From autonomous to autonomist geographies

Autonomist Marxist ideas and concepts are resurgent and, with their latent spatiality, are well placed to contribute to radical geographical debates. In particular, the methodology of ‘class composition’ analysis provides a rigorous, materialist critique of transforming capitalist social relations. This paper first provides vital historical–theoretical context from the milieu of Italian Operaismo, before emphasising the value of autonomist Marxist analyses of three contemporary geographical frontiers: labour process, migration, and social reproduction. It ultimately argues that the laudable motivations of the autonomous geographies project, explored in this very journal, would be better served through an explicitly materialist autonomist geography.

Geography witnessed an autonomous turn in the mid-to-late 2000's. Following Hardt and Negri's best-selling Empire (2000) and Multitude (2006), the English-speaking academy was introduced to other Italian thinkers (Berardi, 2011;Lazarato, 1996;Virno and Hardt, 1996;Virno, 2004) and radical political theorists taking inspiration from the same rich seam of praxis (Cleaver, 1979;Holloway, 2002Holloway, , 2010. For geographers this provocative work spoke against 'capital-centric' analyses of social relations, and to the seemingly multitudinous forms of agency and resistance emerging in the contemporary alterglobalisation movement (Cuninghame, 2010). The triad of Hardt/Negri, Cleaver and Holloway became key reference points for an emergent 'autonomous geographies' literature (Chatterton and Pickerill, 2010;Hodkinson and Chatterton, 2006;Pickerill and Chatterton, 2006) and other agency-oriented critical geographical approaches (Cumbers et al., 2010;Clough and Blumberg, 2012;Marks, 2012). Yet strikingly, this work tends to obviate the category of class and rarely mobilises the key theorists and categories of Operaismo (which translates as 'workerism' and has latterly become known under the 'autonomist Marxism' umbrella term, a lineage we unpack below), instead drawing on more diffuse post-autonomous currents. For us, this is a critical lacuna, because careful scrutiny of Operaismo's powerful conceptual lexicon can help re-invigorate geographical debates around class and activism as well as autonomous-inspired political praxis more generally.
Definitionally, we distinguish between what we refer to as autonomous geographies, which embraces an expansive notion of autonomy, from autonomist geographies, which suggests a closer link to Operaismo's materialist theoretical categories and class perspective (often elided in autonomous geographies). In a paper in this journal, the autonomous geographies collective derives a definition of autonomy from the Greek autos-nomos ('self-legislation'), noting how this broad, inclusive definition helpfully generates synergies with numerous heterodox political trajectories including: autonomist Marxism, anarcho-syndicalism, regional separatism, anarcho-primitivism, Zapatismo, ecologism and anticapitalism. They also acknowledge that this broad definition potentially presents 'autonomy' as a 'dangerously fuzzy concept' (Pickerill and Chatterton, 2006: 732). We share certain reference points/emphases with the autonomous geographies collective, including a stress on collective agency, a critique of capital-centrism (albeit with different emphases) 1 and the conviction that academic research cannot claim a position of neutral objectivity. Yet, we believe that closer attention to the conceptual insights of Operaismo, can help address some of the issues observed by the collective.
In an admirably self-reflexive discussion, Chatterton and Pickerill (2010: 487) identify an important limitation in autonomous practices in the UK as they define them: 'political visions and values were often implicit or taken for granted rather than rigorously interrogated. This often led to a neglect of discussions of the wider social-institutional arrangements or political contexts groups find themselves in'. We argue that reference to the core categories of Operaismo, and especially the concept of class composition, can help remedy this. With its acute understanding of the close relationship between forms of struggle and forms of production (Battaggia, 2018;Kolinko, 2001;Mohandesi, 2013), class composition provides an eminently materialist methodology for understanding the relationship between autonomist agency and the social institutions of capital and the state. We address the concept of class composition more fully in the next section. 'Autonomy', within Operaismo, means autonomy from capital and autonomy of the waged and unwaged to define their own interests and organise autonomously from official organisations (often but not always including Leftist political parties and trade unions), and it sometimes means autonomy from other groups within the movement (e.g. women from men). Although much, if not all, of Operaismo was 'autonomous' from the 'official labour movement' (Negri, 2005), it was not from Marx's categories and his critique of political economy, to which it returned to renew and develop (Wright, 2017). The critique of 'capital-centrism' in Operaismo thus does not mean less focus on capitalist relations, but a partisan perspective placing workers' recalcitrance at the centre of the analysis of class relations and capitalism as a social form (Cleaver, 1979;Tronti, 2019).
'Autonomy' is a contested term containing divergent strands of praxis with origins in Operaismo (Wright, 2008). Within this milieu, distinctions can be made between Operaismo, autonomia (post-Operaismo), post-autonomy and autonomist Marxism. Operaismo (c. 1960-1973 refers to the original phase of Italian autonomy and primarily, but not exclusively, to the worker-student struggles associated with the concept of the 'mass worker' (the relatively undifferentiated, unskilled assembly line worker of Taylorism-Fordism) from the early 1960's until economic recession and industrial restructuring in the early-to-mid-1970's. The 'autonomist Marxism' of Zerowork, Midnight Notes, Big Flame, Wildcat, Notes from Below, and other groups retaining and modifying Operaismo's categories, in our view continues this lineage. Autonomia (post-Operaismo) (c. 1973-1980) refers to the wider diffusion of struggles in the 'area of Autonomia' during the 1970's and the emergence of marginalised 'new social subjects' (women, migrants, the unwaged, youth) in the wider 'social factory,' culminating in the pro-revolutionary 'Movement of 1977' (Berardi, 2007;Cuninghame, 2005). Post-autonomy (1980-present) succeeds from post-Operaismo's engagement with poststructuralism, via Foucault, Deleuze and Guattari (Lotringer and Marazzi, 2007;Thoburn, 2003), breaking from Operaismo's antagonistic class position for a more affirmationist, nondialectical perspective (Noys, 2010). Post-autonomy is evident in contemporary theory developing Negri's 1970's conception of the 'socialised worker' for studies based on the 'cognitive,' 'immaterial' and 'affective' labourer (see Berardi, 2011;Lazzarato, 1996;Marazzi, 2008;Virno and Hardt, 1996), and this is perhaps the most well-known and most used iteration of autonomy in geography.
The limits to such schematic periodisation must be acknowledged (Wright, 2008), since historical demarcation tends to obscure continuities, eclipse alternative histories and homogenise internal differentiation within and across temporal periods (Tomba 2013). Yet, it retains heuristic value, not least for showing divergences between Operaismo and post-autonomy, which are often carelessly collapsed. Indeed, post-autonomous theory has long been criticised for a tendency to abandon the grounded, materialist, historically conjunctural analysisclass compositionso characteristic of Operaismo, for alluring yet imprecise and totalising concepts that tend to flatten out class reality and erase important internal differences, contradictions and divisions within social formations (Battaggia, 2018;Federici, 2004;Wright, 2017). The fuzzy concept of 'the multitude' (Hardt and Negri, 2006) is perhaps the most exemplary iteration of this problem. We apply here a grounded, materialist reading of 'autonomist Marxism', stressing the centrality and importance of the class composition methodology to geography and the social sciences, conveying how a spatialised autonomist class composition perspective can provide a more rigorous materialist basis for critical geographical thinking (cf. Clare, 2019Clare, , 2020Gray, 2018aGray, , 2022Toscano, 2004).
This project is part of a diverse, ongoing research agenda re-engaging with the underexplored current of Operaismo, including recent high-profile translations of classic work from Mario Tronti (Anastasi, 2020;Tronti, 2019), Balestrini and Moroni (2021), and a second edition of Steve Wright's (2017) seminal movement history. Long before 'autonomist' thought was belatedly introduced to academia, activists in search of practical insights were busy scouring translations and interpretations of Operaismo from radical magazines and journals, some representing political organisational structures themselves. From the early 1970's onward, these include but are not limited to: Zerowork, Midnight Notes, Radical America and Telos (US); Red Notes, Common Sense, Big Flame and Aufheben (UK); Camarades, Futur Antérieur (France); and Wildcat (Germany). And from the 2000's, Multitudes (France), Colectivo Situaciones (Argentina), Precarias a la Deriva (Spain), GurgaonWorkersNews (India), Gongchao (China), Viewpoint (US) and Ephemera, Notes from Below and Angry Workers of the World (UK). Such groups have applied Operaismo's innovative conceptual lexicon in updated forms to their own particular circumstances. More recently, autonomist Marxist inspired theory has proven itself very pertinent in the current conjuncture both within and outside of the academyproviding politically engaged analyses of call centre work (Woodcock, 2017), the commons (De Angelis, 2017; Huron, 2018), platform capitalism (Cant, 2019), women's social-reproductive struggles (Mason-Deese, 2016;Norton and Katz, 2017;Toupin, 2018), migrant activism and resistance (Mezzadra and Neilson, 2013;Scheel, 2019), urban unrest (Gray, 2018b;Mason-Deese, 2012Risager, 2021;Vasudevan, 2015Vasudevan, , 2017, new municipalism (Thompson, 2020), distribution and logistics (Curcio, 2014;Angry Workers, 2020a), and resistance to racial capitalism (Baker, 2020;Ekers and Loftus, 2020;Olmos, 2019). Italian autonomist Marxism speaks, therefore, not only to an innovative lineage of radical praxis but to crucial political issues in our own time and to a range of key contemporary debates within geography.
It is our contention that class composition analysis addresses a number of vexed and long-standing issues in contemporary human geography: the relation of (heterogenous) class analysis to contemporary struggle; internal differentiation within movements: the construction of agency-oriented approaches that neither forsake determinate analysis nor romanticise resistance; a focus on social reproduction that is not removed from class and class struggle; and the relation between particular and universal forms of struggle. Class composition is of course not the only approach addressing such critical questions, but we argue that it is well-placed to comprehend and act upon such concerns within a single analytical framework. To fully explore the potential of an autonomist geography, we must necessarily briefly return to Operaismo, whose principal works have not been widely translated and therefore remain somewhat opaque outside of Italy. We then expand on the movement's core theoretical category, class composition, which must be understood within its historical emergence in movement praxis for its full meaning to emerge. This is crucial, for when severed from this context and its militant perspective there is a tendency for a rather loose usage of key categories and ideas originating from Operaismo. Following this brief outline, we then explore how class composition analysis speaks to three interrelated themes: (1) labour process, (2) migration and (3) social reproduction. 2 In conclusion, we summarise the benefits of a renewed autonomist Marxist class composition perspective to geographical thinking.

I Operaismo and Class Composition
Operaismo emerged in the early 1960's when young dissidents in the Italian Communist Party (PCI) and Italian Socialist Party (PSI) applied Marx's critique of political economy against both capitalism and the social democratic Left (Wright, 2017). Operaismo formed a recalcitrant 'workerist' response to the planned integration of workers within a popular-front programme of rapid industrial development, 'unitary' national economic reconstruction, managerial prerogatives in the factories and a notion of 'neutral' technology eliding antagonistic class relations at the point of production (Panzieri, 1964(Panzieri, , 1976Tronti, 2019). In Quaderni Rossi, the first Operaist journal in 1961, Raniero Panzieri advocated for a renewed investigation of working-class reality 'from below and in forms of total democracy', as a means to challenge the PSI's 'dogmatic conception of socialism' (cited in Wright, 2017: 16 (Bologna, 2007b;Wright, 2008). The common 'theoretical matrix' of Italian autonomist Marxism (Borio et al., 2005), derives from core conceptual innovations established in Quaderni Rossi and Classe Operaia.
Important strikes at the Olivetti and Fiat autoplants in the early 1960s forged an autonomist workers' path, where the first 'workers' inquiries' were conducted by Operaists and assembly-line workers as 'co-research' (Alquati, 2013a(Alquati, , 2013b(Alquati, , 2013cWright, 2021). These 'non-collaborationist' workers' inquiries inaugurated an investigative, militant co-research tendency (Wright, 2021) which challenged prevailing idealist notions of class identity and broke down the researcher-researched dichotomies of Taylorism and much sociological inquiry (Mohandesi, 2013). The inquiries marked a radical disruption of the Fordist compact between capital, unions and labour in Italy, underpinning Operaismo's emerging notions of 'class composition' and the 'mass worker' during the 1960's (Wright, 2017). The development of widespread worker-student agitations in 1968 produced a new composition of struggle in the 'Hot Autumn' of 1969, when five and half million workers (25% of the labour force) went on strike (Brodhead 1984;Lumley, 1990;Wright, 2017). The 'mass worker' generated an array of rebellious strategies to gain massive wage increases and reduced hours, including wildcat, checkerboard and hiccup strikes (short heavily disruptive strikes at sporadic intervals in and between plants), sabotage, 'working to rule', base committees, occupations and internal factory demonstrations (Brodhead 1984;Lumley, 1990;Wright, 2017).
From the late 1960's, antagonism shifted beyond the factory walls to the education sector, urban struggles at the level of consumption (housing, transport and utilities), and feminist struggles in the arena of social reproduction (Dalla Costa and James, 1972;Gray, 2022;Lotta Continua, 1973). These struggles rejected a politics of liberal rights discourse, for an antagonistic politics of appropriation and social force (Gray, 2018a). 'Autonomia' emerged as a mass movement in the 1970'suniting women, young people, cultural workers, migrants, the unwaged, students, and others living on the margins of societyand coming to a head in the 'Movement of '77' (Cuninghame, 2005(Cuninghame, , 2007. The Operaist project was significantly undermined between 1979 and 1983 by sweeping state repression, with a crackdown on the Brigate Rosse ('Red Brigades') acting as an alibi for the exceptionalist imprisonment of 15,000 movement activists in 'special prisons' (Cuninghame, 2005: 79;Lotringer and Marazzi, 2007). The defeat of the workers at Fiat in 1980 popularly dates Operaismo's demise in its original Italian phase (Revelli, 1996;Virno, 2004).
Beyond this brief historical reconstruction, our main concern is developing the concept of class composition, which in our view is Operaismo's most important and enduring category (see also Gray, 2022;Toscano, 2004;Wright, 2017). An expanded and spatialised form of class composition analysis retains the exemplary form of autonomist praxispartisan, seeking after commonality, nonpresumptive, investigativewhile guarding against some of the limits of post-autonomist thinking, including conceptual over-determination. For Operaismo, class struggle is immanent to capitalist relations and the working class 'should materially discover itself to be a part of capital if it wants then to oppose the whole of capital to itself' (Tronti, 2019: 31). If capital is understood to be both dependent on, and a consequence of, labour-power -'that unique particular commodity which is the condition of all other conditions of production' (Tronti, 2007: 30) then the capitalist class can be conceived as subordinate to the working class (Tronti, 2007: 31). Labour-power within the 'organic composition of capital'the relationship between 'constant capital' (the value invested in the means of production) and 'variable capital' (the value invested in labour-power)is thus understood as the 'material lever of the dissolution of capital, rooted in the decisive point of its system' (Tronti, 2019: 13). Thus, 'the struggle against work' or 'the strategy of refusal' became primary objectives for Operaismo (Tronti, 2007(Tronti, , 2019, generating a distinctive break from the orthodox labour movement (Cleaver, 1979).
In a militant and partisan historiographical endeavour (Bellofiore and Tomba, 2011), Operaismo sought to grasp a trajectory of autonomist struggle by and for the working class, which required capital to contend with the working class as an immanent driver of development and workers' recomposition: a higher unification, massification and diffusion of workers' organisation. This famous 'Copernican revolution' of class perspectiveswhich articulated the workers' movement as the driver of capitalist development (Cleaver, 1979) was premised on a reading of Marx's historical investigations into workers' demands for the reduction of the working day in the mid-nineteenth century, which forced capital to develop its technological basis and further concentrate capital in the factory system, thus inadvertently creating the conditions for the massification of workers' struggles (Tronti, 2019: 21-23). Likewise, Sergio Bologna saw the development of the 'mass worker', who laid waste to workers' councils and the skilled 'professional worker' of the 1920's, as a shifting hegemonic tendency in the changing structures of capitalist development that would ultimately generalise the potential for immanent revolt within and against developing cycles of capital accumulation (Bologna, 1976). The mass worker (Baldi, 1972;Bologna, 1991Bologna, , 1992 became the key figure of struggle in classical Operaismo. But this social figure was not merely a gift from mass industry but instead was actively created by Operaisti to collapse divisions and hierarchies within industry and conceptualise new forms of antagonism with the new class composition of capital and labour (Gray, 2022).
Class composition theory is an immanent method of inquiry and praxis that aims to negotiate internal differentiation and help generalise self-conscious, reflexive movements which are the antagonistic social expression of general tendencies in new class formations (Bologna, 2007a). Suturing stubbornly schematic universal-particular, vertical-horizontal and subject-object dichotomies in academia and Left activism (Nunes, 2021), class composition is premised on a 'theory of praxis or subjectivation', on the one hand, and 'a theory of historical change' on the other (Toscano 2004: 198). This is reflected in the two main aspects of class composition: 'technical composition', the technological and supervisory means by which capital and state control and divide workers in the workplace and in the community; and 'political composition', which refers to how workers and the unwaged seek to overcome these divisions through processes of unification, massification and recombination (Midnight Notes, 1992). As such, it makes a powerful link between the forms of struggle and changing forms of production (Kolinko, 2001;Toscano, 2004;Battaggia, 2018), class composition's signature contribution. The poles of 'technical' and 'political' composition merge two, often separated understandings of class in Marxist thought. 'Technical' relates to class as an economic category, defined by structural position in historically specific production processes (class-in-itself). 'Political' refers to class as a political subject, which refers to how a class constitutes or composes itself in a historically specific conjuncture (class-for-itself) (Mohandesi, 2013). In emphasising the 'strict bond between subjectivity and objectivity, between mode of production and mode of rebellion' (Battaggia, 2018), the methodology of class composition guards against both capital-centrism and notions of 'autonomous' agency without reference to prevailing socio-economic institutions (cf. Chatterton and Pickerill, 2010). It also avoids anachronism with an emphatic insistence on the historical (and, as we explore below, geographical) specificity of all struggles. With each shift of class composition engendered by class struggle and structural changes in production, the appropriate form of counterorganisation must change in continual processes of composition (unification, massification and collectivity), decomposition (division and fragmentation) and recomposition (re-unification and remassification) (Mohandesi, 2013).
The US journal, Zerowork, provides an exemplary four-fold framework for addressing the relations between labour, capital and socio-economic institutions. First, and primarily, comes the analysis of the struggles themselves: their content, their direction, how they develop and how they circulate. Second, the study of the dynamics of the different sectors of the working class: the way these sectors divide and affect each other and thus the relation of the working class with capital. Third, the relation between the working class and 'official' organisations: trade unions, 'workers' parties' welfare organisations etc., never identifying the working class solely with these organisations. Fourth, all these aspects related to the capitalist initiative in terms of general social planning, investment, technological innovations, employment and to the institutional setting of capitalist society (Midnight Notes, 1992: 111-122). Here, we see a distinction between class composition and other forms of workers' organisation such as associational power and community unionism, since the autonomy of workers' struggle is considered primary over 'official' workers' organisations, including unions (though this does not preclude working in and with unions and alliances between unions and non-labour groups). Hence, the frequent use of 'wildcat' strikes and innovative workers' tactics beyond the control of bosses and unions (Wright, 2017). Additionally, autonomist Marxism's 'refusal of work' strategy (Tronti, 2007(Tronti, , 2019 borne from Marx's fundamental critique of the wage labour relationdistinguishes autonomist Marxist approaches from typical union practices predicated only on achieving better conditions within work (see Holloway, 2010;Federici, 2012). Zerowork's 'model' necessarily requires renewal for contemporary production conditions and local specificities (see Dyer-Witheford, 2008), yet remains highly suggestive for a renewed class composition methodology.
If Tronti's 'Copernican revolution' has sometimes led to a rather invariant, tautological vision of autonomist praxis as a driver of capitalist development (Mezzadra, 2013), we caution that the problem of decomposition is 'every bit as real as that of recomposition' and that compositional analysis should be just as attentive to processes of fragmentation, division and particularisation as it is to processes of unification, socialisation and generalisation within class formation (see Wright, 2017: 208). Capital will always seek to decompose workers' collectivity. Yet, decomposition always offers the potential for political recomposition in new conditions (Bologna, 2007b). However, it would be a mistake to view this as some simple ping-pong process or perennial swing between a 'balance of forces.' Processes of de-and re-composition were understood within Operaismo as configured within the internal contradictions of capitalism, class struggle, and the 'tendency' of capital towards crisis (Gray, 2022). The contemporary geographical utility of Operaist praxis will only be evident with careful attention to the mutable relationship between 'subjective forms of political action and the shifting configurations of space' (Toscano, 2004: 197-198). In the following sections, we show how autonomist thought and a class composition framework has been and can be applied to develop research and practice in three broad overlapping areas in the geographies of labour, migration and social reproduction.

II Labour process
Labour geography aims to exist in-and-beyond economic geography, attempting to counter a latent 'capital-centricity' through a focus on worker agency (see, Coe, 2013;Coe and Jordhus-Lier, 2011;Hastings, 2016;Lier, 2007;Strauss, 2018Strauss, , 2020aStrauss, , 2020b. But despite functioning as an important corrective, autocritiques of labour geography have long highlighted the overwhelming focus on the official labour movement (Rutherford, 2010), white male workers (McDowell, 2015), and struggles in the global North (Bergene and Endresen, 2010), while often overlooking the role of migrants and migration in labour struggles (Buckley et al., 2017). Overcoming such lacunae, we argue that the fundamental motivations of the autonomist projectin particular the Copernican inversion, class composition analysis, and workers' inquiryreveal autonomist geography as a labour geography par excellence. Emphasising how an autonomist approach confronts critiques of labour geography, we then introduce how contemporary autonomist work approaches questions of precarity and logistics, key horizons in labour geography.
As noted, class composition analysis is central to autonomist Marxism, providing not only a detailed reading of the expanded labour process (technical composition) but also the class struggles these processes enable and constrain (political composition). The inverted, dialectical relationship between these two concepts is crucial, with the latter privileged as the 'basis of analysis and political strategy' (Shukaitis, 2013: 56). It is therefore an unashamedly one-sided yet rigorous 'working-class science' (Ciccariello-Maher, 2006;Tronti, 2019). Similarly, labour geography has played an important role in emphasising workers' agency in global production networks (Coe and Hess, 2013;Cumbers et al., 2008;Rainnie et al., 2011) and in shaping multi-scalar economic landscapes (Herod, 1997(Herod, , 2005(Herod, , 2008. This tendency has, however, been accused of overplaying and, counter-productively, romanticising what is typically a 'constrained' agency (Hastings, 2016;cf. Coe and Jordhus-Lier, 2011). By privileging political composition, always framed inagainst-and-beyond technical composition, autonomist analyses focus closely on how different methods of struggle are immanent to modes of production (Battaggia, 2018). But the commitment to detailed consideration of de-and not just re-composition (see Wright, 2017;Clare, 2020) means that an autonomist labour geography, though brazenly partisan, evades romanticism. Relatedly, despite its purported focus on working-class activity, labour geography has been criticised for lacking a robust class analysis, fetishising 'success,' and displaying an aversion to analyses of failure (Das, 2012). Discussion of class in geography, while valuable and important, is often relegated to an identity rather than a dialectical relationship (see Dowling, 2009), blunting the force of much labour geography (Das, 2012). Yet, class composition analysis addresses these issues, with a robust, non-teleological form of class analysis that prioritises working class activity and, as explored in the following sections, is highly attuned to informal and wage-less labour.
Despite important contributions, much labour geography has been criticised by feminist scholars for a latent 'Fordist Marxism' (Smith, 2016), prioritising white, male and unionised labour. This fails to speak to the reality of the majority of the global working class, both currently and historically (Neilson and Rossiter, 2008), and so it is vital that a critical labour geography engages with this much more fragmented and precarious reality. There is much here to learn from autonomist Marxism, which, from its inception, has been hyper-alert to transformations in the labour process and whose very name emerged from the idea that labour struggles can be, and often are, autonomous from, and even antagonistic to, formal labour movements. With class composition analysis, Operaismo provides a conceptual framework and methodology for uncovering commonality within heterogeneity and complexity. Primary attention on 'the struggles themselves' (Midnight Notes, 1992) provides an awareness that the most important developments in class composition may come from (seemingly) unexpected sourcesmigrants, the unwaged, students and the feminist movement, for examplean understanding which labour geography is becoming increasingly attuned to (Griffin, 2021a;2021b;Strauss, 2018). It is no surprise that we have witnessed a resurgence of autonomist thinking in a period increasingly defined by precarity, since the concept was first developed within the Italian autonomist Marxist milieu. This framework and methodology have proved insightful for thinking about self-employed labour (Bologna, 2018), zero-hour contracts (Hastings and Cumbers, 2019;Woodcock, 2017), the gig economy (Cant, 2019) and 'digital workerism' (Englert et al., 2020;Woodcock, 2021). Detailed compositional analysis emphasises agency and strategies of anticapitalist refusal (Tronti, 2007(Tronti, , 2019 often overlooked by more traditional labour geography, yet it does this by foregrounding not romanticising worker agency, and with close attention to the prevailing, yet constantly shifting, social-institutional context of capitalist relations. Another key site of analysis and struggle over the last decade is logistics. In this 'little-known and much discussed (almost always inappropriately) sector' (Bologna, 2014), Italian autonomist praxis has been pivotal, with the autonomist rank-and-file trade unions, Adl Cobas and Si Cobas, engaging with the low-skilled, migrant workers typically left behind by the major traditional unions (Benvegnù et al., 2018). Meanwhile, autonomist workers, scholars and activists have undertaken detailed investigative studies of the sector (Bologna, 2014;Curcio, 2014). Recognising that 'logistics is the logic of capital' (Curcio and Roggero, 2018), more recent autonomist work beyond Italy has explicitly adopted workers' inquiry and class composition as methodologies to investigate contemporary logistical struggles (AngryWorkers, 2020a; Notes from Below, 2020; Ovetz, 2020; Workers Inquiry Network, 2020). An autonomist geography also has great promise for speaking to geography's 'logistical turn' (Coe, 2020), with classic historical-geographical materialist (Swyngedouw, 1999) interpretations of circulation and the spatial configurations of capital -'capital switching', the 'annihilation of space by time', the 'spatial fix' and 'necessary turnover time' (Harvey, 2006)providing essential generative material for grasping the 'lineaments of the logistical state' from more antagonistic perspectives (see Toscano, 2011Toscano, , 2014. If capital's 'logistical fix' (Danyluk, 2018), consolidated by market authority and violent forms of securitisation (Cowen, 2010(Cowen, , 2014a, has had disastrous effects for the environment and working conditions, geographers have stressed the potential for a critical oppositional engagement with logistics (Chua et al., 2018) and potential new avenues for workers' disruption, resistance and translocal solidarity in the 'global social factory' (Cowen, 2014b). In a review of contemporary logistical geographies, Coe (2020) observes a seeming paradoxnamely the relative lack of successful logistical worker struggle given their strategic power and importance in the contemporary economy. When we are witnessing an increase in such resistance, he argues, more research is needed into logistical 'choke points' (Alimahomed-Wilson and Ness, 2018), focussing on the areas where capital (technical composition) is at its most vulnerable and where labour (political composition) has most traction. Yet, for a truly autonomist geography, the matter is not merely technical and objective, it can only be tackled once workers are in actual and contradictory movement (AngryWorkers, 2020b).
The rigorous yet unromantic working-class perspective that comes from class composition analysis has clear synergies with labour geography, speaking to contemporary geographical debates and responding to a series of important critiques, including, as we explore below, the need for labour geography to have a robust engagement with social reproduction (see Strauss, 2020a). An autonomist geography, stressing worker agency and detailed analysis of the dialectical relationship between mode of production and mode of rebellion (Battaggia, 2018) can, therefore, be understood as an exemplary labour geography. It can also be understood as an exceptionally powerful tool in the much-needed reanimation and radicalisation of working-class geographies (see Emery, 2019;Stenning, 2008;Strangleman, 2008), not least around issues of migration and labour mobility, one of the most pressing challenges facing (labour) geography today. The next section explores the relevance of autonomist praxis in this area.

III Migration
Autonomist geography's compositional perspective contributes to key debates in this journal emphasising the importance of: migrant-focussed activism (Ehrkamp, 2019a), migration's politicisation and contestation (Smith, 2019), 'autonomous solidarity' and migrant resistance (Bauder, 2019); moving beyond state-centric analyses (Ehrkamp, 2019b), and foregrounding the links between migration and carceral geographies (Martin, 2020). This section focuses on the links between autonomist thinking, migration and race, with a particular focus on the 'autonomy of migration' current, before ending with an emergent body of work that productively develops autonomist ideas beyond their original Italian context.
Operaismo's praxis emerged in a close relationship with migration and the 'Southern Question'. Ferrucio Gambino, a crucial vector between Operaismo and autonomist organisations in the UK and the U.S., notes how approximately half of the Italian population (25 m of 50 m) moved their place of residence between 1955 and 1975, with most movement from South to North (Gambino and Ruckus, 2018). If trades unions in the Northern factories viewed these migrants with suspicion and even hostility (Bologna, 1977), arguing that they lacked organisational discipline and work ethic, Operaisti recognised that their tendency towards unruly mass refusal formed an antagonistic expression of political recomposition (Wright, 2017). Indeed, by the 1970's, the term 'migrant worker' and 'mass worker' became virtually synonymous within Operaismo (Wright 2008: 125). This was crucial for political organisation since Southern migrants comprised almost 60% of such strategically important workforces as Fiat Mirafiori in Turin (Vasudevan, 2017).
From its inception then, Operaismo was sensitive to internal migration and race within Italy. But the issue of international in-migration became more pressing from the mid-to-late 1970's when a shift from internal-to cross-border migration began for the first time in Italy's history (Gambino and Davis, 2019). Reflecting on experience developed from international engagement with racial struggles in the U.S. (with the Facing Reality group, Zerowork and the Sojourner Truth Organisation) and in the UK (with CLR James and Selma James), Gambino has observed that migration remained a blindspot in the 1970's for much of the Operaist milieu. As such, within Operaismo, the Italian working class was often problematically elided with the working class in Italy, obscuring essential tendencies and differences within the new working-class composition (Gambino and Davis, 2019).
Beyond the likes of Gambino, Sergio Bologna and Lapo Berti within Operaismo, critical engagement with migration also came from the autonomist feminist movement, with Lotta Feminista (1973: 18-19) arguing that anti-racist, anti-nationalist and anti-sexist struggles were concerned with discovering the organisational weakness that permits division within the working class -Selma James' classic 1975 text, Sex, Race and Class (see James, 2012), made very similar arguments. Meanwhile, Mariarosa Dalla Costa wrote presciently on 'Emigration and Reproduction ' in 1974 and, in 1980, on 'Emigration, Immigration andClass Composition in Italy in the 1970s' (see Dalla Costa, 2019), while Sylvia Federici undertook detailed work on migration and class composition with the Operaismo-influenced Zerowork collective in the US (cf. Ramirez, 1977) and the New York City Wages for Housework (WfH) chapter, which helped birth black autonomist feminism (Ekers and Loftus, 2020;Federici and Austin, 2017). Operaismo's incipient relationship with migration as a political concern (Gambino and Davis, 2019) has since been developed more robustly through a growing body of work focussing on the role of migration in changing class composition (Cleaver, 2004;Mezzadra, 2005;Marks, 2012;Pizzolato, 2013;Zanini, 2010).
The autonomy of migration approach (AoM) is arguably the most evolved synthesis of autonomist ideas and migration. Scheel (2019: 44-47) picks out six key features: (1) migration's socio-subjective dimension; (2) migration as constituent force, changing things 'from below'; (3) migrations temporally precede attempts to control and valorise them; (4) state-centred perspectives are inverted; (5) borders are key sites of struggle and contestation; and (6) borders cause differential inclusion and the 'multiplication of labour' (see Mezzadra and Neilson, 2013). Extending Tronti's 'Copernican inversion', AoM focusses on the dialectical relationship between migrants and borders, arguing that the latter emerge as responses to migrant's autonomous movements (Casas-Cortes et al., 2015). Migration is thus a refusal that destabilises methodological nationalism (Mezzadra, 2004). By privileging migrants' subjectivities, rather than focusing on the supposed omnipotence of states, two further key elements of autonomist thinking are transposed to analyses of migration: the immanence of political to technical composition, and the former's tactical privilege (Roggero, 2010;Tronti, 2019). AoM work has explored migrant struggles in the UK (Alberti, 2014), Latin America (Casas-Cortes et al., 2015) and across Europe (Papadopolous et al., 2008), introducing and developing key concepts like the 'mobile commons' (Papadopolous and Tsianos, 2013), the 'deportability of everyday life' (De Genova, 2002), and 'borderscapes' (Altenried et al., 2018). AoM recognises migrants' key role in labour movement renewal (Alberti, 2016) and combats agencystripping narratives in asylum discourse (De Genova et al., 2018), thus functioning as an important primer for autonomist thinking, while expanding the subjects of struggle that drive recomposition.
AoM has, however, received criticism. It seemingly romanticises migration's revolutionary potential (Lan, 2015), promoting affirmative agency while downplaying the power of borders and flattening heterogeneous processes (Scheel, 2019). These plangent critiques are arguably targeted at a postautonomous AoM (cf. Leonardi, 2016). An explicitly compositional AoM, on the other hand, enables detailed structural analysis that is equally attuned to recomposition and decomposition (Clare, 2020). Far from romanticising, a compositional focus on migration's 'subjective dimensions' provides a critical position from which to launch revolutionary structural analyses, while a focus on migration and 'the production and reproduction of labour power' helps rethink and revive class composition analysis (Altenried et al., 2018: 292-298) in the context of migration's centrality to global (labour) struggles (Choudry and Hlatshwayo, 2016). A compositional AoM is thus central to the enrichment of autonomist ideas through engagement with race, class, gender and national identities (Camfield, 2004;Thorburn, 2017).
To avoid 'ossification' (Shukaitis, 2014) autonomist thinking should remain 'heretical' (Tronti, 2019). Focusing on migration helps avoid any provincial (re)production of autonomist thinking. A range of work has applied autonomist ideas in non-Western contexts, exploring changing class composition in, for example, Laos (Brown, 2019), China (Marks, 2012), Thailand (Campbell, 2016(Campbell, , 2018a and Hong Kong (Wong and Au-Yeung, 2019). But arguably the synthesis of autonomist praxis with Latin American decolonial thinking -'two of the most intriguing trends in social and political theory' (Luisetti et al., 2015: 1)has the richest rewards. Latin American struggles have had a lasting influence on (post)autonomist thinking (see Eden, 2012), with 'societies in movement' such as the Zapatistas, Argentina's unemployed workers' movements and Brazil's landless workers' movement providing sterling examples of politics from 'below and to the left' (Mason-Deese, 2016;Vergara-Camus, 2014;Zibechi, 2012). This engagement is reciprocal. Autonomist ideas have been applied to localised struggles (Fishwick, 2019) and continent-wide waves of re-and de-composition (Webber, 2019), but perhaps more importantly, autonomist concepts have been challenged, interrogated and revised. In particular, detailed focus on indigenous struggles has added important nuance to class composition analysis (Neill, 2001), and urban migrants' socialreproductive struggles (Gago, 2017) have reinvigorated key autonomist themes.
The dynamic synergy of autonomist and (Latin American) decolonial thinking illustrates that compositional analyses can travel and adapt. Exploring migration forces class composition analyses to engage with race and national identities as evidenced in geographically diffuse work beyond Italian borders. Such engagement is nascent, yet it provides some of the most exciting and engrossing contemporary autonomist work. A re-engagement with radical class politics is needed in geography, and a truly intersectional and decolonial class composition analysis (that does not forsake the relationship of class to race and gender) is an ideal tool to respond to the complexity and heterogeneity of contemporary class formations. As the next section shows, autonomist class composition analysis also provides crucial tools for radical praxis at the level of social reproduction and expanded urban contestation.

IV Social reproduction
Another enduring conceptual tool derived from Operaismo is the 'social factory' (Pizzolato, 2013;Thoburn, 2003), first developed in Tronti's 'Factory and Society', 1962, and 'The Plan of Capital', 1963, collected in Workers and Capital (Tronti, 2019. Drawing on Marx's concepts of absolute and relative surplus value, on the total process of capitalist reproduction in Capital Volume II, and implicitly on Marx's theory of subsumption (Gray, forthcoming), Tronti attempted to grasptwo decades before neoliberalism gained traction in the US and the UKthe ways in which the post-war state capitalist form of production was penetrating 'all the other spheres of society, invading the whole network of social relations' (Tronti, 2019: 12). The social factory thesis seemingly offered profound potential for conceptualising diffuse class composition beyond the factory walls (Cleaver, 1979), yet in practice Operaismo focused almost exclusively on the industrial workplace (Wright, 2017). However, recent work in the social sciences has attempted to bring the social factory to bear on architecture and urban development (Aureli, 2012), platform capitalism (Muldoon, 2022), warehousing (Delfanti, 2021), and communicative capitalism (Mumby, 2020).
Within geography, the social factory thesis is most renowned through the work of Negri (2000, 2006), where the term is closely aligned with the post-autonomous concepts of 'immaterial labour' and the 'multitude' (Gill and Pratt, 2008), drawing on Negri's long-term engagement with the thesis (Negri, 1987(Negri, , 1991(Negri, , 2005. Yet, Negri's use of the term has been challenged for its totalising conception of real subsumption (Gray, forthcoming), its Euro-centric, teleological conception of capitalist development (Campbell, 2018b;Tomba, 2013) and its neglect of the material 'affective' and gendered nature of labour in the domestic sphere (Dalla Costa, 2012;Federici, 2004Federici, , 2012. Nevertheless, the concept's innate spatiality (Toscano, 2004) and potential to diversify the anti-capitalist front in the sphere of social reproduction (Dalla Costa, 2012) remains tantalising for geographers (Clare, 2019;Gray, 2022, forthcoming;Vasudevan, 2017). As Cleaver (1979: 58-59) has observed, autonomist groups like Lotta Continuaoperating at the intersection of the factory, the university and the metropolisutilised the concept to initiate a wide range of struggles in the education, housing, transport, welfare, health and leisure sectors (Gray, 2018a(Gray, , 2022Lotta Continua, 1973;Quirico, 2021). Meanwhile, Italian autonomist-feminists, 'grasped not only the theoretical concept of the social factory but also the key role of the struggle of non-factory workers--most of whom are women' (Cleaver, 1979: 59). The thesis was central to autonomistfeminist theories of domestic labour and social reproduction (Dalla Costa and James, 1972;Fortunati, 1995;Federici, 2012), and to the conceptual lexicon of Lotta Femminista (LF) and the international WfH network, formed in 1972 (Dalla Costa, 2012;Toupin, 2018). In turn, LF and WfH were seminal protagonists in the broader domestic labour debate, as recognised in key geographical surveys of feminism and social reproduction (Mitchell et al., 2003;Norton and Katz, 2017;Winders and Smith, 2019).
The founding work of autonomist-feminism precedes recent feminist scholarship on the inextricable relation between production and social reproduction (Strauss and Meehan, 2015;Schling, 2018;Winders and Smith, 2019) and studies in social reproduction theory (SRT) (Bhattacharya, 2017;Ferguson, 2019). Yet, the influence of Operaismo's categories is often excised in contemporary references to autonomist-feminism. Although important to acknowledge the feminist 'rupture' from Operaismo for a lack of action on the social condition and roles of women (Curcio, 2020), it is also crucial to recognise the self-acknowledged influence of Operaismo's categories within the radical anti-capitalist milieu of autonomist-feminism (see Dalla Costa, 2012;Fortunati, 2013;Federici, 2018 for personal accounts). Such work loses much of its distinctive meaning when this influence is redacted. Indeed, SRT arguably attempts to recuperate autonomistfeminism into a broader socialist-feminist remit in ways that diminish the radical anti-capitalist, antistatist perspective of autonomist-feminism (Mezzadri, 2019(Mezzadri, , 2020. For instance, Federici's (2012) slogan: 'Wages against Housework'borrowed heavily from Operaismo's 'refusal of work' strategysimultaneously formed a critique of unwaged domestic labour, waged labour and the concept of emancipation through paid work in general (see Dalla Costa and James, 1972;Federici, 2012;James, 2012). This position went much further than most socialists and feminists would accept in the 1970's (Federici, 1984).
The social factory thesis was evident in the formation of Lotta Femminista, which was borne from the Potere Operaia branch in the Veneto region, home to the Porto Marghera petro-chemical complex, where Maria Dalla Costa, Alisa Del Re and other female activist-intellectuals presciently conceptualised 'capitalist noxiousness' beyond the factory walls and into the neighbourhood (Feltrin and Sacchetto, 2021). With reproductive labour, the community and the home, recognised as part of the social factory, the entire sphere of social reproduction was recognised as a potential arena of political contestation (Dalla Costa and James 1972: 38). Yet, in practice, the potential for generalised struggle in the social factory, was understandably sidelined by the milieu behind a primary focus on establishing women's social reproductive labour as pivotal to the reproduction of labour-power (Fortunati, 1995;Mezzadri, 2020). Additionally, there was resistance to political strategies that might re-entrench the family form and state power over women in the area of social services. Indeed, Dalla Costa (2019: 25) considered women's defence of the working-class wage in the early 1970'sliving costs, utility tariffs, social services and rentas 'struggles in defence of a family structure' rather than struggles addressed to 'winning back and redefining one's own individuality, space and levels of wealth'. Despite such reservations, Dalla Costa also acknowledged the importance of social security provision for opening up a degree of autonomy for women and a new terrain of confrontation with the state (see Federici, 2015).
The tension between 'restoring or defending the welfare state, or constructing more autonomous forms of reproduction' (Federici, 2015: vii) remains highly pertinent in contemporary struggles over the 'urban commons' (Cumbers, 2015;Harvey, 2012;Huron, 2018;Joubert and Hodkinson, 2018), the 'new municipalism' (Janoschka and Mota, 2021;Thompson, 2020;Russell, 2019) and other areas of radical urban contestation (Mayer et al., 2016;Vasudevan, 2015Vasudevan, , 2017. Autonomist-feminist work on this issue remains tremendously productive for the contemporary era. For instance, Federici provides an incisive take on the problematic, advocating the appropriation and self-organisation of communal services before demanding the state pay for it, rather than letting the state organise communal consumption on its own terms: 'In one case we regain some control over our lives, in the other we extend the State's control over us' (Federici 2012: 21).
If not always directly involved in urban and social service struggles, autonomist-feminist theorising around the social factory was influential in the wave of often women-led urban struggles across Italy in the 1970's (Gray, 2018a(Gray, , 2022Lotta Continua, 1973;Quirico, 2021;Vasudevan, 2017). A potent geographical perspective can be brought to bear here since such struggles can be seen as an immanent response to the contradictions of tumultuous postwar migration and urbanisation (Ginatempo, 1979;Gray, 2022;Marcelloni, 1979). New fronts and subjects of anti-capitalist urban struggleepitomised by rent strikes, squatting and innovative autoreduction ('self-reduction') strategieswere often led by women (Cherki and Wievorka, 2007;Lotta Continua, 1973;Ramirez, 1992;Quirico, 2021) and addressed key elements of the gendered 'social infrastructure' of social reproduction: housing, social services, utilities and transport (cf. Gray, 2022;Hall, 2020). As such, they provide heuristic concrete examples for geographical debates elaborating the links between feminist social reproductive thought and urban form (Katz, 2008;McDowell, 1983). Adapting the crucial class composition concept from a geographical perspective, such struggles have been identified as emerging from a new urbanised spatial composition of capital (see Clare, 2019Clare, , 2020Gray, 2015Gray, , 2018aGray, , 2022, and here we see significant opportunities to develop closer bonds between urban-geographical theory and autonomist theory and practice. The theoretical tools of the social factory and class composition analysis, and the lessons from Operaismo-influenced urban and social reproductive struggles, provide a rich resource for geographers elaborating a radical geographical anti-capitalist praxis between urban space and social reproduction across the globe (Brown, 2019;Campbell, 2018a;Gray, 2018b;Mason-Deese, 2012Risager, 2021). Operaismo's focus on the industrial labour process remains essential (AngryWorkers, 2020a;Silver, 2003), yet thinking through the concept of class composition, Operaismo's innate spatialities, and the work of autonomist-feminists, offers significant potential for contemporary radical geographical praxis in the sphere of social reproduction and at the intersection of formal/informal labour.

V Conclusion
Over a decade ago the autonomous geographies project helped reinvigorate radical, agency-oriented thought in the discipline (Chatterton and Pickerill, 2010;Hodkinson and Chatterton, 2006;Pickerill and Chatterton, 2006). We suggest it is possible to develop this perspective further through an explicitly autonomist geography, that addresses some of the limits we perceive in autonomous geographies and in the wider post-autonomous literature that is better known in geographical circles. With its precise and grounded conceptual lexicon, there is significant value in a spatialised autonomist (re)turn (see Clare, 2020;Gray, 2018a;Risager, 2021). In particular, we have argued that class composition provides a persuasively practical materialist framework and methodology for carrying out radical autonomist geographical research (Gray, 2022), but this time with a closer understanding of the relationship between mode of production and mode of struggle (see Battaggia, 2018;Mohandesi, 2013). In geography, the class composition concept is already being enriched through a focus on both social (Castellini, 2021) and spatial (Clare, 2019(Clare, , 2020Gray, 2018aGray, , 2018bGray, , 2022 compositional analysis. Here, there lies great potential to radicalise geography through autonomist praxis, and to further develop the innate spatialities of autonomist praxis through geographical thought. We have argued that an autonomist geography speaks directly to three key overlapping areas in geographical research: (I) labour, (II) migration and (III) social reproduction. In section I, we argued that autonomist geography is labour geography par excellence, that its commitment to detailed and nuanced class composition analysis helps develop debates on labour agency (Coe and Jordhus-Lier, 2011), capital-centricity (Hastings, 2016), precarity (Strauss, 2018) and logistics (Coe, 2020), and can help in the necessary task of renewing working-class geography (see Emery, 2019;Stenning, 2008;Strangleman, 2008). Furthermore, in autonomist Marxism's sustained focus on migration (section II) and social reproduction (section III), we see important overlaps with contemporary theory-building in geography. Too many papers have shown that radical geographies often fail to engage properly with gender, race, migratory status and class in a truly meaningful way (Das, 2012;Smith, 2016;Werner et al., 2017). Some of the most exciting and important work in geography pushes against this with a clear recognition of explicitly feminist, anti-racist and decolonial class struggles (see Coulthard, 2014;Gidwani, 2015;Hart, 2018; see also Strauss 2020a, 2020b for overviews). This is again where the benefit of an autonomist geography is clear. Operaismo was forged from active struggle, with inquiries and compositional analyses keeping concepts and conclusions firmly grounded in the industrial workplace. Meanwhile, the autonomist-inflected social-reproductive urban movements of the 1970's were immensely generative for expanding the anticapitalist front beyond factory walls and national borders. It is crucial therefore to continue expanding the scope of autonomist inquiry into new geographies that reflect the changing material conditions of capitalist relations, strengthening, sharpening, and where necessary revising, its core conceptssomething that is especially important as we confront the shifting configurations of global capitalism (Toscano, 2004).
A final word of caution, however. Academia can have sanitising tendencies. Workers' inquiry and class composition analysis are, and must remain, active practices if they are not to merely reproduce 'the detachment of revolutionary (or at least combative) theory from everyday working-class struggles' (AngryWorkers, 2020b). Although there is 'no politics without inquiry' (Emery, 2018), the act of inquiry must remain politicised through rigorous commitment to meaningful class composition analysis. By this we mean an immanent materialist method of social inquiry and analysis that aims to discover and help generalise self-conscious, reflexive movements which are the antagonistic social expression of general tendencies in the new class composition (see Bologna, 2007b). Moreover, there remains a need for a working-class geography that properly embraces and understands working-class heterogeneity, and this paper has argued that an autonomist geography is very well placed to fulfil this role. The autonomous geographies collective sought to imagine a life beyond capitalism, and to provide a toolkit for getting there. That task remains as important as ever, but we believe it is better suited to an autonomist geography.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs
Neil Gray  https://orcid.org/0000-0003-4795-2095 Nick Clare  https://orcid.org/0000-0001-5425-4012 Notes 1. The position of the autonomous geographies collective was influenced by the notion of 'capitalocentrism' introduced by Gibson-Graham (2006), which in our view tends to disavow negation and antagonism in the name of researching 'diverse economies'. Our position is closer to that of the 'Open Marxist' critique of capital-centrism and their insistence on the openness of categoriesan openness to practicewithin and against capitalist relations. This immanent position has a starting point of class antagonism between capital and labour (Bonefeld et al., 1992). 2. Limited space necessitates the omission of key themes within the autonomist Marxist milieu. The commons and cognitive capitalism are two prime examples and we might also have considered the profusion of struggle and research in the platform and gig economies. But these themes, we believe, frame some of the most important autonomist theoretical currents today, while also corresponding with pressing contemporary geographical debates that have rarely been viewed through an autonomist class composition perspective.