n a t i o n a l a c a d e m y o f s c i e n c e s
Any opinions expressed in this memoir are those of the author(s)
and do not necessarily reflect the views of the
National Academy of Sciences.
carl iver Hovland
1912—1961
A Biographical Memoir by
roGe r n. s HePard
Biographical Memoir
Copyright 1998
NatioNal aCademies press
washiNgtoN d.C
CARL IVER HOVLAND
June 12, 1912April 16, 1961
BY ROGER N. SHEPARD
YALE PSYCHOLOGIST Carl Hovland made singularly
important contributions to experimental, social, and
cognitive psychology (focusing respectively on human learn-
ing, attitude change, and concept acquisition). In the pro-
cess he worked unremittingly "to improve the standards
and quality of research in psychology and related fields,"
earning (in the words of one of his longtime coworkers)
universal recognition as a "statesman of the social sciences"
(Janis, 1968, p. 530).
Hovland also served as an insightful and trusted consult-
ant to numerous governmental and educational agencies,
industrial organizations, and philanthropic foundations. All
this he did within a life lasting not quite forty-nine years.
He could hardly have foreseen how limited would be the
time available to him (both his parents lived into their
nineties). Yet he compensated, in effect, through his re-
markable precocity, quickness of mind, and productive use
of every waking moment－along with his extraordinary ability
to bring together bright young researchers with widely dif-
fering theoretical perspectives, to provide them with sup-
port and subtle guidance, and to formulate coherent syn-
theses of the emerging results. A man of unsurpassed
gentleness and moral integrity, he left a deep and perma-
nent mark on everyone who knew him.
I first met Carl Hovland when I arrived for graduate
study in Yale's illustrious Department of Psychology in the
fall of 1951. Hovland's title, Sterling professor, seemed
wonderfully euonymous for this tall, distinguished man,
endowed as he was with rare personal qualities and wavy
hair turning to silver. Now, over forty-five years later, I am
astonished to realize that this revered member of the de-
partment, who had been serving as chairman of the de-
partment and director of the Laboratory of Psychology,
was at that time only thirty-nine years old!
Particularly striking were the apparent ease and efficiency
with which Hovland managed all the many things in which
he was always engaged and his constructive use of every
moment of time. While showing genuine interest in every-
one with whom he had contact, he had a way of keeping
administrative interactions brief and to the point. His ex-
traordinary memory enabled him to carry out much of the
department's business through chance meetings in the hall
or stairway－venues that minimized the risk of someone
plunking down in a chair in his office for more than the
time needed to resolve whatever issue was at hand. If Hovland
did not encounter a graduate student sufficiently soon con-
cerning some matter, the student would find a slip of pa-
per in his or her departmental mailbox with the succinct
notation: "See me. CIH." More than once, discussions of
my own research were carried out as I tried to keep up
with Hovland's rapid stride to the New Haven railway sta-
tion where he would be catching a train to New York－
perhaps to consult with AT&T, Bell Laboratories, or the
Rockefeller or Russell Sage Foundations.
On those occasions when I did actually sit down in
Hovland's office, he would also be reading his mail and
talking with someone else on the telephone. When I called
him on the phone, I could hear someone else in his office
and the occasional rattle of a letter being opened. And,
when I sent him a note, I imagined that while he was
perusing it, he would also be talking with someone in his
office and someone on the phone. I fantasized having the
delivery of my written letter, the playing over the phone of
my recorded voice, and my physical entrance into his of-
fice converge upon him simultaneously－thus gaining, for
once, his undivided attention! In truth, however, I wel-
comed the brief hiatuses that Hovland's time sharing en-
tailed as I was striving to marshal my ideas for his assess-
ment.
Another Hovland student, Herbert C. Kelman (now Cabot
professor of social ethics at Harvard), described to me
how the drafting of his 1953 paper with Hovland began:
"In consultation with Carl, I designed and carried out an
experiment on the sleeper effect [in which the tendency to
endorse a proposition from a low credibility source in-
creases as the source is forgotten]. When the data were
collected and analyzed, I . . . told him that I would like
him to coauthor the article reporting the research. In his
customary generosity, he told me that this was my experi-
ment and he was not expecting coauthorship. But I in-
sisted－whereupon he pulled out a yellow pad and started
writing! Right then and there!" (Kelman, letter of March
25, 1997).
Hovland was the most efficient and organized individual
I have ever known. But the efficiency and organization
was all in his head; it did not depend on external aids. He
conducted classes and chaired meetings in his quiet, infor-
mal manner without notes, while the desk and side table
in his office remained piled with papers in no visible or-
der. When another of my fellow graduate students inquired
whether he might retrieve a term paper to correct an error,
Hovland briefly rummaged through papers piled on
the side table. Then turning to my waiting friend, he re-
marked, "You may think there is no order here. Actually,
there is an order; it's just not an order designed to meet
that particular type of request." And order there evidently
was; Hovland's secretary, Jane Olejarczyk, told me, "Quite
often he would call and ask me to retrieve some document
with instructions like: it's in the third pile from the left on
the table by my desk, about a third of the way down, and
there's a Russell Sage report, printed on blue paper, just
before you get to it . . . Amazing! He was always on target"
(personal communication of May 29, 1997).
Hovland was a master of the Socratic method. Seem-
ingly without any prepared agenda, he would ask the graduate
students around the seminar table for their comments on
the (always seminal) readings he had assigned, or for their
proposals concerning an illustrative problem of experimental
design or data analysis he was working through on the
chalk board. At first, this evoked frustration or anxiety in
students accustomed to more structured styles of instruc-
tion. (A student who had volunteered to calculate－in those
days, by means of a slide rule－a number called for by the
illustrative problem might find that, before he or she was
able to come up with the answer, Hovland was already
writing it on the board, apparently having arrived at it by
his own swifter, purely mental calculation.) Former Yale
student Philip Zimbardo (now a professor of social psy-
chology at Stanford) remarked that the combination of
Hovland's shyness and intellectual mastery may have pre-
vented him from even suspecting that some students found
him intimidating (personal communication of April 3, 1997).
Nevertheless, out of our bumbling efforts a coherent pic-
ture would gradually crystalize, to be succinctly articulated
by Hovland at the end of each class session. It was the goal
toward which Hovland evidently had been subtly guiding
us all along.
I asked Hovland to serve as my dissertation advisor not
only because I valued his quick intellectual grasp but also
because he seemed uniquely free of commitment to any
particular theoretical position and, hence, supportive of
the exploration of promising ideas, wherever they might
lead. Because of the great respect everyone had for him,
Hovland was also able to give my career a couple of unex-
pected boosts at its very start. He endorsed the suggestion
of a younger member of my dissertation committee, Bur-
ton Rosner, to take the unusual step of recruiting a math-
ematical psychologist from outside Yale to serve on the
orals committee of my more-than-usually mathematical dis-
sertation. One consequence was that the up-and-coming
outside examiner selected, George A. Miller, invited me to
join him a year later as a postdoctoral associate at Harvard.
Then, following those two postdoctoral years, both Hovland
and Miller recommended my appointment as a member of
technical staff in a small basic research group that Hovland
had been instrumental in establishing in the Bell Tele-
phone Laboratories, Murray Hill, New Jersey. The research
I was able to carry out during my two postdoctoral years
at Harvard (where I first learned to program on the Univac
1, just given to Harvard) and during the next eight years
at the Bell Labs (where I had access to a major computer
facility) undoubtedly contributed to my own ensuing ap-
pointment to a professorship at Harvard.
In 1957 I participated－along with both Miller and
Hovland－in a Summer Institute on the new computer simu-
lation approach to modeling human cognitive processes
organized by Alan Newell and Herbert Simon at the RAND
Corporation in Santa Monica. Simon, who remembers
Hovland "with great fondness," mentioned that Hovland
and Miller had "co-opted" him to join their small ad hoc
committee of the Social Science Research Council, which
had some Ford Foundation money for work in cognition.
It was this money, Simon said, that made possible their
Summer Institute (personal communication of May 27, 1997).
Over a lunch with Hovland in Santa Monica that summer,
I recalled how my doctoral research at Yale only two years
before had necessitated my approximation of the eigen
roots and vectors of matrices by hours of tedious computa-
tion on mechanical desktop calculators. "When," I won-
dered, "would Yale obtain a programmable electronic com-
puter?" With a wry smile, Hovland replied that he was on
a committee that had just been established at Yale to re-
ceive the gift of such a computer－in case one should be
offered! Only three years later, the 1960 papers on com-
puter simulation of thinking and concept attainment
authored by Hovland, alone and with his student Earl Hunt,
were already appearing.
It was shortly after joining the Bell Labs that I began my
one direct research collaboration with Hovland. Herbert
Jenkins and I had undertaken a study of classification learn-
ing in which human subjects learned by trial and error
which of two responses was correct for each of the eight
possible stimuli having either of two values on each of
three binary dimensions (for example, square or triangu-
lar, large or small, and black or white). Jenkins and I sought
to determine the number of trials required to learn differ-
ent classifications in which correct responding required
taking account of values on just one, on two, or on all
three of the stimulus dimensions.
When we mentioned this study to Hovland, we learned
that quite independently, he and two research assistants
had just begun presenting subjects with explicit classifica-
tions of just such binary-valued stimuli into two groups of
four (one displayed on the left, the other on the right).
They, however, were measuring subjects' speeds and accu-
racies of reconstruction of the two groups from memory,
and recording how the subjects described the rules they
found to govern each classification. We quickly agreed to
join forces and, during our ensuing collaboration, Jenkins
and I (often together with the Bell Labs learning researcher
Ernst Rothkopf) would meet with Hovland－usually at his
home in Hamden, outside New Haven.
On these visits, the Hovlands' longtime housekeeper Eliza-
beth would serve us lunch, elegantly presented with fine
china, silver, and linens in the Hovland's formal dining
room. I must have been seated in Mrs. Hovland's custom-
ary place. For, under a slight bump in the rug there was a
button that I sometimes inadvertently hit with my foot,
summoning the housekeeper, to my mounting chagrin.
At about this time, a growth in Hovland's neck (in the
parotid gland just below his right ear), which had been
diagnosed as benign some years earlier, had recurred and
was now determined to be malignant. Both the advance of
the cancer and the measures undertaken for its treatment
(surgery, radiation, and a then highly experimental che-
motherapy) were soon exacting a toll on Carl's previously
inexhaustible energy, entailing a temporary loss of his full
head of hair, which had rapidly turned entirely white, and
a total deafness in his right ear.
Long before, Carl's wife Gertrude, like himself, still rela-
tively young and universally regarded with admiration and
affection, had been increasingly afflicted with rheumatoid
arthritis. Anticipating his own impending death, Hovland
became deeply concerned about his wife's growing help-
lessness. Her neck was now so fragile that she had to wear
a neck brace whenever she was up and about.
On August 26, 1960, my two colleagues and I made our
last scheduled trip from the Bell Labs to the Hovlands'
home to discuss the final stage of our collaborative project.
We were met at the door by housekeeper Elizabeth, who,
tearfully and barely able to speak, informed us that Mrs.
Hovland had an accident earlier that morning and that
Mr. Hovland would not be able to meet with us. We got in
the car and headed back to New Jersey.
I later learned that Gertrude, having gotten out of bed
without her protective collar, stumbled and fell. Her weak-
ened neck snapped and she died instantly. A few days later,
Carl called me to apologize for not being able to meet
with us after our long drive. When I assured him that no
apology was necessary and expressed my heartfelt sympa-
thy, he became, for the only time in my experience, choked
with emotion and was briefly unable to speak. The loss of
his beloved wife was a terrible blow to this most caring
and responsible of men－left, as he now was, with two
children in their late teens and with less than a year re-
maining of his own life.
Right up to the end, Hovland continued doing (to the
extent that he was physically able) just what he had been
doing even before he learned that he was mortally ill.
Apparently, Hovland had always proceeded each day with
what he regarded as most important－as if that day might
be his last. To avoid the stairs, his final weeks were spent
in a bed that had been set up in the same dining room
where my colleagues and I used to talk with him over
lunch. He was cared for by his son David, then an under-
graduate at Yale, and by his daughter Kathie, who, having
just entered Wellesley College, traveled down from Massa-
chusetts to be with her father during the weekends. Carl
died on Sunday night, April 16, 1961, just after Kathie left
for her trip back to Wellesley. Coincidentally, problems
arising in their necks had cut short the lives of Carl and
Gertrude alike, near the ends of their forty-eighth years.
The 1961 Shepard, Hovland, and Jenkins study "Learn-
ing and Memorization of Classifications" appeared in the
Psychological Monographs in that same year－but not in time
for Hovland to see it in print. Along with Hovland's own
last book Social Judgment, written in collaboration with
Muzafer Sherif (who completed it after Hovland's death),
our monograph was thus one of the last publications on
which Hovland appeared as an author. Some thirty years
later, this monograph attracted renewed interest among
cognitive scientists, who have used our results to test alter-
native connectionist or "neural net" models for classifica-
tion learning; or to elucidate the roles of stimulus dimen-
sions called perceptually "separable" (like size and shape－as
in Shepard, Hovland, and Jenkins, 1961, p. 3) versus those
called perceptually "integral" (like lightness and satura-
tion of colors－as in Shepard and Chang, 1963, p. 96). And
the three students who served as research assistants in this
work－Albert Bregman and Earl Hunt (with Hovland) and
John Gibbon (with Jenkins and me)－have all gone on to
make their own influential contributions at three major
universities (Bregman in auditory perception at McGill,
Hunt in human cognition at the University of Washington,
and Gibbon in timing behavior at Columbia).
www.psychspace.com心理学空间网
