BRANDON P. REINES
ON THE LOCUS OF MEDICAL DISCOVERY*
Key Words: accident, anomaly, continuity, experiment, logic of discovery,
psycholinguistic realm
I. INTRODUCTION: AN ANOMALY OF MEDICAL HISTORIOGRAPHY
Discovery in medicine has long fascinated mankind. While
medical historians have written copiously about the ostensible
discoverers themselves, historians have contributed relatively
little to the literature on the methodology of discoveries in pathology, physiology, therapeutics, and prophylaxis. On the other
hand, philosophers of science have begun to seriously consider
both the structure of theory and the logic of discovery in medicine
(Engelhardt, 1979; Gorovitz and Maclntyre, 1976; Caplan, 1986). In
particular, Schaffner (1980a, 1980b, 1980c, 1985, 1986) has
grappled with both of those traditionally philosophical themes.
He argues convincingly that medical theories as formally
presented in medical texts are 'theories of the middle range' that
are based largely on analogical extension from infrahuman
Brandon P. Reims, D.V.M. Director, The Center for Health Science Policy, 904 N.
Wayne Street, Suite 104, Arlington, VA, 22201, U.S.A.
The Journal of Medicine and Philosophy 16:183-209,1991.
© 1991 Kluwer Academic Publishers. Printed in the Netherlands.
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
ABSTRACT. A search for consensus about the methodology of discovery among
physicians and physiologists led the author to identify a crucial anomaly of
medical historiography: in general, physicians stress the significance of clinicopathologic method, while physiologists emphasize the experimental. Hence,
physicians and bench scientists might be perceived as members of epistemically
distinct research traditions. However, analysis of the historical development of
discoveries in medicine, exemplified by case studies in physiology, bacteriology,
immunology, and therapeutics, reveals that the epistemic dichotomy is illusory.
Both physicians and bench scientists discover in the same way: by identifying
and explaining clinical anomalies. It is argued that the sociological role of
experimentation is to dramatize clinical hypotheses and not test them in a
Popperian sense.
184
Brandon P. Reines
II. DISCOVERY OF AN EXPLANATION FOR THE ANOMALY
Accounting for the drastically different 'world views' of bench
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
species. The structure of pedagogic medical theories are clearly
anomalous relative to physical theory. Less compelling, however,
are Schafmer's attempts to identify an inherent logic of medical
discovery.
Laudable as are the continued efforts to trichotomize discovery
(e.g., Schafmer, 1980b, p. 191), the emphasis on structure and logic
renders such work vulnerable to genetic fallacy and leaves little
room for evaluation of regularities in cognitive dimensions other
than logic. In a critique of Schafmer, Maull argues that Schafflier's
logical account of discovery slights "the process by which the
problem itself is generated and first recognized as a problem"
(Schafmer, 1980c, p. 211). Hence, philosophers have identified
certain logical characteristics of medical theory and discovery
without elucidating the historical process of medical discovery.
Medical scientists themselves, however, have long attempted to
define that process. Physicians and bench scientists have written
extensively about discovery. In particular, there is a large
literature on the methodology of physiological discovery. A
search for consensus about the methodology of physiological
discovery among the de facto historians led the author to identify
an anomaly of medical historiography. While there is no interprofessional consensus about methodology, there is surprising
agreement within the professions. In general, physicians stress the
significance of clinico-pathologic method (Garrod, 1919;
McQuarrie, 1944; Peller, 1967; Good, 1967; Beecher, 1960; Beeson,
1979), while physiologists emphasize the experimental (Comroe,
1977; Dale, 1948; DeKruif, 1926; Cannon, 1945). In an exploration
of clinical and laboratory epistemology, G. Canguilheim (1989)
reviews historical evidence that physicians hold that human
pathology enlightens physiology, while physiologists maintain
the opposite position: i.e., physiology enlightens pathology. The
extent to which methodological accounts of discovery conflict
across professional lines has rarely been highlighted, although the
dichotomy in epistemology between physicians and bench
scientists constitutes an extraordinary anomaly of medical historiography. It is perhaps the central paradox in the intellectual
history of medicine.
On the Locus of Medical Discovery
185
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
scientists and physicians is facilitated by sociological analysis. In
particular, an exploration of the distinctive sociological roles of
bench scientists and physicians clarifies the origin of their apparently conflicting epistemological views. Bench scientists and
physicians occupy distinctive sociological niches in the medical
research community. The sociological role of the bench scientist in
the community of medicine has been the focus of scholarship by
Dubos (1959), and more recently Latour (1988), while the competition between bench scientists and clinical investigators has been
investigated by Geison (1979) and Maulitz (1979). An emerging
view of the role of bench scientists is that they are often Aristotelian experimenters who use showy laboratory experiments in
an attempt to dramatize hypotheses. In a recent work, for example, Latour writes that "Pasteur's genius was in what might be
called the theater of the proof... Pasteur invented such dramatized
experiments that the spectators could see the phenomena he was
describing in black and white" (1987, p. 85). The heretofore
unexplored implication of the bench scientist-qua-dramatist
hypothesis is that discovery in physiology and medicine occurs
prior to laboratory studies in an entirely different locus, in
'Nature': the clinic, postmortem room, and/or the field. Further
potential justification for such an hypothesis derives from 20th
century philosophy of science including Kuhn (1962), Humphreys
(1986), and Root-Bernstein (1989) on the crucial role of anomaly in
discovery. Such philosophers argue convincingly that the identification of anomaly within an existing theory is often the first
step in discovery. On a priori grounds, human pathological
evidence would seem to constitute an abundant source of potential anomalies to reigning doctrine. On the basis of sociological
analysis and recent work in the philosophy of science, therefore,
we may hypothesize that discovery usually occurs in the clinical
setting.
On the other hand, bench scientists might argue that anomalies
are more often identified in the laboratory, based upon the
pedagogic works of Claude Bernard and his followers. Among
Claude Bernard's more famous dicta is his exhortation to the
budding bench scientist that "on entering the laboratory to
perform the actual job of experimenting, he should leave his
imagination in the coat room with his overcoat..." (see Bernard,
1957, p. 6). The implication of Bernard's dictum is that preconceived notions or hypotheses are often contradicted by laboratory
186
Brandon P. Reines
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
testing: unexpected observations are common in the physiological
laboratory. Indeed, in An Introduction, Bernard (1957) stressed the
role of chance or serendipity in the genesis of his own physiological discoveries. The rhetorical impact of Bernard's An Introduction
on medicine, helps explain why discovery in medicine has been
attributed so often to accidental discovery in the laboratory. In
fact, the historical literature on 'accidental discover/ in the
laboratory is so vast that it is tantamount to a genre. The 'lo and
behold' genre was inaugurated by Bernard and passed on to
physiologists and provivisection propagandists alike.
The extent to which provivisection tracts are written in the
accidental discovery genre has been unrecognized by historians of
physiology. Sir Richard Owen's account of John Hunter's
apocryphal experiment on the stag of Richmond Park is a classic
of the genre (e.g., Nuland, 1988), with its emphasis on the role of
unexpected experimental results in ushering in Hunter's ostensible discovery of collateral circulation. In keeping with the
historical genre established in the 19th century, practically all of
the bench scientist-historians of the 20th century including
Cannon (1945), Dale (1948), DeKruif (1926), and Comroe (1977)
rhapsodize on the theme of 'happy accidents', 'The Three Princes
of Serendip', and other laboratory 'surprises'. A chorus of 'Eureka'
would seem to emenate from the laboratories of the leading lights
of 19th and 20th century physiology and bacteriology. Many
major medical discoveries of the 19th and 20th centuries including
adrenal physiology (e.g., Comroe, 1977, p. 54), pancreatic physiology (e.g., Comroe, 1977, p. 53), penicillin (Fleming, 1929), the twolimb theory of the immune system (e.g., Schaffher, 1980), the
azygous or 'low flow' principle of heart-lung bypass (e.g.,
Johnson, 1970), the myogenic theory of the heartbeat (e.g., Geison,
1978) anaphylaxis (e.g., Comroe, 1977, p. 56), the vitamins (e.g.,
Eijkman, 1965), the effect of corticosteroids on fetal lung development (e.g., Liggins, 1969), the function of the carotid body (e.g.,
Comroe, 1977, p. 57), the discovery of heparin (e.g., Comroe, 1977,
p. 90), the chemical theory of nervous transmission (e.g., Dale,
1948), and the Rhesus factor have been attributed to accidents in
the hands of bench scientists (e.g., Rous, 1945-1948).
Yet the laboratory accident genre is untenable as an account of
discovery for several reasons. First, the notion of accidental
discovery is not explicable in an hypothetico-deductive
framework (i.e., accidents may happen in the laboratory but not in
On the Locus of Medical Discovery
187
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
the mind). Second, the notion is inconsistent with the growing
literature stressing the continuity of discovery with prior theories,
arguments, and data. Finally, revisionist historians of science have
repeatedly shown that virtually any ostensibly accidental discovery can be construed as a stage of an ongoing research
progrem (e.g., Root-Bernstein, 1989). It is not accidental that
Bernard himself construed his discoveries as products of 'chance'.
While Medawar (1975) has argued that Bernard was among the
few 19th century biologists to reject Baconianism, Medawar (1969)
has also pointed out that the notion of 'chance discovery7 is
inexplicable in a hypothetico-deductive framework.
Given Bernard's repeated reference to chance discovery, by
definition, he was at least part Baconian inductivist. A more
accurate portrayal of Bernard qua philosopher of science is that he
was a hybrid: part inductivist and part hypothetico-deductivist.
Bernard's curious blend of inductivism and deductivism is
explicable on sociological grounds. We may posit that his primary
commitment was to perpetuate the physiological laboratory. The
surest path to that goal was to convince physicians of the ontological status of laboratory phenomena in the medical domain.
Bernard's strategy was to dramatize the power of the laboratory
by 'discovering' in it. There is little if any evidence that Bernard
actually discovered in his laboratory. He nonetheless denies any
connection between his discoveries and prior data or hypotheses.
In An Introduction, Bernard (1957) contends ex cathedra that his
discoveries occurred by chance in his laboratory and further that
his laboratory discoveries were automatically to be perceived as
medical discoveries. In a famous passage, Bernard wrote that
"Experimental ideas are often born by chance, with the help of
some casual observation" (1957, p. 151).
Characterizing the laboratory observations as 'born by chance'
serves three primary functions for Claude Bernard: 1) it allows
him to avoid accounting for the genesis of his discoveries; 2) it
assures that the locus of discovery would be construed as the
laboratory; and most importantly, 3) it apparently severs
laboratory epistemology from clinical epistemology. In order to
ensure the success of his program, Bernard had to convince the
medical world that physiology enlightens pathology and not vice
versa. He nonetheless hoists himself on his own petard on that
crucial point: he contradicts himself repeatedly in his various
works. While conceding that physiological investigation always
188
Brandon P. Reims
III. EXAMPLES OF MISCHARACTERIZATION OF THE PROCESS
OF MEDICAL DISCOVERY
A. Digestive Function of the Pancreatic Juice
Anyone attempting to discover the function of the pancreatic juice
in the mid-19th century would have to contend with prior theory
and observations. The principle views were shaped by: 1) Brunner's (1683) extirpation experiments had apparently led him to
conclude that the pancreas is not essential to digestion; 2) anatomical studies had led anatomists to postulate that the pancreas
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
begins with pathological observations in the clinic and postmortem room, in tracing the genesis of his own discoveries, he usually
neglects to mention the clinical source of his inspiration. He does
not dismiss the influence of pathological observations entirely but
he does 'downplay7 them. In An Introduction, Bernard nonetheless
insists that the physiologist must always begin his investigation
by reading the human pathological literature. He writes: "...we
should first of all state the medical problem as given by observation of the disease, then try to find the physiological explanation,
by experimentally analyzing the pathological phenomena. But in
this analysis, medical observation must never disappear or be lost
sight of; it must remain as the constant basis or common ground
of all our studies and explanations'' (1957, p. 199). Bernard does
not explain why it is essential that the physiologist know the
clinical literature before experimenting but the implication is clear
enough; that is, it is pathology that enlightens physiology. In his
memoir on the pancreas, however, he insists that it is physiology
that enlightens pathology (Bernard, 1856, p. 84). In his accounts of
discovery, in fact, he utilizes the tactic of 'temporal inversion': he
consistently argues that the results of his laboratory experiments
led him to search for equivalent natural experiments in the clinical
literature, though the natural experiments had been published
many years prior to Bernard's animal experiments. Later
physiologists and historians frequently inverted the relative
temporal sequence of clinical and laboratory 'discovery'. An
example of Bernard's tactic of temporal inversion and examples of
its influence on subsequent accounts of discovery in bacteriology,
immunology, and pharmacology follow.
On the Locus of Medical Discovery
189
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
functions as an 'abdominal salivary gland' (i.e., digests only
starch); 3) Brodie's (1823) experiments had apparently led him to
conclude that bile is essential to the formation of chyle. The
received view therefore held that the pancreas is not essential to
effective digestion. Any scientist attempting to refute that assertion would have to adduce evidence that 1-3 are false. It is
generally assumed that Claude Bernard discovered that the
pancreas is essential to the digestion of fat. Bernard insists that he
discovered the fat-emulsifying properties of the pancreatic juice
during the winter of 1846 in experiments in which he had fed
meat to rabbits. Bernard writes: "lo and behold a fact presented
itself which I had not remotely thought of, but which became, as
we shall see, my starting point in a new piece of work' (1957,
p. 153). The fact to which Bernard refers is his observation that the
intestinal lymphatics of a rabbit were filled with clear fluid
throughout the duodenum but were filled with milky white chyle
near the jejunem. He contends that he was intrigued by the
observation because it was anomalous relative to the position of
chyle-filled lymphatics in meat-fed dogs (throughout the
duodenum). His explanation of the anomaly was that the
pancreatic duct empties much lower down in the intestine of
rabbits than in dogs and the emulsification of fat, as well as its
incorporation into cyle, could not take place until after mixing
with pancreatic juice.
Holmes (1974), however, was unable to corroborate Bernard's
classic account by examination of his laboratory records. While
Holmes nonethless thought that Bernard may have discovered the
function of the pancreatic juice from an observation in the winter
of 1846 but had "nearly forgotten about it", Bernard's laboratory
records contradict that notion. In early 1847, according to Holmes,
Bernard first recorded his interest in exploiting the anatomical
peculiarities of the rabbit to explore the function of the pancreatic
juice (Holmes, 1974, p. 383). Not until April of 1848 did Bernard
finally perform an experiment on a rabbit the results of which
approximate that of his classical account. Bernard expresses no
surprise at the results of this or any other of his animal
experiments. In attempting to explain that fact, Holmes posits that
Bernard's prior in vitro observations would have led him to
anticipate seeing whitened chyliferous vessels downstream of the
pancreatic duct. As Holmes writes, however, "the very fact that he
did try an [in vitro] experiment with tallow in the midst of the
190
Brandon P. Reines
of relevant cases, and since my first memoir appeared, my attention has
been drawn to the signs and symptoms of human pancreatic disease...."
(emphasis added, 1856, p. 76). Bernard's credibility is nonetheless
strained by the temporal relationship between the clinical cases he
reviews and his animal experiments since many of the clinical
cases he cites were reported in the British and French clinical
literature over a decade before he undertook his animal experiments. For example, several of his case studies were reported in
an early British monograph (Lloyd, 1833) on fatty stools.
Not only does Bernard's analysis of human pathology cast
doubt on the significance of his laboratory experiments in the
discovery that the pancreas is essential to fat digestion, it belies his
contribution to the discovery. The reason is that Bernard never
clearly identifies which of the human case studies is clearly
anomalous relative to the received views 1-3. Given that most of
the cases involve blockage of both the bile duct and the pancreatic
duct, Bernard should have identified the case(s) in which only the
pancreatic duct is obstructed. He is careful, although entirely
unconvincing, in his attempt to demonstrate how experimental
evidence refutes 1-3. With regard to clinico-pathologic studies,
however, he blithely asserts that: "All these cases, in which an
autopsy enabled the state of the pancreas to be assessed, show
clearly that functional disease of the gland is associated with the
excretion of fat, in exactly the way that it is seen in dogs after
destruction of the pancreas..." (1856, p. 84). While Bernard had
read a careful analysis of the human cases by Sir James Paget prior
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
others implies that he already had some expectation that the fatty
material would be modified" (Holmes, 1974, p. 383). Bernard must
therefore have believed that the pancreas is essential for fat
absorption before undertaking any laboratory experiments.
Yet, in his Memoir, Bernard (1856) carefully avoids implying
that he elucidated the function of the pancreatic juice by reading
the clinical literature. He seems to go out of his way to 'put the
cart before the horse7. In reviewing cases in which pancreatic
damage is linked to malabsorption of fat, he assures the reader
that his animal experiments led him to read the clinical case
studies and not vice versa. "So it seems essential," Bernard
emphasizes, "for us to ascertain if the presence of fat in the
excreta, which we know is found in animals in which the pancreas
is destroyed, is present in man secondary to pathological changes
in the same organ.... There already exist in the literature a number
On the Locus of Medical Discovery
191
B. Discovery of the Principle of Vaccination
A more celebrated case of accidental discovery is Louis Pasteur's
discovery of the principle of vaccination. The classic version of the
discovery was apparently created by Emile Duclaux in 1896 and
then appeared in subsequent Pasteur biographies as well as
DeKruif s the Microbe Hunters. The tale always begins in the same
way. Pasteur had forgotten to throw out some cultures of chicken
cholera germs when he went on vacation during the summer of
1879 and the germs sat on the laboratory counter for two long
months fully exposed to the air. When he returned in the fall, in
the hope of enlivening any surviving microbes, he had his assistant injected the old cultures into healthy chickens. The chickens
were unaffected indicating that the germs were either dead or
nonvirulent. Pasteur then inexplicably told his assistant to inject
the same chickens with fresh cholera bacilli. When his assistant
injected fresh cholera organisms into the same chickens, lo and
behold, the chickens did not contract the deadly disease. Pasteur
had discovered that oxygen exposure can attenuate a germ
sufficiently to render it an effective vaccine. It was this experience
that supposedly led Pasteur to coin his famous dictum that
"Chance favors the prepared mind".
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
to publication of the 1856 memoir, he failed to appreciate its
cogency. Paget had noted that most of the cases were complicated
by concurrent liver damage "but in at least one the liver was
healthy, and there appeared nothing but the absence of the
pancreatic fluid from the intestines to which the excretion or nonabsorption of fatty matter could be ascribed" (1848, p. 57). It was
Bernard's failure to adduce clinico-pathologic or experimental
evidence that clearly contradicted the received view of pancreatic
function that led his Scottish critics to conclude that his publications on the pancreas amounted to dramatization of his preconceived opinion (Buchanan et ah, 1858). The particulars of their
critique stand to this day. While both Bernard's experimental and
clinical work on the pancreatic juice effectively dramatized Paget's
discovery, the net effect of Bernard's publications on the pancreas
was to begin to canonize the vivisectional element of his experimental medicine at the expense of clinical analysis. His later
pedagogic works led to further dimunition in the rhetorical
impact of clinical studies, and corresponding augmentation of the
drama of already alluring animal experiments.
192
Brandon P. Reims
Without wishing at this time to assert any relationship between the viruses of
smallpox and cowpox, it seems from the preceding facts that in fowl cholera
there exists a state of the virus, relative to the most virulent virus, which
functions the same way as cowpox virus in relation to smallpox virus. Cowpox
virus, properly stated, brings about benign illness, cowpox, which protects
against a more serious illness, smallpox. In the same way, the fowl cholera agent
can occur in such a state of attenuated virulence that it induces the disease but
not death, and in such a way that after recovery, the animal can withstand an
inoculation with the most virulent agent. Nevertheless, in certain respects the
difference between the two orders of facts is considerable, and it is not amiss to
remark with respect to knowledge of facts and principles, a course of studies on
fowl cholera will probably be more helpful. Whereas there is still a dispute about the
relationships between smallpox and cowpox, we know for certain that the
attenuated agent of fowl cholera is derived directly from the most virulent agent
of this disease, so that their natures are fundamentally the same (emphasis
added, Pasteur, 1880, p. 673).
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
However, Caddedu (1987) has examined Pasteur's notebooks
for a more coherent account of the discovery. Caddedu found that
Pasteur's assistant, Emile Roux, had actually begun deliberate
attenuation experiments while his mentor was on vacation.
Probably at Pasteur's suggestion, Roux had systematically attempted to attenuate the cholera bacilli. Roux was able to attenuate the bacilli by altering broth acidity by directing a current
of oxygen over the cultures. It is nonetheless true that Pasteur did
leave cholera cultures out when he went on vacation and he did
have them inoculated into chickens when he returned; and the
chickens were unaffected. When Pasteur had Roux inoculated the
same chickens with virulent germs, however, they actually died.
Between October 1879 and March 1880, apparently failing to
appreciate the potential of Roux's method of acidification by
oxygen exposure, Pasteur had Roux treat the bacilli in myriad
ways including heating them, exposing them to air, passing them
from one animal to another, and growing them in different media.
Based on Caddedu's reconstruction of events, it is nonetheless
clear that Pasteur (or perhaps Roux) must have discovered the
principle of vaccination before he left for his vacation in the
summer of 1879. The principle was the guiding light of both
Pasteur and Roux's work on attenuation. Without assuming its
influence on them, it is impossible to explain their behavior. In
fact, while not highlighted by historians, Pasteur was quite frank
about the actual origin of his discovery in his paper of 1880 in
which he wrote:
On the Locus of Medical Discovery
193
C. Discovery of the Two-Limb Theory of the Immune System
Like the discovery of vaccination, the discovery of the two-limb
theory of the human immune system has been attributed to
laboratory accident. Schaffher contends that: "The theory had
what might be called its protosource in a chance finding of Glick,
Chang, and Japp (1956) that a hindgut organ in birds known as
the bursa of Fabricius had an important immunological func-
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
While Pasteur had gone one step beyond Bernard in including the
'prepared mind' in the equation of discovery, it was not 'chance'
that he had read Jenner's paper on vaccination. Pasteur wanted to
develop practical means of protecting animals and human beings
against disease and had internalized the law' of acquired resistance that preceded Jenner's clinical studies: i.e., that a person
who survives an acute disease rarely contracts that same disease
again (based on enumerative induction). Pasteur recognized
Jenner's observations as posing an epistemic threat to the law
because cowpox is apparently a different disease from smallpox.
Like Leverrier encountering the anomalous orbit of Uranus,
Pasteur attempted to account for the apparent anomaly within the
received theoretical framework. While not "wishing to assert any
relationship between the viruses of smallpox and cowpox,"
Pasteur assumed that just such a relationship exists. It was the
essence of his discovery: he assumed that the cowpox virus is an
'enfeebled' form of the smallpox virus. That assumption implied
that any disease might be prevented by vaccination with a less
virulent form of it. Hence, Pasteur was able to save the received
view of resistance by positing an invisible similarity between the
viruses of smallpox and cowpox much as Leverrier assumed the
existence of the planet Neptune. While the molecular similarities
between pox viruses were not revealed until many decades later,
Pasteur's conservative interpretation of the cowpox/smallpox
phenomenon rested on a solid evidential basis before the chicken
experiments. He developed attenuation methods in order to
further test his principle and to create practical means of vaccination. His ability to effectively vaccinate chickens against chicken
cholera constitute a legitimate test of his theory for chickens but it
was hardly conclusive for man. The chicken experiments
dramatized the potential of his principle of vaccination for
preventing human disease.
194
Brandon P. Reines
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
tion../' (1980a, p. 72). In fact, however, the two-limb theory
evolved from attempts to overthrow the humoral theory of
immunity which reigned throughout the first half of the 20th
century. The dogma that all immune reactions are essentially
antibody-mediated proved extraordinarily resiliant. Technical
obstacles to demonstrating that cells alone effect human immune
responses in Metchnikoff s original sense proved insurmountable.
While there were sporadic attempts to decisively refute the
humoral doctrine with laboratory evidence, neither in vitro nor in
vivo studies of white cell function produced results that proved
'naturally anomalous' (Humphreys, 1968). The various studies of
white cell transplantation including Murphy's (1924) chick
embryo studies and Landsteiner and Chase's (1942) and Chase's
(1945) classic animal studies were unconvincing for two primary
reasons: 1) the cells were not studied in situ in an integrated
organism; 2) it was not clear that the cells did not produce antibody after transfer. The latter drawback for decisively refuting
humoral theory was transmuted into an apparent advantage for
identifying the cellular source of antibody. Apparently on the
basis of various animal studies, Chase (1951) and others (e.g.,
Harris, 1954) argued that lymphocytes produce antibody, while
Kouluch (1947) pointed to the plasma cell. Hence, the focus
shifted from determining whether cells alone mediate immune
responses a la Metchnikov to the cellular source of antibody
without a resolution to the first enigma. The understandably
confused state of the field is mirrored in the statement by Harris
that "Because of the wide variation in experimental conditions
and cytologic emphasis in the studies in this field, the total mass
of data available thus far does not lend itself to unifying
generalizations" (Harris, 1956, p. 128). The identification of a
natural anomaly to the humoral theory nonetheless led not only to
elucidation of the cellular source of antibody but also to the first
theory of the cellular mechanisms underlying both classical
immune responses and delayed type hypersensitivity reactions; it
was the dissociation of classical immunity from allergy that led to
the two limb theory of the immune system.
A crucial event occurred in 1952, when O. Bruton described a
human case characterized by near absence of antibody, agammdglobulinemia. R.A. Good examined a series of agammaglobulinemic patients between 1954 and 1956. While the patients
had no plasma cells in their bone marrow, they had normal
L
On the Locus of Medical Discovery
195
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
lymphocyte populations in their blood, lymph nodes, and spleen.
Agammaglobulinemic patients were vulnerable to a wide variety
of extracellular bacterial pathogens but were resistant to most
intracellular pathogens and viruses. While the patients were
unable to mount any antibody-mediated responses to antigenic
challenge, their lymphocyte populations would expand and they
were capable of normal bacterial type hypersensitivity to tuberculosis and other intracellular microbes. From those phenomena,
Good 'deduced7 several tenets of cellular immunology including:
1) cellular immune responses are alone effective without the
mediation of antibody at any step (i.e., the humoral theory is
false); 2) it is the plasma cell that is the source of antibody; 3)
resistance to extracellular pathogens is mediated by antibody
produced by plasma cells; and 4) both delayed type hypersensitivity and resistance to viral infections are mediated by lymphocytes. The third and fourth statements are the foundation of
the two limb theory of the human immune system. In 1956, Good
was already arguing that: 'The capacity of patients with agammaglobulinemia to develop bacterial-type hypersensitivity... dissociates bacterial type hypersensitivity from the classical immune
response../' (1956a, p. 145). In the same year, Good said that:
"From these studies our tentative conclusion has been that
mechanisms other than those involving circulating antibody must
play an important role in the development of resistance to virus
disease. Our demonstration that the capacity to develop bacterial
type allergy is intact in these patients renders this phenomenon
..."(1956b, p. 158).
Later in the discussion, based on the fact that agammaglobulinemic patients responded normally to tests of delayed
bacterial allergy with the standard intracellular pathogens, Good
posited that "It is possible that these patients resist infection with
intracellular bacterial pathogens in a manner and with an effectiveness similar to that with which they resist virus infections (1956b,
p. 159). Good also wrote that "Certainly our studies are compatible with the concept that lymphocytes have an important if
not determing role in the production of the state of delayed
bacterial type allergy..." (1956b, p. 142). Good (1960) provided
further evidence that the lymphocytes mediate cell-mediated
immunity to intracellular bacterial and viral pathogens by historical and contemporary studies of Hodgkin's disease patients (in
which lymphocytes are abnormal). His team (1960, p. 191) redocu-
196
Brandon P. Reims
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
mented that the antibody system is functional in Hodgkin's
disease and that, therefore, the patients suffer from deficient
lymphocyte-mediate immune responses. He reviewed several
immunologic studies of Hodgkin's patients (e.g. Schier, 1956) that
demonstrated both that they fail to respond to intracellular
pathogens with delayed type hypersensitivity reactions and that
they are highly vulnerable to infection with such pathogens
including bacteria and fungi.
By I960, therefore, Good had discovered that the human
immune system is comprised of at least two distinct cellular limbs:
1) the antibody-mediated plasma cell limb that clears extracellular
pathogens; and 2) the cell-mediated lymphocyte limb that
eliminates intracellular pathogens including viruses and fungi. It
was at that time that Good heard of Glick's observation that
bursectomy in the neonatal period effects immunity in chickens.
This finding resolved a technical (and sociological) problem for
Good's team: how to create an 'animal model' of immunodeficiency. Intrigued by a prior natural experiment in which
a thymic tumor had been linked to broad based immunodeficiency, and assuming that the bursa is homologous to
the mammalian thymus, they began removing the thymus from
rodents in the neonatal period. These studies 'confirmed' their
hypothesis that the thymus is essential to both cellular and
humoral immunities. Miller (1961), who was well aware of the
natural experiment that had stimulated Good's work, independently discovered an immunoglical role for the thymus. He was
able to share in the credit by performing convincing animal
experiments. Neither Good nor Miller were able to determine
whether the thymus processes both the plasma cell series and the
lymphocyte series or only one limb. DiGeorge (1968) found that
the thymus exclusively produces lymphocytes (by studying
patients born without a thymus). While the ontogeny of the two
limbs of the human immune system is still not altogether clear,
both cellular limbs likely arise from bone marrow stem cells, with
the lymphocyte-mediated (T-cell) limb being processed exclusively in the thymus. While Good has clear priority in the
discovery of the two limb theory, Warner and Szenberg (1962) did
propose the notion of dissociation of immunological responsiveness. They dramatized a two limb theory by linking it to distinct
organs in the body of the chicken (bursa and thymus), although
contradicting the basic cellular mechanisms that had led Good to
r
L
On the Locus of Medical Discovery
197
D. Discovery of Anti-Manic Action of Lithium
Among the better known tales of serendipitous laboratory discovery in pharmacology is the identification of the anti-manic
action of lithium by John Cade. Cade's (1949) discovery is
generally attributed to 'accidental' observations which occurred
during experiments on his pet guinea pigs. A fairly typical account of the discovery appears in a paper read by a Dr. G.P.
Hartigan before the Royal Medicopsychological Society of Canterbury in 1959. Hartigan (1984), who was among the first physicians
to utilize lithium for treating bipolar disorders, said:
Some Australian physiologists, working on some recondite project whose exact
nature I regret I am unable to recall, found it expedient to introduce a lithium salt
into the peritoneal cavities of guinea pigs. It was observed that for some hours
after this outrage the animals became thoughtful and preoccupied. This really
seems hardly surprising, but the phenomenon prompted the Australian
psychiatrist Cade to the use of the substance therapeutically in a small group of
excited psychotics. The results were unexpectedly gratifying ... (1984, p. 183).
While F.N. Johnson has attempted to put Cade's experimental
findings in a more realistic historical context, he too has contended that the Cade's discovery of the antimanic action of
lithium was a result of the celebrated laboratory accident. Johnson
wrote: "Instead of killing the guinea pigs, both lithium urate and
lithium carbonate made the animals less timid, calmer, and less
responsive to stimulation. Cade had discovered the calming effect
of the lithium ion. He wondered whether it might have the same
effect on patients that it had in guinea pigs" (1985, p. 4). In order
to appreciate the nature of his discovery, it is necessary to understand the received view of psychopharmacology within which
Cade has to contend.
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
postulate the theory in the first place. In order to reinstate the
cellular basis of the two limb theory, Good's team carried out a
series of chicken experiments that 'refuted' Warner and Szenberg's and 'proved' Good's theory. The fact that various investigators used Glick's chicken system to 'prove' divergent theories
of immune function supports the notion that laboratory
experiments are often performed to dramatize ideas.
198
Brandon P. Reims
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
Despite abundant clinical evidence from the 19th century
literature that chemotherapy affects the mood of human beings,
the dogma that chemotherapy does not induce mood altering side
effects that might be exploited for psychiatric indications held
sway throughout the first half of the 20th century. Cade's discovery grew out of his awareness of the 19th century clinical
literature. F.N. Johnson (1984) has traced Cade's rationale for
lithium therapy to Garrod's work on the 'uric acid diathesis'
theory of disease that was often espoused in the 19th century.
Garrod (1859) postulated that many diseases including gout and
rheumatism are caused by an excess of uric acid in the
bloodstream. He attempted to explain the symptomatology of
various diseases by reference to uric acid deposition in the joints
and/or internal organs. Based on simple experiments on 'gouty7
joints in vitro, in which Garrod demonstrated the ability of lithium
salts to dissolve the ostensible concretions of uric acid, he argued
that lithium would prove therapeutically valuable for gout, renal
calculi, and other diseases. In fact, he and his followers used
lithium salts to treat the various conditions then classified under
the rubric of 'gout'. Garrod included under that heading a wide
variety of ill-defined disorders of mood including "gout retroceding to the head," further noting that "when retrocedent gout
attacks the head, apoplexy is commonly induced, but maniacal
symptoms occasionally arise" (1859, p. 441). He also referred to
gouty mania and complete mental derangement. Garrod found
that treatment with lithium cured many of the symptoms of such
disorders including their psychic manifestations, leading to
increased feelings of well being. Johnson (1984) hypothesizes that
Cade attempted to explain manic psychoses in terms of the uric
acid diathesis framework.
Indeed, Cade's experiments with the urine of manic patients
lend credence to Johnson's suggestion. Cade had been looking for
a toxin in the urine of manic patients and had been using guinea
pig tests as an assay. He began to think that uric acid might either
be the sought-after toxin or an enhancer of the toxin's action. In
order to inject uric acid into the guinea pigs, however, he had to
dissolve it. In order to dissolve it, he had to add lithium salts to
produce lithium urate, the most soluble of the urates. Cade
injected guinea pigs with lithium urate and then lithium alone.
The reason Cade injected the guinea pigs with lithium alone is not
clear from his technical report. He simply reports that "A notewor-
L
On the Locus of Medical Discovery
thy result was that after a latent period of about two hours the
animals, although fully conscious, became extremely lethargic and
unresponsive to stimuli for one to two hours before once again
becoming normally active and timid" (Cade, 1949, p. 350). Most
accounts of the discovery begin with this 'accidental' observation
and suggest that the guinea pig results were the crucial evidence
that led Cade to try lithium on manic patients. More precisely,
such accounts suggest that lithium's CNS depressant effects in
guinea pigs led Cade to hypothesize that lithium might also have
CNS depressant effects in man. Cade himself testified against such
a simplistic interpretation in his original report when he wrote: "It
may seem a long distance from lethargy in guinea pigs to the
excitement of psychotics, but as these investigations had commenced in an attempt to demonstrate some possibly excreted
toxin in the urine of manic patients, the association of ideas is
explicable" (Cade, 1949, p. 350). Cade seems to be saying that he
had been looking for a treatment for mania and that the image of
guinea pigs rendered inactive by lithium impressed him, but he
leaves the connection between those two 'ideas' to the
imagination. Just which hypotheses Cade had in mind that
connected lithium effects on guinea pig behavior to potential
effects on man is the primary enigma of Cade's work.
In interpreting his guinea pigs' behavior, he had to have been
guided by two classes of hypotheses: 1) hypotheses relating
guinea pig behavior to human consciousness; and 2) hypotheses
about the potential effects of lithium on human consciousness.
The second class will be examined first because it will likely shed
light on the first class. Careful linguistic analysis of Cade's statements about the guinea pig results is revealing. Cade's statement
that the guinea pigs became 'extremely lethargic' following
injection of lithium is anthropomorphic. He was imputing the
human feeling of lethargy to guinea pigs, based on his observation
of their unresponsiveness to stimuli. In the following paragraph,
he refers to the unresponsiveness in more-or-less behavioral
terms, as the "sedative effect." His prior usage of the term
'lethargy7 reveals the extent to which his assessment of the slowmoving guinea pigs constitutes 'projection' (in the psychoanalytic
sense). In Webster's Third International Dictionary, the second
definition of lethargy is the more obviously cognitive: "the quality
or state of being lazy or indifferent." The fact that it is now known
that the guinea pigs were more likely in a state of acute lithium
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
.
199
200
Brandon P. Reims
Having found that not only lithium urate, but lithium carbonate too, produced
effects on guinea-pigs Cade unhesitatingly transferred his attentions to hospitalized patients. Why was he so positive in taking such a decision? Did he have any
reason, other than the results of his guinea pig studies, for believing that a
successful outcome in his patients was likely? Probably not - at least not in a
formal, explicit way; but it seems hardly likely that the various claims which had
been put forward for over a hundred years for the therapeutic benefits of lithium
in a wide range of disorders, including mental affections, were either totally
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
intoxication (Kline, 1969) than sedation clarifies the extent to
which Cade was projecting on the guine pigs. Cade must have
brought to the guinea pig experiment the expectation of 'seeing'
an alteration in the mentation of the guinea pigs (which may have
been subconscious). So he must have had a hunch that lithium
had mood-altering properties prior to undertaking the experiment. He must have had at least a vague working hypothesis
about the potential effect of lithium on human cognition prior to
the experiment (or he projected on the imagined scene of the
guinea pigs at some later time). The most likely source of his
hypothesis is the abundant clinical literature on the side-effects of
lithium in the clinical literature from the 19th and early 20th
centuries. In addition, lithium was actually used expressly as an
antidepressant by physicians prior to Cade's time, and he may
have known some of them personally.
Indeed, the first page of Cade's classic paper reviews the long
standing clinical evidence that lithium exerts powerful effects on
the human CNS. He begins by referring to Garrod's work in 1859,
without explicitly citing its relevance, and then leaps to the 20th
century, when lithium treatment was often attended by sideeffects such as "mental depression, nausea, and giddiness..."
(Cade, 1949, p. 349). He continues by pointing out that "Culbreth
(1927) says of lithium bromide that it is the most hypnotic of the
bromides...". Based on the clinical evidence that lithium affects
human mood, coupled with his wish that it prove a useful treatment for human manic psychoses, Cade hypothesized that lithium
would prove useful as a therapeutic agent for mania. His discovery grew out of the recognition that the 19th century clinical
literature on the mood-altering effects of lithium clearly refuted
the early 20th century doctrine that chemotherapy does not
induce therapeutically-exploitable side effects in man. Johnson
edges toward this notion in his seminal account but stops short,
when he writes:
On the Locus of Medical Discovery
201
unknown to Cade or failed to influence his thought, at least in a general way
(1984, p. 43).
IV. DISCUSSION AND CONCLUSION
Practically all of the leading 20th century philosophers of science
including Popper, Kuhn, Feyerband, Laudan, Pierce, Hanson and
Lakatos recognize that understanding the response to empirical
anomalies is central to any account of scientific process (Laudan,
1988, p. 21). Kuhn maintains that in practice scientists ignore
anomalies until they pile up and generate a crisis for the paradigm
which leads to generation of new but incommensurable theory.
Laudan and Feyerband contend that many anomalies are not
identified as such until the discovery of new theory that explains
them. Much controversy has appeared in the literature over
whether the usual sequence is anomaly identification-crisis-new
theory or new theory-recognition of anomaly-crisis. While
Nicholas' analysis of the early quantum theory weighs in favor of
the latter schema, Humphreys had already explained why both
views are essentially correct. The sequence is irrelevant to the
central role of anomaly in assessing a theory's explanatory power,
according to Humphreys. While Laudan (1974) weighs empirical
anomalies as to the degree of 'epistemic threat' they pose to
reigning doctrine, Humphreys argues that the notion of anomaly
is strictly qualitative. He defines a unique class of observations
that are "objectively in need of explanation" as "natural
anomalies". Humphreys notion of natural anomaly as applied to
intertheoretical explanation edges toward the ideal of the crucial
experiment, at least in the sense of decisive refutation modus
tollens. He argues convincingly that Duhem (1954) had misinterpreted the approximative character of physical theory in
concluding that decisive refutation is impossible. The implications
for biomedical discovery of Humphreys' refinement of Duhem's
thesis are intriguing. Recalling Duhem's (1954, p. 180) argument
that the relatively simple logical structure of biomedical theory
allows for true experimental cruci, it is apparent that biomedical
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
The fact that later historians regularly cited the guinea pig results
as the source of discovery - and not prior clinical evidence indicates that animal experiments are more memorable than are
clinical hypotheses and that the former dramatize the latter.
202
Brandon P. Reims
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
discovery ought be more intelligible than physical discovery. Not
only is the logical content of biomedical theory usually much
simpler than physical theory, but the epistemic warrant for
refutation on the basis of anomaly is simpler to determine for
another reason: the crucial experiments are largely or entirely
nonexperimental. The natural anomalies that derive from human
pathology are 'natural' not only in Humphreys' sense of the term
but also in the sense of 'nonartifactual'. Nature knows no artifact.
The 'experiments performed for us by Nature' may constitute the
closest approximation of the Baconian ideal of crucial experiments
in the history of science. In this light, the process of biomedical
discovery is much simpler than might be inferred from the historiographic anomaly documented in the "Introduction". The
process of scientific discovery in general is increasingly comprehensible, as a result of refined analysis of the regularities that
underly scientific process.
Due to the inability to identify an inherent logic of discovery,
there is increasing emphasis on the semantical evolution of
theories (e.g., Kitcher, 1988). The illusory immanence of logic is
slowly being replaced with the real continuity found in another
cognitive dimension: i.e., the psycholinguistic realm. Nickles'
(1985) approaches an explicit recognition of that dimension in his
espousal of 'discoverability7. Whether a given discovery is
achieved by deductive or inductive means is irrelevant to Nickles'
definition of discoverability; it is the semantical continuity with
prior theory and data that is emphasized. Nickles writes:
"Generative justification usually requires setting out what
amounts to a rationally reconstructed discovery path, a derivation
of the new claims from data and theory already established.
\
Clearly, it is not discovery in the sense of original conception of an
I
idea which is important to justification here. Rather, justifying a
>
claim establishes its 'discoverability' in the sense that regardless of
[
how or why it was first thought of - it could have been discovered
\
in the rationally specified manner..." (Nickles, p. 195). Something
]
akin to Nickles' notion of discoverability was not only applied in
j
the analyses above but is increasingly used by philosophers and
historians to rationally reconstruct discoveries achieved by
luminaries such as Darwin (Kitcher), Harvey (Reines), Leverrier
^
(Humphreys), Kepler (Lugg) and Newton (Hattiangadi). Grmek's
J
(1980) neologism omnis theoria et theoria captures the essence of the
(
discoverability thesis. This growing continuist literature conr
1
\
203
stitutes a clear epistemic threat to the accidental discovery model
of scientific process. In a statement indicative of the growing
suspicion of the chance discovery genre, Gruber (1980) writes:
"renewed interest in discovery processes means that we must get
beyong 'Aha!' Once we free ourselves from the view that discovery is summed up and wrapped up in the mysteries of the
'Eureka Experience/ the direction to which we naturally turn is
the conceptualization of scientific thought as protracted, purposeful, constructive work." (Gruber, p. 113). Discoverability in the
biomedical sciences is generally a temporally inductive but
logically deductive process that consists of three successive
semantic steps: 1) internalization of the semantic nuances of prior
theory, arguments, and data; 2) identification of a natural clinical
anomaly; and 3) invention of an hypothesis that accounts for the
anomaly.
Additionally, the notion of discoverability clarifies the role of
laboratory model system studies in biomedicine. With generative
justification of new hypotheses as the rule, the traditional
requirement for laboratory 'verification' lingers as a sociological
artifact. It derives from the widely held view among biomedical
scientists that 'experimental method' is synonymous with scientific method. Peller (1954) observed that the results of model
system studies in modern biomedicine command the respect
afforded a quotation from Aristotle in the Renaissance. Hence, the
dramatic impact of laboratory experiments derives not only from
the visually arresting character of animal and in vitro studies but
also the faith in their results. The notorious variability of the
results of such studies ensures that virtually any clinical
hypothesis can be dramatized with an appropriate choice of
species, strain, and experimental protocol.
Assuming that this analysis of the de facto relationship between
clinical and laboratory studies is correct, it is therefore apparent
that both bench scientists and clinical investigators discover in the
same way: by identifying and explaining clinical anomalies. This
explains both the logical and historiographic anomalies identified
in the "Introduction". Most pedagogic medical theories do refer to
a large extent to 'animal models' and are therefore in some sense
based on analogical extension. For instance, Schafmer indicates
that part of the two-component theory of the immune system is
the negative statement "that there is no mammalian 'bursa'"
(1980a, p. 84). He nonetheless contends that the two component
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
t
On the Locus of Medical Discovery
204
Brandon P. Reines
NOTE
* The research for this paper was supported by a grant from the Medical
Research Modernization Committee, Inc.
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
theory originated in studies of the bursa. The theory is denoted
'biomedical' which connotes that the referent of the theory is at
least mammalian. On strictly logical grounds, one would assume
that a theory of mammalian physiology could not possibly
originate in studies of nonmammaUan organic function. Then, if
the identity of the two component theory is clarified as medical, as
documented by the historical analysis above, then the referent is
not the mammal at all but the human being. In this light, the
inclusion of nonhuman referents in most textbook medical
theories appears gratuitous. Further support for this view derives
from the recent work of Patel, Evans and Groen (1989). In a
fascinating study, these authors determined that training in the so
called basic biomedical studies actually interfered with the ability
of medical students to accurately diagnose human diseases. They
were forced to entertain the hypothesis that "the basic biomedical
sciences and the clinical sciences might constitute theories of
distinct domains" (Patel, p. 32). The linguistic fact that the subject
matter of the basic biomedical sciences is distinct from that of the
clinical sciences, and the practical problems induced by the 'cliniclaboratory gap7, led Blumenson and Bross (1973) to explore the
potential of mathematical analysis to bridge the gap. In contrast to
the theories explored by Schaffner (1980a), medical theories
discovered by such medical statisticians as I. Bross and S. Peller
are unabashedly based on human data. Predictably, however,
Peller was unable to convince the medical world of the verity of
his 'principle of inverse association' to a large extent because it
was based on 10,000 case studies and not animal experiments
(Peller, 1979). Bross (1990) has encountered similar difficulties, as
has H. Heimlich (1990). These sociological facts add evidential
weight to the notion that laboratory studies constitute Aristotelian
experiments that dramatize clinical hypotheses (Reines, 1986). It is
hoped that this analysis sheds light on the process of medical
discovery and means of fostering it in the future.
On the Locus of Medical Discovery
205
REFERENCES
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
f
Beecher, H.K.: 1960, Disease and the Advancement of Basic Science, Harvard
University Press, Cambridge, Massachusetts.
Beeson, P.B.: 1979, The gain of knowledge about a disease: Hepatitis', American
Journal of Medicine 68,366-370.
Bernard, C: 1856, Memoir on The Pancreas and on the Role of Pancreatic Juice in
Digestive Processes, Paris, France.
Bernard, C: 1957, An Introduction to the Study of Experimental Medicine, Dover
Publications, Inc., New York, New York.
Blumenson, L. and Bross, I.D.J.: 1973, 'Use of a mathematical model to bridge the
clinic-laboratory gap: Local spread of endometrial cancer', Journal of Theoretical
Biology 38,397-411.
Brodie, B.: 1823, Quarterly Journal of Science 14,341-344.
Bross, I.: 1972, 'Scientific strategies in human affairs: Use of deep mathematical
models', Transactions of the New York Academy of Sciences 34(3), 187-199.
Bross, I.: 1980, Scientific Strategies to Save Your Life, Marcel Dekker, New York,
New York.
Bross, I.: 1990, personal communication.
Brunner, J.: 1683, Experimenta Nova Circa Pancreas, apud H. Wetstenium, Amsterdam, Holland.
Bruton, O.C.: 1951, 'Agammaglobulinemia', Pediatrics 9,722-728.
Buchanan, G., Cowan, J. and M'Ghie, J.: 1858-59, Review of 'Lecons de
Physiologies Experimentale appliquee a la Medecine, faites au College de
France, 1856', Glasgow Medical Journal, 6,81-98.
Cade, J.: 1949, 'Lithium salts in the treatment of psychotic excitement7, The
Medical Journal of Australia 36,349-352.
Cade, J.:1970, The story of lithium', in Discoveries in Biological Psychiatry,
Lippincott, Philadelphia, Pennsylvania.
Cadeddu, A.: 1987, 'Pasteur et le cholera des poules: Revision critique d'un recit
historique', History and Philosophy of the Life Sciences 7,87-104.
Canguilhem, G.: 1989, The Normal and the Pathological, Zone Books, New York,
New York.
Cannon, W.B.: 1945, The Way of an Investigator, Hafner Press, New York, New
York.
Caplan, A.: 1986, 'Exemplary reasoning? A comment on theory structure in
biomedicine', The Journal of Medicine and Philosophy 11,93-105.
Chase, M.W.: 1945, The cellular transfer of cutaneous hypersensitivity to
tuberculin', Proceedings of the Societies of Experimental Biology and Medicine 59,
134.
Chase, M.W.: 1951, 'Development of antibody following transfer of cells taken
from lymph nodes of sensitized or immunized animals', Federation Proceedings
10,404.
Comroe, J.: 1977, Retrospectroscope: Insights into Medical Discovery, Von Gehr Press,
Menlo Park, CA.
206
Brandon P. Reims
L
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
Dale, H.H.: 1948, 'Accident and opportunism in medical research', British Medical
journal 2,451-455.
DeKruif, P.: 1926, The Microbe Hunters, Harcourt Brace, New York, New York.
DiGeorge, A.: 1968, 'Congenital absence of the thymus and its immunological
consequences', in R.A. Good (ed.), Immunologic Deficiency Diseases in Man,
National Foundation Press, New York, New York.
Dubos, R.: 1959, The Mirage of Health, Harper and Row, New York, New York,
pp. 105-106.
Duhem, P.: 1954, The Aim and Structure of Physical Theory, Philip Wiener (trans.),
Princeton University Press, Princeton, New Jersey.
Eijkman, C: 1965, 'Antineuritic vitamin and beriberi', Nobel Lectures in Physiology
and Medicine, Elsevier Press, Amsterdam, Holland.
Englehardt, H.T., Jr.: 1979, 'Philosophical problems in biomedicine: Towards a
philosophy of medicine', in P.D. Asquith and H.E. Kyburgh, Jr. (eds.), Current
Research in Philosophy of Science, Philosophy of Science Association, East
Lansing, Michigan.
Fleming, A.: 1929, 'On the antibacterial action of cultures of a penicillium, with
special reference to their use in the isolation of B. influenzae', British Journal of
Experimental Pathology 10,226-236.
Garrod, A.L.: 1859, The Nature and Treatment of Gout and Rheumatic Gout, Walton
and Maberly, London, England.
Garrod, A.E.: 1919, 'The laboratory and the ward', in C.L. Dana (ed.),
Contributions to Medical and Biological Research Dedicated to Sir William Osier,
Paul Hoeber, New York, New York.
Geison, G.: 1979, 'Divided we stand: Physiologists and clinicians in the American
context', in C. Rosenberg and M. Vogel, (eds.) The Therapeutic Revolution,
University of Pennsylvania Press, Philadelphia, Pennsylvania.
Geison, G.: 1978, Michael Foster and the Cambridge School of Physiology, Princeton
University Press, Princeton, New Jersey.
Glick, B., Change, T.S. and Jaap, R.G.: 1956, The bursa of fabricius and antibody
production', Poultry Science 35,224-226.
i
Good, R.A.: 1954a, 'Agammaglobulinemia - A provocative experiment of
>•
nature', Bulletin of the University of Minnesota Medical Foundation, 26,1-19.
I
Good, R.A.: 1954b, 'Absence of plasma cells from bone marrow and lymph nodes
i
following antigenic stimulation in patients with agammaglobulinemia', Revue
•
d'Hematologie 9,502-503.
1
•• f
Good, R.A.: 1956a, 'Disturbances in gamma globulin synthesis as experiments of
nature', Pediatrics 18(10), 109-417.
Good, R.A.: 1956b, 'Morphological basis of the immune response and hypersensitivity', in H.M. Felton (ed.), Host-Parasite Relationships in Livin Cells, Charles
C. Thomas, Springfield, Massachusetts.
Good, R.A. et ah: I960, 'An investigation of Hodgkin's disease with respect to the
problem of homotransplantation', Annals of the New York Academy of Sciences
87,187-195.
On the Locus of Medical Discovery
207
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
Good, R.A., Peterson, R. and Cooper, M.: 1965, 'Delineation of the thymic and
bursal lymphoid systems in the chicken', Nature 205,144-146.
Good, R.A.: 1968, 'Keystones', Journal of Clinical Investigation 47,1466-1471.
Gorovitz, S. and Maclntyre, A.: 1976, 'Toward a theory of medical fallibility', The
Journal of Medicine and Philosophy 1,51-71.
Grmek, M.D.: 1980, 'A plea for freeing the history of scientific discoveries from
myth', in M.D. Grmek and R.S. Cohen (eds.), On Scientific Discovery, D. Reidel
Publishing, Dordrecht, Holland.
Gruber, H.: 1980, 'The evolving systems appraoch to creative scientific work:
Charles Darwin's early thought', in T. Nickles (ed.), Scientific Discovery: Case
Studies, D. Reidel Publishing, Dordrecht, Holland.
Harris, S. and Harris, T.N.: 1954, 'Studies on the transfer of lymph node cells: I.
appearance of antibody in recipients of cells from donor rabbits injected with
antigen', Journal of Immunology 72,148.
Harris, T.N. and Harris, S.: 1956, 'The genesis of antibodies', American Journal of
Medicine 20,114-132.
Hartigan, G.P.: 1984, 'Experiences of treatment with lithium salts', in F.N.
Johnson (ed.), The History of Lithium Therapy, The Macmillan Press, London,
pp. 183-187.
Hattiangadi, J.: 1980, 'The vanishing context of discovery: Newton's discovery of
gravity', in T. Nickles (ed.), Scientific Discovery, Logic, and Rationality, D. Reidel
Publishing, Dordrecht, Holland.
Heimlich, H.: 1990, personal communication.
Holmes, F.L.: 1974, Claude Bernard and Animal Chemistry, Harvard University
Press, Cambridge, Massachusetts.
Humphreys, W.: 1968, Anomalies and Scientific Theories, W.H. Freeman Company,
San Francisco, California.
Johnson, F.N.: 1984, The History of Lithium Therapy, Macmillan Press, London.
Johnson, F.N.: 1985, The early history of lithium therapy', in R.O. Bach (ed.),
Lithium: Current Applications in Science, Medicine, and Technology, John Wiley
and Sons, New York, New York.
Johsnon, S.L.: 1970, The History of Cardiac Surgery, Johns Hopkins Press, Baltimore, Maryland.
Kitcher, P.: 1985, 'Darwin's achievement', in N. Rescher (ed.), Reason and
Rationality in Science, University Press of America, Washington, DC.
Kitcher, P.: 1989, 'Explanatory unification and causal structure', in P. Kitcher and
W. Salmon (ed.), Scientific Explanation, University of Minnesota Press,
Minneapolis, Minnesota.
Kline, N.S.: 1969, 'Lithium: The history of its use in psychiatry', Modern Problems
in Pharmacopsychiatry 3, 75-92.
Kouluch, F. et al.: 1947, 'The reticulo-endothelial origin of the bone marrow
plasma cells in hypersensitive states', Journal of Laboratory and Clinical Medicine
32,749.
Kuhn, T.S.: 1962, The Structure of Scientific Revolutions, University of Chicago
Press, Chicago, Illinois.
208
Brandon P. Reims
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
Landsteiner, K. and Chase, M.W.: 1942, 'Experiments on transfer of cutaneous
sensitivity to simple compounds', Proceedings of the Societies of Experimental
Biology and Medicine 49,688.
Lange, C: 1886, Om Periodiske Depressionstilstande og deses Patogenese, Jacob Lunds
Forlag, Copenhagen.
J
Latour, B.: 1988, The Pasteurization of France, Harvard University Press,
\
Cambridge, Masssachusetts.
Laudan, L.: 1974, Progress and Its Problems, Routledge and Kegan Paul, London,
England.
Laudan, R. et ah: 1988, Scrutinizing Science: Empirical Studies of Scientific Change,
Kluwer Academic Publishers, Dordrecht, Holland.
Liggins, G.C.: 1969, 'Premature delivery of foetal lambs infused with glucocorticoids', Journal of Endocrinology 45,515-523.
Lloyd, E.A.: 1833, Discharge of Fatty Matter from the Bowels and Contracted State of
the Duodenum, quoted by Elliotson in Medico-chirurgical Transactions of
London, Vol. XVIII.
Lugg, A.: 1985, The process of discovery', Philosophy of Science 52,207-220.
Maulitz, R.: 1979, 'Physician vs. bacteriologist: The ideology of science in clinical
medicine', in M. Vogel and C. Rosenberg (eds.) The Therapeutic Revolution,
University of Pennsylvania Press, Philadelphia, Pennsylvania.
Mclean, J.: 1959, The discovery of heparin', Circulation 19, 75.
McQuarrie, I.: 1944, The Experiments of Nature and Other Essays from The Porter
Lectures, University of Kansas Press, Lawrence, Kansas.
Medawar, P.: 1969, Induction and Intuition in Scientific Thought, American
Philosophical Society, Urbana, Illinois.
Medawar, P.: 1975, 'Scientific method in science and medicine', Perspectives in
Biology and Medicine.
Miller, J.F.: 1961, 'Immunological function of the thymus', Lancet 2, 748-749.
Mueller, A.P., Wolfe, H.R. and Meyer, R.K.: 1960, 'Precipitin Production in
Chickens', Journal of Immunology 85,172-179.
Murphy, J.B.: 1926, The Lymphocyte in Resistance to Tissue Grafting, Malignant
Disease, and Tuberculous Infection: An Experimental Study, Rockefeller Institute,
monograph no. 21.
Nickles, T., 1985, 'Beyond divorce: Current status of the discovery debate',
Philosophy of Science, 52,177-206.
Nuland, S.: 1988, Doctors: The Biography of Medicine, Alfred Knopf Publishers,
New York, New York.
Paget, S.: 1848, The pancreatic juice', in W. Kirkes (ed.), Handbook of Physiology,
Taylor, Walton, and Maberly Publishers, London.
Pasteur, L.: 1880, 'De 1'attenuation du virus du cholera des poules', comptes
rendus de I'Academie des Sciences 91,673-680.
)
Patel, V. et ah: 1989, 'Biomedical knowledge and clinical reasoning', in V. Patel
^
and D. Evans (eds.), Cognitive Science in Medicine, Massachusetts Institute of
i
Technology Press, Cambridge, Massachusetts.
On the Locus of Medical Discovery
209
f
Downloaded from http://jmp.oxfordjournals.org/ at Pennsylvania State University on September 11, 2016
Peller, S.: 1954, Cancer in Man, International Universities Press, New York, New
York, p. 47.
Peller, S.: 1967, Quantitative Research in Human Biology and Medicine, John Wright
and Sons, Ltd., Bristol, England.
Peller, S.: 1979, Not In My Time, Philosophical Library, New York, New York.
Reines, B.P.: 1986, Cancer Research on Animals: Impact and Alternatives, National
Antivivisection Society, Chicago, Illinois.
Reines, B.P.: 1990, 'On the role of clinical anomaly in Harvey's discovery of the
mechanism of the pulse', Perspectives in Biology and Medicine 34(1).
Root-Bernstein, R.S.: 1989, Discovering, Harvard University Press, Cambridge
University Press, Cambridge, Massachusetts.
Rous, P.: 1945-48, 'Karl Landsteiner', Royal Society of London, Obituraries, Vol. 5.
Schaffner, K.: 1980a, Theory structure in the biomedical sciences', The Journal of
Medicine and Philosophy 5(1), 57-97.
Schaffner, K.: 1980b, 'Discovery in the biomedical sciences: Logic or irrational
intuition', in T. Nickles (ed.), Scientific Discovery: Case Studies, D. Reidel
Publishing Company, Dordrecht, Holland.
Schaffner, K.: 1980c, 'Reply to Maull', in T. Nickles (ed.), Scientific Discovery: Case
Studies, D. Reidel Publishing Company, Dordrecht, Holland.
Schaffner, K.: 1985, Logic of Discovery and Diagnosis in Medicine, University of
California Press, Berkeley, California.
Schaffner, K.: 1986, 'Exemplar reasoning about biological models and disease: A
relation between the philosophy of medicine and the philosophy of science',
The Journal of Medicine and Philosophy 11,63-80.
Schier, W. et ah: 1956, 'Hodgkin's disease and immunity', American Journal of
Medicine 20,94-99.
Warner, N. and Szenberg, A.: 1962, 'Effect of neonatal thymectomy on the
immune response in the chicken', Nature 196,784-785.