The Building of Empathy: Conceptual “Pillars” and Conversational Practices in Psychotherapy

Empathy can be considered a special type of cooperation between therapist and patient. This exploratory study compares psychoanalytical, depth-psychological and behavioural therapy, in each case using transcriptions of audio recordings of initial, mid-term and late sessions. For each school of therapy, five treatments are included, creating a database of 45 sessions. We describe the project and the method of conversation analysis using examples of these transcripts and hypothesise that while all three schools of therapy are faced with common fundamental problems concerning the realisation of empathy, one can observe empathy profiles specific to each school. Here, we introduce theoretical groundwork and the terminology of conversation analysis. The topic may be of particular interest to clini- cians, since everyday problems are examined through the prism of microanalysis.


Introduction
Empathy is not a binary variable that can be reduced to "yes" or "no". What one person experiences as a moment of empathy another might not. Hence, empathy cannot be defined a priori or in abstract terms, rather it is something that is created-through conversation.
An approach based on conversation and its analysis represents a departure from other important studies which examine empathy as an individual "ability" [1]. Tests have already been developed [2] which have proved useful for studying autism for instance. "Empathy" and "evil" are often perceived as a continuum; lack of empathy is connected to violence, while empathy prevents it.
The social scientist Randall Collins [3] has developed a different approach; he focuses on "situations", both of violence [4] and of empathy [5,6]; the rhythms of the linguistic and the physical [7] are "on the same track". In his extensive analysis, which cannot be sketched in here, violence occurs in situations the dynamics of which can be precisely described by microanalysis. Collins' point of departure was not violence however, but the linking of interactive practices via which violence is virtually avoided through empathetic rhythmisation.
These two approaches are methodologically quite different; one seeks to examine individual skills or their absence, while the other employs microanalysis of situations. However, both recognise a continuum of empathy and violence. "Situations" are "objects at border" [8]-they are as much a part of conversation as they are of cognition and experience. The question of which side "counts" is negotiated during the conversation and does not have to be decided in advance. Thus, a question emerges: If violence can be described as a property of mutually coconstructed situations, is this possible for empathy, too?
Empathy is considered a mysterious achievement of gifted therapists. Its origins are sought in happy circumstances in childhood, or successful self-experience shaping therapeutic work. An intelligent survey [9] shows how the psychoanalytical understanding of empathy has developed. She provides a knowledgeable summary of a long discussion within the field: (E)mpathy will here be viewed from an epistemological perspective, as a way of knowing, a way of arriving at an understanding of another's feeling state, potentially generating a bidirectional interactive field of considerable power ( [9], p. 713).
This is similar to what Ed Tronick has called a "dyadic state of consciousness": "This dyadic state organization has more components-the infant and the motherthan the infant's (or mother's) own self-organized state. Thus, this dyadic system contains more information and is more complex and coherent than either the infant's (or the mother's) endogenous state of consciousness alone. When infant and mother mutually create this dyadic statewhen they become components of a dyadic systemboth fulfill the first principle of systems theory of gaining greater complexity and coherence. The gesturing mother-held-infant performs an actiongesturing-that is an emergent property of the dyadic system that would not and could not occur unless the infant and mother were related to each other as components of a single dyadic system". [10] If an individual's consciousness is unable to solve certain problems, it has to couple with another's consciousness (mind) so that both can take on a "dyadic state" and find a solution. They can then uncouple in order to recouple at a later opportunity.
"Theory of mind" research long assumed that children do not develop these skills until the age of three-in experimental situations. Significantly, observation of natural situations [11] has identified these skills at a much earlier stage; they are documented in children as young as 12 months [12,13]. What was conceptually difficult in some experimental approaches to studying empathy becomes clear: in investigations into mirror neurons [14], one monkey is conceived of as the "observer" of an "object" (another monkey); the idea that the observer creates a propositional hypothesis, in analogy to science, was soon dropped [15] and developed into a theory of intersubjectivity [16].
In contrast to experimental situations, in natural situations we have not only an epistemological perspective, but also a shared existential situation. For children, it is not only important that their fathers play with them "correctly", but they also want to feel that their fathers enjoy playing with them. It is not just epistemologically important to the observer to be a mirror correctly reflecting an observation; rather it is of existential importance to both parties to have understood correctly-for life itself, or certainly, its development can depend on it [17]. It is this theoretical idea of a common field that makes empathy a cooperative achievement of two interacting parties. It is they who make empathy possible for one another or who prevent it, and they both depend on sufficiently understanding or being sufficiently understood. That can only happen through conversation, which must obey its own forms-every act of understanding has to pass through the eye of the needle of conversation (of course, not only verbal!) if it is to be effective in the social world [18]. It is more than just a question of epistemological understanding, and it is the social dimension of shared being.
This conception takes a critical view of the influential, still entirely individualistic idea of an empathiser who due to special sensory organs and fine sensitivity is able to look into or even penetrate the internal worlds of others. It is claimed that this can be achieved not only due to such personal abilities, but also with the help of psychoanalysis. To cite just one example: [We] can compare our theories to an optical device that enables us to perceive the innermost core of the latent content of the patient's discourse. [19] Here, countertransference is conceived as a kind of monitor on which the therapist can read the unconscious and suppressed sides of the patient. Interestingly, Kleinian [20] warned of this overestimation of theory and a renaissance of the "x-ray" vision theory of empathy when she criticised a "certain tendency" among psychoanalysts.
"…to be wrapped up in checking their own feelings as the crucial reference point for the session's events; this occurs at the cost of immediate contact with the patient's material" [20].
The empathetic field must thus overcome two risks: either contact is lost, or we have to believe, if we follow Paul Denis, that theory already knows everything. This too would entail a loss of contact; a patient who feels treated merely as a "case" in a general theory must feel permanently overlooked as an individual [21]. Therapy is then unable to do what it is capable of.
Here we can only point to the heterogeneous discussion of empathy [22][23][24][25][26][27] and provide specific analyses of therapeutic sequences, in line with a research strategy suggested elsewhere [28,29]. Using the "empiricism of conversation" (in the form of transcriptions of genuine therapeutic conversations), we expect to be able to provide a clearer description of how empathy is created. This forms part of an attempt to re-establish psychoanalysis as an observational science [30]. Psychoanalysis can look back on an "empiricism of conversation" without recourse to other methods and empiricisms that might be considered quite foreign.
The empiricism of conversation would be able to compare psychoanalysis with other therapies; such contrasts would then bring its own potential into sharper focus.
The key questions posed by our study are thus: how can we describe "preconditions of empathy", in which "sites of empathy" develop and how are "empathy achievements" generated in cooperation? What are the conversational practices realised in each fashion to create the preconditions, to recognise the sites and to acknowledge empathy achievements? Which conversation tools do the participants use to these ends and to signal to each other that the empathetic field is developing and deepening?

Method: conversation analysis
In order to get a grip on these questions, we decided to use conversation analysis (CA) as the most nuanced tools originating from the work of Goffman [31,32] and Garfinkel [33]. About recent developments in CA readers are best informed in Ref. [34]. CA has a keen eye for the microanalytic subtleties of therapeutic discourse creating atmospheres-or not. This is best understood by some illustrative examples.
Conversation analysis (CA) has long been concerned with empathy, without using the term, however. In the early days of the discipline, Harvey Sacks (1992, Lecture I from the winter of 1970) [35] stressed that CA had to consider one aspect of conversation he termed "my mind is with you"; in other words, we always consider the cognitive + emotional information available to another person. When we speak of "my brother Peter", this is a common designation for a member of our family-for the benefit of someone who does not yet know Peter. We tailor the content and form of our utterances to the "knowledge" (epistemic state) of the other person; conversation analysis calls this "recipient design" (more on this later). If we were to use the same phrase when speaking to our sister, such a designation would seem strange. We implicitly provide the exact balance of contextual knowledge required to continue the conversation [36,37]. It is not a matter of establishing consensus of world views, as "knowledge" would suggest, rather it is a question of the small particles of "knowing" necessary to allow the other person to be part of and make useful contributions to the "project" of a shared conversation. Collins [38] describes this weighing up of "epistemic states" as an example of the above process of "entrainment"-if such differences are not compensated for, strong feelings of aversion quickly develop. If on the other hand the correct balance is struck, warm and friendly feelings emerge. Although we speak of "knowledge", precisely these more affectively relevant consequences that are part of conversation analysis; it would be wrong to assume a focus solely on the "rational". Rather, it is a question of "situated knowledge practices". If this is achieved, knowledge is shared, combined with the feeling of being in a (small) world shared with a conversation partner. If situated knowledge can be shared, this sharing has emotional consequences.
Influenced by the empathy boom following the discovery of mirror neurons, some authors have begun to investigate empathy in conversations [39,40]. Only recently, these studies have been extended to prosody and melody of the voice [29,41,42]. The obvious premise is that empathy on the human level must be studied not only in the neuroscientific context, but above all as an interactive social practice. All of these authors express their surprise, however, at how little empathy is valued and how little it is used during conversation.
CA has established a firm place in the social sciences by the assumption that social structures are not only "at hand", not "just a given", but social realities are constructed, better coconstructed by participants in local circumstances. Participants use "practices" to contribute from moment to moment to the course of interaction, thus generating a trajectory of conversation while drawing back to cultural and common knowledge of many kinds. The medium of such practices is to a large part talk-in-interaction; thus, CA does not reduce talking to "exchange-of-information". It includes all kinds of bodily movements participants display mutually to each other. Everything hearable and viewable can become part of conversation. CA has studied in microanalytic detail the formal practices of turn-taking, the production of "trouble" (when participants, e.g. interrupt each other) and the practices of repair, how a topicshift is arranged, the organisation of laughter, how questions are posed and how they are responded (in court, police interrogation and in medical practice), how storytelling changes in dependence from the recipient (recipient design) and a lot of many other things occurring in human conversation; an overview is available [34]. Psychotherapy process research has attracted CA-authors because they are attentive to the "sweet little nothings" during conversation mostly overlooked by more generic "coding and counting" approaches [43][44][45][46][47][48]. The advantage of observing many highly relevant details of voice and prosody [42] or eye movement for coordination [49] is accompanied by the disadvantage that only small numbers of cases can become studied. Up to now, there is no generally accepted solution for this problem.
We want to present an exploratory design of a psychotherapy process research study. The type of process research described here entails language unfamiliar to clinicians. However, without expanding our terminology, we cannot do justice to our observations. We hope to find a language for the main therapeutic tools here: treatment through speaking (a "talking cure", in the words of "Anna O."). Our work is based on constructive collaboration at the International Psychoanalytic University (IPU) in Berlin and is a comparative, exploratory study of three therapeutic procedures (psychoanalysis = PA, depth psychology = DP and cognitive behavioural therapy = CBT). Treatment by five therapists in every procedure is examined at three stages: in an initial, a mid-term and a late session (from the final stage of treatment). We thus have, overall, a corpus of 45 transcribed sessions (3 therapeutic orientations Â 5 therapies Â 3 sessions). In this way, processes within the procedure can be monitored, opening or closing situations can be compared across therapeutic orientations and the influence of therapeutic personalities can be compared within one and the same orientation. All treatments involve patients diagnosed with depression. Almost all of the patients were judged to have had a positive "outcome" by independent assessors in the Munich study; adherence to specific procedures was checked: cases labelled PA (or DP or CBT) did indeed relate to PA (or DP or CBT) [50]. Differences can be ascribed to the therapeutic procedures. Here, the process is constructed with regard to empathy.

Pillars of empathy
In order to reduce the enormous complexity of empathy in therapeutic discourse, we found it useful to erect conceptual "pillars" that might help to present some of our data in an organised fashion. We start with what CA-authors have observed as empathic practice in everyday conversations.
3.1. First pillar: empathy in everyday talk-the CA contribution by John Heritage Heritage [37] calls verbalised participation in everyday conversations via common phrases such as "The same thing happened to me recently …" "parallel assessment" ("I'm-like-you experiences"). In professional contexts however, this type of empathetic participation hardly ever takes place. In an extensive set of records of conversations with general practitioners and homoeopathic therapists, Ruusuvuori [51] finds just one single example of such "parallel assessment". In professional context, the sentiment "I feel the same way as you" is hardly ever expressed. There are conversation practices that promote or prevent empathy on the everyday colloquial level. Hence therapeutic empathy clearly requires conversational preparation. CA has developed "tiered" list of practices through which people communicate empathy with each other in everyday life [39,40,[52][53][54][55], that is, the way they realise the "silent" but continual communicative dimension of "my mind is with you". These practices can be listed as "everyday empathy" [37] as follows: • "Response cries" [56]: exclamations such as "Aah", "Oh dear", "Oh no", used to convey emotional sympathy. For Goffmann, "response cries" are not signs of sympathy, but apparently involuntary expressions in reaction to unexpected events.
• "Pre-announcements" can announce a narrative: "A wonderful/crazy/funny thing happened to me yesterday". Listeners are informed in advance what sort of reaction is expected of them. More than just creating expectations, such "pre-announcements" provide the other person with a key to interpreting the ensuing utterances.
• "Ancillary questions": these everyday questions invite the listener to recreate the imagined scenery and to explore the necessary details.
• "Parallel assessments" denote the utterances one produces when a conversation partner relates stories, for instance about visiting the dentist, or attending a wedding or a funeral. The "parallel" moment is articulated in utterances such as "I feel the same way", "I've had that too" (e.g. an illness).
• "Subjunctive assessments": by way of illustration, Heritage [57] presents scenes in which two friends talk in the kitchen and imagine the ingredients they will use next time they cook something. And then make noises of enjoyment and pleasure to convey to each other how at this very moment they are imagining very similar sensations of taste on their tongues. Subjective assessments are tantamount to anticipation sensed together empathetically although they have yet to be experienced.
• "Observer responses": Heritage uses this term for those comments used to directly mention an external characteristic or a situation to someone, such as "You're speaking so quietly" or "You look exhausted". Such utterances are vulnerable to nonempathetic interpretations on the part of the recipient. In the therapeutic context, "observer responses" include phrases like "I've noticed that you …" Heritage provides several illustrations for this list in the form of excerpts from transcripts. They are too extensive to consider here, but they are very persuasive. We could observe practices of everyday empathy in our material, but there is more of relevant empathic practices in therapeutic conversation.

Second pillar: motive constructions
In test runs, we sought to encode the "student's" transcripts with Heritage's list and found it to be insufficient. One initial and important (re-)discovery was that therapists construct motivation in various ways [58]; they suggest motivations to their patients using expressions such as "because …" [59]. An "in order to" construction is also frequently used: "Then they went there again, in order not to be alone". In everyday life, motive constructions are very rare when addressed to others; only in very exceptional circumstances, one can ascribe motivations to other people without them feeling violated in their personal autonomy and going on the defensive. Thus, a therapeutic situation opens up for different types of empathic practice that are more or less tabooed in everyday life. The permission to use practices that were not tolerated in everyday life can be viewed as a specialty of psychotherapy as a form of institutional talk, permitting other forms of conduct. Such a permission must be granted by the patient which is done by conversational preparations for the emergence of an empathic field. If this fails, patients will tend to hear motive constructions as attacks, accusations and so on. Empathic achievements by therapists need active preparation.
Another empathic achievement is not to make motive constructions until such active preparation is agreed by the patient in order to make the difference transparent between everyday situations and therapeutic contexts. Such conversation practices secretly convey "my mind is with you".

Third pillar: observing expectations
In order for this "silent" dimension of conversation to unfold, conversation partners must develop practices establishing "common ground". For "talk-in-interaction" to emerge, it is essential that people convey that they have something or other to do with one another; for instance, one of them takes on the role of the speaker (commander, narrator) and the other adopts the position of the recipient of a command or the listener-and does not permanently "butt in". Analysis of the beginnings of telephone conversations [60,61] shows how tiny particles ("Hi!") articulate an expectation that the listener will perceive the speaker as someone with a certain identity. Such expectations are a relevant but silent dimension of patient's talk in therapy. To open one's ears for this hidden conversations is a relevant dimension of preconditions of empathy in therapeutic situations. To find answers that make expectations transparent and go over their restraints is a part of empathic achievement.

Fourth pillar: establishing "common ground"
When people meet each other for the first time, however, such an implicit suggestion of familiarity cannot be expected. They must first establish the cooperation that gives rise to reciprocal "commitments". Conversation analysis describes this as "adjacency pairs": a greeting is followed by its return; a question is followed by an answer. This is "conditionally relevant". Violations of such rules require good reason, and under normal circumstances accounts are presented; or they result in severe social rejection.
The concept of "adjacency pairs" has also given rise to "projective pairs" [62]. This term has nothing to do with "projection" in a clinical sense; rather it is derived from the idea of a "project"; participants demonstrate to each other that they are engaged in a common project. Someone who is moving house and wants to put together a cupboard with someone can say, for example, "We have to screw it together here", and without speaking the other person passes the screws within his or her reach-the project requires cooperative actions and at the same time forms a semantic frame. Clark (p. 129) speaks of "collateral communication". How a common ground is enacted by both participants is analysed in detail in Ref. [43].
To have an open ear to how common ground is achieved is a precondition for empathy, the response will be assessed by the feeling of being understood-or not.

Fifth pillar: deontic authority
The founder of speech act theory, John Searle, proposes a most useful distinction [63,64]. For the skill that arises "when the words fit the world" he coined the term "epistemic authority". Those in a position to aptly express the conditions of the world using words have epistemic authority. Distinct from this, we also have "deontic authority", "where the world fits the words". Those who can determine "what is going on (in the world)" through words possess the kind of authority Searle terms "deontic". This distinction has already been applied fruitfully in CA [65]. A priest, for instance, can suggest to the choirmaster which hymns might be sung on Sunday, but his deontic authority requires agreement, he cannot force the choirmaster to make any particular choices. Agreement can be withheld or rejection can be concealed.

Sixth pillar: Rupture-repair cycles
Some of the early literature on conversation analysis is concerned with "repairs" [66]. This focus extends to a variety of phenomena [67]. In psychotherapy research, detailed studies of transcribed therapy material [68] in particular have demonstrated where therapeutic working relationships fail to develop and how they can be "repaired". These authors describe two possible developments: the patient either withdraws or becomes aggressive and accusatory. The decisive element is whether or not the therapist notices or blithely carries on. If a "rupture" is noticed, "repair activities" can be employed. However, it is not precisely clear, what "rupture" means: rupture of what? Of the conversational tissue? In clinical language, rupture can be used metaphorically in some helpful ways; however, if you want to study conversation, a metaphoric use is insufficient. Here CA has developed a rich register of rupture and repair activities.
Repair activities are also used in everyday conversations [69], for instance, when someone restarts a sentence [70] or is corrected on a statement regarding time or place [71]. Repair activities are employed very frequently and are regarded by many authors [72][73][74] as the basis of survival from infancy onwards. Thus, the detection of a rupture is a precondition of empathy, to find a helpful answer is a therapeutic achievement.

Seventh pillar : typical problematic situations (TPSs)
Some clinicians assume that a given treatment's chances of success depend on how "problematic situations" are handled. Such situations are accusations made by patients, postponed appointments, cancellation fees, late arrival, applications to extend health insurance claims and expert reviews and other "performance defects" [75].
But there are also difficult situations that can only be detected by microanalysis: a patient after having told a dream asks the therapist "Did my dream help me?" and the therapist responds with some confusion [76]. A patient does not finish a story, but breaks off mid-sentence with rising intonation-such a "border tone" gives the therapist a clear signal that the patient wants to continue speaking but has paused for thought. But if this pause lasts more than 3 s and stretches to 27 s [77], it becomes difficult. In an example published elsewhere [43], the therapist felt the need to finish the patient's sentence for him. In this case, the intention to help the patient get over a "stumbling block" resulted in the escalation of a fight for the right to speak.
More complex TPSs arise when patients seem to communicate "I urgently need your help, but nobody can help me, not even you" or "I urgently need your help, but on my conditions". Details of such TPS are presented elsewhere [43].
To detect a TPS is a precondition for empathy, and to find a suitable answer is an empathic achievement.

Findings
Let us begin with an initial interview in which the "student", who has already undergone various examinations, is in conversation with the therapist but whose description of his symptoms makes it particularly difficult for the latter to form even the slightest impression of what he is talking about.

Transcription symbols
Reading transcripts requires some practice, just like reading statistical tables or diagrams. Here is a key to the symbols used: • Words in square brackets are spoken at the same time.
• Colons: pronunciation of a letter is stretched out.
• Underlined words or letters: spoken with emphasis.
• Words in UPPER CASE with ! are spoken loudly.
• Quiet words or sentences are indicated by .
• Numbers in brackets indicate pauses in minutes: The precise transcript provides a helpful explanation of what happens here. The arrows show the sections to which our commentary refers. Before the patient reacts with a quickly spoken ">dann net<" (">not then<"), he pauses ever so slightly. The effect of this micropause is that it is impossible to distinguish whether "dann net" is a negation of the content or whether the "net" is a small Swabian tag. A "tag" is a signal often added to the end of sentences to seek agreement, commonly encountered as German "ne", Hessian "gell?" or English "isn't it?". Initially, the localisation of the compulsion is obvious, but then we recognise that the therapist is having difficulties coming to a clear interpretation. He checks (second arrows) and receives another answer that makes it unclear whether the patient's answer is to be understood as agreement or whether the "yes" is merely a particle used to introduce the next statement. A further difficulty is encountered in the negation in the patient's reply. The utterance "When you go in the door of your house", seeking clarification, is ignored not directly, but indirectly (via a description of what he does not do). The therapist senses this irritation too when he asks the patient again (third arrow) if he checks (from outside). Now he receives the answer "no from inside". Here an asynchronic dance develops in which it is not clear who is supposed to take the lead. A further difficulty is added by the patient's delayed explanation of his direction of movement. Despite this imprecision, they continue with the conversation and thus mix up the new question regarding the patient's location when the compulsion occurs with the previous question regarding his direction of movement, a clear answer to which has yet to be provided.
One of the practices attempted by the therapist in this example is known as "ancillary questions" [53]. Such questions are designed to form a picture of when and where the patient's compulsion arises-but in this case this is precisely what they do not do. It is not empathy that fails here, but one of its precursors. Actively but unconsciously, the patient paralyses the therapist's hermeneutic abilities-we speak of "communicative stun grenades" [79]. Of course, there are much more severe examples. Clinicians might be reminded here of Bion's descriptions such as "attacks on linking"; unfortunately, to our knowledge, transcripts have yet to be made available. We hold that this brief example gives rise to a number of issues: the idea that patients want to be understood does not do justice to the whole picture. Nor is the complementary idea that therapists can understand empathetically without the active contribution by the patient is the whole story-yet this is exactly the description we repeatedly find in theories of empathy.
By way of illustration, let us further differentiate the concept of the interactive field using a small diagram (see Figure 1).
There are preconditions of empathy (E-P). The minimum requirement is seeing or hearing (e.g. on the telephone), that is, the co-presence of the parties involved. Here it becomes clear that there is a productive and a receptive aspect; one person speaks, the other hears that person's voice. "Talking to each other" is another prerequisite, but this alone does not constitute empathy. It is possible for two people to talk to each other and understand each other only on the linguistic level; there are also many situations (with waiters, at counters, etc.) in which people talk to each other and empathy is practically neglected. The therapeutic situation, on

PrecondiƟons of Empathy (E-P)
Achieving Empathy (E-A) the other hand, specialises in coproducing "sites of empathy" (SE) in order to observe them. In particular, this includes irritations in everyday communication and physical and other sensations that arise. As the process itself develops, these aspects are in turn produced by speech; the act of talking to each other becomes increasingly self-reflexive and produces new irritations and ruptures and difficult situations. The findings described allow us to categorise our expansion of conversation practices in the diagram above.

An example for common ground and nonresponding to expectations
We can now apply the above considerations to the beginning of a depth-psychological (DP) therapy. The patient has received prior instruction that her treatment will be recorded; at once, this part of the project becomes an opportunity to create "common ground" with which to begin a conversation: Second example: DP, initial session The therapist states the date of the session for the tape; the recording is turned off (we do not know for how long), and the patient makes a first quietly spoken comment about something "visible": " Oh I see you have already (.) got the microphone set up; ".
Opening in this fashion with "joined attention", discussing a little thing that is present and perceptible to both people, is a common tool with which "common ground" in sensory The Building of Empathy: Conceptual "Pillars" and Conversational Practices in Psychotherapy http://dx.doi.org/10.5772/intechopen.69628 perception to the zone of conversation can be transformed. P sees something, sees that the other person sees something, and addresses it, an example of "silently" communicating that "my mind is with you". This is part of the patient's empathy for the therapist, delivered as searching for opportunities for common affective regulation, for instance overcoming initial awkwardness. The patient displays a memory of what was agreed (audio-recording the session) and displays an expectation a) that her pointing to the sensual object (microphone) is confirmed by the therapist and b) that the therapist remembers the agreement, too. Answers to both expectations would share the experience of a common ground. The therapist reacts with a loud "YES!" and loudly breathes in and out.
This minimalist utterance acts as a context marker. In an everyday situation, one would expect a complementary reply, commenting on the preparatory setting up of the microphone. The brief, loud and accentuated "YES!" marks a different, professional, or, to be even more precise, procedure-specific context. A note on communicative markers: While doctors communicate directly and consistently via a white coat and a stethoscope, for instance, that they are doctors, a therapist has no such material context markers. She has to accentuate the project known as "therapy" using communicative markers and to distinguish it from similar types of conversation-such as gossip [80,81], conversations with friends or colleagues, or an interrogation. The psychoanalytical concept of "neutrality" must be realised communicatively. Such realisations fulfil multiple functions.
The therapist's "YES!" is recognisably short and hence can be interpreted in many different ways by the patient. It contains the briefest possible confirmation of her first comment; "YES!" could mean that she has been "heard", but it might also mean more and it might convey acknowledgement of and agreement with the content, that is, the comment about the microphone. Since both interpretations are left open, her response communicates that a particular kind of conversation is going to develop it points forward. The common project of "therapy" has been opened.
The patient now reacts by talking to herself; she expresses that she has "forgotten again". The rhetorical figure of ellipsis 1 in line 8 can be employed because it is clear what she has forgotten: that there would be a microphone. She knows that the therapist knows what the unspoken part of her utterance refers to: "My mind is with you" is the therapist's tight response. But the patient offers to pay a price for the "common ground": she diagnoses a momentary failure of her memory. In her first utterance, however, she had recognised the microphone as something that reminded her of the agreement for being audiotaped.
A small sliver of "common ground" is beginning to develop, although it is still quite fragile. It is not yet clear to the patient whether the converse is true or not: "Is your mind with me?" As in her opening remark in line 8 her attributing to herself some little weaknesses-forgetting- 1 Ellipsis [82] is a rhetorical figure enabling economy of expression. If one goes into a bar and says "a beer please", it is clear from the context that one means "I would like a beer". A certain part is omitted because it is clear from the context what is meant. Hence, it is sufficient for the patient to express herself in this abbreviated form. She can make use of the "common ground" that has been established and builds on it.
brings something to the communication that is already visible to both parties. At the same time, it communicates something new, more than just having forgotten something: the patient positions herself [83] as nondominant in relation to the previous sharp and markedly loud "YES!" The therapist quickly concludes with "Yes"; she acknowledges what the patient has said about her forgetfulness.
To conclude: the "common ground" is sufficiently established for ellipsis to be used, but the question of who is in charge or how the relationship will otherwise be defined has yet to be answered. The patient falls in pauses. By raising her voice slightly, the therapist tries to restart the "interaction engine" [84] and says her "Yes" (line 11), which, when one listens to the tape, is reminiscent of many utterances therapists use to prompt their patients to speak. Now, the patient says: Third example: DP, initial session, continuation: Again, the first remark takes up something that has already been spoken about. Again the therapist responds in the minimalistic way possible. The patient now takes the initiative (line 13), makes sure that the therapist knows what she is talking about and concludes this clear series of articulations of "my mind is with you" with a final "=Okay:.", drawn out and with falling intonation. In clinical terms, one would speak of "clarifying the referral context". However, something else occurs here to allow empathy to be established. Besides the referral context, the question of whether and to what extent the therapist knows why a recording is being made is negotiated. Once this has been established, the patient knows that the therapist knows, but again she does not yet know whether the therapist will also realise an attitude of "my mind is with you". The fact that the patient, despite her many attempts to initiate a conversation, cannot know if this is the case could be perceived to be a difficulty arising from a rigid approach to the concept of neutrality. This concept demands a therapeutic attitude that shows the patient nothing more than the patient has shown herself-as Freud puts it with his metaphor of the mirror. The therapist marks the professional difference from everyday communicative practices, and at the same time she makes it difficult to establish "common ground"; the patient cannot know anything about the relevant dimension of the question "is your mind with me?" At this point, we do not wish to present any more material from this conversation; it suffices to establish that here we have a special realisation of the therapeutic concept of neutrality-in contrast to a "silent" response to the question "is your mind with me?" We can recognise some problems pertaining to this particular realisation, but also to the concept per se.
By way of comparison, we will now turn to an opening scene from a second CBT session. We are not privy to the first session. We have a male patient and a female therapist: Fourth example, CBT, second session 1 ((Recording begins, silence, someone can be heard slowly turning pages)) 2 T: exactly. T: genau.  T: >I mean we will certainly come back to this, but then<= T: >Also wir kommen sicher hier auch immer wieder drauf, aber dann<=

(-)
The Building of Empathy: Conceptual "Pillars" and Conversational Practices in Psychotherapy http://dx.doi.org/10.5772/intechopen.69628 We can immediately recognise an entirely different structure to the conversation. The speakers shift rapidly (shown by =). It is the patient who makes the first pressing remark (line 4). It is the therapist who speaks much more than the patient; the subject is the completion of a questionnaire which "saves us a lot of time" (line 13)-here too the focus is on being pressed for time. In quick succession, the patient gives off signals that he is listening, which seems to spur on the therapists' fast talking. The therapist plays down the questionnaire on his life story in two ways: they will then have a "condensed" version and can "get into it much quicker"-into the project of common therapeutic work.

Deontic authority
"Deontic" is derived from the Greek "deon" and means the cohesive force, the binding effect of utterances, and is thus an apt term for the topic of "commitment". This effect must be expressed, employed and articulated via conversation, and then, in a second stage, it must be answered by the listener, with either agreement or resistance or disagreement. The archetypical deontic modal verbs come into play here: "must", "should", "may". Deontic authority can look backwards, if it is a question of what one should have done, but it has greater significance when it comes to decisions concerning shared future activities. Then, the implicit question arises as to who has the right to announce or suggest decisions and ultimately to make them. The second speaker's answers become significant in steering further interaction. This second turn can entail agreement or disagreement and can conceal it or delay it.
We have already seen an example of this in the first DP session we examined, when the therapist (line 12) expressed her loud "YES!" with its emphasis and the patient did not follow, but asked another question (line 13). In response to the next attempt to begin with the project (line 23), when the patient receives the invitation to "simply talk about" what "is most important" to her, she reacts by saying that Doctor Innerst recommended the therapist to her.
A similar constellation illustrating the significance of deontic authority can be observed in the opening sequence of the "student" case: Fifth example: opening of the "The Student", initial interview  The German "so" with which the therapist begins here is a somewhat prototypical particle of deontic authority [85]. It performs a variety of social functions. One of them is ending one sequence and opening the next-in this case the act of both people entering, which we clearly recognise from the footsteps on the tape, comes to an end and something else begins. The therapist is entitled to use "so"; this marker of completion/opening requires agreement however. A patient could respond entirely differently. The therapist's second, louder "SO!" firmly reinforces the impulse to begin the therapeutic conversation; he then makes a suggestion, the introductory "perhaps", which suggests implicit knowledge that this form of authority is dependent on the listener; he can suggest something, but will it be accepted?
With the words "perhaps (2) >>let's talk about what=brings<< (1.2) you=here?" the therapist initiates the project of therapy, and given our discussion of conditional relevance, the patient would now have to obey the friendly command. But he reacts with a question that forces the therapist into the complementary role of the respondent: he cannot avoid answering. Deontic authority's potential to make decisions relating to the (social) world is dependent on cooperation, confirmation and collaboration and is thus constantly attempting to negotiate. In clinical terms, we can speak of the first signs of a power struggle.
We can observe similar patterns in the CBT session, since here the patient takes the initiative through his first remark (line 4) " shall we start now? ". Before they can start, the therapist is still focussed on a "different project", the questionnaire, and pushes this project through hurriedly while attempting to justify it. From line 33 onwards he begins the "project of therapy" with an expression that seems more appropriate for the friendly private context than the professional: ".hhh yes now [ho:w] have you been this week". Her deontic authority receives no more ratification than elsewhere, however: the patient does not answer the ques-  question then becomes a relevant element of the interaction. It is not about the missing content, but about who is allowed to speak. There is usually something that we would label "compromise" in clinical terms: the person asking the question repeats it and deals with the lack of an answer by reacting as if the other person had not understood the question properly. Inevitably, that impacts on deontic authority and its balance throughout the interaction.

Rupture-repair
Just as mothers cannot always respond with complete empathy to the demands of their children, but make a number of adaptations, therapists too adapt when they notice that they have got things wrong. Repair activity is explicitly oriented around something that has already occurred between the participants and has been perceived by both of them; it is an attempt to develop a level of conversation about these events. If repair is successful, the patient can have a multi-dimensional experience that goes a long way to promoting empathy: he notices a. that another person notices his retreat; b. that this person shows he is prepared to reset himself and his contributions; c. further, how the other person responds; and finally d. whether that person's activities are helpful and clear something up. This is of great importance for the development of cooperative trust. Often, the patient cannot know whether or not the therapist is actually trustworthy; at the same time, everything hinges on this question. However, if the patient can recognise that the therapist has not missed something that was bothering him, and that the therapist has introduced repair activities, then he can begin to form a positive impression of the therapist's "mindfulness" [86].
The term "rupture" describes a situation in which the cooperation described above fails, in which the participants cannot anticipate which local role functions their counterpart will adopt, which "pair" they will become. Of course, "ruptures" can also occur at later stages of therapy, for instance, if a patient recognises disloyalty or similar violations on the part of the therapist. "Ruptures" are not connected to opening situations. If therapy starts with difficulties that does not mean that the entire treatment is doomed to failure. Rather, repaired ruptures can often lead to a better relationship between the patient and the therapist.
Let us examine an example of depth-psychological treatment from the same initial session. The patient had somewhat hesitantly embraced her therapist's attempts to start the session and had told her something about her life. She concludes this narration with a brief coda: Sixth example: the same DP as in the third example, initial session 146 P: ((some words omitted)) And then I also realised that I: yes (1,7) that has come to mean quite a lot to me professionally how I am feeling: (-) that I (---) h (through the nose) (--).hhh (1.5) P: ((einige Worte ausgelassen)) Und ich hab dann auch festgestellt, dass ich: To use Heritage's terminology, the therapist makes several "observer responses" ("that you really pulled yourself together", "that you tried to be very controlled"). In both form and content, these responses are much more extensive than they would be in the everyday setting. The specificity of the professional conversation is being constructed. It is introduced by softening expressions such as "I get the impression" (line 150), which are used in "portioned" fashion. The therapist waits for the patient's ratification, sometimes expressed hesitantly as "yes well" (line 152), sometimes as a decisive "su::re" (line 155). Ultimately, the therapist discovers another reason for her patient's reticence: the transfer mode, "always (.) starting to >re:late> everything a:: gain,>" (line 165).
Here the important thing is not only the information about the reasons for this understandable discomfort, but the fact that something can be "cleared up" which the patient would otherwise have left unmentioned-and thus unheard-throughout the rest of the conversation.

TPS
Here too we can see the difference between the clinical description and what can be observed via microanalysis. For more details, see [43]. Let us examine an example from the middle stages of the above DP treatment. We hold that it is instructive to study such examples in this level of detail, since insights can be gained that can then be applied to similar situations. The example in question is a disagreement caused by a potential error regarding a cancellation fee.
Seventh example: same DP, 50th session, opening. We have what appears to be a trivial difference. The invoice was for a session, but the therapist had made a mistake and readily admits to it. The patient for her part wants to inform her that she has made a kind of discovery. If we take a close look at this opening to the 50 th session, we can observe various small relevance markers. The patient "just" (line 2) wants to come back to it, she plays down the matter and wants to talk about her discovery; the therapist makes it highly relevant (line 6), culminating in self-flagellation: "Especially as you had drawn my attention to it as well. (---) in the previous session". She then says she hopes that the patient is not angry and finally stresses in a much louder voice with strong emphasis that she is sorry and that it is "not a problem!" Only once she has been "absolved" by the patient the latter can proceed with what she has to say.   We can clearly recognise how the patient speaks of finding "a kind of trust", but this trust is then badly damaged; it further shows how she associated that with an earlier experience ("destroyed what I had built up as it were") and how she noticed that she does not have to constantly justify herself. This piece of productive therapeutic work shows that even a TPS that appears to have set in early on in the treatment can be dealt with in such a fashion that the patient is able to experience self-enlightenment. This example can be understood as a meeting of two "troubles tellers", each of whom find the complementary position unoccupied. The therapist is seeking exoneration, the patient is seeking an "ear" for her discovery. However, two "troubles tellers" cannot form a conversational pair-but it was possible to "re-pair" this local irritation.

Discussion
There are moments of profound reciprocal merging which give voice to physical processes of "limbic resonances" [87,88] and in which people's knowledge about each other becomes transparent-although usually described as "unspoken"-and which would not have been considered possible without this experience. Most people, be they therapists or not, would describe such moments using the word "empathy". In fact, however, clinical experience tells us that the path to such "sites of empathy" is stony and full of obstacles. Does empathy not occur via these routes?
Here, we could show how sites of empathy can be recognised and responded to. We could describe preconditions of empathy, e.g. repairing activities, not to early motive construction and how empathic achievements are generated in cooperation of therapist and patient. A TPS seems to be a most relevant precondition for empathy, therapists should be trained in recognising and responding to them.
There are situations in conversations in which empathy is clearly strived for, while at other times it is directly impeded-but people continue or begin to talk with each other nevertheless. As long as one is "in conversation", empathy has a chance. Conversation consolidates empathy, which can only be created "in conversation"-the coproduction of empathy. Accordingly, one can distinguish the conversational preconditions of empathy from its actual realisation. Further, there are certain "sites" in which empathy must prevail in a particular fashion. Ultimately, the conversation partners must work together if their conversation is to be deepened and their relationship is to develop.
Let us now represent and in doing so summarise our project of an "architecture of empathy" by use of a diagram (see Figure 2).
Empathy needs preconditions (E-P), as simple as, e.g. being in a conversation in contextual circumstances as therapeutic or mother-infant situations. Such preconditions can result in an empathetic "path" or, on the other hand, in its termination. We posit further "sites of empathy" (SE), in particular TPSs and RRCs (rupture-repair cycles). Such "sites" outline a special opportunity where empathy can be performed. This is bidirectional.
Finally, we have "empathy achievements" (E-As) that are by no means the work of the therapist alone. They are based on sites, on preconditions arranged by the institutional rules of conduct and by the patient's offers. Insufficient research has been conducted on patients' achievements of empathy for their therapists and their mistakes [89]. Sites, preconditions and empathetic achievements constitute the "empathic field" that is communicated to the participants via the growing security of significant "common ground". Common ground goes beyond the empathic field into the domains of language, culture and societal discourse. This field is represented as an oval surrounding the above diagram. The empathetic achievements are communicated in the form of motivation constructions, individualised recipient design and RRCs. We must emphasise once again that this is by no means a one-way affair communicated by the therapist to the patient, but a reciprocal process.
We hope to have made it clear that this research into the psychotherapeutic process has the capacity to examine empathy as an aspect of therapy consisting of many different parts. The whole we term the architecture of empathy as it is realised via conversation.
Many authors quite rightly stress that in inquiries into empathy it is the tone that makes the music, and thus, prosody must also be examined. To date, our own studies have proven unusually complex [42]. Our methodological objective will be to use the CA approach described here to identify some striking passages in which something seems to be "happening". Those passages will later be examined through the prism of prosody in order to explore whether something is indeed happening and what that might be.
We hold that this type of process research is of benefit to clinicians. While we have used unusual terminology, it is necessary if the study is to be useful. Clinical readers may consider this use of terminology to be unreasonable-the benefit is that one day we will have a better understanding of our main therapeutic tool, treatment via talk-in-interaction. Clinicians should merely recognise that there is a supplementary set of conceptual tools which, we hope, can help provide comprehensive analysis of the complex conversations that take place in everyday clinical practice.
indebted to Prof Dr Dorothea Huber, Munich/Berlin, for providing this material from the Münchner Psychotherapie-Studie (Psychotherapy Study) [90]. The first example is taken from the Ulm corpus of texts (Ulmer Textbank), "Der Student", Initial Interview.