‘You Always Need at Least Two Tones to Produce a Harmonious Sound’: The Value of Arendt’s Ideas on Friendship for Thinking in Social Psychology

In this article, we focus on Arendt’s ideas about the relationship between thinking, dialogue and friendship to make the argument that friendship, although undertheorised in its relationship to thinking in social psychology, is a productive concept that captures something important about the argumentative and dialogical character of thinking (both on one’s own, and with other people). We work through Arendt’s ideas and discuss them in relation to social psychological theorising to consider how the concept of friendship can deepen our understanding and analyses of the relationalities that underpin thinking. We specify that whilst thinking in existing social psychological accounts may be read as adversarial in nature (e.g. through a focus on its oppositional character), the relationship between thinking and friendship has been an important idea underlying the perspectives presented in such works. Distinguishing between thinking as friends and thinking in groups, we suggest that there may be something special about the role of friendship in thinking. We draw out this idea by turning to Arendt, and simultaneously use the work of social psychologists to reconsider aspects of The Life of the Mind, in which thinking is mostly conceptualised as a solitary activity.


Introduction
As Zadeh et al. (this issue) described in the Introduction to this special issue, The Life of the Mind is a complicated yet compelling book that, over the course of the last 18 months, became the focus of discussions among a group of social psychologists of which we are a part. In this article, we bring to the fore one of the many themes in the three volumes that comprise the text: that of friendship, and its relationship to thinking. Perhaps unsurprising among a group of social psychologists who share interests in dialogicality, the fact that Arendt presented in her final work ideas about the 'two-in-one' in thinking was immediately of interest to us all (Marková et al., this issue).
This article was developed by two of us thinking togetheras friends and collaborators about the implications of Arendt's ideas about friendship for social psychology and vice versa. In what follows, we document our dialogues about the book, which inevitably also involved putting the book into dialogue with other books. First, we summarise the ideas on friendship that Arendt presents in The Life of the Mind, and explain how she conceptualised the activity of thinking and the self in thought, how this relates to her ideas about thinking as a dialogue that takes place between friends, and how this allowed her to develop her ideas about the moral imagination. Second, we reflect on how far the activity of thinking, as Arendt conceptualised it, is possible in contemporary societies, where opportunities for solitude are scarce. Third, we connect Arendt's ideas to those of social psychologists whose work has similarly conceived of dualities in thinking, and outline the similaritiesand differencesbetween them. Fourth, despite these differences, we retain Arendt's notion of friendship by outlining its potential in deepening social psychological insights on thinking, using empirical examples that illustrate the generative and creative potential of thinking as friends. Finally, we arrive at a social psychological perspective on friendship and its role in the social process of thinking, and address the implications of this for moral action.

Key Ideas from The Life of the Mind: Friendship in Individual Thought
In The Life of the Mind, Arendt suggests that thinking (in contrast with her definition of cognition) is an activity precisely because of what she calls the 'two-in-one' ('the duality of myself with myself', Arendt, 1978, Part I, p. 182), emphasising the importance of friendship in this relationship (-within-the-self). The 'two-in-one' is most fully articulated in Thinking, the first part of the book. Here, Arendt explains that, '[wo]man exists essentially in the plural' (Arendt, 1978, Part I, p. 185, emphasis in original), a fact that is evident in the process of thinking, which itself involves a 'withdrawal' from the 'world of appearances'; the 'common-sense world' of everyday life (see also Zittoun, this issue). For Arendt, thinking is the actualisation of consciousness in solitude, and it is through this actualisation that the duality of human existence becomes clear, although it is 'already implied in the fact and the word "consciousness," or syneidenaito know with myself' (Arendt, 1978, Part I, p. 74). What distinguishes consciousnesswhich is characterised by an inherent 'original duality or the split between me and myself' (Arendt, 1978, Part I, p. 75)and thinking is that the latter is an activity. Indeed, Arendt claims that thinking is an activity precisely because it actualises self-awareness (the 'I-am-I'): Nothing perhaps indicates more strongly that man exists essentially in the plural than that his solitude actualises his merely being conscious of himself… into a duality during the thinking activity. It is this duality of myself with myself that makes thinking a true activity, in which I am both the one who asks and the one who answers. … This questioning and answering process …is soundless and therefore so swift that its dialogical structure is somewhat difficult to detect. (Arendt, 1978, Part I, p. 185).
The idea that the 'two-in-one' is actualised through the activity of thinking has several implications for Arendt's ideas about the life of the mind. Firstly, it is through this perspective on thinking that Arendt emphasises 'difference' and 'otherness' as 'the very conditions for the existence of man's mental ego' (Arendt, 1978, Part I, p. 187): the idea that the mind is not in fact singular, but plural, and in its plurality, not several mirror images, but rather, two-in-one. Secondly, the notion of a duality actualised in thinking underpins Arendt's distinction between solitude and loneliness, the latter being the awareness of '[being] deprived of human company' (Arendt, 1978, Part I, p. 73) that leads to the experience of oneself as singular, which she claims as possibly characteristic of some statessuch as of dreaming or madnessbut not of most others. Solitude, then, is by contrast the experience of the two-in-one in thinking: the individual who is not only not lonely, but is, in fact, 'keeping [one]self company' (Arendt, 1978, Part I, p. 74). Arendt (1978) maintains that keeping one's own company through thinking is about having a particular type of dialogue, albeit 'soundless', with oneself: one that is fundamentally characterised by harmony. In order to develop this argument, she draws upon Socrates, whom she importantly acknowledges not as a philosopher (and indeed she is highly critical of that enterprise) but rather as an example of someone who thinks. She explains that: To Socrates, the duality of the two-in-one meant no more than if you want to think, you must see to it that the two partners who carry on the dialogue be in good shape, that the partners be friends. The partner who comes to life when you are alert and alone is the only one from whom you can never get awayexcept by ceasing to think. (Arendt, 1978, Part I, p. 187-188, emphasis added). Indeed, later, in Willing, the second part of The Life of the Mind, Arendt again emphasises that 'the two-in-one are friends and partners, and to keep intact this harmony is the thinking ego's foremost concern' (Arendt, 1978, Part II, p. 64; she here distinguishes between thinking and willing on this basis, see also Cabra, this issue; Marková, this issue).
Yet apprehending precisely what Arendt meant by friendship, at least in this text, is complicated. Only once does she offer a definition, and it curiously appears in Willing, not Thinking: Two men who are friends can be said to be 'independent substances' insofar as they are related to themselves; they are friends only relatively to each other. A pair of friends forms a unity, a One, insofar as they are friends: the moment the friendship ceases they are again two 'substances', independent of each other. This demonstrates that somebody or something can be a One when related only to itself and still be so related to another, so intimately bound together with it, that the two can appear as a One without changing their 'substance', losing their substantial independence and identity… The point here is that such a mutually predicated relationship can occur only among 'equals'. (Arendt, 1978, Part II, p. 98). So for Arendt, the two-in-one are 'friends and partners' (Arendt, 1978, Part II, p. 64), and, given her definition of friendship, this implies a relationship between interdependent 'equals' (Arendt, 1978, Part II, p. 98) who are bound to one another at the same time as they maintain something apart from (or independent of) their relationship. Recall that Arendt also suggested that this partner or friend is the one who 'comes to life when you are alert and alone [and] is the only one from whom you can never get away' (Arendt, 1978, Part I, p. 187-188). It is for this reason that she emphasises that the 'foremost concern' of the 'thinking ego … is to keep intact this "harmony"' (Arendt, 1978, Part II, p. 64). This significance Arendt accordsto not contradicting oneself, and not being one's own adversaryis perhaps clearest in her invocation of Socrates: even he, she says, 'has to go home where he will be alone, in solitude, in order to meet the other fellow' (Arendt, 1978, Part I, p. 190).
Arendt's emphasis on harmony and the importance of not being one's own adversary also underpins an important connection between the two-in-one, the 'Self', and morality in her work. Although she claims that 'the thinking ego is not the self' (Arendt, 1978, Part I, p. 43), and makes a clear distinction between the life of the mind and the 'world of appearances', she also uses the hypothetical example of the murderer and sufferer to suggest that it is better to suffer wrong than to be the perpetrator: 'who would want to be the friend of and have to live together with a murderer?' (Arendt, 1978, Part I, p. 188). Here, we see that there is a connection between the activity of the two-in-one in thinking and the actions of the Self in the 'world of appearances', with some authors (e.g. Kateb, 2009) describing Arendt's version of thinking as a practice that shapes a person's 'moral imagination' (see also Brinkmann's discussion on the moral landscape, this issue). Elsewhere in the text, she suggests that although thinking is not about the Self, it is concerned '…with the experiences and questions that this Self, an appearance among appearances, feels are in need of examination' (Arendt, 1978, Part II, p. 64). Although thinking, like flute-playing, 'produces no end result that will survive the activity' (Arendt, 1978, Part I, p. 123), the dialogue that one has with oneself in thinking thus appears to be in some sense connected to everyday life and action. This is despite the fact that, by its very nature, the activity of thinking involves a 'withdrawal' from the 'world of appearances', or, as Arendt puts it, a 'stop-and-think' (Arendt, 1978, Part I, p. 175), and is perhaps most evident when one considers the overarching puzzle she seeks to solve in this text, to which she refers several times throughout: namely, the psychological processes that underpin the banality of evil (see also Zadeh et al., this issue).
Arendt also elucidates the relationship between the 'world of appearances' and the life of the mind in her ideas about dialogue, in particular that the 'soundless' or 'silent dialogue' that individuals who are thinking have with themselves is made possible by the fact that they are born into a 'world of appearances'. This is the case, she claims, insofar as 'I first talk with others before I talk with myself' (Arendt, 1978, Part I, p. 189): '…discursive thought is inconceivable without words already meaningful, before a mind travels, as it were, through them' (Arendt, 1978, Part I, p. 99). Her argument here is that thinking relies on language for its activation, such that the dialogue of the two-in-one is reflective of dialogues carried out between friends in everyday life. Yet it is also that which belongs to the 'world of appearances' that disrupts the process of thinking altogether: The two-in-one become One again when the outside world intrudes upon the thinker and cuts short the thinking process. Then, when he is called by his name back into the world of appearances, where he is always one, it is as though the two into which the thinking process had split him clapped together again. (Arendt, 1978, Part I, p. 185).
The implications of this proposition are that individuals in the 'world of appearances' are not thinking. It is to the implications of Arendt's conceptualisation of thinking as a solitary affair for understanding thinking in contemporary societies that we now turn, before considering some of the points of connection between The Life of the Mind and the treatment of thinking in social psychology.

Two-in-One (Individual) Thinking in 21st Century Life
The world has changed quite significantly since Arendt's lifetime. With the rise and expansion of the internet, social life has become increasingly fast-paced and distributed, holding implications for two-in-one thinking. There is much debate about how today's technologized society impacts on individual psyches. Gergen (1991) described these societal changes as resulting in the 'saturated self', who is 'multiphrenic', exposed to such intense multiplicity that notions of a stable and authentic identity are being eroded, along with a diminishing of time and attention for interpersonal relationships. At the same time, research with marginalised communities highlights that online platforms bolster psychological wellbeing by providing new affordances for finding, building and fostering support around identity work (see e.g. Hanckel et al., 2019). Arendt's discussion of the 'two-in-one' provides an interesting addition to these debates. Through her framing, a withdrawal from the 'world of appearances' via the two-in-one in thinking is essential for the development and maintenance of the moral imagination, and friendship in this relationship is key for ensuring that we do not feel lonely in this withdrawal. Yet in today's world, opportunities for solitude are becoming increasingly scarce, and our internal dialogues, too, are arguably increasingly complex. In Arendtian terms, we are all the more susceptible to 'sleepwalking' into totalitarian regimes and global health crises. The global rise of populism (Mudde, 2007), and increasing mental health challenges worldwide (Cosgrove et al., 2020), are all too apparent.
Yet it is not simply the rise of internet technologies that underlie the challenges associated with withdrawing into the two-in-one and maintaining a friendship in this regard. In dialogical approaches within social psychology, which emphasise the human condition as one of interdependence between the Self and Others (e.g. Marková, 2016), the discussion of bureaucratisation and dissident literatures illustrates the institutionalisation of 'sleepwalking': that we are entangled in systems that deprive us of opportunities for responding to the world around us and acting in thoughtful and intentioned waysand too, the very real costs of trying to act against such systems. For example, dissidents often described becoming marginalised from their friends and families, and that their 'only reward' was having preserved their own dignity and integrity (Marková, 2016).
Bakhtin, whose dialogism is taken up in dialogical social psychology, also wrote about the human tendency to try and renounce responsibility (see Marková, 2016, p. 160)how there can be no 'alibi in being', there is no stand-in for the Self to give respite from responding to Others and 'Being-as-event', yet so many of us nevertheless try to pretend we are someone else, or otherwise operate through avoidance. This therefore implicates the everyday nature of actions that can disrupt the two-in-one 'living in peace'. Recall that Arendt wrote about extremes such as murder, but what about the smaller, everyday and more incremental actions that can factor into us loathing and avoiding ourselves, such as when we don't speak up when another person is getting discriminated against, when we lash out at others who have what we do not, when we become so consumed by our own self-image (e.g. at work, on social media) that we lose ourselves and potentially also some of our relationships, or when our work is undertaken in contexts that prioritise profit and efficiency at the cost of people's wellbeing?
The everyday nature of 'sleepwalking' and attempts to renounce our Being-as-event together highlight the work involved in thinking, in maintaining friendship to ourselves, and importantly, how context and Others are implicated in our ability to do this. We will therefore now connect Arendt's theorising on the two-in-one in individual thought with social psychological theorising on dualities in thinking as a social process.

Key Ideas from Social Psychology: Thinking as a Social Process
There are indeed some interesting parallels that can be drawn between Arendt's ideas about the two-in-one, thinking, and dialogue and the treatment of thinking in social psychological theory. We focus here on the work of Michael Billig, as an example of (antiquarian) rhetorical social psychology (1987/1996), and the work of Ivana Marková, as an example of (interactional) dialogical social psychology (2003,2016). These two authors provide us with the most in-depth treatment of thinking in social psychology from a perspective that is not only shared with Arendt (i.e. that thinking is characterised by a 'two-sidedness' -Arendt's two-in-one), but also moves beyond it, by illustrating how the duality of individual thought is also intrinsically connected to dualities in societal thinking, and in the very nature of knowledge itself. Billig and Marková, like Arendt, situate their positions on thinking in contrast to dominant, discipline-specific ideas about what it means to think, and largely by reference to works of philosophy. In that sense, their works are of a genre similar to that which Arendt chooses to communicate with: in explicit dialogue with other theorists (Zadeh et al., this issue), and in contrast to ideas about the life of the mind that they identify as characteristic of much contemporary work.
First, both scholars share with Arendt an emphasis on thinking as characterised by duality. In Billig's (1987Billig's ( /1996 rhetorical social psychology, thinking consists of logos (discourse) and anti-logos (counter-discourse): in other words, it has a two-sided property, without which it could not exist. Indeed, in this view, the possibility, if not actuality, of counter-discourse is always present in human thought, such that thought is characterised by the spirit of contradiction. This discourse, Billig clarifies, is a form of internal argument that is, first, modelled on outward dialogue, such that 'the dialogic form can be found in the writings of a single person, whose thoughts reproduce the forms of everyday argument' (Billig, 1987(Billig, /1996. Second, thinking in its duality takes not only its form but also its content from the social world, or, to put it differently, from the 'commonplaces of rhetoric' (Billig, 1987(Billig, /1996 those common-sense values and notions that are themselves characterised by contrary aspects. In fact, Billig (1987Billig ( /1996 argues that common sense is by its nature dilemmatic, such that even in circumstances where meanings appear to be shared, the 'seeds of argumentative logoi' can be found. 'The dilemmatic aspects of common sense…', he writes, '...might fill our minds with the controversial things which make much thought and argument possible' (Billig, 1987(Billig, /1996). Marková's (2003) work similarly specifies the role of antinomies, the basic characteristic of human thought in its oppositional nature, in thinking. Through her development of the concept of themata (Holton, 1975;Moscovici & Vignaux, 2000), she outlines the conditions under which the antinomies implicit in common sense thinking are made explicit, and therefore give rise to social representations: 'in the course of certain social and historical events, e.g. political, economic, religious and so on, they turn into problems and become the focus of social attention and a source of tension and conflict' (Marková, 2003, p. 184). These tensions and conflicts may, per Billig, be seen as the 'dilemmatic' aspects of common sense. However, Marková also specifies that such antinomies are an expression of the dialogical mind (Marková, 2003(Marková, , 2016. As outlined above, dialogical approaches take Self/Other interdependencies as foundational to the human condition. This perspective is perhaps more implicit in Billig's (1987Billig's ( /1996 work, which concerns itself with the Other in the expression of attitudes (e.g. in taking or rejecting the side of the Other in discourse). On the other hand, Billig, like Marková, perceives the activity of thinking in terms of Others: he specifies that 'we think to ourselves, as if addressing someone else' (Billig, 1987(Billig, /1996, and describes the 'bonds of association [that] are necessary for … dialogue to continue' (Billig, 1987(Billig, /1996, thus alluding to interdependencies within and between speakers. In that sense, his work may be said to share with the dialogical perspective a sense that 'antinomies in language and communication...are hiding the dialogicality of the mind' (Marková, 2003, p. 85), 1 and in that sense, both scholars share with Arendt an understanding of thinking as characterised by an interdependent two-in-one.
It is worth clarifying that the relationship between the two-in-one in all three authors' works is characterised by tension, and crucially, not fusion. Above, we saw that Arendt perceives 'difference' and 'otherness' as 'the very conditions for the existence of man's mental ego' (Arendt, 1978, Part I, p. 187). We also saw that antinomies reveal the dialogical capacities of the human mind, and so the ontology of the Ego/Alter which is the 'basic' (Marková, 2003) or 'epistemological' (Marková, 2015) thema upon which all other antinomies are based. The Ego/Alter relationship is clarified as one that is characterised by tension through mutual interdependence and autonomy, contra cases in which the Self desires to fuse with the Other, documented in psychoanalytic reports (Marková, 2016, p.133). In Billig's (1987Billig's ( /1996 work, tension is most clearly articulated in relation to the importance of negation to rhetorical thinking, specifically in the conflict between categorisation (i.e. to make something general) and particularisation (i.e. to treat something as special) that he describes as the most basic of rhetorical strategies. However, one can also identify in Billig's work an emphasis on the significance of difference in dialogue, insofar as he makes clear that sameness through 'agreement' prevents or prohibits dialogue: 'When speakers merely agree with each other, there is nothing left to say. The result is happy, wordless smiles' (Billig, 1987(Billig, /1996). This nothingness is also a prospect contemplated by Arendt (1978, Part II, p. 70), who in distinguishing thinking from willing argued that, '...for the thinking ego a 'healing' of the split [of the two-in-one] would be the worst thing that could happen; it would put an end to thinking altogether'.
Yet there are also important differences in what is meant by difference within the twoin-one in Arendt's work, and difference in dialogical and rhetorical perspectives. For Arendt, the duality of the two-in-one is activated in thinking and relates to there being a two-in-one in terms of inner dialogue only (i.e. that there is one who questions, and one who answers). For Marková and for Billig, it is also the case that thoughts themselves are dual in nature, or rather, that knowledge itself has a two-sided property that manifests as a clash of perspectives in dialogues with ourselves and dialogues with others. 2 Similarly, there are no claims to the resolution of difference or duality in dialogical or rhetorical approaches (i.e. that the two-in-one are in the social world 'clapped together again', as Arendt claims in The Life of the Mind), and there is no such 'world of appearances' from which one withdraws to think. Moreover, unlike Arendt's withdrawal from the 'world of appearances' in which the thinking activity takes place, implying a stasis or absence in thought in everyday life, the social psychological theories under consideration take as their starting point that the social world, like the life of the mind, is dynamic. In this sense, both Billig and Marková maintain a continuity of perspective in theorising about the person who dialogues (and so thinks) alone and the person who dialogues (and so thinks) with others. We see this most clearly in Marková's (2003Marková's ( , 2016 theorisation of the dialogical capacities of the human mind, which are fundamentally oriented towards the Other (such that the very notion of being alone is something of a misnomer), and in Billig's emphasis on the connection between individual and social dialogues, where, through Protagoras, he suggests that the former (individual 'reasoned discourse') is connected to the latter (social 'dispute[s] between people'; Billig, 1987Billig, /1996. For Billig, this is crucially because the spirit of contradiction flows through both individual and social dialogues (or 'arguments'), such that the former are always either in actuality, or potentiality, a part of the latter.
When we come to consider the manner in which such dialogues are discussed in dialogical and rhetorical social psychology, however, we begin to see the value of Arendt's ideas about the relationship between thinking, dialogue and friendship to these works. Indeed, thinking in such social psychological accounts may be read as adversarial in nature. Marková (2003, p. 81), for example, defines dialogue as 'a communication in which the co-authors dispute, fight about ideas and negotiate their antinomies in thinking', while Billig (1987Billig ( /1996) talks about dialogue as comprised of 'the continual struggle for argumentative victory against rival thinkers'. There is, as we have suggested, a 'clash' of perspectives. It would therefore be fairly straightforward to take from these accounts a view of individual and social thinking as adversarial. But oppositional does not mean adversarial. In fact, in the second edition to Arguing and Thinking, Billig (1987Billig ( / 1996 explains that this was a misdirected critique of the first edition of his book. Importantly, in the second edition, he clarifies that what he means by arguing is not quarrelling or a discursive battle but rather 'reasoned discourse' (he states that these are the two senses in which 'arguing' is used in the English language, which may account for some of the misunderstandings of his ideas). He goes on to say that perhaps he meant something close to (Bakhtin, 1986) dialogue in his notion of argumentation: 'In dialogue, the destruction of the opponent also destroys the very dialogic sphere where the word lives' (Bakhtin, 1986, p. 150). In so doing, Billig not only brings rhetorical social psychology in line with dialogical social psychology in terms of a shared intellectual ancestry, but also draws upon an entire new literature with which to understand his original work. Consider, for example, the parallels between Bakhtin's ideas and Billig's discussion of 'argument' as non-communication, as it appears in the first edition: When two people are said to be having an argument -that is, they are ill-disposed to each other -they might, in point of fact, be refusing to communicate by argumentation. (Billig, 1987(Billig, /1996 And in the second edition, on the purpose of communication: The goal of communication is not accord, as if all voices should seek to be similar, and, being similar, can then stay happily silent. The image is one of chatter and discussion. Any accord which is reached is to be breached: one companion, with a playful smile, will say, "But, on the other hand…". All will start discussing without constraint, enjoying the continuation of a dialogue which moves forward creatively and endlessly. (Billig, 1997, p. 17).
Billig here seems to be suggesting that 'if you want to think, you must see to it that the two partners who carry on the dialogue be in good shape, that the partners be friends' (Arendt, 1978, Part I, p. 187-188). Although the notion of 'companionship' is perhaps more implicit in Marková's (2016) work, it too features in a brief discussion of forms of trust and distrust, and the different forms of trust that may arise in relation to different Others (among them, friends). Moreover, the vision of dialogues as open-ended, and the idea that they 'move forward creatively and endlessly' (Billig, 1997, p.17), are significant features of Marková's dialogical perspective. That creativity and unfinalisability may thus be two components of dialogue among friends is an idea to which we shall later return.
At this point, it is evident that the companions or friends sketched from rhetorical and dialogical accounts in social psychology are not one and the same as the 'friends' of Arendt's The Life of the Mind who emerge in the individual activity of thinking and last only as long as it does. But can Arendt's notion of friendship, extended to the social context, nevertheless deepen our understanding and analyses of the relationalities that underpin the social psychological process of thinking? Do social accounts of thinking require that the interlocutors be friends? What are the implications of thinking as friends? We turn to address these questions next.

Arendt and Others: Ideas on Friendship in Social Thinking
While we have emphasised in this article the focus of The Life of the Mind on thinking and friendship in solitude, it should be reiterated that thinking is not, in Arendt's final work, depicted as entirely separate from public life. Neither did Arendt, in other writings, restrict herself to a wholly individual understanding of friendship. In relation to thinking, some of Arendt's ideas in her final work suggest a complex relationship between the social world and the life of the mind (e.g. in terms of thinking being reliant upon language for its activation, and the 'stuff' of thought being directed by the Self's experiences). On the other hand, Arendt was not a social psychologist, and it was not her intention to develop in this work a theory of the relationship between individuals and society from which an account of thinking could then be developed (see also Zadeh et al., this issue). Nevertheless, in her identification of 'boundary situations' in which thinking 'ceases to be a politically marginal activity' (Arendt, 1978, Part I, p. 192), she reflects on the implications of thinking for public life (or, put differently, on the relationship between activity and action). In so doing, Arendt offers an account of thinking that here appears to resemble Marková's (2003) ideas about thematisation and the genesis of social representations. Arendt explains that, When everybody is swept away unthinkingly by what everybody else does and believes in, those who think are drawn out of hiding because their refusal to join in is conspicuous and thereby becomes a kind of action. In such emergencies, the purging component of thinking (Socrates' midwifery, which brings out the implications of unexamined opinions and thereby destroys them -values, doctrines, theories, and even convictions) is political by implication. (Ibid., emphasis added).
Arendt also theorises a connection between the faculties of thinking and judging, explaining that the latter is the 'by-product of the liberating effect of thinking [that] realises thinking, makes it manifest in the world of appearances… [as] the ability to tell right from wrong' (Arendt, 1978, Part I, p. 193). 3 This claim is perhaps the point in The Life of the Mind at which the 'moral imagination' is most fully articulated, indicating the social implications of thinking (as an individual processsee Benhabib, 1988, for further discussion on this point).
Yet in Willing, she claims that: Action, in which a We is always engaged in changing our common world, stands in the sharpest possible opposition to the solitary business of thought, which operates in a dialogue between me and myself. Under exceptionally propitious circumstances that dialogue, we have seen, can be extended to another insofar as a friend is, as Aristotle said, another self. But it can never reach the We, the true plural of action. (An error rather prevalent among modern philosophers … is to believe that the intimacy of the dialogue, the 'inner action' in which I 'appeal ' to myself or to the 'other self', Aristotle's friend, Jaspers' beloved, Buber's Thou, can be extended and become paradigmatic for the political sphere). (Arendt, 1978, Part II, p. 200).
Despite having developed an account of the relationship between thinking and judging as a bridge between the individual and society, then, in Willing Arendt (1978) stops short of extending her concept of friendship (here 'intimacy') in thinking to the public sphere (the 'We' of action in the 'world of appearances').
There are nevertheless several other works, written long before The Life of the Mind, in which Arendt articulates the importance of friendship to the political sphere. One such example is in her acceptance of the Lessing Prize of the Free City of Hamburg, a speech she delivered in 1959 that was published in 1968. Drawing on Greek philosophy, Arendt there argued that 'the discourse of friendship, philanthropia, "love of [wo]man"', is essential for ensuring the humanness of the world, in that humanity 'manifests itself in a readiness to share the world with other [wo]men', (Arendt, 1968, p. 25) through discourse and debate. 4 She claimed that Lessing's 'greatness' lay in his understanding of the significance of 'the unending discourse among [wo]men' (Arendt, 1968, p. 27) to human experience, and maintained that without this discourse across difference, the world would become inhuman and stagnant. In fact, she argued in this work that 'solitude', characterised by an absence of dialogue between 'many voices': might result that all [wo]men would suddenly unite in a single opinion, so that out of many opinions one would emerge, as though not [wo]men in their infinite plurality but [wo]man in the singular, one species and its exemplars, were to inhabit the earth. Should that happen, the world, which can form only in the interspaces between [wo]men in all their variety, would vanish altogether. (Arendt, 1968, p. 31). Arendt (1968) thus seems to have suggested that unending dialogue between different interlocutors who are friends is characteristic of public life: it is what makes the world human. In some ways, Arendt's biography provides several good examples of this type of dialogue across differencesin her friendships with Martin Heidegger, Karl Jaspers and Mary McCarthy, for instance (Nixon, 2015). Yet curiously, Arendt's life also documents her resistance to legislative and social changes that would have fostered these sort of dialogues, most notably in terms of her responses to the desegregation of Little Rock schools and to the growing civil rights movement in the American south (see Gines, 2014). Scholars such as Owens (2017) have interpreted these biographical facts as part of Arendt's lack of understanding of African-American experiences of discrimination within their colonial context, and her refusal to see imperialism, in her words 'the one great crime in which America was never involved' (Arendt, 1958, p. 46), beyond the European nations. It would, in our view, be remiss to dedicate a special issue to Arendt's workwhich also takes into account her biographywithout drawing attention to the anti-Black racism that is present in her writings (e.g. Reflections, On Violence), particularly given that this fact has often been neglected in the existing scholarship on her ideas.
While interpreting the meaning of 'dialogue across differences' in Arendt's life and work is therefore not without its challenges, we propose that the concept of 'friendship', rather than, say, people in dialogue, can deepen our understanding of thinking as a social process, and that this concept particularly highlights the generativity of dialogue across differences. Throughout history there are numerous examples of groups of friends who have sparked creativity in thinking and changed the way that we look at the world. Goodwin (2011), for instance, described that the Bloomsbury Group 'looked like a large sandbox in which a collection of brilliant people amused themselves with their friends in off hours. And to some degree that is what it was', (Goodwin, 2011, p. 65), but that it also became akin to one of the major policy think tanks of the early 20th century, innovative in its intersticing of the sciences and social sciences with the humanities and the arts. White and Peters (2017) have suggested that Bakhtin's Circle, which included Ivan Sollertinsky, Valentin Voloshinov, and Pavel Medvedev, was also closely connected to and in dialogue with a wide range of scholars and artists located in the lively cultural small town of Vitebsk, who all shared a 'certain unanimity' in their desire to explore philosophical positions on creativity and understanding. Accounts of anti-colonial movements also highlight that the sometimes fragile friendships between individuals from both coloniser and colonised groups were significant to the fracturing of colonial power (e.g. see Gandhi, 2006;Gopal, 2019).
These examples illustrate the creative potential of thinking as friends, but they also raise questions about what makes 'thinking friendships in the world [of appearances]' distinct from, say, thinking as a social group. Dialogical theorising on the dynamic coconstruction of knowledge foregrounds mutuality with asymmetries in which interlocutors reciprocally 'attune' (Rommetveit, 1992) to one another in dialogue, yet still maintain their unique positions and differences (Marková & Foppa, 1991). We pose the question of whether there is something about friendship that ignites or catalyses such attunement, or, in Arendtian terms, creates the many 'tones that produce a harmonious sound'. The answer to this question depends on how friendship is defined, with specific distinctions relating to ideas about Self-Other differences and how these are understood and negotiated.
What is friendship? In Ancient philosophy, Aristotle's reflections are perhaps the most widely known, drawn upon by Arendt and indeed other scholars such as Hans-Georg Gadamer. Here, 'being in common', and the self-knowledge that can be achieved through this, is foregrounded: Via another, a person becomes one with himself. The other, the friend means much to the person, not because of the person's need or lack, but for the sake of his own self-fulfillment… each see a model in the other… they understand one another by reference to what they have in common and so succeed in reciprocal co-perception (Gadamer, 1999, p. 138-9).
Other scholars have described this Aristotelian conception of friendship as 'nativist', holding homophilic loyalties (Gandhi, 2006), and as a 'mirror view' of friendship in terms of its narcissistic elements (Cocking & Kennett, 1998). Another common conception is what Cocking and Kennett (1998) describe as the 'secrets view' of friendship, which centres upon the importance of self-disclosure for cementing bonds of trust and intimacy. However, these scholars also argue that secrets alone do not sufficiently capture the quality of friendship or its implications for the Self, as self-disclosure is also an aspect of other relationships, such as those we may have with a priest, or a therapist (Ibid).
Instead, Cocking and Kennett (1998) call for a 'drawing account' of friendship in which the Self is understood as relational (different from the 'mirror' and 'secrets' views of friendship that construct the Self as rather static and discrete), reflecting an openness and curiosity towards being directed by the interests of the Other (precisely because they are the interests of the Other), as well as being interpreted through the gaze of this otherness. Such a 'love for' otherness is also reflected in Gandhi's (2006) discussion on the Epicurean conceptualisation of friendship, which is philoxenic, relying on one's capacity to leave oneself open to guests, strangers, and foreigners. In Gandhi's (2006) analysis, the Epicurean conception holds an element of risk that she argues is crucial for nurturing change and transformation, namely 'the inventiveness of friendship': She is ever willing to risk becoming strange or guestlike in her own domain, whether this be home, nation, community, race, gender, sex, skin, or species… at this time in world politics, when our solidarities simply cannot be fixed in advance (America or Iraq; liberalism or the agonism of the particular; citizen or refugee), a utopian mentality shows the way forward to a genuine cosmopolitanism: always open to the risky arrival of those not quite, not yet, covered by the privileges which secure our identity and keep us safe (Gandhi, 2006, p. 31).
Yet in the field of social psychology, less has been said about friendship as a concept. It may be that this is partly because of the (masculinist and other) assumptions that permeate the discipline (see e.g. Squire, 1989), with assumptions of friendship as a 'feminine enterprise', and/or because of assumptions about friendship as being conceptually woolly. Important contributions have considered children's friendships, particularly across identity-based differences (Iqbal et al., 2017), and the relationship between friendship and dialogue among pairs of friends and non-friends (Miell & MacDonald, 2001). Specifically, Miell and MacDonald's (2001) research has shown that communication among friends is characterised by mutual engagement with the Other (i.e. responding in a way that acknowledges the communication partner's ideas, and responding to their questions), while communication among non-friends is less likely to include these features. In other work, we have seen that asymmetries in status among children (e.g. according to gender) may limit or extend the scope for innovation in argumentation, and so the development of knowledge (Leman & Duveen, 1999;Psaltis, 2015).
We propose that thinking as friends, both on our own, and with others, reflects the dialogical and rhetorical capacities of the human mind. Friendships revolve around openness, and a curiosity towards being directed by the interests of the Other, as well as being interpreted through the gaze of this otherness (Cocking & Kennett's [1998] 'drawing' account). As such, there are fundamental differences between dialoguing as friends and dialoguing with less familiar Others. We do not here intend to suggest that the latter is always non-dialogical. We have seen, for example, in teacher/student dialogues, the significance of the human capacity to trust/distrust the Other in epistemic terms (Marková, 2016). 5 We are also all too aware of examples of dialogues that are characterised by mis-recognition or non-recognition that leads to a breakdown of communication (Gillespie, 2008). Crucially, such dialoguesin which speakers do not try to listen to or hear one anothercannot take place among friends.
Friendship is the continued commitment to the Other in their wholeness and uniqueness (Lugones & Rosezelle, 1995), and it involves deliberately embracing the 'Otherness' of the Other (Gandhi, 2006) in an open-ended collective pursuit of something that is important to at least one person. 6 The various groups detailed in this article (e.g. the Bloomsbury Group, Bakhtin's Circle, friendships within anti-colonial movements, and to some extent our own group) are but a few examples of such friendships, through which otherness within collectives is embracedand actively nurtured. Friendships across differences may additionally have implications not only for creativity and generativity in thought, but also for the moral imagination, which scholars have argued is cultivated in contexts where 'the self-centered perspective of the individual is constantly challenged by the multiplicity and diversity of perspectives that constitute public life' (Benhabib, 1988, p. 48). In Benhabib's (1988, p. 44) reading, it is noteworthy that 'Arendt emphasised harmony as the morally relevant experience, but she regarded plurality as the political principle par excellence'. Whether or not the unfinished third volume of her work, on Judging, would have served to more cohesively bridge her ideas on the relationship between the two-in-one in thinking and the moral imagination can only be wondered. From a social psychological perspective, it seems nevertheless appropriate to emphasise that the pluralities in two-in-one thinkingthat Arendt herself acknowledgedare intrinsically grounded in the extent to which we recognise, appreciate, and engage with diversityamong friends in particular, and in the social world in general.

Conclusions
In this article, we have used the relationship between thinking, dialogue and friendship articulated by Arendt in The Life of the Mind to reflect upon what the concept of friendship might mean for social psychological accounts of thinking. Our reflections have, first, enabled us to clarify what is (and is not) intended by the emphasis on the 'oppositional' character of thinking in social psychology, which we have explained to refer to thinking as it is characterised by different, but by no means warring, perspectives. It is possibly owing to the adversarial character that has been 'read into' such accounts that scholars have generally overlooked the role of friendship in thinking. By using and expanding on Arendt's (1978) ideas, we have begun to develop an account of friendship that reflects the rhetorical and dialogical character of thinking, and have proposed that there are some key differences between thinking as friends and thinking with less familiar others. Our definition of friendshipas a continued commitment to the Other in their wholeness and uniqueness in an open-ended collective pursuit of something that is important to at least one personshares clear connections to friendship as it is defined by feminist scholars of colour Lugones and Rosezelle (1995), who also take difference or duality as a starting point, and maintain that friendship entails an open-ended end-point that is guided and practiced through a concern for the particularity of the friend: Friendship is a kind of practical love that commits one to perceptual changes in the knowledge of other persons. The commitment is there because understanding the other is central to the possibility of loving the other person practically. Practical love is an emotion that involves a commitment to make decisions or act in ways that take the well-being of the other person into account. … A commitment to perceptual change is central to the possibility of bonding across differences and the commitment is part of friendship. (Lugones & Rosezelle, 1995, p. 141). 7 In this article, we have identified different groups of friends who may indeed be seen as examples of people 'bonding across differences', and have suggested that friendships by definition hold and even nurture differences at the same time as they catalyse the 'attunement' of Selves to, and with Other(s). Friendships are thus a generative ground for thinking. At the same time, one does not need to look very far to identify the innumerable examples of non-friends, or of communications that break down because differences are held up as important to dispute, rather than held, in contemporary societies. This is because, as Lugones and Rosezelle (1995) have forcefully argued, 'friendship across positions of inequality has to be worked for… [one needs to dislodge] the centrality of one's position in the racist, ethnocentric, capitalist, patriarchal state in one's own selfconcept'.
Arendt's conceptualisation of the two-in-one in thinking provides an interesting perspective on how we might 'dislodge' such self-centred perspectives and instead learn to become friends, establishing harmony across differences in knowledges, values and experiences. Nevertheless, our ideas in this article also suggest that a more contextualised framing of the two-in-one than Arendt undertook is necessary, insofar as our capacities to recognise otherness also depends on the dynamics and 'genre' of our interactions with diverse others (Gillespie, 2012), which are continually constrained by oppressive structures and forces. Such insights reflect what has been our greatest translational challenge in bringing The Life of the Mind into dialogue with social psychology, in that our understanding of 'the social' is inherently different to Arendt's own structuring of public life. In The Human Condition (1958), for instance, Arendt warned against 'the rise of the social', conceptually separating it from the political sphere. Arendt (1958) viewed this rise as having resulted in the public realm being taken over by those activities formerly restricted to the private sphere, owing to the expansion of the economy (i.e. the commodification of values in which consumption is prioritised). This bounded understanding of politics and (economic) sociality makes it difficult to unpack the connections between these 'spheres', and therefore complicates our social psychological reading of The Life of the Mind, given our understanding of social positioning as inherently political, and ultimately implicated in broader contextualised struggles for social justice.
We have nevertheless shown that Arendt's ideas on friendship in thinking are valuable for social psychologists, particularly among those whose work may have implicitlybut heretofore not explicitlyaddressed the concept. In conclusion, we propose that friendship may be best understood as a dialogical activity that, at its core, can explain how individuals and collectives make sense of, ethically engage with, and creatively build on differences in the social process of thinking.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/ or publication of this article. reunited through the particulars by which we judge; and that judging and acting are joined by this principle of humanitas: 'We must act and judge in ways that do not violate the actually existing solidarity of mankind. "A secret trust in man, in the humanitas of the human race" animates action and judgement. Trust in humanitas is another way of saying trust in the love men have for meaning [i.e. that they think], the love they have for the existence of things and people, and the communicative pleasure they take in reflecting on those things and people' (Young-Bruehl, 1982, p.302). 4. Other scholars have identified a similar focus on dialogue among friends as the basis of Arendt's political vision (see, for example, Singer's (2017) discussion on The Promise of Politics, which Arendt wrote in the latter half of the 1950s). 5. 'Epistemic trust … takes two mutually interrelated forms. First, it refers to the participants' trust that they live in a temporarily shared social world comprising a common ground for understanding and interpretation of their social reality. … Second, epistemic trust refers to the capacity and readiness of Selves and Others to learn, share and accept knowledge, experience, ethical evaluations or otherwise from one another' (Marková, 2016, p.127). 6. It is perhaps noteworthy in this regard that our dialogue with Arendt, with each other, and with the other authors of this special issue, was ignited by Tania Zittoun's interest in us jointly pursuing The Life of the Mind (a book that we took up precisely because of Tania's interest in it). 7. Arendt wrote to James Baldwin to 'raise objections' about him associating love, unlike friendship, with the political: 'love, in distinction from friendship, is killed, or rather extinguished the moment it is displayed in public… Because of its inherent worldlessness, love can only become false and perverted when it is used for political purposes such as change or salvation of the world' (Arendt, cited in Gines, 2014, p.5). Nevertheless, we would suggest that the practicality of the love discussed by Lugones and Rosezelle (1995) indicates a qualitatively different form of love to that which Arendt talks about. James Baldwin's conceptualisation of love is also different, and was ultimately misunderstood by Arendt (see Butorac, 2018).