Navigazione – Mappa del sito

HomeNumeri57EssaysDocumentality: A Theory of Social...


Documentality: A Theory of Social Reality

Maurizio Ferraris e Giuliano Torrengo
p. 11-27


Nelle società con un grado di complessità più che elementare, si trovano istituzioni, ruoli sociali, promesse, matrimoni, corporazioni, imprese, e una larga varietà di ciò che possiamo chiamare “oggetti sociali”. Da un lato, comunemente parliamo e pensiamo come se tali entità godessero di un’esistenza non diversa da quella dei tavoli e delle persone. Dall’altro, c’è chiaramente una connessione fra il modo in cui le persone pensano e si comportano e il dominio sociale. Nel presente articolo argomentiamo a favore della tesi secondo la quale l’approccio “riduzionista”, molto diffuso in ontologia sociale contemporanea, non è in grado di cogliere entrambi questi aspetti, e proponiamo un nuovo approccio. L’idea principale sottostante è quella di porre a fondamento della realtà sociale un particolare tipo di oggetti sociali, i documenti — e più in generale le tracce di atti sociali. La differenza fondamentale rispetto all’approccio riduzionista è che il contenuto delle intenzioni collettive non risulta più centrale nello spiegare la varietà ontologica delle realtà sociali che ci circondano.

Torna su

Testo integrale

1. The Social World

  • 1 For a realist approach see Thomasson 2002, for a reductionist one Tuomela 1995.

1Societies are made up of people behaving towards one another in certain ways. That is a truism that few would deny. It is therefore tempting to claim that persons and behaviors are if not the only, at least the basic, categories of social reality. However, in societies with a non-elementary degree of complexity, we find entities that do not belong to those two kinds. Consider, for instance: institutions, social roles, promises, marriages, associations, enterprises, states and the large variety of what we can label “social objects”. As is well known, there is a wide spectrum of positions in philosophy and sociology with respect to the ontology of the social: from the realist, who considers social objects to have an existence that is to a large extent autonomous from those of the individuals, to the reductionist, who considers social objects to be at bottom “made of” people and their behaviors1.

2Both extremes seem to have some appeal in everyday intuitions. On the one hand, (I) we commonly speak and think of promises, financial crises, and bank accounts as if they existed on a par with entities such as tables and persons. On the other hand, (II) there is a clear link between what people think and how people behave and the social domain. For instance, it is hard to think that you could be a husband if nobody – not even yourself – believes that you got married. This is not the case for mountains, dogs, and plants: they would exist even if nobody believed in their existence.

  • 2 See Searle 1995. The theory has been revised to face certain problems (such as those raised by Smit (...)
  • 3 In the latest version of the theory, all such intentions have the ultimate form of the attribution (...)

3The mainstream view in social ontology is probably Searle’s theory. Searle’s theory, especially in its latest formulation, is a form of reductionism: social objects are reduced to collective intentions2. More precisely, social objects are the outcomes of attributions of social functions to persons and objects. Such attributions are the content of collective intentions that are shared within a certain community of people3. Thus – in a sense – social objects are just material objects insofar as they are collectively “seen” and “considered” by a collection of social agents. The theory catches the idea that the social sphere is not as real as the concrete, physical world, and it is somehow more dependent on people’s minds. Thus, it is backed up by intuition (I), but it is in tension with intuition (II).

4In this paper we shall advance a approach alternative to Searle’s, one that is sensitive also with respect to the intuition (II) of the “independence” of social entities from individuals. We shall argue that a particular kind of social objects, namely documents – and more generally records of social acts – are the ground of social reality. The fundamental difference from the reductionist approach is that the content of collective intentions will turn out to be not so important in accounting for the ontological variety of complex social realities such as ours.

2. The Categories of the Social World

5The first problem in inquiring into the social realm is that of individuating the general categories involved in it. The social world seems to be characterized by its own kind of properties. Typical examples of “constituents” of the social sphere are the function of being a president, the role of giving order, the quality of having monetary value, to name a few. Can we account for social reality only through reference to properties or do we need a category of social individuals? The reductionist maintains that the whole metaphysics of the social world has to be handled through social properties. The aim of the reductionist is not to enlarge her ontology, when we start considering the social sphere along with the physical one. However, whether the ontology of the social differs from that of the non-social depends also on how we construe the properties in question. It is essential to the reductionist to regard social properties only as “projections” of people’s beliefs. Since if social properties are a class of “tropes”, then they are a further kind of entity to be accepted in our ontology.

6Yet, admitting tropes in one’s ontology does not seem to be enough to vindicate intuition (I) along with (II). Indeed, even a theory with a domain of social tropes can turn out to be very close to a reductionist approach. Tropes in general depend on their bearers (the person who is president, the piece of paper the banknote is made of). It is then reasonable to claim that the distinguishing feature of social tropes is that they depend also on shared beliefs. In order to have a president, you need people who believe that there is a president, and someone of whom they believe that. The overall picture does not seem to be substantially different from Searle’s idea of a collective attribution of social properties to individuals in a domain which is at bottom physical. Social tropes are very “light” entities that disappear as soon as we take away either their bearers or the collective beliefs about them.

  • 4 For general and specific dependence see Simons 1987. Roughly, x depends specifically on y =df x and (...)

7Note that social individuals would not fare better than tropes with respect to vindicating intuition (I), if the same patterns of dependence relationship are preserved. Consider a theory according to which there is a domain of social individuals, but each social individual is not only generically dependent on beliefs and conventions, but it has also a specific existential link to a given material entity – i.e. its bearer4. The idea that each social entity has a “material bearer” at its back is at the core of Searle’s theory. The well-known constitutive rule of social reality, through which the whole social domain is constructed, makes that clear:

(S) X counts as Y in C

8The formula says that any social object Y is at bottom an object X as regarded in a certain context C. (S) can be recursively iterated, and it can thereby yield the wide variety of social entities.

  • 5 See Smith 2003a, 2003b and the reply in Searle 2003.
  • 6 Note that electronic money does not depend on credit cards. For, if I destroy the credit card, I do (...)

9The main problem for the thesis that each social entity has a material bearer is the existence of free-standing Y terms. Things such as debts, electronic money, collateralized obligations and many others seem to lack any specific material bearer5. Free-standing Y terms are entities that (i) depend generically on people, (ii) that are not existentially dependent on any specific material individual, but that at the same time (iii) are not “general” entities such as abstract types (e.g. the presidential role). In so far as I can sell and buy things with electronic money, it seems to be as concrete as professors and banknotes are. But since electronic money does not specifically depend on given material bearers, such a banknotes6, it shows far more “independence” from non-social reality.

  • 7 An alternative is to embrace an asymmetric treatment between trope-like social entities – e.g. the (...)

10The reductionist approach, then, requires us to show that, appearances notwithstanding, in those cases too there is a specific material bearer7. The 1995 version of the theory was at pains to fit free-standing Y terms in the mould of (S), but the 2010 fares better with that. The idea is that (S) is only a first approximation, social entities – not only the usual banknotes and professors, but also more ethereal things such as electronic bank accounts and corporations – more generally require collective declarations in which specific powers are attributed to persons. Thus, the material bearer of social entities are at bottom always persons, and the difference between social entities is always a difference in social powers. It is easy to see that, with respect to intuition (I), the new version of the theory does not add much to the picture.

11Our proposal aims at preserving both intuitions. In so doing, there will be two main differences from the reductionist approach. Firstly, although the existence of social reality depends on the existence of people, the identity of social objects should not be straightforwardly associated with people’s beliefs and intentions. Of course, the creation and continued existence of an institution X requires some sort of agreement among peoples’ intentions, but it would be a mistake to think that all that matters for X to possess certain distinguishing features is what people believe about X. The way in which social entities – such as marriages, promises, laws, enterprises – bind and influence behaviors of social agents is largely independent of personal beliefs (and intentions), and in many cases it is independent also of collective or shared beliefs (and intentions). As we shall see, social objects are grounded in documents and records in general, rather than in collective intentions. Secondly, although social entities are very often linked to material entities – e.g. a professor or a president cannot fail to be also a person, an enterprise requires partners, employees and offices, the economic value of a piece of some commodity seems to be a property of that very piece of the commodity, and so on – this is not a structural consequence of their constitution, namely a consequence of the constitutive rule through which they are constructed.

3. Dependence, Social Intentions and Documents

12Let us start with the first difference from the reductionist approach. In a sense, social objects possess a nature that is independent of what people think about them. More precisely, social reality does not depend on individual subjects as individual experiences and volitions do. Social reality does require social acts, namely interactions of a certain kind among individuals. But social acts are essentially different from individual intentional acts and personal experiences, in that they always involve more than one subject. Even more to the point, it is not in the power of any individual subject to alter most of social reality, not even with respect to the social facts in which she or he is directly involved.

  • 8 See Reinach 1913; Mulligan 1987.

13The binding we create by promising is a well-studied example of this phenomenon8. When I promise you to pay you five Euros tomorrow, you and I are taking part in an event, a social act. This social act produces a social object, namely a promise. This social object has a certain independence of you and me, in that you and I are both constrained by the existence of the promise. The constraint is determined by the content of the act, i.e. that I pay you five Euros tomorrow. In this way, the promise binds or influences both you (you can claim five Euros from me tomorrow) and me (I have to give you five Euro by tomorrow).

14Now, once we have brought the promise into existence, the promise’s existence will not depend on us as our individual volitions and intentions do. For instance, I cannot dismiss my obligation by changing my mind. If I do not give you the money, there will be consequences, if only that I have betrayed your trust. And you cannot waive my obligation merely by deciding to do so – you have to communicate your decision to me, and in more complex cases, there may be official acts you have to carry out to dissolve my obligation. The binding power of the promise, and in general the causal efficacy of status functions, is not reducible to intentional contents concerning the status function (obligations, rights, etc.) created by a certain social act.

15This means that the binding power of the promise cannot be reduced to our individual beliefs about it. Where is the binding power of the promise, then? Of course, partly it is in the agreement of the parties to keep faith to the promise. However, generally speaking, such an agreement works just as a precondition for the existence of social constraints. Facts that concern who is bound and how and with respect to what (s)he is bound depend on the content of the social act that the two parties have performed.

16The content tells us what has been promised to whom, and thus determines the nature of the binding between the parties. The relevant social facts are determined to the extent that the content is. Now, in Searle’s theory, there is a strict parallelism, if not identity, between the content of the social act that brings the social object into being and the content of the intentions of the participants. The nature of the social object depends utterly on shared intentions. Whose intentions? If I have promised something to you, at least your and my beliefs about the promise are relevant. And if there are witnesses, the promise and the obligations probably depend also partly on their intentions. It is not always clear who has to share the relevant intentional content in order for the social object to be there. Yet the general idea is clear: people have to acknowledge that a certain status function has been ascribed, that certain powers have been bestowed on someone, in order for those powers to be effective and thus for the social object to “exist”. Every level of social reality follows the pattern of the very simple example of the promise that we gave.

17Despite the attractions of the idea that that very simple model of content applies through the whole of social reality, it is fundamentally wrong. Shared intentional contents are just accidental elements that sometimes (typically, in the very simple cases) are attached to social entities, but they are not their ultimate ground. Consider a more institutional situation. I buy a house from a building firm. They promise me to build my house in one year, if I pay them a certain amount of money. Of course, without my intention to pay them and their intention to build it, we would not explain how we came to an agreement. However, if what makes social entities “real” entities is their causal efficacy, i.e. the fact that they often have consequences for us, then neither my intentions, nor the building firm’s seems to be the crucial element. What is binding, and what settles the terms of both my and the enterprise’s duties (and rights), is the content of the social act, which is what can always in principle (and often is) be recorded on a “external” support – i.e. a paper document or a digital document. It is the content produced in the social act and then recorded somewhere that establishes the nature of the actual constraints, and guarantees the endurance of the social object. We shall call this recorded content a document, and defend the thesis that documents are the ontological ground of social objects.

18We can now state our thesis more precisely: documents are the basic source of the “independence” of social reality. That is, they are what makes social reality resistant to individual or collective beliefs. One obvious objection is that the thesis that documents are the ground of social reality seems open to many clear counter-examples: there are many social acts in which there are no documents. All very simple cases, such as the one of the promise that we gave above, seem to be like that. The obvious move here is to claim that individual memories are like written records, in a sense, and in so far as they are shared, they can have the role of determining the nature of the social object at issue. Consider a case in which I promise you to pay you five Euros by tomorrow in front of two witnesses. The memories of the witness count as the external record of the content. We have to make sure that they agree on what they remember, and that they are able to communicate it to us if required, precisely because their memories have the role of an external, public support. What about if there are no witnesses? In such a case we have to rely on our memories, and we can be sure that the social bond has actual binding power only in so far as we agree on what we remember. The way we construe the situation in the very simple cases resembles Searle’s construal only superficially. It is true that the existence of the promise in this case is closely linked to our shared intentional content. But if, according to Searle, this is an essential characteristic of every social situation, from our point of view, very simple situations are in a sense defective. The more a social status or bond is subject to the whim of the participants, the further it is from possessing social reality. In a sense, situations in which we do not find external documents to fix the nature of the social object are social only in so far as they mimic the presence of the document somehow. For instance, the honesty of the participants is a substitute for the publicity of the content.

4. The Dependence and the Essential Involvement Relations

  • 9 The importance of documents for social reality is also stressed by Smith 2006 and De Soto 2000.

19We turn now to look at our second difference from Searle’s theory. In order to do so, we have to get to the core of our position9. What are the relationships between social objects, documents, and the persons (or objects) who are affected by the society’s existence? Here are our three fundamental theses concerning fundamental relations in the social realm:

(Object-Document Duality) For every social object O, there is a document D on which O specifically depends.

(Act-Document Duality) For every document D, there is an instituting event (i.e. a social act) E such that D and E specifically depend on each other.

(Validation) For every document D on which O specifically depends, O generically depends on subjects who are disposed to acknowledge D as valid, and to act accordingly.

  • 10 If social acts are processes, as it seems reasonable to assume, then it is probably more accurate t (...)

20Social objects depend on recordings of social acts – that is to say, documents – rather that on the collective beliefs of a community of people. Social acts are always addressed to someone and have a content10. Their content may involve other entities such as persons or events. The persons involved in the content of a social act, and those to whom the social act is addressed do not need to be the same. For instance, if I promise you that I will give John ten Euros tomorrow, you are the addressee, and John is the person involved by the content of the social act.

21In order to make clear the difference between an approach à la Searle and the one advanced here, involvement relations have to be clearly distinguished from dependence relations. Firstly, not all involvement relationships imply dependence of the social object on the entity involved. Secondly, even when there is both involvement and dependence – a case of what we may call “essential involvement” – our account differ from that of the reductionist. We take it as a general truth that social objects only generically depend on persons, in the sense that although no social entity requires any specific human being (or any other material entity) to exist, without human beings who interact with each other, no social reality would ever exist. This claim is unproblematic with respect to “higher-order” social objects, such as the institution of marriage or the legislation regulating bank accounts. However, there are social objects, such as marriages and bank accounts, that seem essentially linked to specific people or objects. Does not the marriage between George and Anne require the existence of both George and Anne? And isn’t my bank account essentially linked to me?

22We do not want to deny that there is a specific link between the marriage of George and Anne on the one hand, and George and Anne on the other. But this link is not a consequence of how the social object is constructed “out of” the material objects involved. In the reductionist approach, social objects are the outcome of collective attributions of functions to objects or persons; therefore, it is essential to any social object (let alone higher order social objects) to specifically depend on some object. In the approach we are envisaging here, the only entities that social objects can specifically depend on are documents. The content of such documents may refer to specific entities (material objects, persons, or other social objects), as is usually the case in a marriage certificate. If the content of the document implies that the social object cannot exist without these entities, then the social object will specifically depend on those entities. Those individuals are essentially involved in a social object (more on the relation of involvement in general below). However, there may be no object essentially involved in a social object, even if it is a first order one, since whether an entity is essentially involved in a social object or not depends entirely on the content of the document, and there is no structural reason for there to be essential involvement relations.

5. Validation

23Another important distinguishing feature of the theory is that general dependence of the social on individuals and their behaviors takes the form of a validation relation. Documents and records in general have to be acknowledged as valid to have a social function. Even in very simple cases, as when I promise you something and we decide to write it down on a note, that note has no binding import if we do not agree to see it as a record of what has been decided. In more general and institutional cases, social acts are addressed to the society as a whole, because they require that people of that society regard the associated document as valid and thus its content as binding. The function of addressing the social act is that of allowing the validation of the social object produced by the act. The validation is a form of dependence between a social object and the persons belonging to a society, whereby the society acknowledges the constraints that the social object enacts. This is a form of generic dependence, and it is so in two senses. On the one hand, social institutions are binding in so far as a certain collectivity of people regard certain documents as valid, although no specific person is required to regard them as valid. On the other hand, in institutionalized society, the validation is carried out by a procedure that is acknowledged as possessing the power of validating the document that is in turn produced by following the procedure. In other terms, in order for a document issued in a certain established way to be validated within a certain society it is required only that the people in that society consider as valid all documents that have been issued in that way.

6. What Are Documents?

24We are now in a position to focus on the main feature of the social ontology we are sketching: the status of documents. Documents are social objects of a very peculiar kind. The first striking feature of documents is that every social object depends on a document. This fact leads us quickly to a dilemma. If documents are social objects, but every social object existentially depends on a specific document, then either documents existentially depend on themselves or every document depends on an infinite chain of documents. Neither of the horns looks appealing. On the one hand, self-dependence does not seem to provide us with an explanation of why documents do not require specific dependence on another entity to have their identity determined (as other social objects do). That is why this looks like an ad hoc solution (granting that self-dependence here is not meant to be a trivial notion that any entity fulfills). On the other hand, if we are aiming at modeling finite societies like ours, then concluding that every document is grounded in an infinite chain of documents would not be very plausible.

  • 11 By “content” here we mean propositional content. Note, however, that nothing compels us to claim th (...)

25We start to see the solution to this predicament by looking at the other distinguishing characteristics of documents. Documents are not only the more basic social objects, they also are the only one that possesses a socially relevant content11. Being “socially relevant” means that the specific content of a document determines the identity and nature of the social object that depends on it. For instance, suppose that a valid document certifies that I have promised you to pay you five Euros by tomorrow. In virtue of the specific content of such a document, there exists a promise involving you and me. In general, social objects require contents to exist, because what determines their identity conditions are the contents that people regard as valid and binding. As we have seen (and we shall continue arguing in what follows), shared intentional contents cannot be the ground of social objects. But what else can have a socially relevant content? Records of social acts is the only sensible alternative. If that is true, records of social acts (i.e. documents) are the only social objects with contents, and that is why social objects specifically depend on documents.

26This clarifies the situation, but the dilemma is still here, though. If documents are social objects, and each social object has its nature determined by the content of a document, what is the content that determines the nature of a document? It can’t be its content, because its content determines the nature of the social object that depends on it (and we assume that the document is a distinct social object from the social object that depends on it – a marriage certificate is a distinct social object from the marriage that it brings into being), and it cannot be the content of another document, because that will trigger an infinite regress.

27The solution, we maintain, is to look at the validation. Documents are the most basic social objects, because they require only generic dependence on people, and no specific dependence on further documents. The specific content on which they depend is the shared willingness of a certain community to regard as binding every document that has been produced according to a certain procedure.

28In a sense, it may seem that the paradigm of social reality proposed by Searle 1995 (and refined by Searle 2010), which our theory is challenging, holds only for documents, and no other social entities. The fact that documents depend on people in a certain context being willing to acknowledge them as valid, sounds like a variation on the formula

(S) X counts as Y in C

  • 12 Of course, every content – being a semantic entity – depends on people. But this is again just a fo (...)

29When X is the inscription of a certain content (the material support), Y is the document (the social object) and C is the context that comprises the relevant persons. However, the resemblance is only superficial. That a certain piece of paper counts as a document in certain context in virtue of the fact that the people in that context regard it as valid does not mean that its content depends on what such persons believe about the document. The identity of the document is given eo ipso with its content – we do not need to look into people’s heads to individuate it12.

30The infinite regress is thus defused by the fact that documents have contents, as beliefs and intentions do, and thus we de not need to look somewhere else to ground the identity of documents, as it is necessary to do in the case of other social objects. Documents require only generically people who acknowledge their validity, as any social object does. But what is socially relevant is the content of the document, and not the content of the belief associated with the acknowledgment. Memories and traces in people’s heads can be documents in the sense of being the material support on which is inscribed a content that determines the identity of a social object (as when the memory of the witness of a oath is the document on which that oath depends). But, nota bene, the memory contents on which documents depend are not in turn documents (and that stops the threat of regress). Because the identity of the document – like the identity of the social object that depends on it – lies in its content. Thus, the identity of a document does not depend on anything else apart from its own content, even though its existence generically depends on people (which is not surprising, since it is a social object). For instance, the content of a marriage certificate individuates both the certificate itself (the social object “document”) and the marriage (the social object “marriage of x an y”). But the certificate depends generically on people who regard it as valid, whereas the marriage depends generically on people and specifically on the document.

31Of course, the content of people’s beliefs are relevant for the acceptance of the validity, but we do not have to postulate collective intentions with many and differentiated specific contents in order to explain the ground of the identity of social objects. The identity of social object is to be found in the content of documents, and not in shared intentions. The psychological motivations for the acceptance of the validity of a document – or, more often a class of documents – are surely a precondition to its validation, but they fall outside of social ontology. Interesting as they may be, they are the topic for sociological inquiry, politics and possibly economics, but they are of no relevance for ontology.

32This is the deepest difference between Searle’s approach and the one we are advancing here. According to Searle, social objects specifically depend on what people believe about them. Collective intentions do not have only the role of grounding the existence of the social sphere, but also that of grounding its whole nature: what social objects are like depends on the intentions that are shared within a certain community of people. Contrariwise, in the present approach, the only role of the collective intentions is that of validation. Certain documents have to be regarded as valid, by an explicit acknowledgment or, more often, by the implicit acceptance of a procedure of document production. For, in our complex modern society, what is relevant for validation is how a document is issued. If this condition is fulfilled, then the identity of the social object will depend solely on the content of the document (and the identity of the document itself too) – and thus all the constraints on the people and material objects involved in the social object will depend on the content of the document. Again, the identity of a social object, and therefore the social status of people and things that are involved in it (e.g. who has the duty to do something, who has a certain role, what is the value of a certain commodity and so on), depends only on the content of the relevant documents.

7. Type and Token Instituting Documents

  • 13 Their relationship to space is more complex and it will be sidestepped here.

33Let’s get a bit more into the detail of the relations between social objects and documents by introducing temporal considerations. Generally speaking, all social objects need to have a finite temporal duration: they begin with an act and an inscription thereof, and sooner or later come to an end. If I promise you something, for instance, an obligation begins to exist as soon as I make the promise, and it ends when I accomplish it. Quite obviously, here we are talking about what we normally call tokens (of social objects), rather than about types: tokens of social objects always exist in time13. However, the familiar type-token distinction seems to be too coarse-grained here. Our proposal is to distinguish between two “kinds” of types. The eternal types are the abstract possibilities of there being certain individual social object, i.e. of there being a token of a social object that potentially involves people and other entities. Thus, eternal types are possibilities, which we regard as abstract entities not existing in time (although the ontological status of possibilities in general is not very important here).

34It seems plausible to think that such eternal types have properties that we can discover a priori, namely that there are material necessities concerning them. For instance, a promises has to have a propositional object – something has to be promised – and a certain “expiry date”, if only a vague one. Those material necessities are inherited by their concrete tokens – the actual promises involving subjects. Among the tokens of such abstract types, though, there are certain ones that have a “type like” role with respect to other tokens of social objects. What we may call the temporal types of social objects are token of social objects that have a less concrete nature than individual token, such the marriage of George and Anne, or the University of Zurich, in that they do not involve persons or other concrete individual entities. Rather, temporal types are tokens that enable the existence of further tokens of individual social objects such the marriage of George and Anne or the University of Saarland. In other terms, the content on which they depend is always general (and not singular).

  • 14 The relation of realization is simply the converse of dependence between individual token and tempo (...)

35An example is the law governing marriages in a given society (i.e. the institution of marriage), or the national regulation of contracts, and so on. Now, the type/token relation, as normally understood, holds between the eternal type and an individual token, but in order for an individual token to exist, there must be a temporal type (such as a law or a custom) that allows its institution, which is to say, its existence. Thus, there is a specific dependence relation between temporal types and individual tokens. Every individual token (potentially involving persons, events, and other entities) has a temporal and an eternal type “correlated” to it. Take the marriage between George and Anne. As a concrete token, it is a social object whose existence begins with a ceremony, taking place somewhere, but without the abstract possibility of its existence and a law allowing its institution, it would not have existed. To fix terminology, we shall say that such a token marriage is the instantiation of the abstract possibility that two persons get married. But it is also the realization14 of a certain law, declaring that two persons by doing so and so, will form a married couple. Connected to the temporal type of the marriage and a particular token of a marriage, there are two instituting events. Firstly, there is the declaration (or enactment) that the citizens of a certain country, by doing certain things, become a married couple, and secondly the ceremony thereby two particular citizen of that country become married.

36The two events are in turn connected to two distinct kinds of document (by Act-Document duality), which are records of the content of the correlated acts. Connected to the temporal type of marriage there is a type-instituting document: the actual law governing marriages in a certain country. Connected to an actual token marriage, there is a token-instituting document: the certificate bearing the signatures of the spouses. (Along with these two kinds of document, there are simple records, which do not contain declarations instituting social objects, but which may be required for the institution of individual tokens of social objects).

37Most of the features of an individual token – those that are relevant for providing a taxonomy of social objects – are partly determined by the type-instituting document and partly by the token-instituting document. For instance, the social status of George and Anne as being married for a certain interval of time T depends on there being a type-instituting document that is valid for the period T, which allows the institution of individual marriages. And it also depends on the existence of (at least one) token-instituting document, an act of marriage, which is issued at some time.

38The relation between type-instituting documents and token-instituting documents is – like that between temporal types and individual token – a relation of specific dependence. Thus documents can depend on other documents, and indeed this pattern is general and recursive. There may be a hierarchy of type-instituting documents, according to the generality of their purpose (e.g. regional laws depending on national laws, in turn depending on federal laws). What is relevant here is that only social objects that depend on token-instituting documents can involve persons and things. And that the dependence relation between documents, and the dependence relation between documents and social object is not the relation of involvement that holds between social objects and material objects. When the involvement relation does not imply any form of existential dependence of a social object on further entities, we speak of simple involvement (as opposed to essential involvement). Yet, what is involvement in general?

8. Simple Involvement

39The basic idea behind the involvement relation is very simple. All socially relevant consequences of the existence of a social object are articulated in the content of documents. Something is involved in a social object when it is affected by one of its socially relevant consequences. More precisely, a person, a thing, or a further social object (such as a group or an institution) X is involved in a social object O when the content of the document, on which O depends, is such that X is ascribed a power, a role, a function, a duty, or some other socially relevant property. It is important here to stress again that the involvement relation is not a relation of dependence. Even if in certain cases the social object ends up as depending on one or more of the objects that it involves (essential involvement), this is not simply because there is an involvement relation, but rather it is because of the specific content of the document at issue. Contrariwise, in the reductionist approach, any social object is a material object as considered in a certain context, and thus cannot fail to depend on its “bearer”.

40Therefore, even in case in which it is very tempting to identify a social object with a material object, we should resist that temptation. We think it important to stress this feature of the theory, because the reductionist approach seem to be based on a wrong construal of apparently simple cases, such as that of paper money. Consider a five Euro banknote. That piece of paper is the material object involved in the social object, which is why it has economic value and its possessor (who is also involved in the social object) can use it. Indeed, the case of paper money, which is simple and paradigmatic in Searle’s theory, turns out to be rather tricky in our framework. And this is good news, firstly because the case of paper money is only superficially simple, and secondly because our theory turns out to have the explanatory power to provide a uniform account of apparently very different bits of social reality–for instance, to account for what paper money and electronic money, which superficially are very different, have in common (more on this in a minute).

  • 15 What the content of a document says may depend on the content of further documents, for instance, t (...)

41Let us see a bit closer our story about paper money. Each banknote is a token-instituting document, which depends on a type-instituting document (a law that enables a bank to issue legal tender), whose associated social object involves the very piece of paper on which the content of the document is recorded as well as its possessor. It is only because the content of the document says that this piece of paper has a certain value that a banknote depends on its material support (if you destroy the material object, you destroy the social object too)15. And this is only a consequence of a feature of the particular kind of content at issue: paper banknotes are documents that essentially involve their material support. However, generally speaking, it does not need to be so. A document stating that I owe a certain amount of money, does not need to state that my possessions are essentially linked to the existence of any specific material object. That should be clear if we consider credit cards and electronic money in general. The function of credit cards is that of confirming the validity of a certain document: a file in the bank’s computer that says that I can spend, monthly and under certain conditions, a certain amount of money. But since the material piece of plastic is not essentially involved in the social object, my money does not depend on it: if I destroy the credit card I do not destroy my money – although, by so doing, I usually impair my capacity to prove the validity of the relevant document (the one recoded somewhere in my bank).

42Sometimes the relation between the bearer of the social object and the social object may be even closer than that of paper money, but again that is only because of the particular kind of content at issue. Take the marriage between George and Anne. It seems natural to say that the marriage between George and Anne depends on George and Anne. If either George or Anne goes out of existence, there will no longer be the marriage between George and Anne. But this is not because their marriage is just a certain relationship between George and Anne that we collectively recognise. The marriage between George and Anne is an individual social object that depends on a document that George and Anne (together with a officer, and two witnesses–usually the best man and the maid of honour) have signed. George and Anne are involved in that social object, and they are essentially involved in it only because the institution of marriage (a type-instituting document on which individual marriages depends) requires so. The content of the token-instituting document (the marriage certificate) either implicitly or explicitly states that were either George or Anne to pass away, the marriage would no longer be there, and this is the only reason why it is true to say that the marriage depends on the existence of a wife and a husband. It is only because the document at issue has a certain content that it is correct to speak of dependence and essential involvement here. But involvement, in general, it is not a form of dependence and therefore we do not have any reason to identify the social object with one or more material object.

9. Conclusions

43The theory of social reality that we have sketched here, we believe, has strong explanatory power, and allows us to provide a unified framework for the understanding of complex modern societies. As we have noted, social objects seem to come in two quite different kinds. On the one hand, we have entities that have their own “body” (although they may change it perhaps more easily than material objects), such as husbands, presidents, banknotes. On the other hand, we have what look like disembodied entities, such as debts, loans, shares, electronic money, and so on. If we say that the former depend on their material bearers (as it is tempting to do), then we find a fundamental asymmetry in the social realm. “Positive entities” such as paper money or professors are material objects that possess socially relevant properties (in virtue of a collective attribution of them), whereas “negative entities” are abstract things that depend only on shared representations. Asymmetries are not necessarily bad things, but we have argued that that asymmetry is based on a misconception.

44The misconception lies, in the end, in regarding the relation between a positive social object and its bearer as a relation of dependence for formal, structural reasons. If the social object is nothing over and above the material object as considered by a community of persons, then whatever the social object in question is, it follows that it could not exist, were the underlying material object not to exist. But this is wrong, and not because reductionism in the social realm is necessarily wrong, but because even on a reductionist agenda that would be to go in the wrong direction. If there is dependence between a social object and its bearer, this is only because it is mediated by the specific content of the document on which the social object depends. There are no structural reasons, so to say, to claim that a positive social object depends on its bearer, whereas there are indeed formal reasons to claim that any social object depends on the documents whose content determines its identity.

Torna su


Correia, F.
– 2005, Existential Dependence and Cognate Notions, München, Philosophia

De Soto, H.
– 2000, The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else, New York, Basic Books

Ferraris, M.
– 2009, Documentalità. Perché è necessario lasciar tracce, Roma-Bari, Laterza; Eng. tr. by R. Davies, Documentality: Why It Is Necessary to Leave Traces, New York, Fordham University Press, 2012

Mulligan, K.
– 1987, Promising and Other Social Acts: Their Constituents and Structure, in K. Mulligan (ed), Speech Acts and Sachverhalt: Reinach and the Foundations of Realist Phenomenology, Dordrecht, Nijhoff: 29-90

Reinach, A.
– 1913, Die apriorischen Grundlagen des bürgerlichen Rechtes, “Jahrbuch für Philosophie und phänomenologische Forschung”, 1; Eng. tr. by J. Crosby, The a priori foundations of civil law, “Aletheia”, 1983, 3: 1-142

Searle, J.
– 1995, The Construction of Social Reality, New York, Free Press
– 2003, Reply to Barry Smith, in B. Smith and J. Searle, The construction of social reality - An exchange, “American Journal of Economics and Sociology, 62: 283-309
– 2010, Making the Social World: The Structure of Human Civilization, Oxford, Oxford University Press

Simons, P.
– 1987, Parts. A Study in Ontology, Oxford, Oxford University Press

Smith, B.
– 2003a, John Searle: From Speech Acts to Social Reality, in B. Smith (ed) John Searle, Cambridge, Cambridge University Press: 1-33
– 2003b, The Ontology of Social Reality, in B. Smith and J. Searle, The construction of social reality - An exchange, “American Journal of Economics and Sociology”, 62: 283-309
– 2006, Document Acts [ms],

Thomasson, A.
– 2002, Foundations for a social ontology, “Protosociology: An International Journal of Interdisciplinare Research”, 18-19

Toumela, R.
– 1995, The Importance of Us: A Philosophical Study of Basic Social Notions, Stanford, Stanford University Press

Torna su


1 For a realist approach see Thomasson 2002, for a reductionist one Tuomela 1995.

2 See Searle 1995. The theory has been revised to face certain problems (such as those raised by Smith 2003b and Ferraris 2009) in Searle 2010.

3 In the latest version of the theory, all such intentions have the ultimate form of the attribution of powers to persons.

4 For general and specific dependence see Simons 1987. Roughly, x depends specifically on y =df x and y (do not have parts in common and) necessarily, if x exists then y exists; x depends generically on an object of type S =df necessarily, if x exists then some objects of type S exist. See also Correia 2005.

5 See Smith 2003a, 2003b and the reply in Searle 2003.

6 Note that electronic money does not depend on credit cards. For, if I destroy the credit card, I do not thereby destroy the money in my account. See below §8.

7 An alternative is to embrace an asymmetric treatment between trope-like social entities – e.g. the present president of the United States – who depends both on what people believe and on a specific person – and individual-like social entities – e.g. electronic money, debts, shares, which depend on what people believe, but do not require a specific material bearer.

8 See Reinach 1913; Mulligan 1987.

9 The importance of documents for social reality is also stressed by Smith 2006 and De Soto 2000.

10 If social acts are processes, as it seems reasonable to assume, then it is probably more accurate to say that social acts result in contents, rather than have a content.

11 By “content” here we mean propositional content. Note, however, that nothing compels us to claim that propositional content has to be linguistic. This is a further difference from Searle’s approach.

12 Of course, every content – being a semantic entity – depends on people. But this is again just a form of generic dependence.

13 Their relationship to space is more complex and it will be sidestepped here.

14 The relation of realization is simply the converse of dependence between individual token and temporal token.

15 What the content of a document says may depend on the content of further documents, for instance, the content of money does not have to be written on it, there may be a law that says that each piece of paper produced in a certain way by certain authorized entities is legal tender (i.e. is a document with a content that says that the bearer is the owner of its economic value).

Torna su

Per citare questo articolo

Notizia bibliografica

Maurizio Ferraris e Giuliano Torrengo, «Documentality: A Theory of Social Reality»Rivista di estetica, 57 | 2014, 11-27.

Notizia bibliografica digitale

Maurizio Ferraris e Giuliano Torrengo, «Documentality: A Theory of Social Reality»Rivista di estetica [Online], 57 | 2014, online dal 01 novembre 2014, consultato il 12 juin 2024. URL:; DOI:

Torna su


Maurizio Ferraris

Articoli dello stesso autore

Giuliano Torrengo

Articoli dello stesso autore

Torna su

Diritti d’autore


Solamente il testo è utilizzabile con licenza CC BY-NC-ND 4.0. Salvo diversa indicazione, per tutti agli altri elementi (illustrazioni, allegati importati) la copia non è autorizzata ("Tutti i diritti riservati").

Torna su
Cerca su OpenEdition Search

Sarai reindirizzato su OpenEdition Search