DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 120

1/2013

‘God(S)’ AS A CoMPARATIVE
CATEGoRy
STEVEn EnGlER & MARK Q. GARdInER

I denne artikkelen diskuterer vi hva slags meningsinnhold
et komplekst konsept som “gud” kan ha, og setter søkelys
på risikoen ved at andre konsepter assosiert med “gud” i
én kulturell sammenheng blir uaktsomt overført til bruk i
andre. Den sentrale posisjonen til “gud” i vestlige og
kristne kontekster gjør denne risikoen spesielt påtrengende,
og på bakgrunn av dette argumenterer vi for at termen bare
kan bli hensiktsmessig anvendt som et komparativt konsept
innenfor snevert definerte kontekster.

håkon Tandberg raises an important question when he asks, ‘Is the fire
a god, in Parsi Zoroastrian ritual?’ (Tandberg 2012).1 More than clarifying the status of one element of one religion, he calls into question
the value of ‘god(s)’ (and ‘goddess[es]’) as a comparative category in
the academic study of religion. (We will refer to ‘god’ to make the point,
but this broader set of terms should be kept in mind.) We might ask further whether other categories and concepts should be interrogated in
the same light: do concepts such as deity, soul, sacrifice, salvation,
heaven and others warrant similar critiques? Taking a closer look at
‘god’ as a comparative category can cast some valuable light on these
matters.
The core issue is how might we weigh the risks and benefits of using
certain first-order (insider religious or emic) terms as second-order
(scholarly comparative or etic) terms. This question needs to be answered on a case-by-case basis: the risks and benefits – above all confusion and clarity – vary from term to term and context to context. In
the case of ‘god(s)’, the down-side outweighs the up-: the term is best
avoided as a broad category in the comparative study of religion.2
120

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 121

1/2013
We are not just arguing against the viability of ‘god(s)’ as a crosscultural, pan-historical concept. It hardly warrants arguing that the categories and concepts used in the human and social sciences cannot
aspire to such absolute application. It is a false dichotomy to insist that
either ‘god(s)’ is a narrowly circumscribed insider category (e.g., rife
with Christian presuppositions) or it can serve to describe relevantly
similar phenomena in all possible ‘religious’ contexts. As always, the
theoretical and methodological meat of the matter lies in the messy
ground between these extremes of particular and universal. By way of
making an initial foray into this no-gods-land, we will highlight two
distinct problems with the word ‘god’ as a comparative category. The
first arises from contingent and historical matters-of-fact, whereas the
second is a result of improper attention being paid to the nature of meaning in general and of the meaning of ‘god’ in particular. These two
sources are related, but we will only here hint at their connection. The
full argument would require more in-depth theorizing that we will leave
for another time and place.
Problem 1: semantic baggage. In English-language usage, ‘god’ is
associated most frequently with the Christian God. From this statistical
fact emerges a conceptual and normative bias with the insider’s usage
spilling over into the scholarly texts and discourses of the study of religion. A good example is the one chapter on ‘God’ in the series of ‘companions’ and ‘handbooks’ to the study of religion that has been
published over the last fifteen years. The discussion in that isolated
chapter is almost entirely limited to the Christian concept:
other terms, such as the “holy,” the “sacred,” “divinity,” the “supernatural,” the “mystical,” and so on, have been proposed as articulating what
is at the heart of all religion, and thus as best denoting the proper object of
religions. none of these common nouns, however can or does contain either the comprehensiveness or the specificity of meaning to which the
proper name “God” adverts, with its positing of an ultimate point of reference in terms of which all realities must be understood, and its claim that
all aspects of human life therefore…should be oriented in terms of this reference point. (Fiorenza and Kaufman 1998: 154)

121

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 122

1/2013
of course, the distinction between ‘God’ as proper name and ‘god’
as common noun is implicit here, though this also serves to highlight
our point. Whether or not etymologically accurate, the proper name is
taken by most speakers of English to be predominant while the common
noun is seen as derivative, just as the classical Greek term διός (dios)
elides with θεός (theós) with both appearing to be derived from the
proper name Ζεύς (Zeus). In other words, most English speakers tend
first to be exposed to the Christian connotation of ‘God’ and subsequently understand ‘god’ in a derivative, usually deviant or deficient,
sense. Conceptually associated with ‘God’ but lacking in ‘god’ is a sense
of necessary existence and perfection, but as a derivative concept many
of the other monotheistic associations – principally a sense of ultimacy
or absolute transcendence – tend to be misleadingly preserved in contexts involving ‘god’. The near syntactic identity of ‘God’ and ‘god’ encourages this tendency, and without some sort of universally recognized
special quotation marks or other syntactic devices for distinguishing
between them, the danger of association-transfer is significant.
The problem of semantic baggage is a relative one: some first-order
terms will bring more associations than others when they are taken up
by scholars as second-order terms, and some scholars will be more sensitive to the dangers of illicit transfer than others. nonetheless, where a
term is as central and distinctive as ‘god’ is in the Christian context, and
where the preponderance of theorizing in religious studies has been anchored in Western (and largely Christian) traditions (as argued, for example, by Vásquez [2011]), the weight of that usage is sufficient to
count against our using the term as a general comparative category.
‘God’ is more trouble than it is worth because of the difficulty in firewalling the term against its most common monotheistic meanings. In
their article on ‘God,’ Fiorenza and Kaufman effectively grant this
point: after discussing the biblical ‘strand of meaning’, they note that
‘the word “God” itself, as employed today, inevitably carries traces of
all this complex weight of meaning’ (1998: 141).
The problematic transference of association from ‘God’ to ‘god’ is
also replicated in the use of the same type – ‘god’ – in different contex122

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 123

1/2013
tual tokens. For example, hollywood (e.g. Thor, The Mummy,
Clash/Wrath of the Titans) capitalizes on a common association of ‘god’
in much Western mythology with anthropomorphism (hence liam neeson can portray Zeus without much CGI) and anthroporationality (deliberative will or intentional agency). That these facets of the characters
do not need to be explained to the transcultural audiences is telling. Applying ‘god’ to contexts not connected to the western monotheistic
canon raises more than a mere worry of unwarranted transference of
association.
It is easy to distinguish further and deeper senses of this first problem, e.g., historical or post-colonial critiques that see ‘god’ as carrying
pre-modern or western presuppositions of more specific sorts. We trust
that readers can rehearse the trajectory of such arguments. The basic
problem remains the same: the term ‘God,’ and by transference of association ‘god,’ have semantic inertia that brings with them predominant
associations in a manner that distorts the attempt to engage in historical
and cross-cultural research. For example, as david Chidester (1996)
notes for the case of South Africa, many Europeans did not see ‘religion’
in the beliefs and practices of colonized peoples because they failed to
find something that corresponded to their Christian preconceptions of
the meaning of ‘God,’ whereas, at the same time, others insisted that
these people must have religion. The basis of the latter view was that –
as dutch traveler, Johan nieuhof, put it – ‘according to the unanimous
opinion of all theologians, no folk in the world is so barbarous that it
does not honour some Godhead, be it true or false’ (cited in Chidester
1996: 39). This is more than a case of Europeans filtering their understanding of other cultures through their own presuppositions. Both sides
of this debate – those who saw religion among certain groups and those
who did not – tended to share a particular sense of what it was they
were looking for in ‘god.’ As signified by nieuhof’s use of the term
‘Godhead,’ the issue was not whether the sense of ultimacy was appropriate but whether ultimacy was present in some supernatural being postulated by the culture. Both sides of the debate took an overly narrow
meaning of ‘god’ for granted.
123

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 124

1/2013
It might be thought that the problem of semantic baggage will arise
only amongst those untutored in the academy – i.e., that scholars develop habits of caution that inoculate them against this problem. We are
skeptical of such optimism. We see the same potential problem – albeit
more subtle in form – even within communities of scholarly usage. It
is perhaps not surprising that the chapter on ‘God’ in Critical Terms for
Religious Studies is biased toward Christian usage, given that it was
written by Roman Catholic theologian Francis Schüssler Fiorenza and
Protestant theologian Gordan d. Kaufman. But that very point underlines a corollary to the problem of semantic baggage. The academic
study of religion both includes various sub-disciplines (e.g. textual
hermeneutics and study of ritual) and is pursued by various super-disciplines (e.g. sociology and psychology). like it or not, ‘theological’
discourses appear in the same books, journals and conferences as do
‘religious studies’ discourses, more narrowly defined. Given the value,
at least in certain contexts and circumstances, of emphasizing rather
than eliding such sub- and super-disciplinary distinctions, we should
choose our technical vocabulary in a way that allows us to distinguish
the respective theoretical and methodological agendas.
our pessimism about the academy successfully policing itself is
based more fundamentally on what the term ‘god’ is being asked to do.
It is expected to do double duty as a specific marker of monotheistic
ultimacy and as a more generic comparative term for some sub-set of
supernatural beings. no single term can do both of these things. This is
not a failing in our academic training, but rather is a consequence of
what sort of meaning words can have and what implications this has
for their usage. This leads to the second problem.
Problem 2: semantic indeterminacy. The term ‘god’ is a complex
term, especially when used as a broad comparative category. It reflects
a variety of other concepts that might or might not be associated in a
given case. In our discussion of problem 1, we highlighted ultimacy,
anthropomorphism, and anthroporationality, but a host of others appear
on reflection: e.g., absolute, all-encompassing, compassionate, controller, creator, divine, animal-spirit, embodied, disembodied, incorpo124

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 125

1/2013
real, energy, essential, eternal, extraterrestrial, father, mother, formlessness, genderless, good, evil, grace, hidden, holy, immortal, immutable,
incarnate, ineffable, infinite, in-relation, beyond-relation, intelligence,
judge, lawgiver, mind, mystery, nature, omnibenevolent, omnipotent,
omnipresent, omniscient, one, necessary existence, possessing, posthuman, transhuman, punisher, ruler, sacred, self-revealing, set apart, divinely simple, trans-historical, spirit, supernatural or transhuman power
and knowledge, supreme being, transcendent, immanent, transpersonal,
personal, impersonal, trinity, true self, unknowable, revealed, unmoved
mover, vengeful, whole, willful, worshiped, wrathful, etc. The fact that
so many terms on such a list resonate with Christian conceptions is a
further illustration of the problem of semantic baggage.
The word ‘god’ itself does not specify, at least not a-contextually or
trans-contextually, whether such an entity is good or evil, gendered or
genderless, hidden or revealed, personal or impersonal, etc. Any use of
the term, then, must be supplemented by additional, more specific, contextually associated terms in order to clarify its meaning. Why? Because
the meaning of ‘god’ – indeed of any complex term or concept that has
comparative aspirations – is not given by analytic relations to these
other associated concepts. By this we mean that the term ‘god’ is not
defined as the entity that satisfies the conjunction of these other terms.
They do not specify the essence of god(s), if ‘essence’ (another complex
term!) is understood as logically strict, necessary and sufficient conditions. They are not predicates that are somehow, in Kantian fashion,
contained in the subject-expression ‘god(s)’.
This is clear, for one reason, because many of the terms in the list
are incompatible with each other and so cannot coherently be bundled
into necessary and sufficient conditions. More significantly, though,
such a conception of the meaning of complex terms would preclude
their being successfully used in comparative ways. That is, terms displaying any deviation from the ‘defining’ set would have to be regarded
as literally different concepts; yet, numerous perfectly sound and rigorous inquiries have shown there to be such ‘deviant’ but otherwise perfectly good uses of comparative terms. If we were to define ‘god’ in
125

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 126

1/2013
analytic or essentialist terms, its use as a comparative concept in religious studies would be ruled out of bounds by linguistic fiat: not because it would not be useful, but because it could not be comparative.
An essentialist definition of ‘god’ would tie it rigidly to whatever list
of concepts were said to characterize it, and this would undermine its
comparative use: i.e., it could not be extended to encompass other elements not on that list. of course, there is a sense in which the term could
be comparative under these conditions: it could simply be forced to
apply in cross-historical and cross-cultural contexts. however, this
would require that one particular set of network associations be taken
as canonical or privileged or definitional. This would elevate the problem of semantic baggage to the level of ethno-lingo-centrism.
The more palatable view, theoretically and methodologically, is to
recognize that the meaning of distinct uses of ‘god(s)’ is best understood
in terms of specific networks of associated concepts that fully reveal
themselves only in concrete contexts of usage. no single network of
associations is found across all contexts. This ‘network association’ conception of meaning – or what has come to be called ‘semantic holism’
– offers a more nuanced way of making sense of the danger of semantic
baggage, and it underlines that specific historical facts about the hegemony of Christian and western associations make this semantic possibility an actual threat to the comparative project of the study of religion.3
It is important to note that the problem is not that ‘god(s)’ is ambiguous or ambivalent – i.e., that it simply means different things in different contexts. ‘God’ does not have multiple meanings in the way that
‘bank,’ ‘crop,’ or ‘pitch’ do. Rather, the comparative researcher will discover substantial overlap in the associated networks between any two
contexts of usage despite the fact that no two networks will be identical.
Some semantic holists – e.g. W.V. Quine and donald davidson (e.g.
Quine [1960] and davidson [1985]) – have described this consequence
of meaning as semantic indeterminacy. For terms that have this type of
semantic complexity, because no specific network will have semantic
privilege, no single determinate meaning can be assigned to them. But
this does not mean that they are meaningless, or that their meaning is
126

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 127

1/2013
arbitrary, or that their meaning is strictly relative to particular contexts.
The guaranteed overlap of subsets of associations will preclude these
views. Indeed, it is this overlap that makes it possible for such terms to
be used in non-incommensurable, and hence comparative, ways. To give
a trivial example, distinct meanings of ‘grain’ are clear from associations with other words in the following two phrases: ‘cut against the
grain’ and ‘harvest the grain.’
To summarize, our point is not that ‘god(s)’ cannot be used comparatively but that the term’s disadvantages outweigh its advantages for
broad comparative work in the study of religion. It certainly can be useful in more limited comparative contexts. After all, ‘God’ works for
Christian theologians of both apophatic and kataphatic stripes. In this
case, however, the core commitment to the ultimacy of God offers a delimiting characteristic that keeps the semantic domain sufficiently narrow for scholarly comparison, even across a range of relevant
variations. however, in the case of the broader cross-cultural comparative use of ‘god’, the various uses that scholars could be expected to
lump together have only very general characteristics in common: e.g.,
supernatural being or culturally postulated superhuman agent. So why
not just use those more general terms? The semantic domain of ‘god’ is
simply too broad to be of much comparative value. The Christian use
of ‘God’ (as the Muslim use of ‘Allah’) works because it leaves open
certain issues, but within a relatively constrained field of meaning. The
broad comparative used of ‘god’ is too open. Too much is left to specify.
As a result, it is best to drop the term ‘god’ in comparative discussion
and to move directly to the more specific concepts that characterize
cases in comparative studies.
To give a concrete example of this problem, we can turn to early
twentieth-century French Indologist Sylvain lévi who, according to
Ivan Strenski,
showed how Vedic and Brahmanic sacrifice assumed that ritual itself actually produced the gods. This meant, first of all, that the definition of religion could be separated from a belief in the existence or even the idea of
God. Sylvain lévi says that the nature of the religion revealed in the Brâh127

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 128

1/2013
manas is constituted by sacrificial ritual. Thus sacrifice “is God and God
par excellence” (Strenski 1996: 33 [text and note], original emphasis).4

’God’ here is equated with ‘the nature of the religion.’ This is so general
that the term ‘god’ itself does no useful work, apart from, perhaps, an
implicit othering of Christian conceptions of the term. It is either so
general that it might as well be dropped or, if we credit the latter motivation, it constitutes a reductio ad absurdum of the cross-cultural use
of the term.
There is an additional aspect of semantic indeterminacy that further
limits the usefulness of ‘god’ with respect to available definitional and
consequently translational choices. not only will the networks differ
from context to context, there will be overwhelming practical difficulties in delimiting even a single network. Given that the meaning of a
complex term is given in relation to elements of the associated network,
in cases where each of those associated nodes is also complex, then
their meaning will be given by a further extension of the network, and
the same holds for all such complex nodes in the extension. This raises
the methodological difficulty of completing a network for even a single
context of use, i.e. of usefully providing a definition relative even to
one particular context. ‘God’ is especially prone to this problem. For
example, The HarperCollins Dictionary of Religion (HCDR) defines
‘God, Goddess’ as ‘common term for the supreme deity’; ‘god’ as ‘common term for a male deity’; and ‘goddess’ as ‘common term for a female
deity’ (Smith 1995: 389). It defines ‘deity’ as ‘general term for a god or
goddess; in modern Indo-European languages, a synonym for god’
(Smith 1995: 310). This tight circle avoids begging the question of the
general meaning of ‘god’ only in the subtle shift from ‘common term’
to ‘general term.’
on the one hand, ‘deity’ is perhaps more useful than ‘god’ as a general category in that it may have less semantic baggage. It thus also perhaps avoids the problem of communities of scholarly usage to a greater,
if not total, extent. In addition, the fact that it is such a close synonym
for ‘god’ raises the question of why one would use a different term at
all, and this difference is easily interpreted precisely as the move toward
128

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 129

1/2013
a less Christian-specific term. That is, ‘deity’ may be, in practical terms,
a way of saying ‘god as a more general term.’ Minimally, its use is
preferable for that reason. on the other hand, ‘deity’ remains just as
complex and hence semantically indeterminate a term as ‘god.’
More usefully, the HCDR’s entry for ‘God, Goddess’ includes the following reference: ‘See also term question,’ (Smith 1995: 389). This draws
our attention to the issue of translation. The ‘term question’ was a nineteenth-century Protestant theological debate over the relation between
‘God’ and two terms from Confucian thought: ‘Shang-ti’ and ‘T’ien.’
‘God’ is an English word used in English-language discussions of religion
(leaving aside dutch). Given the overwhelming probability of differing
associated networks we cannot take for granted that the meanings of the
following words are the same: allah, batara, bóg, boh, Bože, bůh, deus,
dieu, dio, dievs, dios, бог, got, gott, gud, hyjni, isten, Jumala, θεός, zeu.
This applies even more so with terms such as Shang-ti’, deva/dewa, orixá,
kisemanito, Waaq or Jaumirawo. Consequently, any attempt to inter-translate them should be regarded with deep skepticism.
In one sense, this is an unavoidable problem with any appropriation
of first-order terms for use as second-order terms. ‘Sacrifice’ in English
and French and ‘sacrifício’ in Portuguese all diverge in meaning due,
for example, to distinct nineteenth-century Western-European discourses around individual responsibilities to the nation and to the latetwentieth century emergence of neo-Pentecostal discourses on tithing
in Brazil. In this light, we might consider using ‘god’ tentatively, awaiting confirmation that the usages in different languages – and in distinct
historical and cultural contexts – conform to our hypothesis of a relatively stable core of meaning.
This is not what we advocate. Given the problem of semantic indeterminacy noted above, the meaning of ‘god’ – even across Englishonly contexts – does not serve as a stable basis for translation. on the
hypothesis that comparable terms in other languages are equally indeterminate, the attempt to use ‘god’ as a comparative category reduces
to a futile attempt to match up some indeterminate sub-set of one broad
bundle of meanings to some indeterminate sub-set of another.
129

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 130

1/2013
More fundamentally, we need to move past a facile and spurious
conception of translation as a one-to-one mapping between the elements
of two distinct languages (see Engler and Gardiner, forthcoming). one
potential solution would be to analyze terms like ‘god’ in family-resemblance terms, though in ways more complex than Wittgenstein’s original
notion envisioned (Wittgenstein [1953]). This would provide a way to
conceptualize the problem of semantic baggage: monotheistic conceptions of ‘god’ constitute a workably constrained family; broad comparative uses of ‘god’ constitute an unworkably extended family.
noTES
1. during a recent visit to Bergen, one of the authors (Steven Engler) was able to sit down
with håkon Tandberg and Knut Melvær, the editors of this special issue, to discuss
these questions in greater depth. he thanks them for stimulating his thinking on these
matters, and for pointing out the Ketola and Pyysiäinen chapter.
2. Kimmo Ketola and Ilkka Pyysiäinen (1999) analyzed ‘the concept of “God” as a category in comparative religion.’ They argue that ‘such concepts as “tabu”, “sacred”,
totemism”, etc. have all been lifted from religious contexts and have already been quite
carefully problematized and thus made into etic categories for comparative use. only
“god” is still used without any well formed criteria for its operationalisation’ (1999:
208). They concluded that the term ‘god’ cannot escape from its ‘implicit Judeo-Christian assumptions’ in order to become a properly ‘scientific’ term. We go beyond Ketola
and Pyysiäinen in more fully assessing the nature of this problem, in noting correlated
problems, and, more fundamentally, in bringing out the semantic dimensions of these
issues.
3. For a brief discussion of semantic holism and its implication for the study of religion
see Engler and Gardiner 2010.
4. Citing Sylvain lévi, La Doctrine du sacrifice dans les brahmanas, Paris: lerous, 1898.
REFEREnCES

Chidester, david. 1996. Savage Systems: Colonialism and Comparative
Religion in Southern Africa. Charlottesville and london: University
Press of Virginia.
davidson, donald. 1985. Inquiries into Truth and Interpretation. oxford: oxford University Press.
Engler, Steven, and Mark Q. Gardiner. 2010. ‘Ten Implications of Semantic holism for Theories of Religion.’ Method and Theory in the
Study of Religion 22/4: 283–292.
130

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 131

1/2013
Engler, Steven, and Mark Q. Gardiner. Forthcoming. ‘Translation.’ In
Kocku von Stuckrad and Robert A. Segal, eds. Vocabulary for the
Study of Religion. leiden: Brill.
Fiorenza, Francis Schüssler and Gordan d. Kaufman. 1998. ‘God.’ In
Mark C. Taylor, ed. Critical Terms for Religious Studies. Chicaco:
University of Chicago Press. 136–159.
Ketola, Kimmo, and Ilkka Pyysiäinen. 1999. “Rethinking ‘God’: ‘God’
as Category in Comparative Religion.” In Approaching Religion:
based on papers read at the Symposium on Methodology in the Study
of Religions held at Åbo, Finland, on the 4th–7th August 1997, edited by Tore Ahlbäck. åbo: The donner Institute, 207–214.
Quine, W.V. 1960. Word and Object. Cambridge: The M.I.T. Press.
Smith, Jonathan Z., ed. 1995. The HarperCollins Dictionary of Religion. new york: harperCollins.
Strenski, Ivan. 1996. ‘Zionism, Brahmanism and the Embodied Sacred:
What the durkheimians owe to Sylvain-lévi.’ In Thomas A. Idinopulos and Edward A. yonan, eds. The Sacred and its Scholars: Comparative Methodologies for the Study of Primary Religious Data.
leiden, new york, Köln: E.J. Brill. 19–35.
Tandberg, håkon. 2012. ‘Is the Fire a God? The Case of Parsi Zoroastrianism.“ Paper presented at the annual conference of the European
Association for the Study of Religions (EASR), Stockholm, August
24.
Vásquez, Manuel A. 2011. More than Belief: A Materialist Theory of
Religion. oxford: oxford University Press.
Wittgenstein, ludwig. 1953. Philosophical Investigations. oxford:
Basil Blackwell.
ABSTRACT

In this paper we argue that, despite the fact that the term ‘god’ may be
used effectively as a comparative concept in the study of religion within
narrowly circumscribed contexts, the risks of doing so as a broad crosscultural category outweigh any possible benefits. We advance an account of the kind of meaning that complex concepts, like ‘god’, have.
131

DIN 2013-1 ombrukket 2_DIN 23.01.13 15:38 Side 132

1/2013
This account guarantees a risk that certain further concepts that are associated with ‘god’ in some cultural contexts will be illicitly transferred
to its use in others. The centrality of ‘god’ in western and Christian contexts makes this risk particularly acute, to the point of not being worth
the trouble.
KEyWoRdS: God; meaning; holism