Poisoning children: A reflection on the social constraints of children’s agency and narratives of sexual abuse

Introduction

At the end of the 1990s, a story of abuses involved 16 children in two small towns in a rural and wealthy area of Northern Italy. This story had great resonance on the Italian media and attracted the attention of the Italian political institutions. The story began when a child, known by the pseudonym Dario, reported sexual abuses. At the end of 1993, when he was 3 years old, Dario was separated from his family, who had financial and housing difficulties. At first, he lived in an institution for children and then in a foster family, visiting his family twice a month. In early 1997, Dario started to report sexual abuses to his teacher, his foster mother and then a psychologist working in the local social services. He claimed that other children were involved in the abuses, and these children confirmed the abuses, mentioning in their turn other children, in a snowball process that shocked the two towns.

The children were separated from their families, sent to dedicated institutions and then to foster families. They were interviewed by a psychologist, their abuses were certified by gynaecologists and later were interviewed by judges assisted by other psychologists as court counsellors. These children reported sexual abuses within families, sexual abuses in other places with parents’ complicity and satanic rituals in two cemeteries of the area, with the participation of parents and under the guidance of a person, who was identified as the priest of a local Catholic church. The children reported that other children had been killed and either buried in the cemeteries or thrown in a nearby river. They reported on themselves being abused and obliged to abuse and kill other children.

The parents, the priest and some neighbours were indicted. All these people claimed their innocence. A few of them, including the priest, died of heart attack, and a mother committed suicide. At the end of several trials, the court dismissed the stories of satanic rituals as no evidence was found about them. However, many parents were sentenced for sexual abuse. The parents, whether found guilty or not guilty, no longer met their children.

This story has been (re)told, with new documentation, by an audio-recorded podcast, called Veleno (Poison), hosted in the online version of an important Italian newspaper, in seven episodes in October–December 2017. The journalists collected the court procedural documents, interviewed many people involved in the case, and discovered new video and audio recordings of conversations between the children, the judges and the psychologists. In the light of this documentation, Poison’s conclusion is that the children’s reports were fictitious, the psychological evaluations wrong, the gynaecological visits very dubious and that the court had therefore arbitrarily sentenced many parents.

Poison is a second order observation (Luhmann, 1995), that is, the observation of another (first order) observation that proposes ‘facts’. Second order observations deal with the way in which first order observations are produced. Poison is a documented second order observation of the institutional observation of children’s abuses as facts. Poison does not provide any methodological account of the ways of selecting the published documentation. However, the point here is not whether Poison has adopted an accurate methodology of second order observation. The point is that this documented observation makes a sociological second order observation possible: in particular, this article proposes a sociological second order observation of all video and audio recordings, made public in Poison, regarding interactional procedures through which the children disclosed the abuses. Since the 1990s, the Sociology of Childhood has been looking at children as social agents (James et al., 1998). The analysed recordings highlight the ambiguity of children’s agency, especially their accounts of abuse, and the way in which social institutions influenced these accounts.

Section ‘Truthfulness of children’s accounts’ deals with the truthfulness of children’s accounts through the lenses of an approach combining a theory of narratives and the concept of agency. Section ‘The construction of children’s narratives in institutional interviews’ analyses the institutional context in which the children’s narratives and agency were constructed and validated. The final section summarises the analysed social constraints on children’s agency and narratives and draws some conclusions based on this analysis.

Truthfulness of children’s accounts

Poison questions the truthfulness of the children’s accounts of abuses as based on stories of satanic rituals. The court documents show that Dario introduced the story of satanic rituals during his conversations with the psychologist. The psychologist reported to the court Dario’s story regarding a ‘funeral’, adding, however, that he was not sure whether that was a true funeral: he had been closed in a coffin with two other children; they were very scared because it was dark and ‘he was not able to go out because the lid was too heavy’. These children were thus ‘turned into children of the devil’. In the following months, many other children told the story of satanic rituals. Some of their reports to the court are shown below (translated in English): the first is a court record, and the others are taken from video recordings discovered by Poison:

They made us throw them in the air, they fell down and then they put them in a tarpaulin, they let the children’s blood drip and obliged us to drink it. Then they told us that we were children of the devil because we committed murder. Then, once I killed a baby with a knife. My father too was holding him and I stuck a knife into his heart. (Episode 4)

We also had some cats and we killed them. My father made us drink the cats’ blood. (Episode 5)

Then they opened (.) some children here and all the blood came out, we could see some bags […] Then they made us kill them (1.0) I had to kill five of them okay, more, but five in a night. Then eh (1.0) I had to kill many of them (3.0) But we went there three times a week and after those three times a week all nights I had to kill five of them. (Episode 5)

I had to stick a knife in the heart of the first child, then to another one, I had to tie the second to a block of concrete, I had to leave a fire stamp in the fire for one hour and then I whipped the third one and then he died and I had to stick a knife in his back. (Episode 6)

Poison assesses these stories as clearly not credible, as the court also recognised: although the children reported 15 children killed per week, no corpse was found; no children drinking cat blood and tortured were sent to hospitals; no teacher or paediatrician noted children’s abuses or wounds. How is it possible to explain the narration of these stories?

Narrating stories is a feature of all communication processes (Fisher, 1987). Communication produces situational, as well as historically and culturally grounded narratives, based on specific sets of categories. Somers (1994) describes four different types of narratives: ontological narratives making sense of people’s lives, public narratives, and conceptual narratives based on scientific concepts and metanarratives concerning ‘the epic dramas of our time’ (p. 619). According to Somers, ontological narratives are embedded in public narratives, and metanarratives embed both public narratives and ontological narratives.

In this view, the children’s stories of abuse are ontological narratives. These narratives are usually transformed in clinical (medical and psychological) conceptual narratives, which in their turn are the basis for the public narrative of the abuses. A metanarrative of children as trustful and innocent victims, combined with a child-saving metanarrative (Wells, 2015), is in the background of these conceptual and public narratives. This metanarrative supports the clinical narrative, leading to treat children as in need of radical intervention. In the case documented by Poison, the court provided a final public narrative, accepting the clinical narrative of sexual abuses but denying the story of satanic rituals. Thus, the court denied the causal relation between the clinical narrative, on one hand, and the children’s ontological narratives, on the other. The court, however, did not re-evaluate the truthfulness of the children’s ontological narratives.

Poison produces its own metanarrative: children are psychologically fragile and exposed to collective imagination of blood and violence, drawn from a variety of media; the belief in children’s imagination depends on panic for Satanism. To support this metanarrative, Poison cites similar cases in the United States; other cases could be cited, for instance, in Scotland. Poison’s additional explanation, however, focuses on the children’s traumatic separation from families, which convinced them of their parents’ misbehaviour, and on the psychologists’ inadequate methods of collecting children’s stories. A sociological analysis can reframe this explanation, and the corresponding metanarrative, focusing on the institutional conditions of production of children’s narratives as social constrains on children’s agency.

Agency means availability of choices of action, which opens up different possible courses of action (Van Langenhove and Harré, 1999), thus enhancing social change. Agency implies the attribution of authority in accessing and producing knowledge, that is, the attribution of epistemic authority (Heritage and Raymond, 2005). Children’s epistemic authority is displayed through ontological narratives regarding their lives and through their contributions to public narratives. Children’s agency is thus displayed as availability of choices about what is narrated and how it is narrated. The distribution of epistemic authority is shown through participants’ epistemic stance, that is, through their specific actions that enhance communication. The social construction of children’s narratives depends on the distribution of epistemic authority in the social relations.

These relations constrain children’s agency (James and James, 2008). A hierarchical generational order of social relations determines the difference between children’s limited opportunities and adults’ much wider opportunities to practice agency (Alanen, 2009; Bjerke, 2011; James, 2009; Leonard, 2015). Children’s agency is impeded by ‘unresolved tensions, ambiguities and social power relations’ (Fitzgerald et al., 2010: 293). Children’s participation is instrument for the smooth functioning of institutions, based on control over children’s actions (Hill et al., 2004; Thomas, 2007) and children’s adaptation to adult-driven frameworks (Prout, 2003). Against this background, children’s agency is effective only if it involves children’s ‘everyday life arenas and practices’ (Percy-Smith, 2010: 118) and if it is in line with children’s personal life (James, 2013).

Moving from the data made available by Poison, section ‘The construction of children’s narratives in institutional interviews’ analyses the way in which epistemic authority was distributed in institutional interactions as shown through participants’ epistemic stance, in particular observing (1) how the social conditions of children’s agency and the social construction of their ontological narratives were linked, and (2) whether these social conditions were in line with the children’s personal life.

The construction of children’s narratives in institutional interviews

The interactional construction of narratives does not only concern their content but also, and above all, the rights of narrating (Bamberg, 2012) and therefore the attribution of epistemic authority. Three aspects are particularly relevant to analyse epistemic authority. First, each participant may contribute to produce and negotiate narratives in the interaction as a teller, a co-teller, a listener, or an elicitor of new narratives (Norrick, 2007). Second, narratives can be either first-person narratives or vicarious narratives, the latter being narratives produced about or on behalf of someone else (Norrick, 2013). Third, narratives can show the meaning of the identity of the narrating person (Bamberg, 2005). In this perspective, it is not important why the children told their stories, but how these stories were produced.

The narrative production of child abuse is formalised through institutional interviews. Forensic studies, including both forensic psychology and analysis of conversation, stress that interviews with children who are witnesses and/or victims of sexual abuse require particular care to check reliability of information. This is shown through the analysis of interactions (Childs and Walsh, 2017), as well as the review of previous studies (Goodman and Melinder, 2007). Checking reliability minimises influence on children and children’s reluctance to disclose abuse (Hershkowitz et al., 2006). Interviewers and children are co-authors of these interviews, and interviewers’ actions are important in co-telling narratives of abuse (Deckert, 2010). In particular, interviewers’ questions show the distribution of epistemic authority (Jol and Stommel, 2016): on one hand, children know more than interviewers; on the other hand, interviewers exercise the authority to check children’s knowledge. While children’s knowledge is obviously important, interviewers’ questions highlight the interactional asymmetry between adult experts and children, their different mastery of situation and interview (Deckert, 2010). Moreover, the narratives produced in these interviews ‘are inherently shaped by the legal context’ (Deckert, 2010: 202), which constructs the identities of witness, victim and perpetrator through the categorisation of different actions. The victim is ‘acted upon’, the perpetrator is the agent of action and the witness is the agent co-constructing both victims’ and perpetrators’ identities during the interview. In these interviews, children are both victims (acted upon) and witnesses (agents): the failure of their agency in narrating determines the failure of the construction of identities of victim (and perpetrator). The construction of children’s identity as witnesses is therefore fundamental in the interview.

In this article, a conversational analysis of institutional interviews focuses on the design of participants’ turns and the organisation of the sequence of these turns (Heritage and Clayman, 2010), in particular analysing how these turns are intertwined in the interaction. This type of analysis focuses on the ways in which, by responding to each other, interlocutors make sense of the interactional dynamics. The analysis of interviews is, however, developed beyond the conversational dynamics, to show (1) the structural presupposition of these dynamics, and (2) the stories that are constructed in conversation.

The data that Poison makes available for this analysis are fragmented, but relevant: they include a short audio-recorded conversation involving a child, the psychologist working for the social services, the prosecutor and a police officer, and short video recordings of conversations between the children, judges and psychologists. Here, the transcription of these recordings is based on a simplified version of the conventions of Conversation Analysis (see Appendix 1).

The first extract is taken from an interview that was audio-recorded in a police car (episode 1). In the first phase of this interview (turns 1–5), the prosecutor invites the child to choose where to go. The child answers without hesitation, but she does not provide any specific indication (turn 4). The psychologist suggests that they head to go to the cemetery (turn 6), a destination which is confirmed by the child. After asking for directions, the psychologist also suggests that the child may be scared and invites her to express her fear to make it understandable to the adults (turn 8). Thus, she indirectly suggests that the child could conceal her fears. The police officer provides a third suggestion, about the possible actions linked to the bridge indicated by the child (turn 11), and the child immediately provides a story of killing. However, after providing a few details, the child hesitates. The officer repeats the gist of the child’s answer to confirm it and asks for further details (turn 13). This time the child’s answer is uncertain, but the officer ignores this uncertainty and produces a new question (turn 15). His following minimal response (turn 17) signals his waiting for further details, which are provided by the child with explicit uncertainty (turn 18: if I remember well). Once again, the officer ignores the child’s uncertainty, focusing on the adults who participated in the ritual (turn 19). After a short pause, the child mentions his parents:

1. Prosecutor: Quindi questa strada mi dicevi che la la conosci?

        So, you were telling me that you know this road

2. Child:    Sì

        yes

3. Prosecutor: Scegli tu i posti dove vuoi che vuoi andare a vedere

       You can choose the places where you want that you want to see

4. Child:   Dritto

       Straight

5. Prosecutor: I posti che ci vuoi raccontare diciamo, indicare

        The places that you want to tell us about, let’s say point out

6. Psychologist: Possiamo andare al cimitero?

        Can we go to the cemetery?

7. Child:    Sì

        yes

8. Psychologist: Ci fai vedere dove si va? Quando non vuoi più, che hai

        paura, lo devi dire, altrimenti

        noi non possiamo capire

        Can you show us the way ? When you do not want to anymore, because you are scared, you have to say, otherwise we cannot understand.

9. Officer:    C’è qualche posto particolare, che ti ricorda qualcosa?

         Is there any specific place that reminds you of something?

         (.)

10. Child: Quello lì

      That one

11. Officer: Questo ponte? Perché di lì che facevate?

      This bridge? Why what did you do from there?

12. Child: Lì (.) ammazzavano dei bambini e o ballavamo facevamo

      tutti quei gesti brutti con i

      vestiti (..) poi

      There (.) they killed children and or we danced we did all those bad gestures with the clothes (..) then

      (3.0)

13. Officer: Ammazzavano i bambini come eh?

      They killed the children how eh?

14. Child: Con quella (1) saetta (.) non so (.) quella per tagliare le teste.

       With that (1) arrow (.) i don’t know (.) the one to cut heads

15. Officer: aha. Questo succedeva di giorno o di notte?

      Did this happen during the day or at night?

16. Child: Quando c’era (..) buio.

       When it was (..) dark

17. Officer: mhm

18. Child:  Poi (1) quel pratino (.) Lì, se mi ricordo bene, hanno

       scavato dei bimbi e hanno

       messo dei bambini lì.

       Then (1) that small meadow (.) there, if I remember well, they dug

       some children and they put some children there.

19. Officer: Ti ricordi chi ci stava quando succedevano queste cose?

      Do you remember who was there when these things happened?

      (1.0)

20. Child: mhm mio padre (1) tutti i bambini

      My father (1) all the children

21. Officer: mhm

22. Child: qualche volta mia madre (2) e

      sometimes my mother (2) and

In this interview, the production of the narrative is clearly guided by the suggestions of the psychologist and the officer, acting as co-tellers and elicitors. They never question the uncertain turns of the child: this shows the background metanarrative of the child as a reliable witness.

The next interview (episode 6) takes place between a psychologist, acting as a court counsellor, and a child. The psychologist suggests a discrepancy between the child’s scary story and her serenity in telling it (turn 1). The child provides a minimal response, and the psychologist insists with a question, which openly suggests that the child is pretending to be calm. The child provides an uncertain answer, which is partially in line with the psychologist’s suggestion. Finally, the psychologist provides an assessment of the child as pretending to be calm (turn 5):

1. Psychologist: Stai raccontando delle cose molto spaventose e fanno

        molta paura e tu sei tutta

        tranquilla

        You are telling very frightening things and they are very scary and you are very calm

2. Child:    Eh::

3. Psychologist: Sei tranquilla davvero, o fai finta di essere tranquilla?

        Are you really calm or are you pretending to be calm?

4. Child: beh non tanto tranquilla

        Well not so calm

5. Psychologist: Eh fai finta di essere tranquilla

        You are pretending to be calm

The psychologist’s final assessment is an ‘upshot formulation’, that is, a turn that adds ‘significance’ to the content of a previous turn (Heritage and Watson, 1980: 249) – in this case, the child’s turn 4. An upshot formulation proffers an articulation of what was not said by the interlocutor (Bolden, 2010). It shows the presupposition that the previous turn must be interpreted through the elicitation of a new narrative (see Antaki et al., 2005). The psychologist articulates what the child did not say, eliciting the narrative of ‘real’ fear against the child’s pretence of serenity.

The third extract (episode 6) confirms the psychologists’ use of suggestive questions and upshot formulations regarding children’s emotions. The child says that she went back to her town. The psychologist first repeats her statement and then enquires about the child’s feelings (turn 2). The child discloses an emotional reaction and the psychologist enquires about the type of emotion (turn 4). The child mentions happiness: the psychologist repeats the term twice, with an interrogative intonation, and asks for confirmation, showing some surprise (turn 6). Then, she insists on questioning what seems self-evident (Jol and Stommel, 2016), which is confirmed by the child with a minimal response. In turn 10, the psychologist suggests an alternative emotion, without specifying it. However, the child denies any other emotion, confirming happiness. The psychologist repeats the child’s confirmation and, after a short pause, proffers an explicit suggestion: the child’s emotion might include suffering (turn 12). After the child’s new minimal response, the psychologist provides an upshot formulation: it is difficult to admit suffering because the memory of what happened might determine suffering (turn 14). The child simply nods:

1. Child:    Siamo passati anche per la piazza

        We also went through the square

2. Psychologist: siete passati per la piazza (.) e che effetto ti fa rivederla.

        You went through the square (.) and what effect does it have on you

        to see it again

3. Child:    Un po’ di emozione.

        A bit of emotion

4. Psychologist: Un po’ di emozione (.) Sapresti dirmi dare un nome a

        questa emozione?

        A bit of emotion (.) Could you tell me give a name to this emotion?

5. Child:    Di gioia!

        Of happiness!

6. Psychologist: Di gioia? Di gioia (.) Quindi ti ha fatto piacere?

        Of joy? Of joy (.) so were you pleased?

7. Child:    Sì

        yes

8. Psychologist: Vedere di nuovo questa piazza?

        To see this square again?

9. Child:    Hm hm

10. Psychologist: Forse c’è anche un’altra emozione insieme alla gioia? (.)

         C’è un’altra emozione

         oppure no?

         Maybe there is another emotion together with joy? (.) Is there another emotion or not?

11. Child:    No, solo un po’ di gioia.

        No, only a bit of joy

12. Psychologist: Un po’ di gioia. (1.0) Forse ci può essere anche un

         briciolo di sofferenza a tornare qui

         (2.0) può essere?

         A bit of joy (1.0) Maybe coming back here may cause also a bit of

         suffering (2.0) could it be?

13. Child:     mhm

14. Psychologist:  Solo che per te è difficile dirlo. Forse sono anche

         accadute delle cose (.) che ti fa

         soffrire ricordare

         But you find it difficult to tell. Maybe some things happened

         (.)whose memory makes you suffer

15. Child:     nods

The psychologists’ elicitations of narratives also concern ‘facts’. In the following short conversation (episode 6), after a formulation of a turn in which the child suggested taking his younger brother away from her family, the psychologist suggests several alternative interpretations of ‘bad things’, the last of which is ‘bringing the brother to the cemetery’. The child immediately confirms this alternative: the final position of the alternative in the psychologist’s utterance elicits the child’s narrative:

1. Psychologist: Sì ma mi hai detto gli fanno delle cose brutte ma le cose

        brutte possono essere non

        dargli da mangiare dargli due schiaffi e che so non cambiargli il

        pannolino oppure portarlo al cimitero ci sono

        tante cose [brutte

        yes, but you told me they do bad things to him but bad things can be not feeding him, slapping him and what can I say not changing the diaper or taking him to the cemetery there are many bad

        [things

2. Child:       [portarlo portarlo al cimitero

           [taking taking him to the cemetery

These fragments show that while children can find it difficult to make sense of some questions, related to an institutionalised inference system, they provide an answer anyway, thus showing their orientation to the obligation of the institutional context and their recognition of asymmetrical epistemic authority (Jol and Stommel, 2016). In these fragments, the institutionalised inference system does not produce accurate checking of children’s knowledge. Interviewers’ checking is based not only on suggestive questions but also on the elicitation of narratives through upshot formulations. The children adapt to this interview process and accept the proposed subject matters. While children’s identity as witnesses should determine their identity as victims (Deckert, 2010), in this case, it is their identity as victims that determines their identity as witnesses.

The institutional power of eliciting ontological narratives in the interaction is amplified by the psychologist’s statements of certainty, as in the following video-recorded utterance:

Di sicuro qualcuno ti ha fatto male al sederino e alla patatina, e questo è proprio sicuro perché l’ha detto la dottoressa. (Episode 6)

Surely someone hurt your little bottom and your little pussy and this is for sure because the doctor said that

The children’s acceptance of the psychologists’ certainties is shown by a video recording in which the child asks for confirmation of the correctness of his ontological narrative:

Quando cadeva del sangue, mia madre subito a pulire perché non voleva che rimanessero delle tracce (.) Va bene quello che ho detto? (Episode 6)

When blood fell, my mother immediately cleaned because she did not want that to leave any trace (.) Is what I said alright?

The first legal interview with Dario, with the assistance of the psychologist working for the social services (episode 7), shows a different type of pressure. This interview is reported in a fragmented way as the video recording is damaged. However, the available fragments make the organisation of this interview rather clear. The first fragment shows that Dario is more interested in his future trip to the seaside than in the interview. Here, the judge pays attention to Dario’s interest (turn 2), building preliminary rapport, as suggested by protocols of interview (Goodman and Melinder, 2007; Hershkowitz et al., 2006):

1. Dario: Domani vado al mare

     Tomorrow I’m going to the seaside

2. Judge: Dove vai al mare?

      Where are you going to the seaside?

3. Dario: Sul Gargano

      To Gargano ((seaside resort))

Later, however, it becomes clear that Dario is not focused on the interview. For instance, he introduces the topic of his cat. The judge provides some feedback, but Dario’s lack of motivation is becoming a problem in the interaction:

1. Dario: Sai che la mia gatta che si chiama Luna, ha fatto dei cuccioli

     Do you know that my cat, called Luna, had kittens

2. Judge: Ma va?

     Really?

3. Dario: Che che ha portato via

     Which which she took away

This problem is stressed by the judge, who resumes the topic of the trip to the seaside to motivate Dario to tune in the interview. Doing this, the judge provides unsupportive comments, suggesting that the trip to the seaside depends on Dario’s participation in the interview. These unsupportive comments seem to be typical when children do not disclose abuse (Hershkowitz et al., 2006):

1. Judge:   guarda che il mare si allontana eh?

       Look the seaside is going away eh?

2. Dario:   Quando ho finito tutto, vado subito via?

       When I finish, will I go away immediately?

3. Judge:   Subito via, vai al mare

       Immediately away, you can go to the seaside

 Judge:   Più sei veloce a raccontare, più il mare si avvicina sai?

       The faster you tell, the nearer is the seaside, you see?

Later, the judge suggests the psychologist’s support ( ‘vuoi stare in braccio a: alla V.?’ [do you want to be n V’.s arms?]); this support is also acted by caressing Dario’s back. After that, the instrumental reference to the trip is pursued with the psychologist’s help; thus, building rapport becomes instrumental to the disclosure of abuse:

1. Psychologist: C’è un mare azzurro lì

        There is a blue sea there

2. Judge:    Però prima devi darmi una mano, eh, se no

        But first you have to help me eh, otherwise

3. Dario:    Eh.

The psychologist’s support becomes pressure when she links the taken-for-granted truth of Dario’s narrative ( ‘what really happened’) with his satisfaction for his trip and, metaphorically, with his strength.

se riesci a dire tutto quello che è capitato davvero quando vai fuori da quella porta ti godi di più la pallavolo il mare e tutto quello che fai e guardando i tuoi muscoli dirai però sono stato proprio forte.

If you are able to tell me what really happened to you when you go out of that door you will enjoy more volleyball the seaside and everything you do and looking at your muscles you will say wow I have been really strong.

Nevertheless, Dario does not show interest in the interview and is therefore told off by the psychologist, once again using the future trip for the purpose of the interview, thus repeating the unsupportive comment previously delivered by the judge:

Non stai ascoltando, non stai ascoltando! E questo non ti è utile, non è utile, altrimenti stiamo qui tanto tanto e vai dopo

You are not listening, you are not listening! And this is not useful to you, it is not useful, otherwise we’ll stay here for a long long time and you’ll leave later

The problem of children’s reluctance in disclosing abuse is certainly important in this type of interview (Hershkowitz et al., 2006). However, here the problem is rather the interviewers’ influence on the children, based on ‘a preconceived notion that children rarely if ever lie about child sexual abuse’, leading to ‘ask suggestive questions or take ambiguous statements or debatable behavioural indicators by the child’ and thus determining ‘that the child is indeed an abuse victim’ (Goodman and Melinder, 2007: 2).

In the analysed fragments of interviews, the problem of influence is very evident, although it can take ambiguously supportive forms. The interviewers systematically select the children’s turns, deserving attention and feedback. They do so through leading questions and upshot formulations, explicit or implicit suggestions of favourite answers, and normative interpretations of emotions and behaviours. Thus, the interactional construction of narratives is based on the interviewers’ co-telling and often on their explicit elicitations (upshot formulations). The interviewers deal with inaccuracy of autobiographic memory and difficulties in retrieving memories (Goodman and Melinder, 2007) through suggestions and elicitations of narratives. It is possible that, in other circumstances, the interviews took a different form. However, it is clear that the children were exposed to strong social pressures.

Frequently, the children aligned with the interviewers’ epistemic authority, thus displaying their suggestibility, through both affirmative responses to suggestive questions and change of responses to please the interviewers (Goodman and Melinder, 2007). Sometimes, however, they rejected this epistemic authority, changed topics (e.g. Dario’s mentions of the trip and the cat) and displayed minimal responses which do not confirm the interviewers’ suggestive questions or upshot formulations explicitly. These turns, as well as the children’s uncertainties, were either ignored or denied by the interviewers.

This analysis shows heavy influence on the production of children’s ontological narratives. Children’s ontological narratives become vicarious narratives, first in the interviews and then in the public context where the experts speak on behalf of the children. The structure of this interview is a specific version of the hierarchical structure of institutional interactions between professionals and children, according to which children are subordinate to adults with higher epistemic authority (Jol and Stommel, 2016). This structure of interview fulfils the function to explain events and facts according to the interviewers’ expertise. This function presupposes the primary relevance of the principles of causality and rationality, as typical of adult experts. Interviews are based on expectations of a rational construction of truth, breaking, if necessary, the metanarrative of children’s truthfulness.

This function determines the failure of supportive actions. According to Hershkowitz et al. (2006), ‘interviewer supportiveness has a positive effect on the amount of information provided by children in the interviews’ (p. 765); therefore, it is important to establish rapport before talking of abuse, to ‘prepare’ children ‘so as to maximize their willingness and capacity to be informative’ (Hershkowitz et al., 2006: 755). Rapport reduces children’s anxiety, which is ‘a barrier to effective evidence gathering’, and promotes children’s comfort in disclosing themselves (Childs and Walsh, 2017: 190). According to a child-centred metanarrative, the production of children’s narratives must be based on empathy, meant as action showing concern and sensitivity for the interlocutor’s words, and active listening (Mearns and Thorne, 1999). Empathy and active listening aim to construct children’s trust in professionals as well as children’s self-esteem. In the analysed fragments of interview, the interviewers provide some minimal responses, repetitions and gestures that indicate their active listening. However, active listening is overshadowed by leading questions, upshot formulations and explicit suggestions, and therefore loses the meaning of empathic support. The interviews show that the explanation of facts and events is much more relevant than listening and enhancing the children’s ontological narratives. The interviewers use rapport as a tool to achieve logical explanations of abuse at any cost and thus fight against parents’ inadequacy and abuse. For this purpose, the children’s ontological narratives are transformed into a first order observation of facts and the children’s epistemic authority is used to reproduce hierarchical structures. Goodman and Melinder (2007) warn against ‘simplistic views of children’s capabilities’ (p. 14): in the case shown here, the view of children’s epistemic authority was used to support external explanations of facts.

This structure of interview was coupled with the structure of institutionalisation. Institutionalisation takes the form of ‘panopticon’ (Foucault, 1977), separating children from their everyday life arenas and practices and monitoring them. This structure nullifies the empathetic child-centred approach: in total institutions, children are treated as non-persons (Goffman, 1981). The child-centred approach is replaced by rational explanation (Goffman, 1961): children’s institutionalisation is explained by establishing the rational knowledge of children’s bad conditions. This rational knowledge is acquired through the children’s ontological narratives regarding their experiences and emotions, reshaped as conceptual narratives.

Through the coupling between structures of interview and institutionalisation, children’s acceptance of professionals’ transformation of their ontological narratives can be a way of finding a recognition as persons-with-an-identity, initially the double identity of witness and victim (Deckert, 2010) and then new personal identities. This coupling of structures can reshape children’s everyday life arenas and practices and personal life, in new families and places. Children’s new identities are the outcome of a generational order based on the paradoxical enhancement of children’s agency as subordinate to the function of providing experts’ rational explanations. Children are empowered as having access to the ‘right’ knowledge, but this empowerment is denied in the interplay between children’s ontological narratives and institutional systems of inferences. Empathy is thus used to produce a paradoxical mélange of empowerment and denial, defined as rational in conceptual and public narratives.

Conclusion

This article has provided a sociological analysis of the ways in which some institutions (social services and courts) dealt with stories of children’s abuse in the late 1990s. The following conclusions, however, are not limited to this case and that time, since they are re-considered through a sociological theory that highlights three general results of the analysis, regarding children’s narratives, interaction between experts and children, and the organisation of institutional generational order. First, children’s allegations of abuse can be produced through different types of narratives (ontological narratives, conceptual narratives, public narratives, metanarratives). The connections between conceptual narratives, metanarratives and public narratives can transform children’s ontological narratives.

Second, interviews, as interactions involving experts as well as children and following an institutionalised inference system (Jol and Stommel, 2016), can affect children’s agency and ontological narratives through a variety of interviewers’ actions, including leading questions, upshot formulations, explicit suggestions, lack of consideration for rejection, uncertainty and ambiguity of minimal responses, and instrumental comments on children’s interests.

Third, the hierarchical generational order can explain children’s alignment and acceptance of the institutional inference system. In particular, it is interesting to observe that this hierarchical order can be displayed as a paradoxical construction of children’s agency, through the subordination of rapport (empathy) and empowerment to the function of explaining facts. The coupling of structures of interview and institutionalisation leads to a new construction of persons-with-an-identity. The children’s ontological narratives of abuses, produced through this hierarchical institutional order, become public through conceptual and public narratives in which the children have no voice.

These results may unveil the need for a different, dialogic form of communication, which can align the institutional system with children’s everyday life arenas and practices (Percy-Smith, 2010) as well as personal life (James, 2013). Children’s agency is enhanced if ‘both children and adults are co-constructors of knowledge and expertise’ (Hill et al., 2004: 84). Therefore, its enhancement requires specific relations with adults (Wyness, 2013), in which adults’ actions enhance children’s autonomous choice of action (Baraldi, 2014; Shier, 2001). Dialogic forms of communication are based on equal opportunities of participation and require adults’ sensitivity for and empowerment of children’s autonomous actions as enrichment (Baraldi, 2012, 2014). Dialogic forms of communication might be useful to challenge hierarchical structures of institutional intervention on children’s abuse, use of empathy to improve explanations, metanarratives of the innocent and trustful child, and institutional constructions of the child’s need to be a person-with-an-identity.

The need of dialogic forms of communication resonates with some protocols of interview regarding sexual abuse (Childs and Walsh, 2017; Goodman and Melinder, 2007; Hershkowitz et al., 2006). These protocols suggest rapport building and promotion of free-recall before possible specific and focused questions that clarify and check children’s accounts and allegations (Goodman and Melinder, 2007). In a dialogic form of interview, ‘before’ is not only understood in the temporal dimension, as rapport and free-recall are never abandoned during the interview. In this view, only a dialogic form of interview can give meaning to both disclosure and non-disclosure of sexual abuse, ensuring that they are based on children’s effective exercise of agency.

In conclusion, this article has shown that ‘poisoning children’ means either that psychologically fragile children may poison their social context or that children’s ontological narratives may be poisoned through institutional systems of communication. In the latter view, it is also possible to observe that the institutional systems can be ‘poisoned’ through dialogic forms of communication.