Regimes of motherhood: Social class,the word gap and the optimisation of mothers’ talk

Abstract

The role of working class mothers’ talk in explaining their child’s ‘impoverished’ language development and the resulting ‘word gap’ between social classes is hotly debated. Academic research in this area spans decades, crosses continents, and gathers up a wide range of disciplines and positions, ranging from research seeking to intervene in and optimise mothers’ talk, to research that vigorously criticises any attempt to do so. Through an extended analysis of Jacques Donzelot’s seminal study The Policing of Families, and Michel Foucault’s concept of a ‘regime of truth’, we explore how motherhood is constructed by academic debate as something to be endlessly optimised. Academic debate functions by (1) reducing expectations concerning the role and remit of experts so as to place the onus on mothers to implement findings which the former will facilitate, (2) complicating the contributing factors to language delay, thus avoiding apportioning blame too directly whilst giving endless cause to do further research, and (3) committing mothers to a permanent labour in which they are expected to better themselves as measured by the manifest language development of their children. Its strongest critics remain within the constraints of this regime of truth to the extent that they argue for humility of expertise, for recognition of broader sociocultural factors, and for the importance of privileging the expertise and agency of mothers. This article considers how all parties are, in effect, obliged to declare the truth of motherhood and will find themselves implicated in its governance.

Introduction

In this article we consider the ongoing academic debate concerning the relationship between social class, mothers’ talk and children’s language development. This debate shows no sign of approaching resolution. It spans decades, crosses continents, and extends across disciplines and a host of professional practices. Our argument is this: all sides are invested in a debate they will not end by demolishing, out-thinking, or out-evidencing others. We claim that this is not simply because researchers speak from disciplinary points of view that are fundamentally irreconcilable, although that is undoubtedly part of the problem. Rather, we suggest that the debate concerning the improvement of children’s language and the role of mother’s talk (and motherhood more generally) will not resolve because all sides are invested in a ‘regime of truth’ that constitutes the ground of the debate itself. Our analysis builds here on the work of Michel Foucault (1980/2014) as well as the work of Foucault’s colleague and co-researcher, Jacques Donzelot (1977/1979), whose study of motherhood and parenting helps unpack some of the prescriptive effects and governmental logics of that debate.

In this article, our overall objective is not to take a position within this debate, but to draw attention to the investment of all concerned in its continuation. We argue that the debate ties all parties into a set of commitments to refine motherhood, either by intervening further, or by reducing the demands of intervention. It produces a basic conception of motherhood and mother–child relations as something that can be endlessly optimised, where the term ‘optimisation’ ranges in scope from direct professional assistance designed to augment or even correct mothering, to the appeal for mothers to be allowed to resource themselves without professional intervention. That is to say, it produces an obligation to act on the so-called ‘truth’ of motherhood that is suffered by all parties.

The debate: An initial overview

The debate about mother’s talk and children’s language development is not simply unresolved but has, over time, served to proliferate dispute and uncertainty. It is notable, moreover, that while there have been some attempts to investigate fathers’ language input (Johnson et al., 2014; Pancsofar & Vernon-Feagans, 2006) and the influence of other family members (De León, 2011), the overwhelming drive of research endeavour is directed at mothers. Most often these mothers are thought to be ‘at risk’ based on demographics alone (with the effect that social class is only very narrowly configured ¹ ), and are prejudged on that basis as having insufficient or deficient resources to invest in their children (Gillies, 2007).

The debate has roots in the 1970s and 1980s (Bernstein, 1977; Elardo et al., 1977; Heath, 1982; Snow, 1977; Tizard et al., 1983; Wells, 1986). However, the common reference point for the current obsession with maternal language and a ‘word gap’ in both research and policy, is Hart and Risley’s seminal study Meaningful Differences (1995). This study famously claimed that by the age of three, poor children hear 30 million fewer words than advantaged peers. The enduring influence of this study is remarkable: it remains widely cited (with over 11,000 citations on Google Scholar at the time of writing) and debated in academic work, the media and policy documents, as well as books and training for teachers and public education and health initiatives. ² The existence of a so-called word gap remains a matter of controversy (also referred to as a ‘language gap’, Johnson & Johnson, 2021), and the role of working class mothers’ talk in explaining their child’s ‘impoverished’ language development remains hotly debated (e.g. Avineri et al., 2015). However, the persistence of the debate itself suggests that motherhood is something to be endlessly worked on by academic research.

The internal dynamic of that debate is complex but has, we claim, the following broad features: It functions by (1) reducing expectations concerning the role and remit of experts so as to place the onus on mothers to implement findings which the former will facilitate, (2) complicating the contributing factors to language delay, thus avoiding apportioning blame too directly whilst giving endless cause to do further research, and (3) committing mothers to a permanent labour in which they are expected to better themselves as measured by the manifest language development of their children. Its strongest critics remain within the constraints of this regime of truth to the extent that they argue for humility of expertise, for recognition of broader sociocultural factors, and for the importance of privileging the expertise and agency of mothers. This article considers how all parties are, in effect, obliged to declare the truth of motherhood and will find themselves implicated in its governance.

Regimes of truth

Foucault’s concept of a ‘regime of truth’ offers a framework for understanding the dynamics of the debate about motherhood and language development. ³ The concept is outlined in Foucault’s 1980 lecture series, On the Government of the Living (Foucault, 1980/2014; for an earlier use see Foucault, 1977, p. 13). For Foucault, the connection between any given ‘truth’ and its effects is not automatic. Human beings are not innately programmed to obey or act on a truth once it has been demonstrated. There must be an accompanying set of practices which must be produced (or fabricated), like the truth itself. Foucault claims that the most rigorously constructed argument requires this last step, which is not a logical step, but involves learned submission; it ‘consists in saying: if it is true, then I will submit’ (Foucault, 1980/2014, p. 96). Or, more gently, it involves an obligation which consists in saying: if this is true, then I will oblige.

So the question becomes: ‘How [in this debate about motherhood] does the truth oblige, in addition to the fact that it is manifested?’ (Foucault, 1980/2014, p. 94). We suggest that there are two basic levels of obligation: firstly, the obligation of researchers to continue researching as they repeatedly return to similar problems and questions further complicating their understanding of motherhood and language development; and secondly, the consequences of continued academic debate for those mothers who are picked out by research and obliged to act via the various institutions that intervene at the level of everyday parenting.

In this article we are primarily concerned with the debate itself, and so with the first level of obligation, that is, how researchers are obliged and oblige others to persistent debate, although we recognise that this first level of obligation cannot be discussed without exploring its relation to the consequences for mothers.

The subjective attachments of those involved are crucial. This is because, for Foucault, there are no truth-effects without a corresponding subject that recognises the truth in question and fulfils its obligations. Hence, the researcher must recognise their activity as the work of someone who treats motherhood as a serious concern in order to fulfil the obligations of such research, and the mother must recognise herself in relation to a discourse of motherhood in order to be obliged by it. Given these preliminaries, the problem established here, by Foucault, is one of investigating ‘the types of relations that link together manifestations of truth with their procedures’, where these procedures depend on ‘the subjects who are their operators, witnesses, or possibly objects’ (p. 100).

The motherhood debate extends from psychology and developmental neuroscience to linguistic anthropology, sociolinguistics, sociology and education, and hence must stretch notions of truth across different disciplinary articulations of how truth is best arrived at. Yet despite this scope, academic debate does not succeed in demolishing the idea that there is a truth about motherhood and children’s language development worth pursuing, and not simply for its own sake, but for the benefit of all those involved. Indeed, although all academic actors in this regime of truth may be fulfilling their disciplinary obligations and will approach the construction of truth in different ways, they do so with regard to a wider debate they can never conclude to the satisfaction of all involved.

As Foucault argues, these commitments to tell the truth (as it is construed), and act according to that truth, might involve ‘truth obligations that impose acts of belief, professions of faith’ and perhaps even ‘confessions with a purifying function’ (Foucault, 1980/2014, p. 94). Within the context of academic debate, these are, of course, curiously emotive terms. And yet, motherhood is an emotive concept. To extend Foucault’s argument, one might suggest that the motherhood debate may also have a ‘purifying function’ for those engaged in it. Debate offers an opportunity to right the wrongs of previous misunderstandings or epistemic wrongdoings. This debate is ongoing and seemingly inexhaustible. With so much still to be corrected or at least critiqued, those involved will find satisfaction within the debate itself rather than in its resolution. This is the kind of satisfaction that may be found within the ‘constraint of the unverifiable’ (Foucault, 1980/2014, p. 95).

Policing families, schooling mothers

There is no easy escape from a regime of truth: to discuss it is to already take part. Even to critique research on mothers’ talk and its potentially harmful effects is to perpetuate that debate. Critique of this sort is, of course, important, and might extend the work of key studies attending to the subjectivity of motherhood (Gillies, 2007; Skeggs, 1997, 2013; Smart, 1996, 1992; Walkerdine & Lucey, 1989; Walkerdine et al., 2001). These studies show how the mother is constructed as a governed subject, who must commit, according to her obligations, to her own ‘manifestations of truth’ (Foucault, 1980/2014, p. 100). An important, though peripheral, reference point in these key studies is the work of the French sociologist, Jacques Donzelot, in The Policing of Families (Donzelot, 1977/1979). We argue that Donzelot’s conceptual work is applicable to the debate surrounding working class mothers and their children’s language development. We return to this study in particular to explain why debate about motherhood is driven to extend itself indeterminably and builds that commitment into its architecture.

With its focus on ‘policing’ families, Donzelot’s analysis of the family as an object of social policy and intervention is indebted to Foucault’s (1975/1991) work on disciplinary power. But it also coincides with the gradual displacement of that schematic, and Foucault’s subsequent work on the question of ‘subjectivation’. Foucault defines the latter as an investigation of ‘the way[s] in which the individual establishes his [or their] relation to the rule and recognizes him[or their]self as obliged to put it into practice’ (Foucault, 1984/1992, p. 27). Donzelot’s study can, then, be read through Foucault’s subsequent work on subjectivation, understood now as an analysis of the ways in which bourgeois and working class families were differently recruited, from the nineteenth century onwards, serving as agents of a developing regime of truth concerned with the so-called liberation of children from the ‘malpractices’ of traditional childrearing.

According to Donzelot’s analysis, ‘the modern family is not so much an institution’ (as Smart, 1996 claims), but serves ‘as a mechanism’ (Donzelot, 1977/1979, p. 94, original emphasis). As an institution it would function as a bulwark against change. As a mechanism it serves as an agent of adjustment. Indeed, Donzelot argues, ‘this familial mechanism is effective only to the extent that the family does not reproduce the established order’ (p. 94). The old social order of feudal patriarchy was reproductive in that sense and would have to be replaced by a new politics of family life, one that positioned childhood, and the development of children at its centre. About this central concern mothers function as agents of adjustment, helping undermine feudal patriarchy with a regime that rendered family life porous to the changeable configurations of modern power. Extending throughout the nineteenth and into the twentieth century, this regime depended on the development of a ‘privileged alliance’ between experts and mothers, wherein: ‘The doctor prescribes, the mother executes’ (p. 18).

The ‘advancement of [bourgeois] women’, Donzelot claims, was tied to growing ‘recognition of their educative usefulness’ as agents of change in and beyond the household (p. 18). This reflects their status as members of an ‘educated’ middle class, and hence, their positioning as subjects of power who are considered both reasonable and reasoning, and therefore worth working with on this basis. Indeed, it was on this foundation of their assumed cooperation, and their presumed ability to assume responsibility, that ‘this promotion of the woman as mother, educator, and medical auxiliary’ would serve ‘as a point of support for the main feminist currents of the nineteenth century’ (p. 21). This so-called promotion of the bourgeois mother as a reasonable agent is at its very best a mixed blessing, since it was also the route to a new mode of submission. It placed the bourgeois mother within an adjusted network of familial governance, in which she is assigned her role as agent of adjustment within a structure that remains, of course, patriarchal (as Walkerdine & Lucey, 1989, also observe).

Donzelot claims that the development of a unified school system also produced a shift in how the family functioned. The growing expectation that all children had a right to some sort of state funded education turned the bourgeois family inwards as it sought to optimise its internal ties and reinforce its advantages against social inferiors who were now guaranteed a basic education and might work their way up an educational structure that was becoming increasingly accessible to social climbers. The bourgeois family would, for this reason, pioneer techniques to augment the family environment. This ‘tactical withdrawal into itself’ brought about ‘an intensification of family life’, as Donzelot argues, as the family became ‘an avid consumer of everything that might help it to “realize itself”’ (p. 224, original emphasis).

Working class women were subject to different tactics (see, for example, Smart, 1996; Walkerdine & Lucey, 1989). For Donzelot, a key transition was in the development of jobs figured as ‘a natural extension’ of working class women’s domestic duties (p. 39). Social agencies training women for these roles could then begin to influence the practices of the family, with ‘feminine’ jobs figured as a kind of preparation for family life. Heavy state intervention was hereby avoided by promoting the ‘autonomy’ of family practices. The mother would be shaped as a key agent by an institutional network that coupled material assistance with moral guidance. As Donzelot argues, such assistance was increasingly tied to the ‘painstaking investigation of needs’, entailing ‘surveillance of the family’ and ‘full penetration into the details of family life’ (pp. 68–69).

These enquiries did not generate conclusive findings, and that, for us, is the key insight to be found in Donzelot’s analysis of motherhood. They led to the inauguration and growth of modern disciplines (including sociology) that set out in pursuit of a form of understanding, a full understanding of the human condition, that would always elude them. This is what Foucault (1975, p. 312) describes in The Order of Things as the ‘analytic of finitude’ which governs and besets the modern empirical sciences. From the perspective of power this is not a problem; indeed it is functionally useful. The endless pursuit of truth, in this case the truth of motherhood, ensures that the figure of the mother remains forever indeterminate. As she assumes a central role in the regulation of social life, the makeup of the mother, her best characteristics, and in our case, the principles underlying the perfection of mother’s talk, remain disputed. She must be endlessly worked upon and adapted. This denies motherhood the kind of permanence or security that might allow mothers to fix onto and draw reassurance from a stable configuration of their role. As social agents, mothers are instead expected to remain supple, if not also insecure in their position, so that they are ready to change their practices and effect social transformation via the family unit in response to changing norms and requirements.

In Donzelot’s analysis, the development of early twentieth century psychoanalysis, with its entourage of educational, sexual and marriage counsellors, symptomises the basic logic of this regime of truth. It operated by (1) reducing expectations of expert intervention, where those engaged in psychoanalysis are expected to direct the course of their own recovery (‘the first job’ of this relational technology consists in ‘discouraging the demand for expertise’, p. 211, original emphasis); (2) setting in motion the ‘floating’ of social norms and family values (p. 8), where these norms and values were no longer tied to an established canon of fixed laws and standards but are up for perpetual renegotiation between actors and experts; and (3) treating the family environment as a relational mechanism that may suffer from ‘communication gaps’ between those concerned (p. 212, original emphasis). This approach ‘does not point the finger at any particular person or wrong behaviour; it places the blame, rather, on the relations that obtain within the family and on the unconscious mental representations of its members’ (p. 214).

We believe that a broadly analogous strategy can be seen reflected in the research summarised below which, as it has become more sophisticated, has moved from measurement of mothers themselves or features of their own language, towards greater examination of dyadic features of the interaction between mother and child (see, for instance, Hirsh-Pasek et al., 2015).

This mode of governance is effective because it reforms judgementalism so that it becomes indirect, circuitous and cautious. In Donzelot’s scheme, when judgement is careful to avoid a simplistic identification of fault, it is most effective in convincing the family of its usefulness. This overcomes the defensiveness that a straightforward attack on mothers might produce. The open-ended nature of psychotherapy, its careful positioning so that it does not seek ‘any other mandate than the subject’s demand’, its way of conducting its operations ‘only on the basis of what is furnished by [its] clients’, and its determination to ‘reject beforehand any possibility of intervention’ or promise of finding a solution (Donzelot, 1977/1979, p. 170), helps explain how it was able to interpolate itself into family life and social consciousness so effectively. It operates a regime of truth that is necessarily indeterminate, and for that reason, all the more obliging. Indeed, one might say that the inadequacy of expertise is here employed as a governmental strategy. Perhaps the last and most important legacy of early twentieth century psychoanalysis was its inability to ‘meet the demand’ it helped create, with its ‘little army of counsellors and psychologists’ (p. 219). Instead, it threw back the challenge (it could not meet) so that it became a commitment to permanent work on the self, the endless labour of self-work, or in our case, the endless (and unforgiving) labour of seeking to improve one’s family environment.

In the next section we detail how this is all reflected by and further extended within the motherhood debate. But it is already worth noting here that there is no expectation that widespread access to high quality, expert-led language interventions (as might be offered to a middle class family, e.g. Rose et al., 2020) is the solution. Instead, research seeks to narrow down the precise reasons for deficiencies in mothers’ interactions (Rowe, 2018), or offers intervention programmes specifically designed to respond to large demand within low socioeconomic populations (Knight-McKenna et al., 2022).

Our overall claim is this: The motherhood and child language debate operates within a regime of truth, or a realm of power of the kind Donzelot schematises, where expectations are reduced concerning the role and remit of experts, and the role of mothers is correspondingly enhanced (or further burdened) as it is enlarged. Perhaps most disturbingly, even arguments made on ethical grounds for privileging the agency and expertise of mothers risk only further extending the reach of this regime. Instead, all parties work to ensure that the motherhood debate avoids closure (which might be found in the ‘truth’ of motherhood, or the perfection of evidence-based intervention), ensuring that motherhood remains an open and changeable category, flexible and supple before power.

The motherhood debate as regime of truth: An outline account ⁴

The influence of the Hart and Risley (1995) study on policy and practice, combined with its considerable methodological shortcomings, inherent racism and sociocultural biases, has only fuelled its continued citation and discussion. Some have objected to its small and selective sample size and called for larger scale studies. These studies have added statistical power to the idea that maternal language input is differentiated across socioeconomic groups, with significant consequences for children’s vocabulary, syntax and language processing skills (Gilkerson et al., 2017; Hoff, 2003; Huttenlocher et al., 2010; Rodriguez & Tamis-LeMonda, 2011; Rowe, 2012, 2018) as well as cognitive skills such as executive function (Sarsour et al., 2011), mathematical skills (Levine et al., 2010) and social skills (Connel & Prinz, 2002). It is claimed that this impacts on school readiness and the perpetuation of education inequalities (Hoff, 2013), ultimately facilitating the transmission of socioeconomic status (Farkas & Beron, 2004).

The rationale for intervening in motherhood, from this perspective, could not be more urgent, and supports the basic drive Donzelot (1977/1979, p. 18) observes, where: ‘The doctor prescribes, the mother executes’. Statistical power affords the necessary justification and preliminary motive, in claiming that (to adapt Donzelot again): The mother condemns by her actions, and so the doctor must prescribe. Maternal language remains firmly centred by all this research activity as the point where work addressing disadvantage and inequality can be leveraged. At its most individuating, this research clearly articulates well with the logic of neoliberal governance, where the mother is firmly positioned in terms of her personal responsibility. This is worth noting, since research of this inclination and outlook is clearly operating at a far higher level of overt collusion with the status quo than some of the cultural critics (whose interest lies with social structure) mentioned below.

Very much continuing this focus on the mother as the site of responsibility and adjustment, other researchers have sought to refine Hart and Risley’s approach, shifting attention from the total number of words towards differences in: the number of unique words (Hoff, 2003; Huttenlocher et al., 2007; Pan et al., 2005; Rowe, 2008), discourse features (Hoff, 2003) and syntactic features (Huttenlocher et al., 2007). This research outlines a far more complex terrain of intervention in what Smart once described as ‘the calibrations of good motherhood’ (1996, p. 46), where maternal language is to be improved in its qualitative detail.

There have also been efforts to refocus from lexicon towards distinct features of the mothers themselves including the impact of mothers’ education levels (Hoff, 2003; Hoff et al., 2018; Huttenlocher et al., 2002; Levine et al., 2020; Rowe & Goldin, 2009). The mother’s sensitivity to her child has also been studied, involving a determination of ‘prompt and appropriate responses’, and how closely aligned the mother’s talk is to the child’s gaze (Leigh et al., 2011; Tamis-LeMonda et al., 2009).

Recent work has focused even more tightly on the interpersonal nature of that talk, on the apparently measurable quality of mother–child interactions (Hirsh-Pasek et al., 2015). Crucially, what a ‘quality mother’ looks and sounds like remains in question and is subject to further study and debate. This helps maintain the ‘floating’ of norms Donzelot (1977/1979, p. 8) observes, where norms of quality motherhood are not to be tied to fixed values but are to be treated as changeable according to variations in context and setting.

Research has also helped drive technological developments, in particular the production of wearable devices that hugely increase the amount of data that can be collected. Directly inspired by the Hart and Risley study, the Language Environment Analysis System (produced by the LENA Foundation) audio-records the language heard by a child for up to 16 hours and then automatically produces data such as number of adult words, number of adult–child conversational turns and number of child utterances (Ganek & Eriks-Brophy, 2018). This system was originally intended for parents to monitor their own language – and so has that governmental potential – but its remit was also extended to examine variation in maternal language practices (Weisleder & Fernald, 2013; Zimmerman et al., 2009).

A further technological advancement to the field is the integration of brain imaging into the study of associations between language input and child language outcomes (Brito et al., 2020; Garcia-Sierra et al., 2016; Merz et al., 2019; Sheridan et al., 2012). This emerging area of research examines the relationships between pre-recorded measures of maternal language input and brain activity as shown on functional MRI. This currently very small research field has had a disproportionate influence in policy (Edwards et al., 2015) and influential agencies such as the LENA Foundation (e.g. ‘Early talk is one of the most important factors shaping children’s brain development during the first few years of life’, www.lena.org/about/).

In the shift from the simple message ‘talk more’ to a more complex directive to ‘talk differently’ (if not relate differently to one another), interventions have been trialled for mothers to instruct them in how to better interact with their babies, infants and children in order to accelerate language development (Greenwood et al., 2020; Heidlage et al., 2020; Landry et al., 2008; Leung et al., 2020; Suskind et al., 2013). Treating the family as a mechanism (Donzelot, 1977/1979, p. 94) that can be optimised, interventions range from small intervention studies (e.g. Gibbard & Smith, 2016; Heymann et al., 2020; Hindman et al., 2016; Laundry et al., 2008; McGillian et al., 2017) to large-scale application of policies (Lieberman, 2018) and programmes such as Sure Start (Potter & Hodgson, 2007) and Head Start (Lipsky, 2013). The focus of these interventions – the vast majority of which have been in the USA and Europe (Roberts et al., 2019) – is often on changing the language environment, despite a lack of evidence for what constitutes an ‘appropriate environment’ and what aspects of poverty affect language (Marshall et al., 2007, p. 552). Indeed, critics have argued that such interventions might have the opposite effect: by apportioning blame and stereotyping they may be causing harm to those families they are designed to help (Adair et al., 2017; Baugh, 2017).

Research in the areas of anthropology and language socialisation has raised concern about the emphasis on working class mothers, who are then held responsible for the apparent verbal deprivation of their children (Avineri et al., 2015; Blum, 2017; Dudley-Marling & Lucas, 2009; Johnson, 2015; Miller & Sperry, 2012; Sperry et al., 2019a). It has been argued, even within developmental psychology (Bishop et al., 2016; Pace et al., 2017), that any associations between maternal social class, maternal language features and child language outcomes may well be just that – associations rather than causal pathways:

Nowhere has it been established that rate of language development bears any relation to later linguistic or educational competence. This is simply an unanalysed assumption that is over-determined by the technologisation of development and the drive to maximise and accelerate it. (Burman, 2008, p. 210)

Meanwhile, those who argue against the existence of a word gap (Sperry et al., 2019a, 2019b) are still vigorously rebutted (Golinkoff et al., 2019), with arguments made that there are attendant risks to school achievement in all subjects (Pace et al., 2019), in self-control (Roben et al., 2013) and in executive function (Matte-Gagné & Bernier, 2011).

Cultural critics have argued that much of the above debate is severely narrow in its cultural outlook, pointing out that the kind of infant-directed, contingent reciprocal conversation between mothers and infants researched above is anomalous in terms of global cultural practices (Blum, 2017; Henrich et al., 2010; Sperry et al., 2019b) and is socioculturally defined (Brown & Gaskins, 2014; Miller et al., 2005a, 2005b; Ochs et al., 2012; Rogoff, 2003). It is observed that children in societies where there is very little direct interaction with infants nevertheless reach linguistic milestones at a similar rate (Brown & Gaskins, 2014). Maternal language input has been shown to vary widely in both quantity and quality across cultures, with linguistic anthropologists restating that mothers in some cultures do not routinely direct speak to babies with no lasting consequences for children’s language skills (Brown & Gaskins, 2014; Henrich et al., 2010; Miller et al., 2005a, 2005b; Ochs et al., 2012; Rogoff, 2003). It is likely, some claim, that children in contexts of very little directly related language will attend to others’ interactions in a way that is different from children who learn to see directed interactions as informative based on their social experiences (Shneidmann & Woodward, 2016).

These observations are important, and yet, in making them, cultural critics have added their own impetus to the argument that norms of good mothering must remain in question. This again supports what Donzelot describes as the ‘floating’ of norms, where the government of motherhood in his analysis relies upon the fact that these norms are kept unstable, and hence open to governmental intervention and adjustment. To precisely ‘fix’ the norms of good motherhood would, in Donzelot’s analysis, reflect the logic of a system of feudal patriarchy which has been surpassed, where the family served as an institution, and not a mechanism. We would also claim that such cultural critics have also (and again perhaps despite themselves) served to considerably expand the scope of the motherhood debate, extending the range of what is rendered to understanding, and hence available to adjustment. It helps furnish a degree of local sensitivity to the operations of power and the manipulation of motherhood that might, otherwise, be unavailable. The findings of cultural critique are of course still to be welcomed and hold the most potential to tear away at some of the debate’s foundational assumptions, and yet for all the cultural critics’ laudable investment in challenging assumptions and indeed transforming the social relations in which motherhood is figured, a future is still imagined in which they have a continued role as its critics and advocates.

Perhaps unsurprisingly the work of cultural criticism has remained largely unheeded within large segments of the motherhood and language development debate. Research has, nonetheless, worked to expand its understanding of the immediate environment of mothers’ talk. More recent research has focused on the dynamics between an ‘active child and responsive parent’, examining dyadic features (Gilkerson et al., 2017; Hirsh-Pasek et al., 2015; Rowe et al., 2005; Weisleder & Fernald, 2013). This research helps ensure that responsibility is not simply attached to the mother, but the larger relational environment (which is a product of multiple conditioning factors) is now in question. Such a research trajectory again opens the way for a potentially more subtle form of intervention, that more effectively co-opts mothers by avoiding directly apportioned blame.

Much research is now focused on mothers’ contingent responsiveness (Bornstein et al., 2008; Hirsh-Pasek et al., 2015; Kuchirko et al., 2018). While some research looks at the interactive features between mother and child as standalone factors, others repackage this as ‘sensitive parenting’ arguing that ‘a global description of a caregiver’s provision of warm, responsive, and stimulating engagement with his or her child predicts language outcome’ (Hirsh-Pasek et al., 2015, p. 1072). The claim here – that ‘well-intentioned interventions may not maximise their impact if they do not improve the quality of language children hear and how it is woven into the fabric of early caregiver–child interactions’ (p. 1072) – offers another example of how the scope of intervention has been considerably expanded. The very fabric of care – the warmth of the mother (as viewed by others) – is at stake.

Ours is, necessarily, a highly selective summary of a vast and growing literature. It serves to show that the basic trajectory observed by Smart (1996), from defining the ‘good mother’ in terms of external, easily observable characteristics (such as word usage), to judging mothers in terms of their adequacy to ‘love properly’ (p. 46), can be observed again in the specific context of the motherhood and language development debate recounted above. But there are some crucial, additional features worth observing. These extend, or bear out our discussion of Donzelot’s genealogy of parenting. Chiefly, they involve (1) attempts to complicate judgemental categories so that experts retain a certain kind of ‘diplomatic-epistemic’ uncertainty concerning their findings, and (2) the strategic positioning of mothers in research discourse as key operators of the mechanism of motherhood, which must remain open to new understanding, shifting judgement and new points of potential access and manipulation.

On shifting the accent, by way of a conclusion

In this article we have worked towards an objective similar to that described by Foucault (1980/2014), who sought to show through his conception of a regime of truth that truth-effects are not inevitable. The position he advances does not allow that ‘truth, by right and without question, has a power of obligation and constraint over us’. Rather, by ‘shifting the accent’ from the pursuit of truth to ‘the force we accord truth’, he draws attention to the ways in which we are bound, and bind ourselves, to particular manifestations of it (Foucault, 1980/2014, p. 101). This article explores the effects of that commitment in a research community that remains wedded to its ‘truth’ and does not see in each critique, however well intended, an argument for greater intervention. The research community is committed to a constellation of activity that, for all its internal rivalry and disagreement, remains obligated to the ongoing necessity of objectifying motherhood.

Our purpose in confronting the presumed importance of this debate is not simply to demonstrate its insatiable appetite to weigh in on the truth of motherhood. We draw attention to its existence as historically formed, and hence inessential, to make the point that its regime of truth does not have to be the only one available. It might still be possible ‘to constitute a new politics of truth’ as Foucault (1977, p. 14) once put it. This would require a mode of practical enquiry that ‘alters both our relation to truth and our way of behaving’, that seeks to detach us from what is accepted as true and change how we relate to ourselves and others as beings who bear witness to the truth we accept in our actions and everyday conduct (Foucault, 1980/2000, p. 327). We use the words ‘us’ and ‘we’ advisedly, and to refer to the research community first and foremost, before and above their objects of interest. We seek to question the implicit assumption that it is necessary to submit mothers and their children to a form of study that involves endless scrutiny and inspection, and invite, as we do so, all those engaged in the debate described in this article, to consider their own complicity in the refinement of a regime of truth that takes motherhood as its (endlessly augmentable) object.

It is both remarkable, and utterly symptomatic, we think, that an entire academic machinery, spanning several decades and disciplines, has remained committed to that objective, seeking to define and redefine, contest and define again the most favourable conditions of motherhood as if that will lead to ‘the salvation and deliverance’ (Foucault, 1980/2014, p. 75) of mothers, their children, and by extension society. We do appreciate just how difficult it is to avoid the temptation to be right about motherhood in research and in everyday discussion. But it is worth recognising how this temptation to constantly adjudicate on motherhood is the product of a longstanding interpolation of mothers as governmental agents of change. The entire machinery of judgement that orients itself to motherhood and then rebounds to an endless critique of itself, is a consequence of the positioning of both motherhood and research as mechanisms not institutions (in Donzelot’s sense), that is to say, as functional devices within a social order that is built on flux and uncertainty. Researchers and mothers are figured within this regime as flexible, indeterminate agents, a system that encourages insecurity for the suppleness that insecurity creates. The problem, then, is not one of ending or redirecting a particular set of academic debates. Rather, it is a matter of foregrounding the social relations in which these debates are embedded, debates in which mothers are persistently centred as among the most important (and culpable) agents in their children’s upbringing. At issue here is the continued role of even the most radical critics in this regime of truth, who, by their continued participation as critics, even as dissenters, must function to uphold and perpetuate a debate they set out to challenge. This is done through the very act of introducing greater grounds for disagreement into its workings, and greater reason for exercising doubt. This regime of truth functions, as we have observed, because of its ability to remain unresolved, and its capacity to invite continued interest in debating such an emotive topic. This entire focus on motherhood within academic debate might indeed be judged as a systematic evasion of responsibility, where academic researchers tend to focus on the culpability of others (and one another), rather than question their own role as agents of government and truth.