Gestalt structures in multi-person intersubjectivity

In this paper I argue that there are gestalt principles underlying intersubjective interactions and that this means that intersubjective ‘units’, can be recognised as unified gestalt wholes. The nub of the claim is that interactions within a ‘plural subject’ can be perceived by others outside this plural subject. Framed from the first-person perspective: I am able to recognise intersubjective interactions between multiple others who are not me. I argue that the terminology of gestalt structures is helpful in framing and understanding the non-reducible make-up of these relational units. I consequently defend the legitimacy of the claim that we can attend to more than one other person at once, holding multiple others as a single focus of attention insofar as we can attend to multiple others as a gestalt whole. I argue that it is therefore legitimate to talk about attending to, perceiving and addressing multiple others at the same time, in the second-person plural. I argue that this can be identified in the phenomenology of such interactions and in an analysis of the core underlying structures of these interactions.


Introduction
Gestalt theory and phenomenology are historically and conceptually intertwined, with both methodologies genealogically rooted in Franz Brentano's interest in the perception of objects and events not just as loose bundles but as unified 'wholes' (Grossman 1984, pp. 41-59;Heinämaa 2009, pp. 268-273). Both are premised on the insight that subjectivity, as embodied and embedded in its environment, 'is not a concealed interiority, but an open world relation' (Zahavi 2008, pp. 664-665). While the relationship between the schools of thought has been historically complex, with critiques issued on both sides [e.g. Husserl 2012, p. 45;Koffka 1935, p. 73)], I will focus here on ways that the methodologies can be employed in overlapping and complementary ways, demonstrating this by example. 1 There are various debates in phenomenology where the language of gestalt theory has been largely absent. One such debate, in which this paper is situated, is in the phenomenology of intersubjectivity. My paper thus has a dual hope: it purports to be valuable for its own sake in its contribution to the philosophy of intersubjectivity, but it also functions as an example of how phenomenological and gestalt approaches can mutually complement one another, towards a non-reductive analysis of a specific phenomenon.
In this paper I specifically argue that there are gestalt principles underlying intersubjective interactions and that this means that intersubjective 'units', such as 'plural subjects', 2 can be recognised as unified 'wholes'. The nub of my claim is that intersubjective interactions between multiple subjects can be perceived by others, a claim relevant to both detached observation and engaged participation of and with these interactions. Framed from the first-person perspective: I am able to recognise intersubjective interactions between multiple others who are not me. I argue here that the terminology of gestalt structure is helpful in framing and understanding the nonreducible make-up of these 'plural subjects' or other relational units: intersubjective awareness between subjects has a set of basic 'shapes' or structures that can be understood as gestalt.
In this paper I hence defend the legitimacy of the claim that we can attend to more than one other person at once, holding multiple others as a single focus of attention insofar as we can attend to multiple others as a unified whole. I will argue that it is legitimate to talk about attending to, perceiving and addressing (as a participant as well as an observer) multiple others simultaneously as subjects-together with a gestalt form. I argue that this can be identified in the phenomenology of such interactions and in an analysis of the core underlying structures of these interactions. This is underexplored terrain, in part because little has been done to apply gestalt principles to the metaphysics of intersubjectivity, but also because analysing the perception of relationships is distinct from the task of analysing (i) the perception of other subjects or (ii) perception as part of a group. Far more work needs to be done, I claim, on what it is like to recognise and interact not just with another person, but with another relationship. This work on the recognition of others-in-relation as gestalt wholes can be applied to an observer's recognition of a plural subject-that is, to recognition from the third-person perspective-but much more interestingly, I argue, it opens up space for thinking about more complex forms of intersubjectivity proper, namely how a subject might engage a plural subject in a genuinely second-person plural address, something underexplored in the philosophy of intersubjectivity.
The need for this analysis is symptomatic of the fact that, to date, the philosophy of intersubjectivity has focused on either the second-person singular or the first-person plural (as contrasted with the third-person standpoint, singular or plural), but nowhere do we find an in-depth analysis of the second-person plural. This is part of a more general trend in the philosophy of intersubjectivity which has been slow to look at forms of at multi-person interaction (that is, interactions that involve more than two subjects) beyond the first-person plural. I focus here on the possibility of secondpersonal plural address, but there are a whole range of multi-person structures that the exclusive focus on the 'we' and the 'I-you' are apt to miss. 3 As part of thinking about what is involved in a robust account of the secondperson plural I will first outline some of the relevant background assumptions about the nature of (different types of) intersubjectivity and will identify in more detail this gap in the literature. I will then lay out some of the basic principles of gestalt theory and will identify aspects I take to be relevant to my analysis. Having laid out these presuppositions I make the case that multi-person intersubjectivity, specifically the second-person plural, can be understood to exhibit gestalt structure and gestalt qualities by looking at the phenomenology and the structure of specific interactions.

Multi-person intersubjectivity: situating the discussion
In this first section I will lay out some of the basic conditions of an intersubjective interaction and will outline the two main varieties of intersubjectivity discussed in the literature, namely the first-person plural and the second-person singular. Having done this I will lay out examples of multi-person interactions which cannot be fully understood in terms of either of these forms of intersubjectivity alone, and I will argue that non-reductive analyses of these kinds of multi-person phenomena, such as the second-person plural, are needed.
'Intersubjectivity' refers to the relations and sets of relations between subjects qua subjects. These are not just spatial or temporal relations as may exist between objects, but are relations of mutual awareness of one another as subjects (Gallagher 2008;Reddy 2010;Zahavi 2014). It is possible for me to perceive the subjectivity of another without this being mutual, such as when I perceive someone getting frustrated with a parking meter, without their being aware that I am there, for example, where my standpoint towards them is third-personal. However, for a relation to be properly intersubjective, on my understanding, it must involve mutual recognition of one another as subjects, and this recognition, in Naomi Eilan's phrase, must be 'out in the open' between us, 'manifest to both participants'. (Eilan 2005, p. 1. See also Gilbert 1992;Schutz 1972, p. 168). Two people may perceive each other as subjects but not yet be aware that the other is aware of them, meaning that the recognition is not yet mutual. Imagine two people who know each other in a crowded room of others. At different points they notice each other across the room but they are not sure if the other has seen them. Eventually their attempts to catch the eye of the other are successfully co-ordinated-their mutual awareness is now 'out in the open' between them. 4 Dyadic intersubjectivity, then, has the following formal structure: There are a number of different structures or types of intersubjectivity which exhibit the structures of mutual awareness outlined above. The literature has essentially focused on just two types of intersubjectivity, namely 'I-you' intersubjectivity and 'we' intersubjectivity.
Reciprocal intersubjective awareness takes a particular 'shape' in 'I-you' intersubjectivity'. This kind of interaction is characterised structurally by a direct, 'frontal confrontation' (Zahavi 2001, p. 157), which involves 'mutual attention' (Reddy 2005, pp. 85-109) or 'mutual recognition' (Gilbert 1992, pp. 217-219) between a pair. In this configuration one other person is the focus of my attention, and I am the focus of theirs, I am subject to 'the look' of the other (Sartre 2003, pp. 276-326). This orientation is said to be 'face-to-face,' or en face (Levinas 1994), with the quality of directed and undivided focus towards one another crucial to this standpoint. In such an encounter we take up what we would call the second-person (singular) standpoint: in this configuration I am oriented towards this one other person such that I can intelligibly address them by the second-person pronoun 'you', and they reciprocally are able to address me with a 'you'. This is the standpoint we take up in most forms of everyday conversation, characterised by the exchange of looks and a 'turn-taking' structure of dialogue (Stawarska 2009, pp. 105-114). Developmental, psychological, philosophical and moral aspects of the second-person singular have been discussed extensively in the literature (Thompson 2001;Darwall 2009;Schilbach et al. 2013;Eilan 2014).
By contrast, the structure or shape of 'we' intersubjectivity has all participant subjects' attention focused towards a common 'object' (which might be a physical object, a fact, an event, or even the relationship itself) with some peripheral awareness that the other participants are also attending to the same object or event. This basic structure is captured by the much-discussed phenomenon of joint attention (Eilan 2005;Cavell 2006;Seemann 2012). This basic subject-subject-object structure is foundation of more complex varieties of collective intentionality and joint action. This basic structure can be identified where an infant and her caregiver watch a squirrel together in the park, or in the more complex joint action where two people carry a table down some stairs together. Crucially this mutual peripheral awareness of the other changes the phenomenological structure of the experience for both parties. Watching a squirrel 'together' with someone is different to a case in which both subjects attend to the squirrel simultaneously, but are not aware of each other's awareness, so do not experience this as something we are 'doing together.' In such a configuration someone 'comes to share a focus of attention with someone else.' (Hobson 2005, p. 185). This awareness of 'sharedness' or 'jointness' transforms how the world (object, event, etc.) is experienced. To be part of a 'we' is to experience a shared world. 5 The basic definitions and distinctions between 'I-you' and 'we' intersubjectivity are recognised by developmental psychologists and philosophers (Bakeman and Adamson 1984;Trevarthen 1979;Reddy 2010;Rochat 2014;Zahavi 2014Zahavi , 2015Stern 2018) and widely accepted as the two basic forms of intersubjectivity. Much of this work identifies the need to think about these intersubjective interactions (both 'I-you' and 'we') as non-reducible to the 'aggregates' (Gilbert 2013, p. 59) of the individual subjects that constitute them. In short, the mutual recognition in either an 'I-you' or a 'we' interaction births something that is more than the sum of its parts. An 'intersubjective approach' to social ontology hence rejects 'singularism' (Gilbert 2013, p. 42) or 'methodological individualism' (Zahavi 2014, p. 242) as a viable approach to analysing these kinds of interpersonal interaction.
Intersubjectivity theorists, then, have championed non-reductive social ontologies. The reasons given have been phenomenological and structural. Phenomenological insofar as there is a discernible difference in, for example, the quality of an experience where we attend to something together, versus one where we attend in parallel, as already outlined. The reasons are structural insofar as clear structures of reciprocal awareness can be identified as constitutive of these phenomena, namely 1-3 above (I am aware of you, and you of me, and this is 'out in the open' between us.) While at first glance it might look like a 1-3 checklist is the paragon of a reductive analysis, the fact that (3) constitutively includes reference to (1) and (2) precisely demonstrates the non-reducibility of the whole to separate unrelated parts.
While there is a huge amount of valuable work which discusses the nature of I-you and we interactions, and this work provides the framework for thinking about more complex interactions, these complex 'group dynamics' (as psychology might have it), have not been much analysed by those thinking philosophically about the nature of intersubjective standpoints. It is these more complex intersubjective structures and standpoints that emerge in triadic and multi-person interactions that I am interested in here. As I have already indicated, it is the second-person plural standpoint that I will focus on here.
Reflection on our ordinary experience should make clear the kinds of interactions that are purportedly second-person plural in structure. When we say to the couple who have hosted us 'Thank you so much for your hospitality', or we say to a pair of friends 'see you next week', we are addressing more than one person at once with this 'you'-which might be more helpfully referred to as 'yous'. This is not just a playing with words but, I argue, there is a phenomenology attached to recognising and relating to a pair as a pair, which cannot be reduced to two separate I-you relationships in parallel. There is a quality to receiving the hospitality of a couple, for example, in a home shaped by their joint attention, joint intention, joint commitment and joint action, that means that this 'thank you (yous) for your hospitality' means something different to the 'thank you' that might be spoken to each one separately (and in turn, these separate singular addresses capture something distinct from the plural address). I want to call this the second-person plural address or an I-yous structured address. If this is right, then in my address I am not only perceiving and addressing each of my friends as subjects, but I am also perceiving that which is between them, intersubjectively-the 'in-between' (De Jaegher et al. 2017)-and this is constitutive of the possibility and meaning of my address to them as a plural subject.
My claim is that this I-yous structure, which minimally involves at least three subjects, should be properly understood as a unified intersubjective whole, just as an I-you encounter or a collective experience as a 'we' are intersubjective wholes. Not only this, but that the plural subject addressed as a 'yous' is recognisable by the addresser as a unified whole, capable of receiving a single address.
For clarity, as above, the argument that we can perceive multiple others as plural subjects via gestalt perception is a point that can be applied to the observational (thirdpersonal) stance of 'I-they' as well as the participatory (second-personal) stance of 'I-yous'. I am more interested to think about how this gestalt perception might work in cases where I take a second-personal stance towards the plural subjects I identity. A full defence of the second-person plural requires more than I have space to consider in this paper, for, as I go on to outline, a second-person plural requires not only that I am able to perceive and address multiple others as a unified 'plural subject', but also that the members of this plural subject are co-ordinated such that they can jointly attend to me, can receive my address as directed to them as a plural subject, and can co-ordinate a joint reply-that is, can reply to me with an address as a plural subject-which I argue elsewhere is possible in a number of ways. This part of the second-person plural is the 'other side' of my recognition of multiple others as a plural subject, namely it is their recognition of me (as a subject) and their joint response to me (second-personally) as such. This I call the 'we-you', which is the inverse or the 'other side' of the I-yous. As analysis, explanation and justification of the possibility of the we-yous is a whole distinct section of the project of analysing the second-person plural, I will not focus on this 'side' of the interaction here for reasons of space. I look at this as part of the bigger project that this argument is situated in, which seeks to give a full justification of the substantiality of the second-person plural. The focus in this paper is the role that gestalt perception plays in the second-person plural, so I am here focusing on the I-yous, which specifically involves this gestalt perception. 6 To be clear, I am not only claiming here that in engaging another subject I recognise them as a gestalt whole, nor that our 'I-you' or 'we' intersubjective recognition of one another is a non-reductive whole, although both of these are true. My emphasis is that the intersubjective awareness between an other and another other is also something that I can perceive as a gestalt whole. This is not something that can be taken for granted. While it is obvious that we can in some way observe interactions between people other than ourselves, the question for the philosopher of intersubjectivity is whether these interactions are recognised as genuinely intersubjective, and-my interest here-whether they can be brought into my own intersubjective interactions as such. There is a line of thought which denies or ignores the possibility of our recognition of other plural subjects as properly intersubjective. This line of thought can be traced in the intersubjectivity literature, with its focus on the second-person singular but not the second-person plural.
The position that I'm arguing for therefore stands opposed to the claim that these various multi-person structures can be analysed sufficiently in terms of their constituent dyads. This dyadic approach may claim that to address a group or a pair is, as a matter of surface grammar, to engage in a second-person plural address, but when we actually look at what is going on we find that we are addressing one person. This person is a spokesperson (or receptive equivalent) for the others. When I thank my friends for their hospitality, this is in fact an address directed to each of them separately, in parallel, with each address referring to 'the other other' in its content, but not its form. This interaction can hence be given in terms of the dyadic interaction I have with the addressee, and their dyadic interactions with the other or others in the group, which can be considered separately. We find this dyadic approach in Sartre and Buber's accounts of intersubjectivity, as well as explicitly in the contemporary analysis given by Beata Stawarska (Buber 1996;Sartre 2003;Stawarska 2009).
The dyadic approach asserts that while I may be able to be peripherally aware of other others when I address the primary other, or while I might be able to engage other others indirectly by bringing content from other interactions into the content of this one, I nevertheless can only have one focus of attention, and so I can only directly attend to and address one person at a time. I might be able to switch between secondperson singular standpoints, giving the impression that I address multiple others at the same time, but, goes this line of argument, this is not strictly true. This is an important concern worth taking seriously. If a second-person plural address is genuinely second-personal, it must involve a quality and focus of direct mutual attention, so the question arises as to how I can attend to more than one person at once. If my attention is split or broken, then it cannot be genuinely second-personal. I face this worry by arguing that the second-person plural is not simply a matter 6 See fn.3. of surface grammar, but is as metaphysically substantial as the first, second and third person singular standpoints. Taking seriously the fact that the second-person standpoint must involve focused mutual attention between addresser and addressee, I will argue that this is possible in part because plural subjects exhibit gestalt qualities. This means that the kinds of multi-person interactions I have in mind not only cannot be reduced to the mere addition of the participating individual subjects, but that they also cannot be reduced to the mere addition of any of the participating dyadic interactions within the multi-person structure.
Again, to clarify, there are two aspects to this. First, when we observe or analyse multi-person interactions third-personally, we can point to the structure of such an interaction and see that this is a unified whole, composed of parts-in-relation, i.e. that these interactions have an underlying structure and as a result exhibit properties that are more than the sum of their parts-both their parts as the individuals involved, and the dyads involved. Second, thinking about the phenomenology of perception through the lens of gestalt theory also helps us think about how we engage as participants in multi-person intersubjective interactions, namely, as I have already mentioned, how it is that we can recognise and address multiple others as a unified relational unit or whole. Having laid out the claim I want to argue for, I now turn to the some of the theory underpinning gestalt psychology so as to establish the legitimacy of this application.

Gestalt structure
Gestalt theory taken broadly argues that perception is not only of individual things, but also of the relations that hold them together, which comprise a thing's form or structure. (Grossman 1984, pp. 41-59). As Köhler puts it: 'Parts of molar perceptual units often have characteristics which they do not exhibit when separated from those units. Within a larger visual entity, a part may, for instance, be a corner of this entity, another part is contour or boundary, and so on. It now seems obvious, but nobody in psychology had seen it before: the same happens in any physical system that is pervaded by interactions.' (Köhler 1959, pp. 4-5) By way of illustrating in more detail the nature of gestalt perception, I want to focus on the famous example of perceiving gestalt qualities aurally in listening to a piece of music, used by Christian von Ehrenfels in his analysis of 'temporal gestalt qualities' (von Ehrenfels 1988, pp. 97-100). 7 When we listen to a piece of music we don't hear a set of notes in isolation from one another, otherwise we wouldn't hear them as a piece of music-as a melody-which is a unified whole. Each note has a place in the structure of the whole. The phenomenology of each part is constituted not just by the immediate impression but by its place within the whole. The musical case brings to attention more clearly how a multiplicity of 'individual' perceptions-in this case musical notes-stand in a complex set of relations to one another, creating patterns. It is these patterns, these relations and sets of relations, that we perceive as the form of the music. 'Patterns…clearly demonstrate that it is relations among…objects which decide what objects become group members, and what others do not, and where, therefore, one group separates itself from one another,' says Köhler (1959, p. 3).
It is notable that this musical example is also used by Husserl in his analysis of the lived experience of temporality or 'time-consciousness' itself. Husserl famously articulates the phenomenology of time-consciousness as having a threefold structure, that of impression, retention and protention. (Husserl 1991, pp. 5-25 andHusserl 1977. See also Zahavi and Thompson 2007, pp. 77-78). As we've noted, when we listen to a piece of music at any moment in time we experience the immediate note, but we also retain in consciousness what has just passed, and we protend (reach forward) in consciousness to what is next. At any moment there is a note that we immediately perceive, but this perception in some way includes the perception of the previous note within the horizon of the current perception itself. We recognise events unfolding in time as unified wholes rather than a set of disconnected moments. This is not to say that the individual notes are not heard as separate notes. Clearly, they are. The unity of a piece of music is not the unity of a homogenous sound: if the individual notes were not heard as distinct from one another then the piece wouldn't be heard as a piece of music, but as invariant noise! As Solomon Asch puts it, when discussing the perception of the unity of persons: 'It is…far from the observed facts to describe the process as the forming of a homogeneous, undifferentiated "general impression."' (Asch 1946, p. 278). When listening to a melody the unity perceived is rather the unity of notes-in-relation, notes-in-a-pattern.
Insofar as we think about listening to a piece of music, it is because of the structure of time-consciousness that we perceive notes in the past as building towards those playing in the 'immediate now' of the present. These notes 'are continually being stamped with new characteristics,' (Cohen and Moran 2012, p. 227). It is for this reason also that we can know that an unfamiliar piece of music has been cut off in the middle, because we perceive the notes together as a whole, rather than just the individual notes in isolation from one another. We are capable of being surprised or startled because we 'reach forward' into what is coming next.
It is worth saying that retention is not the same as recollection from any other point in the past, and protention is not the same as an expectation of something further off in the future. Retention and protention are constitutive features of every lived experience which make possible the recognition of unfolding events as unfolding and connected. Unlike remembering what happened yesterday or anticipating what will happen tomorrow, retention and protention are automatic, unconscious and necessary features of the lived experience of events as temporal. Lanei Rodemeyer gives a helpful summary: In a phenomenological analysis of the experience of a series of musical notes, we notice [that]…these notes influence one another; they are not experienced as a series of individual, independent notes that happen to be played and heard. In other words, the perception of these notes is not simply of each individual note while it is immediately before consciousness. Instead, my experience gives the notes as reflecting on one another, playing in relation to one another, creating harmonies etc.….I actually experience several notes in their different qualities at once: The experience of the last few that have been played is held onto by consciousness. (Rodemeyer 2006, pp. 8-9).
This leads Ehrenfels to conclude that a melody is not a 'mere sum [Zusammenfassung] of elements' but is 'something novel in relation to this sum, something that certainly goes hand in hand with, but is indistinguishable from, the sum of elements' (von Ehrenfels 1988, p. 83).
The two key points I want to focus on here are that (i) gestalt qualities are nonreductive properties, and (ii) gestalt qualities are grounded in underlying structures which really exist. The first of these emphasises that there is something still 'left over' once all the notes that make up a melody have been 'counted'. The structure of the piece cannot be taken out of the perception without destroying the whole. In Grossman's words: What we perceive, to put the theory in a nutshell, are not mere sets of perceptual stimuli, but are structures of stimuli. Consequently, what we experience are not mere sets of sensations, but structures consisting of sensations. This basic insight was often summed up in the slogan: a whole is more than the sum of its parts! In our terminology, it becomes the ontological law: a structure is not the same as the set of its parts. (Grossman 1984, p. 59).
The second point, that Gestalt qualities are grounded in underlying structures which really exist, is to say that there really are, for example, a set of relations of notes that make up the musical form that we perceive. This is what Ehrenfels called 'the foundation [Grundlage] of that quality' (von Ehrenfels 1988, p. 93). The consequence of this is that if gestalt qualities are given by the form as a whole rather than merely the component parts, a similar (or isomorphic) form (but different parts) should therefore nevertheless give a similar set of perceivable qualities. For example, a piece of music is recognisable as having the same form even if all the component parts are different-if the whole piece is an octave lower, for example. Ehrenfels states this explicitly: '…Proof of the existence of Gestalt qualities is provided, at least in the sphere of visual and aural presentations, by the similarity relations…which obtain between melodies and figures having totally different tonal or positional foundations' (von Ehrenfels 1988, p. 90).
We cannot just help ourselves to gestalt qualities wherever we feel like it, of course. Ehrenfels is clear that gestalt qualities have to be earned, as is Kurt Koffka, who notes that 'just as the category of causality does not mean that any event is causally connected with another other, so the gestalt category does not mean that any two states or events belong together in one gestalt.' (Koffka 1935, p. 22). The characterisation of gestalt structure given above prevents unregulated appeal to gestalt perception, because justified appeal to gestalt structure requires that the phenomena in question fulfil the following criteria: (i) The qualities in question must be given in the phenomenology of perception, (ii) It needs to be shown that similar structures have a similar phenomenology, even if the component parts are different.
My claim that multi-person phenomena, like the second-person plural, exhibit gestalt qualities hence needs to be held to account by these criteria. Having outlined these key features of gestalt perception, I now move to arguing in a more targeted way that a non-reductive account of multi-person intersubjectivities can indeed be understood in these gestalt terms.

The phenomenology of others-in-relation
There is plenty of precedent for applying gestalt principles to interpersonal relationships and groups, indeed, I would suggest that the work that we find on this maps on to the work in the philosophy of intersubjectivity on 'I-you' and 'we' interactions. There is, for example, precedent for the claim that we can perceive other subjects as gestalt, rather than as a set of characteristics in parallel with one another. Asch, for example, notes that 'the subject can see the person only as a unit, he cannot form an impression of one-half or of one-quarter of the person.' (Asch 1946, p. 276). Another person might be engaged as gestalt on a number of levels, but engaging them as a 'whole' is one of the characteristic marks of the second-person standpoint in the literature. This 'engaging the other as-a-whole' is contrasted with the third-person standpoint, which is liable to deconstruct a person into their attributes or parts. While this analysing or observational stance is fitting to the third-person perspective, the second-person cannot fall into this reductive standpoint without destroying the second-person quality of the interaction. Compare, for example, Buber's articulation of the I-you with the language of the gestalt thinkers: When I confront a human being as my [Thou]…then he is no thing among things nor does he consist in things. He is no longer He or She, limited by other He's and She's, a dot in the world grid of space and time, nor a condition that can be experienced and described, a loose bundle of named qualities…Even as a melody is not composed of tones, nor a verse of words, nor a statue of lines-one must pull and tear to turn a unity into a multiplicity-so it is with the human being to whom I say [Thou]. I can abstract from him the colour of his hair or the colour of his speech or the colour of his graciousness; I have to do this again and again; but immediately he is no longer [Thou] (Buber 1996, p. 59) We also find explicit reference to intersubjective groups as gestalt insofar as this pertains to the first-person plural. Indeed, we find an explicit exploration of this in the gestalt thinker Koffka, who defines the first-person plural very much in line with dominant understandings of this phenomenon in the intersubjectivity literature: The reality of the psychological group finds its expression in the pronoun "we", "we" means not simply a plurality of persons which includes myself, it means in its most proper sense a unified plurality of which I and the others are true members….The speaker [of the pronoun "we"] experiences himself as part of a group, and his actions as belonging to this group…[T]he word "we" refers to a reality. It is never a mere abbreviation of "they and I", or "he and I". (Koffka 1935, p. 651) Koffka briefly considers the group as a gestalt whole from the first-person plural perspective. This to say, we find in the gestalt psychology literature precedent for thinking about the single other as gestalt in a face-to-face encounter, and for thinking about the first-person plural group as something more than the sum of its parts. Kurt Lewin's 'field theory' takes seriously the idea that small groups can be understood 'as 'natural dynamic units' or wholes.' (Lewin 1943(Lewin -1944. 166) Lewin makes use of gestalt principles specifically to think about ways that individuals influence and are influenced by their place in and experience of social groups. He analyses what he calls a subject's 'life space' or 'hodological space' as shaped by group dynamics (Burnes and Cooke 2013, p. 414), which in some cases must implicitly involve concrete multi-person interactions in a group, though this stands in need of fleshing out further. However, even where some philosophical considerations of the nature of intersubjectivity as gestalt have been attempted, to my knowledge this has not been with an eye to thinking about the implications for a more in-depth analysis of the formal structure of intersubjective standpoints, i.e. contrasting first, second and third-person standpoints, singular and plural. It is this kind of analysis that I am interested in here, specifically a defence of the possibility of the under-discussed second-person plural standpoint.
In order to earn this claim, I will need to demonstrate, as above, that (i) gestalt qualities are exhibited in the phenomenology of the perception of intersubjective interactions between others, and (ii) that there is a structure or a form underlying the phenomena in question that produces similar gestalt qualities even if the component parts are changed. I will start with the phenomenology.
The phenomenological claim is that just as a piece of music is unified in consciousness, so too are certain intersubjective interactions between oneself and multiple other parties. Let us build up to this claim by starting with the more straightforward case of a dyadic second-personal interaction. Consider an everyday conversation over coffee between oneself and another. Just as we don't merely hear the discrete notes of a piece of music, but the whole, so too when we perceive and engage with the other in a conversation we don't merely perceive each (verbal and non-verbal 8 ) contribution in isolation, but we follow the 'ordered flow' of the interaction, perceiving its structure and meaning or sets of meanings as a whole. We 'hold onto' the last contribution in consciousness, whether this is a movement, a gesture or a verbal communication, and as such, the meaning of the immediately present contribution makes sense, creating an 'intersubjective melody' (in a potential range of different ways) with what comes immediately before and after. In this way perception and engagement of intersubjective reality always involves more than an immediate impression: it involves the retention and protention of the rest of the communicative interaction.
If we shift the thought experiment to think not of a dyadic conversation, but a triadic one, in which I am engaging with two others, it becomes clear that this retention and protention is not only identifiable in cases of dyadic intersubjectivity. The ongoing retention and protention of the movements, gestures and verbal articulations of, not only the 'first other', but also the 'other other', are part of our experience in this kind of small group. These are not merely aggregated and understood post hoc, but contribute to the unfolding meaning of the whole interaction. For this to be true, structures of reciprocity between subjects other than me must also be perceived and engaged as 'wholes'. Not only is the reciprocity between self and other non-reducible, but so is my perception of the reciprocity between others and other others. If we were not able to hold the meanings of the communicative words and actions of more than one other subject within the horizon of our present experience, we would not be able to respond coherently in many multi-person interactions. Imagine guiding two others engaged in the project of moving a table up some stairs. I address them from the sidelines, "It needs to tilt to the left", etc. In this interaction, I perceive them (yous) as a whole, and I participate in the joint project by guiding their joint actions with directive addresses. We are all aware of one another, and aware of this mutual awareness, which makes it possible to be responsive to one another in this joint task. I do not process each of their actions and perceptions separately, but as a whole. A complex co-ordination of our three subjectivities takes place here: no-one is a mere observer, everyone is a participant. It would be right to say that there is a straightforward first-person plural going on here, as we are all jointly committed to a specific task, but I submit that it is also the case that there is an asymmetry here between me and those moving the table: I am also addressing them second-personally in terms of their joint action. As such there is also a clear I-yous in play. (In addition, they respond to my address, responding primarily with their co-ordinated action, or perhaps with questions directed back at me-this is the 'we-you' that makes up the other 'reply' aspect of a second-person plural).
Just as retention is not recollection and protention is not anticipation, holding multiple points of awareness in attention simultaneously is not the same as adding together different intersubjective interactions. Rather, there are elements of the awareness of each person which play a constitutive role in the awareness of the others. Not only does the awareness of each friend's awareness of me 'stamp' my response with 'new characteristics' creating new meaning, but their awareness of each others' awareness 'stamps' both of their movements and responses with emergent meaning as well. That which emerges between them is also part of what I am engaging with, as well as with each of them as subjects. I perceive their behaviour as an intersubjective interaction, as a pattern of subjects-in-relation-to-each-other. In such contexts each individual's behaviour is made meaningful as a part in the whole interaction. Indeed, I see behaviour as 'theirs', and it's possible to engage with it as such, namely to engage with it as 'yours' (plural).
Of course, there is also straightforward recollection and anticipation in group interaction as well, and this is also part of everyday intersubjective interactions played out in time. My emphasis here, however, is to note that it is not simply recollection and anticipation that are at play in intersubjectivity. Recollection and anticipation bring content about others and other others from 'other times' into the present, but intersubjective retention and protention are structural features of the interactions themselves, not extrinsic add-ons to an otherwise self-contained experience, but intrinsic.
Consider another everyday example. Imagine being on a train and finding yourself in proximity with a pair of others. They are clearly friends or a couple, one of them is crying, the other comforting. You rummage through your bag and offer some tissues to the one doing the comforting. This is a clear I-yous interaction-the comforting-and-being-comforted drama that's playing out is one in which (from the first-person perspective) I can see the structures of intersubjective reciprocity between these others, and as such I can meaningfully engage with them as a 'yous', a second-person plural subject. All parties understand what I'm doing when I hand the tissues to the comforter rather than the comfortee: I am giving the tissues to 'them', or rather, to yous, to your relational unit, and to the interaction playing out playing out between the two. At some level this is very obvious. Just as Köhler notes that it is obvious that we see shapes and hear melodies as wholes, so too with interpersonal interactions.
We can see why, therefore, the second-person plural can have a quality of attention that is not problematically split or broken. Just as multiple notes can be contained within the 'horizon' of the focus of my attention when listening to music, so too can multiple subjectivities and their contributions to the unfolding whole. Understanding subjects-in-relation as gestalt wholes hence speaks to the worry that it is not possible to attend second-personally to more than one other at once. 9 The musical analogy also helps us note differences in the way that this argument about gestalt perception of plural subjects opens up space for thinking about the second-person plural as well as the third-person plural. We can perceive plural subjects as gestalt wholes from a detached, spectatorial perspective just as we can listen to a piece of music that we are not a participant in contributing to. Music is also something that we can engage and participate in, however. We can consider scenarios in which we participate musically in multi-person contexts. Playing in a jazz trio will involve a first-person plural perspective, of course, as 'playing together' is something that the whole group does together. But I would argue that we also find more complex sets of interactions within the group-the double-bass player, for example, waiting to join in, listens not only to the drummer and the pianist separately, but also engages with the whole complex sound that the two of them are already producing. She works out how to enter the music as a participant, not only insofar as she is attuned to each of her musical colleagues, but also to the whole that is already emerging between them. 10 These analyses of concrete interactions provide phenomenological evidence for the possibility of a substantial second-person plural. I turn now to look at the second aspect of gestalt perception that I take to be relevant-that of underlying structure.

The structure(s) of others-in-relation
Just as a musical form has an underlying structure which determines how it is perceived, I will now lay out how something comparable is at work in intersubjective interactions. There are certain structures that underpin interpersonal reality, namely a very specific set of structures of interpersonal reciprocity that interact with one another to give a certain pattern that can be perceived. Just as the structure of music is built out of the relationship between different soundwaves (experienced phenomenologically as tone, pitch, key change and so on), or the structure of a visual scene is a spatial relationship between shapes and lines (experienced phenomenologically as distinctions between 'figure' and ground', and so on), so the 'structure' of intersubjectivity is 'built out of' the relationships of mutual awareness between subjects. And just as different arrangements of notes give rise to different relationships or patterns between the notes, so too when human subjects are 'arranged differently' and are mutually aware of each other in different ways, different phenomena emerge.
My claim is that addressing a pair or a group is not the same as addressing each member separately in parallel precisely because the object addressed (i.e. the interpersonal 'object' of the others-as-a-plural-subject) has an underlying structure or form, which the addresser perceives, attends to and addresses as a unified whole. I can recognise a pair as a system of two other subjects held together as a whole by a structure of reciprocity between them. It is this structure which holds multiple subjects together as 'system' that can be attended to without this involving a problematic division of attention between multiple parts. Whether I see that they are engaged in mutual direct attention of one another or whether they are engaged in joint attention together towards some object or task, I can engage them not simply as two isolated subjects doing their own thing, related to one another by mere contiguity-rather, I perceive and engage them as an interpersonal whole structured by specific patterns of reciprocity. 11 In a second-person plural address, I must be aware of the others as a plural subject who are jointly committed to attending to me, that is, to giving and receiving communications together.
Building on the dyadic formulation above, an example of the structure of intersubjective awareness in a case where C attends directly to A-and-B in addressing them, such as when I address the couple on the train, is: Which is to say: 7. [A and B under the form 1-3] attend to C 8. 5, 6 and 7 are 'out in the open' between A, B (as A-and-B jointly) and C As in the dyadic case, this formulation is designed to demonstrate that these different elements of intersubjective awareness are constitutive of each other in their form, and not just their content. It is for this reason that the form of such an interaction is nonreducible to the mere addition of isolated parts. This is why I am able to attend not simply to 'you' and to 'you' in parallel, but I can attend to 'yous': to you-and-you-inrelation-to-one-another as a whole. This analysis is not a vague claim invoking a 'more than the sum of its parts', but a recognition of the specific structures of interpersonal reciprocity in play. Different forms of interpersonal reciprocal awareness produce different intersubjective structures, which give different gestalt qualities. This is why an I-you is different to an I-yous, which is in turn different to a we-yous. As the forms of intersubjective awareness change, this is a change in the underlying structures, and this gives rise to a change in the phenomenology of the interaction. The jolt of realisation that there is a third person in the room, following our conversation, when I (or we) thought it was just the two of us in the room, highlights this phenomenological shift. This means, as per the requirements of gestalt structure, that even where the component members of an interaction are changed, we can identify similar structures which give rise to similar gestalt qualities. We recognise similar structures in different contexts with different people: I-you structures have identifiable features in all kinds of different social contexts, as do I-yous structures. We can identify I-you, we, we-you, I-yous, we-yous interactions which rely on (I have argued) the recognition of different interpersonal wholes. These structures really exist and provide the foundation [Grundlage] of the quality of these different intersubjective standpoints. These interactions involve completely different people and can have a completely different relational content: congratulating a couple on their wedding day is very different to facing a panel for a job interview, for example, but there is a similarity of I-yous interpersonal structure to both that is immediately visible.

Conclusion
In summary, I have outlined some relevant aspects of the phenomenology of multiperson structures, including a phenomenological analysis of second-person plural address. I have argued that there is good reason to think that we can adopt the language and understanding of parts and wholes used by gestalt theory to understand these phenomena non-reductively. We have as a result good reason to reject the claim that we can only attend directly to one person at a time. This analysis functions as an example of how phenomenology and gestalt theory can complement one another, towards a non-reductive analysis of our engagement with the world beyond us.
This analysis also opens up further debate at the intersection of gestalt theory and intersubjectivity and, more widely, encourages further consideration of how an account of the second-person plural as substantial feeds into other discourses that trade in the importance of the second-person perspective. To return to Lewin, for example, it is worth making explicit, in the light of my argument here, that nuanced attempts to do a cartography of a subject's 'life space' must recognise that the forces that shape this space are not only forces from the individual on the group-as-a whole, or the groupas-a-whole on the individual, but also that interactions within the group influence the group and the individual, insofar as these interactions between members of the group are complex, and effect other others in the group, which then also reverberates through the whole. Further research in field theory with these more complex intersubjective structures in mind could then prove fruitful. In this vein I have elsewhere started to look at how mindfulness of multi-person structures such as the second-person plural can itself facilitate nuance in multi-person discussions themselves, such as in university seminar contexts (Pawlett Jackson 2016), an idea which invites further research and consideration. There is likewise further work to be done applying this consideration of the second-person plural into 'second-personal ethics' (Darwall 2009)-where the possibility of the second-person plural suggests that the intersubjective structures which underpin moral obligation may be more complex than is currently reflected in the literature. I suggest that these, and numerous other possible applications, indicate that this analysis has much to offer ongoing research across a range of disciplines. 12