Molecularisation and metaphor

Introduction

Geneticisation indicated a future in which ‘individual existence is to a large extent affected and permeated with predictive genetic information’ (ten Have 2001: 302).

According to Kay (1993), the ‘molecular vision of life’ could be traced to the 1930s when the older ‘molar’ image of life began to be replaced by an analysis based on molecules. This shift was further promoted, it was claimed, by ‘the political and cultural aspects of the Cold War, its emphasis on information and control, and its permanent state of scientific mobilization in preparation for war’ (Quirke and Gaudilliere 2008: 444). In its turn this process introduced a new ‘regime of knowledge production’ based around the experimental laboratory and the clinic (Berlivet 2008). Clarke et al. (2003: 162) also identified the parallel process of ‘biomedicalization’ in which ‘transformations of both the human and non-human (is) made possible by such techno-scientific innovations as molecular biology, biotechnologies, genomization, transplant medicine, and new medical technologies’. The resulting emphasis on pharmacological/molecular treatments also gained its own label of ‘pharmaceuticalisation’ (Abraham 2010). Niewohner (2011: 279) claimed these changes had an important effect on the body of the patient: ‘the practice of doing epigenetic biology contribute to a molecularisation of biography and milieu, suggest the configuration of somatic sociality and produce a different concept of the body’.

While the broad process of molecularisation (and geneticisation) in medicine has not been challenged, the thesis has been qualified in a number of ways ranging from Latimer's (2013) claim that it was too totalising to Arribas-Ayllon's (2016) argument it was too simplistic. So far, its impact on social life has probably been over-played (Condit and Williams 1997) and the process has not been sufficiently grounded in empirical investigation (Hedgecoe 1998). Even though studies of genetic nosology (Hedgecoe 2003) and the work of genetic scientists (Darling et al. 2016) have since emerged, the broad sweep of medicine's involvement with genes and molecules cannot be captured by case studies alone. This article tries to map that wider medical engagement by exploring how molecules have been framed in the laboratory and clinic over time. The method involved analysing reports in the New England Journal of Medicine (NEJM) a weekly American medical journal first published in 1812 that today is the most highly cited medical journal in the world. This resource provides a chronicle of how clinicians and scientists viewed molecules in relation to the practical activity of patient care.

All copies of the NEJM are now digitalised and word/phrase searchable. Preliminary searches for the word molecule, genetics and associated concepts, such as genotype and phenotype that emerged during the analysis, were used to develop the account provided here. Due to the high number of references identified and the number of quotations provided below only the relevant date is provided rather than the full citation. Interested readers can search for the details in the NEJM database (http://www.nejm.org/medical-search) using the provided text and year of publication.

Molecular medicine

In the opening years of the 20th century the NEJM believed that future progress in clinical medicine would depend on chemistry: ‘we want a knowledge of the atomic and molecular structure of cells, of the changes which take place in the atoms and molecules in health and in disease, and of the effect of medicines and remedial procedures upon them’ [1901]. As predicted, chemistry did indeed provide the core science for molecular discovery in the first half of the century, but in the second half it was overtaken by genetics. The latter had its origins in Mendel's classic work on the genetics of inheritance, first published in 1865 and re-discovered in 1900 – though it was another six years before reference to its findings first appeared in the NEJM. Over the following decades, Mendelian genetics then provided a framework for examining those diseases that medicine already knew had a strong inherited component.

The NEJM first used the term ‘genotype’ in 1917 to describe the patient's genetic inheritance. In time, the characteristics of the disease resulting from the genotype came to be described as the patient's ‘phenotype’, but it took several decades for this meaning to be standardised. In the 1930s, for example, the phenotype referred to the non-genetic basis of disease: a study of the prevalence of hypertension claimed it had ‘all the hallmarks of being phenotypic rather than genotypic in origin’ [1935] and in another report the patient's personality or constitution was found to be ‘phenotypic’ [1935]. By the middle of the century, however, the meaning of genotype and phenotype achieved some stability when they were construed not as alternatives but as two sides of the same coin: the phenotype was the outward manifestation of the genotype.

Huntington's chorea provided a well-recognised model of how a Mendelian dominant gene could determine a specific disease in that patients with the gene always came to have the disease. In 1955, the NEJM published its first article reporting a ‘genetic association’ in which a relationship between genotype and disease was found, though not everyone, as in Huntington's chorea, had the disease in question. For the rest of the century and into the next, medical researchers reported large numbers of these genetic associations as they attempted to determine the heritable basis of both known ‘genetic diseases’ and multifactorial common diseases. The emerging relationship between genotype and phenotype, however, was far from simple. On the one hand, a single phenotype could have two different underlying genotypes [1944], particularly for blood groups [1949]. On the other hand, a particular genotype could be related to several different phenotypes [1960]. More challenging still, chromosomal analyses revealed ‘cases in which a normal chromosome pattern that does not agree with the phenotypic sex of the patient’ [1960]. For the next few decades, a succession of studies recorded the poor degree of association between genotype and phenotype. ‘The homozygous state does not incur the anticipated thrombophilic phenotype. It is intriguing to speculate about how this disparity between phenotype and genotype may have arisen’ [1994].

The lack of a consistent association between genotype and phenotype seemed puzzling. Sometimes the genotype might predict one aspect of the phenotype but not another: ‘Despite the strong association between the cystic fibrosis genotype and the pancreatic phenotype, this study demonstrates that the severity and course of pulmonary disease are not predicted by the genotype’ [1993]. At other times, a range of phenotypes was predicted: ‘In earlier studies of the association between genotype and phenotype, large variations in the severity of lung disease among patients with cystic fibrosis and the same genotype have been noted’ [1995]. The phenotype could also be present without the genotype: ‘A previous report described a patient with a non-classic phenotype in whom no CFTR mutations could be found’ [2002]. Finally, there were instances in which the predicted association was totally absent: ‘We found no correlation between phenotype and genotype; all four mutations abolished 1α-hydroxylase activity, and the clinical features in all four patients were similar’ [1998].

Given the independent generative status of genetics how was medicine to explain these poor levels of association? In 1933, in response to the ‘discrepancies between theoretical and practical studies of hemophilia’ (in which abnormalities were not transmitted to offspring), it was suggested that ‘modifying genes’ may have been responsible [1933]. In 1965, the idea of mosaicism in which ‘tissues of different genetic constitution lie adjacent to or are intermingled with each other within one individual’ was advanced to explain inconsistencies in the genotype-phenotype axis: ‘This attractive theory accounts for various observed facts’ [1965]. Another possible explanation was ‘intermediateness’ which ‘represents a “gray” area between a clearly recessive form of transmission and clearly dominant from the point of view of the characteristics of the phenotype’ [1964]. In 1965, the term ‘polygenic’ appeared in the NEJM in recognition that many different genes could be involved in determining the phenotype.

A more long-term solution to the problem of the poor correspondence between genotype and phenotype entailed the application of four key metaphors: gene penetrance, gene expression or expressivity, (gene) regulation and (molecular) pathway. The concept of penetrance was first used in the NEJM in 1949 to explain the way in which the patient's genotype varied in its ‘penetration’. At first, the question was whether penetrance was ‘complete’ or ‘incomplete’ [1957] but then estimates of the degree of penetrance were calculated: ‘this difference was consistent with a hypothesis of a single autosomal recessive gene as the basis of susceptibility to poliomyelitis, and a penetrance of approximately 80 per cent was calculated’ [1958]. Here, then, was a concept that might explain discrepancies between genotype and phenotype such as when ‘the penetrance was too low to ensure the phenotypical appearance in both twins’ [1964].

The idea of gene expression or expressivity was another way of describing penetrance but in an individual patient. Penetrance of 50 per cent indicated that half of patients with the gene also had the corresponding phenotype, but it was the degree to which the gene expressed itself that determined the severity of the phenotype in individual patients. The concept of ‘expressivity’ first appeared in the NEJM in 1955 and that of ‘gene expression’ in 1961. As the phenotype was the result of that expression, the term ‘phenotypic expression’ was commonly used: ‘Such lesions may be chance findings, but they are more likely to represent a general lowering of the threshold of phenotypic expression. This is a feature of many genetic disorders’ [1957]. The idea of phenotypic expression also allowed the interplay of non-genetic factors. ‘Hormonal and metabolic inter-relations’ [1975] might alter phenotypic expression as might ‘age, sex, alcohol and a number of other environmental factors’ [1977]. In effect, the metaphor of gene expression provided an account of why two patients with the same genotype could have different phenotypes. ‘Through further studies on genetic variation, we may begin to understand characteristics that are “multifactorial” and “continuously distributed” such as intelligence and height, the molecular meaning of “low penetrance” and “variable expressivity”’ [1968].

In descriptions of the molecular process the gene held attraction as the obvious explanatory starting point. But if the genetic origin somehow determined the resulting phenotype (in the context of factors such as penetrance, expressivity and regulation) then to what extent could subsequent events be attributed to genetics? It was suggested that perhaps the idea of a ‘gene’, and genotype, could be extended to include some of the contextual information about the physiologic and pathologic behaviour of related proteins in the form of gene ‘annotation’ – though ‘thorough and extensive annotation must be in order to provide a basis for a mechanistic connection between a gene and the clinical picture’ [2004].

Where the molecular process ‘finished’ was also less than clear. The phenotype was the downstream outcome but it could refer to the distant result of personal observable characteristics such as ‘fair-skinned … blue-eyed persons with blonde, golden or red hair’ [1961] or it could be the manifestations of a specific disease such as phenylketonuria [1991] or it could refer to more immediate ‘epigenetic phenotypes’ [2013]. In effect, there was a proliferation of ‘strata’ at which the phenotype might be described, which pointed to the importance of a further metaphor, molecular ‘pathway’, that connected these various phenotypic levels.

Molecular pathways led between the genotype and the phenotype. A pathway had a beginning, a point of origin: that was the gene (including its various parallel structural concepts such as DNA, SNPs, RNA, haplotypes, etc). The pathway was then subject to the activity of the gene in terms of its penetrance and the ability to express itself through systems of regulation. The end-result was the phenotype – though this would vary depending on the length and complexity of the pathway. Pathways interacted with each other, wove in and out, producing their varied phenotypes at different levels. So how were those phenotypes to be described given that one might subsume another? Were they to be described as the proximal consequence of the pathway, such as insulin resistance, or were they to be identified as more distal effects, such as diabetes?

One approach was to collect all the phenotypes under a broad diagnostic label such as ‘metabolic syndrome’ that seemed to capture many different pathways. Ironically, when diagnosis was reportedly becoming more individualised under precision or stratified medicine, the biggest shift in diagnostic practice in the early 21st century was the recognition of widespread molecular syndromes, the most common being the metabolic syndrome (or Syndrome X). The latter was a descriptor for a collection of population and molecular risks ‘a clustering of medical conditions that may include these risk factors [such as type 2 diabetes, hypertension, hyperlipidemia] plus abdominal obesity, hypertriglyceridemia, and insulin resistance’ [2002]. These risks could be explained in terms of pathways that were in turn underpinned by gene expression and regulation.

While the metabolic syndrome subsumed a number of molecular pathways it also incorporated non-molecular environmental factors. Metabolic syndrome captured ‘the interrelated characteristics of obesity, a sedentary lifestyle, hyperglycemia, hyperinsulinemia, hyperlipidemia, and hypertension’ [1999]. It was related to the pathways underpinning diabetes – diabetes was part of the syndrome – yet its ‘polygenic determinants of obesity, insulin resistance, and the metabolic syndrome may be distinct and are highly likely to differ from those determining type 2 diabetes’ [2006]. In its turn, the insulin resistance pathway characterised both the metabolic syndrome and polycystic ovary syndrome so that ‘the polycystic ovary syndrome might thus be viewed as a sex-specific form of the metabolic syndrome’ [2005].

In summary, while the logic of genetic medicine could easily be set out – ‘a person's complete genomic sequence (genotype), acting in concert with environmental influences, creates individuality (phenotype)’ [2010] – the relationship between that genotype and phenotype was far from straightforward. It was the deployment of several powerful metaphors, however, that served to make the confusing and inconsistent intelligible. The fact that the relationship between genotype and phenotype did not follow the basic rules of Mendelian genetics could be managed by claiming that relationship was affected by penetration, expression and regulation along a pathway, without necessarily implying any literal understanding of what those various concepts actually meant.

Molecules and metaphors

Metaphor has been recognised as an important mechanism for filling gaps in scientific theories where literal terms lacked conceptual precision (Haack 1988). Their use also opens up a space for creative interpretations of the underlying metaphorical analogy: ‘it is precisely because metaphorical statements are unspecific or open-textured that they are apt for representing novel conjectures in their initial and undeveloped stages, and for prompting investigation of what might be significant respects of resemblance’ (Haack 1988: 299). Science might even be construed as the process of literalising metaphors as once agreement on meaning was achieved the resulting ‘dead metaphor’ could no longer inspire and fertilise new ideas (Brown 2003, Rorty 1989): in that sense, ‘metaphor is a ladder that science aims to kick away’ (Haack 1988: 299).

In the field of medicine, Sontag (1978) pointed out the ways in which metaphors could affect the experience of illness but it was later investigators who identified the affinity between the metaphorical language underpinning biological and medical theories and the social world. In her study of the history of immunology, for example, Martin (1994) recognised the resonance with the larger body-politic of metaphors such as flexibility and training in how the body built up resilience to external threat. Similarly, in her history of 20th century biology, Keller (1995) identified the significance of descriptions of genetic code as a blueprint in a wider social context that was concerned with information technology. Overall, the discourses of (bio)medicine could be judged as thoroughly metaphor-laden (Segal 2005).

In their seminal analysis of metaphor as part of cognitive linguistics, Lakoff and Johnson (1980) argued that rather than treating metaphor as ornament or ‘mere language’, it was possible to detect patterns of translation from source to target domains as one thing was conceived of in terms of another. Many metaphors used to describe ‘argument’, for example, such as defend, attack, demolish, were derived from the broader metaphorical concept of war. Further, they argued that ‘the most fundamental values in a culture will be coherent with the metaphorical structure of the most fundamental concepts in the culture’ (Lakoff and Johnson 1980: 22) and the concepts in a scientific theory ‘are often – perhaps always – based on metaphors that have a physical and/or cultural basis’ (Lakoff and Johnson 1980: 19). In other words, choice of metaphor and metaphorical concept were unlikely to be arbitrary.

The four metaphors of penetration, expression, regulation and pathway that have played essential roles in the explanatory framework of genetics and molecular medicine are in the process of being literalised. The Oxford English Dictionary now accords all four terms literal meanings connected to genetics and metabolism. Penetrance/penetration is recognised as the production of an expected phenotype by a particular genotype; gene expression is defined as the appearance in a phenotype of a character attributed to a gene; regulation is described as ‘control of gene expression’ while pathways include the sequence of biochemical reactions within the process of metabolism. But these are not (yet) ‘dead metaphors’: their prior use in the NEJM indicates earlier meanings drawn from social domains that still resonate today.

The word penetration (penetrance was a 20th century derivative) was used almost exclusively throughout the 19th century in the NEJM to refer to the physical piercing of tissues. This use continued through the 20th century and into the 21st but, as described above, the term has now also been used to describe the way in which the genotype is related to the phenotype. Similarly, expression had wide use in the 19th and early 20th centuries to capture the expression of a belief, an opinion, or an emotion. Occasionally, it was used to indicate a representation (‘an expression of intoxication’ [1921]) or something being squeezed out (‘the cyst will express small quantities of fluid’ [1940]). In the second half of the century, these latter uses became more common relative to the idea of expressing a view. There was more reference to representation (‘expression of sensitivity’ [1951] or ‘expression of therapeutic effect’ [1952]) along with ‘fluid expressed from the wound’ [1961] or ‘expressed the placenta’ [1961]. So, when the term was applied in genetics (‘unbalanced X-chromosome would provide an increased opportunity for direct expression of certain genes’ [1961]) it marked both a transfer from contemporary use but also an echo of past meanings. These metaphorical antecedents suggested that when a gene ‘expressed’ and ‘penetrated’ there was an anthropomorphic intimation of speaking and acting.

Although the term regulation has been layered with metaphorical associations, its literal meaning, derived from the Latin regulare, is to control by rules. It occurred several hundred times in the 19th century issues of the NEJM where it had two main applications. First, in line with its literal meaning, it was related to governance through the application of rules, such as how organisations, particularly the medical profession, were ordered and managed. For example, reference to the regulation of dissecting schools [1829] or concern with the regulation of noxious trades and occupations [1854] or the regulation of prostitution [1867] reflected this concern with ordering social life. Equally, ‘port regulations’ [1831] to control the importation of disease and ‘sanitary regulations’ [1861] were both aspects of 19th century public health approaches to managing danger. Regulation, however, was also used in a more metaphorical way to refer to ‘regulation of the diet’ and ‘regulation of the bowels’. In that sense the individual had to follow the ‘laws of health’ – usually through adopting good habits – to regulate their body functions. In effect, the metaphorical use of ‘regulation’ construed the body as a potentially disruptive social form that needed to abide by rules if good health was to be secured. Medicine therefore promoted the ‘more scientific regulation of all the habits of life’ [1895].

These literal and metaphorical uses of regulation extended into the early 20th century when another target appeared, namely the regulation of some aspect of internal body function such as ‘regulation of the rate of oxidation within the body’ [1922] or the ‘normal regulation of fluid balance in the body’ [1924]. This metaphorical use was apparent in the occasional inverted commas that surrounded the term: ‘There is no evidence that any other non-volatile metabolites are concerned in the “regulation” of the coronary flow’ [1925]. This period also witnessed the appearance of the term ‘self-regulation’ to describe how body functions were managed through internal feedback loops. In the late 1920s the term homeostasis (later popularised by Cannon (1932)) appeared in the NEJM to reflect the way in which the regulation metaphor was contained within a broader theory of metabolic stability. The 19th century concern with regulation of habits had also treated the body as an organism but those ‘rules’ came from without; in the early 20th century, regulation came from within as the body was reconstrued as a self-correcting object.

By mid-century, reference to the regulation of body habits had almost disappeared while regulation of the body's internal systems increased, particularly in terms of the regulation of physiological processes involved in metabolism: for example, ‘regulation of carbohydrate metabolism’ [1938], ‘regulation of acid–base balance’ [1947] or ‘regulation of water metabolism’ [1955]. When, in 1967, regulation first appeared in relation to gene expression, it became a powerful explanatory term to describe the uncertain relationship between genotype and phenotype. The underlying analogy was to liken the genotype-phenotype axis to a social system and to see the power of the gene as a potential threat to underlying order. The ‘rules’ exerted and imposed by other molecules ensured those effects were constrained and managed.

The pathway metaphor showed a similar trajectory. Its use throughout the NEJM was almost entirely metaphorical (as the pathway of life, the pathway of great physicians, etc). In the 19th century, it appeared relatively infrequently and, when applied to clinical matters, was mainly concerned with pathways outside the human body such as the routes in the community along which bacteria spread. In the early decades of the 20th century, particularly from the 1930s, it was used increasingly to describe internal anatomical structures that appeared to be involved with transmission, either nerve impulses (as in nerve or neural pathways including optic, visual, sensory, sympathetic and parasympathetic pathways) or blood along vascular pathways. The metaphor of a ‘metabolic pathway’ first appeared in the 1940s and increased rapidly over the next two decades. In parallel, reference to ‘common pathways’, which mainly involved metabolic and molecular processes, also became more frequent. In short, the idea that the body contained pathways was a 20th century one. At first, it was applied to those anatomical structures that acted as conduits for transmitting flows around the body but in the second half of the century, when metabolic pathways appeared, it was less a physical movement and more a chemical/molecular one that captured the directionality and constraints that characterised the relationship between molecules.

In summary, the four metaphors of penetrance, expression, regulation and pathway transferred elements of their origin in a social world into the molecular. Through anthropomorphism and social rules, through representations and control, through physical trauma and sentiment, metaphors carried the constructs of one domain into another. In that way, the terminology derived from the governance of individuals and groups, together with pathways (that encompassed notions of flow, progress, direction and limits), came to be applied to the body's sub-cellular world to make it legible.

Molecularisation revisited

The pages of the NEJM provide an extensive resource for mapping over time the appearance of concepts and metaphors in clinical medicine. They do not offer a precise method for dating a new introduction nor of studying the vicissitudes underlying the discoveries of ‘science’ but rather show how and when these ideas could be presented to a broad medical audience. Moreover, these texts do not provide direct access to contemporary events: everything is refracted through the lens of NEJM writers. Nevertheless, there seems little doubt that over the last few decades or so medicine has developed a fascination with the molecular, particularly in the form of genetics and biomarkers. Between, say, the first appearance of the term ‘molecular disease’ in 1955 (with reference to sickle cell anaemia) and the introduction of the term ‘molecular medicine’ in 1995, a new molecular discourse emerged.

This molecular focus has been maintained even though the genetic promise, so forcefully proclaimed about the time of the Human Genome Project, has become more muted. ‘What is becoming clear from these early attempts at genetically based risk assessment is that currently known variants explain too little about the risk of disease occurrence to be of clinically useful predictive value’ [2010]. Indeed, by the end of the first decade of the 21st century, barely 10 years since the triumph the Human Genome Project, the limited potential for the clinical application of genetic research was becoming apparent. Despite the investment of ‘vast resources … the number of tangible benefits to patients remains small’ [2012].

Even though the ambitions of molecular medicine seem to have become more modest, there is little evidence that its ascendency has been threatened (Joyner et al. 2016). One factor maintaining that position could be the way in which metaphorical concepts of how the body works play into the molecular vision. In many ways, it is the hermeneutic tension behind terms such as penetration, expression, regulation and pathway that enables molecular medicine to maintain its promise. The existence of a pathway confers foundational status on its (genetic) point of origin while penetrance and the regulation of expression indicate disorderly forces that need control. These metaphors, as proposed by Lakoff and Johnson (1980), carry a physical and cultural domain into a scientific narrative.

The four metaphors described in this article enabled the problematic relationship between genotype and phenotype to be stabilised, analysed and comprehended. There were other metaphors used in genetics, such as mapping, but these did not address the genotype-phenotype ‘problem’. There were also many metaphors that could have been deployed to account for the relationship between molecules. One molecule could transmit or communicate with another as suggested by early understanding of genetic ‘code’ (Keller 1995; Rose 2006); one molecule could ‘train’ another therefore building flexibility and resilience (Martin 1994); one molecule could be ‘assembled’ or built alongside another as in a factory (Arribas-Ayllon 2016); or one molecule could invest or borrow from another, or attack and defend, and so on. Yet the metaphors of penetration, expression, regulation and pathway over-shadowed alternatives to account for how molecules were related in time and space. As Rorty (1989: 9) has observed, science progresses through ‘increasingly useful metaphors rather than of increasing understanding of how things really are’.

While clinical discourse might be infiltrated by molecularisation (or geneticisation), it is less molecules that have been given primacy than the relationship between them. Construing relationships as pathways in need of regulation implies a particular view of how the body functions. Genetic risk (a term first used in the NEJM of 1989) lies at the origin of the molecular pathway; it might affect other risk factors, events and illnesses, but was itself less amenable to being transformed by antecedent risks. In their turn, these pathways have the flexibility of both enabling external risks to be read in terms of internal changes (Darling et al. 2016) and allowing non-molecular factors to be integrated into an explanatory web (particularly in the form of external ‘risk factors’). To be sure, the emphasis continues to focus on the molecules within those pathways but it is more the connections between molecules, their relationships, the spaces between them that seems of greatest interest to clinical science.

Metaphors, largely drawn from the social world, are then deployed to describe – and explain – molecular relationships. In fact, it might be better to describe this process as ‘metaphorisation’ rather than molecularisation if this was not such a clumsy term. In a post-modern audit society in which regulation of the potential excesses of the free market features strongly and protocol-driven clinical pathways underpin the routinisation of patient care, the metaphors of molecularisation are not value free. If a society is preoccupied with regulating social life, with pursuing modes of self-expression and with organising social activity in predefined pathways, it is no accident that the molecular world seems to be arranged along similar lines. In this light, the reductionist thesis that underpins the molecularisation process seem to be incomplete: medicine, through its metaphors, reads the molecular world through a social lens. Perhaps even the fundamental reorganisation of the gaze of the life sciences that Rose (2001, 2006) identified might be viewed as being underpinned by the metaphors of social life.

In summary, the emergence of molecular pathways and ideas about the regulation of gene expression over the last half century point to both the strengths and limitations of the molecularisation thesis. Certainly, over this period, medicine has increasingly been involved in the study and promotion of the molecular and new forms of clinical practice – personalised medicine, individualised medicine, precision medicine, stratified medicine or molecular medicine – are said to be imminent. Molecularisation, however, needs to be seen in the context of the metaphors that enable it to exist. Truth, as Nietzsche famously observed, is a mobile army of metaphors, metonyms, and anthropomorphisms. It is the application of metaphors derived from the social world that underpins sub-cellular understanding and it is that process, in its turn, that tempers the apparent dominance of the molecular in social life.