Social ontology is often said to study and account for the nature of the social world. But what exactly is ‘the social world’ and what does it contain? What does it even mean to ‘give an account’ of it?

Although they are not always discussed explicitly, these questions are fundamental to social ontology. In this chapter, I will sketch some common answers to them in order to provide an overview of the field of social ontology (Section 1.1) before I reconstruct Heidegger’s answers (Sections 1.2–1.4).

1.1 Scope and Method

In most definitions, the social world comprises all those things that ontologically depend on human beings and their mental representations. To take just two examples, social ontology studies the entities ‘that arise out of, and depend necessarily upon, human interactions …; those, if any, that could not exist in the absence of human beings and their doings’ (Lawson Reference Lawson2019, 31); alternatively, it studies ‘social and institutional entities or facts’, which are ‘collectively constructed’ and therefore ‘mind-dependent’ (Tuomela et al. Reference Tuomela, Hakli, Mäkelä, Tuomela, Hakli and Mäkelä2020, 1). This way of thinking about sociality restricts the scope of social ontology by – either explicitly or implicitly – distinguishing its proper object from other potential objects, the social from the non-social.

There are, of course, many ways to draw such a distinction. An influential and commonsensical way of doing so is by distinguishing between the social world (e.g., institutions, groups, and artefacts) and two non-social worlds, namely, the mental world (e.g., beliefs, desires, and memories) and the physical or natural world (e.g., electrons, geological materials, and biological organisms).

A social ontology that implicitly or explicitly operates with such a distinction between the social and the non-social has what I will call a restricted view of its own scope. As in the above, the restricted view is often accompanied by the assumption that the social world not only is distinct from some other world or worlds but also ontologically depends upon these. This has the methodological implication that the task of social ontology is to account for the relation of dependency that holds between the social world and the world(s) on which it depends. Social ontology must provide ‘an account of how the social world is built. What are its building blocks and how do they come together to build it?’ (Epstein Reference Epstein2016, 149). If we adopt a restricted view of the scope of social ontology and claim that the social world depends on, for example, the mental world and the natural world, social ontology must account for the basic mechanism by way of which the social world emerges out of a set of non-social building blocks.

A highly influential example of such an approach to social ontology claims that the basic mechanism that produces the social world must be found in the mental world, that is, in the psychological states of individual human beings. Let us call this psychologism. Psychologism is often considered to be the ‘standard model of social ontology’ (Guala Reference Guala2007). It suggests that ‘social entities are constituted by beliefs about beliefs’, and that they, unlike natural entities, must be ‘constantly re-created (or “performed”) by individuals by way of collective intentionality’ (Guala Reference Guala2007, 960–963). The general idea is that the social world constitutively depends on or is generated by something that is not itself part of the social world, namely, the mental or psychological states of individuals. Gilbert formulates the point well when she notes that ‘individual human beings must see themselves in a particular way in order to constitute a collectivity’ (Reference Gilbert1992, 12). Psychologism and the standard model of social ontology holds, in other words, (1) that collective intentionality creates the social world and (2) that the perceptions and thoughts of individuals ground collective intentionality.Footnote ¹ Tollefsen summarises the psychologistic method in the following way: ‘If social facts are not natural kinds but made up or constituted by individuals’ perceptions of their world, then an explanation of those social kinds needs to appeal to individuals’ perceptions (i.e., individual psychology) of themselves vis-à-vis others’ (Reference Tollefsen, Jankovic and Ludwig2017, 393).

Alternatively, one might claim that a successful account of the social world must appeal to entities and properties contained in the natural world. Let us call this naturalism. There are, of course, many forms of naturalism. Different theories offer different demarcations of the natural world, and they also disagree on whether the appeal to the natural world complements or substitutes the appeal to the mental world. Behaviourism, for instance, claims that we need not refer to the psychological states of agents in order to explain social phenomena and takes agents’ behaviour, which is in principle available to third-personal observation, to explanatorily exhaust the social. Others, such as John Searle, are committed to both psychologism and naturalism. For Searle, social ontology studies the ‘portions of the real world’ that exists only by virtue of ‘human agreement’ (Reference Searle1995, 1) – that is, the psychology of individuals – but he also insists that social ontology ‘must respect the basic facts of the structure of the universe’ that are recorded by ‘physics and chemistry, by evolutionary biology and the other natural sciences’ (Reference Searle2010, 4). For him, the central question of social ontology concerns the relation of dependency between the social world, the mental world, and the natural world:

Few contemporary authors are as explicit as Searle, but it is not uncommon to combine psychologism and naturalism in this way.Footnote ²

This assumption about the restricted scope of social ontology is, however, not logically tied to the assumption that social ontology must explain the social in terms of the non-social. It is possible to insist that the social, the mental, and the natural are ontologically distinct, but that the order of dependence or, at least, the order of explanation goes in the other direction so that, say, the mental must be explained as a function of the social (as is the case in, for instance, structuralism and functionalism). Alternatively, one might agree that social and non-social entities are ontologically distinct but that we cannot explain one in terms of the other. Then we arrive at a form of social ontological dualism (or – if we hold that the social, the mental, and the natural are all irreducible to each other – a form of trialism).

These different methodological approaches agree that social ontology is limited in scope. Some, however, doubt that we can make a sharp distinction between the social and the non-social. For instance, Sally Haslanger thinks that it is ‘unlikely that there is a non-circular definition’ of what makes something ‘social’ (Reference Haslanger2016, 8); Epstein writes that ‘it … may be pointless to engage in a lengthy exercise to pin [that which circumscribes social facts] down’ (Reference Epstein2016, 150f); and Gilbert notes that ‘the phenomena aptly thought of as “social” are a motley crew’ (Reference Gilbert1992, 441). This might be a simple conceptual difficulty (which definition of ‘social’ will accommodate all of our intuitions?), but it could also be a symptom that there is no substantial difference between the social and the non-social. The latter is an ontological claim, but from it derives a methodological one, namely that we get off on the wrong foot if we attempt to explain the social in terms of the non-social.

If we reject that the social world is a distinct domain of reality, we have an unrestricted view of the scope of social ontology. The unrestricted view poses the following methodological questions: What then does social ontology do? Which kind of knowledge or explanation does it provide? The unrestricted view does not necessarily claim that it is nonsensical to distinguish between, for instance, artefacts that are causally created by human beings and natural kinds but argues, instead, that allegedly non-social entities, properties, or facts exist (or are conceived to exist in this or that way) because of a latent social process. This broadens the scope of social ontology to include entities that are otherwise categorised as mental or natural. The aim of social ontology is, accordingly, not to explain the social in terms of the non-social but to describe the way in which a given (latently or manifestly social) entity depends upon a particular social process. There are two versions of this approach.

The first version claims that a given entity depends upon an empirical social structure. In its most radical form, this approach claims not just that a specific entity or category (e.g., gender or race) depends for its subsistence or its properties upon a particular empirical social formation but that our conception of reality as such somehow depends on empirical social factors. Consider, for instance, Peter Winch, who, inspired by the later Wittgenstein, postulates that ‘[r]eality is not what gives language sense. What is real and what is unreal shows itself in the sense that language has’ (Reference Winch, Frankenberry and Penner1999, 346). Given that language is a product of a contingent and empirical social formation, Winch claims that the very distinction between the real and the unreal is somehow socially constituted. Another argument to the same effect can be found in practice theory, which claims that entities depend for their intelligibility upon the social practices of which they are part. The term ‘social practice’ refers, in this context, to a mesh of organised bundles of human activity and material arrangements of organisms, artefacts, and things (cf. Bourdieu Reference Bourdieu1990; Giddens Reference Giddens1986; Schatzki Reference Schatzki2003a, Reference Schatzki2003b). In short, entities are only intelligible within a particular social practice or context.

The second version claims that entities, properties, and facts are intelligible only due to a transcendental social structure. On this account, there is no clear-cut distinction between the social and the non-social because entities exist or appear to us in the way that they do because subjectivity itself (or, to be precise, the correlation between subject and object) implies a set of necessary and a priori social relations. On this view, the aim of social ontology is to account for these necessary and a priori social relations. Phenomenologists such as Husserl, Merleau-Ponty, and Heidegger have pursued this type of social ontology in the greatest amount of detail.

1.2 Fundamental Ontology and Social Ontology

Let us now turn to Heidegger. What does his general conception of ontology imply for the idea of social ontology?

Both psychologism and naturalism assume that the social world depends on another domain of reality. That is, they both claim that social entities, properties, and facts owe their existence to another independent and, hence, more fundamental substratum of reality in terms of which we must then account for their existence. As a phenomenologist, Heidegger is highly critical of this approach. In fact, his conception of ontology undercuts this very distinction between dependency and independency, for example, between mind-dependency and mind-independency. In this regard, he follows Husserl’s epoché, the methodological ‘bracketing’ of the question whether or not a given entity exists in itself.Footnote ³

For Husserl, the epoché is the ‘method by which I appreciate myself purely: as ego, and with my own pure conscious life, in and by which the entire objective world exists for me and is precisely as it is for me’ (Reference Husserl and Cairns1960, 21). This does not mean that phenomenology is only concerned with introspection, with the inner life of the phenomenologist. Rather, the basic idea is that once we bracket the question whether a particular entity is so and so in itself, we get a way out of classical metaphysical puzzles because it now becomes possible to study the ‘conditions that make possible not the existence of entities in the world (the issue of existence has been bracketed), but their meaning as existing, and indeed their being given as anything at all’ (Crowell Reference Crowell2013, 47). In other words, we study how we ourselves must be in order for the entities to appear to us as they do. In Husserl’s vocabulary, phenomenology studies how entities are constituted in their correlation with the subject. Unfortunately, the term ‘constitution’ is somewhat misleading. The claim is not that the ego somehow ‘produces’ entities or that it provides the ‘material’ of which entities are made (in the way that we, for instance, just saw Guala claim that ‘social entities are constituted by beliefs about beliefs’ [Reference Guala2007, 961, my italics]). Instead, phenomenology studies what we might call intentional constitution. In McManus’ description, intentional constitution is ‘the structuring of an intentional agent that allows other entities to manifest themselves to it’ (Reference McManus and Wrathall2021, 184).

Heidegger’s phenomenological inquiry thus concerns a different kind of dependency than the one central to psychologism and naturalism. Following the epoché, he writes that his phenomenological ontology does not concern entities as such but being in the sense of ‘that which determines entities as entities, that on the basis of which entities are already understood’ (SZ, 6/25f). Heidegger is not interested in whether a given entity depends for its subsistence or properties on another entity or set of entities. His question concerns the being of entities. The term ‘being’ refers, roughly, to the intentional structure that allows entities to become intelligible. In short, Heidegger’s fundamental ontology examines how entities depend for their intelligibility (rather than subsistence or properties) upon a taken-for-granted intentional structure.Footnote ⁴

For Heidegger, the aim of ontology is not to explain a given domain of entities by showing how it causally or constitutively depends on another domain of entities. Nor is it, as Quine would have it, to take inventory of the objects in the world. Phenomenology is ontology, for Heidegger, because phenomenology describes what subjectivity must be in order to encounter different kinds of objects (SZ, 37/61).Footnote ⁵ Heidegger’s ontology describes the being of entities by accounting for the implicit intentional structure that makes entities appear to us in the way that they do. This, however, is not simply a description of what subjectivity is. The point is, rather, that when describing what something is, we necessarily take our point of departure in our understanding of what it is. These two elements – being and understanding of being – are inseparable. As Schear puts it, ‘Heidegger holds that part of what you understand when you understand what it is to be something is what it is to access it’ (Reference Schear and Wrathall2021, 92, my italics). For instance, to say that something is a rook is also to say something about the conditions under which this entity is accessible as a rook (rather than, say, a small piece of wood). Phenomenology is ontology because it describes not just subjectivity but the basic correlation between human being and world by way of which entities are meaningful to us.

Phrased differently, rather than studying the entity as something that causally depends on some other thing or as a higher-order object that constitutively depends on a set of lower-order constituents, Heidegger’s ontology explicates how the appearance of a given entity phenomenally depends on an unthematic and taken-for-granted intentional structure (cf. SZ, 31/54f). This ontological project is transcendental like Kant’s insofar as it aims to make explicit those correlative structures (for Kant, these are the categories and the forms of intuition) that acts as conditions of possibility for our experience of entities, although Heidegger’s account ultimately differs from Kant’s by claiming that the transcendental structure that binds human and world together must be described in conative, affective, and indeed social terms.

There is a stark contrast between this phenomenological conception of ontology and the naturalistic view advocated by someone like Searle. When arguing that it must respect the ‘basic facts’ of the natural sciences, Searle stipulates a naturalistic constraint on social ontology, thereby suggesting that the ‘basic facts’ of the natural sciences are the ultimate explanans. Yet, for Searle, the basic facts have simply fallen from the sky and are beyond philosophical dispute. Heidegger, by contrast, holds that all forms of cognition must be subjected to a phenomenological analysis. He would hence point out that the explanation pursued by someone like Searle is only possible because of an unquestioned and overlooked transcendental intentional structure. Indeed, natural entities and scientific facts are only intelligible as natural entities and scientific facts on the basis of what Heidegger calls a ‘projection’. The projection is the way in which certain features of entities are dimmed down (e.g., utility, beauty, sacredness, sentimental value), while others are brought to the fore (e.g., motion, force, location, time, universal accessibility).

For Heidegger, the ‘basic facts’ of natural science are intelligible only on the basis of a specific engagement between human being and world. This means that the naturalist outlook is itself a ‘distinctive way of making entities present’ [ausgezeichneten Gegenwärtigung] (SZ, 363/414) rather than the ultimate explanans of all of reality. Heidegger thus finds the very notion of ‘basic facts’ problematic, because these facts depend on a set of scientific practices that are themselves in need of phenomenological clarification.

Turning to the question of scope, it seems, at first, that we might be able to adopt a phenomenological method and still insist – with, for instance, Searle – that reality is divided into three different domains, and that social ontology studies the subset of distinctively social entities. It appears that we can reject the naturalistic constraint on our methodology and still hold that social ontology is a subdiscipline of ontology because it studies only a distinct subset of what there is.

But Heidegger also rejects this idea. He argues, as I will show briefly, that the intentional structure that allows entities to appear to us by itself implies a form of intersubjectivity, that is, that being is always a ‘being-with others’ (SZ, 118/155). If this is correct, sociality cannot simply be a subset of what there is. Rather, sociality is itself a transcendental condition of entities appearing in the first place. In Merleau-Ponty’s memorable phrase, phenomenological analysis ‘slackens the intentional threads which attach us to the world and thus brings them to our notice’ (Reference Merleau-Ponty2010, xv). Sociality is one of these threads.

Since they are part of our intentional make-up, Heidegger believes that social relations characterise human mindedness as such. If this is correct, the psychology of individuals, the mental world, cannot be the explanans of the social world, since the psychology of individuals itself presupposes a form of intersubjectivity. For this reason, psychologism fails. Similarly, the naturalistic impulse to explain the social world with reference to a set of basic facts, the natural world, fails because the basic facts can only appear as meaningful on the basis of a set of necessary and a priori social relations. It follows that social ontology should not restrict itself to the study of a domain of existence but must analyse what I’ll call a dimension of existence. Borrowing another formulation from Merleau-Ponty, we should conceive of ‘the social world, not as an object or sum of objects, but as a permanent field or dimension of existence: I may well turn away from it, but not cease to be situated relatively to it’ (Reference Merleau-Ponty2010, 421).

We can summarise Heidegger’s approach to social ontology in two theses. The first concerns the scope of social ontology, and the second concerns the method of social ontology.

1. (1) Sociality is a dimension of the world rather than a domain within the world.

2. (2) Sociality must be accounted for by reference to the de jure or transcendental social structure of the correlation between human and world rather than some non-social level of reality or some de facto or empirical social formation.

Taken together, these theses support the stronger claim that human mindedness and agency is what it is only by virtue of being embedded in and engaged with a shared world. This is the central claim of Heidegger’s social ontology. Although it will take the entire book to defend these claims, the rest of this chapter provides a preliminary clarification of these two theses, thereby setting the scene for the chapters to come.

In the next section, I provide an overview of Heidegger’s early attempts to integrate sociality into his phenomenological ontology. I focus on the period from 1919 to the publication of SZ in 1927, because Heidegger at first, like Searle, distinguished between three domains of entities: the surrounding world [Umwelt], the self-world [Selbstwelt], and the with-world [Mitwelt]. I then outline how Heidegger comes to his mature view that sociality, conceived phenomenologically, cannot be restricted to such a domain of entities but is, instead, a constitutive dimension of the world as such, which is therefore necessarily a shared world.

This implies that we cannot and should not explain the social dimension of existence by reducing it to an allegedly non-social level of reality. The shared world is the holistic cloth from which our understanding of specific entities is cut. Yet, this idea has been misunderstood in the contemporary literature, and I will therefore, in the last section, provide a preliminary clarification of what Heidegger means by ‘world’ and by ‘shared’. In anticipation of the arguments in Chapters 2 and 3, I argue that pragmatic conventionalists like Dreyfus and Okrent fail to recognise the transcendental gist of Heidegger’s social ontology, although they correctly understand the scope of it. In short, Dreyfus and Okrent believe that we share the world if and only if we share certain de facto social formations, namely, conventions or social practices, but I argue that Heidegger is committed to the de jure claim that we share the world if and only if we comport ourselves in accordance with wholes of significance that are responsive to each other’s behaviour and we are capable of intending the same entities. This claim is considerably stronger, methodologically speaking, since it does not take world sharing to be a function of empirical social formations but a transcendental condition of human existence as such.

1.3 Domains and Dimensions

In contrast to what he considers to be Husserl’s overly scientific aspirations, young Heidegger envisions a phenomenology capable of illuminating concrete everyday life. For example, in 1923, he introduces a lecture course by criticising what he deems to be the fundamental inadequacy of traditional ontology. ‘From the very start’, he writes, ‘its theme is being-an-object, i.e., the objectivity of definite objects, and the object as it is given for an indifferent theoretical meaning’ (GA63, 3/2). In contrast, he believes that ontology should focus on a particular kind of existence: ‘[T]hat entity which is decisive within philosophical problems: namely, Dasein, from out of which and for the sake of which philosophy “is”’ (GA63, 3/2). Dasein, he then points out, does not simply name a universal that can be defined in abstraction from its particular properties. Rather, Dasein is factical in the sense that Dasein is ‘in each case this Dasein for a while at a particular time’ (GA63, 7/5). Thus, Dasein cannot be defined by its what; it cannot be the thematic object of cognition, but refers essentially to the how of a ‘factical life’ (GA63, 7/5).

In light of this ambition, it is unsurprising that Heidegger tries to incorporate an investigation of social life into his ontology. Sociality is an integral part of factical life that profoundly affects how we live both as a species and as individuals. Ontology – understood as an investigation into the fundamental structure of Dasein – must include social ontology.

In his early work, there are two distinct ways in which Heidegger tries to integrate social ontology into fundamental ontology. In the period from 1919 to 1924, he distinguishes between three different ‘worlds’ – the surrounding world, the self-world, and the with-world. He understands these ‘worlds’ partly as irreducible background conditions for factical life and partly as distinct domains of innerworldly entities. However, as I will show, he failed to clearly distinguish between these two conceptions of sociality, that is, sociality understood as a dimension of the world versus sociality as a domain within the world. Thus, his early approach to social ontology was essentially ambiguous and therefore systematically flawed. Heidegger realised this in 1924 and improved his social ontology conceptually and methodologically by accounting for the crucial difference between these two conceptions of sociality. It is the second approach that would finally find its way to the important but brief chapter on being-with in SZ.Footnote ⁶

In place of an extensive textual analysis (see, however, Knudsen Reference Knudsen2020), a few representative paragraphs will help us understand the ambiguity in Heidegger’s first approach to social ontology. In his earliest reflection on how sociality can become an object of phenomenological inquiry – a lecture course from the winter semester 1919/1920 – Heidegger illustrates the concept of the surrounding world by mentioning ‘landscapes, regions, cities, deserts’ and so on (GA58, 33). The with-world, in contrast, contains ‘parents, siblings, acquaintances, superiors, [and] teachers’ (GA58, 33). Lastly, the self-world concerns everything that I encounter based on the rhythm of my personal life (GA58, 33). According to this outline, the surrounding world refers to publicly accessible material, while the with-world covers the domain of human objects and the specific role that these humans play in the life being uncovered. The self-world covers that which is only accessible to me personally. As Christian Ferencz-Flatz says, each of the three senses of the world is ‘merely meant as a type of inner-worldly being’ (Reference Ferencz-Flatz2015, 491). Heidegger does not say so directly, yet his illustrations of the three aspects of the world reveal that he has something like three different domains in mind – the natural, the mental, and the social world.

The surrounding world, the with-world, and the self-world are thus supposed to illustrate different parts of the lived world. Yet when Heidegger tries to formulate what it is to have a world, the distinction between the structural elements disappears:

Each human being lives within a horizon of intelligibility and familiarity. This familiarity, however, is not confined to the individual alone but is carried by a ‘group of people’, who have thus made a part of the world accessible to themselves. They have a shared understanding of the tools of everyday life, for instance, and of various public institutions. This already contradicts the separation between the with-world and the surrounding world, between the natural and the social, since both of these domains are required to describe how an environment (as a collection of tools, buildings, institutions, and places) is intelligible for a group of people. The distinction seems to disappear right after it is made.

Similarly, he explains in 1920/21 that the world has three distinct aspects:

While denying that the world is an object, Heidegger nonetheless illustrates the three aspects of the world by listing various objects. The surrounding world is the domain of material things that we can encounter; that is, it is the totality of non-human objects that can become intentional content for us. The with-world is the totality of other people in their factical roles. The self-world is no longer defined as private, but now as a form of locus that anchors the lifeworld in the individual (cf. GA60, 13/10). To be sure, Heidegger emphasises that there is no hierarchy between these three worlds (GA60, 12/9) and that they cannot be sharply distinguished from each other (GA60, 11/8). In fact, he explicitly states that the methodological utility of these concepts is to characterise the ‘manner, the how, of the experiencing of those worlds’ (GA60, 12/9) rather than what is inside them.

This points to a fundamental tension in Heidegger’s thought. On the one hand, the three worlds refer to a what, and, on the other hand, they refer to a how. In Heidegger’s terms, there is a tension between content sense and relational sense (GA60, 12/9, 88/61) that characterises the entire period from 1919 to 1924. He constantly illustrates the three worlds by referring to a list or domain of entities that makes up the potential content of experience, but at the same time he maintains that he is, in fact, not interested in the content, but how it is given. This points to the conceptual and methodological shortcoming of the early approach: Heidegger simply lacks the resources to distinguish between sociality as a domain (e.g., the list of other people, social roles, and so on) and sociality as a dimension that constantly overflows any such domain thus also affecting how we experience ourselves in a ‘self-world’ or how we experience material things in the ‘surrounding world’.

John van Buren also notes how the three ‘worldly spheres’ constantly overlap. He explains this by saying that ‘[i]n the flow of experience we are always oriented primarily to one of these “worlds of caring”, while the other two are there in the background’ (van Buren 1994, 258). While this certainly makes clear why Heidegger’s examples always complicate his distinctions, van Buren’s distinction between a foreground and a background is by itself insufficient to solve the ambiguity in Heidegger’s account. The problem is that van Buren’s formulation suggests that an entity of the with-world appears in the foreground, while the entities of the surrounding world and the self-world are in the background. On this account, however, we cannot explain what makes a foregrounded entity a social entity, since the foregrounded or thematic entity is only social by virtue of backgrounded or non-thematic references to other people.

It is, therefore, necessary to say that the with-world is simultaneously in the (thematic) foreground and the (non-thematic) background – along with the backgrounded surrounding world and self-world, of course. Van Buren’s distinction between an experiential foreground and an experiential background does not, therefore, solve the ambiguity of Heidegger’s concept of the with-world. It is, in short, not radical enough in its distinction between the thematic and the non-thematic structures of human comportment, between domains within the world and dimensions of the world.

Heidegger draws this conclusion in 1924. Influenced by his simultaneous lectures on Aristotle’s practical philosophy, he gives a talk called ‘The Concept of Time’ for the Department of Theology at the University of Marburg. Here he abandons the terminology of the with-world in favour of the concept of being-with-each-other [Miteinandersein]. This new terminology indicates a change in the social ontological architecture of his thought. He now claims that being-with-each-other characterise Dasein regardless of whether other people are actually present or whether they are actually intended. In other words, our being-with-each-other is in principle independent of the entities previously used to illustrate the with-world:

This reformulation finally solves the tension by insisting that sociality, as a dimension of experience, goes beyond any domain of objects. Sociality is posited as being equiprimordial with being-in-the-world and thus given a transcendental status: When there is Da, there is also a Mit.

A little later, in a lecture course from 1925, Heidegger even comments on his earlier terminology and explains that he now realises that his focus on entities (as opposed to the world as such) was confused. The concepts of being-with [Mitsein], being-with-each-other [Miteinandersein], and fellow Dasein [Mitdasein] remedies this fault:

In this passage, Heidegger offers one formulation of his central thesis, namely, the idea that we encounter ourselves, others, and things (‘world-things’, as he calls them) on the basis of the phenomenon of the world. As he explains, he earlier tried to formulate this thesis by calling the domain of other people for ‘the with-world’. Now, however, he realises that this terminology confuses different ontological levels. The entities, with which Heidegger illustrated the surrounding world and the with-world, are given within a world. That is to say, these domains of entities are innerworldly and do not, therefore, have the ontological structure of the world. They do not have ‘worldly being’ [weltliches Sein].

Innerworldly entities are meaningful due to the place they occupy within the world. The world is the realm of intelligibility in which other Dasein or tools appear to me. Therefore, the world must necessarily be of a different ontological order than these innerworldly entities. The terminology of the with-world blurs this distinction and leads us to mistakenly believe that others, who are strictly speaking within the world, can be conceptualised as a form of world.

To avoid this confusion, Heidegger coins the term ‘being-with’ [Mitsein]. The benefit is that being-with cannot be separated from other aspects of the world. Furthermore, the concept of being-with makes clear that sociality is equiprimordial with being-in-the-world and thus not something that occurs occasionally when we encounter another Dasein. Most importantly, being-with finalises the separation between sociality as a transcendental condition that cuts across all domains and sociality as a term designating a certain type of innerworldly entities. This is, of course, not to say that Heidegger completely disregards that there is a domain of objects that is ontologically similar to ourselves. This is what he calls ‘fellow Dasein [Mitdasein]’. As he states programmatically in SZ:

Thus, his attempts to integrate sociality into phenomenological ontology lead Heidegger to claim that sociality is a dimension, not a domain, of the world.

1.4 Sociality De Facto, Sociality De Jure

This historical analysis makes clear that the object of social ontology is the shared world, and that we should be careful not to understand the term ‘world’ as referring to a domain of entities. Moreover, it displays how a recognition of the ontological difference between entities within the world and the being of the entities determined by the structure of the world widens the scope of social ontology. Before we can explain the meaning of the shared world, we must, however, clarify what exactly the world is if it is neither an entity nor a sum of entities.

Heidegger distinguishes between (1) an ordinary and (2) a phenomenological concept of world (GA24, 235f/165f). The ordinary concept of the world means something like a collection of entities. SZ calls this concept of the world categorical as opposed to existential (SZ, 64/92). It divides into two subcategories:

1. (1a) The ontic-categorical sense of world designates all particular entities taken together. This is what we usually call the universe.

2. (1b) The ontological-categorical sense of world designates a set of individual entities in terms of the essential properties that define the set, for example, the ‘world’ of mathematics in the sense of the domain of all possible mathematical objects.

The phenomenological concept of the world is existential as it describes the world as pertaining to human existence (SZ, 64/92). This is also divided into two subcategories:

1. (2a) The ontic-existential sense of the world designates ‘that “wherein” a factical Dasein as such can be said to “live”’ (SZ, 65/93). World, in this sense, refers to a realm of familiarity like when we talk of ‘the world of commerce’ or ‘the world of the native’. This world is a system of practices, instruments, and roles that are taken for granted while making our everyday dealings with each other and our environment possible.

2. (2b) The ontological-existential sense of the world is what Heidegger calls the world, in contrast to the plurality of worlds in (2a) (SZ, 86/119), or simply worldhood (SZ, 65/93). This is the manner of being of all the particular realms of familiarity designated by the ontic-existential sense of the world. It is the a priori structure that characterises all Dasein and all of Dasein’s particular worlds.

It is important to note that the two subcategories of the phenomenological concept of the world do not amount to a type/token-distinction, since the worlds (2a) are not innerworldly entities. Instead, we might say that worlds (2a) are factical modes of the world (2b) (SZ, 65/93).

Dreyfus explains the distinction between (2a) and (2b) in the following way:

This is telling for Heidegger’s general approach to a priori claims: He does not suggest that a particular proposition is apodictically true, nor does he assert that the existence of a particular entity is necessary tout court. Rather, he analyses a mode of existence from within. He starts with a concrete experience and attempts to uncover what must be the case in order for this experience to be the way it is. This means that although his starting point is always something factical – say, my experience of using a hammer – he aims to uncover something that ‘already’ or ‘a priorily’ structures this experience and others like it. The ambition is, then, not to issue de facto claims about a world in the ontic-existential sense (2a) but de jure claims that hold for all such worlds (2b).

This fourfold distinction reveals that when ‘the social world’ or ‘the with-world’ refers to the totality of social facts or the domain of other people and their social roles, the term ‘world’ is used in the ordinary ontological-categorical sense (1b). Furthermore, it suggests that it is possible to conduct a phenomenologically inspired ethnographical or sociological investigation of a specific social world (in the sense of 2a), say, the world of the Inuit, which would then investigate the particular realm of familiarity inhabited by the Inuit as a constellation of practices, rituals, instruments, social roles and so on. Yet, social ontology, as Heidegger understands it, must describe the a priori social structure of any such given world. Hence the transcendental claim, ‘the world is always the one that I share with others’ (SZ, 118/155).

The next question is of course: What exactly is the a priori structure of the world? SZ claims that the world consists of a ‘whole of references’, which is presupposed in our everyday involvement with various entities (SZ, 84/115). The hammer is meaningful only if we already understand a whole set of other things (the nail, the boards, the windows to be covered, the oncoming storm). This whole lets the hammer appear as a hammer by allowing us to encounter it within a structure of what Heidegger calls ‘in-order-to’ [Um-zu] relations. The hammer is what it is by appearing within a field of practical possibilities. It appears as a hammer in order to drive in the nail, which we do in order to board up the windows and so on.

To understand how this ‘whole of references’ functions, we need to distinguish between two types of awareness. In our dealing with things, we tend to orient ourselves towards or focus upon a single entity or a single relation between entities (e.g., the position of the hammer). We are, as I will say, thematically aware of this particular aspect of our experiential field. By ‘thematic’, I do not mean that we necessarily perceive an entity for the sake of perception alone nor that we necessarily think about it in explicit terms. Thematic awareness can be both practical, cognitive, and affective; it names simply a way of being intentionally directed towards an object. Importantly, Heidegger claims that thematic awareness presupposes a whole of references that we do not direct ourselves towards in this manner. While the hammer might be the object of my thematic awareness, I am also aware of a network of entities and relations between entities that I do not direct myself towards. Our relation to these things and relations is, hence, non-thematic. Such non-thematic entities and relations are still tacitly operative in our thematic awareness of an object. Without a tacit understanding of the oncoming storm, my comportment towards the hammer would be very different. For this reason, and in contrast to thematic awareness, we also have non-thematic awareness. The ‘whole of references’ is largely non-thematic with only a single element being the object of our thematic awareness.

One might object that ‘non-thematic awareness’ is an oxymoron – that all awareness is necessarily thematic and that the processes on which our awareness relies are not themselves matters of awareness. To see that this is not the case, we must distinguish between two different kinds of processes on which our thematic awareness might rely. On the one hand, our awareness of an object clearly depends on causal processes, for example, neurological functions. Usually, we only have a very poor understanding of our neurological functions, if any at all. On the other hand, our awareness of an object also depends on another kind of process that is distinguished from the first by the fact that we do, in fact, understand these. Let us call these hermeneutic processes. Of course, hermeneutic processes are not present to mind as this would render them instances of thematic awareness. Yet, they differ from purely causal processes, since we necessarily have a background understanding of hermeneutic processes. By this, I mean that if a hermeneutic process were to fail, we would not only cease being thematically aware of the object in question but would, instead, become thematically aware of the failed hermeneutic process itself. To use Heidegger’s formulation, the hermeneutic process, on which we non-thematically relied, would suddenly become ‘conspicuous’ (SZ, 73/102). If the nails I have chosen for my project are too short for the task at hand or if I suddenly glimpse through the window that the storm has passed, I no longer direct my awareness to the hammer in the same way as I did before; instead, my attention is drawn towards an element in the whole of references on which my previous comportment relied non-thematically. In contrast, if a purely causal process fails – say, I have a mild stroke – this also disrupts my hammering, but the failed process need not come to my attention.

For Heidegger, all understanding is holistic because any act of thematic understanding takes place within a non-thematic whole of references. Heidegger uses the term ‘signifying’ to refer to the relation between what is thematically understood and what is non-thematically presupposed in this act of understanding. The hammer is significant by virtue of its relations to the nails, the window, and so on. Heidegger then says that ‘[t]he relational whole of this signifying we call “significance”. This is what makes up the structure of the world – the structure of that wherein Dasein as such already is’ (SZ, 87/120). Clarifying the initial definition of the world as a ‘whole of reference’, we now see that the world is the relational whole of significance that makes our involvement with entities possible.

In this context, ‘whole of significance’ means both (1) that it is a whole that signifies in the sense of (tacitly and non-thematically) referring something (the hammer) to something else (the nails) and (2) that this whole is significant in the sense of being important to someone. This latter point follows from the fact that the in-order-to relations cannot go on ad infinitum. In the end, the in-order-to relations must also refer to something for-the-sake-of which these relations matter; that is, the relations of significance must ultimately refer to someone for whom they are significant.

The ‘world’ is hence neither subjective nor objective. It is not subjective as it clearly involves relations between non-mental entities. Hammering becomes meaningless if I run out of nails. But it is not objective either since the relations of significance that hold between these non-mental entities are established by the specific projects of an agent. Using a contemporary term, we can say that the in-order-to’s are affordances. Affordances are neither objective nor subjective but a way in which the environment appears to an agent based on the abilities and dispositions of this agent. The for-the-sake-of, on the other hand, indicates that not all affordances are salient at the same time. In order for an affordance to motivate me to act, it must be significant to me. This is established through the tacit or non-thematic self-understanding that accompanies an act of understanding: The act of understanding places an entity within a field of possibilities (qua the in-order-to relations or affordances) and this necessarily refers back to an agent for whom to actualise such a possibility is also to actualise a future possibility of being this or that (qua the for-the-sake-of relation).

We might say, then, that the for-the-sake-of is a commitment because my ongoing engagement in a relational whole – my ongoing attempt to be this or that – is normatively significant to me. It offers me a way of understanding myself in which I can succeed or fail. If, for instance, I engage in the project of hammering in order to board up the windows so that I can protect my family from the storm, I use the hammer for the sake of being a good father and husband. I am thus committed to the activity of hammering through a normatively significant self-awareness. I measure myself and my activities in light of being a good father and husband, and I am aware – even if I do not think about it, and even if I might not be able to specify exactly how – that I can succeed or fail in this regard. This commitment is self-referential, since it, in the end, comes down to my self-understanding. There is no further level of explanation of why I want to be a good father. It is simply a matter of who I take myself to be. This is one reason why Dasein is an entity ‘that in its being has this very being as an issue’ (SZ, 42/68).

Significance is a relation that involves two elements: (1) the environmental affordances inflected by an agent’s abilities and dispositions and (2) the agent’s self-referential commitment to a project that is furthered through some of these affordances. Based on these two elements, a whole of significance offers a set of practical possibilities to Dasein. Let us call the set of possibilities thus outlined for an existential projection. Similarly, let us call the type of selfhood defined by its participation in a whole of significance for existential selfhood.

Mark Okrent provides a classical formulation of the connection between existential projections and existential selfhood that capture some of the same features that my definition of significance does:

I agree with Okrent that the world is, at the same time, a teleological and a normative structure and that it weaves together practical utility and self-understanding. However, it is not clear how this gloss on the phenomenological concept of the world helps us understand social life. Indeed, formulations like Okrent’s seem more or less blind to the possibility of a world containing people and not just tools. Surely, we miss something crucial if we try to describe social life in terms of functional relations and existential self-understandings.

Heidegger points out that the priority given to the tool analysis is a purely didactic one (GA26, 233/181; cf. GA9, 155n55/370n59; GA29/30, 263/177) and says that it would be a ‘violent constriction of the analysis of the world’ if we were to leave out social relations (GA20, 327/237). We should, hence, be hesitant to take the ‘structure of functional relations’ to be a kind of ‘basic layer’ [Urgrund] (cf. SZ, 131/170) capable of providing the (non-social) building blocks needed to understand all features of social life, as some critics take Heidegger to do.Footnote ⁷

We need then to clarify the sense in which our being-with-others modifies the structure of being-in-the-world. How is the world, as a relational whole of significance, shared with others? An influential line of interpretation, which I will call the pragmatic conventionalist interpretation (PCI), answers this question by reference to Heidegger’s analysis of the Anyone. Dreyfus, for instance, writes

The defining feature of PCI is that it takes the sharedness of the world to be a function of our socialisation into a society, understood as a more or less well-defined and stable set of social practices. This set of social practices is prior to our explicit agreement and provides a taken-for-granted condition of possibility for our interaction with our environment and each other by defining what counts as right and wrong use of tools and right and wrong social interaction. PCI is pragmatic because it emphasises the importance of social practices rather than explicit agreements and individual beliefs; it is conventional because it takes social practices, understood as relatively stable patterns of social interaction, to determine how we comport ourselves to entities.

For Dreyfus (Reference Dreyfus1991, 155), ‘[w]hat we share’, when we share the world, ‘is simply our average comportment’. Thus, two people share the world if they comport themselves to entities according to the same typified pattern, that is, when they have a pragmatic convention in common. Drawing on Wittgenstein, he then argues that ‘once a practice has been explained by appealing to what one does, no more basic explanation is possible’ (Dreyfus Reference Dreyfus1991, 155).

Similarly, Okrent argues that

The suggestion is, hence, that the behaviour of the social groups in which we live crystallises into a world without us being aware of it. In this way, pragmatic conventions are the ‘source of intelligibility’ (Dreyfus Reference Dreyfus1991, 154ff). We should, therefore, take it quite literally, when Okrent writes that ‘I am with others insofar as my behaviour is proper within my community’ (Okrent Reference Okrent1991, 49) and that ‘[t]he others with whom I share a world are those who are like me’ (Okrent Reference Okrent1991, 48). This does not mean that I share the world with those who are ontologically like me (and, hence, that I share the world with all Dasein) but simply that I share the world with those who are sociologically like me, that is, those who belong to the same community (Okrent) or society (Dreyfus) as I do.

Although Heidegger’s analysis of the Anyone certainly plays an important role in his social ontology, this line of interpretation has several shortcomings. First, the attempt to explain the world as a function of social groups begs the question of social ontology. After all, a – if not the – central aim of social ontology is to explain how groups, communities, and societies are possible. This being the case, we cannot simply presuppose a specific social formation to be the source of intelligibility. If social ontology is a meaningful endeavour, it must be possible to somehow get behind such social formations, to conceptualise their necessary structure, and to explain how this dimension fits with the other dimensions of human life.

Dreyfus might still insist that the appeal to certain typified modes of comportment is the best form of explanation that we can hope for. This, however, is not Heidegger’s view. His central social ontological claim is not that some factical social formation – like the pragmatic convention regarding right and wrong ways of comporting oneself – constitutes the ultimate source of intelligibility. Rather, he claims that being-with is a transcendental condition – a necessary structure of all human understanding – and that this transcendental condition, in turn, makes concrete social formation such as a typified comportmental pattern possible.Footnote ⁸ The fact that being-in-the-world and being-with-one-another are equiprimordial enables ‘the various possibilities of community as well as of society’ (GA20, 333/241).

In short, we need to explain particular social formations in terms of the world rather than the other way around. Although still unfulfilled, the task is clear: Social ontology must account for the de jure relations between human beings that make various de facto social formations possible.

Second, PCI’s claim that we share the world with those that comport themselves to entities according to the same typified pattern as us leaves us wondering: What, then, do we have in common with those who are socialised into different societies? What sort of common ground or shared understanding can be reached between people that do not share the same pragmatic convention? One might worry that Heidegger’s holism – his insistence that all acts of understanding rely on a holistic and socially constituted background – leads to an untenable relativism that renders intercommunal understanding impossible. I discuss this issue in Chapter 3.

A third problem is that the conventionalist reading threatens to make a caricature of not just intercommunal understanding but all forms of social cognition. If we accept that the Anyone is the source of intelligibility, we might be unable to understand others as anything but instances of their social roles. McMullin formulates the point well: ‘The problem … is that Heidegger’s account seems to fall into the danger of viewing other Dasein merely as interchangeable representatives of the public norms and meanings through which we all pursue our particular abilities to be’ (McMullin Reference McMullin2013, 58). Although anonymous social norms certainly permeate social life, we also need an account of how we experience others as concrete and unique fellow Dasein rather than just embodiments of a public norm. I turn to this issue of interpersonal understanding in Chapter 4.

In what follows, I will argue that we need a more complex account of the world and a more sophisticated understanding of what it is to share it if we are to solve these problems. To put it briefly, PCI claims that

• Two individuals, A and B, share a world if and only if A and B comport themselves towards entities in accordance with the same whole of significance that is afforded to them by a specific (set of) social practice(s).

I, on the other hand, argue that

• An individual, A, shares the world with another individual, B, if and only if

  1. (1) A comports himself in accordance with a whole of significance that is responsive to the behaviour of B,

  2. (2) A tacitly assumes B to be capable of intending the same entities as A.

Similarly,

• Two individuals, A and B, share the world with each other symmetrically if and only if

• (1)(a) A comports himself in accordance with a whole of significance that is responsive to the behaviour of B and (b) B comports himself in accordance with a whole of significance that is responsive to the behaviour of A,

• (2)(a) A tacitly assumes B to be capable of comporting himself towards the same entities as A and (b) B tacitly assumes A to be capable of comporting himself towards the same entities as B.

Chapter 3 argues in more detail for this account of world sharing and shows why our capacity to intend the same entities is crucial. Here it should be noted that I deliberately emphasise our capacity to intend the same entities, since we should not confuse world sharing – a very basic form of sociality – with something like joint attention in which this capacity is enacted. It is also on purpose that the reliance on other people in (1) is vaguely described in terms of being ‘responsive’. I intend this to include both those kinds of dependency of which we are explicitly or thematically aware, for example, the thematic other-awareness of a face-to-face encounter, and non-thematic relations like those between strangers sitting next to each other on a bus. In Section 4.5, I discuss the possibility of asymmetric world sharing by considering human–animal interaction in which the animal, at least if we follow Heidegger, is structurally incapable of intending the same entities as us.

There is a very minimal sense of the word to share at stake here. Indeed, what we share is not something empirical that can be measured and formulated in terms of an average or a shared norm. Rather, what we share is a transcendental structure – a common condition of being a human agent or self. For Heidegger, this condition renders us fundamentally responsive to each other, always already related to each other. As a minimal kind of sharing, this allows a great degree of differentiation. We, of course, share the world with those with whom we share a personal history, a convention, a language, or a common project. But we also share the world with those with whom we do not have a personal history, who are raised according to different conventions and in different languages, and who partake in different – perhaps even opposing – projects. This differentiated notion of sharing resonates well with the German word for sharing, teilen, which means both to have something in common but also to divide.

Although somewhat overlooked by Heidegger scholars, it is not unprecedented to extrapolate such a minimal and differentiated notion of world sharing from Heidegger. In fact, it is central to both Hannah Arendt’s and Jean-Luc Nancy’s work on being-with. Nancy claims, with an ambiguity reminiscent of Heidegger’s German, that ‘we share what divides us [nous partageons ce qui nous partage]’ (Nancy Reference Nancy1999, 95). Similarly, Arendt describes her notion of the world by using the image of a table:

Rather than designating a specific social formation, the world is the common condition that relates and separates us at the same time. Only if we keep this in mind can we account for the common ground of the many different and more robust forms of interaction that make up human coexistence.