Skip to navigation – Site map

HomeThematic Issues18On Varanasi’s Tiny Temples


In this paper, I present the results of a comprehensive survey of the public temples of the city of Varanasi in north India. Considering the city’s total public temple environment, I argue that an emphasis on major centers of pilgrimage and other large temples has obscured the prevalence of the “tiny” temples which are far more numerous and widespread. These temples do much important work of Hinduism, bringing the power of divinity and the legal weight of the Hindu tradition to bear on the lives of their devotees as well as the many others who encounter them each day.

Top of page

Full text

  • 1 Eck (1982:109), citing the Kāśikhaṇḍa section of Skandapurāṇa, 58.63.

“In Kāshī it is said that there is a linga at every step, and that there is not a place as big as a sesame seed that does not sprout a linga.”1

  • 2 Research for this paper was generously funded by the Metzger Fund and the Faculty Development Commi (...)

1Varanasi is, by practically every estimation, the most sacred Hindu city and the greatest center of Hindu pilgrimage.2 Hindus travel from all corners of India and, today, from all over the world, to visit the ghats and dozens of famous temples. For over a thousand years now, Varanasi’s sacred geography—the focus of so much of that pilgrimage—has received no small amount of attention. From Sanskrit authors in the first millennium to a number of modern academics, temples, pilgrimage routes, the Ganga, and several other features have been cataloged, mapped, described, lauded, and more. For many residents of the city, though, the unsung, local, unostentatious “tiny temples” which dot the city play just as great a role in day-to-day religiosity.

  • 3 We in fact located only a half dozen or so temples north of the Indian Railways rail lines to the n (...)

2After completing an initial and informal survey in 2013, in the summer of 2015, I carried out a comprehensive survey of conspicuous public temples in Varanasi, Uttar Pradesh. Along with three undergraduate assistants and Varanasi resident Binit Kumar Misra, I canvassed the city and documented the location and basic details of every temple we found using GPS and smartphone technology. While there are a number of different versions of what constitutes “Varanasi,” this project graphed the area that most would think of as constituting the modern city, bounded by Banaras Hindu University in the south, the Ganga River in the east, the Grand Trunk Road and Indian Railways lines in the north, and the Chunar / DLW Road to the west. While we also documented some shrines beyond these boundaries,3 some 99 percent of our records are from within that circumscribed area. We have rendered the data into graphic form using data formatted through the Fulcrum Application, and compiled and visualized it using ArcGIS (Figure 1).

3This paper makes one primary claim, that tiny temples have been significantly underrepresented in both academic and traditional depictions of Hindu religious landscape, and that they merit greater attention than they have received thus far. By observing where temples occur and the other features of landscape with which they frequently coincide, we can learn how temples instantiate the presence of deities at important junctures of shared space. This initial presentation of our data also suggests a number of insights about the roles that tiny temples play in making Hinduism for the residents of Varanasi.

Figure 1

A basic Google map of the area covered by the survey

4In sum, we recorded 3,347 public religious structures in Varanasi (see Figure 3). By public, I here mean publicly accessible; thus, while we would map a temple in the corner of a shopping mall parking lot, we would exclude one inside the gate of a private home. While I expect this wealth of data will continue to provide the basis for a number of insights, in this paper I examine the implications of one of the more obvious facts that the data revealed, namely, that in Varanasi, what I call “tiny temples” are vastly more numerous and prevalent than their larger counterparts.

Figure 2

ARCGis Map of Varanasi depicting distribution of all temples located during survey

5To date, numerous studies of Hindu temples have been carried out. For the most part, these have addressed single temples, and usually ones that are large enough to serve as the center of a community, sustain a staff of attendant priests, etc. A cursory review quickly reveals scholarship such as Fuller’s work on the Mīnākṣī temple in Madurai (2003) or Afsar Mohammed’s on Veṇkaṭeśvara (2010), both of which focused on a single temple. Although treating new temple communities in new areas, both Waghorne (2004) and Dempsey (2006) also anchored their studies to one or a few significant major temples, for example. In assessing the role of temples in patrons’ lives, much as this paper seeks to, Appadurai and Breckenridge, too, focused on major large temples (1981). These are but a few examples. Of course, each of these works considers primarily not the temple structure itself, but instead the activities that happen within, and the relationships and experiences occurring among members; however, we are left with the distinct impression that the large central temple, in the model of the ancient impressive monoliths, serves as the center of Hindu community life.

6This very notion of the temple has dominated scholarly thinking about what temples are and the ways that they function. The two landmark works in the field, by Stella Kramrisch (1946) and George Michell (1977), are each entitled The Hindu Temple; a more recent work by Tiziana Lorenzetti, Understanding the Hindu Temple: History, Symbols, and Forms (2015) reflects the continued influence of this pronominal proclivity. Each of these studies recognizes that there is a variety of styles of temples (here and elsewhere across the literature, this variety usually refers to there being North Indian and South Indian temples.) They also frequently make mention of the existence of the incredible number of shrines of all sizes that can be found anywhere one goes in India. However, these and other studies rely heavily on textual Sanskrit sources for information about large, important, classically styled temples. This tendency results from a certain preconception which is nicely exemplified by Lorenzetti as follows:

  • 4 Lorenzetti 2015, 15; while n.15 notes that “[i]t should be remembered that Hinduism (a term coined (...)

Before speaking about the origin and evolution of the Hindu temple, we must allude to certain fundamental philosophical concepts, common to most speculative thought, which—as is the case with the artistic expression of any religious creed—are intimately linked to temple architecture and structure.4

This sort of sentiment is neatly echoed elsewhere as well:

  • 5 Michell 1977, 61.

The fundamental preoccupation of Hindu thought is with man’s release (moksha) from an illusory world into which he is recurringly born. The architecture of the Hindu temple symbolically represents this quest by setting out to dissolve the boundaries between man and the divine.5

  • 6 These are two genres of Sanskrit text. Māhātmya means greatness; these are panegyrics of persons, p (...)

7A number of problems in these claims will be readily apparent to the trained eye of a religious studies scholar; others will emerge in the course of this paper. Lorenzetti here assumes that doctrine, and indeed, “philosophical concepts,” are the guiding source of and provide the impetus for both individual temples and for the “origin” of Hindu temples. The notion of “Hindu” did not, of course, exist at the point of the origin of even the major monolithic temples, let alone their predecessors. Further, the assumption that philosophical concepts, which are abstract, highly developed, strongly textual, and usually the exclusive province of a literary elite, are the fundamentals and the effective essential DNA of Hindu temples is just that—an assumption, and one that tends not to hold up under scrutiny in this case or in that of “any religious creed.” Many Hindus lack any real fluency in these elite textual philosophies, and many know nothing of them at all. We might also wonder how philosophical doctrines that are often diametrically opposed could have produced so many similar temples, or how, if they are intimately linked to temple architecture, so many temples are completely free of the elements that these doctrines mandate. In the end, this thinking is born of a sort of essentialism that insists on a correlation between “Hindu” persons and “Hindu” texts. In other words, Hindu identity somehow produces a Hindu-ness within a Hindu person that also imbues her with a full knowledge of Hindu theology. A number of scholars have problematized such an emphasis on texts. Ronald Inden has pointed out a deterministic and Orientalist tendency that conflates persons with theology or books denies individual actors agency (1991). In the field of Buddhist studies, Gregory Schopen has argued that texts are misleading because they are often the province of the elite, for example, and that textual sources are too often given epistemological priority over archaeological and other “on-the-ground” evidence (1997). A central concern of this paper is to consider those temples that māhātmyas, vastuśāstras,6 pilgrimage guides, and other texts ignore, and to point out that for many residents of Varanasi, smaller local temples form a greater part of their experience than do the larger, more famous counterparts.

8Scholarship on Hindu temples has unpacked features such as parikrama (circumambulation path), garbhagṛha (innermost chamber housing the main deity), porticos and columns, śikāra (crest over the central chamber), vimāna (main temple building), maṇḍapa (entrance hall), and mūrti (statue of the deity), sacred geometry, and geomancy, all the while emphasizing the role of the temple as a tīrtha which allows a crossing into the realm of the gods, and often with the understanding that this is a critical element of the essential Hindu desire for mokṣa. These elements are, no doubt, prevalent and significant in a great many temples, and are frequently expounded as bearing a precise equivalence in spatial symbolism to the concerns for complete release from the phenomenal world. I do not seek to negate those claims in this paper. Instead, I wish to point out an overlooked mass of Hindu religious structures which may consist of little more than a mūrti itself and the barest of enclosures. I argue that the emphasis on grander Hindu temples has obscured our view of these smaller, more quotidian houses of gods, or has, perhaps unintentionally, resulted in the notion that they are deviations and departures from the norm, and somehow not truly “temples,” unquestioned though that term itself has remained.

  • 7 The day after he was elected Prime Minister of India in 2014, Narendra Modi visited the Kāśi Viśvan (...)
  • 8 Gutschow 2006 and Gengnagel 2008 address the Pañcakroṣi yatra and its associated mandir.

9A number of factors prompt a re-examination of the idea that the smaller temple is the departure from the norm of the major temple. According to both archaeological textual records, small temples or god-houses (Sanskrit: devagṛha or devālaya) predated stone monoliths by several centuries (Olivelle 2010), with the advent of the major monolithic temples coming only in the fifth century (Kramrisch 1946). That date also means that the birth of the major temple coincides neatly with the Gupta dynasty, the first long-lived and widespread imperial formation in India. From the outset, and up to the present moment,7 major temples have always been associated with ruling power and polity (Sears 2014). The local, village, neighborhood, wayside, or other small shrine is much more the province of the average Hindu. Moreover, while virtually all studies have focused on individual temples, pilgrimage circuits,8 or clusters of temples (often as outlined in Sanskrit texts—for example see Singh and Singh 2006) in Varanasi and elsewhere, Hindus (and others) encounter multiple discrete temples in their environments. To date, no study that I know of has tried to analyze the possible impact of a total religious landscape of variegated temples, of which, perforce, most will be smaller.

  • 9 Maṇḍalas are circular geometric designs associated with divinity and ritual. The term is also used (...)

10While there have been many maps and lists made of Varanasi’s temples, a number of studies have pointed out that depictions of the city of Varanasi, maps of the temples within and of circuits to be traveled by pilgrims, and illustrations of the city’s sacred geography have been shot through with political and economic concerns, and have relied heavily on textual sources. Religious maps like the Kāśīdarpana were more concerned with cataloguing those temples that their creators thought most important than with giving an accurate or reliable rendition of the location of those structures (Gengnagel 2003). They were often stylized depictions of the city as maṇḍalas,9 and in many cases constitute visualizations of textual depictions (Gengnagel 2011). Authors of guides to the city’s temples omitted some shrines and lauded others in efforts to direct traffic in certain directions; these depictions were composed in Sanskrit by Brahmans with vested interests in promoting one vision of the city’s power over others (Desai 2017:19). Even if mapping could somehow be a value-free practice, and those creating maps could lay aside their own mundane interests, different groups bring different religious orientations to their ideas of the organization of space:

  • 10 Gaenszle and Gengnagel 2006. They point to two further studies in their 2006 volume, Gengnagel (200 (...)

Brahmans and other high-caste experts of the written tradition tend to organize a complex and disparate reality according to holistic and encompassing patterns … whereas lower castes have their own set of ideas and are generally less concerned with such abstract models …10

The textual sources which inform and shape religious mappings of the city thus always already owe allegiance to the Brahman conception of well-ordered and hierachicized power. British colonizers supported and encouraged Brahmanic views of ritual and pilgrimage in an effort to curry favor with Hindus from all of India (Desai 2017:10). Varanasi has been destroyed and rebuilt, and virtually nothing in the city predates 1600 (Desai 2017); moreover, since the time of Independence, the city has grown over 500% in population, expanding west and south away from the Old City (Singh 2008:46–50). All this means that earlier maps were not composed with accuracy in mind, and the city has evolved since the time of their production. Our effort to comprehensively chart the city’s temples using modern technology is thus both unique and necessary.

11Even the great megalith and pilgrimage center temples are encountered against a present backdrop of the other smaller shrines in their vicinity, and the remembered history of a worshipper’s experience with the other smaller temples that they likely see in their day to day lives. Thus, we cannot truly comprehend the work of even the major temples without knowing something about how the more numerous and prevalent smaller temples condition the subject’s encounter with them. As Bakhtin would have it, the meaning of each individual temple-text that a worshipper encounters comes into dialogue with all other temples that she has seen before, and creates a new and unique meaning in that subject’s mind on that basis (1986:106–07; 1981:276).

12In this paper, I argue that we ought to widen our gaze to include a broader number of smaller, humbler, less central temples; at the same time, we should sharpen our focus to bring these smaller structures into view, and to recognize the various and important roles that they play in Hindus’ lives. In the course of so doing, we will also gain a richer and more detailed understanding of the work that Hinduism does in Hindu worlds.

Survey Results

  • 11 There are a few limitations to our data. First, my three undergraduate student assistants did not s (...)
  • 12 Results of the reports of The Association of Religion Data Archives can be found online at http://w (...)

13Our survey found 3,347 public Hindu temples.11 Assuming the boundaries of Varanasi as given in the opening paragraph, in an area of approximately 17.69 mi2 (45.82 km2) we then have a temple density of approximately 190 per square mile, or 74 per square kilometer. According to the 2011 Census, the population of Varanasi is 1.435 million (2012, 3); if we can assume, for the sake of argument, that all those persons live within the same boundary that we have described, we find a ratio of one temple per every 438 persons. By comparison, for example, San Antonio, Texas, in the United States has 1,189 churches, mosques, etc., among a population of 1.714 million spread out over 1,256 mi2, for a density of one church per 1,441 residents and only 1.05 church(es) per square mile.12

  • 13 Haberman’s recent monograph, People Trees (2012), more thoroughly investigates the nature of trees (...)

14In surveying temples, we especially sought to note the relative size of each, and the results here are illuminating. We identified 175 temples as “large,” 644 as “medium,” and 2034 as “tiny” or “small”. Of the 523 remaining temples, we also located 157 trees that were identifiable as temples,13 in addition to and apart from the several hundred temples of all sizes that were built around trees.

15While the “tree” designation is fairly unambiguous, the size descriptions merit some further explanation. We designated as large those temples that were able to accommodate a significant number of persons. These included such well-known temples as Saṇkat Mocan, Durgā Kuṇḍ, and the New Kāśi Viśvanāth Mandir on the Banaras Hindu University campus, as well as less well-known centers of congregation like the Mānav Dharm Mandir in Sigra. These often served as the central or main temples in temple complexes. Medium temples were those capacious enough to allow one or more worshippers to easily enter, but not to serve as a site for any sizable gathering. Tiny temples may have allowed access for a single worshipper, though usually not while standing, and often did not have an actual interior beyond the space that contained the deity image[s]. See the following Figures (3, 4, and 5) for examples.

Figure 3

The New Kāśi Viśvanāth Temple on the campus of Banaras Hindu University: a large temple

Figure 4

A temple complex between Varanasi and Sarnath. From left to right, we would categorize these as a tree, a tiny temple, and a medium temple

Figure 5

A tiny temple, set in a wall in the Old City, with the inscription sahāra bhairavāya namaḥ (reverence to Bhairava the Supporter)

Figure 6

A tiny temple to “Śrī Narvā Bīr Bābā,” near Bhelupur.

  • 14 This assumes a temple in the exact center of each 600x400ft2 area, and a pedestrian walking directl (...)
  • 15 Please note an important distinction: tourists and pilgrims, who tend to stay near the Ganges in th (...)

16Even if we allow for some ambiguity between categories, it becomes readily apparent that smaller temples greatly outnumber large temples. The widespread distribution of smaller temples makes this difference even more significant. While there are almost exactly 10 major temples per square mile in Varanasi, there is an average of 115 small temples per square mile [see Figure 7]. Rendered differently, there is one temple, on average, per 242,420 ft2, or approximately one temple for every 600’ x 400’ square. Our data does demonstrate a clear clustering effect, with groups of temples interspersed among larger temple-less areas. If we imagine an even distribution, though, this means that we might expect a pedestrian to encounter a temple every two or so minutes when walking at average pace.14 Given further that our data and a graphically expressed map of it show that, outside of the Old City, temples tend to be located on main thoroughfares, this increases the likelihood that city residents will regularly and repeatedly encounter temples as they go about their daily business.15

17The sporadic mentions of smaller temples that occur throughout the literature often characterize them as local, folk, jāti, regional, village, wayside, or in other ways that our research empirically contradicts. While a tremendously wide number of deities feature in Varanasi’s temples (including a rare Brahmā temple and the only temple I know of dedicated to Hanuman associate and monkey king Valli), smaller temples do not specialize in lesser-known or “little” tradition deities. It is in this regard that concrete numbers can tell a different and more accurate picture. For example, Coccari opined that temples dedicated to Bīr Bābās “appear to be the most numerous” (1989:130) while our research only found about sixty of these. On the other hand, and as we might expect in the “city of Śiva,” 1,039 small temples featured Śiva. While Śivaliṅgas abound throughout the city, more than half of these temples (565) held images of Śiva together with Pārvatī or with his family of Pārvatī and Gaṇeśa (and occasionally Skandha/ Murugan). Furthermore, more than 2,000 temples featured mainstream, pan-Indian, “classical” Indian gods, with Śiva and Pārvatī, Rāma and Sītā, Hanuman, Ganesh, and Kṛṣṇa being most widely represented. Perhaps also somewhat curiously, tiny temples seemed to have no sectarian loyalties, with Śaiva and Vaiṣnava deities regularly appearing together in the same temple. This and other factors suggest that these temples and those who build them are less concerned with maintaining theological loyalties or other hierarchical ordering structures, concerns which sometimes motivate the Brahman composers of texts and their temples.

Some Conclusions We Might Reach

18This section addresses the social function of temples, such as I have been able to observe and record in Varanasi over the last ten years, though mainly in the summers of 2015 and, to a lesser extent, 2013. In this and other research, I have sought to give sustained attention to the possibilities for new insight that are afforded by reading against the grain, both of widespread popular assumptions about religion, and of individual traditions’ orthodox, textual, philosophical, and otherwise idealized views. Thus, for example, when attempting to understand the “work” of temples in the modern urban environment, I attend to intentional use by devotees who make planned visits for the purpose of worship, as well as to the impact that the physical form has as an element of material and visual culture for any and all who encounter it in whatever way they do. Moreover, because I have come to expect a range of “uses” beyond that which may have been envisioned by the person[s] who commissioned or built it, I will try to distinguish the dynamics at work in the use of temples from those that may have motivated their construction. Furthermore, given the overall aim of this project and my critique above of understandings of temples that are anchored in or bound to text-derived frameworks, I try here to present temples of any variety as they are experienced by Hindus who lack philosophical or textual expertise.

Spatial Distribution of Temples

19As Figures 2 (above) and 6 (below) make readily apparent, temples of all sizes in Varanasi are much more highly concentrated in the Old City around the Kāśi Viśvanāth Mandir. Many of these comprise parts of pilgrimage circuits, or rings of protectors around the Viśvanāth temple or others in the area. While there is no way of determining this conclusively at present, a greater number of temples may be found there because the Old City has a denser population and a longer history, giving more Hindus more time to construct temples and for these to accumulate; it may be because those who work around the Maṇikarṇikā and Dāsāśvamedh Ghats choose to live there because they have greater piety; it may be because the presence of several large temples encourages further religious expression; or, there may be other reasons altogether.

Figure 7

Map depicting distribution of tiny, medium, large, and tree temples in Varanasi.

20Tiny temples are not evenly distributed throughout Varanasi. Outside the Old City, north of Madanpura, a fairly even mix of large, medium, and small temples cover the city. South of the Old City, however, in the more newly settled and developed areas (see Singh 2008:46–50) around Benares Hindu University, there are only 12 large temples and 33 medium, but over 300 tiny temples. Throughout the rest of the city, tiny temples are clustered on public thoroughfares, and are often found near wells and crossroads and under trees.

The Tiny Temple in Hindu Life

  • 16 In Sanskrit this word is mandira while in Hindi it is often rendered as maṃdir. The Hindi conventio (...)
  • 17 Olivelle (2010) lists six variants for “temple” that occur in the dharma literature; Kramrisch (194 (...)

21Though the research on which this project is based was quantitative, we did have ample opportunities to speak with a number of persons who happened to be in the vicinity of temples. Uniformly, without exception, persons we spoke to in Varanasi used the same words to describe small, medium, and large structures containing deity images (mūrtis). In English, they always called these “temples;” in Hindi, they always used the word “mandir.”16 While Sanskrit texts, along with other regional languages, use a number of other words, this is the only Hindi word I have ever heard used in reference to a temple.17 While few of our respondents are likely to have been aware of this, the word mandir(a) comes from a verb (Sanskrit: √mad or √mand) meaning to tarry or stop; a mandir / temple is, thus, literally, a place where a god has come to rest. We may draw two important conclusions from this opening discussion. First, inasmuch as they are referred to by the same word, small temples are understood to be the same thing as any other temple. Second, what seems to constitute a temple, in the understanding of a Hindu in Varanasi, is that it is a dwelling place for a god; in other words, a temple is a place where a deity is present. With that salient element firmly in mind, we can begin to consider some of the reasons why Hindus choose to install so many small temples throughout Varanasi.

22The early classical sense of the reason for building a temple is expressed in the following passage:

  • 18 Kramrisch 1946, 139, citing Bṛhat Saṃhītā LV 2. This expression, squarely located in the Sanskrit a (...)

Let him who wishes to enter the worlds that are reached by sacrificial offerings and the performance of religious obligations (iṣṭapūrta) build a temple to the gods, by doing which he attains both the results of the sacrifice and the performance of religious obligations.18

While this may indeed be the motivation for Brahmins who emphasize the yajña, a range of other concerns may also prompt believers to erect temples. We might begin by considering those that seem most directly opposed to the, literally, other-worldly aspirations of elite practitioners.

  • 19 Aside from a fear of backlash or censure from one’s neighbors, or what we might term a superstitiou (...)
  • 20 On the ubiquity of public toilet use in India, see Chakrabarty (1992:541).

23On a surprisingly large number of occasions, Varanasi residents asked about our research and then told us, with a dismissive wave of the hand and roll of the eyes, that people build temples only to mark their territory and to prevent encroachment on it, or to solidify their encroachment onto other people’s territories. They pointed out that no one would even consider removing or destroying a temple.19 (This interpretation seemed to be particularly favored by older, well-educated, successful men.) The frequent occurrence of temples with trees, another facet of the landscape usually deemed inviolable, further supports this. Temples serve other functions as well. On one occasion my team and I were asking a woman about her conception of what constituted a temple and what did not. While we had seen a number of temples built into walls, we had also observed a few areas where colorful two-dimensional images of deities on ceramic tiles—i.e., pictures instead of statues—were embedded in walls. The woman laughed and said, “No—those are just to keep men from urinating there!”20

24While both of these explanations of the roles of public religiosity may strike readers as reductive at best and cynical at worst, an important understanding of the nature of Hinduism underlies them. To begin with, Donald Davis has argued that we should understand Hinduism as a legal tradition, with “law” meaning “a socially determined set of rules enforced by authoritative sanctions” (2007:243), a claim that scholars have begun to take seriously. Much of the “law” Davis refers to here is in the form of dharmaśāstra and dharmasūtra, of which we have a fairly good understanding; less clear, to me at least, is how this law comes to have relevance in the ongoing actions and transactions of Hindus who are mostly ignorant of those laws, and who also live in a secular state. How is it possible for a completely de-centralized and diffuse Hinduism to come to have real social impacts in that environment, when it seems to lack both the personnel and the infrastructure to even partially and weakly bring that law to bear? How, too, does the deep legal literary tradition integrate with devotional temple Hinduism and practical or “lived” Hinduism, and come to have relevance in the lebenswelt of Hindus in, for example, Varanasi? How could the Mānavadharmaśāstra’s pronouncements on where one should or shouldn’t urinate (4.46; Olivelle 2010:195) and how property should be marked and treated (8:246–48; Olivelle 2005:180) come to have any power?

25We may begin by noting that the smaller temple is not only of ancient provenance in India, but that its supposed present role as boundary marker is quite old as well, and legally sanctioned, as Manu prescribes that the king should establish boundary trees, and other immovable markers on boundaries, including temples (Olivelle 2005:180). Aside from their immovability, temples also bring an additional power factor into the equation to discourage violation of boundaries. The gods serve as guardians of dharma and the moral order. Images of them serve not only to remind potential wrong-doers of that moral order, but also make the power of the deity manifest in the relevant domain, establishing that the place (and its boundaries) come under the sway of dharma. The well-attested visual aspect of Hindu devotional worship also means that, when instantiated in mūrti form, the deity can see and observe behavior. Indeed, since a tiny temple is often little more than a god in a shelter, the presence of a watchful deity can comprise the majority of a temple’s function, providing a perch for the god—and thus Hinduism and the Hindu dharma—to supervise and regulate local activity.

26The recurring presence and view of the gods also points out another element of how tiny temples function through unintentional encounter rather than use per se. We discovered that worship occurred regularly in many temples. In these cases, neighborhood residents visited a local shrine near their homes, often each morning, to make the same sort of pūjā offerings that they would at any other temple. In some cases, we observed worshippers whose daily observance included a round of temples. One woman in the Lanka area visited as many as 15 each day, stopping to remove old offerings, sweep, and clean each before making her offerings and gestures of devoted supplication. These intentional visits do not exhaust the possibilities, however, as each day dozens more Hindus encounter temples. They may see a particular temple each day as they exit their neighborhoods, or a visitor to the area may see a temple for the first time. Then, too, some Hindus will make gestures toward a temple, offering namaskar or touching the ears, or even stopping to gaze briefly in a moment of darśan. Others will make no noticeable gesture, but may well take notice of the temple and its resident deity. Most who notice the temple will be sure not to spit or urinate in its vicinity, or to otherwise show a lack of respect. In this way, any temple can bring the force of Hinduism to bear on individual subjects, but because tiny temples are more ubiquitous, they meet Hindus closer to home, in places that they think of as their own.

  • 21 Marijuana is sold in several government shops (sarkāri bhang dukhāns) across Varanasi, and is viewe (...)

27Another unsuspected phenomenon we regularly observed was the use of local Śiva temples as gathering spots for small groups of men to meet and smoke chillums full of marijuana. The men always made some gesture of offering the first portion to the liṅga, and often took the time to inform us that ganja is, in fact, the prasād of Lord Śiva. Many also said that while under the influence of the drug they felt closer to Śiva. Sometimes a sādhu was present, but most often, these were 25–50 year-old married men. Whether or not a spot to imbibe with friends was, at some point, the intended purpose for any given temple, the local temple has come to serve that purpose in Varanasi. This kind of behavior would never be possible at the larger, more crowded temples overseen by Brahmins. Moreover, since these gatherings seemed often to happen around 8 am, it would hardly have been manageable to go across town to a major temple for a smoke. Clearly, too, the sanctity of the temple lent an aura of respectability to what might otherwise be seen as a disreputable activity.21

28Of course, the cost and convenience of constructing a smaller temple close to home makes it much more likely that those temples will actually get built. Locating small spaces for religious observance in their own neighborhoods is not merely a matter of convenience, nor one of cost, though those concerns surely motivate some decisions, at least in part. However, we should understand the prevalence of tiny temples, especially in Varanasi, over and against the several dozen more prominent temples for which the city is so famous. Not only are those temples larger and more ornate, they are also crowded and staffed by Brahmins who often demand hefty fees. Whatever power these temples and their liṅgas or other mūrtis might hold, one gets only a very short time in the presence of the image. One must also worship in a strictly proscribed fashion, or by proxy, and cannot truly act according to one’s own proclivities and desires. At a small temple near home, though, no priest serves as intermediary, one often has individual and unlimited access to the space, and as a result, worshipers are free to worship in whatever way they choose. Even if they do not choose to do anything at all different from what they might do at a larger, more prestigious temple, it is the matter of choice that is significant. Resonant with an emerging ethic informed by globalization, Westernization, and capitalism, independent worship makes free and individual autonomy possible. Hindus can thus act out of their own unfettered agency. Moreover, as Ortegren (forthcoming) points out, it is this ability to do as one wishes—to act according to one’s own desire (icchā)—that is the hallmark of modern urban middle class status. Many Hindus believe that lower class and rural Hindus are not as free to act out their own religious inclinations. By allowing those with fewer resources a venue for expression of their religion, small urban temples stand in defiance of that notion, and I would argue that at least part of their success lies in that capacity.

  • 22 See Michell (1975:100, 119).
  • 23 From an unpublished lecture delivered by Prof. Singh to a group of students on January 20, 2016.

29Small temples are often termed “subsidiary,”22 with the understanding that they are later accretions added on the periphery of the primary major temple. Many mahātmyas and sacred maps of Varanasi also make tiny temples out to be part of maṇḍalas, pilgrimage routes, or rings of protection around larger temples (Gutschow 2006; Singh 1993, 1995), thus mapping the city as a “sacredscape” or “faithscape” (Singh 1997). While it is difficult in most cases, in Varanasi at least, to determine much of a history of any smaller temple, there are good reasons to question accounts which make tiny temples into “part of” a larger edifice. Larger temples, associated as they are with political influence, Sanskritic and Brahmin authority, and great wealth, wield a great deal of power, and while there may be benefits to affiliation, many Hindus would rather retain their independence. Many Banarsis strongly resist regulation. As Prof Singh so aptly told another group of my students in 2016, “Each and every Indian follows one and only one rule, and that is: ‘I shall follow no rule.’”23 We often found layer upon layer of construction on top of temples, suggesting successive attempts to preserve small temples, and even convert them to serve as the centers of larger ones. Except when they were in the same complex, we found very little direct evidence, such as architectural or iconographic continuity, or inscriptions to connect most small temples to larger ones. We also found very little pattern to the location of most temples, suggesting that by locating temples according to their own preferences, Hindus in Varanasi neatly achieve the kind of resistance to organization imposed from above or outside that de Certeau found to characterize the movements of urban-dwellers in Europe (1984). In the end, while the temples (numbering more than 100) of the Pañcakroṣiyātra or the seven rings of Vināyaka protectors are, of course, important parts of visions of sacred geography and powerfully connected to central temples, many hundreds more have individual and independent value and play a role in Varanasi residents’ lives that is separate from that of the major pilgrimage temples.

30It is widely accepted that Hinduism is decentralized, free from large-scale denominational formations, and widely variable. This all affords Hindus a certain degree of freedom, within certain constraints, to believe, act, and practice as they choose. It is in the tiny temple, and the ability to construct one where and when one chooses, though, that Hinduism can be seen to reach its democratic pinnacle. Hinduism has been widely criticized for a sort of burdensome determinism born of the synergy of karma and caste. Tiny temples suggest to us that the opposite is true: in the choosing of who, where, how, and with whom to worship, the ability to construct one’s own temple means an incredible amount of freedom and autonomy. In Assi, quite near to a large Brahmanic temple, I found the only “under construction” tiny temple that we were able to locate in the course of this research. While I hoped for an exciting tale of a discovered image or a prophetic dream, a wish granted or a demonic force that needed to be repelled, the man that local residents identified as the builder of the simple three foot pedestal topped by a simple enclosure told me that he was building the temple there “so that [he] could do pūjā.” It was a few steps from his front door, and he told me he wanted to bring the deity, whom he only called Devī, into the area where he lived. Although this brief anecdote only concerns one temple of many, it does illustrate the possibility afforded to Hindus, and the real democratization of religious power that Hinduism affords.

31I will offer one last set of thoughts on the possible roles of the small, local temple in the making of Hindu worlds.

  • 24 Inadvertent pollution of a “hidden” temple can act as the cause for a spiritual or supernatural aff (...)

32First, while many temples are clearly pratiṣṭhita, or established, many more are understood to be svayambhu, or self-arisen. In fact, it is the deity itself there enshrined that is usually understood to have naturally occurred, and the temple then serves to honor the deity, and to make sure that those who come near to it know that a deity is present.24 As the epigraph to this paper shows, many believe that Varanasi is particularly shot through with the power and presence of Lord Śiva, and that is why so many liṅgas appear throughout the city. Tiny temples act as visible reminders of this notion.

33Second, in a more Buddhist milieu to the north, Mills has observed that Ladakhis use sacred monuments as ways of “nailing down” territory and claiming it as Buddhist (and thus as non-Muslim) (Mills 2003). We initially wondered if, in keeping with other uses of Hindu temples as boundary stones, we might find temples serving as markers between Hindu and Muslim neighborhoods. This turned out not to be the case. Muslim neighborhoods were usually devoid of tiny temples, as one might expect, but there were no clear instances of temples marking “Hindu territory.” In fact, in some very solidly Muslim areas, Hindu temples continued despite the obvious way that they interfered with traffic. Observant Muslims I spoke with in this area were very well-informed about the temples in their midst, and also appeared to be entirely accepting of their presence. So, while tiny temples seem not to be used to enforce Hindu ownership of a place, they can serve as an indication of a strong presence of Hindu divinity.

34Third, the notion of “seeing God everywhere” has a long history in Hinduism, perhaps going back to Bhagavadgītā 6.30. The ability to see God everywhere and in all things, though, was already recognized there as a difficult attainment, and one that one must work to achieve. It is far more of an ideal than a description of what the average Hindu does—if anything, it seems that Hindus appreciate being reminded of the possible presence of a divinity. The very concept of a temple means that it is a special place where divinity abides, and thus implies that there are other places in which divinity is either not present, or less strongly so. At the same time, the prevalence of temples throughout Varanasi means that city-dwellers get continual reminders about the presence of the divine, along with all that entails. The thousands of temples give added vitality to the idea that the gods dwell in Varanasi, the holiest of Hindu cities, as they do nowhere else. The temples make Hinduism repeatedly and visibly manifest. Shampa and Sanjoy Mazumdar have argued for the importance of religious institutions in making Hindu space (2009); for both residents and visitors to Varanasi, the presence of thousands of temples urges them to understand that they are in Hindu space, space that is permeated with Hindu gods who are available to them.

35In sum, the three ideas outlined above explain the mechanisms by which tiny temples operate in visual culture to create the perception of the powerful presence of Hindu society and Hindu divinity. In so doing, I argue that they make Hinduism into a real and discernible force in the lives of those who encounter them.

36In this paper I have perhaps opened more lines of inquiry than I have been able to close, and prompted more questions than I have answered. I also concede that much of what I offer here is hypothetical, and will need to be corroborated by further research. I should be sure to point out that my observations are also limited to the situation in Varanasi; however, I hope this paper may prompt us to ask whether the predominance of tiny temples in Varanasi prevails anywhere else in India, or indeed, beyond. While informal observations, anecdotal evidence, and hurried reassurances from colleagues have all suggested that, indeed, throughout India the small local shrine occurs nearly everywhere one cares to look, the method of this paper demonstrates that a rigorous quantitative survey can establish the facts of the matter with far greater certainty. I hope that I might also have convinced the reader that the tiny temple plays a much greater role in Hinduism than we have previously supposed, and that we ought to give local, wayside, tiny temples more attention than we have thus far.

Top of page


2011 Census of India. 2012. Retrieved July 1, 2017 (

Appadurai, Arjun, and Carol Breckenridge. 1981. “The South Indian Temple: Authority, Honour and Redistribution.” Contributions to Indian Sociology 10(2):187–211.

Bakhtin, Mihail. 1981. The Dialogic Imagination. Translated by C. Emerson and M. Holquist. Austin: University of Texas Press.

Bakhtin, Mihail. 1986. Speech Genres and Other Late Essays. Translated by V. W. McGee. Austin: University of Texas Press.

Bellamy, Carla. 2011. The Powerful Ephemeral. Berkeley: University of California Press.

Chakrabarty, Dipesh. 1992. “Of Garbage, Modernity and the Citizen’s Gaze.” In: Economic and Political Weekly 27(10/11):541–47

Coccari, Diane M. 1989. “Protection and Identity: Banaras’s Bir Babas as Neighbourhood Guardian Deities.” Pp. 130–46 in Culture and Power in Banaras: Community, Performance, and Environment 1800–1980, edited by S. Freitag. Berkeley: University of California Press.

Davis, Donald. 2007. “Hinduism as a Legal Tradition.” Journal of the American Academy of Religion 75(2):241–67.

de Certeau, Michel. 1984. The Practice of Everyday Life. Translated by S. Rendall. Berkeley: University of California Press.

Desai, Madhuri. 2017. Banaras Reconstructed. Seattle and London: University of Washington Press.

Dempsey, Corinne. 2006. The Goddess Lives in Upstate New York. Oxford: Oxford University Press.

Eck, Diana. 1981. Darśan. Chambersburg, PA: Anima Books.

Eck, Diana. 1982. Banāras: City of Light. New York: Knopf.

Fuller, C.J. 2003. The Renewal of the Priesthood. Princeton: Princeton University Press.

Gaenszle, Martin, and Jörg Gengagel, eds. 2008. Visualizing Space in Banaras. New Delhi: Oxford University Press.

Gengnagel, Jörg. 2008. “Maps and Processions in Banaras: The Debate Concerning the Pañcakrośīyātrā.” Pp. 145–64 in Visualizing Space in Banaras, edited by M. Gaenszle and J. Gengagel. New Delhi: Oxford University Press.

Gengnagel, Jörg. 2011. Visualized Texts. Wiesbaden: Harassowitz Verlag.

Gutschow, Niels. 2006. Benares: The Sacred Landscape of Varanasi. Berlin: Edition Axel Menges.

Haberman, David. 2013. People Trees. Oxford: Oxford University Press.

Inden, Ronald. 1990. Imagining India. Cambridge, MA: Basil Blackwell.

King, Richard. 1999. Orientalism and Religion. New York: Routledge.

Kramrisch, Stella. [1946] 1976. The Hindu Temple. 2 vols. Delhi: Motilal Banarsidass.

Lorenzetti, Tiziana. 2015. Understanding the Hindu Temple. Berlin: EB-Verlag.

Mazumdar, Shampa, and Sanjoy Mazumdar. 2009. “Religion, Immigration, and Home-making in Diaspora: Hindu Space in Southern California.” Journal of Environmental Psychology 29(2):256–66.

Michell, George. 1977. The Hindu Temple. Chicago: University of Chicago Press.

Mills, Martin. 2003. Identity, Ritual and State in Tibetan Buddhism. London: Routledge.

Mohammad, Afsar. 2014. “The Sri Venkateswara Temple in Tirupati.” Pp. 232–44 in Contemporary Hinduism, edited by P. Penumala. London: Taylor and Francis.

Olivelle, Patrick. 2005. Manu’s Code of Law. Oxford: Oxford University Press.

Olivelle, Patrick. 2010. “The Temple in Sanskrit Legal Literature.” Pp. 191–204 in Archaeology and Text: The Temple in South Asia, edited by H.P. Ray. New Delhi: Oxford University Press.

Olivelle, Patrick. 2016. King, Governance and Law in Ancient India. Oxford: Oxford University Press.

Ortegren, Jennifer. Forthcoming. Dharma of Desire: Examining Religion and Class in Urban North India.

Schopen, Gregory. 1997. Bones, Stones, and Buddhist Monks. Hawaii: University of Hawaii Press.

Sears, Tamara. 2014. Worldly Gurus and Spiritual Kings. New Haven, CT: Yale University Press.

Singh, Rana P.B. 1993. “Cosmic Layout of Hindus’ Sacred City, Varanasi.” Architecture & Comportement / Architecture & Behaviour 9(2):239–49.

Singh, Rana P.B. 1995.Varanasi: Gaṇeśa Images and Shrines in the Cityscape in Dubey.” Pp. 209–20 in Pilgrimage Studies: Sacred Places, Sacred Traditions, edited by D.P. Dubey. Allahabad: Society of Pilgrimage Studies.

Singh, Rana P.B. 1997. “Sacred Space and Pilgrimage in Hindu Society: The Case of Varanasi.” Geoscience & Man 34:191–207.

Singh, Rana P.B. 2008. Banaras, the Heritage City of India. New Delhi: Indica.

Singh, Rana P.B. 2014. “Banāras: Encountering the Experiences and Expositions of the Spirit of Place.” Pp. 77–93 in Banāras Revisited, edited by I. Keul. Wiesbaden: Harrassowitz Verlag.

Singh, Ravi S., and Rana Singh. 2008. “Goddesses in Kāśī (Vārāṇasī): Spatial Patterns and Symbolic Orders.” Pp. 41–69 in Visualizing Space in Banaras, edited by M. Gaenszle and J. Gengagel. New Delhi: Oxford University Press.

Times of India. 2014. “At Kashi Vishwanath Temple, Modi Offers 1,001 Lotuses to ‘Baba,’” May 18. Retrieved July 10, 2017 (

Waghorne, Joanne Puzo. 2004. Diaspora of the Gods. Oxford: Oxford University Press.

Top of page


1 Eck (1982:109), citing the Kāśikhaṇḍa section of Skandapurāṇa, 58.63.

2 Research for this paper was generously funded by the Metzger Fund and the Faculty Development Committee of Centre College, as well as by a Collaborative International Research Grant from the American Academy of Religion. I received invaluable research assistance in Varanasi from several Centre students: Nathaniel Deaton in summer 2013, and Gray Whitsett and Robert Widener in summer 2015. Amy Jackson-Smith of Oberlin College and Binit Kumar Mishra of Varanasi also assisted my research in summer 2015. I wish to thank Dr. Rana P.B. Singh of Benares Hindu University, Drs. Borayin Larios and Raphaël Voix, and the anonymous reviewers from SAMAJ whose comments tremendously improved this paper.

3 We in fact located only a half dozen or so temples north of the Indian Railways rail lines to the north; another eleven to the west of the DLW Road in the Mandua Dih neighborhood; and, in the south, we took the demarcation as the Highway Two Varanasi Bypass Road, and located only five temples south of that boundary.

4 Lorenzetti 2015, 15; while n.15 notes that “[i]t should be remembered that Hinduism (a term coined by Western scholars) is an aggregate of cults, philosophical systems, traditions, and rituals—often of very ancient origins—that underline and regulate the social structure, culture and private lives of Hindus,” this only furthers my critique. Diverse forms of decentralized Hinduism have diverse forms, and sizes, of religious space.

5 Michell 1977, 61.

6 These are two genres of Sanskrit text. Māhātmya means greatness; these are panegyrics of persons, places, deities, etc. Vastuśāstra or “treatise on substances” are often called the Indic equivalent of Feng Shui; these dictate the siting and building of temples (and other buildings) to maximize power and auspiciousness.

7 The day after he was elected Prime Minister of India in 2014, Narendra Modi visited the Kāśi Viśvanāth Temple in Varanasi, where he had prayed earlier for divine help in becoming PM (Times of India 2014).

8 Gutschow 2006 and Gengnagel 2008 address the Pañcakroṣi yatra and its associated mandir.

9 Maṇḍalas are circular geometric designs associated with divinity and ritual. The term is also used in classical political literature to describe “circles of kings” surrounding the overlord (see Olivelle 2016:46–49).

10 Gaenszle and Gengnagel 2006. They point to two further studies in their 2006 volume, Gengnagel (2006) and Singh and Singh (2006) which address higher- and lower-caste spatializing tendencies respectively.

11 There are a few limitations to our data. First, my three undergraduate student assistants did not speak Hindi or read Devanāgarī, the script often used to inscribe names of temples, or of deities. They could not always ask locals for information or record the full information about each site. Second, Varanasi, especially the old city and the areas nearest the river, is well-known to be a veritable maze of narrow alleyways without street signs or addresses. Both GPS and conventional maps sometimes fail to depict all the local streets accurately or thoroughly. This results in two possible shortcomings: we may well have overlooked some sites; and, in some cases, our technology for gathering data was hampered by the canyon-like terrain of eight foot wide alleys with three- or four-story buildings on either side. When we compiled and rendered the data, we discovered that several temples had been digitally graphed as situated in the very middle of the Ganges. Our GIS technology even managed to locate one temple off the coast of West Africa! All this may well mean that data for some temples contains some degree of error, mostly concerning precise location. Finally, because I am, I believe, the first scholar to carry out this kind of survey, and because this is something of a new enterprise for me, we conducted our first week’s worth of surveys with an experimental template, and collected information on 481 temples. We then assessed the results, re-configured our survey questionnaire, and canvassed the rest of the city. Ideally, we would have returned to re-survey the initial 481 temples and collected the same data on them as we did with the later entries, but time and logistical realities prevented this. I hope to be able to return to Varanasi to refine the results of this initial survey soon, and to remedy many of these shortcomings.

12 Results of the reports of The Association of Religion Data Archives can be found online at, including for cities and counties in the USA. The ARDA counts “congregations,” and so it is not surveying the same kinds of public temples that are under consideration here; however, there are virtually no public shrines in San Antonio. The survey results for San Antonio, TX, actually cover the entirety of Bexar County.

13 Haberman’s recent monograph, People Trees (2012), more thoroughly investigates the nature of trees as objects of worship than can this paper. Though a multitude of interpretations emerge, and in this case, from a popular rather than a textual domain, for now it will suffice to say that trees are understood both as deities and as abodes of deities. As I remark below, in Varanasi several “temples” comprise only a deity image with no other structure, and so I would argue that even if the tree is understood as a deity rather than a temple, there is sufficient ambiguity to say that it merits consideration as both.

14 This assumes a temple in the exact center of each 600x400ft2 area, and a pedestrian walking directly from each temple to the next at a pace of 3 mph. Anecdotally, I will attest that while traveling by motor vehicle, I traversed large swathes of the city where I saw a temple at least once every ten seconds. It was this experience which prompted my initial interest in this research.

15 Please note an important distinction: tourists and pilgrims, who tend to stay near the Ganges in the Old City, will often intentionally visit the major temples there, and often stay in that area.

16 In Sanskrit this word is mandira while in Hindi it is often rendered as maṃdir. The Hindi convention does not reflect any change in pronunciation or meaning.

17 Olivelle (2010) lists six variants for “temple” that occur in the dharma literature; Kramrisch (1946) gives a list that overlaps with Olivelle’s, but includes perhaps a dozen more. These, however, are all given as other, less-popular alternatives to vimāna or prasāda, which Kramrisch gives as the two principal terms (131–38).

18 Kramrisch 1946, 139, citing Bṛhat Saṃhītā LV 2. This expression, squarely located in the Sanskrit and Brāḥmanical tradition, seems to very closely echo the svargakāmo yajeta, “he who desires heaven should sacrifice” of the Mimāṃsa school.

19 Aside from a fear of backlash or censure from one’s neighbors, or what we might term a superstitious fear of baleful consequences, normative legal texts also admonish followers not to move (Reddy 2016) or damage temples (Olivelle 2010).

20 On the ubiquity of public toilet use in India, see Chakrabarty (1992:541).

21 Marijuana is sold in several government shops (sarkāri bhang dukhāns) across Varanasi, and is viewed with varying degrees of tolerance. Many informants insisted that marijuana was preferable to alcohol because alcohol led to violence. One pānwalla told me he smoked because it was a good way to “enjoy my time with my family.” However, apart from holy men whose use of marijuana is sanctioned as a legitimate religious practice, users usually make efforts to smoke in remote places, or to do so clandestinely, so as to avoid being observed.

22 See Michell (1975:100, 119).

23 From an unpublished lecture delivered by Prof. Singh to a group of students on January 20, 2016.

24 Inadvertent pollution of a “hidden” temple can act as the cause for a spiritual or supernatural affliction (ḥāẓirī), on which see Bellamy (2011:138).

Top of page


Electronic reference

Christian Haskett, On Varanasi’s Tiny TemplesSouth Asia Multidisciplinary Academic Journal [Online], 18 | 2018, Online since 25 April 2018, connection on 24 February 2024. URL:; DOI:

Top of page

About the author

Christian Haskett

Assistant Professor of Religion, Centre College, Danville, KY, USA

Top of page



The text only may be used under licence CC BY-NC-ND 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search