IDENTIFYING DISEASE IN THE PAST:

CUTTING THE GORDIAN KNOT

Andrew Cunningham
Department of History and Philosophy of Science
University of Cambridge

I: DISEASE IDENTITY

How should we historians approach the issue of the identity of disease in the past?
Can we legitimately identify past diseases? Can we legitimately identify past dis-
eases with present diseases? Can we legitimately identify particular epidemics in the
past? Can we legitimately talk of the evolution of diseases or pathogens? Can we
legitimately reach past people’s experience of disease by identifying what they were
suffering from? Are our attempts at retrospective diagnosis legitimate? Is retrospec-
tive diagnosis either possible or desirable?
These questions —surprisingly— virtually never crop up in the work of historians
of medicine, who for the most part assume that the identification of past diseases is
simply not a problem, since they assume the continuous identity of past diseases with
modern diseases. They just get on with identifying past diseases in modern terms –
whether it is logically, philosophically or historiographically possible or not. Hence
my use of the term ‘legitimately’ in my questions. Certainly we can make such iden-
tifications, and we do. But do they mean anything? Do they mean what we want
them to mean? Are they logical, sensible, and coherent things for us to do? Most
important of all, do they tell us anything at all about the past, or are they simply pro-
jections backwards of present-day issues and concerns?
Before we can explore these issues we have to ask a fundamental question: what
is disease? There has been a lot of confusion on this issue, based on misplaced and
un-inspected assumptions. What we can say is that, at its most fundamental, disease
is (1) an experience – an experience of debilitation, pain, suffering, together with (2)
the spontaneous appearance of non-customary phenomena with respect to the body,
such as spots, vomiting, sweating, aches, and (3) with outcomes of recovery, death or
disability. In this, disease is something that humans and animals have in common.
But there is one big difference. For unlike other animals, humans seem to insist on
seeking reasons or causes for disease: for its incidence, its origin, its course, its out-
come. Some of the reasons given in the past and in other present-day societies seem

Asclepio-Vol. LIV-1-2002 13
 ANDREW CUNNINGHAM

to us to be either unreasonable or irrational – something which prejudices them in our

eyes and could exclude them from our considerations. So I prefer to speak in terms of
causes, since this is a more inclusive term than reasons or rationales. It is this ‘cause’
dimension in human disease which means that disease is always experienced so-
cially, that it is not just a biological phenomenon but just as much a social phenome-
non. It is not just that we have to use language, itself essentially social, as our only
means to think and express ourselves about disease. It is rather that we seem to be
unable to talk about or even conceptualise disease without invoking cause in one
sense or another. Different societies, separated culturally by space or time, will have
different views as to what states constitute disease and what its causes are. But at all
times and places, in all societies, disease identity, and especially the cause dimension
of it, is going to be an expression of how people in that society think the world func-
tions. Cause, with respect to disease, is thus expressed in many ways and thought to
be many different kinds of thing, from djins and the evil eye, to humoral imbalance,
to germs and vitamin deficiency, to poor DNA and social disadvantage. For we can
only think about our experience of disease —as of anything else— in the terms and
categories of whichever particular society we are in.
For almost 150 years the dominant model of disease we have had in Western de-
veloped societies has been that many diseases are caused by some encounter of the
human body with some dangerous element (or a failed encounter with some neces-
sary element) ‘out there’ in Nature. This model applies primarily to our categories of
infectious disease and deficiency disease, but it is pertinent also to a large extent also
to degenerative diseases, which occupy a large part of our present-day disease land-
scape. As we also believe —since this is part of our modern scientific world-view—
that Nature is pretty consistent in its behaviour, and the elements in it (whether good
or bad with respect to humans) are pretty constant in their behaviour, so it feels obvi-
ous to us that these encounters with Nature or the environment are —and always
have been— constant too. Thus when talking about disease identity we tend to make
a little logical leap at this point, and assume that disease identity has been pretty
constant over time, and that diseases are themselves a-cultural and a-social, that they
have an identity quite separate from the social circumstances in which they are ex-
perienced. It seems obvious to us, looking through our scientific medicine spectacles,
that of course social interpretations of disease (and what ‘counts as’ a disease) do
and have varied from society to society, but it seems to us that these just express
greater or less success in coping with the underlying constant disease reality ‘out
there’ in Nature. As we assume that our own success in coping with disease has been
the greatest, we naturally take our models of disease identity as the final, and thereby
the only legitimate, models. In particular, for those diseases where we believe some
minute material living entity to be the cause, we believe we in effect encapsulate the
whole disease experience by talking about the encounter between microbe and hu-
man organism. So when we come to doing the history, when we come to trying to

14 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

identify past outbreaks of plague for instance, we assume that what we need is the
best modern thinking about the disease and its manifestations. Armed with this su-
preme form of knowledge we are able, we believe, to correctly identify outbreaks of
plague in the past, even down to pronouncing on the presence or absence of the bacil-
lus, and we correct the people of the past in their identifications of plague, telling
them when they were right and when they were wrong, since our form of knowledge
is clearly superior to theirs.
Yet in fact this is just our society’s way of thinking: true for us and our world, but
not necessarily true for other societies and other times.
Some of the large issues about the problems in making retrospective diagnosis
have been raised very ably in a recent extended critique of other historians’ ap-
proaches1. Using Ludwig Fleck’s analysis of the history of the concept of syphilis, it
is there made clear that, through all their many changes, disease concepts are always
social products. But, even more strikingly, it is shown that the maintenance of the
stability of disease concepts is also a social phenomenon, achieved by social rein-
forcement. As a necessary consequence of this social reinforcement, ‘the historicity
of [disease] concepts is necessarily eliminated’. This elimination of the historicity of
disease concepts is ‘a necessary by-product, within «popular science» of the modern
concept itself’. This is how diseases are attributed eternal reality by us: ‘the modern
concept is extended backwards in time: the disease as presently conceived is seen as
a permanent entity, and it is assumed that it can be diagnosed retrospectively’. In the
light of this account of the social construction of disease identity and of the perma-
nent being of diseases, it can now be quite easily understood why the assumption that
retrospective diagnosis is both possible and desirable, has so dominated the history of
medicine. Medical historians have turned to modern medical knowledge to help their
investigations of the past, and modern doctors have felt themselves specially well-
placed to make retrospective diagnosis themselves. But the assumption of the persis-
tence of disease identity is simply that: an assumption – and one which is not open to
proof or disproof, because of the incommensurability of old disease concepts with
new disease concepts.
Once we understand the sources of our unjustified assumption about the validity of
retrospective diagnosis, we can stop trying to do it. But where should we go from there?
The author of this particular critique recommends that instead of concentrating on dis-
eases (as real entities), we should instead concentrate on disease concepts (as thought
entities), and write the history of these2. However, this alternate approach —that we
should be studying the history of disease concepts— is also subject to fatal criticism,
primarily that concepts, as things thought, don’t actually have histories. Concepts are
————
1 WILSON, A. (2000), «On the History of Disease Concepts: The Case of Pleurisy», History of Sci-

ence 38, 271-319; the quotations are from page 275.

2 Ibid., especially the conclusions, pp. 303-6.

Asclepio-Vol. LIV-1-2002 15
 ANDREW CUNNINGHAM

the product or outcome or perhaps the elements of thinking; they do not have their
own histories separate from the thinking act. This thinking act, by contrast, certainly
does have a history because it is a human activity3.
The main thrust of the present paper is to follow this line and to offer an alterna-
tive approach by cutting through the Gordian knot of disease history. It will involve
turning our attention away both from diseases (the old way) and also away from
disease concepts (the proposed new way), and turning it instead towards how diagno-
sis happens. In other words, to people thinking and acting in particular cultures,
situations and times. It will be noted that looking at people thinking is not the same
as looking at their mental concepts, although of course mental concepts are involved.
Looking at people thinking and acting will, I believe, give us a properly historical
view of disease history, placing past disease firmly in the past, and interpreting that
past experience of disease in such a way that people of the present may empathise
with that past experience, but not turning it into some early version of modern dis-
ease and hence of modern experience. By making how diagnosis happens central to
our historical investigations, we are using the only sure thing we have, the only thing
which we can rely on. For it is by the act of diagnosis that disease identity is given or
established. The operations that humans perform in making diagnosis are not just the
key to disease identity, but the source of disease identity. The only identity disease
has is this operational identity. I see this as the equivalent of Alexander the Great’s
solution to the complexities of the Gordian knot. He simply sliced through it, rather
than seeking to disentangle it. If we concentrate on how diagnosis happens, we need
no longer worry about disentangling disease entities, disease concepts, linguistic and
conceptual incommensurability, germ evolution or anything else.
Diagnosis always proceeds operatively or operationally, by people asking and an-
swering one or more specific questions about the patient and his or her affliction. But
these sequences of operations differ from one medical system to another. And while
these operative sequences are what actually give the disease identity at all times and
in all cultures, we all tend to think that they are merely procedures enabling us to
recognise the disease, with its (supposed) pre-existing identity.
Here I shall make two points about the identity of disease as it affects the practice
of the historian, and which are built on this principle that how diagnosis happens is
the source and key to disease identity at any time and in any society.
The first of these two points is extremely simple. It is that the identity of any disease
is made up of a compound of elements, of which the biological or medical is only one,

————
3 This criticism of the pursuit of a history of concepts of disease, I hope to justify at length in the

larger work of which the present essay is a preliminary part. A second problem with trying to write a
history of disease concepts is that when we ask ‘What is a disease concept a concept of?’ the answer still
has to be disease. So disease —the very category which is problematic here— remains the focus in this
approach, and its ontological status still goes unexplored, or appears to be begged.

16 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

and sometimes the least important one. This is to reiterate the point that disease is al-
ways experienced socially.
The second point is as simple as the first one. It is that you die of what your doc-
tor says you die of. Your cause-of-death certificate is not negotiable. While this
might seem a reasonable thing to say about people dying today, I want to argue that it
also applies to everyone in the past. They died from what their doctors said they died
of. Their cause-of-death certificates (as it were, for of course such certificates are
very modern and very western) are equally not negotiable, neither by the modern
medic, whether clinician, pathologist, epidemiologist or psychiatrist, nor by the modern
historian.
Of course, in the matter of specifying cause of death, the doctor or other practitio-
ner (even a witchdoctor) is simply a bystander who has been given or conceded spe-
cial authority by the other bystanders to speak, and whose pronouncements are thus
accepted as locally definitive. So for situations and societies where there are no ac-
knowledged doctors or other practitioners with equivalent special authority, it is the
bystanders, whoever they may be —those whom John Graunt in 1672 called ‘the
generality of the World’— who define cause of death and thus determine the array of
available diseases in any particular time and culture. The more general form of my
point that you die of what your doctor says you die of can therefore be reformulated
as People die of what their bystanders say they die of, with it being understood that
there is often a special class of bystander (the doctor) who makes claims to, or is
ascribed, special authority based on specialised knowledge. But in general the speci-
fication of cause of death requires no medical expertise: if there is no medical practi-
tioner present, then anyone can do it, and does, and what they say goes. No matter
how ignorant we may regard such bystanders as being, nevertheless the cause of
death they ascribe is not negotiable afterwards.
Only if a doctor is involved does specification of the cause of death require medi-
cal expertise. But in all cases, including those where a doctor is involved, what the
specification of the cause of death really requires is social expertise. That is to say, it
requires full immersion in and acquaintance with the mores and beliefs of the perti-
nent society. This is what everyone brought up in a given society possesses as second
nature, and it is precisely what the outsider to that society, whether anthropologist,
missionary, visitor from Mars or even historian, does not possess. What the outsider
does, spontaneously, is translate what he or she sees or hears from the bystanders
into his or her own language and culture. And as we all know, traduttore traditore.
To illustrate these points I need a modern moment of death, a volunteer from the
audience as it were. The particular volunteer I have chosen is my own father, partly I
suppose as a personal memento mori. My father died at about midnight on the longest
day of the year in 1987, at home in Swansea in Wales. He had been ill for some
months, and had spent a period in hospital where he had undergone abdominal sur-
gery. He was 76 years old, but did not die of old age because, although it was once a

Asclepio-Vol. LIV-1-2002 17
 ANDREW CUNNINGHAM

regular cause of death, it is a very difficult thing to die of these days. The illness from
which he died is what might be called ‘the disease which dare not speak its name’ (if I
may paraphrase Oscar Wilde). In my father’s case this was quite literally so. For my
mother reports that just before he died, and fully aware that these were his last mo-
ments, my father in his very weakened voice told her he loved her, spoke of the last
matters that needed to be attended to when he’d gone, and then said, ‘Is it —(pause)—
cancer?’ My mother said, yes, it was. Then he died.
So my father had experienced a long illness which had made him progressively
weaker and which over a number of months had caused him to waste away in front of
his own eyes; he had been into hospital for what was probably only the second time
in his life, and he had received major invasive surgery for this illness; he had dis-
cussed his condition with his doctors, his wife, his family and visitors over a period
of months; and yet no-one had felt able to tell him what he was suffering from, and
nor had he asked. He was not a shy man about such matters. If he had wanted to
know, he would certainly have asked. But he didn’t want to know. He really didn’t.
He waited until the very moment of his death to have his worst fears confirmed – and
they were. Because it was cancer. In fact it was cancer of the bowel.
How do we know this? Why can we trust this? Well, we take the doctors’ word
for it (at least I did). But how did the doctors themselves know what my father died
of? They did so by following their training as modern doctors. How else? Thus, my
father complained of certain problems, which happened to be rectal bleeding, severe
abdominal pains and weight loss. This, together with my father’s age, raised bowel
cancer as a possible clinical diagnosis in their minds, something they had learned at
medical school. They then went and tested the provisional diagnosis, by using the
modern methods of diagnosis they had been taught. To distinguish with certainty
between a benign and a malignant (cancerous) tumour or growth, a biopsy was per-
formed. These tests were at the cellular level, because this is the primary level of
medical understanding today, that organic diseases are cellular phenomena. Thus
carcinomas are nowadays defined as products of abnormal cell activity, or as some-
one has said, cancer is ‘the misguided cell’4. With this evidence from the laboratory,
the doctors could be certain of the identity of the disease in biological terms: they
now knew what my father was suffering from. As far as treatment was concerned, the
doctors again naturally turned to what they had been taught when they were trained.
Thus, as the tumour seemed to them to be so far confined only to a portion of the
bowel, so their next move was to submit my father to surgery to remove the affected
————
4 PRESCOTT, D. M. and FLEXER, A. S. (1986), Cancer: the Misguided Cell, Sunderland, Massachu-

setts, Sinauer Associates Inc. Another author calls cancer ‘the wayward cell’; see Richards, V. (1978),
The Wayward Cell: Cancer, its Origins, Nature, and Treatment, Berkeley, Los Angeles, University of
California Press. GRAHAM, R. M. (1963), The Cytologic Diagnosis of Cancer, Philadelphia, W. B. Saun-
ders Company, notes that methods for securely distinguishing cancerous from non-cancerous cells have
been developed only since the 1940s (see Preface).

18 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

section. In this way the doctors hoped to stop the cancer spreading, either by direct
extension from the primary site, or by the cells entering the vascular system by in-
vading the lymphatics or blood vessels. So a large length of bowel was removed, and
the healthy ends of the intestine were joined up. My father did not respond as hoped.
So the doctors concluded that the cancerous cells had spread and that there was nothing
practical more that they could do for him, especially given his age. So my father
went home, where he grew weaker and gradually faded away, all his actions and
responses slowing down until even watching him eat, dress or speak required great
patience on everyone’s part5. Eventually he took to his bed and never got up again.
So in the present example, my father died of cancer of the bowel because this is
what his doctors said he died of. And that’s that. His cause-of-death certificate is not
negotiable, either today or in the future. The diagnosis was offered, tested and con-
firmed, through all the procedures that modern Western hospital-based medicine
requires. The cells had been found guilty. On the basis of the diagnosis the treatment
was given. The sequence of diagnostic steps tells you that this is the answer, that this
has to be the answer. There is no alternative diagnosis possible in this case, nor any
other possible cause of death available.
With respect to how diagnosis happens today, and that you die of what your doc-
tor says you die of, this must suffice for now. But what of the other point, that the
identity of any disease is made up of a compound of elements, of which the biological
or medical is only one, and sometimes the least important one? Of all modern disease
categories, cancer is perhaps the one most straightforward for making this point
about how far the social element makes up the disease’s identity. It is clear from my
father’s own reaction —and this is something widely shared in our modern society—
that cancer has a very special place in the modern disease spectrum. It is the most
feared of modern diseases, in most of its forms. The social meaning with which we
load cancer, and lead each other to load it with – seems to be the reason why cancer
has this special position of being unmentionable in our social experience of disease
in today’s society. Susan Sontag, the celebrated modern commentator on the social
dimensions of cancer, has remarked that no-one is embarrassed to say that they have
had a heart attack if they’ve had one. Similarly, no-one will hide the fact of their
heart surgery if they have had heart surgery. But few people want other people to
know that they have cancer. In many cases, my father’s included, they themselves do
not want to know that they have cancer. There is something about our attitude to this
disease which makes it different. It comes with overtones of dirt and shame, and of
blame and punishment, which few other diseases have. Lester and Devra Breslow
————
5 All this was rationalised by my father in a way which made cancer irrelevant as a possible cause.

As he knew sections of his bowel had been removed (though did not ask why) it was obvious to him that
there was not enough intestine left to detain and absorb the food long enough to nourish him. As he
noticed that his food just ‘fell straight through’ him, he was not at all surprised he was fading away. He
explained all this to me. Very slowly.

Asclepio-Vol. LIV-1-2002 19
 ANDREW CUNNINGHAM

have written that ‘Cancer was and is still perceived largely as a disease that attacks
individuals one at a time, each uniquely, rather than as a public burden requiring
large-scale public efforts’6. This focus on the individual as responsible for his or her
cancer has, according to the Breslows, been inadvertently maintained, at least in the
U.S.A., by the action of interested parties such as the private medical business, the
private industrial business and the biomedical research establishment. So strong is this
social element in the identity of cancer in our society today, that it over-rides and ren-
ders nugatory the medical view that it is just a few cells misbehaving, and that five-
year survival rates from cancer treatment are improving all the time. My father did not
die from a disease whose identity is purely biological. His experience (demonstrated
precisely by his refusal to voice his concerns), together with the experience of the doc-
tors around him, the experience of his family, and the experience of all the other by-
standers, was that he was dying from a shameful disease, the disease which dare not
speak its name. All this human experience was part of the identity of the disease.
The social component of the identity of cancer may be the most easy to point out,
thanks to the work of sociologists and other commentators over the last few decades.
But every disease has its social component, and the social component (like the medi-
cal component) of a disease can and often does change over time. This applies equally
to all conditions labelled as diseases in all societies. I hope to make that case at
length elsewhere sometime.
I have here only given one case history tracing the steps of diagnosis and treat-
ment appropriate to the late 20th century. I chose it because of its familiarity: we all
know how the suspected cancer patient is treated today. Now imagine tracing the
steps of diagnosis and treatment of any episode of disease of the past. Here we will
be on unfamiliar territory. The conditions will be different, the content of diagnosis
and treatment will be different, the thought patterns guiding diagnosis and treatment
will be different. And yet, the same conclusion will have to be reached by us: that the
patient died of what his doctor said he died of. Because there is nothing else in the
encounter. There is no ‘real’ disease, with an identity separate from its sufferers at
any given time, which can be separated out as a timeless entity for us to give our
modern labels to, years —centuries— after the events.

II: THE ARRAY OF AVAILABLE DISEASES

The present conference paper is part of a longer argument about disease identity
in the past that I am currently engaged on, and there can be room here only to discuss
one particular phase of the longer argument. So to support my major claims about the
————
6 BRESLOW, L. and D. M. BRESLOW (1982), «Historical Perspectives [on cancer epidemiology and

prevention]», Cancer Epidemiology and Prevention, Schottenfeld, D. and J. F. Fraumeni, eds., Philadel-
phia, W. B. Saunders Co. 1039-1048, 1040.

20 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

identity of disease, and in particular the claim that you die of what your doctor says
you die of, I shall look here only at what I shall call the array of available diseases in
any given historic or modern society, that is to say at what set of diseases is thought
to exist at a given time in a particular society. Other questions that arise naturally
from this, such as how diseases are added to or subtracted from these arrays, or how
individual cases of disease are mapped onto these arrays in any particular historic or
present society, as also with most questions about the operative steps involved in
diagnosis, I will have to leave for another occasion. For convenience I shall use here
only examples of diseases that are or were believed to kill people, but the analysis
could in principle be deployed also for all non-fatal diseases too.
Different societies and different periods in any one society have different sets or
arrays of ‘available diseases’. Such arrays are different both in what they include and
also —and this is even more important— in the theoretical underpinnings of what
constitutes the identity of a particular disease. Hence what you can die of at any par-
ticular time differs. To illustrate what I mean by this I shall compare three cause-of-
death analyses drawn up at intervals of about 160 years from each other. They are
(1) the London Bills of Mortality, kept from the 1590s to 1849, as analysed by
John Graunt in 1672;
(2) the first comprehensive report on causes of death in England, viz. the statistics
drawn up by the Registrar-General in 1839, as commented on by the Compiler
of Abstracts, William Farr; and
(3) what is at the time of writing the most recent edition of Health Statistics Quar-
terly for England and Wales, the issue covering the year 1999.
At first glance the three documents and the information they contain seem rea-
sonably comparable, given that they cover much the same area and society (England,
London), are in the same language (English), were drawn up for much the same rea-
sons (to track contemporary causes of death), were drawn up by people with a strong
interest in ‘political arithmetic’ or statistics as a means of assessing the state of the
State, and were drawn up within a period of less than four hundred years. That is, they
seem like successive attempts to solve the same problem, and indeed this is how they
have been generally treated.7 But they are not comparable at all because, given the
changes in disease conceptualisation that had occurred between the first and the second
and between the second and the third of them, they were, in effect, drawn up in differ-
ent societies and cultures. In a word, they are incommensurable. Thus causes of death
ascribed in the first two are not amenable to subsequent diagnosis by the modern medic
or historian. And the causes of death ascribed in the most recent one, in its turn, may
well not be amenable to later re-diagnosis by future medics or historians.
————
7 See for instance GREENWOOD, M. (1948), Medical Statistics from Graunt to Farr, Cambridge,

Cambridge University Press.

Asclepio-Vol. LIV-1-2002 21
 ANDREW CUNNINGHAM

1. The London Bills of Mortality

The Bills of Mortality were kept in London sporadically from the 1592 plague, and
regularly from the 1603 plague. They are believed to have been started in order to in-
form the royal court and the rich residents of London when plague or other epidemics
had reached such a state that flight from town was necessary. Outside plague times,
however, they were consulted by Londoners ‘so as they might take the same as a Text
to talk upon, in the next Company’, that is as dinner-table conversation8. They were
drawn up by parish clerks, as part of their duties, and the records were kept in their
guild headquarters, Parish-Clerks Hall. They were printed every week on Thursdays,
and a general account of each year was printed on the Thursday before Christmas Day.
They covered the 97 parishes within the walls of the city of London, plus the 16 par-
ishes in the liberties but outside the walls (together with their pest-house), plus the nine
adjoining parishes. A few other parishes were added in later years. Because of the way
they were conducted, they recorded only the births and deaths of members of the
Church of England, omitting Catholics and Dissenters9.
How were the causes of death gathered? This is what John Graunt says about the
procedure:

When anyone dies, then, either by tolling, or ringing of a Bell, or by bespeaking of a

Grave of the Sexton, the same is known to the Searchers, corresponding with the said
Sexton. The Searchers hereupon (who are ancient Matrons, sworn to their Office) repair to
the place, where the dead Corps lies, and by view of the same, and by other enquiries, they
examine by what Disease, or Casualty the Corps died. Hereupon they make their Report to
the Parish-Clerk, and he, every Tuesday night, carries in an Account of all the Burials, and
Christenings, happening that Week, to the Clerk of the [Parish-Clerks’] Hall10.

————
8 GRAUNT, J. (1672), Natural and Political Observations Mentioned in a following Index, and

made upon the Bills of Mortality. By John Graunt, Citizen of London. With Reference to the Government,
Religion, Trade, Growth, Ayre, Diseases, and the several Changes of the said City, London, John Martin,
James Allestry and Tho. Dicas, 1.
9 For an extensive series of them, and information about their history, see [Heberden, W., Ed.]

(1759), A Collection of the Yearly Bills of Mortality from 1657 to 1758 inclusive. Together with several
other Bills of an earlier Date. To which are subjoined 1. Natural and Political Observations on the bills
of mortality: by Capt. John Graunt, F. R.S. reprinted from the sixth edition, in 1676. II. Another essay in
political arithmetic, concerning the growth of London; with measures, periods, causes, and consequences
thereof. By Sir William Petty, Kt. F.R.S. reprinted from the edition printed at London in 1683. III. Obser-
vations on the past growth and present state of the city of London; reprinted from the edition printed at
London in 1751; with a continuation of the tables to the end of the year 1757. By Corbyn Morris Esq;
F.R.S. IV. A comparative view of the diseases and ages, and a table of the probabilities of life, for the last
thirty years. By J. P. esq; F.R.S., London, A. Millar.
10 GRAUNT (1672), (note 8 above), 11. Here and in all other quotations from 17th century sources I

have silently modernised spellings.

22 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

For Graunt’s purposes in making his natural and political observations, it mattered
little ‘whether the Disease were exactly the same, as Physicians define it in their Books’
(p. 14). As Graunt points out, what was sometimes reported was the predominant
symptom, rather than what a doctor would count as the disease proper. Sometimes the
two old women ‘after the mist of a Cup of Ale, and the bribe of a two-groat fee’ could
perhaps be persuaded to report all deaths from emaciation as consumption, and not
specify whether it was a phthisis, a hectic fever or an atrophy. But even so, in general,
Graunt concluded, these diagnoses of cause of death could be pretty well trusted:

To conclude, In many of these cases the Searchers are able to report the Opinion of the
Physician, who was with the Patient, as they receive the same from the Friends of the
Defunct, and in very many cases, such as Drowning, Scalding, Bleeding, Vomiting, making-
away themselves, Lunatics, Sores, Small-Pox, &c. their own senses are sufficient, and the
generality of the World, are able pretty well to distinguish the Gout, Stone, Dropsy, Falling-
Sickness, Palsy, Agues, Plurisy, Rickets, &c. one from another. (pp. 14-15)

Graunt reported that the Bills of Mortality revealed that some new diseases had
appeared in the early 17th century in London. Rickets first appeared in 1634, and
Graunt believed it was a new disease, not just a hitherto unreported one. The
‘stopping of the stomach’ first appeared in 1636.
What is particularly striking about the Bills of Mortality is that the establishing of
causes of death was barely at all a medical affair. The whole business was a concern of
church administration (the parish clerk system), since it was something which had
simply been added on to the arrangements for funerals and burials in the parish. It was
not the concern of any medical institution. Neither the College of Physicians, nor the
Barber-Surgeons Company of London, nor the Society of Apothecaries were involved.
Nor were even individual doctors directly involved. The opinions of the bystanders at
the bedside of the dead person, and those of the two old women themselves, about cause
of death were at least as significant as the opinions of any doctors.
The Table (TABLE 1) shows what diseases were effectively available to die from
in London in a randomly chosen sample year, 1632: that is to say, they are the
diseases which constituted the only official current form of listing of causes of death.
The diseases are listed alphabetically, and no-one now seems to know who first
chose this set of available diseases, nor how the list came in time to include one or
two new diseases such as stopping of the stomach and rickets. Presumably the listing
was systematised by the parish clerks, so that in practice the Searchers were
effectively limited to choosing a candidate disease from those already on the list.
Quite a number of these available diseases have no modern equivalent whatever
as possible causes of death, even in name. Fright, ague, bloody flux, canker (a
spreading sore, not to be confused with cancer), consumption, fever, grief, fallen jaw,
the King’s Evil, lethargy, livergrown, piles, planet, purples, quinsy, rising of the lights,
suddenness, surfeit and worms, are none of them recognised possible causes of death in

Asclepio-Vol. LIV-1-2002 23
 ANDREW CUNNINGHAM

Courtesy of the University Library, Cambridge.

24 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

the early twenty-first century in the West11. Yet it seems that, according to the Searchers,
in London in 1632 no less than 3,783 people —over a third of all the deaths that year—
died of one or other of these diseases. Of the four major kinds of disease which the
majority of people die from today in England and Wales —the degenerative diseases:
cancer, heart disease, respiratory disease and cerebrovascular disease— the name of
only one of them (cancer) is even present in this list. The big causes of death in 1632
London, by contrast, were 1. ‘chrisomes and infants’, that is children dying within the
first month of life or shortly after, the chrism being the baptismal robe in which the dead
infant would also be buried (2,268 cases); 2. ‘consumption’ (1,797 cases); 3. ‘fever’
(1,108 cases). The next biggest causes of death were 4. ‘old age’ (628 cases); 5. ‘flocks
and small pox’ (531 cases); and 6. ‘teeth’ (470 cases).
This system of drawing up the weekly Bills of Mortality continued into the early
19th century, without substantial change12. In time it came to be replaced by our next
kind of report.

2. 19th century Registrar-General reports

We now move on one hundred and sixty-seven years, to the first governmental
listing of causes of death throughout England13. The office of Registrar-General of
Births, Deaths and Marriages in England, was set up by the Registration Act of 1836
and its first annual report appeared in 183914. The Benthamite sanitary reformer,
Edwin Chadwick, claimed that it was at his suggestion that the Registrars had to
collect causes of death, and he had even hoped that the Registrars themselves would
be medical men15. The point of collecting causes of death, according to the Registrar-
General, was that ‘if the cause of death be correctly inserted [in the registers], there
————
11 I appreciate that some of these are conventionally taken to have been renamed later, and thus are
conventionally taken to correspond to modern diseases (for instance, both consumption and the King’s
Evil were supposedly later reclassified as tuberculosis). I shall deal in the larger version of this paper
with what happens in such reclassifications, and show that as a consequence of how reclassification takes
place, the earlier disease does not in fact correspond with the modern one.
12 For as complete a series of Bills as could be assembled in 1749, see [Heberden, W., Ed.] (1759),
(note 9 above).
13 For some of the interests in counting people and their causes of death in Britain in the interval be-
tween Graunt and the Registrar-General, see Greenwood (1948), (note 7 above), and Glass, D. V. (1973,
repr. 1978), Numbering the People. The Eighteenth-Century Population Controversy and the Develop-
ment of Census and Vital Statistics in Britain, London, Gordon and Cremonesi.
14 (1839), First Annual Report of the Registrar-General of Births, Deaths, and Marriages in Eng-
land., London, Longman, Orme, Brown, Green, and Longmans. For a comparable attempt to register and
tabulate causes of death in Ireland, see (1843), Report of the Commissioners appointed to take the Census
of Ireland for the Year 1841. Presented to Both Houses of Parliament by Command of Her Majesty,
Dublin, Her Majesty's Stationery Office.
15 GLASS (1973), (note 13 above), 149.

Asclepio-Vol. LIV-1-2002 25
 ANDREW CUNNINGHAM

will exist thenceforward public documents, from whence may be derived a more
accurate knowledge, not only of the comparative prevalence of various mortal dis-
eases, as regards the whole of England and Wales, but also of the localities in which
they respectively prevail, and the sex, age, and condition of life which each princi-
pally affects’. The Act itself had specified only that ‘some person present at the
death, or in attendance during the last illness’ should give this information to the
Registrar. But to avoid reliance on ‘persons ignorant of medicine, and of the names
and natures of diseases’, the Registrar-General ‘earnestly recommended that every
practising member of any branch of the medical profession who may have been pre-
sent at the death, or in attendance during the last illness of any person, shall, immedi-
ately after such death’ place in the hands of the person who would report to the Re-
gistrar ‘written statements of the cause of death’. For London, the Presidents of the
Royal College of Physicians and of the Royal College of Surgeons, together with the
Master of the Worshipful Society of Apothecaries, were prevailed upon to pledge
themselves ‘to give, in every instance which may fall under our care, an authentic
name of the fatal disease’, and, they said, they entreated ‘all authorized practitioners
throughout the country to follow our example, and adopt the same practice’. The
popular or common name of the disease was to be preferred to the technical name
known only to medical men16.
Thus the causes of death collected by the Registrars were intended to furnish new
knowledge about the incidence of mortal diseases, and the co-operation of medical men
of all kinds meant that the causes given should be highly reliable for statistical analysis.
After the information on causes of death had been collected each year, William
Farr, who held the post of Compiler of Abstracts from 1838 to 1880, wrote a long
letter to the Registrar-General analysing the material, and this letter was printed in
the Registrar-General’s annual report, and came to form the largest section of it
throughout Farr’s forty-two years in the job. Farr himself was a Paris- and London-
trained medical man, and by the time of his appointment at the age of 32 he was also
a fairly accomplished statistician. As a liberal reformer who was also a medical man,
his concerns were with promoting the sciences in medicine, especially hygiene, and
with investigating ‘all the laws of vitality capable of being observed in masses of
men, expressed in numbers’17.
So Farr had to construct a nosology suitable for statistical purposes, and based on
some medical principles of the day. (TABLE 2) A nosology is any classification of
diseases, usually presented in tabulated form. That is to say, it is a list of ‘available
diseases’ of a particular period and place, drawn up by an individual or organisation,
established on the basis of some medical principles or other, with the diseases ar-

————
16(1839), First Annual Report, (note 14 above), 77-9.
17EYLER, J. M. (1979), Victorian Social Medicine. The Ideas and Methods of William Farr, Baltimore,
Johns Hopkins University Press, 8, quoting Farr’s own aims for his journal British Annals of Medicine.

26 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

Courtesy of the University Library, Cambridge.

Asclepio-Vol. LIV-1-2002 27
 ANDREW CUNNINGHAM

ranged usually according to some theory about cause, and where that is not possible
or desirable, by some other supposedly meaningful classification, such as anatomical
location in the body. The term ‘nosology’ was popularised in 1763 by Boissier de
Sauvages for his tables of diseases18, and the term was rapidly adopted by other
medical teachers in the century, and it is still in use, though the expression ‘classifi-
cation of disease’ is more common today. Nosologies are not usually very stable. If
someone came along who had a different view of the range of possible diseases, of
the relative importance of different diseases, of the natural relations of diseases, or of
the causal relationships of different diseases, then he could simply draw up his own
nosology if he felt strongly enough about it. By 1838 there were many possible exist-
ing nosologies that Farr could choose from if he wished – he himself mentioned
those of Pinel, Richerand, Bichat, Parr, Young and Mason Good19. However, Farr
had his own medical principles and a strong view about the doctor’s role. In this
respect he was particularly concerned with the diseases which he believed were most
amenable to prevention. So of course he had to draw up his own, new, nosology20.
This is what is employed in the present table.
The disease taking pride of place —the most sudden and spectacular killing dis-
ease of the time— is cholera, which had scourged Europe in 1832. It headed the ‘first
division’ of diseases in Farr’s table, which consisted of the epidemic, endemic, and
contagious diseases. In Farr’s view these all constituted a natural group, because he
believed that they all had similar causation – what he was soon to call zymotic, a
poisonous effluvium given off by decay and dirt21. This primary division is therefore
a classification by cause: all these diseases, for Farr, had similar causes, and hence
were amenable to similar management. Their ‘exciting causes’, were, for Farr, the
insalubrity of the local conditions. Action could be taken to reduce their incidence,
by undertaking engineering measures to introduce clean air, clean water, remove
wastes, and so on. This belief on Farr’s part in ‘filth diseases’ and their causation
preceded his classificatory activity, and was as much a political belief —that is, about
how society ought to be structured and run— as it was a medical one. In 1866 Farr him-
self stated that ‘No variation in the health of the states of Europe is the result of chance;
it is the direct result of the physical and political conditions in which nations live’22.
Indeed, the Registry was for Farr in effect an instrument to advance this politico-medical
agenda. He did not draw up the array of available diseases in an innocent manner and
————
18 This was a development of his Nouvelles Classes des Maladies, first published in 1731. See Martin, J.
(1990), «Sauvages's Nosology: Medical Enlightenment in Montpellier», The Medical Enlightenment of the
Eighteenth Century, Cunningham, A. and R. French, eds., Cambridge, Cambridge University Press, 111-37.
19 (1839). First Annual Report (note 14 above), 92.
20 EYLER, (1979), (note 20 above), 53-60.
21 (1842), Fourth Annual Report of the Registrar-General of Births, Deaths, and Marriages in Eng-
land, London, Longman, Brown, Green, and Longmans, 199-205.
22 Eyler (1979), (note 19 above), 199.

28 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

then make supposedly Baconian inferences from it. No: the necessary structure of the
table and the array of diseases was given to him by his pre-existing beliefs about disease
causation. And because Farr for forty-two years personally drew up the tables and statis-
tics from the material he was sent, he continued to be the gate-keeper of ‘what counted’
as a disease, and the arbiter of which diseases were more important than others, even
though his positions on these matters were actually controversial.
So what could one die from in William Farr’s England in 1838-9? By far the sin-
gle greatest cause of death was consumption (20, 247), followed at a great distance
by convulsions (10, 729) and typhus (9,047) – the first in killing-power of the ‘first
division’ diseases. Indeed, as a group, Farr’s ‘epidemic, endemic and contagious
diseases’, comprised just under a quarter of all deaths that year, with smallpox, measles,
whooping-cough and diarrhoea following typhus in their mortality. But it needs to be
noted that, in grouping this ‘first division’ together, Farr’s statistics show that over
32,000 out of almost 150,000 deaths were actually caused by the zymotic products of
‘filth’. Cancer, a disease classified as ‘of uncertain seat’, killed only 1,228 people.
Meanwhile, debility killed more than cancer did, and atrophy, and intemperance (a
morally defined disease) too claimed their victims as causes of death.

3. The 1999 statistics

This most recent table of causes of death reflects how internationalised medicine
has become in recent decades. In Western countries it is no longer optional which
nosology you use for cause of death if you are a medic. What you use is the Interna-
tional Classification of Diseases (hereafter ICD). First drawn up in the 1890s by the
Chef des Travaux statistiques de la ville de Paris, Jacques Bertillon, at the initiative
of the International Statistical Institute (a statisticians’ pressure group), it was soon
adopted by the registrars of the United States, Canada and Mexico. It has been regu-
larly revised since, approximately once every ten years until recently, and the 1979
revision, the ninth (hereafter ICD9), though technically out of date, is the one cur-
rently in use23. Since 1946 the World Health Organisation has been entrusted with
revising the classification and getting governments to adopt it. As the use of the code
numbers reveals, it is the classification used in the U.K. statistics in TABLE 324.

————
23 On the history of the International Classification, see (1977), International Classification of Dis-

eases. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death.
Based on the Recommendations of the Ninth Revision Conference, 1975, and Adopted by the Twenty-
ninth World Health Assembly, Geneva, World Health Organization, Introduction.
24 «Report: Death Registrations 1999: cause England and Wales», Health Statistics Quarterly

(2000), n. p. The table is five pages long, and only pages one to three are reproduced here.

Asclepio-Vol. LIV-1-2002 29
 ANDREW CUNNINGHAM

30 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

Asclepio-Vol. LIV-1-2002 31
 ANDREW CUNNINGHAM

32 Asclepio-Vol. LIV-1-2002
 IDENTIFYING DISEASE IN THE PAST: CUTTING THE GORDIAN KNOT

According to the authors of the official U.S.A. clinical version of the ICD, it
‘represents the best in contemporary thinking of clinicians, nosologists, epidemiolo-
gists, and statisticians from both the pubic and private sectors’25. All nosologies, of
course, are intended to reflect the best in contemporary thinking, and this one there-
fore, like its predecessors, reflects perfectly the dominant medical thinking of its age.
In particular it expresses the interests of the epidemiologists and statisticians in the
management of public health, because the categories it uses are designed to be ones
most useful for general statistical use. In fact they descend from Farr’s distinction
between general diseases on the one hand, and those localized to a particular organ or
anatomical site on the other. This is probably because Farr was on the first committee!26
Most strikingly the great scare disease of the 1830s —cholera— still takes pride of
place among the general diseases, being number 001. However, in other respects, how
the diseases are distributed between these categories differs greatly from Farr's ar-
rangement, especially with respect to the transfer, to the ‘infectious’ grouping, of sev-
eral diseases which Farr had distributed according to anatomical seat.
What could one die of in 1999 in England and Wales? Of all the many differences
between Farr’s (first) version of 1839 and the tables of the year 1999 based on ICD9,
the greatest is the enormous presence in the latter of ‘neoplasms’ which, like ‘carci-
nomas’, is a medical euphemism for cancers. The tables reveal that cancer now is the
cause of death of 25% of the population of England and Wales. In 1632 in London,
just ten deaths out of over nine thousand were attributed to ‘cancer or the wolf’. In
Farr’s 1839 statistics ‘carcinoma’ was a disease ‘of uncertain seat’ and killed only
1228 out of a total of almost a hundred and fifty thousand deaths, which was less
than 1%. Today cancer kills one person in four.
We could extend this analysis of the array of available diseases to other periods,
other cultures: we could even take the practice of a witch-doctor. But for the moment
I will limit the analysis to England over this four hundred year period.
We find people in these three periods able to die only of a limited, specifiable and
specified, array of diseases. These arrays are different. In the second two, it is the
doctors and epidemiologists who specify what diseases are possible, whereas in 1632
it was two old women in each parish and then the parish clerk. In 1632 and 1999
death from zymotic diseases was impossible, while in 1839 almost a quarter of the
people who died were killed by filth diseases. Other comparisons will occur to each
reader. While some historians may choose to read these three different tables and
their disease categories in a direct and naive way, making the 1632 cancer and plague
the same as the 1839 and 1999 cancer and plague, to most medical historians it will
be self-evident that these three moments are separated not just in time but also in
————
25 (1980), ICD.9.CM The International Classification of Diseases 9th Revision Clinical Modifica-

tion, Washington, U.S. Department of Health and Human Sciences, vol. 1, Preface.
26 EYLER (1979), (note 20 above). 58.

Asclepio-Vol. LIV-1-2002 33
 ANDREW CUNNINGHAM

ways of thinking. For a host of innovations in medical thinking and transformations

in social attitudes separate each of these different cause-of-death tables. The early
19th century views of pathology and disease causation were radically different from
those obtaining in the mid-17th century. Similarly, the nosology of William Farr,
though it seems to correspond more to the modern ICD classification, had to give
ground to a new way of thinking about causation —the germ theory— in the latter
decades of the 19th century. What counted as a disease and, more particularly, pre-
cisely how a disease was diagnosed, had changed beyond recognition27. That is to
say, the disease identities changed. More properly, the disease identities were
changed, for this is something which happens entirely through human action. Hence
it remained and remains true, that in all these periods the people died of what their
doctor (or their bystander) said they died of. And that’s that.

————
27 CUNNINGHAM, A. (1992), «Transforming Plague: The Laboratory and the Identity of Infectious

Disease», The Laboratory Revolution in Medicine, Cunningham, A. and P. Williams, eds., Cambridge,
Cambridge University Press, pp. 209-44.

34 Asclepio-Vol. LIV-1-2002