solution” is that computer science is not a
suggested that we assign the highest priority to the de- homogeneous and unified field – at least not at this time
velopment of computer science in EE. But what it was – and that, in paraphrased words of Professor A.
obvious today was not so obvious then. The reaction to Oettinger of Harvard, «… it has some components
my memo was mixed and some influential faculty which are the purest of mathematics and some that are
members objected strongly to my proposal.” [22] the dirtiest of engineering.» [25] This split personality
After some initial difficulties, Zadeh was finally suc- of computer science makes it very difficult to create a
cessful in changing the name of the department to single academic unit within the university structure
EECS. In his article “Electrical Engineering at the where mathematically oriented automata theorists, for-
Crossroads” published in 1965 [23] (Figure 6) he de- mal language experts, numerical analysts and logicians
scribed the problem as such: “The slowness with which could establish a comfortable modus vivendi with non-
many electrical engineering departments have reacted mathematical oriented hardware designers, systems
to the rapidly growing demand for computer scientists programmers and computer architects. […] In essence,
and engineers, and their unwillingness to make substan- the Berkeley “solution” provides a partial answer to the
tive changes in their curricula to meet the need for spe- dilemma by dividing computer science not into two
cialized training in computer sciences and related fields, non-overlapping parts but into two overlapping parts
is generating strong pressures on some campuses to es- which differ from one another mainly in degrees of em-
tablish separate computer science departments.” ([23], phasis places on various subject areas.” ([29], p. E164f)
p. 30) In the same article he presented the new EE cur-
riculum at Berkeley that “reflects the fact that, today,
electrical engineering is no longer an aggregation of a
small number of subject areas sharing a large common
body of concepts and techniques – as it was in the thir-
ties, forties, ant to a lesser extent, in the fifties. Rather,
it is an assemblage of a wide range of subjects, falling
into three major areas which have a relatively small
common core. […] If this premise is accepted, then the
only logical conclusion is that the student must be pro-
vided with a choice of several basic programs, which
could permit him to focus his studies in one of the ma-
jor areas falling within or nearest to his main field of
interest.” ([23], p. 31)

Figure 8: Excerpt of Zadeh’s article in1965, [23].

At Berkeley three programs (A: Electronics, Fields,

and Plasmas; B: Systems, Information, and Control and
C: Computer Sciences) and a “General EE program” D
were established ([23], p. 31f).
Three years later Zadeh gave a talk on “Education in
Computer Science” at Israel’s 4th National Conference
on Data Processing that took place in the Hebrew Uni-
versity Jerusalem (Figure 8). Initially, he claimed that
computer science “as a collection of concepts and tech-
niques which serve to systematize the employment of
the means with which modern technology provides us
for purposes of stage, representation and processing of
information.” ([24], p. E157) In the printed version of
this talk he affirmed that “Computer science cuts across
Figure 9: Reprint of Zadeh’s paper in1968, [24].
the boundaries of many established fields. It is glamor-
ous; it draws a large number of students – many of them In his article “Computer Science as a Discipline”,
from other departments; it is hitched to the bandwagon that appeared in the same month in the Journal of Engi-
of computers and the information revolution.” ([49], p. neering Education [26] (fig. 10), he also characterized
E158) CS this way. He wrote what was “pointed out in the au-
In the same paper he discussed some details of thoritative reports of the ACM Curriculum Committee
“Berkeley’s solution” to the CS-problem that bred final- on Computer Science [27, 28] computer science is not
ly a Department of CS in the College of Letters and simply concerned with the design of computing devic-
Science and a program in CS in the College of Engi- es, nor is it just the art of numerical calculation. Essen-
neering within the Department of EECS. By way of ex- tially, computer science is concerned with information
planation he added: “Essentially, the main premise of

437
in much the same way that physics is concerned with Under the heading “Containment Table for Computer
energy; it is devoted to the representation, storage, ma- Science” (Figure.19) he arranged the most relevant
nipulation and presentation of information in an envi- “subjects in question and their degrees of containment
ronment permitting automatic information systems.” in computer science”. Zadeh explained: “Clearly, such
([26], p. 913) numerical values of degrees of containment represent
merely this writer’s subjective assessment, expressed in
quantitative terms, of the current consensus regarding
the degrees of inclusion of various subjects in computer
science.” ([26], p. 913.) He also emphasized “that a
high degree of containment of a particular subject in
computer science does not imply that it cannot have a
Figure 10: Excerpt of Zadeh’s article [26], 1968. similar high or even higher grade of containment in
some other field. For example «automata theory» has
Concerning his view that CS “cuts across the bound- the degree of containment of 0.8 in computer science; it
aries of many established fields” and also that the parts also has the same, or nearly the same, degree of con-
of CS differ from one another “in degrees of empha- tainment in system theory. Also, the subjects listed in
sis”, Zadeh then presented a new idea in the field of CS the table may have substantial overlaps with one anoth-
education: because of the broadness and vagueness of er. This is true, for example, of «Automata theory» and
the statement “to convey even a rough idea of the «finite state systems».” ([26], p. 914.)
boundaries of computer science and its relation to
mathematics, electrical engineering and other neighbor-
ing fields”, he introduced his idea of fuzzy sets, to em- 3. Outlook
ploy a new approach: “Specifically, let us regard com-
The historical aspects described in this paper are only
puter science as a name for a fuzzy set of subjects and
hints of the complete history of relationships between
attempt to concretize its meaning by associating with Fuzzy Sets, Computer Science, Electrical Engineering
various subjects their respective degrees of containment
and the other branches of science. This work will be
(ranging from 0 to 1) in the fuzzy set of computer sci-
continued by generalization of Zadeh’s “fuzzy concep-
ence. For example, a subject such as «programming
tion” of Computer science to other scientific disci-
languages» which plays a central role in computer sci-
plines. In the hopes of the authors, this should breed to
ence, will have a degree if containment equal to unity.
a new concept in philosophy of science.
On the other hand, a peripheral subject such as «math-
ematical logic» will have a degree of containment of,
say, 0.6.” ([26], p. 913.) Acknowledgment

We would like to thank Lotfi A. Zadeh for his support

to this work.

References

[1] A.M. Turing, Computing Machinery and Intelli-

gence. In: Mind, vol. LIX, 1950, pp. 433-460.
[2] L.A. Zadeh, Thinking machines - a new field in
electrical engineering, Columbia Engineering Quar-
terly, January 1950, pp. 12-13, 30-31.
[3] L.A. Zadeh, System Theory, Columbia Engineering
Quarterly, Nov. 1954, pp. 16-19, and 34.
[4] L.A. Zadeh, From Circuit Theory to System The-
ory, Proceedings of the IRE, vol. 50, 1962, pp. 856-
865.
[5] L.A. Zadeh, Ch. A. Desoer: Linear System Theory:
The State Space Approach. New York, San Francis-
co, Toronto, London: McGraw-Hill Book Comp.,
1963.
[6] L. A. Zadeh, E. Polak, System Theory, Bombay,
New Delhi: McGraw-Hill 1969.
[7] N. Wiener: Cybernetics or Control and Communi-
cations in the Animal and the Machine, Cambridge,
Massachusetts: MIT Press, 1948.
[8] L.A. Zadeh, J.R. Ragazzini: An Extension of Wie-
Figure 11: Illustration in Zadeh’s article [26], 1968. ner’s Theory of Prediction, Journal of Applied
Physics, vol 21, 1950, pp. 645-655.

438
[9] L.A. Zadeh: Theory of Filtering, Journal of the So- [19] L.A. Zadeh: My Life and Work ̶ a Retrospective
ciety for Industrial and Applied Mathematics, vol. View. Applied and Computational Mathematics,
1, 1953, pp. 35-51. vol. 10, Nr. 1, 2011, pp. 4-9.
[10] R. Seising: Interview with L.A. Zadeh on July, 26, [20] R. Seising: The Fuzzification of Systems. The
2000, University of California, Berkeley, Soda Genesis of Fuzzy Set Theory and Its Initial Appli-
Hall. cations – Developments up to the 1970s, Studies in
[11] R. Seising: Interview with L.A. Zadeh on June 15, Fuzziness and Soft Computing Vol. 216, Berlin,
2001, University of California, Berkeley, Soda New York, [et al.]: Springer 2007.
Hall. [21] L.A. Zadeh: Fuzzy Sets, Information and Control,
[12] R. Seising: Interview with L.A. Zadeh on June 16, vol. 8, 1965, pp 338-353.
2001, University of California, Berkeley, Soda [22] L.A. Zadeh: History of Computer Science at
Hall. Berkeley, Manuscript from November 1, 1992,
[13] O. Lowenschuss: Multi-Valued Logic in Sequential Personal Archive of L.A. Zadeh, page 2.
Machines. Ph.D. Thesis. School of Engineering, [23] L.A. Zadeh: Electrical Engineering at the Cross-
Columbia University, New York, 1958. roads, IEEE Transactions on Education, Vol. E-8,
[14] O. Lowenschuss: A Comment on Pattern Redun- Nos 2-3, June-September 1965, pp. 30-33.
dancy, IRE Transactions on Information Theory, p. [24] L.A. Zadeh: Education in Computer Science, Pro-
127, December 1958.W.F. Atchison et al.: An Un- ceedings of the National Conference on Data Pro-
dergraduate Program in Computer Science— cessing, Jerusalem, Information Processing Asso-
Preliminary Recommendations, Communications ciation of Israel, 1968, pp. E157-E167.
of the ACM, vol. 8, pp. 543-552, September 1965. [25] A.G. Oettinger: The Harvard-Software-
[15] O. Lowenschuss: Restoring Organs in Redundant Complementarity, Communications of the ACM,
Automata, Information and Control, vol. 2, pp. vol. 10, October 1967, pp. 604-606.
113-136, 1959.W.F. Atchison et al.: Recommenda- [26] L.A. Zadeh: Computer Science as a Discipline,
tion for Academic Programs in Computer Science, Journal of Engineering Education, Vol. 58, No. 8,
Communications of the ACM, vol. 11, March 1968. April 1968, pp. 913-916.
[16] W. Ulrich: Nonbinary Error Correction Codes, Bell [27] W.F. Atchison et al.: An Undergraduate Program
Systems Technical Journal, pp. 1341-1142, No- in Computer Science—Preliminary Recommenda-
vember 1957. tions, Communications of the ACM, vol. 8, pp.
[17] L.A Zadeh: Toward an Institute for Research in 543-552, September 1965.
Communication Sciences. IRE Transactions on In- [28] W.F. Atchison et al.: Recommendation for Aca-
formation Theory, Editorial, p. 3, March 1960. demic Programs in Computer Science, Communi-
[18] L.A. Zadeh: A Critical View of Our Research in cations of the ACM, vol. 11, March 1968.
Automatic Control, IRE Transactions on Automat-
ic Control, AC-7, No. 3, pp. 74-75, April 1962.

439