These experiments establishing bat echolocation were reported in Griffin and Galambos’s two seminal papers and formed part of the latter’s doctoral thesis. Griffin’s thesis, by prior agreement, was on bird navigation, the problem he had originally planned to study in graduate school experiments. The central question was whether birds released in unfamiliar territory immediately determined the homeward direction and flew directly back to their nests. He captured petrels, gulls, and terns and transported them, often in rotating cages, in different directions from the site of their capture, then released them and timed their return home. However, their flight times home were consistent with both a search until they found familiar landmarks and a leisurely but direct route home.
Directly tracking them should disambiguate these possibilities, he hoped. So he got Alexander Forbes (professor of physiology at Harvard Medical School and one of the founders of modern neurophysiology) to take him up in Forbes’s single-engine plane to try to track some gulls. Later, Griffin took flying lessons and bought his own two-seater with funds from the Harvard Society of Fellows. The results were again consistent with both a search pattern and true homing.
The Society of Fellows awarded three-year Junior Fellowships with generous research funds. The fellowship was originally supposed to be a super elite substitute for a Ph.D. with no required courses, teaching, exams, degrees, or requirements except for attending candlelit dinners along with the senior fellows. In practice, when the junior fellows went on the job market, say in distant Berkeley, they were told, in effect, “no degree, no job” and had to go back and get conventional doctorates. Today most junior fellows earn their doctorates first and it is a kind of fancy postdoc club, imitated predictably at such places as Princeton and Columbia. Griffin was fortunate to get elected to a Junior Fellowship, since his undergraduate grades had been too poor for a conventional graduate fellowship.
WARTIME
With the onset of war in 1941, Griffin became involved in war research at Harvard. His first assignment was to S. S. Stevens’s psychoacoustic laboratory. (Stevens was the founder of modern psychophysics.) There Griffin worked on auditory communication problems and acquired valuable familiarity with acoustic equipment. After a stint in the Harvard fatigue laboratory (working on such problems as the optimal gloves for handling fly buttons) he worked with George Wald (subsequently a Nobel laureate) on problems of night vision.
One rather weird wartime incident was the Bat Bomb project.6 Lytle Adams came to Griffin with the idea of equipping bats with small incendiary bombs and releasing them by plane over Tokyo where they would roost in Japanese “paper” houses and set fire to them. The government was supporting this idea, and Griffin agreed to help until he realized that there was no way bats could carry an adequate payload. In spite of Griffin’s disavowal of its feasibility, the Bat Bomb project continued on, even involving at one point Louis Fieser, the distinguished organic chemist and inventor of napalm. In his account years later, Adams continued to defend the project and claimed that it would have ended the war in a quicker and more humane way than Hiroshima and Nagasaki.
After the war, Griffin moved to the Cornell zoology department for seven years before returning to Harvard for another twelve years. The next paragraphs summarize some of his research interests in those years.
FURTHER RESEARCH ON BAT NAVIGATION
Research on bat echolocation (Griffin’s term) expanded in a number of different directions with an increasing number of collaborators.7 (Indeed by the time of his death, most of the now numerous bat researchers everyplace in the world saw themselves directly or indirectly, implicitly or explicitly as Griffin’s collaborators.) One such direction was to determine the limits of the avoidance and object-detection abilities afforded by echolocation. It was clear early that Myotis could discriminate wires down to a quarter of a millimeter, but could they actually echolocate moving-insect prey in the dark? Field experiments suggested that they could. This was confirmed by combining acoustic recording with ultra-high-speed strobe photography in an enclosure with released fruit flies and then weighing the bats before and after a short period of catching flies. Furthermore, the bats could quickly learn to discriminate pebbles and other inedible objects from flying insects. These experiments were carried out with Alan Grinnell, Fred Webster, and others.
Another direction initiated by Griffin, with his collaborators Alan Grinnell and Nobuo Suga (and encouraged by Galambos), was the neurophysiology of bat echolocation. Today, largely due to the work of Suga and his students, more seems to be known about the organization of auditory cortex in the bat than almost any other animal.8
Whereas the North American bats initially studied by Griffin emitted brief frequency-modulated (FM) signals, in 1950, F. P. Mohres discovered that the European horseshoe bat used longer-duration constant-frequency signals for echolocation. This inspired Griffin, Alvin Novick, and other collaborators to survey the signals produced by different species of bats. As most bat species are tropical, this led Griffin, Novick, and their collaborators to a number of exciting Latin American expeditions and the discovery of many different modes of echolocation, including one specialized for fishing and others in cave-dwelling birds.
In the last weeks of his life Griffin was out “night after night” on Cape Cod, “still trying to learn more about bats.”
BIRDS AND OTHER CREATURES
Griffin continued to work on the mysteries of bird navigation.9 What made this a diffcult problem was that although it became clear that birds (or some birds under some conditions) were using such cues as the elevation of the sun, the pattern of the stars, the magnetism of the earth, their own circadian rhythms, and spatial memory, it was diffcult to sort out the interaction and relative roles of these cues. Griffin pioneered the use of airplanes, radar, and high-altitude balloons to study this problem. (My first publication in a scientific journal, on bird navigation, arose out of a paper I wrote for an undergraduate seminar with Griffin. I then conducted research in his lab on the subject. My most vivid, if irrelevant, memory was the time he asked me to get the car battery from the next room for use as a power supply and I answered, “What does it look like?” He gave this Brooklyn boy a brief strange look and then went and picked it up himself ).
Griffin’s discovery of a “new sense” in bats probably influenced, at least in part, the discovery of other “new animal senses” such as infrared vision in snakes, infrasonic signals in elephants, and orientation and discrimination in electric fish. He played a more direct role in the story of the dancing language of bees. During the war the Austrian zoologist Karl von Frisch had discovered that honeybees could communicate the distance, direction, and desirability of food sources by a dancelike behavior. This work was hardly known in America in 1949 when Griffin arranged for him to give a series of lectures at Cornell, and then across the country, and shepherded their publication through Cornell University Press.10 Griffin had initially been skeptical until he replicated some of the critical experiments himself. (At the age of 72, Griffin published his last experimental paper; it was on bees.)
Griffin was interested in how beavers communicate. The last weeks of his life found him introducing microphones into beaver’s nests near the Harvard Field Station in Concord. Indeed, the number of anecdotes about the field studies he carried out in his last, and eighty-eighth, year that I collected while preparing this memoir are a measure of the man.
THE ROCKEFELLER INSTITUTE AND BACK TO HARVARD
In 1965 Griffin left Harvard to organize a new Institute for Research in Animal Behavior, jointly sponsored by Rockefeller University and the New York Zoological Society. It eventually included a field station in Millbrook, New York. Joined by the leading ethologist Peter Marler, and by Ferdinando Nottebohm, the well-known investigator of bird song and adult neurogenesis, the institute became one of the leading United States centers for the study of animal behavior. Among Griffin’s collaborators and students at the institute were Roger Payne, discoverer of acous
tic hunting by owls and of whale songs and now the leading advocate of whale conservation; Jim Gould, who extended von Frisch’s bee studies; and Carol Ristau, pioneer in the study of intentionality in the piping plover. From 1979 to 1983 Griffin was president of the Henry Frank Guggenheim Foundation and he used this position to encourage research on animal behavior.
When Griffin retired from Rockefeller in 1986 he spent a year at Princeton University and then returned to Harvard, where he worked at the Concord Field station and occasionally taught undergraduates. In this final period of his life he continued his experimental work on bats, birds, and beavers as well as his cognitive ethology advocacy, described in the next section.
COGNITIVE ETHOLOGY
For about the first 40 years, Griffin’s career had been that of the very hard-nosed empiricist and skeptic typified by the following oft-told tale (attributed to Griffin’s students Donald Kennedy, former president of Stanford and FDA commissioner, and Roger Payne, among others): When passing a flock of sheep while traveling in a car, his companion noted that among the flock of sheep there were two that were black. Griffin replied, “They’re black on the side facing us, anyway.” Then in 1976 Griffin began to publish a series of books and papers that contained no new data, no figures, but a host of citations and arguments from philosophers as well as scientists which challenged the contemporary world view of animals. He claimed that animals (and not just chimpanzees or even mammals) were aware and conscious and these properties of their minds should be the subject of scientific study, a field he named “cognitive ethology.”
At least at the beginning, these claims and exhortations were usually greeted by harsh and angry criticism (one critic called them the “satanic verses of animal cognition”) or the sadness of seeing a great experimenter supposedly slipping into premature senility. (He himself even called this interest an example of “philosopause.”) To better understand why imputing awareness or even minds to animals was considered outrageous, or at least, extrascientific, by most of those who studied animal behavior, we need to go back to Charles Darwin and the beginning of modern biology.
One of Darwin’s central points was the continuity of humans and other animals. As evidence of mental continuity Darwin cited examples from animals of humanlike emotions of joy, affection, anger, and terror as well as of what we now call cognitive functions such as attention, memory, imagination, and reason.11 George Romanes continued this tradition in what became known as the “anecdotal school.”12 C. Lloyd Morgan reacted against this approach and formulated what became known as Lloyd Morgan’s canon,13 essentially the application of the law of parsimony (Occam’s razor) to animal behavior: “In no case may we interpret an action as the outcome of the exercise of a higher psychological faculty, if it can be interpreted as the outcome of one which stands lower in the psychological scale.” This quickly came to imply the rejection of animal consciousness and awareness and a wariness to impute any complex cognitive functions to animals. This tendency was reinforced by Jacques Loeb’s theory of tropisms and the Russian school of reflexology,14 which also downplayed or denied consciousness in animals as well as humans. All these “objectivist” tendencies came together in the behaviorist movement, founded by J. B. Watson.15 The dominant figure in behaviorism, indeed in all of U.S. psychology until the rise of cognitive psychology, was B. F. Skinner.16 Skinner and the other “radical behaviorists” flatly denied the validity of the scientific study of consciousness, attention, awareness, thought, and other mental phenomena in humans as well as other animals.
The other principal group studying animal behavior was the ethologists deriving from a European zoological tradition.17 They tended to stress the role of innate wiring in animal behavior, in contrast to the behaviorists who stressed the role of experience. However, they too obeyed Morgan’s canon and were generally uninterested in the role of consciousness, intention, and mental experience in animal behavior. The cognitive revolution against behaviorism starting in the 1960s brought consciousness, attention, and awareness back into human psychology but had left other animals still essentially mindless and unaware.18
Thus Griffin’s plea for studying “the question of animal awareness” (the title of his 1976 book) was fiercely counter to the prevailing ideology in both psychology and zoology. Griffin used a variety of arguments coming from different directions and different fields to attack this view. One central argument was that it was simply anti-intellectual and anti-scientific to deny any subject an objective and experimental inquiry. A second argument was Darwin’s original one: the continuity of humans and other animals. Another argument was that animal communication, albeit admittedly fundamentally different from human language, might provide “a window on the animal mind.”
In his next two books, Animal Thinking (1984) and Animal Minds (1992), these arguments were amplified and supported by a Romanes-like compendium of experiments and observations that greatly enhanced, at the least, the case for animal consciousness and awareness. They included studies of tool construction and use, communication, planning, deception, blind-sight, cooperative hunting, and intentionality. Two new lines of evidence came into prominence. The first were a host of neurophysiological experiments seeking mechanisms of consciousness. Since most of these were invasive, such as single-neuron recording, they could only be done in animals and thus, with all due respect to Lloyd Morgan, they assumed animals were conscious, reflecting the change in the intellectual air that Griffin had helped bring about.
The second line of new evidence, increasingly prominent in Griffin’s last books and papers on cognitive ethology, came from studies done by Griffin’s students—such as Jim Gould on bees or Roger Payne on whales—and by the increasing number of quasi-students, investigators who were never formally his graduate students but readily acknowledged him as their mentor. These included Dorothy Cheney and Robert Seyfarth, who detailed communicative alarm calls and deception in vervet monkeys, and Irene Pepperberg, who trained a grey parrot to answer cognitive questions in English. (Even his formal students are not readily identified as he rarely attached his name to their work).
Although many biologists and psychologists are still uneasy about Griffin’s attribution of consciousness to nonhumans, particularly invertebrates, there is no question that he has radically opened up the field of animal behavior to new questions, ideas, and experiments about animal cognition. Because of his own towering achievements as a meticulous and skeptical experimental naturalist, his cogent and repeated arguments about studying the animal mind, and his support and encouragement of others, coupled with his unusual modesty and soft-spoken nature, Donald Griffin was able to effect a major revolution in what scientists do and think about the cognition of nonhuman animals.
NOTES
The details of Griffin’s early years, his student experiences, his account of the discovery of echolocation, and his continuing work on echolocation and all the Griffin quotes are from three of Griffin’s works: “Early history of research on echolocation,” in R.-G. Busnel and J. F. Fish, eds., Animal Sonar Systems (Plenum, 1980); “Reflections of an animal naturalist” in D. A. Dewsbury, ed., Leaders in the Study of Animal Behavior (Bucknell University Press, 1985); and “Donald Griffin,” in L. Squire, ed., The History of Neuroscience in Autobiography, vol. 2 (Academic Press, 1998). I would like to thank the following of Griffin’s colleagues for providing additional information: Robert Galambos, Alan Grinnell, James Simmons, Roger Payne, Marc Hauser, Greg Auger, Jim Gould, and Herb Terrace. This chapter was published previously in Biographical Memoirs of the National Academy of Sciences (86: 1–20 [2005], “Donald R. Griffin 1915–2003”).
1. Griffin, 1934.
2. Spallanzani, 1932; Galambos, 1942.
3. Gross, 1998a.
4. Watson and Lashley, 1915.
5. Galambos, 2004, personal communication. See also his memoir, Galambos, 1995.
6. Couffer, 1992.
7. Griffin, 1958.
8. E.g., Suga and Ma, 2003.
r /> 9. Griffin, 1964.
10. Frisch, 1950.
11. Darwin, 1871.
12. Romanes, 1882.
13. Morgan, 1900.
14. Loeb, 1900; Pavlov, 1929; Bechterev, 1932.
15. Watson, 1924.
16. Skinner, 1953.
17. E.g., Tinbergen, 1951.
18. Gardner, 1985.
12
THE GENEALOGY OF THE “GRANDMOTHER CELL”
A “grandmother cell” is a hypothetical neuron that responds only to a highly complex, specific, and meaningful stimulus, such as the image of one’s grandmother, that is, to a single percept or even a single concept. This chapter discusses the origin of the term, and the alternative view that complex stimuli are represented by the pattern of firing across ensembles of neurons, rather than that of a dedicated “grandmother cell.”
As originally conceived, a grandmother cell was multimodal, but the term came to be used mostly for representing a visual percept. As we shall see, the term arose because the first such neuron was postulated to represent a grandmother (actually, the very first was a specific mother). There might be many grandmother cells responding to a specific stimulus, such as one grandmother, but their response properties would be the same. Thus “coding by grandmother cells” is at the other extreme from “ensemble,” “coarse,” or “population” coding, in which a grandmother or other stimulus is coded by the pattern of activity over a group of neurons. In ensemble coding each member of the ensemble responds somewhat differently, for example to a granny’s wrinkles, to white hair, or to several different old women; the coding of a specific grandmother is done by a unique pattern of activation across the ensemble.
Starting in the early 1970s the term “grandmother cell” moved from laboratory jargon and jokes into neuroscience journals and serious discussions of the bases of pattern perception.1 The term is now nearly ubiquitous in introductory neuroscience and vision textbooks, where it often plays the role of straw man or foil for a discussion of ensemble or coarse coding theories of sensory representation.2 This chapter considers the origins of the term “grandmother cell” and similar expressions and, more briefly, the roots of ideas about ensemble coding.
A Hole in the Head Page 21