Between the village and the global city: the production and decay of translocal spaces of Thai migrant workers in Singapore

This paper explores the mobilities and structural moorings of Thai labour migrants in Singapore from a translocal perspective. We argue that combining the mobilities paradigm with the concept of translocality offers a fruitful avenue of investigation not only of the production of translocal spaces, but also of their temporality and mutability. Through a multi-sited research approach we shed light on the genesis as well as the decay of translocal connections. This paper shows that translocal structures are important moorings of migration, and raises the question of what happens to translocal spaces when migration flows dissolve.


Introduction
Since the beginning of the debate on transnationalism and translocality (Glick Schiller, Basch, and Blanc-Szanton 1992) an increasing number of publications illuminate the emergence of transnational or translocal spaces and phenomena, such as 'transnational villagers' who remain simultaneously rooted in two worlds (Levitt 2001), or translocal rural-urban livelihoods (Steinbrink 2009). However, little attention has been paid to the temporal dimension of translocality. How do translocal spaces change over time, and how are they maintained? Are they persistent phenomena simply because global interrelations are intensifying? Although this topic has recently been attracting interest , there is little empirical research about the temporality, persistence, or transience of translocal spaces. While the nexus between migration and translocal spaces is acknowledged to be non-linear and complex (Freitag and von Oppen 2010;Brickell and Datta 2011), the translocal approach often seems to overemphasize migration movements as the main factor for the emergence and persistence of translocal spaces, neglecting everyday mobilities. In this paper we combine a translocal perspective with the mobilities approach, allowing us to analyse the relations between (long-distance) migration, quotidian mobilities, and the dynamics of structures and moorings.
Singapore -an important financial and trading hub as well as a significant immigrant gateway (Benton-Short and Price 2008) and 'mobile city' (Oswin and Yeoh 2010, 168) -was one of the main destinations of Thai labour migrants from the 1980s to the early 2010s. Today, the young generation from Thailand turns to other countries such as South Korea, Taiwan, or Israel. In the past five years, the number of Thai workers in Singapore has visibly decreased, and Thai labour migration could even come to an end within the coming years. In this light Singapore is an interesting case as it has a long migration ARTICLE HISTORY history with dynamic shifts of migration patterns, allowing us to study the temporality and evanescence of translocal spaces. To address this we need to examine migration decisions and translocal practices, as well as related structures that have evolved over time (Bakewell 2010;Coulter, Ham, and Findlay 2016).
The combination of the new mobility paradigm (Hannam, Sheller, and Urry 2006;Cresswell 2010) with the concept of translocality (Brickell and Datta 2011;Greiner and Sakdapolrak 2013) provides a powerful framework for understanding how migration, translocal connections, and social embeddedness are interlinked. The inspiring debate of the 'mobility turn' has not only shown that migration is an essential part of the lives of people who have become more mobile in the age of globalization, but also that ' "moorings" are often as important as "mobilities"' (Cresswell 2010, 18; see also Adey 2006;Hannam, Sheller, and Urry 2006). This paper shows that such moorings often find their expression in translocal structures, and explores how movements and everyday practices produce, maintain or corrode such structures.
This paper is based on seven months of empirical field research, conducted from February to September 2015 by the first author. The research was conducted as a multi-sited ethnography (Marcus 1995;Gille 2001;Falzon 2009). Fifty-one semi-structured interviews, nineteen informal talks, and twelve participatory methods (PRA) and group talks were conducted, as well as participatory observations. To understand the transnational migration system between Thailand and Singapore, the first author followed migration trajectories through both space and time. The latter was enabled by a gift from a returning migrant, who had captured his life as a migrant worker in Singapore with his camera in the early 1990s. He provided 930 photos for digitalization and analysis. These photos offer rare insights into the facets of a working man's life. The photos were juxtaposed with the views and perspectives gathered during the field research 20 years later.

Translocal spaces
The concept of translocality has evolved from the debate on transnationalism, and the differences and similarities between both approaches have been discussed by other authors (see Brickell and Datta 2011;Greiner and Sakdapolrak 2013). Today, we can look back on almost three decades of debate, evolving from critical reflections on the production of space, place, and locality (Soja 1989;Massey 1992;Appadurai 1996) to a new perspective on migration processes and local embeddedness (Levitt and Glick Schiller 2004;Steinbrink 2009;Verne 2011). Transnationalism has been criticized for its overemphasis on the nation-state from which everything is increasingly detached (see Vertovec 2009) and where Transmigrants live in the fluidity of globalization. Following this critique (Glick Schiller and Çağlar 2009) we argue that (trans)migrants -though transgressing borders -are nevertheless grounded in localities (cf. Mitchell 1997;Verne 2011). In this vein translocality is a more neutral notion, as it does not refer to one specific entity -the nation-state -that is transgressed, but remains open to application at varying scales (Freitag and von Oppen 2010).
In this paper we use the terms place and locality synonymously (for a more fine-grained differentiation see Massey [1993]). Following Soja (1989), and Massey (1992), in our understanding a locality is a socio-spatial and relational entity that is culturally contextualized, and emerges when subjects give meaning to the physical space. More precisely, a locality is a portion of the physical space that is imagined, defined, and lived by different actors, and which is contested over time. Through functional, imagined or symbolic connections a locality expands beyond its mere socio-spatial dimensions (Callon and Law 2004). The boundaries of most places are multiple, blurred, and contested, depending on the perspective and domain -administratively for example, places are mostly well-defined; socially, however, boundaries are much less clear. Migration is further challenging the conceptualization of discrete places and scales (Porst and Sakdapolrak 2017), as the production of localities is increasingly shaped by migration and related processes.
The emergence of migrant places is often described in the context of migration flows or events. From a translocal perspective, we argue that a wider focus on the underlying processes is needed to understand the emergence of translocal places, as the actual migration journey is just the beginning of ongoing processes that produce close interrelations between different places and people. Concomitantly, we plead for a broader understanding of migration processes, acknowledging the wide range of migration types (long-term, temporal, circular etc.). Furthermore, in the face of globalization, the uprooting nature of migration has decreased (Sterly 2015), and references to the place of origin and the sense of belonging are important aspects of migrants' everyday lives -aspects that are often neglected (Collins 2016(Collins , 1170. Thus we need to understand these everyday interactions of migrants (calling home, sending money etc.) as they help us to understand their social embeddedness and follow-up migration decisions, such as return or chain migration. In this light we understand migration as an extended process which starts with the first idea to migrate. It continues with an information phase, when people discuss their plans with family members or friends. At some point a decision is made whether, where, and for how long to move. This is followed by concrete preparations before the actual journey. The latter is often a brief moment in the migration process, followed by a phase of adjustment to a new daily life at the destination. It is difficult to delineate when migration processes end, because even if migrants stay at their destination for long time they may never completely 'settle' , living in a state of what Collins describes as 'permanent temporariness' (2011). The latter describes migrants' experience of being only temporarily accepted and the reflexive knowledge that such temporariness is permanent or lasting (Collins 2011, 322). This certainly influences migrants' agency and also their translocal social embeddedness, as this temporariness can be a strong reason for migrants to maintain close relations to their places of origin.
A translocal perspective with a focus on connections, relations, and everyday practices widens the view on locality, place, and space (Verne 2011;Pordié 2013, 9). Through the interplay with other localities and the simultaneity of spatial and social relations, a locality becomes part of a wider translocal setting. Translocal connections are actively produced through everyday practices by both migrants and non-migrants. Thus, it is also necessary to look at the quotidian practices and mobilities of migrants, in addition to the migration movement itself. Subjects and their practices are always embedded in a wider societal arena, and through mobility and translocal connections they can be simultaneously embedded in different (local) social fields. Following Bourdieu (1985) we understand social space as constituted by multiple social fields. Based on this we define translocal spaces as composed of multi-layered social fields that are anchored at different localities and constituted by the networks of actors and their embedding in societal structures. In this sense we do not understand translocal spaces as discrete social segments of space, but rather as a 'heterogeneous criss-cross of diverse and distinct relational and material spaces' (Bork-Hüffer 2016, 124).

Mobilities and translocal moorings
The nexus between migration and translocal spaces often remains blurred, and the translocal approach tends to overemphasize migration as the main factor for the emergence of translocal spaces. The mobility approach helps in two ways to readjust the framing of this nexus. Firstly, the mobilities perspective views migration within the wider context of long-distance movements and everyday mobilities of both migrants and non-migrants (Hannam, Sheller, and Urry 2006). Secondly, it acknowledges this wider range of mobilities without neglecting the importance of structures and moorings.
Mobility, as Cresswell emphasizes, involves a fragile entanglement of physical movement, representations, and practices (2010,17). The physical movements represent the 'raw material for the production of mobility' (19) and they similarly represent the basis for the production of translocal ties. However, movements do not happen without representations, which are the meanings and values that subjects give to movements. Mobility is also practiced in everyday life, including 'both the everyday sense of particular practices such as walking or driving and also the more theoretical sense of the social as it is embodied and habitualised' (20). As Cresswell highlights, these three elements of mobility are not easily disentangled, as they are closely interrelated (2010,19). From a translocal perspective this is even amplified, as migration processes imply that these three -movements, representations, and practices -can take place and effect at different localities simultaneously. Cresswell's (2010) central point is that mobility inherently bears questions of power and politics, which find their expression in at least six aspects: (a) reasons for movement; (b) routes; (c) velocity of movements; (d) their rhythms; (e) mobile people's experiences; and (f ) the frictions they encounter. These aspects are only partially subject to free choice -for example people may move freely or they might be forced (reasons), they follow direct itineraries, or travel hidden as stowaways (routes), they may move on express ways 'tunnelling' peripheries, (Graham and Marvin 2001) or use slower but cheaper means (velocity). The rhythms of mobilities are often given by timetables or, in the case of migration, by immigration policies, seasons, and work contracts. These mobilities and movements are also practiced and experienced in everyday life (Cresswell and Merriman 2011). Such experiences are on the one hand shaped by immigration and labour politics, but on the other hand quotidian practices and mobilities are contested and subject to negotiation within the social field of what Collins (2016) frames as the 'politics of migrant lives' . Following this understanding, it is not sufficient to look at national-level immigration politics only, but also at the quotidian -sometimes tactical -practices of migrants in the host society, or what Papadopoulos and Tsianos (2013, 190) call the 'mobile commons' . These practices and mobile commons are embedded in wider societal structures. We distinguish between three different levels of (trans)local structures that influence migrants' lives. First, there is the general structural background of globalization, constituted by the economic and political realities. Second, there are internal structures that are closely linked to the emergence of translocal spaces and become part of them. Finally, there are embodied structures that influence people's behaviour and translocal practices.
The general structural background comprises external structures and developments that influence the translocal setting on a collateral level without having a direct reciprocal relation with translocal spaces. Such structural backgrounds may include the economic or political situation, the cultural and societal setting, or technological contrivances and infrastructures like road grids, airports, internet cables etc. Smartphones have doubtlessly changed the way migrants communicate with their families back home, and they promote the intensity and frequency of translocal ties (cf. Brown, Green, and Harper 2002;Sterly 2015).
The transition between the general structural background and internal structures is sometimes blurred. We distinguish internal structures as having an immediate influence on migrants' practices and mobilities. In this regard they become an integral part of the translocal space. Immigration policies are an example, as they produce and alter migration systems. Other forms of internal structures materialize as specific infrastructures such as dormitories or training centres (Hannam, Sheller, and Urry 2006, 14). Often such material structures are local structures that are adopted by migrants and become part of translocal settings. At these meeting points new social networks evolve, being internal translocal structures of their own that also shape migration flows and directions of movements.
'Embodied structures', or what Bourdieu would call habitus, are formed through past experiences with societal rules and socialization (Bourdieu 1977(Bourdieu , 1985, and condition an individual's practices and behaviour. As dispositions and habitus are often similar for members of a given social group, they also shape collective practices within social fields and networks and their institutions.

The production and temporality of translocal spaces
The emergence and maintenance of translocal spaces is based on connections created through the mobility and (translocal) practices of migrants. Their engagement with other places usually starts before the migration journeys. Migrants gather information e.g. about salary, working and living conditions at the potential destination, and they imagine their future abroad. They ask return migrants about their experiences, from which again they develop their own images of migration (representations). This very first moment of developing geographical imaginations about other places (Thompson 2017) is the nucleus of a translocal space.
Translocal connections then emerge during and after migration, when migrants send remittances -including social remittances (Levitt and Lamba-Nieves 2011) -or call their families to exchange information and share emotions that maintain their sense of belonging and embeddedness (cf. Conradson and Mckay 2007). Being spatially separated demands regular efforts to nurture emotional bonds, and to reassert mutual belonging. Still, in the 'mobile age' (Brown, Green, and Harper 2002) phone or video calls are not sufficient, and visits back home are essential to maintain translocal relations. Translocal connections also have economic functions, as migration is often a strategy for income diversification (Steinbrink 2009). This leads not only to new opportunities but also to dependencies, as the situation or decisions made at one place affect those at the other place(s) and vice versa. However, translocal connections alone do not yet form translocal spaces.
Translocal spaces evolve when these connections and the social embeddedness expand beyond the individual family or household level, and become institutionalized. Migration leads to a spatial expansion of networks and social fields. The embeddedness of migrants within host societies facilitates the formation of translocal spaces, even though this might be less determinate than is often suggested (Levitt 2001). Empirical research shows that social embeddedness varies with different types of migration and the socio-political situation at the destinations. In many cases migrants also find their embeddedness within their own translocal group or diaspora (Peth 2015). The persistence and rhythms of mobility and movements are crucial factors for the embedding of migrants in translocal fields, and thus for maintaining translocal spaces. Only continuous connections will lead to a routinization and institutionalization of translocal practices. The latter are not limited to practices that connect the migrants' destinations with their places of origin, but also include migrants' quotidian practices (praying, cooking, talking and bonding with other migrants, etc.) that refer to their place of origin, or that maintain their sense of belonging and thus enable -to varying degrees -their simultaneous embeddedness in 'multiple sites and layers of the transnational social fields in which they live' (Levitt and Jaworsky 2007, 130).
In the face of globalization, scholars have highlighted an increasing pace of movements that influence subjects, communities, and public and private spaces (Hannam, Sheller, and Urry 2006, 2). Rhythms, as 'repeated moments of movements and rest' (Cresswell 2010, 23), gain momentum within the dynamics of globalization. Rhythms of mobility are simultaneously practiced, intuitively lived, endogenously (organically) and exteriorly (mechanically) shaped, and imposed by technologies and the configurations of globalization (Lefebvre 2004, 6). In this regard the temporality of rhythms seems to intensify their impact on translocal spaces. However, temporality can also comprise processes of change that might lead to a loosening or decay of mobility and connectivity. The direction and degree of change depend on the structural context described above as well as the subjective representations and practices of migrants and non-migrants. This can be specified by looking at the different structures introduced above: Embodied structures are relatively stable, as they are habitualized, and because subjects usually act unconsciously regarding their habitus. Changes of habitus do gradually occur over lifetime and through new experiences that subjects encounter, as well as across generations through social reproduction (Bourdieu 1985, 739). Relative to the time span of migration processes and translocal spaces, however, embodied structures are still relatively stable. Material structures of places are often relatively persistent as well, as their material characteristics can only be changed through substantive inputs of economic capital and time. Translocal material structures (internal structures) are often -by their initial purpose -local structures that have been adopted by migrants and integrated into their practices, and have thus become material anchors of translocal spaces. As such they can be abandoned, thus losing their anchoring nature and translocal representation. Material (trans)local structures tend to evolve with high numbers of migrants; however, this does not mean that declining migration numbers inevitably lead to an instantaneous decline of such structures. As noted above, social institutions and policies constitute other forms of internal translocal structures, which are bound to socio-cultural spaces such as nation-states. Such structures are negotiated from the (inter-)national to the local level. Rules and regulations are not only set by governments or administrations; they are also negotiated in societies, by private actors such as migrants or employers. These policies and social institutions are subject to change, although there is a certain degree of inertia, as social and political negotiations take time.
Finally, practices, representations, and the various structures influence one another. If for example migration policies drastically reduce migration flows, this also influences the rhythms of translocal mobilities and exchanges; networks will alter, and the need for material (infra-)structures will decline. This can result in the disappearance or shift of whole translocal spaces. On the other hand, translocal networks and structures can develop such a strong coherence that they outlast declining migration flows. While (long-distance and long-term) migration is necessary for the emergence of translocal spaces and moorings, it is rather the quotidian mobility and the rhythms of everyday practices that influence their persistence and decay. These everyday mobilities and practices are conditioned by the internal structures (governing migration, mobility, labour, housing, etc.), but also by the embodied structures, the habitus of migrants (preferences for food, language, rituals, experience of being a migrant worker, etc.). To sum up, there are different mechanisms that influence the genesis, persistence and decay of translocal spaces, and moorings: (a) migration (and the related general structural background) is the key underlying reason for creating long-distance translocal spaces (networked connections, flows of resources, knowledge, imaginaries, etc.); (b) quotidian mobility, which is negotiated within the framework of mobile commons (status of migrants, labour and living conditions etc.) -the rhythms, routes and practices of migrants -are important for shaping the moorings, the material places with translocal meanings attached to them. This nexus between mobility, moorings, and the temporality of translocal spaces can best be disentangled by looking at specific cases.

Translocal spaces of Thai migrant workers in Singapore
The political and economic processes in Thailand and Singapore fostered the development of a ramified migration system between the two countries. In the early 1970s Thailand was marked by an increasing income gap and peasant revolts, resulting in an unstable political period. During that time, young men from rural Thailand left the country to seek employment abroad, e.g. in the Middle East. However, in the aftermath of the oil crisis and the Gulf War in 1990/91 and the associated economic recession, migration flows shifted towards Newly Industrialized Countries (NICs) in East and Southeast Asia (Sakdapolrak 2008, 82;Yue 2008, 99). The booming NICs such as Brunei, Malaysia, Japan, Taiwan, and Singapore lacked domestic labour and increasingly demanded foreign workers (Promburom and Sakdapolrak 2012).
Singapore has been attracting immigrants from different Asian countries since its foundation in 1819 (Kitiarsa 2014, 11). Singapore is a crossroads of global flows, and its tremendous development from a small colonial trading post to a global city (Bercuson 1995) was only possible with the contributions of millions of migrant workers (Warren 2003;Yeoh 2004). However, in the beginning Singapore only allowed citizens from so-called 'traditional sources' like China, Malaysia, and Indonesia to work there. In 1968 Singapore opened its labour market, also allowing workers from 'non-traditional sources' such as India, Philippines, Bangladesh, and Thailand to enter (cf. Kitiarsa 2014, 11). Since the 1980s, which were marked by a growing number of undocumented migrant workers in Singapore, there was a rethinking and systematization of immigration policies. Today, Singapore has one of the most elaborated migration regimes in the world and is able to steer immigration flows according to the needs of different sectors. This migration regime includes thorough rules and regulations for the foreign workers and draconian punishments in case of defiance.
This structural background, as well as the changes in migration flows, form the base of translocal spaces of Thai migrant workers in Singapore. In the following we discuss the migration system and the changing translocal spaces, by referring to the six aspects of mobility introduced by Cresswell (2010).

Movements and reasons
Movements between Thailand and Singapore are dominated by the mobilities of labour migrants. Reportedly a group of five hundred Thai contract workers first arrived in Singapore in 1978 to work for electronic and textile companies in Jurong (Kitiarsa 2014, 12). These pioneers established initial relations with their employers and informed friends and relatives about job opportunities, and the number of Thai workers in Singapore rapidly increased (Kitiarsa 2006, 7-8). From the early 1990s to the early 2010s in particular the construction boom in Singapore increased the need for labour, and Thais comprised one of the largest groups of foreign workers in Singapore (see Figure 1).
During our interviews we found that many households in rural Thailand face structural problems such as unsteady income from farming and the lack of alternative income sources, which are the main reasons for migrating. The unsteady income must still meet monthly expenses for household needs, especially educational costs. After finishing school, the young generation are often left with few opportunities in the village, and it is common that even teenagers embark on a longer migration trajectory. This usually begins with internal migration to urban places, often followed by disappointment with income, and the plan to migrate overseas. Seksan represents a typical case: he was 14 and finished school at grade six; 'well, after that I hadn't done any work yet. I was a young man just walking around in the village' . He tried to work in Bangkok and travelled there three times, without finding a proper job: 'Staying here didn't make anything better; no work, no money' , so he agreed immediately to go to Singapore when his uncle asked him to go there.
In other cases labour migration is a way to cope with sudden financial problems due to crop failures, environmental hazards, or price fluctuations. In such cases migration means great responsibility and a burden for the migrants, who express that they feel social pressure. Finally, there are also more individual motives, in particular of young men who see international labour migration as an exciting adventure and as opportunity to 'dig for gold' . One of them is Nut I went to Singapore because I just wanted to go there. I got a strong body so I could go there and it is also easy to enter and to leave Singapore. Thai workers are the number one [best workers]. (Nut, return migrant from Singapore) The cases of Seksan and Nut are typical, and they also illustrate the political dimension of mobility. Thai migrant workers do not belong to the 'kinetic elite' as Cresswell (2010, 22) calls it, because they cannot move freely. Their mobility greatly depends on the work they get offered through their social networks, and without these specific jobs the economic risks of migration would be too high. On the other hand, however, they do not represent a kinetic subaltern either. Thai labour migrants move within the powerful structures of migration systems (cf. Yea 2017), but they also navigate in the fields of global labour markets and negotiate some degree of freedom. Brasid, a 51 year old migrant worker explains important differences between the migration systems and policies. 'In Taiwan you could only work for four years but in Singapore you could work until you are 60 years old. I knew that so that is why I went to Singapore' .
However, not all migration movements were made in such a reflexive manner. Often migrants just follow the crowd. Praew remembers the time when her husband decided to work overseas. Singapore 'was quite popular back then, good salary (…) and he said he would be fine; the city is clean, and no hard work' . Particularly in Northeast Thailand it became common that young men -and to lesser extent also women -went abroad, and labour migration became a social practice as sons followed their fathers and uncles. Though the work opportunities in Singapore were common knowledge and everyone could potentially go there, labour migration still depended on network structures. They provide the right information at the right time, and they help to reduce the migration costs, like in the case of Seksan: His uncle who was already working in Singapore told him when his company was hiring, and he organized the work permit and all necessary paperwork for him. Thus Seksan did not need to involve any agent from Thailand and could save the expensive commission fees of 2000-2500 USD.
The development of the Thai-Singapore migration system increased not only the movements of people, but also the flows of goods and commodities, as the demand for Thai specialities such as fermented fish (bla ra), grilled bullfrogs (yang kob), and chili paste has risen. There are several pickup trucks running daily between the trading city Hat Yai in southern Thailand and Singapore (see Figure 2). They provide shops and restaurants all over the city state with commodities from Thailand, and special orders can reach Singapore the next day.
Besides the trucks, numerous women traders arrive by bus every day to sell Thai products to the Thai workers. The workers often organize in groups to collectively order food ingredients which cannot be found in Singapore. The women commute regularly between Thailand and Singapore to deliver their orders directly to workers' dormitories. These flows of food and goods to Singapore are elements of translocal business relations. In Thailand numerous villagers or relatives are involved in producing and processing the demanded goods, while the traders bring them to the border city of Hat Yai, which serves as intermediate store. From there the goods are brought in smaller packages by bus to Singapore.

Routes, velocity, rhythms and friction
Thai migrant workers in Singapore originate mostly from North and Northeast Thailand, especially from Isan, where over 65% of the workers surveyed by the Royal Thai Embassy in 2015 came from (see Figure  2). The migration routes usually lead first to provincial cities or to Bangkok, where the workers have to pass basic skill training, health checks, and visa applications (preparation movements). The official way to work overseas leads through the Office of Labour Affairs, and the preparation phase usually takes three months. Their route then often leads back to the village where they wait for the final confirmation of the employer from Singapore. This phase of waiting mostly takes a couple of weeks, in few cases some months. Here the politics of mobility play a role again, as it is possible to speed up this process for those families who can raise enough money to involve a migration agent. Arisara is such an agent, and explains that she can speed up the process to one month if 80,000 Baht are paid. Although this might be faster and more convenient for those who can afford it, however, it also bears the risk of being cheated by untrustworthy agents.
The travel route to Singapore always leads through Bangkok, from where two routes reach to Singapore (see Figure 2). The older route, which still exists today, leads overland by train to the border city of Hat Yai and then by direct bus to Singapore via Malaysia. The overland route was highly frequented in the 1990s as it was the cheapest way. Nowadays this route is mostly used by the trading women -who need to carry more luggage -or younger Thai women working in Singapore's night life.
Most migrant workers nowadays take the much faster air route to Singapore, which became economically interesting with the rise of low-cost airlines shortening the journey from two days to 2.5 h. Today there are on average two direct air connections every hour between Bangkok and Singapore, showing the strong interconnectedness of the countries. One could argue that this increased velocity is just a result of globalization and the booming market for low-cost airlines; however, it is also linked to labour migration, which led to an increased demand for faster and more diverse routes.
For Singapore's booming construction sector, Thailand has become a pool of cheap and trained labour, and also of goods and resources consumed by the migrant workers. The rhythms of movements differ however. Traded goods and information for example move freely and with high frequency. Money can be transferred within hours, and the introduction of smartphones has enabled real-time communication between migrants, family members, agents and employers. The movements of migrant workers, however, are less free, and stipulated by contract. Visits back home are not a matter of free choice. The standard work contract in Singapore predetermines one visiting period at the end of each contract period -mostly annually or biennially. As the employer has to cover the flight and continues to pay taxes for the worker during that time, the visits are subject to power relations, and nobody wants to stay too long in the villages. 'It is better to go back early' explains Brasid' who only visits home for two weeks, although he could stay longer. The immigration and working policies also prohibit family members from visiting Singapore.
Such frictions also affect the everyday mobility of migrant workers in Singapore. No family members or friends are allowed to visit them or enter the dormitories. The latter are high-tech camps that are under 24/7 surveillance, and access is only granted with fingerprint or iris scans. Also the daily mobility within the city is constrained, as foreign workers do not move freely to their worksites. Every morning they are loaded onto lorries and transported to the worksites, and are picked up in the evening. The rhythm of daily mobilities is hence conditioned by the 'politics of migrants lives' (Collins 2016). In a similar vein the rhythm of the work itself is regulated. As mentioned above, the standard contracts in Singapore run for one to two years. Immigration laws require the workers to leave the country immediately when their contracts expire. As most construction projects run longer and companies usually run several projects at the same time, they extend the contracts yearly, as long as needed. Most of the interviewed workers were in such short-term contracts for more than 10 or even 20 years. This is a strategy to 'flexibilize' the workers, who are put in a state of 'permanent temporariness' (Collins 2016;Yea 2017). Thus, and in contrast to highly skilled internationals, foreign 'unskilled workers' live permanently at the edge of the Singaporean host society (cf. Yeoh 2006).

Experiences, practices, and representations
How do these rhythms, frictions, and structural marginalization influence the emergence and maintenance of translocal spaces? The transnationalism and translocality debate has emphasized that translocal spaces emerge through the simultaneous embeddedness of people at home and in the host society. Thai migrant workers in Singapore, however, are structurally segregated and controlled by the system. The initial excitement is often met with this harsh reality of drudgery and strict working and living conditions: 'people who go to Singapore have been trained like soldiers, if it is not the time you don't get to eat and sometimes I couldn't sleep at night' says Brasid.
This social non-embeddedness in the host society combined with the insight that 'precariously positioned workers are not often part of a solidaristic community of labour' (Buckley, McPhee, and Rogaly 2017) is compensated with a strong sense of belonging in the place of origin and within the translocal group of Thai labour migrants: The whole life as a migrant worker -although spent abroad -is focused on the village, and the family back home. Thus translocal practices that keep the connections between the village in Thailand and Singapore are crucial, of which financial remittances are the most important: Sending money back home is the monthly 'ritual' of payday and the first thing workers do after they receive their salary, followed by a call to inform their wives that the money has been sent (see Figure 3). For the women, remittances are not only important for daily expenses and for the education of the children; they are also a confirmation of their relationship. Tanawat, a 50-year-old worker from Udonthani, has worked in Singapore for 18 years, and he does not deceive himself. He saw many neighbours going abroad struggling with their relationships, and states: you need to keep the momentum and you should not break the circle. The most important thing is, you have to send the money exactly at the same time every month. And also you need to seek contact actively. (…) I call (my wife) three times a day and (…) always try to give her the feeling that I am not running away.
Besides the family it is also important to show one's belonging to the village, explains Brasid. In Singapore he regularly organizes so-called Pa Bas, Buddhist donation circles, intended for temples or in some cases for village projects, such as new school buildings, or to subsidize poorer children's education. His wife in Thailand organizes the distribution of the money in the name of the respective workers. These Pa Bas take place quite regularly and some migrant workers in Singapore donate every month, also joining Pa Bas from other villages to gain greater merit, while others donate a few times a year.
While experiencing segregated life as labour migrants in Singapore, Thai workers often portrayed themselves as successful men. This portrayal takes place within their translocal group in Singapore as well as in the villages. Photographs taken by a Thai foreman who worked in Singapore in the early 1990s provide a unique insight into the (self-)representation of Thai migrants as work heroes. In many photographs they are staged with items symbolizing their working men's life, masculinity, or economic success, such as cars, sunglasses, walkie-talkies, belt buckles, and wrist watches, or they pose casually on high scaffolding on the tops of skyscrapers (see Figures 1 and 4).
In the villages too, migration of household members is represented as a success story, which frequently leads to investment of remittances in symbolic capital such as new houses or pickup trucks, rather than in productive resources such as land or businesses. Sometimes this symbolic and economic capital from Singapore is used for political action, as in the case of Somsak: he was able to influence local politics while he was working in Singapore, and later became part of the local government after returning to his village. In other cases migrants also managed to take an active political role in their village after returning home, or even while in Singapore. An interesting example is Arthit, who worked in a shipyard in Singapore for 18 years. During that time he managed to become a member of the village committee in Udonthani, despite the geographical distance. If there is any issue, the villagers call him by mobile phone and, in urgent cases, Arthit flies back to his village. This translocal political involvement is a matter of negotiation and shows that Thai migrant workers are not completely heteronomous subjects. Arthit has a good standing with his employer, who supports him by accepting his short trips to the village, though, in most cases Arthit fulfills his responsibilities by phone.
Besides these more individual or family-based practices, Thai migrants also engage in collective quotidian practices that substantiate their belonging to the group of Thai migrant workers. During the scarce free time they meet to prepare Thai food and drink Thai whisky or beer. Everybody contributes by bringing beer or food, sometimes even 'hunted' items, as practiced in rural Thailand. The anthropologist Pattana Kitiarsa (2014, 68) describes how this -often illegal -'hunting and gathering' (e.g. fishing) magnifies the expression of belonging to the group, by providing fresh food the rural way. These gatherings are often accompanied by karaoke with local songs such as Molam, the Thai country music with lyrics that express the feelings and aspirations of the migrants. This sharing of food and alcohol, as well as the exchange of information and emotions, stimulates social cohesion and leads to the creation and maintenance of the translocal group (see Figure 5). It is at the same time an expression of the mobile commons of Thai workers in Singapore. Drinking alcohol or gathering in the public space is strictly regulated in Singapore, and after 10:30 pm it is prohibited to drink alcohol in many places. Thai migrants managed to negotiate a space on the 3rd floor of the Golden Mile, a shopping mall, where they can still meet for drinks until late. This shows that migration and mobility does not take place detached from local moorings or social structures (cf. Yea 2017).

Translocal structures and moorings of Thai workers
The tight and encompassing migration regime of Singapore represents an internal structure shaping migration and translocal spaces. It involves, among other things, a sectoral work-permit system, levy rates for employers, company quotas for foreign workers, the obligation of foreign workers to register with the Ministry of Manpower (MoM), and regular controls by the Singaporean authorities (Huang and Yeoh 2003;Kitiarsa 2014). Employers are obliged to ensure acceptable housing for the foreign workers, resulting in the materialization of the most iconic structures of the labour migration system: the dormitories.
Previously, the dormitories consisted of containers near the work sites, and up to 24 men slept in one container (see Figure 6). Nowadays the dormitories are run by external contractors, who have constructed high-tech dormitories, some for more than 16,000 foreign workers. These camps are like independent urban fragments, with their own infrastructure to accommodate thousands of men from very different cultural backgrounds. There are kitchens, gyms, grocery shops, basic health services, internet and TV rooms, laundries, and sports fields for football, cricket, and sepak takraw (a popular kick-volleyball game in Thailand and Southeast Asia). Tuas View Dormitory, the largest in Singapore, even has a cinema showing films in Hindi. The dormitories are constructed in such a way that there is little reason to leave, except for work, and no outsiders are allowed to enter.
The dormitories are particular cases of translocal structures. On the one hand they are intended to segregate and to prevent social embeddedness. On the other hand they represent a condensation of translocal space where thousands of migrant workers are concentrated. These workers form new networks that become part of a translocal space that is not only connected to their places of origin but also to other destinations of labour migration such as Taiwan, Korea, and Brunei, as the workers also exchange information about alternative destinations and, possibly, better working conditions. These dormitories are translocal structures as they function as 'immigrant gateways' which should not be seen as 'endpoints but turnstiles' of migration movements (Benton-Short and Price 2008, 6).
Another important translocal structure, the Golden Mile, is an old shopping, office and housing complex built in the 1970s which never became part of the city's main centre, as was initially planned. Thai migrant workers adopted the place, and today it is their primary meeting place in Singapore. At weekends, and especially after paydays, the Golden Mile is full of Thai workers who come to send their money back home. A number of remittance services are located here, offering competitive conditions for customers. The Thai supermarket has an assortment ranging from typical Thai groceries to workmen's needs. Numerous small shops are specialized in other goods such as mobile phones, prepaid cards, or Thai massage. There are barbers, a music store with Thai country songs, and, most importantly, small shops selling Thai local whiskey and beer. A clinic with Thai-speaking personnel is located on the third floor, as are the offices of job placement agents, and NGOs such as the Friends of Thailand Association.
Clerical organizations are also present at the Golden Mile, offering worship services, as well as English courses and evening classes in collaboration with the Sukhothai Thammathirat Open University of Thailand. It is thereby even possible for Thais to finish their primary and secondary school education in Singapore. Additionally, the importance of the Golden Mile as a translocal place also becomes obvious in that fact that the Thai Office of Labour Affairs opened a branch office there, consulting Thai workers about administrative and official issues. The main language is Thai, as is evident from the signboards to the Singaporean businessmen, who have learned to speak Thai over time. The Golden Mile is not just a simple shopping mall; it is a place that is completely tailored for Thai migrant workers. In short, Little Thailand, as the locals call it, is a paragon of a translocal place in the global city of Singapore.
Such translocal structures and moorings developed in the context of intensive migration movements, and are shaped by the local and quotidian mobilities, and rhythms of social practices. However, the whole migration system is always subject to transformation, and this in turn has implications for translocal structures.

The shift and decay
The production of translocal spaces is not a linear process: The temporality and the rhythms of the movements, representations, and practices change at different levels, and this also affects the translocal structures that they produce. Firstly, the movements and connections accelerate: People as well as money, goods, and information have become more mobile, and the frequency of movements and connections has increased due to technological change (e.g. smartphones, aeroplanes). However, this does not mean that Thai workers can travel home more frequently, as their work contracts do not allow more visits. Thus, whereas goods, money, and information have become more mobile, labour migrants themselves have not.
Second, in recent years the number of Thai workers in Singapore has dropped significantly. From the beginning of the 1990s to the first decade of the new millennium Thai workers comprised one of the largest groups of foreign workers in Singapore. According to the Office of Labour Affairs at the Royal Thai Embassy in Singapore, the number of Thai migrant workers in Singapore dropped from an estimated 36,000 to 20,000 between 2014 and 2016. This downward trend, which started in 2012, has several causes: In the construction sector companies increasingly employ Indian and Burmese workers, who accept lower wages than Thais. The latter increasingly seek other destinations, as Singapore becomes economically less attractive. Another issue is the regular skill tests for the extension of work permits, which must be passed in English, and which most Thais see as an obstacle. In addition, the younger generation have higher educational levels and can find jobs in more diverse sectors in countries like Korea or Taiwan. However, the decline of Thai migrants has not yet led to the loss of translocal moorings and local structures such as the Golden Mile. Even with comparably low and declining numbers of migrants, such structures play a core role in their translocal lives; however, local actors have started to adjust to the changing migration flows. The Thai supermarket for example has started to stock more Vietnamese products as workers from Vietnam increasingly join the labour market in Singapore -though they have not made the Golden Mile their meeting place. The small shops and restaurants can also adjust to the reduced numbers of Thais as they also have Singaporean customers who appreciate Thai food. Many smaller job-placement agents specialized in Thai migrant workers have closed their business, as Arisara reports. Only the larger agencies like hers can adapt to the new situation by focusing on other nationalities like Burmese, Bangladeshi, or Vietnamese migrants. In a similar way the small remittance services are affected. They initially developed through networks that were specialized in sending money to Thai banks and moneylenders. These networks are now declining, and only the transnational remittance companies are able to adapt.
Of course not all changes in the Golden Mile are related to the decrease in the number of Thai migrants: with digitalization and the rise of smartphones, some shops become obsolete, such as photo shops, or the music store selling Thai pop and country music CDs. The owner of the latter complained that most migrant workers listen to their music on their phones and through internet streaming, and closed his shop in 2016. Thus we can discern at least two bundles of factors that determine the adaptability and persistence of translocal moorings in Singapore: the diversity of functions and clients or users, and the depth of translocal meaning or the agglomeration with other translocal moorings. Businesses serving merely for migration-related needs, such as the agents, or small remittance services have little possibility to adapt to other clients.
Our research in Singapore reveals that social embeddedness can have different effects regarding the temporality of translocal spaces. Permanent migration with deep social embeddedness at the destination places can lead to a loss of translocal connections to the place of origin -or it can produce profound translocal spaces which can outlast migration flows or systems. In the case of Singapore, social embeddedness in the host society is actively prevented. Thus, the migrant workers orient their lives towards their place of origin. Translocal practices such as sending money, calling home, imagining, singing, and the like ensure their belonging with people and the places they come from, and with their translocal group. This case from Singapore shows that migration movements are necessary for the emergence of translocal spaces and moorings, but it is rather the quotidian mobility and the rhythms of everyday practices that influence their persistence and decay.

Conclusion
Previous research has highlighted the key role of migration for the emergence of transnational or translocal spaces -without, however, paying much attention to the temporal dimension of translocality. This study, with a specific focus on the temporality and mutability of translocal spaces, found that it is necessary to look both at migration and the translocal connections of migrants, and also at their anchoring in different structures and moorings. Thus, combining the mobility approach with a translocal perspective offers a fruitful way to explore not only the production of translocal spaces, but also their temporality and mutability.
The context in both the places of origin and the place of destination of Thai labour migrants influence their translocal embeddedness and interconnectedness. Thai labour migrants live between the two worlds of their villages and the global city of Singapore. The notion of living between can be understood in different ways. In one sense, the movements and practices of Thai migrant workers in Singapore are oriented towards their absent lives in the villages where they belong. They toil in Singapore to enable their families to have simple but secure lives. At the same time they can only remotely take part in everyday life in the village via phone calls, chats and, video calls.
In another sense the migrant workers are in an in-between state, as they live fairly segregated from the Singaporean society and are constrained by the structural context of immigration law and social institutions. This lack of social embeddedness in the host communities is compensated by a stronger situatedness within the translocal group of labour migrants, which have their own (trans)local structures and institutions. Social embeddedness, and the resulting local and quotidian routes, rhythms, and practices are key factors influencing the local anchoring, and concomitantly the temporality and mutability, of translocal spaces and moorings. Further research on this nexus between the rhythms of migration, the temporality of translocality, and the intensity of translocal social embeddedness is needed to understand in more detail the nexus between mobility and translocality. A comparison with other migration systems would be a valuable contribution for a better understanding of the role of diversifying migration patterns and processes.

Geolocation information
The research was conducted in Thailand and Singapore following a multi-sited research approach.