Law after Anthropology: Object and Technique in Roman Law

Anthropological scholarship after Marilyn Strathern does something that might surprise lawyers schooled in the tradition of ‘law and society’, or ‘law in context’. Instead of construing law as an instrument of social forces, or as an expression of processes by which society maintains and reproduces itself, a new mode of anthropological enquiry focuses sharply on ‘law itself’, on what Annelise Riles calls the ‘technicalities’ of law. How might the legal scholar be inspired by this approach? In this article, I explore one possible way of approaching law after anthropology, which is to find within law’s own archive a set of resources for an analogous representation of law itself. Drawing on the historical scholarship of Yan Thomas, I suggest that the Roman conception of law as object offers an engaging counterpart to the anthropological take on law as a specific set of tools or, technicalities, or as a particular art of making relations.

In 1928, Claude Le´vi-Strauss took the advice of his professors of philosophy and began to study law on the side, 'simply because it was so easy' (Le´vi-Strauss, 1955: 55). The encounter soon proved disappointing: The study of law is afflicted by a peculiar fate. Caught between theology, with which it then had an intellectual affinity, and journalism, which recent reforms were causing it to resemble, it seems impossible for it to establish itself on ground that is both solid and objective; it loses one of these virtues as soon as it tries to conquer or retain the other. As an object of study for the scholar, the lawyer made me think of an animal purporting to show a magic lantern to a zoologist [Objet d'e´tude pour le savant, le juriste me faisait penser aù n animal qui pre´tendrait montrer la lanterne magique au zoologiste]. (1955: 55-6) The image echoes Florian's 18th-century fable of 'The monkey and the magic lantern [Le singe qui montre la lanterne magique]', in which a performing monkey borrows his master's magic lantern to put on a show for an audience of animals. The monkey -characterized as a 'modern Cicero' -slips a painted slide into the lantern with an apparently wellpractised gesture, and then begins to discourse on the wonders made visible, entirely oblivious to the perplexity of the animals in the audience, none of whom (with the exception of the turkey) is able to see anything at all. The payoff of the fable is that the monkey had neglected to light the lamp, and its moral is addressed to 'those fine minds whose prose and verse are ornate and always admirable in style, but who fail to make themselves understood' [les beaux esprits dont la prose et les vers sont d'un style pompeux et toujours admirable, mais que l'on n'entend point] (Florian, 1843).
Le´vi-Strauss's allusion recalls only the spare elements of the fable: an animal, a magic lantern, and a certain sense of imposture. In the case of lawyers, the imposture consists in the pretension to knowledge. The lawyer-animal suffers from a spectacular failure of what systems theorists call 'second-order observation'. It purports to instruct someone by whose knowledge it is already encompassed, thereby demonstrating a misunderstanding as to what knowledge is really about. 1 And if, stretching things a little, we take the scholar to be the academic lawyer rather than the social scientist, there is a double failure of second-order observation; like the turkey, the academic lawyer sees (or theorizes) where there is in fact nothing to be seen (or theorized). In any case, lawyerly knowledge is at the antipodes of ethnographic knowledge. Le´vi-Strauss's observation that 'the clientele of law repulsed me even more than its sterility ' (1955: 56) pointed up a distinction between two kinds of student: in law and medicine, a brash, assertive and right-wing 'youth' for whom university education was just the anteroom to a professional career; in the arts and sciences, a group of shy, prematurely-aged and left-leaning 'future scholars', for whom teaching and research were ways of practising disengagement. Ethnography was the ultimate expression of the second ethos: 'Without giving up his humanity, the ethnographer seeks to know and to judge man from a perspective that is sufficiently elevated and distanced as to abstract from the particular contingencies of any given society or civilization ' (1955: 57).
Some years later, Le´vi-Strauss conceded that law 'might have changed' (Rouland, 1991: 25-6), but one might venture that developments on the side of ethnography are more interesting. Anthropology is now engaged by the very kind of rhetorical performance that Le´vi-Strauss took to be an epistemological imposture. In part, this follows from a turn away from the modernist assumption that 'law's production and efficacy are mutually derived and, similarly, bound up in (indeed, binding) social circumstances over time' (Greenhouse, 2010: 806). The once-innovative project of analysing 'law in context', or 'law and society', is being eclipsed by projects that do not expect law to function instrumentally (in society), or to have the epistemological qualities that Le´vi-Strauss found to be so wanting. The work of Marilyn Strathern has played a significant part in this reframing of law. Consider the following observations about the value of taking legal rhetoric (as played out in litigation) seriously: If only in order to persuade, the narratives, images, tropes and analogies must at the least communicate what is possible, and anthropological interest in such resources is an interest in the possibilities entailed by what is said and done for what others say and do. It is that possible and potential world that anthropologists abstract as culture. (Strathern, 2005: 50) We can now entertain the possibility of what might almost be called an ethnology of legal rhetoric. And anthropology turns law into an engaging counterpart by valorizing the very quality of legal knowledge that is derided in Le´vi-Strauss's allusion to the monkey and the magic lantern, namely the 'non-epistemological' mobilization of knowledge as a tool or device (Strathern, 2005: 85).
In anthropology, this approach has inspired the work of a surprisingly large group of scholars working on law as a counterpart to anthropology, and many of these scholars are former doctoral students of Strathern's. Here, one should notice the work of Annelise Riles, who has developed a rich account of 'the agency of legal form ' (2005), and a series of analyses of 'technical legal knowledge' as 'a constellation of material and aesthetic features, and forms of expertise that go with them' (2011a: 65). If this is how the inspiration of Strathern plays out in scholarship that seeks to articulate the relation between law and anthropology, how might it be taken up by the legal scholar? If anthropologists are now interested in law as a 'context' for anthropology (precisely because it performs a kind of 'non-epistemological knowledge'), how might legal scholarship turn this interest into an exchange of perspectives? For the lawyer, the object is not to offer an indigenous variation on the 'anthropological exegesis' of legal form. 2 At least one of Le´vi-Strauss's assumptions remains plausible; there is a difference between the exegesis of texts and the practice of ethnography, and when legal scholars draw on anthropological perspectives one suspects that they cannot help but formulate and write up their insights in ways that are deeply inflected by Pottage the experience of working with(in) the law. So one (hopefully plausible) approach is to find in law's own archaeological remains the resources for a reflection on legal technique that works as a counterpart to recent anthropological approaches. These approaches do something that the modern lawyer would least expect anthropology to do; they strip away the modes of sociality that legal scholars over the past few decades have so inventively ascribed to legal form.
If law is not a vehicle for instances or agencies other than itself, then the challenge is to find within law's own (textual) archives an analogous (and hence different) sense of technicality. My hypothesis is that such an analogy can be made by way of the archaeology of Roman law that is developed in the historical analyses of Yan Thomas. The influence that Thomas had acquired in French intellectual life before his death in 2008 was based on the forbidding erudition and thoroughgoing originality of his studies of certain institutions of Roman law. In a series of intricatelyworked articles, Thomas entirely transformed an understanding based on more than a century of heavyweight historical and philological scholarship. His exegesis of Roman law was informed by, and engaged with, a broad range of positions in anthropology, philosophy, and linguistic theory. Occasionally, his engagement with these positions is quite explicit, as in his history of the Roman lawyers' conception of labour, or labour time, which deploys an intensely technical analysis of law in a critique of Hannah Arendt's and Moses Finley's accounts of slavery in classical society. Elsewhere, the engagement is more discreet; the close analysis of the technicalities of Roman legal institutions is turned towards contemporary histories or theories of law in such a way as to set up analogies that are almost Strathernian in character. My somewhat selective use of Thomas's histories might not capture the subtlety of these analyses, and it risks traducing Thomas's own sense of the contemporary relevance of Roman law, 3 but the object here is to use the archaeological remains revealed by Thomas as resources for an exchange with anthropology after Strathern.

Law as Object
Scholars in the English-speaking world are more likely to have encountered the work of Yan Thomas at a remove, by way of the footnotes to some of Bruno Latour's texts (see notably 1999: 88-9;2002: ch. 5). 4 In particular, Latour picks up Thomas's historical reconstruction of the Roman legal form of the thing (res). In Roman law, res was a metonym for the trial process and legal issue which the parties were disputing through that procedure -'res meant first and foremost the trial, or the issue in dispute' (Thomas, 1980: 415). 5 Things took the form of a res de qua agitur, or 'thing in question'. Of course lawyers recognized that legal arguments had to do with things in the world, but the 'real' or 'material' existence of these things was eclipsed by the existence that they came to have within the discursive or rhetorical frame of legal debate: 'The objectivity of a res was simply that which was conferred upon it by a causa [a legal name or definition]: this kind of objectivity resulted from the dialogue in which the partners in the controversy were engaged' (Thomas, 1980: 421). This sense of the thing as res de qua agitur is taken up in Latour's construction of the 'matter of concern', or that which 'brings people together because it divides them' (Latour, 2005: 13). Latour's manifesto for the new art of Dingpolitik assumes a doubling of sociality. On one side there is the now-familiar sense of sociality as assemblage: For too long, objects have been wrongly portrayed as matters-offact. This is unfair to them, unfair to science, unfair to objectivity, unfair to experience. They are much more interesting, variegated, uncertain, complicated, far-reaching, heterogeneous, risky, historical, local, material and networky than the pathetic version offered for too long by philosophers. (Latour, 2005: 9-11) On the other side, there are the publics that might be gathered around or into these issues: 'What are the various shapes of the assemblies that can make sense of all those assemblages? ' (2005: 14). If politics -or 'compositionism' (Latour, 2009(Latour, , 2010 -is the art of articulating this duality, then the archaeological counterpart in Roman law is the relation between res and causa, or between forms and frames, which is articulated by the particular art of enunciation that was materialized in legal formulae: 'c'est la formule qui cre´e le droit' (Thomas, 1974: 112).
It is impossible here to develop even a condensed history of Roman legal procedure, but we should notice the two species of procedure that inform Thomas's sense of law as formularity: the legis actiones, which emerged with the original XII Tables of Roman law (451 BC), and the formulary procedure that complemented and gradually replaced the legis actiones from about 200 BC onwards. The usual illustration of the rigid formality of the legis actiones is the example of the plaintiff who lost his case because (in seeking factual accuracy) he uttered the word 'vines' rather than the prescribed term 'trees'. The formulary action was a written procedure based not on certa verba but on concepta verba [modulated words]: 'the plaintiff now had a much wider choice of formulae, and was himself responsible for the phrasing of the demonstratio (material specification of facts) and the intentio (statement of claim within it) (Meyer, 2004: 82). In the Republican era, when the legis actiones and the forms of action [formulae] coexisted, the plaintiff could choose from a menu: 'The praetor's mode of exposition was to write up the still surviving legis actiones and the formulae on a whitened board (album) so that anyone who chose could come and claim the particular vehicle of his rights' (Greenidge, 1901: 84). The options on this menu were qualified by a specific kind of limitation. Commenting on the survival of forms of action into early 20th-century English law, the historian Frederic Maitland observed that 'the plaintiff's choice is irrevocable; he must play the rules of the game he has chosen, [and in some cases] he may find that, plausible as his case may seem, it just will not fit any one of the receptacles provided by the courts and he may take to himself the lesson that where there is no remedy there is no wrong' (Maitland, 1909: 5). The formalism of the legis actiones survived into the supposedly more liberal forms of action; hence Cicero's characterization of lawyers as 'birdcatchers of syllables', obsessed with 'insignificant things, almost with single letters and the interpunctuations of words' (cited in Meyer, 2004: 83).
For the most part, historical debates about how these formulae worked -or were made to work -are premised on assumptions about law's sociality that are directly problematized by Thomas's reconstruction of Roman law. For Thomas, the essential point about the formularity of Roman formulae was that law was not instrumental; it was not a vehicle for anything other than itself. To begin with, formulae generated their own medium of operation. Law was a matter of verbal performance -it was 'essentially a sphere of verbal action, according to the specific meaning of the word ius, which designated a formula, or effectively pronounced or articulated speech' (Thomas, 1993: 36) -but it was not just a variation on the classical art of rhetoric. Thomas's characterization of the form of action as 'a kind of program of inquiry ' (1980: 417), or of the res as a metonym for 'a network of adversarial debate: res in controversia posita ' (1980: 416), recalls Roland Barthes' characterization of ancient rhetoric as 'a subtly articulated machine, an algorithm [un arbre d'ope´rations], a programme, whose object is to produce discourse ' (1970: 175). And there is a close resonance between and Thomas's sense of the agency of formulae and Barthes' observation (interpreting Aristotle) that 'rhetoric is a techne (it is not a mode of discovery [une empirie]); it is the means of producing things that might indifferently be or not be, whose origin lies in the creating agent, and not in the created object' (Barthes, 1970: 179). 6 But law is also technical in a very particular sense. Commenting on the conventional sense of technicalities as excrescences of ordinary language, Thomas observes that Roman formularity saturated or rearticulated language in a quite specific way: 'Roman law was an inherent structure [une structure de contenu] that manifested itself in a linguistic structure (namely, Latin)' (Thomas, 1974: 111). 7 This observation might appeal to the contemporary sense of law as an autonomous discourse -consider for example Niklas Luhmann's reference (1995) to Thomas's discussion of the language of Roman law -but this association should not be made too readily. In Rome, formulary agency had its origins in the divine, mythopoietic, capacity of legal speech to bring things into existence (Thomas, 1978), but this capacity was gradually translated into what, paraphrasing Max Weber, one might call a more routinized charisma. Roman lawyers used or performed the legal formula as a carmen in the sense of a charm, spell, or chant whose efficacy depended on an absolutely precise and faultless recitation of the prescribed words (Meyer, 2004: 71-2).
In one sense a form of action was a formula in the pharmaceutical or chemical sense; it was a protocol or recipe for making something (happen). To achieve an effect, one had to follow the prescribed procedure to the letter, and one could only realize those effects for which there was a formula. But a formula was not a means to an end. The incantation of a formula measured out a rhythm or periodicity that generated the temporal, existential, frame for the fabrication of res and personae. These artifacts existed within the 'time of the trial' rather than the 'time of the world' (see Thomas, 1977). And, crucially, the 'ends' or 'endpoints' of the legal action -crudely, persons, things, and their qualities -were wrapped into this frame. The ends were just the means seen from a different perspective.
The point can be illustrated by reference to an exemplary legal object or endpoint -the criminal actor. Modern jurisprudence assumes a subject that is answerable for its acts in the world: 'the legal subject, as the legal form of the economic agent, is naturally free to contact infinite relations, just as he is naturally obliged to answer for all his wrongs' (Thomas, 1977: 83). By contrast, Roman law treated acts as things in themselves rather than an expression of subjective intention. The result was a kind of 'eclipse of agency [un de´passement de l'agir]' (Thomas, 1977: 66): The very notion that one might be bound by one's actions was foreign to a way of thinking for which the subject was the accessory of his action. The relation implied in the formula noxae se obligare -to be 'bound to one's act', rather than 'bound by one's act' -is the opposite of engagement in the true sense; in this case, the wrong (noxa) closes around the guilty person and captures him retroactively. The latter is not truly the agent of a wrong, so much as the subject included in it. Tied or bound to his action, he is not required to answer for it, but he is, strictly speaking, 'held' by law; actione teneri means 'to be held by a legal action'. (Thomas, 1977: 71) Roman law addressed wrongs not as the expressions of some complex authorial psychology, but as objects that were related to agents by bonds of possession rather than affect or intention: 'the legal conception of action belonged to the register of possession [l'avoir], or of things [la chose]' (Thomas, 1977: 66). It is true that, in the course of a trial, advocates would draw on a rich discourse of passions, motives, and reputation, but they did so not so as to reach the interiority of the subject but so as to enfold the agent in the action. Rather like the personae of Roman theatre (Dupont, 2000), the qualities of the criminal actor were really the properties of a typified action that entirely eclipsed the person that was annexed to it: 'The relation did not run from the agent to the act, but from the act to the agent. It was an external and synthetic relation, which excluded the psychological and moral continuity on which, etymologically, liability [responsabilite´] is based' (Thomas, 1977: 71).
A similar sense of the reciprocity of frame and form is given by the example of property. The instinct of contemporary scholarship might be to follow the example of Max Weber, whose habilitation thesis (2008) treated property in Rome not by way of an analysis of adjudication but by way of a history of cadastral surveys; property was construed as an effect of social and economic processes that were under way well before interests came to be adjudicated. In an extended footnote discussion, Thomas suggests that the Roman law of property was actually 'relative' because it existed only within the formulary frame of 'adversarial interest, controversy, and exchange ' (1974: 105). Beyond that, with respect to what would now be seen as the assertion or exercise of property rights outside the frame of legal action, Roman law spoke of private powers (potestates) rather than a right or ius. For that reason, Thomas proposes that in the case of ancient Rome it would be more appropriate to speak of the 'property of law' rather than the 'law of property' [il faudrait, pour tenir compte du langage romain, inverser les termes de notre expression: proprie´te´de droit, et non pas droit de proprie´te´]' (Thomas, 1974: 105).
If the endpoint and the procedure, the form and the frame, were actually the same 'thing' seen from different perspectives, then it follows that legal actions were themselves things of a certain kind. The 'typified' persons and things that were brought into being in the course of a trial were reflexes of the objectivity (in the sense of 'thingness') of the form of action itself: Typicality and objectivity are related characteristics. The work of distinguishing, defining and labelling that Roman lawyers performed on recognized actions (there being no other cognizable actions) presupposed that the relevant legal action could be conceived of as a thing that was distinct, neutral, and capable of being catalogued. (Thomas, 1977: 70) The understanding of actions as things is quite suggestive. Modern lawyers see legal procedures as means to ends, or as more or less effective instruments for the realization of social purposes (see Riles, 2005), hence the implicit suggestion of irrationality in Maitland's description of the Victorian forms of action. Roman formulae did not assume the modern premise of instrumentality. The crucial distinction here is between an object and instrument. If a form of action was the peculiar kind of 'object' that is described in Thomas's analyses of legal formulae, then its composition was structured by processes that were initiated and unfolded within the frame of the formula, not by some external social process or instance. As in the case of Latour's 'matters of concern', objects and endpoints are not found in nature, ready to be discerned and acted upon by law through the exercise of cognitive and practical reason, but are instead immanent in the legal operations and transactions that act upon them. 8 And, if 'law' itself consisted only in a set of cognizable actions or formulae, and if those actions were objects rather than instruments, then one might say that there was no such thing as 'law' in the modern sense. Again, whereas the modern understanding of law takes legal forms and institutions to be means to social ends, or as expressions of broader social processes, the set of law-objects existed and were employed without reference to some abstract social function or normative principle. Actions were species without a genus; there was no abstract universal instance called 'law' that was expressed in, or that traversed, these frames.

Form and Frame
If actions were objects, and if 'law' itself was just the set of these objects, then how were the properties of these objects articulated? What were the relations, tensions, or operations that determined the temporality, configuration, and extension of a law-object? How was the duality of form and frame articulated? Thomas's histories of Roman legal institutions focus on three particular techniques: imputation, or the ascription of actions to actors in an operation that defined both the properties of the action and the agency of the actor; equiparation, or the creative analogizing of one legal form to another; and fictions, which created states of fact that did not 'really' exist, or negated facts that did exist in reality. Thomas's reconstruction of Roman and mediaeval fictions is taken up in Bruno Latour's account of law as a regime of enunciation (Latour, 2002). For Latour, what is most interesting is the difference between the legal technique of fiction and 'technologies' properly speaking. By contrast with technological inflections or detours, law cannot 'fold space and time so as to replace its injunctions with another material of expression ' (2002: 293). Instead, a fiction is a very peculiar kind of relay. According to Latour, legal technique is animated by an 'obsession for imputation ' (2002: 297); more precisely, the peculiar art of law is that of connecting or concatenating enunciations: to attach a state of affairs to a text by an operation of qualification; to attach a statement to a speaker by retracing a chain of signatures; to authenticate a document; to impute a crime to a proper noun; to relay texts and documents to one another; to retrace the trajectory of statements; law as a whole can be grasped as an obsessive effort to make enunciation assignable. (Latour, 2002: 294-5) Law is a compulsive and resourceful stenographer, which records and connects statements in something like the manner of a creative child playing with a LEGO set, ingeniously making disparate blocks fit together (for the LEGO metaphor see Latour, 2004).
This reinterpretation of Roman fictions resonates with Thomas's presentation to the extent that it makes the point that legal technique is about making rather than knowing; the 'as if [comme si]' of fiction is not an aspect of 'legal thought' so much as it is an operationalization of the specific technique of law, its way of doing things' (Thomas, 1995: 18). In law, fictions operate not in the dimension of metaphor but as 'effective legal constructions' (Thomas, 1993: 26). What makes a legal construction 'effective'? Again, the important point is that the fictions do not articulate some general process of sociality; instead, they actualize a potentiality that is immanent in legal formulae. Fictions are embedded in the frame of a case or causa; Thomas speaks of an 'enracinement casuistique concret ' (2005: 70). The idea can be illustrated by a specific example -a particular legal questio. 9 If a donor had given property to a church or monastery which was subsequently dissolved, did the property, or what was left of it, revert to the donor, or did it fall to the Pope as bona vacantia? The mediaeval glossators explored a number of variations on this basic question, and proposed a number of possible answers. One answer was that the gift remained valid because the legal beneficiary was the material edifice of the church -the stone and mortar of the building itself -rather than the monks who served it. The effect of treating the grant as a gift to a building was to give the church an existence distinct from that of its human members. Later writers refined this approach by treating the church as a persona ficta; that is, as an abstract entity or name (nomen universitatis) which both represented its members and was represented by them.
These successive solutions emerged from the splicing of analogy into analogy. These analogies could run from things to persons, as in the use of the old philosophical example of the ship of Theseus (which remained the same despite and through the substitution of all its material parts components) to explain how the persona ficta of the city could persist despite or through the substitution of its human constituents. Or, they could run from one kind of collective entity to another. Later answers to the question of the dissolved or uninhabited monastery strung together a number of such analogies. The example of the city as a particular kind of persona ficta, which would persist as a distinct entity even if there remained only one citizen to act on its behalf (see Thomas, 1993), was analogized to the figure of a herd of animals, which remained a 'herd' for the purposes of a legal claim even if there remained only one representative animal, on the basis that what persisted was 'the memory of the thing but not the thing itself [rei memoria, nec tamen res] ' (2005: 60), and then again to the figure of the boat as a kind of collective thing which would continue to exist virtually even when all its components had disappeared, so long as there was an intention to rebuild it. This layering of analogy, added to another analogy between the collective entity and the power of an inheritance (as a particular kind of object) to represent its heir (Thomas, 2005: 65), fashioned a fiction of retroactivity. Not only did the church survive as an abstract persona, but if a dissolved or defunct religious establishment were to be reformed, then the property would be 'restored' to the newly-formed collectivity 'as if' there had been no dissolution and no abeyance of ownership.
From the perspective of 'substantive' legal history, Thomas's analysis makes some quite intriguing suggestions, not least the idea that public agencies, the ancestral persons of public law in the modern sense, were once artifacts of private law transactions. But with the question of technicality in mind, the crucial point is that 'the history of personification was not just the history of an idea, but also the history of a techniquethe legal technique of [temporally] extended property rights [biens jacents]' (Thomas, 2005: 61). Fictions, in conjunction with the techniques of equiparation and imputation, modulated the composition of res and personae so as to give rights a temporal existence or persistence that was thoroughly 'unnatural'. If we want to understand these operations we have to locate the fiction of the church as an enduring persona ficta in the frame of analogical relations from which it emerged. Borrowing Strathern's distinction between the epistemological and the nonepistemological, one might say that the specific technique of law was to turn ostensibly 'epistemological' propositions about things and persons into the vectors of 'non-epistemological' operations.
The agency of legal form lay not in literal, figurative or fictional reference, but in a technique that enwrapped and articulated these forms into a specific technical performance. The being and agency of 'fictitious' statements is given not by the propositional content of the statement but by a specific way that these 'statements' are operationalized within a particular procedure or transaction. This is something different from what lawyers conventionally understand as the process of 'interpreting' rules, or applying them to facts. Thomas insists that the related forms of fictitious personality were not variations on basic normative principles, or samples of the 'fact patterns' to which laws might be applied; they were 'objects' that referred back to the same 'exemplary narrative or originary intrigue ' (2005: 72).
This takes us back to the doubling of process and product, or procedures and endpoints, and it re-emphasizes the idea that the animating principle of legal actions lay within the formula of the action itself.
Rather like the formulae of the ancient form of action, the mediaeval casus consisted in a basic narrative -the premise of the defunct or abandoned monastery -that was progressively complicated, enriched, and constellated by analogies that facilitated the development of specific technical operations. So, for example, the more sophisticated characterizations of the church or monastery as a persona ficta renewed the meaning of the older idea of a grant to a church as a gift to stone and mortar, turning it into a variation on the theme of abstract or fictitious personality. Similarly, each fiction addressed or 'solved' questions that had been cued up by preceding layers of analogy. So each successive solution was looped back into the frame of the casus in such a way that each analogyeach fiction -could be understood as an operation that remade its own context or condition of possibility at the same time as it advanced a particular solution or endpoint. The frame of the casus had a quite obvious limit; casuistic reasoning 'operated neither freely nor speculatively. It operated under the constraint, one might say the urgency, of deciding between one position and another' (Thomas, 2005: 59). But although the technical effect of fictions consisted in the terms of the decision, the logic of fictions is not explained by what came after the decision, its 'instrumental' effects, but by what happened within the frame of the case. As Thomas observes, 'the precise terms of the solution mattered less than the references on which it was based ' (2005: 72).

Contextualization
The idea that Roman legal technique consisted in the instrumentalizing of semantic forms informs Thomas's sense of the continuing relevance of Roman law. This argument was sharpened in the course of Thomas's engagement with an anthropology of a certain kind. From the 1990s onwards, lawyers and legal academics in France became attached to a curious theory of the 'anthropological function' of law. This theory was based in part on the idea that the legal form of the person secured the dignity and subjective expression of the human being (see Edelman, 1999), and in part on the theory of legal persons that emerged from Pierre Legendre's theory of the 'dogmatic function' of law (see, for example, Legendre 1985). Legendre's fusion of legal history, Le´vi-Strauss's anthropology and Lacanian psychoanalysis reconstructs the institutions of western law as the structural frame of a symbolic order that assigns each desiring subject to a gendered kinship role, thereby acculturating the subjective unconscious. Subjects are born twice over; once as beings of flesh and blood, and then again as legal personae. Somewhat improbably, these intellectual currents merged into a conservative critique of biotechnology.
Whereas Strathern takes biotechnology as an effective metaphor for the contemporary art of making relations -'recombinant knowledge' -the critique developed by French lawyers saw the technology of (biological) recombination as the representative of a more generalized corrosion of the old legal armatures of personality, gender and kinship. To state the argument somewhat crudely, if people imagine that they can now 'choose' both their biological sex and the juridical gender that ratifies that 'biological' existence, then law will no longer be able to play the essential role of acculturating subjective desire. Even if legal forms and institutions are artifacts, they are so deeply wired into our social species-being as to have become second nature (Supiot, 2007).
Thomas's reconstructions of Roman law take on this 'anthropological' theory of law by painstakingly reconstructing (in the context of Roman and mediaeval law) the operations of fiction, equiparation and imputation that shape the constitution and agency of the legal form of the person. It follows from Thomas's sense of fictions as a kind of immanent technique that the reality or ontology of personae or res is a reflex of the way that the constellation of a legal formula or casus is mobilized within a specific argument or transaction. Again, the agency of these artifactsthe person and thing -cannot be read off from their semantic or propositional contents; it is an effect of the technical operations into which the particular artifact is articulated.
Thomas directly addresses the theory of law's 'anthropological function' in an extended essay on the decision of Cour de Cassation in a widely-discussed case of 'wrongful birth' (Cayla and Thomas, 2002). In 2000, the court held that a child born with a congenital handicap could bring an action of negligence against the medical practitioner who had advised his mother during her pregnancy. The child's mother contracted rubella early in the pregnancy but her doctor had advised her, on the basis of an erroneous laboratory test result, that she was free of the disease. Although the resulting handicap was directly caused by exposure to rubella in utero rather than medical negligence as such, the court held that the doctor's failure to inform the mother that she had contracted the disease had prevented her from exercising her declared choice to terminate the pregnancy in such a case. The plaintiff was entitled to compensation for the loss of his right to be or have been aborted.
The decision provoked vigorous criticism from a number of quarters: the parents of handicapped children, obstetricians, partisans of the legal doctrine of human dignity, and, inevitably, various species of 'pro-life' campaigner. A group of academic lawyers wrote to the newspaper Le Monde complaining that the court had breached the basic principles of law and legal right, and advocating for 'the anthropological and ethical function of law'. Thomas's commentary is premised on the Roman sense of legal technique. The court may have recognized the right of the plaintiff to sue 'as if' he had been in existence before his birth, but this fictionalization of personality was just a particularly inventive variation Pottage on techniques of fiction that shape the form and agency of legal personality in all its forms. Legal forms do not have an analogue relation to 'nature', nor do they have the kind of stability or integrity that would allow them to function as armatures of subjective existence. As Thomas's analysis of the case of the dissolved or abandoned monastery suggests, persons and things have their existence in legal formulae that are formed and reformed within specific cases or transactions. Again, the semantic or epistemological form of the person (or thing) is eclipsed by the technical operations within which it is actualized. The lesson of Roman law is that we should think of cognitive or epistemological forms as substantives that have been turned into procedures -'contenus de pense´e devenus proce´de´s techniques, en quelque sorte' (Cayla and Thomas, 2002: 168). 10 With biotechnology in mind, Strathern observes that 'one should not overlook the imagination and ingenuity of lawyers in dealing with new issues' (Strathern, 2005: viii): law is 'depicted in various guises, whether contributing to the conceptual resources through which people approach problems entailing ownership or rights, or intervening in disputes, crystallizing certain cultural moments for the sake of advocacy, and so forth ' (2005: viii). Looking at these depictions, we understand, first, that lawyers' ingenuity has already been cued up by the horizon of anthropological inquiry -lawyers are the agents of an art of relation that is invented by the anthropologist -and, second, that their ingenuity pertains to tools (relations) that have a more general social existence or effect. Lawyers' ingenuity is an inflection of a cultural imaginary. Relations have a social life, they are replicated, translated, analogized, and recombined across social fields or disciplines; Strathern (2005) invokes law, science, and anthropology.
Thomas more radically abstracts legal form from any figure of 'society'. Law is not imagined to be an entirely autarchic phenomenon; legal techniques do not 'emerge spontaneously from the speculation of lawyers ' (1993: 26), but the art of recombinant analogy that is described in his archaeology of Roman law is an art that deploys (one might say exhausts) itself in the space between form and frame, within the dimension of law as object. Thomas insists on the historical and institutional specificity of the legal art of analogy. One of the peculiarities of private law, even today, is that it passes through transactions and procedures that have always already translated things in the world into forms with competences that make them available to these transactions and procedures. Latour observes, with reference to Thomas's archaeologies, that 'there is more "society" in law than there is in the society that is supposed to explain the making and operation of law ' (2002: 278). The art of law is that of making objects that are metonyms of a technical frame.
Although this approach resonates quite closely with aspects of anthropological analyses of law, there is an important contrast to be made. For example, Riles turns the chiasmus that binds law and anthropology 11 into a consummate performance of 'law after Strathern', folding 'technicality' into and out of law, turning anthropological tools into legal technicalities, and legal technique into theoretical equipment. This is a sophisticated instance of more generalized theoretical concern with what Niklas Luhmann calls 'second-order observation', or, more precisely, the 're-entry' of the observer into the observed: The observer is part of what he [sic] observes and sees himself in the paradoxical situation of that which he observes. He can observe a company, a society, or a sub-field of physics, as long as he reintroduces the distinction of observer and observed into the object of his observation. (Luhmann, 2013: 120) 12 Once one has this notion of second-order observation in mind one starts to see it everywhere. Strathern's invocation of a 'common capacity or facility in the making of relations that exist in another register altogether ' (2005: viii) resonates with Niklas Luhmann's wonderfully dry characterization of 'culture' as a product of operations of 'duplication': '[C]ulture is a simple duplication of every artifact, including texts. In addition to their immediate meaning of use (gebrauchsinn) they acquire a second meaning precisely as the document of a culture' (cited in Sta¨heli, 1997: 127). The Strathernian tropes of 'seeing twice over', of the reversibility of figure and ground, or the deployment of relations to elicit relations, can be seen as variations on this theme of observational duplication. 13 Anthropological scholarship offers other examples, notably Paul Rabinow's recharacterization of Foucault's analytic of ethopoiesis in terms of subjective 'equipment' (Rabinow, 2003). 14 The reference to Luhmann is a contextualizing move; it draws out a conceptual ecology that one might otherwise miss. In so doing it makes a crucial point about the contextualization of law. As Luhmann's engagement (1998) with 'Old European' thought makes clear, the fact that the paradox of observation has been noticed and aestheticized in social theory is the sign of a 'metaphysical revolution' (Luhmann, 2013: 120). Whether one speaks of 're-entry' or of 'recombination', one has already abandoned the premise of epistemic mastery that is assumed in Le´vi-Strauss's referencing of the fable of the monkey and the magic lantern. Theoretical knowledge is now all too aware of itself, as an adjunct of analytical tools whose operations on the social have to include the observation of the contingency of those operations. The result, so far as law is concerned, is a second-order mode of 'law in context'. Law is not contextualized directly, through the ascription of social agency to legal form, but 'orthogonally', through the distinguishing and recombining of analogizing or contrastive cuts into the social world. So, for example, Euro-American patent law can be switched into the Melanesian form of the malanggan (Strathern, 2001), or the fictional persona of the corporation can be switched into Japanese kinship (Riles, 2011b). Here, contextualization is not a matter of locating objects in contexts; it is the process of situating observations in a theoretical technique of (relational or recombinatory) observation. Although we learn a good deal about the particular aspects of legal knowledge that are drawn into this process, the most vital and visible aspect of an analogizing or contrastive 'cut' is the intelligence of the tool that does the cutting. The effect, as one of the reviewers of a draft of this article observed, may be to turn 'law' into an unbounded object, which unfolds across multiple tangible and intangible dimensions, but this is what one would expect of an object that emerges from a mode of knowledge that aims at self-observation.
By contrast, Thomas hides the technique of historiographical observation within the archaic forms that are being observed; the historian's art is credited to the art of the Roman lawyer. The point is to think with the techniques that are elicited by historical analysis: In a world [such as Rome between 200 BC and 300 AD], in which legal technique [l'art du droit] provided a means for the formal elaboration of human activities such as labour, one risks missing the significance, or even the existence, of that art if one fails to employ the armoury of tools forged for the qualification of these activities, and hence for their production as objects -as objects of litigation and exchange. (2002: 231-2) This comment, made in the course of an analysis of contracts relating to the 'labour time' of slaves, implies that the objects produced by law lend themselves to broader social uses; hence Thomas's reference to 'the juridical morphology of social objects ' (2002: 231). But it also emphasizes that the specificity of legal technique becomes visible only if one remains within the space between form and frame, retracing the recursive analogies that loop the forms of person and thing (back) into the frame of a legal action. There is nothing social about the agency or instrumentality of legal technique. To employ terms that would scandalize Kantians, one might call legal technique a 'means in itself', a means that is its own principle of being. 15 Legal knowledge generates and sustains itself, and is practicable and intelligible without reference to its possible actions upon a social context.
Nor, to sharpen the contrast with anthropology after Strathern, is legal knowledge accessible to techniques of second-order contextualization. For Thomas, the essential operation of contextualization is the involutional process by which legal forms are 'contextualized' within the frame of a legal action or formula. The technicalities that matter are those that can be brought to bear 'within a legal case [en justice]' (2011: 31). 16 This might be just too austere for some tastes. For example, whereas Riles's ethnography of trading rooms in Japanese banks identifies legal technique as an effect of the articulation of documents, legal formulae, and devices such as fictions or notions of 'instrumentality', Thomas says very little about the material dimensions of legal technique, notably the documents (or, more precisely, the tabulae, parchments, and codices) that were essential to the operation of legal procedures (see generally Meyer, 2004). But in a sense this limitation is the object of the exercise. The figuration of legal technique as a 'means in itself' might seem to resonate with Riles's sense of instrumentality as the 'reversible' form that binds the discipline of law to the discipline of anthropology, or as the relation that brings into view the relation between these two disciplines, but the point about a means in itself is that it deconstructs the relationality of instruments, and hence relationality in general. This is where Thomas's history engages the question of law after Strathern. His account of 'formulary' contextualization proposes a sense of 'law as object' that both mirrors and eludes the 'object-ness' that is elicited by relational analyses. So this engagement tells us something about the way in which the inspiration of Strathern has played out in the study of law at the same time as it opens new ways of taking up that inspiration.