Multispecies Interaction in Dog-Assisted Therapy Sessions

Abstract

In the article, we examine dog-assisted therapy sessions in triadic terms, as a constellation between three agents. The prevalent model of the therapeutic relationship is the two-person model. The article contests such dyadic and anthropocentric understanding by emphasising the role of ‘the third’ in the interactions between client and therapist from a multispecies perspective. Our analysis draws on technologically mediated participant observation of dog-assisted therapy sessions supplemented by ethnographic interviews with therapists. By stressing the integral role of the dog in the therapeutic relationship, we ultimately try to bridge together two different traditions of thinking the third, one focused on third-party human agents, the other on third-party things. Ultimately, the article suggests that engaging with dog-assisted therapy offers a means to extend the thinking of triads beyond human relationships into multispecies interaction.

Introduction

The dog jumps on the sofa.

The client begins to tell their therapist how they’re not feeling all that well today and. . .

Thump! Thump! Thump! [a loud sound interrupts the conversation]

The patient stops for a second, but continues to tell how their morning began. . .

Thump! Thump! Thump! (Field notes, 7 November 2022)

At first sight it might be hard to believe that we are in a therapy session here. After all, therapy rooms do not tend to have dogs jumping on sofas and thumping their tails as part of the set-up or habitual routine of therapeutic practice. However, situations like this are quite typical in our ethnographic data concerning dog-assisted therapy, where trained therapy dogs engage voluntarily with clients in both expected and unexpected interactions. In the small, warmly lit home-like room where the episode described above takes place, there is an armchair for the therapist, facing a sofa on which the client sits. The low, piercing sound that interrupts the conversation between the two humans comes from the therapy dog’s thumping tail when it frequently hits the back of the sofa. The dog stands on the client’s lap, holding a fluffy toy in their jaws. The client puts their arm around the dog’s chest in a hug-like gesture, smiles, and then continues to relate what their morning has been like. Therapists say that the dog’s presence may help the client to stay more focused during the session, and the dog is rarely simply just external to the therapeutic relationship or merely disturbing the humans.

The therapeutic relationship is nowadays paradigmatically conceived in dyadic terms (see e.g. Beebe and Lachmann, 2003; Fuertes et al., 2006; Jung, 1985; Mitchell, 1997; Stolorow et al., 2013). The two-person model was developed to challenge and replace the intra-psychic ‘one-person’ model of traditional psychoanalysis (Messer, 2011; Mitchell, 1988). For example, the APA Dictionary of Clinical Psychology (2013) defines individual therapy as ‘treatment of psychological problems that is conducted on a one-to-one basis. [. . .] Also called dyadic therapy’. However, in the vignette presented above, the therapeutic encounter or the ‘therapy dyad’ between the therapist and the client is accompanied by an intervening tail-thumping third: the dog. In this article, we examine dog-assisted therapy sessions in triangular terms, as triadic configurations. In the analysis, we do this by focusing on the actions and predisposition of ‘the third’ in the relations between the therapist and the client. We aspire thereby not only to complicate the prevailing dyadic understanding of the therapy relationship but also to expand the study of triads beyond human relations into multispecies interactions.

In our thinking of triadic configurations, we commence from Georg Simmel’s (1992 [1908]) theorising on dyads and thirds, yet we extend his insights to multispecies relationships. The German sociologist and philosopher was among the first to emphasise the sociological significance of the ‘third’ (Dritte) in and for social relationships. However, as he perceived the third mainly as a third-party human agent, we see it as important in the study of dog-assisted therapy to take triadic thinking beyond the anthropocentrism of Simmel’s contribution. For this purpose, we cross-pollinate Simmel’s thoughts with those of philosopher Michel Serres (2007) and with some of the perspectives of recent human-animal studies (see e.g. Carter and Charles, 2018; DeMello, 2012; Motamedi Fraser, 2024; Taylor and Twine, 2015; Wilkie, 2015) as well as multispecies ethnography (e.g. Kirksey and Helmreich, 2010). We ask: how are the dyadic and triadic interactions between the therapist, the client and the dog played out in dog-assisted therapy, and how may thirds both mediate and interfere or interrupt the therapy dyad?

The article is structured as follows. We start by briefly describing dog-assisted therapy together with our materials and methods. Then we contrast triadic thinking over against the prevalent dyadic perspectives in the social sciences, which consider relations as involving two members. After that, we present our analysis of dynamics between the dyad and the third in dog-assisted therapy in two sub-sections. In the analysis, we introduce the reader to moments of contact between dogs and humans, how they affect and are affected by each other, and how their interaction fluctuates between dyadic and triadic forms. First, we start with the moment when the therapist, the dog, and the client greet each other and examine how the presence of the dog transforms the dyadic therapeutic relationship into a triad. There we also look into the various ways in which the third is simultaneously the condition of possibility and impossibility for the emergence of the therapy dyad, something necessarily both included and excluded. After that, we examine how trust is built in therapeutic relationships with the help of therapy dogs, and how the dogs may mediate the relationship between the two human members of the therapy triad. We conclude by summing up the contributions of the article. Ultimately, for us, investigation into dog-assisted therapy not only helps contest the prevailing dyadic understanding of the therapy relationship but also introduces a multispecies perspective into triadic thinking.

Inviting a Dog into the Therapeutic Space

As a practice, animal-assisted therapy has long historical roots. Historical records show that already in 460 BCE Hippocrates recognised the ‘healing rhythms’ of horseback riding in improving the overall well-being of an individual and, in the 17th century, physicians applied animal therapy on patients to cure their mental and physical ailments (Fung et al., 2024). Dog-assisted therapy itself dates back to the 1950s (Fung et al., 2024); in the article ‘The Dog as a “Co-therapist”’ (1961), American clinical psychologist Boris Levinson writes about how he had been successfully using his dog, Jingles, with certain child patients for almost a decade by then.

Empirically, this article draws from nine months of ethnographic fieldwork conducted in Finland with therapist–therapy dog pairs and their clients, supplemented by six preliminary interviews and three follow-up interviews conducted with individual therapists. The therapists whom we studied include their trained dogs found suitable for therapy work in their therapeutic practice in the fields of psychotherapy and speech therapy. With the term ‘therapist’, we refer in our analysis to the professions of speech therapists and psychotherapists as occupational titles licensed and protected by the national supervisory authority for welfare and health. ¹ The field of animal-assisted therapy is emerging among other animal-assisted interventions, mainly in education, health care and social services, and to date there’s neither a mandatory training program nor an ethical committee yet to lead the way and direct the co-development of this expanding field. At present, institutionally dog-assisted therapy is still considered as a mere ‘method’ in Finland, but the field is going through changes; not only have professionals unified the training of working dogs in rehabilitation and education since 2010, but today (since 2023) the Social Insurance Institution of Finland also compensates animal-assisted therapy in the context of intensive medical rehabilitation.

Therapy dogs who are vividly present in the therapy context are considered neither as mere pets nor a means (therapists, for example, care for the well-being of the dogs, pay attention to their working conditions and regard it as important to make sure that the dogs have enough time free from work). For the therapists and their clients alike, they are present above all as individuals. As for the therapists, the dogs are their companion animals, best friends who share a home with their guardians. And, when it comes to clients, they perceive the therapy dogs as genuine and non-judgmental social agents who have their own language, different from that of humans, and whose presence and touch are integral to the therapy process (see Charles and Wolkowitz, 2024: 5). Given that it is not insignificant how we speak of nonhuman animals – Donna Haraway (2016: 12) has famously suggested that ‘[i]t matters what matters we use to think other matters with’ – it makes a difference how therapy dogs are conceptualised and approached both in the therapy setting itself and when researching them. The fact that nonhuman animals are not automatically regarded as members of human societies, with equal rights compared to their human companions, adds an extra layer of complexity to research. In our materials, therapy dogs are perceived as bodily, spontaneous and surprising beings, both as individual members of the therapeutic interactions and as representatives of a domesticated nonhuman species.

Among other companion animals, dogs are used in health-promoting work life due to their potential stress-reducing capabilities, such as psychological development or motivation for improving social skills (Rambaree and Sjöberg, 2019: 3). But the sole fact of owning a dog, having a dog present or being touched by a dog does not yet automatically improve a person’s mental health or make the human–dog relationship ‘therapeutic’ in itself. To derive health benefits in work or life in general, humans do tend to prefer to be in contact with companion animals when given the chance to do so. However, there is still lack of evidence in research literature on ‘how’ exactly therapy dogs are and can be implemented in work settings. Combining this with the fact that relatively little is known about therapists’ and psychologists’ perceptions of their own professional role in treatment of others’ wellbeing, it remains uncertain whether and to what extent dogs’ potential and human preferences are attributable to the interaction with particular therapy dogs (see Baird et al., 2023: 3–4; Martins et al., 2023: 33; Rambaree and Sjöberg, 2019: 12). In this article, we try to fill this gap with rich ethnographic materials by examining how dogs are integrated into therapy sessions.

Whenever the dogs’ own view of the world is considered in and out of the therapy sessions, their passing interests guide the conversations between therapists and their clients. The role of a therapy dog can be tiring for the dog due to a large variation of tasks they are called upon to perform with different people. In general, their work week is limited to two to eight hours in an attempt to maintain their ability and willingness to participate. When therapy dogs assent to or dissent from interaction, humans engage in meaning-making, trying to figure out, for example, whether the dog gave space to the client when they were sad, or could the dog be encouraged to take part in other ways. Be that as it may, therapy dogs nonetheless have a right to take part or leave without demands or consequences for them (see e.g. Mancini and Nannoni, 2022; Volsche et al., 2022). This is not regarded as a problem, since the dog’s active participation is not always necessary for the dog-assisted therapy sessions to work as therapeutic. Even a deceased dog may continue to ‘virtually’ (Deleuze, 1991) affect the therapeutic relationship. Whilst no longer physically present, the dog is not simply absent from the relationship, as the client and the therapist may, for example, together reminisce about what the dog did in a particular situation, and the client may also put to use in their everyday life new social skills that were originally learned and acquired in bodily contact with the dog.

The first author had already gained access to the field in a previous study and established connections with informants from the field of animal-assisted therapy. They kindly offered their help by connecting him with other practitioners of dog-assisted therapies and distributed the research announcement within their community. Separate research announcements and informed consent forms were composed for therapists and their clients. Six therapists replied via email to the research announcement sent out to therapists, and altogether nine individual 60 to 90-minute interviews were conducted remotely between November 2022 and August 2023. After preliminary interviews, three psychotherapists and speech therapists practising dog-assisted therapy found clients willing to participate in the study, provided information for them, and confirmed their consent. Consent forms were delivered to the first author without direct contact with the clients. ²

The data from technologically mediated participant observation for this study comprise 20 hours in total, of which 14 hours consist of video recordings by therapists without the researcher’s physical presence in the therapy sessions. Ethnographic interviews, each of the duration of approximately one hour, were conducted individually with three therapists in August 2023. A field diary of 164 pages in total was formed from the interview transcriptions, observation notes and video transcriptions of dyadic and triadic moments of contact in dog-assisted therapy sessions. All data was handled and stored in accordance with the guidelines for the responsible conduct of research by the Finnish National Board on Research Integrity. To organise the collected data for the purposes of analysis, the first author implemented selective video coding of micro-level interactions that displayed triadic exchange with its mediation, fluency and interruptions. The materials from observations and transcriptions of in-depth interviews were complementary to each other.

Taking Both Human and Non-human Thirds Seriously

In social theory, relations have been typically considered in dyadic terms, as relations between two. Michael Theunissen (1984: 6) has, for example, argued that the notion of ‘the social’ can be ‘extended to every type of relation [of the self] to the Other, including the presocial’. As for particular relationships, Alfred Radcliffe-Brown (1940: 2), for instance, maintained that kinship structures consist of a number of dyadic relations as ‘between a father and son, or a mother’s brother and his sister’s son’. Further, love is commonly conceived as ‘a direct relation that binds the first and second person, in an ecstatic form that precludes any distance’ (Esposito, 2012: 116). Ethics, too, offers itself to be thought – especially in the Levinasian (1985) framework – in reference to and as responsibility for the Other. But the examples do not stop here. One could just as well mention how exchange is typically modelled as a transaction between two parties, communication is conceived as taking place between the sender and the receiver, and how it is held that there are two sides to every struggle, conflict and war.

In contrast to dyadic social theory, ³ we propose in this article a sociological trialectics, a way of fathoming therapeutic relationships as triangular configurations. This is to foreground the figure of the third as a key category. We argue that it is only through the interplay of twosomes and thirds that it is possible to grasp the multispecies interactions in the context of dog-assisted therapy in their complexity. Overall, dyadic social theory is somewhat poorly equipped to grasp the complexity of the world (Fischer, 2013: 95, 98), but the point is not only that in addition to there being dyadic therapeutic relations there also exist those involving three elements. The third is not only the sign of increased complexity; it is elementary to any relation. There is a third to every relationship between two in (dog-assisted) therapy, either included in it or excluded from it.

As already mentioned, Simmel was perhaps the first to emphasise the sociological significance of third persons for social relations. The difference between the formations involving two members and those involving three, so Simmel (1992 [1908]) argues, is not only quantitative but also qualitative. A triad is comprised not solely of direct relations between one member and another (A ⇌ B, A ⇌ C, B ⇌ C), but between each two members there also exists an indirect relationship mediated by or derived from their common relationship to the third member as well as the relation of each member to the whole Simmel, (1992 [1908]: 114). Formations of three differ from dyads in terms of dynamics, stability, and by their degree of objectivity (see also Pyyhtinen, 2009). What is more, Simmel (1992 [1908]: 117–18) insists that any further expansions – to formations of 4, 10, 100, 100,000, and so on – do not have the same transformative effect as the addition of the third.

While in recent Anglophone scholarship the theoretical discussions on thirds and thirdness that would draw explicitly on Simmel have remained relatively scarce (for exceptions, see e.g. Cooper and Malkmus, 2013; Goodstein, 2012; Kagel, 2012; Krackhardt, 1999; Pyyhtinen, 2009, 2010, 2018a, 2018b), in Germany several authors have argued for a triadic understanding of relations, thus emphasising the importance of the third to our conception of social relationships (see e.g. Bedorf et al., 2010; Eßlinger et al., 2010). The discussions and debates have focused both on triadic or triangular constellations and on the figure of the third itself, and they have also insisted on the importance of Simmel’s work for their investigation. Worth mentioning is especially the work of Joachim Fischer, who has made a serious effort to systematise ‘tertiality’ and the sociology of the third (see e.g. Fischer, 2000, 2013, 2022).

The discussions and debates tend to use interpersonal relationships as the primary model for triadic or triangular relations. However, humans are dependent on and entangled with many more thirds than just their fellow humans. What is more, non-human or more-than-human thirds cannot be considered in the exact same terms as their human counterparts. In fact, if we are to believe Marcel Hénaff (2020: 170), the third-party human agent ‘tells us nothing about the third-party thing that bonds together the partners’. Therefore, to disregard the role of non-human thirds in relationships would unnecessarily narrow down the conceptual breadth of the figure of the third. While the point Hénaff makes is highly important, we nevertheless insist that to fully acknowledge the effectivity of non-human thirds it is crucial to transcend his more or less anthropocentric perspective. Whereas Hénaff acknowledges the significant role played by the things from the world in human relationships, he ultimately reduces them to mere symbols, giving primacy to human agency and privileging language, meanings and culture. This is already betrayed by the concepts third-party agent and third-party thing that he uses; the choice of words is prone to render non-humans passive and inert things, distinct from assumedly active and free human agents. However, far from being passive, non-humans are active, have effective capabilities and do something (e.g. Bennett, 2010; Coole and Frost, 2010; Latour, 1999). When non-human thirds are considered as hardly anything but symbols, it becomes impossible to see their powers and effectivity to the full extent.

We regard the work of philosopher Michel Serres as of key importance here, since Serres fathoms relations as triangular configurations consisting of three elements without reducing the third to either third-party human agents or to assumedly passive third-party things. This becomes clear at the outset of Serres’s (2007) The Parasite, in his recapitulation of Jean de La Fontaine’s fable of the city and the country rat. Serres uses the fable to illustrate his thinking of the ‘parasite’, which is his main concept for the third (le tiers). The story is about the city rat who has invited the country rat for a visit. The rats chew their meal with absolute delight on a Persian rug. The meal consists of nothing but sundry odds and ends, but for the country rat, at least, it makes a royal feast. However, their most genteel repast is cut short, as the rats hear noise from the door. The noise made by their cutting and nibbling, or rather their mirth, has woken up the head of the house, the tax farmer, who has now got out of bed to determine the origin of the disturbing sounds. From this mishmash of abusive animals, social parasites and noise, Serres (2007) unearths three aspects of the word parasite: firstly, an organism feeding on another (biological sense); secondly, an abusive guest (anthropological sense); and, thirdly, noise, static, a break in the message (information theoretical sense).

According to Serres (2007), the parasite is someone or something in-between, in the position of the third, interfering and intervening. The particular usefulness of the figure of the parasite for us here is that it illuminates the duality of any third: while a third may mediate and establish a relation, it may also intervene, interfere and disturb. Any relation is thus dependent on exclusion: to have equitable exchange, undisturbed communication or fair and productive cooperation, for instance, one needs to expel the parasite. No relation can be absolutely inclusive. A completely open, all-welcoming relation would collapse. Yet the notion of the parasite also suggests that the third is constitutive of the dyad even when being excluded, as the twosome establishes itself in relation to the parasite (see also Pyyhtinen, 2018c). As for La Fontaine’s rats, they later chose to deny themselves dainties and feast modestly but without fear.

Ultimately, in our argument we aspire to bridge these two traditions of thinking the third – the first focusing on the third-party human agent and the second emphasising the non-human thirds – which heretofore have scarcely crossed paths. In a given situation, it may just as well be the dog, the client or the therapist who is placed in the position of the third. Even though our analysis does not focus on the institutions partly covering the expenses of the clients’ therapy, those institutions, too, may present thirds, at once mediating and interfering in the therapeutic relationship of the client and their therapist. All in all, our analysis builds on a relational understanding of the changing nature of agency in triadic interactions, where the dog may act as one of three active members present in the setting.

Victory Laps

Let us now go to the field, to the moment from which the triadic interactions between the therapist, the client, and the dog start to unfold upon seeing each other. The people of the therapy triad typically do a short and somewhat formal one-off greeting of, for example, ‘good morning’, making short eye contact with a smile on their face, sometimes also raising (and waving) their upper limb and its extremities in an upright position. However, therapy dogs may prolong their greeting of human fellows and even repeat it multiple times. The dogs may engage in physical contact directly and stare towards the person they are greeting, as if wanting to express an interest in an even closer contact. In our data, therapists verbalise the dogs’ greeting gestures as friendliness and perceive these spontaneous acts as an expression of the dogs’ suitability to interact with humans, but also a way of interrupting human interaction with good intentions.

The following excerpt is from a scene presenting the same therapy triad as the one appearing in the scene depicted in the beginning of the introduction. The following moments of contact happened right before the dog jumped on the sofa:

In the hallway, the dog walks and wiggles thrice around the client with a toy in their mouth as the client scratches their back and greets them by repeating softly ‘Hello’. After three rounds the client asks ‘Still? And again. . .’, the therapist replies: ‘. . .And a victory lap!’ with the client replying: ‘And then around in the opposite direction’, and the therapist continues by saying: ‘. . .Like that!’ The dog stops for a second in front of the client to pick at the toy in their mouth and then starts pacing back and forth between the therapist and the client, right after the client having stood up and praising ‘Like that. That was good!’ The therapist asks, ‘Is it good now?’, indicating that the greeting has come to an end and meaning to continue towards the therapy room. As the client takes the first step, the dog continues their wiggly walk around them. After they’ve floundered to the room, the dog jumps to and from the sofa before the client manages to sit down, and then they jump back to their lap for some more back-scratching. (Field notes, 21 November 2022)

This excerpt shows how the therapy dog may both help to reinforce the therapeutic relationship and interrupt it by coming in between the two humans; in the excerpt, the dog momentarily becomes the centre of attention by wagging their tail, wiggling their body and squeaking their toy. Serres’s (2007) thinking of the parasite helps us conceptualise such interference. When making noise or asking for attention otherwise, the dog comes in between the client and the therapist, momentarily interrupting whatever it is that may be going on between the humans (with e.g. the client sharing something difficult). In that very moment, the dog becomes a parasite or the ‘noise of the system’ (Serres, 2007: 79).

However, it is not only by interfering that the dog may play the position of the third in the therapeutic setting. They may also act as a mediator or ‘catalyst’ (Brown, 2013) between the therapist and the client, reinforcing their bond. In the preliminary interviews conducted by the first author, the therapists expressed that their interest in adding a trained dog into the therapeutic setting is that often a dog’s presence makes a relationship warmer compared to one without a dog. The dog can help people get to know each other faster by simply being present and showing interest in the client, and the dog may also catalyse conversation with the two human participants. As one therapist put it:

My younger dogs, they have this thing with the toys. They always have some toy in their mouth, and that leads to conversations with clients on which toy the dog chose today. Are they holding a rat, a cow, a donkey or what? This gives a spark to conversations about what it reminds the client of and what the dog might want to communicate to us by having that particular toy in their mouth today.

This interview excerpt also points towards another crucial matter: verbalisation of the dog’s actions and intentions. Quite typically, the interaction between the client and the therapist is mediated by the actions of the dog, with the therapist and their client verbalising the dog’s doings and (assumed) thinkings to each other. As a verbal representation and vivid description of physical, non-verbal action and communication, the verbalisation could perhaps be understood as an ekphrasis of sorts. Ekphrasis traditionally means a verbal – often written – representation of a work of visual art. Etymologically, the word ekphrasis comes from the Greek ἐκ, ek (‘out’) and φράσις, phrásis (‘speak’), and the verb ἐκφράζειν, ekphrázein, literally means ‘to proclaim or call an inanimate object by name’. Thereby, ekphrasis can be said to amount to ‘descriptive speech which brings (literally “leads”) the thing shown vividly before the eyes’ (Webb, 2009: 76).

In dog-assisted therapy, the verbal communication between the two humans is focused on and fed by the non-human third placed before the eyes of the human participants, with the therapist working as an interpreter and connoisseur of ‘dog language’, which also in itself inserts a certain power asymmetry into the therapy triad. ⁴ The verbalisation of the dog’s actions serves mainly two ends. On the one hand, it is used to create a shared understanding of the dog’s intentions, providing the human participants of the therapy triad with a sense of knowing the non-human animal other. The therapist works to facilitate the client’s understanding of the non-human animal. In such situations, the therapist is placed in the position of the third between the client and the dog, providing the former with real-time commentary by talking over the dog. Some acts of therapy dogs may indeed require clarification, say, when they restrain the client’s movements or when the client’s act of speaking is somehow interrupted by the dog appearing as noise/parasite. On the other hand, verbalising the dog’s actions is a way for the therapist and the client to get to know and connect with each other. In the quoted excerpt, the therapist and the client use the expression ‘victory laps’ that is easy to grasp for both to describe how the dog victoriously circles around the client when meeting them. The choice of words is crucial for the building of trust between the therapist and the client through a shared understanding of the meaning of the particular figure of speech. At the same time, the dog is being excluded from the bond created by language between the two human participants; they are merely talked about and over.

All in all, while on some occasions it is the dog who acts as the third vis-á-vis the dyad, on others it may be the therapist or the client who are placed in the position of the third. This draws attention to two important things. Firstly, it highlights how ‘thirdness’ is not a fixed quality or position. It cannot be decided in principle, or known in advance, who or what appears as a third in a given situation. This is because thirdness is a relational notion and a circulating epithet passed on from one to the other; basically, anyone and anything can be placed in the position of the third (Serres, 2007). Secondly, provided that thirdness is relational, it follows that there exists a wide spectrum of different kinds of thirds; the third is a general category, which assumes different figures, positions, prepositions and functions depending on the relational constellation in question. The figures of the third that our analysis above described were the intervening third, the mediator and the excluded third. The last of these also reveals the fragility of triads. They may be split into two parties, where a dyad of two members allies against the third member.

Making Noise

Before proceeding to our last lengthy excerpt from the field diary, let us first go back to La Fontaine’s fable of the two rats startled by the noise they hear at the door. They fly into hiding and decide to continue feasting elsewhere – more modestly, to be sure, but with the advantage of not having to fear further disturbance. Conveniently for our topic, in the original version of the fable, written by the Greek storyteller Æsop (around 600 BCE), the rats were chased away by a house dog (and a cat), later to be replaced by a tax farmer in La Fontaine’s adaptation. Furthermore, in Æsop’s original version, the rats also get a second chance to return to their banquet. Combined with Serres’s insights on parasitic relations, this provides us with fascinating clues as to the bodily aspects of interruptions and how they are managed in the triad; interruption does not automatically mean that the interaction is disrupted or terminated for good.

The excerpt we focus on here concerns a situation where the client is sitting on the floor and the therapy dog is lying on the client’s shins with the therapist next to them. Then the therapist, holding a tablet computer in their hands, begins a new task. They start a quiz and collect dog treats in a cup on the floor for every right answer, when the following incidence suddenly interrupts the script and the expected course of action:

In a sudden moment, the tablet almost tumbles from the therapist’s hands onto the dog. The dog jumps backwards from the client’s lap, startling the client, and they all stop for a second or two. [. . .] The humans begin to talk about what happened, whilst the therapy dog returns to lie in between them, glancing at the treat cup, the therapist and the client one at a time. Instantly, a conversation about friendship begins. The client briefly describes how they were startled but then focuses on the consideration of the dog’s perspective on the situation. [. . .] After a minute, the therapist guides the session onward: ‘I think you’re both in need of relaxation after I startled you that way. . . . Do you want to lie on your back or on your belly?’ The client expresses their preferred position and how the blanket is placed on them. The client also decides the quantity of dog treats to be hidden inside the blanket. Lastly, they decide who asks the dog to search for the treats. The task falls on the therapist, who turns towards the dog, who has been lying still between them waiting for their turn. The therapy dog starts snuffling, digging up, and climbing on the blanket whilst the client giggles beneath them. After all the treats have been found, the dog climbs down, the client sighs deeply, smiles, and they decide to repeat the relaxation exercise. (Field notes, 10 May 2023)

In general, therapy can be understood as aiming to restore and reinforce the client’s agency. And, for the therapy relationship to work, the client and the therapist need to have established mutual trust (Crits-Christoph et al., 2019). The above excerpt exemplifies an instance of that trust being tested and jeopardised, with the therapist almost dropping the tablet computer on the dog, thereby startling both the dog and the client. It results in a state of noise. While the English word tends to reduce noise to an auditory stimulus, that is, sound, the French word from which it comes is richer in connotations; in the Old French from which it originates, noise meant ‘furor’ and ‘tumult’, that is, a state of confusion and disorder (Serres, 1995b: 8). Accordingly, for Serres (1982: 66), noise represents interference and obstacles to communication and interaction.

Moments of confusion and disorder may, however, be not only disruptive but also bear unexplored opportunities, and this also holds for the situation described above. What at first seems to present a mere falling into disorder in fact turns out to lead to a slight detour or caesura ⁵ in the flow of action and is actually made use of in the therapy. The first author watched a video recording of the events depicted in the excerpt together with the therapist appearing on it and discussed it with them. The therapist observed:

I make a mistake and, exactly, let’s stop here for a while without just moving forward like nothing happened. If I were a newly graduated therapist, I think I wouldn’t have had the skill to make good use of that situation. This is something that happens especially when the dog is involved in therapy; that I can utilise these situations and see where they take us. (Interview, 15 August 2023)

Instead of perceiving the mistake simply in negative terms, as a nuisance, here the therapist sees the state of noise as an unexpected opportunity, a gift, even. They adapt to the situation caused by the rupture by initiating a new conversation with the client, and later suggest a joint relaxation exercise for the client and the dog (which was something that the therapist had planned for the client–dog dyad in any case, despite the startling interruption). The exercise not only calms down both of them and nourishes their mutual bond but also restores the triad. Analogously, by drawing from Serres’s (1995a, 2007) idea of the duality of the parasite, in the situation the interruption was at once a condition of impossibility and possibility for the interaction (see also Pyyhtinen, 2018c; Watkin, 2020: 227).

The described moment of tumult can also be discussed from the perspective of the therapist’s self-disclosure as pro-relationship behaviour (see Crits-Christoph et al., 2019: 2). In the situation, the therapist accidentally and unintentionally acts like La Fontaine’s tax farmer, who startles and disturbs the feasting rats by squeaking floorboards and thereby becomes a parasite in relation to the rats. Interestingly, it was by accidentally disturbing and making noise – in the sense of creating disorder – that the therapist ultimately managed to enhance the client’s flexibility and agency, and the subsequent relaxation exercise restored the triad. The therapy dog, too, played a crucial role in the restoration of the triad: by showing the client their trust in the therapist in the situation, the dog helped the client regain their trust in the therapist after the state of noise and the rupture of interaction.

Conclusion

In this article, we have examined the triadic interactions between therapists, clients and dogs in dog-assisted therapy to challenge the prevalent dyadic understanding of the therapeutic relationship. The therapy dyad of the client and the therapist was shown by our analysis to be accompanied by a third in each and every moment; there is no relationship between the two without the therapy dog third either making their presence felt by conditioning, mediating, transforming or interrupting the twosome or by being momentarily excluded by the two partners.

By taking seriously that the actions of the dog count in and for therapeutic practice, the article also extended the thinking of dyads and thirds beyond the human realm into multispecies relationships. We wanted to bring together two different traditions of thinking the third which heretofore have remained largely separate. The first consists of authors like Simmel (1992 [1908]), Freud (2010 [1899]), Girard (1989) and Levinas (1985), who have paid attention to third-party human agents, with some of them examining tertiality more explicitly than others. The other tradition, by contrast, has emphasised especially non-human thirds. This tradition has its roots especially in the problem of communication (e.g. Serres, 1982, 2007; Shannon and Weaver, 1949). To disregard the role of non-human thirds in relationships would unnecessarily narrow down the conceptual breadth of the functions and figures of the third. To fully acknowledge the effectivity of non-human thirds, it is also important to transcend the more or less anthropocentric perspectives of treating them as mere passive and inert things or objects. In the context of dog-assisted therapy, therapy dogs are rarely if ever simply reduced to the role of an object or instrument, but they are social agents, transforming the therapy dyad into a triad. The dog may not only interrupt the therapeutic relationship by coming in between the two humans and momentarily becoming the centre of attention but also reinforce the therapeutic relationship. For example, thanks to the therapy dog, as we suggested above, the client and the therapist may learn to understand and trust each other better.

How dogs work with humans in the therapy setting challenges and undermines the ‘umbilical’ species thinking that easily bypasses and erases the particularity of the individual animal. When the dog is perceived by its origin as a loyal companion with unconditional love for their human guardians, their individuality becomes something incongruous with that origin (see Motamedi Fraser, 2024: 185; Watkin, 2020: 164–5). If the dogs were added to the therapeutic relationship as nothing but representatives of their species, with humans in control of their actions and abilities, they would not be able to help in accomplishing the pro-relationship goals of therapists and therapeutic practice. Instead, in reality, they act and are also perceived to do so by both the therapist and the client. As such, the dogs do both what the therapists intend and do not intend them to do. ⁶ The two human members of the therapy triad do not have full control over the actions of the dog, who on many occasions may surprise them by what a canine personality can do, be that by thumping their tail or by making victory laps around the client on the occasion of greeting them, for example. Furthermore, as these examples illustrate and make clear, the dog’s effects are not limited to their linguistic representation in the conversations between the client and the therapist (with the therapist and the client jointly verbalising the dog’s actions to frame the ambiguous canine acts as friendly, joyful and therapeutic), but there were occasions in our data where the dog succeeded in rendering the atmosphere of the therapeutic interaction warmer and more relaxed simply by being there and making their physical presence felt.

Overall, it is not predetermined and entirely clear, however, who or what acts as the third in a given situation. It is therefore not automatically the therapy dog who is placed in the position of the third in the therapeutic space, but thirdness fluctuates, being determined or ‘given’ in and through relations. This also makes it susceptible to change; it varies from one situation to another who or what appears as the third. The third, as we have seen, may not only mediate the relationship between the other two but also interrupt their relationship, coming in between them, as it were, as is expressed by Serres’s (2007) notion of the parasite or Simmel’s (1992 [1908]: 134) tertius gaudens. What is more, the interrupting third does not bear exclusively negative connotations but it can also be productive, beneficial and occasionally even necessary (see Burton and Tam, 2016: 104–5; Serres, 2007). This positive side of the parasite was discernible in our data, too, for example in the situation where the dog as the tail-thumping third interrupted the conversation between the client and the therapist. Already by interrupting and making their physical presence felt, the dog succeeded in rendering the atmosphere of the therapeutic interaction warmer and more relaxed. This made the dog crucial for therapeutic work and healing. In addition, there were also situations, such as that when the therapist startled the client and the dog by nearly dropping the tablet on the dog, where moments of interruption in communication created new and unexpected opportunities for the interaction and helped the participants to build and strengthen mutual trust. Ultimately, no matter who interrupts and distracts whichever dyad, inviting the therapy dog into therapeutic space without doubt changes and redefines what therapists can professionally do with their clients. In addition, while our analytical focus here has been on dog-assisted therapy, we nevertheless believe that sociological trialectics potentially has a lot to offer to the study of animal–human relationships more widely, beyond this context.