the Massachusetts Institute of Technology and the editors of The Journal of

Interdisciplinary History

An Anthropology of Religion and Magic, II

Author(s): Keith Thomas
Reviewed work(s):
Source: The Journal of Interdisciplinary History, Vol. 6, No. 1 (Summer, 1975), pp. 91-109
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/202826 .
Accessed: 14/06/2012 20:15
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The MIT Press and the Massachusetts Institute of Technology and the editors of The Journal of
Interdisciplinary History are collaborating with JSTOR to digitize, preserve and extend access to The Journal
of Interdisciplinary History.

http://www.jstor.org

Journal of InterdisciplinaryHistory vI:I (Summer 1975), 9I-I09.

Keith Thomas
An Anthropology

of Religion

and Magic, II

Southey was a somewhat impatient listener to Coleridge's metaphysical talk. When Southey was engaged on his History of Brazil,
Coleridge said to him, "My dear Southey, I wish to know how you
intend to treat of man in that important work. Do you mean, like
Herodotus, to treat of man as man in general? Or do you mean, like
Thucydides, to treat of man as man political? Or do you mean,
like Polybius, to treat of man as man military? Or do you mean ..."
"Coleridge", cried Southey, "I meanto writethe historyof Brazil."I
Since most working historians tend to be impatient of anything which
looks like methodological discussion I must begin by saying that I am
genuinely grateful to Geertz, not only for so closely reading my text,
but also for formulating her criticisms of it in terms which pose wide
general issues of some profundity.2 It is a salutary experience to have
one's work subjected to probing analysis of this kind and I will readily
admit that, if I had had the advantage of reading Geertz at an earlier
stage, Religion and the Decline of Magic would have been a different
book, though perhaps not very different. Still, as the Red Queen said
to Alice, "When you've once said a thing, that fixes it, and you must
take the consequences." My aim in this brief note will not be so much
to "defend" my book as to reflect on the implications of some of the
important general issues which Geertz has raised.
She begins by objecting to the categories which I have used to
conduct my analysis. In particular, she questions whether there is such
a thing as "magic" at all. By adopting such a concept, and, even more,
by defining it in such a way as to distinguish it from "religion," I have,
she suggests, fallen victim to language which reflects the official prejudices of my own society; for today both scientists and theologians
agree in using the term "magic" negatively and pejoratively, to group
together and disparage such practices as they currently regard as irrational or useless. Worse still, these official prejudices have led me into
Keith Thomas is Fellow and Tutor in Modem History at St. John's College, Oxford.
I Richard J. Schrader(ed.), The Reminiscences
of AlexanderDyce (Columbus, 1972),
178.

2 I am also deeply grateful to the program committee of the American Historical

Associationfor devoting a sessionat its annualconvention (1972) to the discussionof
my book and for making it possiblefor me to be present.

92

KEITH

THOMAS

makingthe historian's
greatesterror-askingthe wrong question.For

she says, "It is not the decline of the practiceof magic that cries out for
explanation,but the emergence and rise of the label'magic."'
Is this criticismjustified? Have I been studying a non-existent
problem? Should I be compared to a pre-Namierite historian who
assumesthat the essenceof mid-eighteenth-centuryBritish politics was
a conflict between "Whigs" and "Tories," or a pre-Freudiandoctor
trying to find the causes of "hysteria"? Is "magic" a concept which
totally dissolveson closer inspection?
Let me say first that I am fully aware that anthropologiststoday,
when discussingthe beliefs of other societies, are chary about using
the Western concept of "magic" tout court.Acutely sensitive to the
danger of ethnocentricity,they emphasizethat an ethnographer'sfirst
task is to arrive at the basic categories or systems of classification
employed by the people whom he is studying. To do this he has to
begin by discardinghis own categories. "Typically he may have to
abandon the distinctionbetween the naturaland the supernatural,relocate the line between life and death, accept a common nature in
mankind and animals."3 It is partly this awarenessof the difficultyof
apprehendingunfamiliarsystems of classificationwhich has led to the
immense current interest among anthropologistsin linguistics, symbolism, and communicationstheory, and to a majorchangeof direction
in socialanthropology as a whole.4 The interestsof the new generation
of anthropologiststend to be not so much sociological, as linguistic,
even philosophical. Their primary concern is the way in which language and symbolism determine human understandingand behavior.
Their object is to reconstruct the various methods by which men
impose conceptualorder on the externalworld. They wish to identify
the "programs," the "grammars,"the "paradigms,"the "cognitive
structures,"on which social behavior,as they see it, is founded. Above
all, they seek to reconstructindividualculturalsystemsin theirentirety,
and to understandparticularnotions by identifying their place in the
system to which they belong. They thereforereject the work of those
earlier ethnographerswho thought it possible to study a society by
simple observationwithout masteringthe language of its people, who
classifiedbeliefs by their functionsratherthan by their inner structure,
3 Rodney Needham, introduction to his translation of Emile Durkheim and Marcel
Mauss, Primitive Classification(London, 1963), viii.
4 For this change and its implications see Edwin Ardener, "The New Anthropology
and Its Critics", Man, VI (I97I), 449-467.

RELIGION

AND

MAGIC,

II

93

and who, by wrenchingparticularaspectsof a systemout of their

culturalcontextand arbitrarilygroupingthem togetherwith superficiallysimilaraspectsof other systems,endedup by studyingnonexistententities,broughtinto being by the ill-consideredapplication
of a singlelabelto socialphenomenawhich in fact differedradically
from societyto society.
Adherentsof thisolderstyleof anthropologyarenow marooned
on a barrenshore,cut off by a fastrecedingintellectualtide.Away on
the
the horizonsailsa trim new craft,bearingthe post-structuralists,
Theirs
a
is
which
and
the
cognitiveanthropologists.
semiologists,
ship
no longer flies the flag of comparativesociology, but is dedicated
insteadto the discoveryof the enduringfeaturesof the humanmind.
To achieveits new speedthisvesselhashad to castoff a greatdealof
once
ballast,amongit manygeneralcategorieswhich anthropologists
usedandwhichhistoriansstill use withouta blush.They includenot

just such anthropological exotica as "totemism," which everyone

regardsas a uselessitem of vocabulary,but other more familiarterms,
which most historiansprobably do not realize are now regarded as
contentious; for example, "ritual,""belief," "witchcraft,""kinship,"
and "religion."5 All of these have been rejected, at least by some
writers, because they are culture-boundcategorieswhich are alien to
the thinking of many societiesand which, if used on a universalscale,
turn out to lack any constant or intrinsic content. ("Primitive," of
course went much earlier because of its condescending evolutionary
overtones;it hasbeen replacedby such debatablesubstitutesas "tribal,"
"traditional,""undifferentiated,""preliterate"or "having a low level
of materialculture.")6 And with this rejectedballasthas gone "magic."
For, as one writer remarks,"If categoricaldistinctionsof the Western
5 See Claude Levi-Strauss (trans. Rodney Needham), Totemism (Harmondsworth,
I969); Malcolm Crick, reviewing J. S. La Fontaine (ed.), The Interpretationof Ritual
5I; Rodney Needham,
(I972), inJournal of theAnthropologicalSociety of Oxford, III (I972),
Belief, Language, and Experience (Oxford, 1972); T. O. Beidelman, "Towards More
Open Theoretical Interpretations," in Mary Douglas (ed.), WitchcraftConfessions and
Accusations(London, 1970), 35I; Malcolm Crick, "Two Styles in the Study of Witchcraft,"Journal of the AnthropologicalSociety of Oxford, IV (I973), I8; Rodney Needham
(ed.), Rethinking Kinship and Marriage(London, 1971), cviii; David M. Schneider, "What
is Kinship all about?" in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial
Year (Washington, D.C., I972), 51; Wilfred Cantwell Smith, The Meaning and End of
Religion (New York, I964). The term "religion" is retained by Thomas Luckmann, The
Invisible Religion: The Problem of Religion in Modern Society (New York, 1967), but
widened to embrace any matter which an individual regards as of "ultimate" significance.
6 For an isolated defence of the term see Mary Douglas, Purity and Danger (London,
I966), Ch. 5.

94 | KEITH THOMAS

mind are found upon examinationto impose distinctionsupon (and so

falsify) the intellectualuniverses of other cultures then they must be
discarded... I believe 'magic' to be one such category."7
The wider perspectiveof the present-dayanthropologist,however,
inevitably distinguisheshis methods from those of the historian. For
the anthropologist is rightly suspicious of any terminology which
looks unsuitable for use in cross-culturalcomparison, whereas the
historian, whose preoccupationsare usually less global, is more obviously culture-bound.He is content to speak of "religion" or "kinship" without worrying whether these terms are helpful in some more
exotic context. Of course he would find it easy to agree with E. R.
Leach that "English-languagepatternsof thought are not a necessary
model for the whole of human society".8 But though unsuitablefor
export they may well be good enough for home. In Religionand the
Decline of Magic (London, 1971), I was attempting to write English
history, not to engage in cross-culturalanalysisand I must plead guilty
to having used languagewhich contemporariesthemselves,or most of
them, would have understood. Of course, there is no single definition
of "magic" elasticenough to embraceall of the differentusageswhich
contemporariesgave the term. Even so, at the end of the period with
which I was concerned, the expression "magic" had come to have a
tolerably clear connotation. It meant the deliberate production (or
attemptedproduction) of physicaleffectsor the gaining of knowledge
by means which were regarded as occult or supernatural.There is
nothing hardand fast about this definition.For contemporariesdiffered
among themselvesas to what was or was not "natural"(255-256), and
many refrained from applying the term "magical" to supernatural
operations which were authorized by the Church (Chs. 2, 4) or the
state (Ch. 7[ii]). Moreover, the word "magic" was relatively slow to
emerge as a single label for a number of different activities. In the
Middle Ages it was more common to speak separatelyof "enchantment," "necromancy," "conjuration" or "sorcery" than to refer
simply to "magic." The word existed both in Latin and in English,but
commentatorsand clericstended to list the magic artsseparately;only
in the sixteenth century did it become common to group them
7 D. F. Pocock, foreword to Marcel Mauss (trans. Robert Brain), A General Theory of
(London, 1972), 2. As yet, however, few anthropologists seem to have managed
MVagic
to keep the word out of their pages, though some use the alternative "magicoreligious".
8 Leach, Rethinking Anthropology(London, 1961), 27.

RELIGION

AND

MAGIC,

II [ 95

together under a single head.9Nevertheless,I felt no inhibition about

using the expression "magic" as a convenient label for bracketing
together a variety of specific practiceswhich contemporariesusually
associatedtogether, which had been classifiedas "magical arts" since
classicaltimes, and which in any case I tried to describe in concrete
enough detail for it to have been reasonably clear at any particular
point as to just what I was talking about.
Perhaps I should have laid even more emphasis on the heterogeneity of these differentactivities,stressinghow the nature of a neoscientific system of divination like astrology was quite differentfrom
that of healing by charms or conjuring spirits; indeed, the distinction
between ars magicaand scientiadivinationiswent back to the classical
period. But I think I describedthe individual practicesand beliefs in
sufficientparticularityfor any seriousconfusion to have been avoided.
Only in the last few pages did I introduce, half frivolously, the quite
different definition of magic as ineffective technique (667-668). But
that sort of magic, which is of courseuniversal,I was not the subjectof
my book. Much of the magic with which I was concernedwas certainly
ineffective,but ineffectivenesswas not part of my definition of it; and
the question of whether Elizabethanswould have accepted such a
definition does not arise.Many of Geertz'sstricturesrelate, therefore,
to the last two pages of my conclusion, ratherthan to the main body
of the book. For theremy working assumptionswere no differentfrom
those of the anthropologist,Nur Yalman, who speaksof "the practical
use of [supernaturalor divine] powers for everyday purposes such as
healing or assuringluck and fertility-which in very general terms we
may refer to as magic"; though, like him, I would readily agree that
it is "not a uniform classof practicesand beliefswhich can be immediately discoveredin every society." I fully recognize that for anthropologists the concept of magic is "one which has been batteredabout out
9 Cf. Robert-Leon Wagner, "Sorcier" et "Magicien." Contribution a l'histoire du
Vocabulairede la Magie, these (Paris, 1939), 26n-27n. On the terminology used in the
late Roman period see Eliane Massonneau, Le Crime de Magie et le Droit Romain, these
(Paris, I933); H. Hubert, "Magia," in Ch. Daremberg and Edm. Saglio (eds.), Dictionnaire des antiquitesgrecqueset romains (Paris, 1877-I919). In England the word "magic"
was established at least by Chaucer's time (see J. A. H. Murray [ed.], A New English
Dictionary on Historical Principles [Oxford, I888-I933]), but its semantic history needs
more investigation.
Io Cf. W. Michael Brooker, "Magic in Business and Industry: Notes towards Its
Recognition and Understanding," Anthropologica, IX (1967), 3-I9, where magic is
defined as repetitive, non-adaptive behavior.

96 [

KEITH

THOMAS

of useful recognition"" and I have no desireto follow their predecessors in a searchfor the universalmeaning of magic, religion or science.
But so long as we are concerned with the sixteenth and seventeenth
centuries,the analyticutility of these terms is surely adequate.
In any case, as my definition of magic implies, I did not suggest
that magic was always distinct from "religion." On the contrary, I
observed that "The line between magic and religion is... impossible
to draw in many . . societies; it is equally impossible to draw in
medieval England"(50). I would agree that "magic" is normally "best
regardedas an aspectof religiousbelief and practicethat takesits special
force from the antecedent and deeply rooted recognition in many
societies of supernaturalor divine power." I2 What I suggested in my
book was that a reclassificationtook placeduringthe period with which
I was concerned,whereby those elements in religion which ultimately
came to be regarded as magical were gradually identified as such,
firstby the Lollards,then by the Reformers(Ch. 3). I furtherurged that
a fundamentalchange took place in the idea of religion itself, as the
emphasiscame to be placed on formal belief ratherthan on a mode of
living (76-77).I3Farfrom ignoring the emergenceof the term "magic"
as something separatefrom "religion," I pointed out that the classic
distinction between the two, normally associated with E. B. Tylor
and other nineteenth-centuryanthropologists,was in fact originally
formulatedby the sixteenth-centuryProtestantReformers (6I). It was
they who first declaredthat magic was coercive and religion intercessionary, and that magic was not a false religion, but a differentsort of
activity altogether. The error of Tylor and SirJames Frazer(but not,
I think, of Thomas)was to make this distinctionuniversalby exporting
it to other societies.I4
Nevertheless, I am sorry if my use of the terms "magic" and
"religion" has caused confusion. As Evans-Pritchardsays, "terms are
only labels which help us sort out facts of the same kind from facts
which are differentor in some respectsdifferent.If the labels do not
11 Yalman, "Magic," in David L. Sills (ed.), International
of the Social
Encyclopaedia
Sciences(n.p., 1968), IX, 522; Needham, Belief, Language, and Experience, 2o8n.
12 Yalman, "Magic," 522.
13 I wrote these pages before I came acrossSmith, MeaningandEndof Religion,but I

think my descriptionclosely parallelshis account of the shift from the concept of

"religiousness"to that of"religion" (ibid.,39).
14 I note that even the reviewer (RandalKeynes) of Religionandthe Declineof Magic
in the avant-gardeJournal of the AnthropologicalSociety of Oxford (III [I972], 154), acquits

me of using "Frazerian"terms of reference.

RELIGION

AND

MAGIC,

II |

97

provehelpfulwe candiscardthem.The factswill be the samewithout

their labels."I' I claim no universalityfor a distinctionbetween magic

and religion, but I do suggest that in Europeanhistory, at least, it is
analyticallyuseful to distinguishthose religions which, like medieval
Catholicism, credited their rituals with physical efficacy from those
which, like eighteenth-centurydeism, did not. I devoted a good deal
of space to describing how the sectarians,by engaging in prophecy
and religious healing, brought back into religion much of the magic
which the Reformers had cast out (Ch.5). But I also statedthat "at the
end of our periodwe can draw a distinctionbetween religionand magic
which would not have been possibleat the beginning" (640).
To that extent I did discussthe emergence of the label "magic."
I explained how churchmenof every denomination used the term to
brand as implicitly diabolicalall unauthorizedattempts to manipulate
the supernatural,including many folk practicespreviously regarded
by their adherents as godly (I92, 256, 265-267); and how Protestants
appliedthe same descriptionto cover any claimsto manipulationmade
by the Churchitself.The dividing line between "magic"and "religion"
was hardened by the parallel attempts of Protestant and Catholic
Reformers to eliminateall popularrites of unauthorizedor ambiguous
status.16The Catholics did not abandon all claims to supernatural
manipulation,but the more austereposition of the EnglishProtestants
generated the very categories which anthropologiststhemselves are
only now beginning to discard(61).
Nevertheless, Geertz has a point. I should have devoted more
space to a proper semanticdiscussionof how the boundariesbetween
"religion," "magic," and "science"shifted and reshifted,according to
the varying outlooks of differentsocialand religiousgroups.17I should
have shown, for example, how the category "naturalmagic" melted
away altogether, part becoming science, the rest being discardedas
obsolete. I should also have paid more attention to the changing
vocabularyin which magical practitionersand magical activitieswere
described;and I should have consideredmore explicitly how far the
practitionersof magic themselves regardedtheir activities as magical.
For I think it quite wrong to suggest, as does Geertz, that the only
15 E. E. Evans-Pritchard, Witchcraft, Oracles and Magic among the Azande (Oxford,
1937), iI.

On this theme see Jean Delumeau, Le CatholicismeentreLutheret Voltaire(Paris, I97I).

This criticism is also well made by Keynes in Journal of the AnthropologicalSociety
of Oxford, 154.
r6
I7

98

| KEITH

THOMAS

contemporarieswho used the term "magic" were those who rejected

it. If that had been so we should never have encountered women
claiming to be "good" witches or intellectuals boasting of their
magical powers. Of course, some wizards saw themselvesas possessed
of a specialkind of "cunning"(i.e. knowledge or technique),but others
unashamedlyconfessed to attempting to manipulatethe supernatural.
To that extent the modern definition of magic was acceptedby many
of its practitionersthemselves.
Even so, I am now more sensitive to the intricate problem of
shifting vocabulary and classificationthan I was when I wrote the
book. I can see a lot of historicalwork waiting to be done in this area,
not just on magic, but also on social class, kinship, age-groups, and
other fundamentalcategories.For, from the anthropologist'spoint of
view, much of what historianscall social change can be regardedas a
process of mental reclassification,of re-drawing conceptual lines and
boundaries.My book was meant to demonstratea hardeningof mental
divisions, between naturaland supernatural,between the moral order
and the naturalorder; which, I take it, is what Max Weber meant by
the disenchantment of the world. I cannot, therefore, agree with
Geertz in dismissingas a boring non-question the problem of how far
the various practiceswhich I identified as magical did in fact decline.
On the contrary, I maintain that in England magic declined in a
double sense: The clergy abandoned all claims to be able to achieve
supernaturaleffects; and the practice of the various magical arts
diminishedin prestige and extent. I also think that this declining faith
in the physical efficacy of religious ritual and in the power of the
cunning men, poses some crucialhistoricalissues.Despite the popular
survival (perhaps even to a greater extent than I suggested (665,
666-667) of many of the practicesand attitudes which I discussed,I8
I remain convinced that what I called "the decline of magic" has to be
regardedas one of the great historicaldivides.
Many of my critics have accused me of being unfair to religion.
Geertz,however, thinksthat I have been unfairto magic, by suggesting
that it dealt with only a limited number of problems, by contrastwith
religion, which was an altogether more elaborate affair, a mode of
living and a coherent system of explanation (153-I54, 636-637). I

should have done more, she thinks, to bring out the hidden conceptual
i8
Cf. E. P. Thompson, "Anthropology and the Discipline of Historical Context,"
Midland History, I (I972), 53-55.

RELIGION

AND

MAGIC,

II | 99

foundationson whichmagicalpracticesrested,to showthattheywere

builtuponan alternative
but equallycoherentcosmology.Instead,she
suggests,I crudelytreatedmagic ritualsas mere psychologicalresponsesto immediateneeds,therebyfailingto seethatmenwouldnot
haveturnedto themin the firstplaceif they hadnot alreadythought
them to possesssome intellectualplausibility.Belief-systems,she
argues,havean independentlife of theirown, whereasI havesimply
discussedthem in utilitarianfashion,implying that magicalbeliefs
aroseto fit immediatepracticalneeds,only to be discardedoncethose
needshadevaporated.
I hope that my pictureof the relationshipof magicalbeliefsto
practicalneedswas not reallyso crude.I fully realizethat modern
cognitive anthropologistsare reluctant to treat mental activity as a

mere epiphenomenonof the socialand economicinfrastructure.

Far
are
more
maintain
from being economicdeterminists,
to
they
likely
that culturesare not so much materialphenomenaas "cognitive
organizationsof materialphenomena."9 Similarly,those influenced
assumethatmodesof thoughtaredeterminedby the
by Levi-Strauss
inherentqualitiesof the humanmind.20I, too, agreethat symbolic
formshave an autonomousreality,that ritualsarenot derivedfrom
sentiments,that psychologicalneeds do not createbeliefs,and that
magicwouldneverhavebeenpracticedunlessit hadfirstbeenthought
plausible.21Magic was stylized and inherited;men did not invent it at

momentsof stressand it did not caterfor every problem(656)any

more thanwitchcraftwas invokedto explainevery misfortune(538539).I certainlydid not suggestthatmagicwas a merealternativeto
technology,the one going out when the other came in.22On the

I9 Stephen A. Tyler (ed.), Cognitive Anthropology(New York, 1969), 3.

20 Cf. Nur Yalman, "The Raw: the Cooked:: Nature: Culture," in Edmund Leach
(ed.), The StructuralStudy of Myth and Totemism(London, I967), 71-89.
21 A classic refutation of the cruder functionalist view may be found in Levi-Strauss,
Totemism, Ch. 3. But it should be noted that even writers of a more functionalist
orientation have been careful to emphasize that situations of tension do not in themselves explain occult beliefs. See, e.g., Max Gluckman "Moral crises: Magical and
Secular Solutions," in Gluckman (ed.), The Allocation of Responsibility (Manchester,
1972), 4, I3.
But contrast
22

E. E. Evans-Pritchard, Theories of Primitive Religion (Oxford, 1965),

113 ("the advances of science and technology have rendered magic redundant"). Since

Geertz (above, 84n) suggests that I have misunderstood Evans-Pritchard's argument, it

is perhaps worth pointing out that Evans-Pritchard remarked of Malinowski's findings
that "his general conclusions as to the function of magic in society are fully borne out
by the Zande data" ("The Morphology and Function of Magic," AmericanAnthropologist, XXXI [I929], 621).

I00

I KEITH

THOMAS

contrary I pointed out that Lollards and Protestantsrejected magic

long before the practicalneeds for which it catered had received any
alternativetechnological solution (77, 656-666). When I wrote of a
shift from reliance on magic to sturdy self-help, I had in mind not a
psychological change so much as a doctrinal one. It was not that
Lollards and Protestants had stronger personalities than their predecessors,but that their convictions about the relativescope for human
action and supernaturalaid were different. The crucial shift was
"attitudinal,"in that it reflecteda changing attitudeto the relationship
of God and man.23 But to say that is not to invoke a psychological
interpretationbut to appeal to something nearer to the ontological
one which Geertz prefers; though, to the extent that religious beliefs
affected men's outlook and behavior, the change, of course, also had
psychological implications.
Neither was my treatmentof witchcraft primarilypsychological
in character.I tried to show that witchcraft beliefswere not private
delusions, generated by situations of stress, but were anchored in a
culturallyacceptableview of reality (Chs. 14, I5). They were part of
a much largercorpusof assumptionsabout the universe.A personwho
believed in witchcraft was not necessarilya paranoiac.On the other
hand, it took a specific social situation to bring witch beliefs into
action; and when they came into action in the form of witchcraft
accusations
they had, like all other human actions, their psychological
dimensions,being rooted in a variety of emotions, uppermost among
which was guilt (Ch. 17). But it was not guilt about turning old
women from the door which generatedthe conceptof witchcraft, any
more than it was a declining sense of guilt which led to its decay. A
psychological explanation of the kind advanced by LeVine may just
possibly help to explain why, in a society holding witch beliefs, some
individuals levied witchcraft accusations while others in a similar
situationdid not. But it certainlycannot explain the growth of skepticism about the possibility of witchcraft as such. If that skepticism
began among the social elite it was not becausethe members of that
elite had stronger egos, but because their social situation (superior
23 A vivid illustration of the change is provided by the medieval story of the Hungerford man who, in the dry summer of I259, set out to water his fields, but was miraculously
paralyzed for blasphemously attempting to mitigate the effects of a divinely-ordained
drought (Henry Richards Luard [ed.], Annales Monastici [Rolls Series, London, I86469], II, 351-352). By the seventeenth century the fatalism implicit in anecdotes of this
kind had been officially repudiated.

RELIGION

AND

MAGIC,

II

I10

education,greatermobility, more accessto news and information)

exposed them to a wider range of intellectualexperience.And if
atthevillagelevelalsodeclined,thatwasbecause
witchcraftaccusations
of a declinein the frequencyof the kindof ambiguoussocialsituation
not becauseof a changein the
which engenderedsuch accusations,
personalitytype of Englishvillagers.24
I thereforesympathizeentirelywith Geertzin her suspicionof
any shallowfunctionalistattemptto treatpopularbeliefsas simple
defencesagainstanxiety, vain compensationsfor technologicalinI canwell understand
why manyanthropologists
preferto
adequacies.
of
as
self-contained
ideas
thought,
systems
concentrating
solely
study
on theirinternallogic and theirontologicalstructure.Yet, thoughI
recognizethat the persistenceof magicalbeliefsis a problemin the
I alsothinkthatthe historianwould be
historyof cognitivestructures,
ill-advisedto separatesuchbeliefsfrom theirsocialand technological
context.It may not be trueto say thatmagicalbeliefsareonly to be
societies.But it is unquestionably
true that
foundin "pre-industrial"
it is the technologicalgap betweenman'saspirationsand his limited
controlof his environmentwhich gives magicalpracticestheirrelevance.As I suggested(637,667),magicalritesmayhavealsohadtheir
England
expressiveaspects,but in sixteenth-andseventeenth-century
doctors
theirpurposeswere usuallystrictlypractical.If contemporary
would
not
and
more
have
hadbeencheaper
successful,
gone to
people
charmers.If therehad been a policeforceto tracestolengoods there
would have been less recourseto cunningmen.25If the Churchhad
been able to caterfor all practicalneeds therewould have been no
wizards.26Counter-witchcraft,
magicalhealing,exorcism,were not
justexpressiveorsymbolicrites;theyweremeantto work.Thecunning
24 LeVine's ingenious addition to my argument is apparently based on my earlier
brief statement, "The Relevance of Social Anthropology to the Historical Study of
English Witchcraft," in Douglas (ed.), WitchcraftConfessions and Accusations, 47-79,
rather than on my book (LeVine, Culture Behavior, and Personality, 255n). It seems to
arise from his assumption that a decline in the belief in the mystical interdependence of
individuals and their neighbors must necessarily have had a psychological cause (ibid.,
264-265). But he does not say why he finds the more conventional social and intellectual
explanations of this phenomenon inadequate.
For continuing resort to a diviner in circumstances when the prospect of police
25
detection is thought unlikely or undesirable see Richard W. Lieban, "Shamanism and
Social Control in a Philippine City," Journal of the Folklore Institute, II (I965), 47-49.
According to a seventeenth-century divine, "People weary of their Christianity because
it easeth them not of the little discontentments of their estate in this world which they
meet with"; they therefore went to magicians (Herbert Thomdike, An Epilogue to the
Tragedy of the Church of England [London, I659], III, 290).

102

KEITH

THOMAS

folk dischargeda limitednumberof functions;peoplewent to them

at times of need, for highly practical purposes and in a distinctly

frameof mind.Theirprestigedependedupontheirsupposed
utilitarian
were rightto point out how the
efficacy,and earlieranthropologists
self-confirmingnature of their activities prevented clients from
realizingthatthey werenot efficacious.
Conversely, a belief which lost its practicalrelevance was likely

to wither.This seemsto have been what happenedwith witchcraft.

Despitewhat Geertzsays, it is by no meansclearthat witch trials
ceasedbecauseof a changein men'scosmologicalassumptions.On
the contrary,prosecutionsstopped,less becauseof disbeliefin the possi-

bility of witchcraftthanbecauseof the difficultyof provingit in any

instance;the well-publicizedexposureof fraudulentaccusaparticular
tionsmademenmoreawareof the epistemological
difficultyof telling
froma falseone;justason theContinentthetraumatic
a trueaccusation
effect of prosecutionswhich got out of hand ultimately sapped men's

And once the trialsstopped,it was

faithin thejudicialprocedure.27
only a matterof time beforethe laws changedand the realityof the
ideaitselfgraduallyfaded(453,573-576).
andanthroWe arestill,I think,very muchin the dark,historians
beliefs
as
to
the
which
collective
mechanism
precise
by
pologistsalike,
changeover long periodsof time. But no satisfactoryfutureinterpretationof the processwill be able to ignore the fact that beliefs
derivemuchof theirprestigefromtheirsocialrelevance.Theirinternal
have theirown logic and thislogic is not utilitarian.But if
structures
understandwhy the beliefsare held or rejected,we must
are
to
we
examinetheir relationshipto the society in which they operate.It
to explainthe declineof
would, for example,be quiteunsatisfactory
Christianbeliefin moderntimesmerelyby indicatingways in which
ancient theology has lost its intellectualplausibility. We should also

haveto considerthe changingfortunesof the Churchas an institution

andtakeaccountof the growthof rivalagenciesof education,welfare,
and entertainment.
Similarly,in askingwhy one type of medicineis
rather
thananother,for example,why osteopathyor
acceptedtoday
lack
prestige,we are dealingless with an intellectual
acupuncture
questionthan a socialone; we have to answerit by following the
which determinewhat the
fortunesof the professionalorganizations
Cf. Robert Mandrou, Magistrats et Sorciersen France au XVIIe siecle (Paris,I969);
H. C. Erik Midelfort, Witch Hunting in Southwestern Germany, 1562-1684: The Social
and IntellectualFoundations(Stanford, 1972), e.g., 162-163.
27

RELIGION

AND

MAGIC,

II

103

reigning orthodoxy shall be. We cannot study belief-systems in a

void; we have to determine what gives them their social credibility.
If it remainstrue that the contentof such beliefs cannot be explainedby
psychological reductionismor by sociological functionalism,it is also
true that changesin belief arevery difficultto accountfor in structuralist
terms. To understand why men's basic assumptions change it is
insufficientto expose the inner logic of their systems of thought; we
have also to take account of the relationshipof those systems to the
external social context, modified though human awareness of that
context may be the persistenceof antique categories of thought. As
Douglas has remarked:"It should never again be possible to provide
an analysisof an interlocking system of categories of thought which
has no demonstrablerelationto the social life of the people who think
in these terms."28
Geertz, however, maintainsthat faith in astrology or spells was
sustainedby a particularview of reality. Such faith could only decline,
she says,when "this deepersubstratumof convictions about the nature
of the universebegins to fall apart."Why then have I not exposed this
substratumin all of its detail? Well, to some extent I tried to do so. I
indicatedsome of the assumptionsunderlyinghealingritualsand witch
beliefs (Chs. 7, 14-16); and I discussed the rationalizations put forward

by Renaissanceintellectuals,with their microcosm and macrocosm,

and their animate universe-rationalizations,however, which I maintained had little to do with the actualpracticeof magic at the village
level (I85, I90, 222-223).
To the wizard, as to his clients, the source of
his power was often unclear. Recourse to him did not necessarily
reflectsubscriptionto some alternativeview of reality, any more than
a visit to an orthodox physicianindicateda cleargraspof the principles
of Galen (I9I, 257, 264). Men went in a spirit of "try anything which
works"; and the symbolism of the wizard's ritualswas highly limited
in its implications.I readilyadmit that I may have been less sensitiveto
the symbolic or poetic meanings of these magical rites than I should
have been.29But I am not convinced that a more sensitive observer
would find behind them a view of reality comparablein coherence to
28 Mary Douglas, "The Healing Rite (review article)," Man, V (1970), 303. Cf. her
assertion (on the basis of John Middleton, The Religion of the Lugbara [London, I960])
that "The only way in which a witch-dominated cosmology can be transformed is by
a change at the level of social organization" (Natural Symbols: Explorations in Cosmology
[London,

1970], I2I).

29 As is urged by Thompson,
Context," 49.

"Anthropology

and the Discipline of Historical

104

KEITH

THOMAS

that offeredby the theologians.I am prepared,in other words, to

underpinnings"
questionwhethermagicalwayshadthe "philosophical
with which Geertzcreditsit.
At this point it should be stressedthat anthropologistsdiffer
greatlyamong themselvesas to how much coherencethey should
expectto find when studyingthe beliefsof otherpeoples.Cognitive
seem to posit a unitary"culture,"albeit one comanthropologists
havealwayslooked
Frenchstructuralists
infra-cultures.
prisingseparate
for coherenceandsometimessurprisetheirmoreskeptical,empirically
Britishcolleaguesby the symmetryof
minded,behaviorally-oriented
the elegantlyarticulatedcosmologicalsystemswhich they claim to
Evenatthelevelof consciously
havefoundamongindigenouspeoples.30
havedeveloped
articulated
beliefsit is clearthatsomeanthropologists
schemeswhichin the opinionof othersgo farbeyondthe evidenceof
data.If thereis roomfor thistypeof argumentwhen
the ethnographic
we are dealingwith contemporaryAfricansor Indians,who can be
observedand questioned,how muchmoreuncertaintymust therebe

when we come to consider illiterate Englishmen who lived three or

four centuriesago. Of course we must persistin our effort to recreate
theirmentalworld. But in the presentstateof knowledge it is impossible
to maintainfor certainthat that world was a coherentone. In my book
I wrote that "what we are faced with in this period is not one single
code, but an amalgamof the culturaldebrisof many differentways of
thinking. Christianand pagan, Teutonic and classical;and it would be
absurd to claim that all these elements had been shuffledtogether to
form a new and coherent system" (627-628). As Levi-Strausshimself
admits, "the nearer we get to concrete groups the more we must
expect to find arbitrary distinctions and denominations which are
explicableprimarilyin terms of occurrencesand events and defy any
logical arrangement."In the sixteenth century even contemporary
intellectualsfailed to produce a genuinely coherent rationalizationof
magical practices.31
30 See A. I. Richards, "African Systems of Thought: An Anglo-French Dialogue
(review article)," Man, II (1967), 284-298.
31 Claude Levi-Strauss, The Savage Mind (London, 1966), I55. D. P. Walker, Spiritual
and Demonic Magicfrom Ficino to Campanella (London, I958), 75, 96. I am equally hesitant about adopting Thompson's suggestion ("Anthropology and the Discipline of
Historical Context," 51-53) that what contemporary authorities regarded as the religious
"ignorance" or "skepticism" of the lower classes was really a coherent alternative system
of religious symbolism. The popular utterances which I quoted (Ch. 6) seem too
heterogeneous to be easily fitted into any coherent alternative (or alternatives) to orthodox

RELIGION

AND

MAGIC,

II

I105

No doubt I should have looked not only for consciousrationalizations (which at village level are obviously seldom to be found), but
also for less conscious underlying structuresof thought. More, for
example, might have been said about the relationship of magical
methods to the widely prevailing conception of all knowledge as a
searchfor resemblancesand correspondences,and thus itself a form of
divination. For at this time the affinity of human beings and nature
was presupposed;and language itself was seen as part of the natural
world, rather than something external to it. Foucault,who has done
most to develop this theme, remarksof the intellectualchanges of the
seventeenthcentury that
This new configurationmay, I suppose,be called"rationalism";
one
might say, if one's mind is filled with ready-madeconcepts[!], that
of the old superstitious
the seventeenthcenturymarksthe disappearance
or magicalbeliefsandtheentryof nature,at long last,into the scientific
order.But whatwe mustattemptto graspandattemptto reconstitute
are the modificationsthat affectedknowledgeitself, at that archaic
level whichmakespossibleboth knowledgeandthe mode of beingof
whatis to be known.32
Here, I admit, my competence failed me. No doubt this abdication was the result of being rearedin an educationaltradition whose
productsmust inevitably recoil from Levi-Strauss'suggestion that the
investigator should attempt to transcendempirical observation so as
to achievea deeperreality.33But it is also the consequenceof approaching my subjecthistorically.Forhistorians,as Levi-Strausshasremarked,
tend to organize their data "in relation to conscious expressions of
social life," whereas anthropologistsproceed "by examining its unconscious foundations."34This dictum is obviously only a half-truth,
theology, though they do suggest a widespread tradition of materialism. Thompson
points to the coherent universe implicit in Thomas Hardy's The Mayor of Casterbridge,
but this is surely a clear example of the difference between art and life. Nevertheless, I
readily agree that my crude concept of "popular ignorance" needs a lot of refiing.
32 Michel Foucault, The Orderof Things: An Archaeologyof the Human Sciences(London,
I970), pt. I, ch. I; 54. It ought perhaps to be added that Foucault denies that he is a
"structuralist," attributing this aspersion to "certain half-witted 'commentators"' (xiv).
33 Levi-Strauss, "Introduction a l'oeuvre de Marcel Mauss," in Marcel Mauss,
Sociologie et Anthropologie (Paris, 1960), xxxiii. On the British empiricist's distaste for
any enquiry into underlying structures or hidden realities see David Goddard, "Anthropology: The limits of functionalism," in Robin Blackburn (ed.), Ideology in Social
Science. Readings in Critical Social Theory (London, 1972), 62.
34 Levi-Strauss (trans. Claire Jacobson and Brooke Grundfest Schoepf), Structural
Anthropology(London, I968), i8.

IO6 I KEITH

THOMAS

since, from the days of Marx, if not earlier, many historians have
sought to uncover the invisible foundations of society.35 But whereas
historians are quite used to dealing with the notion of underlying social
structures, they are much less accustomed to searching for invisible
mental structures, particularly the mental structures underlying
inchoate and ill-recorded systems of thought, which are only articulated
in a fragmentary way. These are structures of which the average member of the society concerned is, almost by definition, unable to give a
coherent account, any more than he can describe the analytical structure
of the language which he speaks. Indeed one anthropologist has remarked of the unconscious thought-structures of Levi-Strauss that they
tend to be "at least three degrees removed from the ethnographic
data."36
At a rather less inaccessible level, however, I would fully agree
that more justice needs to be done to the symbolism of popular magic.
Just as the mythology of witchcraft-night-flying,
blackness, animal
female
us
metamorphosis,
sexuality-tells
something about the
standards of the societies which believed in it-the boundaries they
were concerned to maintain, the impulsive behavior that they thought
it necessary to repress; so we can learn from the language of white
magic-sympathy and antipathy, narrative charms, and the symbolism
of salt or south-running water. But it remains to be established whether
these charms and rituals always constituted a coherent system or
whether, as is implied in the old-fashioned definition of"superstition"
(627-628), they were just unintegrated remnants of older patterns of
thought. At present it would seem common sense to assume that in a
changing society mental coherence is no more to be expected than
social coherence. Just as sociologists have to come to terms with the
fact that nearly every society contains institutions which are obsolete
or dysfunctional, so anthropologists have to be prepared for mental
inconsistencies. They also have to consider the problem of how to
handle the immense range of variations, chronological, social, and
regional, presented by a society as diverse as seventeenth-century
England; for the range of mental sub-universes is much wider than
35 Cf. Maurice Godelier, "System, Structure and Contradiction in Das Kapital,"
in Michael Lane (ed.), Structuralism:A Reader (London, 1970), 341; Levi-Strauss,
StructuralAnthropology,23.
36 Richards, "African Systems of Thought," 297; Aidan Southall, "Twinship and
Symbolic Structure", in J. S. La Fontaine (ed.), The Interpretationof Ritual. Essays in
Honour of A. I. Richards(London, I972), 74.

RELIGION

AND

MAGIC,

II | 107

that postulated by Geertz's simple distinction between literate and

illiterate, and the boundaries between them are far from clear-cut.
Anthropologiststend to feel a priorithat some coherencemust underlie
apparently conflicting symbolisms, just as a grammar is known to
underlie human discourse.But this coherence will not be found for
early modern England until some immense technical problems have
been solved.37 Meanwhile we must, I think, continue to question
whether a seventeenth-centurymagical practice (or for that matter a
modern superstition,such as a refusalto walk under ladders)is necessarily embedded in a closed system of ideas in the way that Geertz
assumes.
I fearthattheseratherdogmaticcounter-assertionsareno substitute
for the detailed discussionwhich Geertz's observationsdeserve. But
having ventured thus far into this rather abstract methodological
domain I would like, before retreatingfrom it, to suggest a few conclusions which seem to have emerged from this exchange of views.
The firstis that historiansmust recognize that much of their work does
not easily lend itself to cross-culturalcomparison(and this, I confess, is
something of which I am now much better aware than I was). This is
not to say that historicaldata should not, where possible, be presented
in a form suitablefor such comparison,but merely that the problems
of such comparison are much greater than is usually appreciated.It
remainshelpful to compare materialaspectsof differentcivilizations;
and it is also possible (though not easy) to compare differentkinds of
social structure.But when one enters the domain of "culture" and
ideas, or indeed that of any behavior in which the actor's intentions
become important, then the work of comparison becomes infinitely
more difficult, primarily because of the absence of any agreed set of
universallyapplicableconcepts. It is significantthat some anthropologists now despairof the possibilityof such comparisonand urge their
colleagues to concentrate on the cultural particularitiesof individual
societies38 while those who continue to offer global comparisons aim

primarilyto isolatethe naturalqualitiesof the humanmind (forexample,

the tendencyto group categoriesin sets of binaryoppositions),qualities
which are so timeless and general as to be of little historical interest.
The essential point, as one cultural anthropologist puts it, is "that
classificationsappropriateto a comparative study are on a different
37 It must be said that none of the sociological and anthropological works to which
Geertz refers (note 2I) even begins to handle variations on a comparable scale.
38 E.g., Needham, introduction to Rethinking Kinship and Marriage, cviii.

108 | KEITH THOMAS

conceptual level, serving different purposes, from the categorical

distinctions that make best sense of phenomena within a particular
society."39
My second conclusionis that historiansare going to have to come
to termswith the methods and approachesof structuralanalysis.When
I wrote my book, the anthropological monographs available were
mostly those written in the older functionalisttraditionand I fear that
Religionandthe Declineof Magicreflectsthat fact. Uncertain though I
remain about the methodology involved, I welcome the prospect of
more work by historianson the hidden structureof ideas.I differfrom
some anthropologists,however, in thinking that attention has to be
paid to the actualcontent of those ideas no less than to their structure.
Historianswill continue to be more interestedin local and temporal
differencesof content ratherthan in structuralsimilarities.
My third conclusion is simply the hope that the next few years
will see a sustainedonslaughton the variousproblemswhich my book
leaves unresolved. The task is both sociological and intellectual.We
need to clarify the human context in which magical practiceswere
invoked and witchcraftaccusationslevied; we also have to account for
the changing formation of mental structures.It would be wrong to
categorize the inquiry as primarilysociological or primarilyintellectual.
At presentit seems obviously both; though in the end we may have a
better idea of whether and how far intellectualchanges are related to
social ones. To that extent we are dealing with the very hardestkind
of historicalproblem, and one which, I suspect,neitherhistoriansnor
anthropologistshave yet directly faced up to.
Finally, there remain my own mistakes and limitations. Even if
some of Geertz's criticisms are off the mark, she is right to detect
sundry minor inconsistenciesof approach and definition.40But it is
fair to say that the main substanceof ReligionandtheDeclineof Magicis
what anthropologistswould call ethnographyratherthan theory; and
the ethnographyat least is, I hope, reasonablysound. I therefore take
comfort from the (possiblysad) fact that humdrum ethnographytends
to outlive even the most dazzling theoreticalconstruction.If I were to
39 Ward H. Goodenough, in Goodenough (ed.), Explorations in CulturalAnthropology:
Essays in honor of George Peter Murdock(New York, I964), 9.
40 I cannot accept her charge (above, 74) that, in crediting some medieval theologians
with a symbolic view of the sacraments, I am "projecting onto them an interpretation
which cannot possibly be theirs." In fact, such a view, though frequently overlaid, had
been in circulation since the time of St. Augustine or even earlier. See, e.g., C. W.
Dugmore, The Mass and the English Reformers(London, I958), 5, 7, I3-I4, 78-79.

RELIGION

AND

MAGIC,

II i 109

rewritemy field notesI thinkthatI shouldprobablynow castthem

into a slightlydifferentconceptualframework.But that is something
whichthe criticalreadercaneasilydo for himself.Besides,asthe great
historiansaid when he belatedlyrealizedthat he had postdatedthe
startof the declineof the RomanEmpireby overa century,"Of what
availis this tardyknowledge?Whereerroris irreparable,
repentance
is useless." 4I

41

EdwardGibbon(ed.J. B. Bury), TheDeclineandFall of theRomanEmpire(London,

1900), I, xxxv.