Empirical and analytical subaltern space? Ashrams, brothels and trafficking in colonial Delhi

ABSTRACT In this paper, I engage with Gayatri Spivak’s writings on the figure of the subaltern, focusing on a recurrent tension in her writings, and in readings of them. The tension is between two seemingly contradictory definitions of the subaltern. One, more empirical definition, has featured in Spivak’s writings for over 25 years and identifies the subaltern as the non-elite, the immobile or the figure beyond the reach of the state. Against this more empirical definition comes the famous analytical definition of the subaltern as he or she who ‘cannot speak’, being defined by their inaccessibility in the archive, as broadly conceived. This paper will argue that these two interconnected definitions have their respective forms of space, which demand different methodologies. I will suggest that an over-emphasis on the analytical definition has led to an over-cautious approach to subaltern spaces, neglecting the compulsion to attempt to find and say something about subaltern spaces, as Spivak insists. The paper demonstrates this approach through the examination of a report into the abuse of women in some of Delhi’s ashrams in the 1930s, so as to suggest how we can use studies of empirically archived subaltern space to think about the analytically subaltern spaces that must always be beyond exploration.

different sub-disciplines, areas of study and period of interest (for recent overviews see Jazeel 2013, Lester 2013, Sidaway et al. 2014. In the wake of Said's (1978) Orientalism: western conceptions of the orient, geographers have largely attempted to bring down the postcolonial edifice of geography from within: recounting the complicity of its practices and representations with historical forms of domination; charting ongoing epistemic violence in contemporary pedagogies of the global education economy; but venturing more rarely into the hidden spaces and meanings of the 'colonial underclass'. The reasons for this perhaps lie in the series of cautions with which Crush prefaced his suggestions. He acknowledged that a proposed alternative to the internal obsession with orientalist representations in postcolonial theory was the turn to the 'lost historical voices of the marginalized' pioneered by the subaltern studies group in India (Guha and Spivak 1988). He immediately stressed, however, that this can simply result in the lives of those in the margins becoming a fixed object of study through which they are spoken for. Spivak's (1985Spivak's ( , 1988Spivak's ( [2000) interventions into the subaltern studies debate helped secure its place at the heart of postcolonial studies. It has been taken up in various global regions and its insights regarding education, representation, power and history have been applied at scales ranging from the individual body to the notion of world history (Chakrabarty 2000). Geographers have responded to many of the challenges laid down by Spivak's penetrating commentaries on the subaltern (Gidwani 2009, McEwan 2009), including works on the spaces of differentiation, diversality and paradox in subaltern theory (Clayton 2011), 'subaltern geopolitics' (Sharp 2011), subaltern spaces of crisis (Chari 2012), the representational assumptions in the models of the world that geographers carry with them (Jazeel 2014), and the international networks and solidarities of subaltern groups (Featherstone 2008(Featherstone , 2012. What is rarer in geography is work that attempts to recover voices of otherness and difference from within the archive (although see Duncan 1999, Driver et al. 2009, Moore 2010, Bressey 2011. The archive is increasingly understood in the discipline as a space of documentary investigation but also as a space of embodied encounter and a discursive phenomenon that intertextually dreams of including us all (Lorimer 2009, Mills 2013. But geographers have not paid as much attention to postcolonial marginal, or absent, voices. This is, in part, a logical response to Spivak's suggestion that the subaltern cannot speak. But she has since both recanted the overwhelming negativity of this response (' … in the first version of this text ["Can the Subaltern Speak"], I wrote, in the accents of passionate lament: the subaltern cannot speak! It was an inadvisable remark'. Spivak 2010a, p. 63) and chastised researchers for not trying to listen to the voices of the oppressed in the archive. In this paper I would like to engage these complex epistemological issues through investigating some evidence which offers up the possibility of saying something about the experience of subalternity in some of interwar colonial Delhi's ashrams. To frame these studies I will provide a brief rereading of Spivak's work to sketch out more clearly the dialectics between compulsion and caution, theory and practice, and the empirical and analytical studies of subalternity, that she insists we enter every time we engage the postcolonial archive which is, for her and us, our every text.
The founder of the Subaltern Studies group, Guha (1982, p. 8), defined the subaltern empirically, as: 'the demographic difference between the total Indian population and all those whom we have described as the "elite"'. The attempt to locate and describe the 'autonomous sphere' (Guha 1983, p. 40) of the subaltern preoccupied the early writings of the group, who were criticized for elevating the subaltern to the status of a romantic and timeless sovereign subject of resistance (O'Hanlon 1988) and for their positivistic essentializing of the subaltern (Spivak 1985). Despite this, in her later commentaries in which Spivak argues against the interpretation that she did not allow the subaltern to speak, she shows how her early work depended on Guha's definition of the subaltern as a space that is cut off from lines of mobility in a colonized country: 'You have the foreign elite and the indigenous elite. Below that you will have the vectors of upward, downward, sideward, backward mobility. But then there is a space which is for all practical purposes outside those lines' (Spivak et al. 1996, pp. 288-289). Again: 'The reasonable and rarefied definition of the word subaltern that interests me is: to be removed from all lines of social mobility' (Spivak 2005, p. 475). And, recently, in asking what the subaltern means: 'It means the group that only takes orders [ … ] It is an absence of any access to the possibility of the abstract structures of the state' (Spivak 2014b, pp. 9-10).
Space and mobility are simultaneously abstract and material concepts for Spivak. It is, of course, possible to be stationary in space yet socially mobile, or to be constantly labouring through space while remaining socially subaltern. In this paper, I focus on the relationship of subaltern subjects to spaces that we might also term subaltern. This is due not to their locational marginality (the ashrams were in the absolute heart of Old Delhi) nor to their exceptional materiality (there is no evidence that the ashrams were outwardly distinguishable from other buildings of similar size). The character of their occupants (widows, orphans and, in some cases, victims of rape, forced prostitution and trafficking) mark them as subaltern in Guha and Spivak's definition of being cut off from the elite, the state and social mobility. It is into that empirical, and experienced, space of immobility and abandonment that we must look to find and hear the subaltern. Against various misreadings, Spivak did not give up on 'retrieving' the subaltern (Morris 2010, p. 2). On the contrary, she has criticized 'the subalternists' for their move towards postcolonial theory and away from any effort 'to touch the subaltern or, with the energy with which historiographic practice is questioned, to question the political strategy that appropriates the disenfranchised' (Spivak 2005, p. 477).
If the subaltern exists as a more empirical category, a space of immobility, even a body we must try to find, what then of the analytical silence of the subaltern? In an early interview, Spivak pinpointed the exact relationships between the empirical existence and the analytical impossibility of the subaltern: 'The subaltern is all that is not elite, but the trouble with those kinds of names is that if you have any kind of political interest you name it in the hope that the name will disappear' (Spivak and Veeser 1990, p. 158, although in a counter-reading of Gramsci, via Green 2002, Nilson andRoy 2015 insist that subalternity only disappears once its conditions of subordination are altered, not once its members are politically mobilized and, indeed, that subalterns are always already, if variably, politicized).
Spivak has clarified that speaking is a transaction between a speaker and a listener (Spivak et al. 1996, pp. 291-293). Her point is that the subaltern, even when taking pains to the death to speak, cannot be heard due to the intervening interpretations, silencing, censoring and appropriations of the archive and of representation more generally. As such the subaltern is a predicament, not an identity; an obstruction from accessing power and voice (Morris 2010, p. 8). This is the condition of their existence as a theoretical category and is the dynamic behind the 'disappearing subaltern'; if you can hear a 'subaltern' then they are, by definition, no longer subaltern. Analytically, this definition of the subaltern turns back to face the researcher, with subalternity as 'a space where the intellectual instrumentalises himself or herself in order to go into learning from below' (Spivak 2014b, p. 10).
How, then, are we to reconcile these two seemingly contradictory definitions of a more empirical subaltern defined by its immobility and a more theoretical subaltern defined by its inaccessibility? I believe we must attempt this reconciliation because, to return to Crush's terms, the fear of speaking 'for' the subaltern has deterred many academics (myself included) from attempting to speak 'of' them. I will argue that caution about speaking for the subaltern has lazily overwhelmed the compulsion to speak 'of' subalternity, and that this can be traced to a failure to think critically about the interlinked but distinct definitions of a more empirical and a more analytical subaltern.
These definitions have been variably read: as ones of theory and empirics, or as moral-or action-based definitions (Masselos 2001(Masselos [1992; as a tension between the approaches of postmodern literary theory and social history (Sivaramakrishnan 2001(Sivaramakrishnan [1995, p. 223); and as a route between the 'positivist euphoria' of subaltern voice discovery and the acceptance of silence as normal (Morris 2010, p. 8). For me, the terms function as a productive and compulsive dialectic, creating each other without necessary or possible resolution (for a complementary commentary on the role of unresolved dialectics in Carl Schmitt's work, see Rowan 2011). Spivak insists that finding the (more empirical) subaltern is not hard; what is hard is entering into a responsible analytical relationship with the subaltern; 'that's the hard part' (Spivak et al. 1996, p. 293).
The two co-dependent definitions of subalternity compel us to work ever harder. We must seek out structured spaces of immobility and dis-empowered populations. In the contemporary world that process can open up potentially radical spaces of engagement and learning between academics and self-representing subalterns (Roy 2015). In writing historical geographies from sparse textual traces, however, the opportunities for engagement with subaltern subjects shrink in the face of the vast representational and epistemic violence of the archive (on the methodologies of this struggle see Pandey 1995). As such, as soon as we find a subaltern individual or group then we must immediately ask what that group has become, what further type of oppression we could look to, and how we must start again: The possibility of subalternity for me acts as a reminder. If it is true that when you seem to have solved a problem, that victory, that solution, is a warning, then I begin to lookit's not a substantive formulabut I always look at that moment for what would really upset the apple cart. And that's quite often the moment when one begins to track the newly created subaltern, out of reach. (Spivak et al. 1996, p. 293) We might, therefore, think of two dialectically interlinked types of subaltern space. One, based on the more empirical definition of immobile subjects, would be an experiential space populated by non-elite groups, which can be historically located through reading along the archival grain (setting the context and working with reports to detail the worlds they investigated: see Stoler 2009) as well as against it (questioning the nature of the document, the conflation of agency and subjectivity, and the silencings of textuality, as well as resistant and resilient languages and acts: see Banerjee 1998). A second type of subaltern space, based on the analytical definition of silent, out-of-reach subjects, would be irretrievable, a series of locations that are hinted at but remain un-knowable. Both will be evidenced through the cases below, in which the urge to explore the empirical space of Delhi's ashrams is balanced against the nihilistic realization that the true subaltern spaces of the historical city remain wholly beyond reach.

Caution, compulsion and the sexual subaltern
The balance, then, is between caution and compulsion. For many, Spivak's subaltern writings have functioned as postcolonial chloroform, silencing and gagging any attempt to speak not only 'for' but even 'of' the subaltern. As a white, male, researcher from the economically exploitative side of the international division of labor, I have carried this theory heavily, haunted by Spivak's oft-quoted statement regarding sati (widow immolation) reformers as 'white men saving brown women from brown men' (Spivak 2010a, p. 48; for a discussion of Spivak's intention behind using this 'grammatical form', see Morris 2010, p. 3 and Balibar and Spivak, this issue). But Spivak has been explicit about the necessity of engagements between researchers and thinkers across postcolonial geographical boundaries. She has defended her ongoing engagement with white, male, Western philosophers (Spivak 1993, p. x) who attempted to study otherness and heterogeneity Adamson 1990 [1986], p. 56). While many find deconstruction's complexity and lack of clear definitions intimidating, she insisted that 'the greatest gift of deconstruction [is] to question the authority of the investigating subject without paralysing him [sic], persistently transforming conditions of impossibility into possibility' (Spivak 1987b, p. 201). Against those who suggest that only women can know women, or only 'natives' can know 'natives' (Spivak 1987b, pp. 253-254), she insists that 'knowledge is made possible and is sustained by irreducible difference, not identity' (Spivak 1993, p. 8). She is at her most compelling when discussing a teaching conversation with a white, male, politically correct student who, conflicted about attempting to research the subaltern, suggested that 'I am only a bourgeois white male, I can't speak' Gunew 1990 [1986], p. 62). She recalls her response: 'Why not develop a certain degree of rage against the history that has written such an abject script for you that you are silenced?' Then you begin to investigate what it is that silences you, rather than take this very deterministic position -since my skin colour is this, since my sex is this, I cannot speak.
[ … ] From this position, then, I say you will of course not speak in the same way about the Third World material, but if you make it your task not only to learn what is going on there through language, through specific programs of study, but also at the same time through a historical critique of your position as the investigating person, then you will see that you have earned the right to criticize, and you will be heard … In one way you have to take a risk to criticize, of criticizing something which is Other-something which you used to dominate. I say that you have to take a certain risk: to say 'I won't criticize' is salving your conscience, and allowing you not to do any homework. On the other hand, if you criticize having earned the right to do so, then you are indeed taking a risk and you will probably be made welcome, and can hope to be judged with respect. Gunew 1990 [1986], pp. 62-63) The very last thing that subaltern theory is about, therefore, is silence. It must not be used as an excuse to not investigate the empirical conditions and experience of seemingly powerless subjects because of one's privilege. But alongside this compulsion to act comes representational caution. The sexual heightens the intensity of this binary. Sexual violence and prejudice compel action, but the intimate and culturally sensitive nature of the sexual compels caution, in historical as much as contemporary studies. Arondekar (2009, p. 1) has insisted that historical studies of sexuality should question why the archive is returned to as a site of truth regarding sex? What compels this return? Subaltern studies are acknowledged, yet 'even as the impossibility of recovery is articulated, the desire to add and fill in the gaps with voices of other unvoiced subalterns remains' (Arondekar 2009, p. 6). While acknowledging these cautions, Gupta (2011) has expertly revisited colonial archives in north India to read them along and against their grain, bringing their materials into dialogue with alternative archives and counter-readings. No unmediated access to sexual experience is possible here, but Gupta acknowledges that state and non-state efforts to occasionally understand the mundane and the everyday mean that colonial archives can and do retain valuable traces of those who have been called the 'sexual subaltern'.
The term sexual subaltern builds upon efforts to think about a gendered subaltern, against the un-gendered and hence male subaltern of the early Subaltern Studies (Spivak 1987a, Visweswaran 1996, Bannerji 2000. This sexualized category refers to 'the disparate range of sexual minorities within postcolonial India, without suggesting that it is either a homogenized or stable category' (Kapur 2000, p. 16). Nor is it necessarily a position of total victimhood. Though Kapur initially aligns sexual subalterns with positions traditionally associated with victimhood, such as gays, lesbians and sex workers, she reminds us that the position of sexual(ized) subalternity can also be ones of pleasure: 'Emphasizing the pleasure of this subject does not deny the violence and exploitation that surround her life, but serves to challenge the representation of her exclusively as victim, or according her partial agency' (Kapur 2000, p. 22).
In what follows, I aim to describe sites of sexualized subalternity in colonial Delhi, but to resist the narrative of total victimhood. The aim is to acknowledge partial agency, through both empirical and analytical subalternist turns. In over a decade of researching prostitution in colonial India, I have come across only a handful of documents that offer up the prospect of insights into the lived experiences of sexual subalternity (see Legg 2016). The approach here is twofold. Empirically, I will use one of these exceptionally rare and rich archival files to detail the constant striving against capitivity and violence in some of Delhi's ashrams. Analytically, I want to show that the very act of entering the archive disqualifies these voices and lives as being subaltern, but that the files contain fragments of other voices unheard and spaces un-documented. These fragments stand here as evidence of women without recorded agency, and also stand as compulsions for us to begin, once again, 'to track the newly created subaltern, out of reach' (Spivak et al. 1996, p. 293).
The environment of that tracking is the colonial Indian capital of Delhi. In 1911 the capital of India was relocated from Calcutta to the south of the walled city of 'Old' Delhi (Legg 2007). 'Prostitutes' had been forced into this space following their ejection from military cantonments across British India in the 1890s (Legg 2009). The government had failed to devise a policy for these women, but had tacitly encouraged their segregation in tolerated brothel zones in most large cities (Legg 2012). These 'red light districts' became objects of scandal in the 1920s in the context of increasingly heated debates about India's treatment of its women (Sinha 2006). The Delhi government's response in the 1930s was to push for new legislation (a Suppression of Immoral Traffic Act (SITA)) that would provide powers to prevent procuration and trafficking, while using municipal regulations to force brothels out of the walled city (Legg 2014). This led to increased attention on the capacities of the city to take care of vulnerable women. The report that is examined below was compiled in 1935 in response to suggestions that traditional ashrams in the city had been exploiting child widows, orphans and abducted women; a group we can confidently depict as empirically subaltern, in terms of being non-elite, beyond the state and with reduced mobility. It is very rare to find cases of alleged rape, trafficking and forced prostitution described in such fine detail. However, in this report, the women that are documented are largely those who exploited the system or who escaped from it. Their mobility and voice make them, by strict analytical definition, not subaltern, though we have rich details of the conditions of subalternity they co-habit. Yet within the report we find fragments of subaltern lives beyond the spaces we can here document, which will be returned to in the concluding reflections on the types of historical exploration we engage in when trying to locate the subaltern.
Empirically subaltern space? The ashrams' 'hue and cry' Before the passing of the SITA, Delhi had its own institutions for assisting certain women and girls in need. These often took the form of ashrams; places of 'religious retreat, sanctuary, or hermitage' 1 which often fulfilled functions of what was increasingly becoming known in India as social service. They blurred public and private space, serving society by taking socially problematic figures inside, but were usually beyond the state in terms of funding, intervention and surveillance. Very few of the ashrams listed were rescue homes in the emerging sense of institutions designed specifically for women rescued, or fleeing, from brothels. The rescue was, more usually, that of orphans and, especially, of child bride widows. The latter were products of arranged marriages, who could find themselves widowed and without prospect of re-marriage or employment at a young age (for a contextual reading of the Sarda Act of 1929 which set marriage ages at 14 for females and 18 for males, see Sinha 2006 andMukherjee 2006. For Empire-wide context and comparisons, see Levine 2007 andPhillips 2006). As wholly non-elite spaces, where the inhabitants faced severe challenges to either social or spatial mobility or accessing the state, the ashrams are empirical subaltern spaces; or, as Crush (1994, p. 337) described them, 'hidden spaces occupied, and invested with their own meaning, by the colonial underclasses'.
Serious concerns were raised about these ashrams by a question proposed in the Legislative Assembly on 16 September 1935, regarding an article in the Amrita Bazar Patrika (a national newspaper) headed 'Scandal of Rescue Homes, Girls Trapped and Sold.' 2 The story was traced to the local Hindustan Times who had run the article in July, alleging that over one hundred 'so-called widow and rescue homes' existed in Delhi where girls from the city and beyond were bought and sold to the highest bidder, under the garb of marriage. While not made public, the Chief Commissioner did procure a report from the City Inspector of Delhi Police, submitted on 4 October 1935, which provides an unprecedented insight into the spaces and experiences of 'rescue' in 1930s Delhi. The 24 handwritten pages detailed ashrams, orphanages and rescue homes in the different police jurisdictions of the city. This was not a systematic survey of Delhi's over one hundred ashrams, but just those nine which had come to the attention of the police.
The original report conflated, in almost every case, ashrams and brothels. The 'outward' aim of such institutions was said to be to rescue girls and women, to provide guardianship and to teach them industries, although the Inspector immediately insisted that they failed even in this, teaching women only how to go begging for alms from door to door. The Inspector's report classified these homes alongside regular brothels, suggesting they were where 'unfortunate women were brought for immoral purposes'. Such brothels were exploited by 'regular agencies of "Barda Faroshes" [an Urdu term translated as "slave traders"] in different cities where they sell the girls who happen to come into their clutches'. Managers of the homes were said to engage persons: of low class + of low morals as their agents for the supply of these girls + women. The agents look in different parts of the city, especially Ry.
[railway] stations, lorries + tonga 3 stands, cinema, hotels, sarais, 4 dharmsalas 5 and jumna ghats 6 and whenever they should happen to come across some deserted woman or girl they by false promises take her to the ring leader of the ashram where she is then confined. 7 The geography of procurement is clear: busy places of transit or spaces of leisure and relaxation in which normal assumptions about being in or out of place become blurred. Once inside the ashram, the report suggested, the women had no access to the outside world and were thus helpless to make a complaint. If they attempted to make a 'hue and cry' 8 they would be 'subjected to harsh treatment' while the ashram staff would play on the harmonium or would sing religious songs 'in order to silence + drown out the wailing of these unfortunate women' so that their cries would not be heard by outsiders. The report enriches the nature of this subaltern space. It emerges as inherently relational, dependent on the movement of women and girls through the city and into the ashram, refusing any blanket association of subalternity with the 'local' or the parochial (Featherstone 2008). Secondly, it explicitly outlines the violent mechanism of silencing through which the ashrams produced their subjects, achieved through punishing attempts to move and shout, and through the production of sound to drown out the 'wailing' of women inside.
But the primary sources cited by the City Inspector (registered cases under the Indian Penal Code, IPC) also give us a much broader sense of female intransigence, resistance, and even complicity with ashram owners and traffickers, while the role of men in this trade is also complicated. Women are not just victims of tradition or circumstance, in this document. They are also vibrant actors, as also suggested by Mani (1998) in her study of victims of sati (widow immolation); perhaps the test case of the presumably silent subaltern. Alongside the suicide of who we later find out was her great aunt, Bhubaneswari Bhaduri (Spivak 2010b, p. 228), Spivak spoke about the attempts to ban sati, and the archival over-writing and over-determining of these widows' actions in nativist and colonial discourses as either, respectively, ancient practice or murder. These women have no free will, either to decide to die or to resist their death: 'The dubious place of the free will of the constituted sexed subject as female was successfully effaced' (Spivak 2010b, p. 275). The widows that Mani studied did not necessarily speak, but they acted, while any impression of their docility was shown to be produced by an (often unwitting) compact between Indian and British patriarchies. The women who are written about in these cases had proven by their vociferous voicing, fighting and fleeing that while they had found themselves in a subaltern space, they were not irredeemably immobile or silent. On the contrary, their naming and emergence in the archive marks out their partial agency and refuses their definition as analytically subaltern. But, the description of the conditions of their emergence does, occasionally, sketch out traces of the silent, and often abandoned, members of the ashrams who did not escape and who remain, in these 24 handwritten pages, mute, civilly abandoned (Legg 2014) and thus, subaltern.

Abduction, escape, prosecution and archiving
The Inspector listed the cases by police jurisdiction and then by organizations suspected of abusing their position, followed by any cases that had arisen from that institution in recent years. Nine organizations were listed in total, and the typical description followed that of the Social Services Bureau in Paharganj. Its aim was said to be trafficking in women, though it claimed to be a shelter for widows and unfortunate women from brothels. No case had successfully been brought against it, though one had recently been registered, concerning two women 'kept' in the ashram. One was a widow of two months, the other was an orphan and a widow. The report clearly stated, without evidence, that the purpose of the ashram was that of selling women.
Where there were details of other 'kept' women in the report, this was because they had somehow escaped or drawn enough attention to themselves for the police to investigate. For instance, the Hindu Anath (orphanage) Sudharash in Bazar Sita Ram, which was described as 'a regular seat of barda faroshi' from the United Provinces in the East to the Punjab in the West, was the subject of an investigation in November 1935 under section 366 of the IPC (kidnapping, abducting or inducing woman to compel her marriage). On the first of that month a 'great noise' had been reported in the premises. On investigating, the police found four females confined there, who were described in varying detail. One, Must 9 Ramkali, was simply described as having been brought from an outside ashram, while Mst Janbi stated that she was abducted from her husband's guardianship in Cawnpore and was brought to the ashram without her consent. More, disturbing, detail was provided by Mst Shanda, who stated that she had been enticed by one Ganja Ram Brahmin, who left her in the ashram without consent. The ashram keepers, Piyari Lal and Munshi Lal, kept her against her will where they, and others, 'committed rape with her'. Twelve men had been arrested but no conviction was noted, nor was the fate of the women (the fourth of whose case appears below. For a legal history of rape in India at this time, see Kolsky 2010).
Further details emerged through the cases of the ways in which the women were enticed into the ashrams and their struggles for escape. Several cases centred on the Brahmukal Raksha Anath Ashram in the Kotwali jurisdiction, which encompassed the centre of Old Delhi around Chandni Chowk. One case, lodged on 27 October 1935, focused on the cheating by impersonation (IPC section 419) and dishonesty (IPC section 420) of a pimp, rather than centring on the entrapment of a woman. The case concerned a man named Qimat Mal from Karachi who passed through Delhi on his way to Lucknow, to which he was heading with the object of getting married (no details were provided of how). In Delhi, he met one Amba Pershad, a guide at the Empress Hotel, who took him to the Anath Ashram 'where girls were kept for marriage'. He selected one 'girl', an 18-year-old named Soshilla, and married her the same day after paying Rs 500. He took her to his native village in Karachi, Sind. She stayed with him for three months and then slipped away one night, taking with her the ornaments and clothes that she had been given. The next morning she was caught and later deposed to the police that she was actually a Muslim girl named Khursheed Begum, daughter of a Hakim 10 from Rawalpindi, from where she had been enticed away by two Hindu young men and taken to Lahore. She left them, but fell into the 'clutches' of one Parma Nand, a Pleader (a court advocate) who brought her to Delhi and left her in the Maharaja Hotel where she was kept for a month before being handed over to the ashram. Her case was, however, cancelled for lack of evidence.
A second case under IPC section 366 concerning the same ashram, lodged in August 1935, concerned Mst Sailan, wife of Ram Nath who was listed as a Jat. 11 Conforming to common depictions of the Jat community as hardy and proud, she alone among the other women listed in the report orchestrated her own escape and then went on to seek justice and revenge. Far from being physically immobile or irretrievable, Mst Sailan here speaks and moves. She had come from Rohtak, in the neighbouring state, and was met by one Lala Ram of Rohtak at the station who drove her in his lorry to the ashram, under the pretense of taking her to see her daughter in Delhi. She was detained and compelled to get married, but refused to do so. 'On getting an opportunity she escaped and visited her home in Rohtak and then returned to Delhi and reported the matter to the police.' In this case five persons were arrested in early September, though they were acquitted in October. Five women from the ashram gave evidence and were then allowed to return to their homes, having been brought to the ashram in similar circumstances.

Subaltern hierarchies: conflict, duped men and agents provocateurs?
The report has shown us how ashram subaltern spaces, marked by their aggressive silencing, were intrinsically related to their outside worlds. But they also emerge here as sites of complexity and internal differentiation, and show us that a group that might be thought collectively subaltern can have internal hierarchies of power and identity; the fine grained articulations of subalternity within itself. That is, not a space of radical alterity alone (Spivak 2014b, p. 10), but of familiar hierarchical disputes between workers of different status, and of perhaps unexpected gender roles. If the impression so far has been of relatively well-run barda faroshi trafficking networks, other cases complicated both the image of efficiency and the image of a strict divide between exploited women being beyond the law and of exploitative men working outside of the law. Both complications remind us of the hierarchies and complexities of subaltern spaces that we find detailed in the archive.
First, the organizers of the ashram and their trade disagreed. During June 1935, in an ashram in Katra Neel under the Hauz Qazi jurisdiction in the centre of Old Delhi, a girl (of unspecified age) was 'given' in marriage for Rs 400. However, on account of a quarrel within the management of the ashram over the sharing out of the price of the girl, the 'inhabitants' of the ashram pressed the landlord to get the house vacated, which he did. The manager left for Agra, 'where he [was] alleged to be doing the same nefarious business'. The 'inhabitants' were not defined but, given the lack of access to the outside world of the confined women that was widely reported elsewhere, we must presume it was the fellow 'managers' of the ashram who took revenge on the owner and got it closed down. A similar incident occurred at the Bharat Sewak Anath Ashram in Chauri Bazar, under the Hauz Khas jurisdiction in the city. On 18 September, a case was lodged on behalf of an unnamed worker at the Ashram along with a woman named Shanti and Piyari Lal (the connection is not made, but this ashram was listed directly alongside the Hindu Anath ashram in Bazar Sita Ram in which a Piyari Lal had also held women captive. The two bazars joined each other in the centre of the city). The crime was that of voluntarily causing hurt by dangerous weapons or means (IPC section 324). As above, the dispute was over being paid the correct share of the 'booty' for the sale of women in the ashram, and it was other ashram workers who had beat the three above with a knife and sticks. The case was acquitted, but Shanti and Piyari Lal will appear again, below.
The second complication of the ashram scene regarded disputes between the ashram organizers and visitors. Take, for instance, a case from the Janatan Dharm Yateemikhana orphanage in Charkewalan, a prominent shopping bazar in the centre of Old Delhi. The orphanage had been established in 1927, was run by a staff of 12 and housed 53 boys and 4 girls at the time of reporting. In February 1935, a case was lodged by one Suraj Bhai who reported that he had married (no financial transaction was recorded) a 14-15-year-old girl named Raimkali in the ashram, but had left her with the secretary of the ashram, Gundilla Lal Bhander. While away he paid the ashram a maintenance allowance for his wife, but during this time Lal Bhander allowed her to be 'enticed' by another man. The case ended in a compromise and Raimkali was returned to her husband. There are several questions which this brief but dense case note provoke. Why did Suraj Bhai leave his wife in such a place? At 14-15 years old, she was above the age of consent for marriage (though her age was not verified) so there was no reason why they could not be openly married, unless he himself was under-age (18)? Was the maintenance less than the cost of keeping her himself? Was he a mobile labourer without home? On what ground was section 363 of the IPC (kidnapping) invoked? Given that he had not paid for his wife, did he believe he was doing a social good by marrying a widow or destitute, and was shocked to find his wife confined and sexually abused? Was he a reform-minded Hindu, following Gandhi's invocation to overcome caste traditions and to marry across social divides? The short report offers no light on Suraj Bhai's motives or understandings of the orphanage, but it does complicate the role of men as simply traffickers, rapists or clients.
Another counter-typical gender role occurred in the case of the brother-inlaw of the fourth woman confined in Hindu Anath Sudharash in Bazar Sita Ram (the other three having been described above). In a confusing statement, the report suggested that Mst Bilaso had been brought by one Mahadeo to her brother-in-law, Madho, both of whom took her to the ashram, under the pretense that they had rented a house there. Mahadeo went back to his house in Sabzimandi, leaving Madho and his sister-in-law in the ashram. The report continues: Whether Madho had been duped by Mahadeo, or whether on seeing the state of the ashram he went back on an earlier deal is unclear. The pulp fictional nature of his disguise and infiltration also begs questions, and this account was explicitly stated to come from Bilaso's statement. Having been returned home, the extent to which her future depended upon Madho's treatment of his now raped (and in some eyes defiled and polluted) sister-in-law must be read between the lines of her testimony.
If these cases force us to resist easy assumptions about the men involved in the ashrams, so too do the actions of the following women. The Inderparasth Vinaushram Ashram situated in Nia Bazar near Lahori Gate, had been established in 1920 by the Arya Samaj and claimed to rescue widows and orphans. Uniquely, for this list, it was run by an elderly woman, named Krishnadevi. A kidnapping case (IPC section 363/366) was registered against her in September 1935 regarding Mst Buddowade. She was a 16-year-old woman who resided in Ghaziabad with her husband. After a quarrel she left him for her own family. On the way to Delhi an old man said he would take her to her father's house, but took her instead to the ashram and kept her there alongside 50 other women. Krishnadevi detained her and showed her to many prospective bridegrooms (during this time many others were sold). They were not allowed to talk or leave the building, but Buddowade did manage to befriend another girl who had likewise been abducted. The girls conspired and, making excuses of being ill, were sent to a hospital, with two guards. Both girls raised a 'hue and cry' in the bazar, shouting that they were being forcibly kept in the ashram. Hearing this, the police were called, and the case was registered against Krishnadevi as well as the chaprasi and her secretary.
A more extraordinary case of female participation involves a woman as collaborator and, even, agent provocateur. The case, from May 1935, centred on the marriage in the Bharat Sewak Anath Ashram between an outside male, Basa Lal, and ashram inhabitant Mst Shanti, on receipt of Rs 400. Shankar Lal of the ashram insisted that his fellow worker, Piyari Lal, accompany the bride and bridegroom during their onward travel. They all spent the night at Ghaziabad in a dharmsala, during which Shanti asked her husband to purchase her 'a shoe'. While away at the bazar, Shanti and Piyari Lal attempted to run away. On his return, Bari Lal raised the alarm and a police constable caught the couple. This was not a case of IPC 363 (kidnapping) but of IPC 420 (cheating and dishonesty). Shanti was obviously in on a deal with Piyari Lal, and it was these two who would later receive a beating by their colleagues for failing to fairly divide up the booty of a sale. Such cases were not uncommon; Punjab police reports of the 1920-1930s document what were known as 'willing wives', who supposedly agreed to be trafficked, and of women who would agree to be sold, only to abandon her husband a few months later and return to her abetters (Legg 2014, pp. 155-159). As the 1935 police report put it: 'The kidnapper is frequently no more than a dealer, and kidnapping does not enter into the matter. The women leave poor hardworked homes for comparative affluence where they are given the status of wives; and they willingly go with the dealer.' 12 No space, of course, was made for the wives' opinion on their own willingness. Though such cases feature regularly in the colonial archive the women remain, analytically, silent.
Analytically subaltern space? Absent women and big men Against the figure of the socially and physically mobile 'willing wife', the report also offers us sketchy outlines of a sexual subaltern defined by her (and possibly his) total immobility. If we accept that the women who are the subject of the cases above are not analytically subaltern, having found action ('hue and cry', or their release), voice (testimony), and even bureaucratic recognition (a case number), then who is? How can a report of cases and testimonies contain traces of a subaltern if the very act of featuring in such a report immediately disqualifies these subjects from subalternity? I would suggest that the following case does just that, through casting a different light over those cases that have gone before, so as to focus our attention on rooms and buildings which remain unexposed in the Inspector's report.
The case concerns the Brahmukal Raksha Anath Ashram, on Kucha (lane) Bulaqi Begam in Dariba, to the south of Chandni Chowk. The ashram was 8-9 years old in July 1935 when a case of wrongful confinement (IPC section 342) and robbery (IPC section 392) was levelled against it by Mst Shibbo, wife of Himat (caste Dhinwar (associated with water carrying) of Bulandshahr in the United Provinces). She reported that she was living with her husband in Qarol Bagh and had come to Chandni Chowk to buy clothes with another woman, named Must. Gomati, who left her near the clock tower outside the Town Hall. The proprietor of the Ashram, Kanwal Singh, met her in the bazar and persuaded her to accompany him. Once at the ashram, and following her refusal to marry, she was confined to the home and allegedly had her ornaments (jewellery) removed. She tried to escape but was forcibly detained by the owner and two men. On her raising a 'hue and cry' neighbours arrived and rescued her, during which time her friend and husband also reached the spot. Like so many of the examples above, the case was dropped in September. It was revealed, however, that there were about 13 women in the ashram at that time, including a minor girl of about 11 years of age who had been abducted, and whose case had been registered. The report states that three of the other women left for their own homes '+ the remaining stayed in the ashram'.
What became of these eight remaining females? This part of the report is the only explicit rendering of an assumption running throughout; unless an inhabitant somehow found voice and attracted the police, and the victims could prove they had a home to go to, the women were abandoned to the ashrams, regardless of whether its owners had been shown to be rapists, pimps or traffickers. They remain nameless, immobile, and wholly subaltern. The challenge here, however, is to refuse representation. Maybe the girls and women refused to leave; they may have preferred the life of the ashram and its opportunities to participate in the system in which it is easy to presume that they were victims. The evidence suggests the ashrams were sites of abuse, but this context does not denote victimhood in all of the females who dwelled there.
The report is also wholly silent on the boys who were listed as resident of these ashrams, outnumbering the girls who were listed: 52 boys to 5 girls in the Vedic Anath Ashram in Chawri Bazar; 56 boys to 11 girls in the Hindu Yatreen (orphanage) Khaana, in Bazar Sita Ram; 12 boys to 4 women in the Bhartia Anath Abija in Daryaganj. There is very little historical research on male prostitution in India (though see Legg 2012, pp. 30-34), although boys raised in brothels were often assumed to become pimps and touts, not prostitutes themselves.
If the report is marked by the absence of many women's voices, it is also marked by the near continual, 'secondary discourse' (Guha 1983) of the Inspector himself. In representing the facts of the cases he also inserted his own interpretative nouns and adjectives; the 'clutches' of the barda faroshes; the 'nefarious' business of the ashram owner; men of 'low morals' and 'unfortunate' women. Concluding his report, the Inspector offered his now unabashed interpretation of the cases gone before: The keepers of these ashrams are like hungry wolves, who in the name of humanity devour + plunder the public, + are leading life of debauchery. The pity is that the public though cheated, has sympathy with these institutes as they hear the religious 'big names' and are shown to have been keen to save the girls of one community going into the clutches of another.
The Inspector had earlier suggested that the women were not only forced to 'cohabit' with the ashram workers, but also with respectable 'big men' who patronized them. Here he was hinting at the spirit behind the formation of the homes which, especially through Arya Samaj and Sanatanist reform organizations, sought to prevent Hindus being converted to Islam.
This conclusion, though not the facts of the report, was contested by the Deputy Superintendent of Police, who appended a note to the report. He agreed that many low-caste Hindu women were kept and sold in the ashrams, but that the absence of 'moral courage' among decent Hindus meant that cases were not pursued, instead 'compromises' among the parties were reached. The Senior Superintendent provided his own note when passing the file on to the Deputy Commissioner, in which he subtly dismissed the emotive, secondary discourse of the original report: The City Inspector has possibly overstated the case against the ashrams, but from his investigations conducted by the police in specific cases it seems that the original charitable intentions of the founders of the ashrams have been subordinated to avarices and immorality.
The Inspector's suggestion that the ashrams be regularly inspected by the police was dismissed because IPC cases were necessary for an inspection to be justified. Given the lack of access of these women to the outside world, this marked the rubber-stamping of their subalternity.
The superintendent concluded that he would be glad to see the municipal committee look into the ashrams, but suspected this would be met with opposition due to the 'vested interests in these ashrams', by which he was referring to the presence of Hindu nationalists on the Municipal Committee who, it was presumed, would protect the ashrams for their supposed role in the forefront of the fight against aggressively (in terms sexuality and politics) expansionist Indian Islam. Others suspected an even more direct interest. Meliscent Shephard, the Indian representative of the London-based Association for Moral and Social Hygiene, was convinced that the 'big men' of the Delhi Municipal Commission protected not only ashrams but brothels themselves, suggesting that 'so many of the Municipal Councillors are themselves no unconcerned with the problems involved'. 13 Whether this was true or not, the ashrams in Delhi remained without substantial state intervention for the remainder of the colonial period, suspending their inhabitants in a state of civil abandonment (Legg 2014).
The nine ashrams investigated by the Delhi police emerge as dense and complex sites of voice and silence, agency and subjectivity, mobility and immobility, and provoke in us hope and despair. They compel us to condemn the subjectification, abduction, rape and selling of these women, but caution us against effacing their agency or will. The women of whom I have been able to write overcame one (empirical) form of subalternity through their voice and their vigorous refusal of silence; that is, through the embodied performance and shocked audience implied by a 'hue and cry'. These furious truth-tellings and ashram-shamings bought abducted women their freedom. The (analytical) subalterns here are the numbers that do not add up; the girls and women listed as ashram occupants but not as those that were freed.
What this reading has suggested, based on Gayatri Spivak's refusal to retract the definition of the subaltern as the figure that cannot speak while simultaneously insisting that we strive to 'touch' and 'hear' him/her, is that we might think of two types of subaltern space. One is an empirically traceable, experiential space of subalternity: a space of non-elite experience; beyond the state and beyond mobility. Such spaces are not rare (Spivak et al. 1996, p. 293); the majority of a population are non-elite and a large proportion of them may lack social mobility. The vast ranks of the subaltern have spaces which are distantly recorded and often retrievable, and which feature etches of immobile lives. But in examining these empirical spaces we catch glimpses and fragments of analytically unfindable spaces, beyond the archive. These irretrievable spaces include the spaces of the women and children who were left in the ashrams, whether at their own insistence or in spite of their pleas for rescue. The irretrievable spaces of the police report also include the non-investigated 90 or so other ashrams in the city, or the homes into which women had been sold and lived out their lives as possibly unwilling wives. Subaltern space here is, therefore, the conditions we can and must investigate and know. But it is also the space over the horizon, by definition unreachable, but a site that demands impossible exploration.

Disclosure statement
No potential conflict of interest was reported by the author.

Funding
The majority of the archival research undertaken for this paper was funded by a Philip Leverhulme Prize (2010).

Notes on contributor
Stephen Legg is professor of Historical Geography at the University of Nottingham. He is a specialist on interwar colonial India with a particular interest in the politics of urban space within imperial and international frames. He has analysed these spaces and frames through drawing upon theoretical approaches from memory scholarship, postcolonialism, political theory and governmentality studies. His publications include Spaces of colonialism: Delhi's urban governmentalities (2007, Wiley-Blackwell), Prostitution and the ends of empire: scale, governmentalities and interwar India (2014, Duke University Press) and the edited collection Spatiality, sovereignty and Carl Schmitt: geographies of the nomos (2011, Routledge). He is currently focusing on Indian constitutional reform in the 1920-1930s as well as continuing a longer project on Spaces of Anti-Colonialism in Delhi.