Lessons from love-locks: The archaeology of the contemporary assemblage

Loss of context is a challenge, if not the bane, of the ritual archaeologist’s craft. Those who research ritual frequently encounter difficulties in the interpretation of its often tantalizingly incomplete material record. Careful analysis of material remains may afford us glimpses into past ritual activity, but our often vast chronological separation from the ritual practitioners themselves prevent us from seeing the whole picture. The archaeologist engaging with structured deposits, for instance, is often forced to study ritual assemblages post-accumulation. Many nuances of its formation, therefore, may be lost in interpretation. This article considers what insights an archaeologist could gain into the place, people, pace and purpose of deposition by recording an accumulation of structured deposits during its formation, rather than after. To answer this, the article focuses on a contemporary depositional practice: the love-lock. This custom involves the inscribing of names/initials onto a padlock, its attachment to a bridge or other public structure, and the deposition of the corresponding key into the water below; a ritual often enacted by a couple as a statement of their romantic commitment. Drawing on empirical data from a three-year diachronic site-specific investigation into a love-lock bridge in Manchester, UK, the author demonstrates the value of contemporary archaeology in engaging with the often enigmatic material culture of ritual accumulations.


Introduction
Loss of context is a challenge, if not the bane, of the ritual archaeologist's craft. Those who research prehistoric ritual frequently encounter difficulties in the interpretation of its often tantalizingly incomplete material record. Although Richards and Thomas (1984: 215) rightly stress that 'ritual is not beyond the realm of archaeological inference', many questions cannot be answered with only fragmentary evidence.
Ritual activity may be evident in the form of material structured deposits, for example, but often the details surrounding their deposition do not survive. As Fontijn (2002: 275) observes in his work on the sacrificial and depositional landscapes of the Bronze Age: We know something about the treatment of the object deposited as well as its earlier history, but many questions remain. How was the actual depositional procedure carried out? On what occasion was it done, which people were present, what further activities did it involve and so on? All these aspects may contribute to a further understanding of the meaning of depositional practices, but they are practically beyond the limits of archaeological knowledge.
Careful analysis of material remains may afford us glimpses into past ritual activity, but our often vast chronological separation from the ritual practitioners themselves prevents us from seeing the whole picture. We may be able to answer the what, where and when of ritual activity (e.g. in identifying and dating a structured deposit) and we may be able to address the how (e.g. in recognizing that it was deposited in a certain environment by a certain method). The specifics of these, however, can often elude or even mislead us -just because an artefact dates to a certain period does not necessarily mean that it was deposited in that period -while the who of these ritual activities often frustratingly evades us. Is a mass deposition evidence of a hoard, deposited by a single individual/ community at one time, or an accumulative assemblage, added to by different individuals/communities over a long period? We can make educated inferences but we cannot know. Thus is the nature of archaeology.
However, what if an archaeologist could have been present at the time of deposition? What if he or she was not separated from the ritual activities by time but could observe them in practice? Even if the archaeologist could not engage with the practitioners themselves (to answer that notoriously elusive question of why), what knowledge would be gained from recording structured deposits as they are deposited, rather than millennia later? These are questions that this article will address by adopting the methodologies of contemporary archaeology. I will compare past approaches to prehistoric deposition with my research on a contemporary depositional practice: the love-lock. This custom involves the inscribing of names/initials onto a padlock, its attachment to a bridge or other public structure and the deposition of the corresponding key into the water below. This is a ritual often enacted by a couple as a statement of their romantic commitment.

Introducing the love-lock
One padlock attached to a bridge could have served a utilitarian purpose. Three padlocks attached to a bridge are less likely to have been functional, but are not prominent enough to be noteworthy. Five padlocks are even less likely to have been functional, but may still be overlooked. Seven padlocks attached to a bridge, however, prove both remarkable and recognizable as an assemblage of structured deposits -at least according to personal experience.
I traversed Manchester's Oxford Road bridge for many years; situated between home and university, crossing it was an almost daily occurrence. As bridges go, this one is small and inconspicuous, consisting only of four sculptured metal panels running alongside the pavement of a busy city-centre road. Indeed, it is barely recognizable as a bridge; only by peering through the metal panels would pedestrians know they were walking over the Rochdale Canal ( Figure 1).
It was on 12 February 2014 that I first noticed the padlocks attached to the bridge. There were seven in total, distributed across the four metal panels. Judging by the level of rust on three of the padlocks, they had probably been there for a significant amount of time, but it was not until the addition of four more that I first took note of them. Three padlocks were apparently not prominent enough to capture the attention of a casual passer-by; seven padlocks, on the other hand, were quite conspicuous.
Five of the seven padlocks bore inscriptions and adornment. Thick black marker spelled out what I assumed to be the initials 'D B' on one side of a padlock and drew a love-heart on the reverse. More probable initials 'J' and 'B' flanked a love-heart on another, whilst initials accompanied the Spanish phrase for 'I love you' (Te Quiero) on the third padlock. The other two inscribed padlocks each bore a pair of names, one a love-heart, and both the number '2013'. If this number indicates year and the year indicates the date of deposition of the padlock (see more in-depth discussion of this below), then those two padlocks had been attached to the bridge for at least 43 days before I had noticed them.
Pairs of initials and names, the adornment of love-hearts and the inscription of Te Quiero all amounted to my identification of these padlocks as love-locks, as defined above. The origins of this practice are unclear but residents of Vrnjačka Banja, Serbia, claim that their assemblage on the Most Ljubavi ('Bridge of Love') dates back to WW1. However, it gained popularity following Italy's adoption of the custom in the 2000s, spurred by Federico Moccia's 2006 romantic novel Ho voglia di te (I want you), in which a character attaches a padlock to the Ponte Milvio, Rome. The subsequent dissemination of this practice was rapid and geographically unbound, with love-lock accumulations emerging in locations as distant and varied as Paris and Taiwan, New York and Seoul, and Melbourne and Moscow, demonstrating a custom's capacity for construction, growth and widespread dissemination without the impetus of authoritative agents. Considering the speed and range of their dissemination, it is unsurprising that love-locks have featured in academic research from a variety of perspectives. Art historian Cynthia Hammond (2010) focused on the accumulation in Pécs, Hungary, which dates to the 1980s, illustrating how it can be perceived as representative of control and dissent in the city. Urban scenographer Jekaterina Lavrinec (2011Lavrinec ( , 2013 classified the custom as an 'urban ritual', considering the emotional and bodily experiences of love-lock deposition. Engineer Christian Walloth (2014), describing love-locks as 'emergent [i.e. unplanned and in principle unpredictable] qualities', explored the influence they have on urban development and planning. Artist Lachlan MacDowell (2015) considered love-locks within the context of street art, exploring deposition through the theory of stigmergy, whereby 'urban practices cluster spatially, without direct coordination' (p. 41). While social scientist Kai-Olaf Maiwald (2016) adopts an objective-hermeneutic approach in his investigation into the symbolic meaning of 'padlocking' at the Hohenzollern Bridge in Cologne.
However, as far as I am aware, there has not yet been any research published on lovelocks that traces the development of an assemblage. Like many archaeologists of prehistory, researchers focus on these accumulations at one static point in time. This is surprising considering the potential value of an ongoing investigation into structured deposition, and also surprising given Duncan Garrow's (2012: 91) assertion that 'Processes of deposition (and thus also their interpretation) are a fundamental element of archaeology', and have been since the early 1980s. The contemporaneity of love-locks does not negate their status as material evidence of structured deposition, as defined by Richards and Thomas (1984: 115): 'formalised repetitive actions which may be detected archaeologically through a highly structured mode of deposition'. That the love-lock is a structured deposit -in that it is deposited via formalized repetitive actions -cannot be questioned, and yet its value within contemporary archaeology has yet to be explored. Following on from my assertion in a previous article that more archaeological attention should be given to the 'wealth of information contemporary deposits can proffer' (Houlbrook, 2015a: 187), I hope to redress this oversight.

Methodology
Having already researched the practice of love-lock deposition from a contemporary archaeological perspective in London and Moscow for a section of my thesis, I was naturally excited to find a fledgling assemblage so close to home. I photographed and recorded the seven love-locks, curious to see if, despite their small number, they would attract more. And they did. Less than a week later, on 18 February, another one appeared. There was a gap of three weeks before the next love-lock was added, on 11 March, but two more had been added by 3 April, another by 16 April and three more by the end of May. Thus is the nature of accumulation, which Gamble (2007: 122) describes as having a 'magnetic-like effect'; deposits attract more deposits.
I decided to continue recording additions of these love-locks in situ as a site-specific investigation into the material culture of contemporary depositional practices. Every addition was photographed and assigned a catalogue number, prefixed with MLL (Manchester Love-Lock), which individual examples will hereafter be referenced by. The plan was to take weekly inventories and photographs of the Oxford Road bridge, either until the assemblage was removed or it ceased growing. Three years later, neither of these had occurred. I have now amassed a catalogue of 409 love-locks, along with 27 'deviant deposits' (see below), which represent 36 months of deposition ( Figure 2). This material is valuable for a variety of investigations, such as into the impact of contemporary deposits on urban space and place; the subjectivity of heritage; and the 'reenchantment' of society, to name only a few. However, I plan to tackle each of these topics in future publications, while the focus of this article is purposefully narrowed to one primary question: what insights can be gained into the place, people, pace and purpose of deposition by recording an accumulation of structured deposits during its formation, rather than after?

The place of deposition
The locations of love-lock accumulations vary greatly. For example, I have recorded love-locks attached to the landslide mesh along Tuscany's Cinque Terre; on a fence within view of Montmartre's Sacre Coeur; and on the flag rings of surviving sections of the Berlin Wall. However, the majority of accumulations do occur on bridges, from Paris' Pont des Arts, the Brooklyn Bridge, and the railings crossing the canal near Lennon Wall in Prague, to the lesser-known structures spanning the River Wye in Bakewell, Derbyshire, and the River Wharfe in Otley, Yorkshire ( Figure 3).
The popularity of the bridge is easy to account for. The symbolic custom of dropping the love-lock's key into a watery place does necessitate a body of water. The significance of the watery place will not, of course, be lost on the archaeologist, with so many prehistoric and historic hoards, assemblages and deposits having been discovered in rivers and marshlands (see Aitchison, 1988;Bord and Bord, 1985;Bradley, 1990;Cool and Richardson, 2013;Fulford, 2001;Laing, 1969;Merrifield, 1987). Focusing on the Bronze Age, David Fontijn (2002: 264) proposes a number of possible reasons for this: 'purity, pollution, regeneration, fertility', as well as the status of the river as a boundary between peoples and worlds, and as a 'central element in people's perception of land- scape ' (p. 263). This is still the case today, with many of the world's cities and towns situated -and orientated -along the banks of a river.
However, it is easy to over-interpret the choice of location for the contemporary lovelock (and, indeed, for prehistoric deposition). Would a depositor of a love-lock cite 'purity, pollution, regeneration, fertility' as the factors motivating his or her decision to attach a love-lock to a bridge and drop the key into a river? Beliefs are notoriously difficult to pin down, the why of ritual activity often being the most elusive question to answer, especially when armed with only remnants of material evidence. It is, however, possible to detect the more pragmatic factors; to consider what water does rather than what it symbolizes. Fontijn (2002: 266), for example, considers the physical qualities of bodies of water: They 'seal off' the invisible parts of the world: the muddy bottoms of streams and rivers … Throwing a gold-glimmering bronze axe into such a place must have been an act whereby the onlookers really got the impression that the object disappeared completely. Sunk to the bottom of the marsh, it could no longer be seen or retrieved anymore.
The dropping of a padlock's key into a river could have the same effect; the depositors can watch as it sinks from view. Richard Bradley (1990), writing of prehistoric hoards, describes how many offerings were deposited in such a fashion so as to be irredeemable. Deposits could either be physically damaged, rendering their economic value moot, or deposited in a location from which they could not be recovered. The depositors of lovelocks employ both methods; the key deposited in the water becomes irredeemable, while the padlock -locked and lacking a key -can no longer serve a utilitarian function. Through this 'fragmentation', to use Brück's (2006) term, the love-lock transitions from functional tool to symbolic deposit.
The symbolism of the bridge itself may also be significant, for its purpose is to connect; in uniting two previously separated sides, the bridge is an ideal platform for a custom typically observed to declare romantic attachment (Maiwald, 2016). However, again we risk over-interpretation. The popularity of the bridge as a place of deposition more likely stems from a far simpler reason: its proximity to the water in which the key is dropped. The depositors can stand on the bridge and watch as the key disappears from view in the river or canal below. The bridge, therefore, is more a matter of convenience than metaphoric potency. This may not have always been the case. The symbolism of the bridge and the water below may, in the past, have motivated their selections as sites of deposition. Thenthrough what Hartland (1893: 461) scathingly terms 'the process of ceremonial decay' -these earlier motivations were skewed or forgotten. It is, however, equally likely that the opposite has occurred; that bridges were initially chosen as sites of structured deposition for pragmatic reasons, but later, as the custom became more widespread and well known, it accrued a symbolic significance (see Davies, 2015: 403).
Whichever theory is correct, it is not difficult to understand the popularity of bridges as places of contemporary deposition, nor is it difficult to understand the popularity of world-famous bridges such as the Pont des Arts, the Ponte Milvio and the Brooklyn Bridge. But why was the Oxford Road bridge chosen? There are other, equally suitable structures in Manchester, a city criss-crossed with waterways. Indeed, a five-minute walk north of the Oxford Road bridge takes you to an almost identical structure, also spanning the Rochdale Canal, consisting of three sculptured metal panels, on Great Bridgewater Street. Despite my recording four love-locks attached to this bridge in August 2015, no more have appeared since. Why has the Great Bridgewater Street accumulation stagnated at four, while Oxford Road has exceeded 400? Aesthetics surely play no part, for the former is more conventionally scenic than the latter. Proximity to tourist sites is probably equally negligible, for Great Bridgewater Street is as close to the Bridgewater Hall as Oxford Road is to the Palace Theatre.
Levels of pedestrian traffic, however, are distinguishable between the two sites. The Oxford Road bridge's location on a busy road running between the city's main shopping area and the university, along with its proximity to Manchester Oxford Road train station, ensure that it is passed by high volumes of people on a daily basis. And the more passers-by who observe the assemblage, the more potential there is for people to continue populating it, with deposits attracting more deposits (see more discussion below). Great Bridgewater Street, on the other hand, is quieter, attracting fewer pedestrians and -apparently -far fewer love-locks.

The pace of deposition
Before archaeologists can consider the pace of deposition, they must determine whether an assemblage is a 'hoard' or whether it is technically an 'accumulation', defined by Laing (1969) as a collection of artefacts deposited over a period of time. This relates directly to the question of temporality, an element that archaeologists -ironically -can be guilty of overlooking. Garrow (2012: 90), for example, criticizes studies that omit 'any serious consideration of the effects of time in creating the patterns observed' (emphasis in original). Time, writes Garrow, 'is often flattened significantly, as the deposits plotted two-dimensionally across causewayed enclosures and henges are compared without full consideration of the temporality of their deposition' (p. 109). Assemblages of multiple artefacts are often presumed, wrongly or rightly, to be singleevent depositions, as observed by Bradley (1990: 6).
Despite the prevalence of such assumptions, there is compelling evidence to suggest that some assemblages were accrued over long periods of time. Some sites appear to have been repeatedly resorted to as places of deposition; Fontijn (2002: 260) terms such sites 'multiple-deposition zones', citing several Bronze Age examples from the Netherlands (e.g. the terrace swamp near Belfeld, the inland swamps between Echt and Montfort, and the stretch of the Waal river near Nijmegen). In the case of the love-lock assemblage, it has already been recognized as a 'community deposit', in that the material evidence indicates multiple depositors. However, how do we determine whether the Oxford Road bridge is a 'multi-deposition zone'? How can we tell if the Manchester love-locks were deposited by a community at one time, or whether they represent an accumulative assemblage, added to by different members of the community over a long period?
Employing the methods of contemporary archaeology, this question is easy enough to answer. By cataloguing the Oxford Road assemblage over a three-year period, observing it grow from 7 in February 2014 to 409 by February 2017, I can proffer indisputable proof that the Manchester love-locks are the process of accumulative deposition. This is one particularly notable benefit of diachronic documentation. However, without the ability to record the growth -if, for example, I were to encounter the Oxford Road assemblage for the first time as it currently stands -how might I recognize it as an accumulation rather than a hoard? How might I determine a timescale for its deposition?
Archaeologists have employed various methods in determining the timescales of prehistoric and historic deposition, usually hinged upon the date of the deposited artefacts. The Hallaton Hoard is one such example. Discovered in 2000 and excavated by the University of Leicester Archaeology Unit (ULAS), this site, situated on a hilltop in Hallaton, southeast Leicestershire, has yielded the largest assemblage of Iron Age coins recovered under controlled archaeological conditions in Britain. Over 5000 Iron Age and Roman Republican gold and silver coins were recovered, along with a Roman iron cavalry helmet (Leins, 2007(Leins, , 2011Williams, 2003), which appear to have been deliberately buried in at least 15 separate hoards at a site that is proposed to have been an open-air gathering place with possible ceremonial significance (Leins, 2007: 39;Priest et al., 2003: 359-360;Score, 2006Score, : 206, 2011. The archaeologists interpreting the finds at the Hallaton Hoard were able to estimate a relatively short time-period of deposition: late pre-conquest and/or the early Roman period (Leins, 2007(Leins, : 25-26, 2011, the majority of the coins having been issued roughly between AD 20-50. For the deduction of a time-frame, they used the testimony of the coins, which are particularly valuable finds for the archaeologist precisely because of this ability to offer a relatively accurate means of dating (Betlyon, 1985: 163). Their date of manufacture is often stamped on the artefact for all to see; numerals incorporated into their materiality thus become central to their interpretation. In this way, love-locks could be similarly analysed.
Of the 409 Oxford Road love-locks catalogued, 196 (48%) bore dates. Some were professionally engraved but most were hand-written, varying in specificity from years (e.g. MLL7: '2013') to precise dates (e.g. MLL294: '26.5.16'). In all likelihood, if an archaeologist recorded these in the future, they would be interpreted as the dates of deposition, and certainly the vast majority of them probably are. However, of the 196 dated love-locks, 95 (48%) bore dates that could not signify the day the love-lock was attached to the bridge. Some are dated too early to be consistent with their deposition.
For an extreme example, MLL180 bears the number '1999' -presumed to indicate a year -but cannot have been deposited until 16 years later; it was recorded as having been added between 29 September 2015 and 8 October 2015. Others are less extreme but still clearly demonstrate a time-lag between date inscribed and date deposited. MLL35, for instance, bore the date '4.12.13' but was first recorded over nine months later, on 29 September 2014, while MLL150, recorded on 25 August 2015, contained the date '26.08/2012'. A date on one love-lock actually post-dates its deposition; MLL17 was recorded 7 July 2014, but bears the date '4.10.14'. Either this was in reference to a future planned event (a wedding perhaps) or it was written in an Americanized sequence (therefore signifying 10 April 2014), in which case it was referring to a date in the past.
Fourteen love-locks contain two dates, only one of which could indicate the date of deposition, if indeed either of them do. MLL49, for example, has the dates '20/10/11' and '21/11/14' hand-written, one on top of the other, while MLL141 has '4/9/15' written on one side and '15/8/15' on the other. MLL373 has '1930' written down one side and '2016' down the other. If one date signifies the time of deposition, then what does the other refer to? Five of the love-locks provide clues. Two appear to pertain to the start of a relationship: 'since 2K8' (MLL223); 'since 16/12/2015' (MLL343), while another may indicate both the beginning and the end: 'January -June 2015 It's never goodbye only see you later' (MLL100). One love-lock, containing words in Spanish, possibly denotes a period of time spent visiting Manchester: 'Manchester 23/10/15 15/11/15' (MLL221), and another, commemorating a birthday, probably signifies the celebrant's year of birth as well as the year of deposition: '1993 2016' (MLL322).
These give us an indication of what the dates on other love-locks could refer to: anniversaries, years of birth, or the time-frame of a trip, and it is likely that there are a variety of other events they could commemorate. This reminds us of an important lesson; that the date marked on a deposit does not necessarily denote the date it was deposited. Likewise, archaeologists have recognized that just because an artefact dates to a particular period does not necessarily indicate that it was deposited during that period (Bradley, 1990: 186). Some hoards, for example, have been found to contain artefacts that were hundreds, possibly even thousands, of years old at the time of their deposition. The 600 items comprising the Salisbury Hoard, for example, date to the Bronze and Iron Ages, but are estimated to have been buried c. 200 BC (Stead, 1998: 123). As Eckhardt and Williams (2003: 142) observe, 'It has been suggested that the antiquity of these objects encouraged them to be valued as sacred items and appropriate offerings to gods.' Additionally, antiquities may be deposited to apply a form of artificial patina to a custom, strengthening a sense of authenticity (Kalshoven, 2010: 68-69). Dating a deposit, therefore, can surely only proffer a terminus ante quem, while its terminus post quem remains elusive -and the date marked on a love-lock cannot even provide the former, as some clearly precede the time of deposition.
However, in the case of diachronic documentation, an accurate timeframe of deposition can be established. Although it is not known when the first seven Oxford Road lovelocks were locked in place, the author has been able to identify the deposition dates (accurate to within 1-2 weeks) of the subsequent 402. Through this systematic cataloguing, the pace of deposition has been recorded, and as the cumulative graph in Figure 4 demonstrates, the Oxford Road accumulation has continued to grow. Over a three-year period, no month passed without the addition of at least one love-lock, with 11.4 being the mean average quantity added per month. This is unsurprising given the 'magneticlike effect' (Gamble, 2007: 122) of accumulations, with deposits believed to attract more deposits. There have, however, been a few surprises.
The first six full months of recording (March-August 2014) revealed an exponential increase in the rate of deposition. One love-lock was added in March, 2 in April, 3 each in May, June and July, and then 11 were deposited in August. Following this, I had expected to see a continuing exponential escalation and had predicted that the Oxford Road assemblage would continue growing until it reached critical mass, at which point there would be marked increase in the numbers of love-locks being deposited. However, three years since its inception, no such critical mass was reached. As Figure 4 demonstrates, although there is a continuous ascending trend, it is a steady rather than a sharp increase. Perhaps, in contrast to the more established assemblages of the Pont des Artes and the Ponte Milvio, the Oxford Road bridge is too modestly sized or inopportunely located to attract immense quantities.
Another surprise occurred when the data was transferred to a non-cumulative graph ( Figure 5). The increase appears far less steady here than it does in Figure 4. Instead of consistency, there are sharp spikes and slumps in the quantities of love-locks added from one month to the next. December 2014 saw only 3 love-locks deposited, while January 2015 saw 11; only 4 were deposited in May 2016, while there were 14 added in June 2016. What Figure 5 demonstrates, therefore, is that there have been periods of intense depositional activity (August 2015 saw the highest quantity with 29) and periods of very little (March 2014 saw the lowest with only one). What might account for these fluctuations? Garrow (2012: 111-113) considers how an archaeologist might interpret similarly irregular patterns of deposition. 'It would certainly be possible', he writes, 'to interpret patterns such as these as having been intentionally (and meaningfully) constituted in the past.' Might the spikes of love-lock deposition in August 2015, for example, have been deliberately instrumented?
Given the casual and unofficial nature of this practice, intentional and meaningful patterns seem unlikely, and Garrow certainly advocates considering more 'everyday' explanations for such fluctuations (p. 113), drawing on a comparison with the patterning of contemporary waste. Analysing quantities of different material types recycled each month in Merseyside during the 2010-2011 financial year, Garrow notes spikes and slumps that are easily attributable to patterns in everyday life, such as the increase of garden waste during the summer months and glass disposal over and after the Christmas period (pp. 110-111). As Garrow concludes: These patterns of variability were not created intentionally in order to convey a symbolic message. The people making these deposits will not have been aware of the patterns they helped to create. However, that is not to say that those patterns are meaningless. They do have something significant to say about the rhythms of everyday practice. (p. 111) So what might the fluctuations in love-lock deposition reveal? If, like Garrow's contemporary waste, deposition was linked to particular seasons or festivals, the assumption might be that Valentine's Day, for example, would see an increase in love-locks; what better day to ritually declare your love for another than on the day traditionally associated with romance? However, February witnessed only minor spikes. Perhaps Christmas, another romantic time of year? Again, December saw little increase (with December 2014 even seeing a decrease). So when do we see the spikes?
I am still at a loss to explain the spikes of October 2016 and January 2017, but can suggest a few possible reasons for the surge in August 2015. Firstly, the summer holidays increase the number of visitors to Manchester, increasing the traffic on Oxford Road, and subsequently increasing the amount of potential depositors. Secondly, August Bank Holiday weekend is when Manchester hosts the Gay Pride parade, passing close to the Oxford Road bridge, which witnesses upwards of 40,000 participants. Not only does the parade further elevate visitor numbers to this particular area of the city, but it also generates a celebratory or commemorative atmosphere that might inspire depositional activity. The spike in August 2015, therefore, may not have been intentionally created, but may nevertheless reveal something significant about the rhythms of city life.
Thirdly, a newspaper article that appeared in Manchester Evening News (Butler, 2015) on 9 August 2015 entitled 'Love locks in Manchester to get a special city centre', details the location of the Oxford Road bridge and explains the custom: 'Around 100 locks are already attached to the bridge on Oxford Road like they have done in cities around the world, including Paris and New York.' It would not be unreasonable to assume that some of the depositors in August 2015 had been inspired by this newspaper article, hinting at the role the media might play in the perpetuation or dissemination of depositional practices.

The people of deposition
In general, these love-lock assemblages are in publicly accessible and visible spaces. This suggests that they were, to use Needham's (1988: 246) term, 'community deposits', in the sense that they are deposited overtly rather than privately, 'in the knowledge … of society at large'. Their sheer quantities also suggest that the assemblages are not the work of one or two individuals, but of many. However, in another sense they are the opposite of Needham's community deposits: they are not attached to bridges 'to the benefit' of society at large. Instead, each love-lock appears to have been employed as a personal deposit, placed for the benefit of the individual depositor(s) rather than for the community.
Indeed, there is a certain individuality to these deposits. Granted, their quantities and the general homogeneity and alienability of padlocks, none of which bear inherent physical qualities overtly linking them to their depositors, contribute to a sense of collective anonymity rather than to conspicuous consumption. Some objects, as Snodgrass (2006: 265) observes, are 'too numerous and too cheap to be seen as motivated by competitive ostentation'. However, as I argued in my research on contemporary British coin-trees, by altering context and use, a common and alienable object -from a coin to a padlock -can become a highly personal deposit (Houlbrook, 2015b). The material evidence of the Oxford Road bridge assemblage demonstrates that the depositors were concerned with distinguishing their love-locks from the masses, with harnessing these padlocks as expressions of their personal identities or relationships. How do they do this? By altering the materiality of the padlocks.
The most popular alteration is the addition of an inscription: 375 of the 409 love-locks (92%) bear some form of inscription, either professionally engraved or hand-written in marker, pen, correction fluid, nail varnish, or scratchings. The majority of these inscriptions (344) contain an overt method of personalization: initials or names, presumably (although not necessarily, see below) of the depositors. This method of personalizing otherwise anonymous deposits through the use of initials is certainly not unique to the love-lock custom. For example, the practice is evident in 12th-century Corinth, where a buried hoard of 30 gold nomismata of Manuel I (1143-1180) was uncovered in the 1938 excavations of Old Corinth, 14 of which had graffiti, such as letters, scratched onto their surfaces. Josephine Harris, who published the find, suggested that this graffiti may have been used as 'identification marks' (Harris, 1939: 273). Likewise, in my work on cointrees, I observed that the coins ritually inserted into trees in modern-day Britain frequently bear initials and other identifying markers (Houlbrook, 2015b: 203-204). Love-locks have been similarly personalized.
Methods other than the inscription of initials and names have also been employed. Over half (234) of the Oxford Road love-locks have been distinguished through the adding of adornments, the majority of which are love-related symbols: love-hearts, crosses, lemniscates and flowers. Some depositors, however, have put greater effort into their adornments ( Figure 6). The depositors of MLL32, for example, drew a stick-figure couple, most likely representing themselves; MLL102 is embellished with delicate lovehearts, starbursts and a cloud; while MLL160 is adorned with white flowers with green stalks. Red love-heart stickers had been added to MLL286 and MLL321, with the latter further distinguished by a piece of silver thread tied to the shackle. A red ribbon had been tied around the shackle of MLL214; a strip of fluffy green material was wrapped around MLL243; decorative, flowery tape covers MLL122; while, interestingly, a plastic skeleton tangled in silver tinsel was attached to MLL402.
Colour has also clearly been utilized in distinguishing one love-lock from many. As Jones and MacGregor (2002: 12) observe, 'Colour is powerful in the construction of difference.' Bright red, blue and pink nail varnish, glitter paint and gold marker are just some of the media employed to demarcate love-locks. While other padlocks may have been selected for deposition by the virtue of their manufactured colour -with brass being the predominant material of the accumulation -those padlocks that are coloured green, yellow, purple, red, pink, blue and zebra-print stand out from the rest. Other padlocks may have been chosen for their size, some being much larger than others; a boast, perhaps, that the depositors' love for each other is 'bigger and better' than that of other depositors? Likewise the decision to commission a professionally engraved love-lock may have been motivated by a desire to demarcate one from the many.
Indeed, such examples of conspicuous consumption are not without precedent in ritual contexts. Bradley (1990: 137), for example, considers the role of prestige -'the common currency of non-market societies' -in Late Bronze Age deposits, when opulent ritual offerings were intended to lend themselves to 'the quest for personal prestige ' (p. 188). Ritual deposits in Archaic Greece were similarly harnessed, with Day (2010: 182) observing that 'competitive self-presentation or social display on the dedicator's part played a major role in dedicatory practice.' Anthropologists and economists have similarly detailed evidence of conspicuous consumption in contemporary societies (see Douglas, 1979;Mason, 1998). Coins are a prime example. What, for example, is being communicated when a person deposits a high-denomination coin in a fountain or a cointree otherwise filled with pennies (Houlbrook, forthcoming)? Evidently even objects as small and homogeneous as coins -and padlocks -can lend themselves to 'personal prestige' and 'competitive self-presentation'.
As well as processes of conspicuous consumption, the adornment of love-locks may also be aimed at creating objects that represent their depositors. By physically associating the love-lock with a couple, the love-lock essentially becomes that couple's 'tag', their expression of identity. Love-locks and depositors become entangled (Hodder, 2012). This entanglement and the desire to associate a ritual deposit with the depositor's identity stretches back to antiquity, at the latest (see Dowden, 2000: 176), but these objects are not designed to only represent the depositor; they are intended to be the depositor. As Tilley (2006: 63) writes, the 'thing is the person and the person is the thing', a merge that Gell (1998: 74) terms the 'objectification of personhood'. This objectification leads to Gell's theory of 'distributed personhood', whereby the deposit becomes a detached part -a 'spin-off' -of the depositor (p. 104). In this way, therefore, each of the love-locks on the Oxford Road bridge is intended as a detached part of the couple who deposited it.
However, by studying the Oxford Road assemblage diachronically, the assumption that each love-lock represents a couple is soon proved erroneous. Interestingly, 25 of the love-locks appear to have been 'hijacked', in that they were deposited at one point in time and then were (re)inscribed or (re)adorned weeks, months or years later (Figure 7). Two examples of this are MLL2 and MLL3, first recorded on 13 February 2014 as bearing no inscriptions. Less than three weeks later, two names had appeared on MLL2; a further month later, two names had appeared on MLL3, accompanied by a date five months later. Occasionally inscriptions are superimposed over the original messages, with new names obscuring the old ones, creating a palimpsest effect. MLL230 has in fact been hijacked twice. First recorded in February 2016 with a professional engraving, it was then recorded in April 2016 with handwritten names superimposed over the original inscription, and then again in January 2017 with a different set of names.
However, in most cases, later inscriptions are made (more respectfully?) on bare padlocks or over heavily-faded messages, in a process of 'ritual recycling' (Houlbrook, 2013). Clearly these are cases of people wishing to make a deposit but being unable or unwilling to source their own padlocks, and so they 'recycle' or 'redeposit' another's, creating a palimpsest of deposition. By adding their names, initials or messages to the love-locks of past depositors, these opportunistic 'recyclers' -for want of a better term -are still participating in this custom, albeit in a distorted form.
However, this not only distorts the custom; it also skews the material evidence. Looking at the Oxford Road assemblage in toto, postliminary to the processes of accumulation, might lead the archaeologist to assume that one love-lock represents one set of depositors. For at least 25, however, this is not the case. These love-locks represent the activities of at least two, possibly three, sets of participants: the original depositors and also the subsequent recyclers. In the more common case of later inscriptions being added to bare love-locks, though, this process of ritual recycling would not be materially evident, and so would be easy to overlook.
Such recycling of past deposits is not without precedent within the archaeological record, with Ian Armit's consideration of the Cnip skull offering one particularly illuminating example. During excavations of Cnip, a 1st-2nd-century AD settlement on the west coast of Lewis, a cache of structured deposits was discovered beneath the foundations of a building. Included in this cache was the upper part of a human cranium, and Armit (2012: 225) notes how easy it would be to interpret this as evidence of a foundation sacrifice -except for the fact that this cranium dates to 1540-1410 BC, nearly a millennium and a half before it was buried beneath the building's foundations. This example not only illustrates that the date of deposition does not necessarily correspond with the date of the item (as observed above), but also that objects previously deposited can be re-deposited at a later time by other individuals or groups.
This teaches a vital lesson: that an object's biography does not end at its moment of deposition. As Fowler (2004: 65) observes, 'Artefacts, like people, are multiplyauthored'; and, as Gruner (2015: 57) notes, they can 'go beyond the intentionality of the person who originally constructed them'. The structured deposit is a product not necessarily just of its original depositor, but of later 'authors', whose engagement can alter the object both tangibly and intangibly. It is not, however, only the recycling of love-locks that demonstrates their continued biography. There is evidence suggesting that some love-locks were in fact later altered by their original depositors.
According to the widely-known 'tradition', by attaching a love-lock to a bridge and throwing the key into the water below, the depositors are affirming their commitment to each other. Should they ever wish to separate, the key must be fished out of the water and the love-lock symbolically unlocked. By studying a love-lock assemblage only after the processes of accumulation, it would be difficult -impossible perhaps -to judge whether or not some love-locks had been removed. However, a diachronic study allows for this assessment, and it has revealed that at least eight love-locks have been removed from the Oxford Road bridge over the course of this 36-month period. Whether they were ritually 'un-deposited' by their original depositors to signify the end of the relationship, or by somebody else for a different reason, is impossible to ascertain, but either way it is doubtful that people would wade into the Rochdale Canal in order to find the key. Perhaps bolt-cutters were used or the key was retained rather than dropped into the water.
In other cases, a different (easier) method of 'un-deposition' is employed: erasure of the inscription. This occurred on MLL17, which was recorded as bearing initials and a date on 7 July 2014. By 23 July, silver marker had been used to obscure the initials. Likewise, MLL66 was recorded as bearing two names on 9 February 2015, but by 1 June the names had been covered by scribbled black marker (Figure 8). Without a diachronic study, the original inscriptions on these love-locks would have been lost, and the possible symbolic significance of these silver and black scribbles may not have been recognized. A diachronic study of an accumulation therefore proves essential to an understanding of how an assemblage not only grows over time, but can experience depletion, alteration and ritual recycling.

The purpose of deposition
At first glance, the purpose of deposition appears obvious. Even without the myriad news features declaring that padlocks are being employed as 'love-locks', deposited to affirm romantic commitment, this function is fairly self-evident in the material record. The symbolism of padlocks as objects of security and steadfast unity, as explored by Maiwald (2016), is inherent to their materiality and mundane utility (locking one thing to another). This symbolism is made explicit by the depositors in their inscriptions. The vast majority of the Oxford Road love-locks contain the names or initials of two people, as well as the romance-related symbols of love-hearts, crosses and lemniscates. Many more make overt reference to love and commitment: 'forever and always', 'I love you', '4 EVER', 'FOREVER and a Day', 'Love Forever', 'Always', 'forever & always', 'LOVE YOU ALWAYS', 'Sarah Loves Peter', 'Soul Mates', 'Te Amo' (Italian for 'I love you'), 'NUESTRO AMOR SERA LEYENDA' (Spanish for 'Our love will be legend'), to list only some. One even draws on an analogy with the love-lock itself, inscribed with the words 'LOCKED OUR LOVE'.
It would be easy to conclude, therefore, that all padlocks attached to bridges are declarations of romantic commitment. However, we must be wary of assuming homogeneity. As I have argued previously: Scholars may seek to unearth the meaning of a custom, but when that custom is observed by multiple practitioners, in numbers that can range from 'several' to several million, how can one single motivation be ascribed to every individual? Humans are distinct, emotionally heterogeneous creatures. Granted, physical actions are widely imitated and homogeneous; participation in folk customs tends to be formulaic and ritualized. However, the reasons behind participation and the 'meanings' ascribed to the custom will be as varied as the practitioners themselves. (Houlbrook, 2014: 41, emphasis in original) Viewed en masse, the love-locks do look homogeneous. Despite some variations in size, shape and colour, their physical attributes are largely uniform: square or rectangular bodies with 'U'-shaped shackles. And, as noted above, the nature of their inscriptions varies little. However, some of them do vary, and it is these divergent examples that evince the mutable nature of this custom.
Some inscriptions speak of a clear romantic purpose for deposition, but refer to specific events. The depositors of one love-lock, for example, appear to have participated in the custom to celebrate '1 Year Together ',while '1ST YEAR','1 YEAR','Happy 1st Anniversary','7 years','10 Years','12 years','21 YRS','wedding Anniversary' and 'GOLDEN WEDDING' are inscribed onto others. The custom of love-locks has clearly developed as a form of celebrating romantic milestones; it also appears to have developed as a method of creating them. One love-lock bears the inscription 'will u marry me?', and another 'WILL YOU MARRY ME?' Did their depositors actually employ these objects to propose to their partners, or were they deposited to celebrate an engagement? Either way, they are clearly being used to mark significant moments in relationships.
These relationships, however, are not always romantic. Sometimes it is friendships that appear to be affirmed ('Friends foreveR', 'BFF', 'BBF'), while others seem to celebrate familial bonds ('sisterly sisters', 'Sisters Forever', 'CHLOE & MUM', 'Grandma'). Others appear to honour individual milestones ('HAPPY BIRTHDAY') or accomplishments ('Study Abroad 2K16'), while some are more commemorative than celebratory: three near-identical love-locks, for example, bear the message 'With Deepest Sympathy', each one bearing a different name but undoubtedly deposited by the same individual/group. Another one is inscribed with the words 'I'll never forget', whilst another declares 'It's never goodbye only see you later', memorializing relationships if not individuals. As Petts (2003: 194-195) observes, durable items 'serve to crystallize into physical form the dynamic act of remembrance', and it is unsurprising that lovelocks have developed memorializing functions alongside romantic ones, with the symbolism of the padlock being particularly appropriate to messages of memory.
The padlocks may be durable, but the inscriptions are relatively ephemeral; constant exposure to the elements means that even those professionally inscribed or produced in permanent marker will become faded or obscured over time. Many inscriptions that were recorded with ease in 2014 and 2015 are now difficult, some even impossible, to decipher. It is likely therefore that, given further years of exposure, the majority of the inscriptions will become illegible. And as the variant messages fade from existence, the variant purposes (celebratory of specific events, commemorative, etc.) may fade from knowledge -hence the importance of current cataloguing attempts in the recognition of the mutability of the love-lock.
It is not, however, only the occasional inscription that proves divergent. Some deposits are divergent in themselves: 27 objects have been deposited on the Oxford Road bridge that are not padlocks. These objects, which I term divergent deposits, vary significantly, from nightclub entrance wristbands to disposable lighters; from a half-eaten cupcake to an uncooked rasher of bacon; from pieces of ribbon to a hyperbolic inflatable penis. Some are more clearly personal and functional -if not valuable -than others. A child's dummy (or pacifier) was attached to the railings via a beaded chain that spelled out what was presumably the child's name, while an intact pair of gloves was placed on the wall of the bridge. Whether these were lost (and displayed by finders) or purposefully deposited is impossible to say, but either way they can be classified as structured deposits. They were either displayed by finders in a prominent way to increase the chances of reunion with their owners, or placed in lieu of a padlock as a more personal deposit.
The other objects, however, are more obscure: the McDonald's drinking straw; the rolled-up pages of newspaper; the hair bobble; the paper napkin; the elastic band; the plastic carrier bag; the piece of chewing gum. If these items were encountered on the floor they would be unambiguously classified as 'rubbish'; objects of discard rather than structured deposits. Fontijn (2002: 33) considers this distinction within the archaeological context: 'The difference between discard and deliberate deposition is that they are steered by different motivations. Discard is defined here as a way of getting rid of an object that is no longer considered to be meaningful and useful'. However, this distinction is not always self-evident.
While the nature of these divergent deposits suggests that they were disposable, alienable objects, not unlikely to have been thrown away, the material evidence of their placements suggests deliberate deposition rather than casual discard. These objects are not dropped to the floor or tossed in the general direction of the bridge; they are carefully draped over, balanced on top of, tied around, stuffed between, or stuck to the railings and the love-locks. Why might the depositors of these objects have chosen to place them on the bridge rather than dispose of them in the rubbish bin standing only a few feet away? Perhaps they were deposited by people unwilling or unable to source a padlock (much like the 'recyclers' discussed above), and so they used whatever objects they happened to be carrying. Alternatively, the depositors may have intended to dispose of these objects in a bin or on the floor but, upon seeing the bridge, chose a more unusual style of discard.
Evidently, as Thompson (1979: 10) avers, 'rubbish' and 'valuable' are 'malleable' categories, a concept that has been most usefully explored by archaeologist Brück. Writing of 'odd' deposits in middle Bronze Age settlements, from a smashed bucket to the carcass of a cow, Brück (2007: 296) asks, 'how can we draw the line between refuse and ritual? Many of these items were considered "rubbish" but were disposed of in a ritualistic style.' The same applies to Oxford Road's divergent deposits: items that are considered 'rubbish' but, because of their atypical context, suggest a ritualistic style of discard. Clearly the line between discard and deliberate deposition is a blurred one, the categories of refuse and ritual existing on a scale rather than as polar opposites.
Much like the inscriptions, however, these divergent deposits prove far more ephemeral than the padlocks. With the exception of the chewing gum and a couple of pieces of thread tightly tied to a railing, none of the divergent deposits have lasted beyond a few weeks. The inflatable penis, for example, retained its place perched in front of the love-locks for less than 36 hours. This is unsurprising when we contrast their methods of placement with the secure attachment of a padlock. These objects are either cleared away by street cleaners, have fallen to the floor or into the canal below, are blown away in the wind, or simply taken by passers-by (be they human, rat, or bird). Without a diachronic study of this assemblage, therefore, it is unlikely that the divergent deposits would be recorded.

Other directions and concluding remarks
This article had one primary objective: to consider what insights can be gained into the place, people, pace and purpose of deposition by recording an accumulation of structured deposits during its formation, rather than after. And the answer? Many. Some insights were beyond the remit of this article, not least the fact that -unlike prehistoric and historic accumulations -the depositors are actually available to question. In the case of love-locks, the researcher can engage with the practitioners and can answer the often elusive question of why. This practice is therefore just as much of interest to anthropologists as archaeologists, who are well aware of the significance of material culture in constituting relationships. 'Objects', observes Daniel Miller in his ethnographic study The Comfort of Things (2008: 38), 'store and possess, take in and breathe out the emotions with which they have been associated.' The love-lock -a simple padlock, but one that is imbued with considerable emotion and intrinsically linked with personal relationships and special events -encapsulates this perfectly.
Over the course of the three-year diachronic study, I have exploited every opportunity to interview depositors of love-locks. Sometimes a practitioner is serendipitously present at the time of cataloguing, while at other times contact was made with people who had deposited love-locks previously or elsewhere. These practitioners were happy to answer questions about why they had engaged with this custom, and therefore provided invaluable ethnographic data on the why of this custom. Ethnographic material has also been sourced through the internet. The custom features heavily on the video-sharing website YouTube, with literally hundreds of videos of people depositing their love-locks uploaded, often as part of travel vlogs (blog posts in video form). 1 According to these interviews and videos, most often the practitioners had deposited love-locks with their partners as statements of romantic commitment, often while on holiday (the deposit becomes an inverted souvenir) or while attending special events, such as shows at the nearby Palace Theatre. These interviews revealed that deposition was occasionally timed to coincide with an anniversary, engagement, or birthday. However, other motivations were made apparent in these interviews; for example, one elderly couple in Bakewell attached their love-lock to a bridge to celebrate a recent lottery win.
Some of the ethnographic material elucidates aspects that are not clear from the material evidence, which reveals a weakness in archaeological inference. One example is the great significance explicit in many of the uploaded videos attached to the throwing of the key into the water, which seems to equal -if not surpass -the ritual importance of attaching the love-lock itself. There are certainly numerous lines of enquiry that the anthropologist could pursue, not least how archaeological methodologies could provide insight into contemporary practices in material culture, with the attention paid to the deposits themselves providing a counterbalance to ethnographic methods. I have plans to publish on this subject in the future.
However, the main aim of this article was to consider these contemporary accumulations as archaeological evidence and to consider what knowledge an archaeologist (who could not engage with the practitioners) might gain from recording structured deposits as they are deposited, rather than millennia later. And as has hopefully been demonstrated, a diachronic approach to love-locks has enabled a much richer understanding of the custom and has highlighted the value of tracing the development of an accumulation, rather than simply viewing it in temporal isolation. This diachronic form of documentation has revealed much about how we might approach and possibly renegotiate interpretations of ritual deposits. Archaeologists may take from this the lesson that accumulations should not be studied at one static point in time, and this will prompt us to question assumptions about the place, people, pace and purpose of historic and prehistoric accumulations. And yet three years and 409 love-locks later, I still believe this custom has more to teach us.