ESRC RESEARCH GROUP ON CARE, VALUES AND THE FUTURE OF WELFARE

University of Leeds

Workshop Paper No 15
Prepared for Workshop Four
Methodology for researching moral agencies
Friday 17 March 2000

Wendy Hollway

MORAL INTERSUBJECTIVITY, METHODOLOGY AND EMPIRICAL RESEARCH

Introduction.

In this paper I aim to sketch out briefly the theoretical characteristics of my understanding of moral subjectivity and then discuss their methodological implications. I discuss these in two parts. The first is properly methodological, in the sense that it focuses on how the principles used to theorise moral subjectivity can be translated into a conceptualisation of the research subject and a set of epistemological principles for knowing about such subjects. The second part is about translating that into research practice: how, given these principles, might one design an empirical study to look at the practices of moral intersubjectivity in partnering and parenting?

Moral (inter)subjectivity

I have set out my approach to theorising moral subjectivity in relation to understanding the capacity to care in a seminar paper for the Centre for Research on Family Kinship and Childhood (working paper 15). Here, I extract some principles which provide an alternative to accounts of unitary rational subjects who are socialised into moral conduct through training and example. I draw on three, interdependent, theoretical perspectives.

The first is psychoanalytic. Put at its simplest, this involves assuming the influence of dynamic unconscious processes in the construction of meaning and a person's relation to reality and the effects of past experience (duly transformed by internal mental processes such as unconscious fantasy) on present and future actions and relationships. The second is biographical, drawing on the above psychoanalytic premises. It means that I work with the principle that a person's (constantly varying) capacities to act morally can best be explained by the way that their past life experiences (and the way that they have come to terms with them) mediate the meanings through which current events and relationships are experienced. The third is 'psycho-social' (the term indicating an attempt to transcend the commonplace individual-social dualisms in most social theory). The implication here is that a person's moral capacities can be understood as social, as long as social is understood as biographical and relational, not just situational and structural. It also depends on the 'psycho' part of that pair being understood through unconscious intersubjectivity (whereby the social is mediated and transformed and rescued from being determining), otherwise we are left with the inadequacies of socialisation theory (1).

For Klein (1988a and b), humans are moral (as opposed to amoral, not immoral) in their fundamental nature. Moral feelings are first precipitated by an infant's experience of another in the process of sculpting a sense of self from its early merging with the mother (or other primary carer):

This capacity is neither guaranteed nor, if it is achieved, is it stable across different relationships and events. If someone continues to operate from a paranoid-schizoid position (2), there can be no concern for others. Experience of self and others as containing both good and bad is overwhelmed by unconscious defensive splitting. More often than not the bad is projected out so that the self can be experienced as good. However, in an attempt to preserve belief in a good parent on whom it depends, a child may locate the bad in itself. Hate is split off from love and while hate predominates, the capacity for concern is unattainable. With access to the depressive position comes a different moral sensibility. Reparation is understood as the urge to repair the damage which one has done (in fantasy or reality) to the loved object. The concepts that Klein is developing as part of her account of necessary human developmental processes involve not only moral sentiments, but moral actions (3).

In brief, Kleinian theory has the characteristics which are needed to theorise moral development:

• It can explain the human capacity for feeling and its roots in, and effects on, relationships.
• Moral sensibilities (and their lack) are explained in the context of an account of the way the self achieves differentiation from primary others.
• Morality in this light is inseparable from the development of subjectivity. The central concepts in Kleinian theory are moral: envy, guilt, love, hate and reparation.
• Development is seen as potentially progressive, but the idea of 'position' replaces that of stage and there is no liberal humanist assumption that there will automatically be progress. There is an account of the conditions under which progress (towards the capacity to feel concern and act with care) can be accomplished.
• The account is psycho-social in the sense that it constantly works with dynamics between inner and outer worlds, reducing to neither one nor the other.
• The capacity for thinking and reasoning is not sidelined but is understood as dependent on emotional processes which affect how one can face reality (subjectivity is not understood as rational, nor exclusively irrational).
• It begins in the infant-mother relationship of dependency, but does not idealise or naturalise this.
• It holds a two-fold view of emotional and relational life, where love and hate are constantly both in the frame.
• The capacity for concern and what constitutes good enough care can be theorised, rather than just described, within this theory and developments from it.

To this Kleinian core, I want to add two further concepts which further express relational aspects of morality: containment and recognition. Containment derives from Bion's developments of Kleinian theory and depends on the principle of unconscious intersubjectivity to refer to how a person can contain parts of another. This capacity begins in the early mother-infant relationship, prior to language, if the mother is capable of tolerating the baby's feelings, neither 'denying them ingress' nor 'becoming a prey to the anxiety which resulted from the introjection of the baby's bad feelings' (Bion 1959:104, cited in Hinshelwood 1991:247). An incapacity to contain painful or bad feelings can lead to projecting them on to another and attacking them there; that is, it can lead to violence and harm to others. Recognition derives from Benjamin's (1984, 1995) account of the early dynamics of differentiation. True as opposed to false recognition depends on having an other who is capable of responding to one from a position which is independent of ones omnipotent desires to control the other's responses. Without anyone who can be trusted to do this, it is hard to achieve a secure sense of differentiated self.

Methodology

There are three primary ways in which this account of moral subjectivity departs from dominant social science assumptions about the research subject which are based on the notion of a rational unitary subject. First, the paradigm for understanding subjectivity is based on the principle of a dynamic unconscious which has determining influences on people's relations and actions (and thus on their moral capacities). Second, these unconscious dynamics are not primarily intra-subjective, but inter-subjective. Defensive and desirous ego-protective dynamics work across porous psychological boundaries between people to affect people's capacities to perceive reality and to identify accurately the situations and meanings of others. Third, moral capacities are biographically inscribed through a history of relationships and continue to be reproduced and changed in relationships.

Dominant social science methodology often ends up assuming that the research subject shares meanings with the researcher, is knowledgeable about themselves (their actions, feelings and relations), can access the relevant knowledge accurately and comprehensively (that is, has accurate memory), can convey that knowledge to a stranger listener and is motivated to tell the truth. Hollway and Jefferson (2000) use the phrase 'the defended subject' to refer to an alternative set of assumptions. These can be summarised as research subjects who may not hear the question through the same meaning frame as that of the interviewer or other interviewees; are invested in particular positions in discourses to protect vulnerable aspects of self, positions which will vary in the accuracy with which they represent a person's actual conduct; may not know why they experience or feel things the way they do and are motivated, largely unconsciously, to disguise the meaning of at least some of their feelings and actions.

There are three problems with translating this alternative ontology into an epistemology. The first is that research subjects cannot, in this view, be relied upon to know, let alone tell, certain things (4) about themselves, especially if these are anxiety-provoking. The second problem is that a theory of the defended subject suggests that any language-based method of research will be limited in what of that person can be communicated to the interviewer. In this case, either we need to develop a method which does not rely on research participants' accounts (for example, observation) or we need to develop a language-based method which is capable of going beyond the conscious intentional information provided by interviewees. So far,Tony Jefferson and I have developed the latter, addressing the premise of defended subjects by using the psychoanalytic principle of free association which enables the expression of unconscious meanings through language. The third problem is that there is no guarantee that the interviewer will make the right or best sense out of the data. To address this, we treat the accounts given in the interview as a co-production (albeit produced from asymmetrical roles) and look in detail at the intersubjective dynamics that will influence this and the subsequent analysis (5).

Method

We ended up basing our interview method on the the biographical-interpretative method, first developed by German sociologists producing accounts of the lives of holocaust survivors and Nazi soldiers (Rosenthal 1993, Rosenthal and Bar-On 1992, Schutze 1992). However, given our understanding of the way that unconscious defences affect the information that is produced in the research relationship and the way that it is interpreted, we wanted to incorporate this idea of the defended subject in our use of a narrative method. Schutze saw 'some intersections between Freud's impressive theory on repression' (1992: 359n1) and his own method, but this insight is not developed. The main principle of the biographical interpretative method is not the defended subject, but the idea that there is a 'gestalt' (a whole which is more than the sum of its parts; an order or hidden agenda) informing each person's life which it is the job of biographers to elicit intact, and not destroy through following their own concerns (Rosenthal 1990).

Data production

The German biographers' strategy for eliciting narratives - which we adopted and adapted - can be summarised in terms of four principles, each designed to facilitate the production of the interviewee's meaning frame, or gestalt: Use open-ended not closed questions; elicit stories of specific events, avoid why questions and follow up using respondents' ordering and phrasing. The first three principles are all about trying to ensure that the interviewee does not answer at the level of explanation, generalisation or opinion, but in terms of some event which s/he has actually experienced. The closer the account stays to events, the more it can elicit emotional meanings rather than an exclusively cognitive account. If interviewees stay faithful to what they actually did, as part of what happened, then their actions are available as data for interpretation, albeit through linguistic accounts. The fourth principle is about remaining within the interviewee's meaning frame (which has not been intruded upon by a structured question and answer format). The gestalt principle suggests that meanings are contained not simply in the units of meaning themselves, but in the links between these: in the structure of the whole. This style of interviewing is designed to frame, elicit and then preserve and analyse a structure of meaning which is an unintentional product of the interviewee's whole account.

Where narrative interviewing and analysis emphasise the coherence of a person's story, we were basing our use of this method on the principle of a non-unitary and non-rational subject. There are similarities between the principle of respecting the narrator's gestalt and the psychoanalytic method of free association. By asking the patient to say whatever comes to mind, the psychoanalyst is eliciting the kind of narrative that is not structured according to conscious logic, but according to unconscious logic; that is the associations follow pathways defined by emotional motivations, rather than rational intentions. According to psychoanalysis, unconscious dynamics are a product of attempts to avoid or master anxiety. This suggests that anxieties and attempts to defend against them, including the identity investments these give rise to, provide the key to a person's gestalt. By eliciting a narrative structured according to the principles of free association, therefore, it is possible to secure access to a person's concerns which would probably not be visible using a more traditional method. While both approaches aim to elicit detail, narrative analysis has a preoccupation with coherence which we do not share. Free associations defy narrative conventions and enable the analyst to pick up on incoherences (for example, contradictions, elisions, avoidances) and accord them due signficance.

Data analysis

These are the principles that we were developing when Tony interviewed Ron, whose case study exemplifies them in use (see workshop 3 paper). Here I will summarise two central principles: preservation of the whole and interpretation. With regard to the first, we avoided the use of any method based on a 'code and retrieve' principle which involves the fragmentation of data according to themes (for example all computer-assisted qualitative data analytic packages). The principles of gestalt and free-associative linking required us to work with the whole data set which we had produced for one person (or for one family). With regard to the second principle, data analysis is necessarily based on interpretation. This is more than a hermeneutic principle but stems from the premise that interviewees (like everyone else) are not necessarily going to be aware of the motivations for their own behaviour. Psychoanalysis explicitly posits that there are areas of blindness in everyone, prompted by defences against anxiety, which other people will not share, unless they share the same unresolved conflicts. We rely on empirical evidence from the interview transcript, but it must necessarily be interpreted and this interpretation will rely on ourselves as instruments of understanding and the theoretical resources we have available to us. We use various safeguards because of the danger of the researcher's subjectivity over-riding better interpretations. In the end, in all hermeneutic research, we are not claiming 'truth' in 'findings', only an analysis which hopefully is more faithful, more complex and produces more relevant insights than another analysis might have done.

Beyond the current FANI method.

While I believe that the method as we have practised it so far takes us a considerable way into an empirical analysis which can address moral intersubjectivities, I would like to develop it in two ways.

First, our work so far has mainly been focussed on individual case studies (with the notable exception of one family case study based on double interviews with three members of the same family). In future work I would like to focus more systematically on families (defined as nexuses of partnering and parenting relationships)(6), negotiating access with several, even all, members. In particular, it would be policy-relevant to select non-standard families (and as many versions of these as possible). Since the design is case-study based, numbers within each type would not be an issue, although of course the limits to generalisability always are. This broad definition of research focus can be narrowed down by identifying key times in the life of family members which illuminate relational transitions in the four ways in which I have 'operationalised' moral intersubjectivity: concern, care, containment and recognition. Two likely transitions are the times when children from a household form their own partnerships and when they have their own children. Both these moments represent, in different ways, relatively sudden transitions from positions of being in receipt of others' concern, care etc to providing this for dependent others.

Second, despite the workings of free association (and therefore unconscious dynamics) in the interview accounts, they are still language-based. Increasingly I am convinced that much of what matters in the arena of moral intersubjectivity does not pass through discourse, but is expressed directly in relational practices. While it is possible to get (fairly faithful?) accounts of such practices through stories of past events, a different order of information can be achieved through observation. There is a well-established method of family observation first developed for observing infants in their family setting by the Tavistock Clinic, which has shown itself, through case studies (7), to be a powerful method for producing insights into non-discursive aspects of family relations. The combination of free association narrative interviews and observation would produce a few depth case studies which could illustrate the detailed and complex workings of moral intersubjectivity in diverse family forms.

FOOTNOTES

1. While this is primarily a psychological level account, it complements more sociological analyses (for example, Sevenhuisen, Honneth, Fraser, Elias). The external realities on which they largely focus are influential, but not determining, in the making of meaning. Back

2. The paranoid-schizoid position is the position of early infancy which continues to be accessible for defending against unbearable aspects of reality. Back

3. The reason for the elision of the terms moral development and moral subjectivity resides in the importance of biography in this perspective and because development is seen not as progressive in any simple sense, but residing in the ongoing challenges of acting morally in relation to others at whatever age. Back

4.Research questions which are close to issues of subjectivity, intimate (parenting and partnering) relationships and morality, fall within this domain. Back

5. See discussion in the Ron case study from workshop 3. Also Hollway (unpub) 'Reflexivity, unconscious intersubjectivity and the co-production of data'. Paper presented at the 18th International Human Science Research Conference, 'Unity and Diversity', in Sheffield, England 26-29th July 1999. Back

6. It has recently occurred to me (thanks to a conversation with Juliet Mitchell) that one disadvantage of the phrase 'partnering and parenting' which 'family' does not share is that it exclude sibling relationships, which may be very important in the development of moral subjectivity. Research designs should bear this in mind. Back

7. See, for example, the 'International Journal of Infant Observation'. Back

BACK