VI - Narrative Sensibility and Argument

17 Anecdotes: Epistemic Switching in Medical Narratives

Brian Hurwitz

[F]rom Procopius’s Anekdota onward […] anecdotes have run counter to the order of imperial authorization.

Peter Fenves, Arresting Language, 153.

The anecdote is a narration that claims to present (whether true or not, verifiable or not) a historical event, usually a single event detached from other events.

Paul Fleming, ‘The Perfect Story: Anecdote and Exemplarity in Linnaeus and Blumenberg’, 74.

17.1 Introduction

This chapter examines the narrative means anecdotes employ to marshal observations and entertain interpretations of them. It also examines the way anecdotes point to explanations of the situations they recount. A term of long lineage that originally invoked ‘a revelation of events previously undivulged’ (Burke 2012: 60), anecdote in the eighteenth century gained wider currency as a striking account of an incident in a person’s life (Johnson 1773); informative, entertaining and often memorable, it took the form of a bon mot on notable happenings or repartee that praised and mocked human accomplishments (Adams 1789).¹

The anecdotes considered here partake in some of the features of such closely worked accounts; however, the main focus of attention will be on their modes of noticing and narrative reasoning and how they bring situations and conceptual frameworks for understanding them into close relationship with each other. In An Essay Concerning Human Understanding (1690), John Locke recounted an anecdote concerning the behaviour of water, which contained a striking twist:

As it happened to a Dutch ambassador, who entertaining the King of Siam with the particularities of Holland, which he was inquisitive after, amongst other things told him that the water in his country would sometimes, in cold weather, be so hard that men walked upon it, and that it would bear an elephant, if he were there. To which the king replied, ‘HITHERTO I HAVE BELIEVED THE STRANGE THINGS YOU HAVE TOLD ME, BECAUSE I LOOK UPON YOU AS A SOBER FAIR MAN, BUT NOW I AM SURE YOU LIE’.

(Locke 1690: vol. 2, book IV, chap. 15, para. 5)

That the surface of water could support the weight of a human being proved too fantastical a particularity for the king to accept. Although a ‘fair man’, the Ambassador failed to overcome what Bernard Williams called the ‘positional disadvantage’ (Williams 2004: 55) of his royal interlocutor, a disadvantage that was geographical and above all perspectival.² To persuade the king of its truth – never having ventured into cold climates – the Ambassador would have to have reframed his notion of the fluid. His failure to do so stemmed from the overly narrow way he approached the task. Whether it was water’s capacity to solidify or the nature of the solid once formed that was inconceivable, the unfamiliar object to the king was ice (not water). Had the Ambassador sought to expand the range of the king’s conception of its physical states, he might have had more success.³ In the event, reliance solely on his own testimony did nothing to transform the intellectual frame the king brought to the fluid. The anecdote offered a pointed example of how the credibility of a witness’s claims turns on their conformity with the experiential and conceptual frames of an audience (Shapin 1995: 244).

In one of the earliest treatments of anecdotes in English, Isaac D’Israeli found this ‘species of composition’ (Anon. 1799: 181) to comprise ‘minute notices of human nature, and of human learning’ (D’Israeli 1793: 80–81). Impressed by its special kind of realism, the twentieth-century literary scholar Jürgen Hein noted the propensity of anecdotes to present ‘a differentiated incident’ and a ‘dramatically shaped action or saying […] through a social story-telling situation’ (Hein 1981: 15–18).⁴ As a stripped-down account claiming import and significance beyond the particulars recounted, the literary scholar Joel Fineman found anecdotes provide ‘pointed, referential access to the real’ in structured formats that carry ‘peculiar and eventful narrative force’ (Fineman 1989: 67), a forcefulness, I argue, that derives from epistemic vantage points, narrative branchings, reversals and punchlines (Beatty 2016; 2017).

The anecdotes to be discussed appeared in medical publications over the past two-and-a-half centuries. Selected for their narrative and epistemic articulations and the way they connect healthcare scenarios to general propositions and frameworks, these accounts effect ‘epistemic switches’, a term Richard Wollheim coined in relation to a thought experiment of A. J. Ayer’s, concerning a man whose knowledge of something was initially based on the testimony of others, until he found the source of it – unchanged in content – in his own memory. For Wollheim (1979: 199), ‘epistemic switch’ referred to a change in the reasons for a true belief. I adapt it here to include a sudden change of evidential base, which alters the explanatory level of a phenomenon, by changing the domain in which it is explained from, for example, a biomedical to a socio-psychological domain or from a normal to a pathological one. By swapping perspectives on their objects of contemplation, anecdotes bring before the mind’s eye scenes viewed in a new light.

17.2 Epistemic Effects of Medical Anecdotes

A dialogue published in The Lancet in 1824 headed ‘Anecdote’, involved Michel-Philippe Bouvart (1717–87), a French physician to Louis XV’s court, and his patient, a Marquis recovering from a severe illness:

‘Good day to you, Mr. Bouvart, I feel quite in spirits, and think my fever has left me.’ – ‘I am sure of it,’ replied the doctor, ‘the very first expression you used convinced me of it.’ – ‘Pray explain yourself.’ – ‘Nothing more easy. In the first days of your illness, when your life was in danger, I was your dearest friend; as you began to get better, I was your good Bouvart; and now I am Mr. Bouvart: depend upon it you are quite recovered’. Ward’s [sic] Nugæ Chirurgicæ

(Anon. 1824: 256)

The proposition advanced is that bedside salutations can be medically noteworthy. In response to the Marquis’s declaration of buoyant spirits, the doctor positioned his greetings – hitherto peripheral to the Marquis – as central to his clinical assessment of him and claimed that his expressions of affiliation were a function of the disruptive effects of illness. The contention could have been based on an antecedent generalization: that familiarity of greeting varies inversely with recuperation, recovery restoring social and inter-personal distance. But how well established was such a generalization? Was it grounded in Bouvart’s personal observations – in contemporary terms, a small, biased sample likely to engender overgeneralization – or in an accepted maxim or hunch, entertained for suppositional purposes? Whatever its source and standing, the claim put a spotlight on social exchanges in clinical practice, elevating their notice to a level of prognostic significance. In proposing that greeting style carried meaning beyond simple etiquette, the dialogue showed how clinical attention could move from biomedical considerations to a domain of socio-psychological observation that encompassed interpersonal aspects of doctoring.

James Wood finds anecdotes in the Enlightenment played an important role in advocating the naturalistic study of customs (Wood 2019). From this perspective, what the Marquis relayed to Bouvart – regarding his spirits and lack of fever – was less important to the medical assessment of him than the Marquis’s recourse to plain ‘Mr.’, which Bouvart located within a series ‘dearest’, ‘good’ and ‘Mr.’, a pattern that pointed to recovery. Even before the Marquis disclosed how he was feeling, the pattern supplied Bouvart with foreknowledge of his well-being.

Within Bouvart’s diagnostic process, an ‘epistemic switch’ takes place, which grants greeting style an evidential value. The scenario shows how medical consideration moves away from reported feelings and symptoms to focus on inadvertently expressed greetings which match a pattern of signs or clues; cognitively, the switch persuades readers that social and interpersonal relations could be a source of valuable information.

In creating a ‘picture through conversation and witness’ (Varillas 1686: unpaginated), the anecdote revealed a doctor drawing out meaning from social interactions which otherwise appeared incidental and mundane. Whatever doubts we may harbour today about its embedded generalization, the acuity of Bouvart’s discernment lends the account plausibility. If we remain sceptical about its central claim, the dialogue prompts less specific considerations, such as that other facets of healthcare exchange could vary with illness; concerning what sort of courtesy it is that intensifies at moments of extreme dependency; and about the levelling of birth and social rank by serious illness. Read in these ways, the anecdote brings a plurality of possibilities to the fore (Paskins and Morgan 2019), including radical and ironical meanings.

Thomas Wakley, founder editor of The Lancet, took the dialogue not from its earlier source in L’Esprit des journaux, but from William Wadd’s Nugæ Chirurgicæ, published in London in 1824 (Wadd 1824: 199). The volume offered a miscellany of medical portraits in which Wadd announced his intention of ‘blend[ing] the “utile” with the “dulce”, the learned and the ignorant, the regulars and the irregulars […] with the Republic of Medicine’ (Wadd 1824: i), and of reforming the hierarchies and outdated medical practices of the Ancien Régime.⁵ By interlinking the social world and health – even in the case of a Marquis – the anecdote captured something of the rationalist universalism of post-Revolutionary medicine, and the cultural authority of medical men. That this socio-political context was occluded in The Lancet’s retelling of the dialogue is characteristic of the anecdotal form, which the literary scholar Paul Fleming takes to be ‘a discrete isolated narrative’ that lacks ‘chronological connection to any surrounding narration of events’ (Fleming 2011: 74). Although anecdotes may cite news reports, witness accounts and informants’ memories as sources for events recounted, the literary scholar André Jolles found their defining characteristic to be a pronounced capacity to condense the ‘flow of events’, enabling diverse elements of a situation to be grasped in a ‘bound factuality’ (Jolles 2017: 161–174). As well as its sources, the anecdote’s credibility for Jolles also derived from its capacity to realize events and people ‘concretely’, qualities it shares with the clinical case and its capacity to delineate particularities and reason about them (Hurwitz 2017).

In 1985, The Lancet’s ‘In England Now’ column recounted an anecdote about a man who upturned the conceptual framework of an X-ray report:

How often have we stumbled on a diagnosis by accident and then taken credit for a good examination? Like the man whose chest we X-rayed routinely to find a peculiar thin line stretching from one lung to the other through the mediastinum. Technical fault, said the report, but lungs normal. Only when the patient commented on the number of technical errors his X-rays caused, did the penny drop. A thoracotomy revealed a bicycle spoke lying within the chest cavity, a relic of a road accident some thirty years previously. It produced some pleural pain and an interesting publication.

(Anon. 1985: 381)

There is something glib and ‘off pat’ about this account of how a patient was briefly recognized as a situated knower after he remarked how commonly a ‘technical fault’ appeared in his chest X-ray reports. The information made it a priori unlikely that yet another technical fault accounted for the X-ray appearance. His disclosure on the contrary suggested that the ‘peculiar thin line’ denoted a persistent lung abnormality. The patient’s mention of ‘pleural pain’ and a history of road accident constituted a decisive first-person vantage point from which the diagnosis inverted: no longer could it invoke ‘technical fault though lungs normal’ but instead ‘adequate image and lungs abnormal’, an epistemic switch that led to the removal of a spoke from his lungs!

The change in what the image signified was mediated through the patient’s attentiveness, memory and agency. His ‘context-dependent knowledge’ (Flyvbjerg 2006: 221) switched the interpretation of the X-ray from that of an artefact to that of pathology and led to publication of the image. Even so, the anecdote’s punchline implied minimal acknowledgement of the patient’s pivotal role in the diagnosis: although his vantage point changed the explanatory part of the explanation – the explanans – that applied to what required explanation – the explanandum – his role in the change appears not to have featured in the report.

An epistemic switch that removed the explanans of a phenomenon altogether occurred in an anecdote recounted by the general practitioner, writer and broadcaster Michael O’Donnell, in 2016:

A paediatrician travelling on a train grew intrigued by the state of an infant girl cradled in the arms of a woman who sat opposite him. He could see there was something abnormal about the child but couldn’t diagnose what it was. He carefully watched the respiratory movements of her chest, listened to the timbre of an occasional cry and whimper, observed the way she moved her limbs. But he remained baffled.

He described in The Lancet how 20 minutes passed before the penny dropped. The child had a condition that rarely came his way. She was perfectly healthy.

(O’Donnell 2016: 10)

Once the object of observational interest – ‘the state of an infant girl’ – had fallen outside the doctor’s frame of reference, the classificatory gaze wielded at the start of the encounter proved nugatory by the end of it. The story is about ‘uncontrolled observation’ and the medicalization of everyday life, through imposition of a narrow conceptual frame on a non-medical context. The change of object, from a sick infant requiring pathological explanation, to one needing no explanation, amounted to a radical reframing that effected an epistemic switch. But, as the punchline makes clear, the apparent rarity of such an encounter for the doctor – even on a train – carried its own compensation: a report in an international journal. In responding to what was unfamiliar by attempting to diagnose it – and only later appreciating that what he thought was ‘something abnormal’ was in fact a healthy child – the doctor fell victim to the insidious power of medicalization. Through publication of a case in which the medical framework had no grip – a case of nothing – the doctor’s mindset was rewarded and seemingly left intact his drive to deploy it again on the next train journey.

Despite its brevity, O’Donnell’s anecdote is quite a layered story whose meanings emerge as the framing shifts.⁶ A prototype for this effect is suggested by Bernard Williams’s discussion of how differences in spatiotemporal observational position justify tellings, on the grounds that tellers were there when the audience was not (Williams 2004: 55). The situated and conversational aspects of such grounds were noted by ethnographer and sociologist Paul Atkinson, who observed anecdotes recounted by UK and US haematologists during clinical ward rounds in the 1970s:

One of the most striking characteristics of speech exchange in daily rounds is the deployment of personal narratives and reminiscences on the part of senior physicians [… who] claim, tacitly, the right to tell stories and to relate medical knowledge back to their biographical experiences. The justification for the story as evidence does not derive from the warrant of textbooks, journals or other sources of biomedical science.

(Atkinson 1995: 137–139)

Caught between personal, experiential accounts of medical phenomena and formal, impersonal accounts of medical knowledge, anecdotes occupy a contested space. On the one hand, they offer experience-based insights into healthcare phenomena, on the other, insights based on haphazardly collected observations. Within this space, anecdotes are viewed as anachronistic reports, ‘the enemy of objective, dispassionate observation […] riddled with bias, faulty memory and “foolish optimism”’ (Campo 2006).

Despite such sampling problems and the lowly position anecdotes occupy in hierarchies of evidential credibility (Murad et al. 2016: 125–126), they continue to be published in medical journals. Since the British Medical Journal (BMJ) initiated its Minerva column in the 1880s and its subsequent ‘Literary Notes’ section a few decades later, it has carried snippets of anecdotal information, anonymous experienced-based testimonies, mirrored by The Lancet’s ‘In England Now’ column, and by its more recent ‘Uses of Error’ section, which featured personal accounts of medical mistakes.

Anecdotes remain prominent features of contemporary medicine’s written and oral culture, a circumstance that led the literary scholar Kathryn Montgomery Hunter to insist that ‘[s]omething so pervasive and so contrary to medicine’s scientific ideal […] must have a function in the everyday business of medicine’ (Hunter 1991: 70). Within the schema of a ‘nutshell narrative’ (Morgan, Chapter 1), these functions include delineating modes of noticing that effect epistemic switches that prime audiences to adjust their prior understandings of situations in the light of experience.

17.3 Colligating Medical Anecdotes

Anecdotes played a distinctive role in William Withering’s discovery of the therapeutic value of the foxglove in the treatment of dropsy, a condition that caused bodily swelling. At the head of his Account of the Foxglove (1785), Withering featured several anecdotes, foremost of which was the following:

In the year 1775, my opinion was asked concerning a family receipt for the cure of the dropsy. I was told that it had long been kept a secret by an old woman in Shropshire, who had sometimes made cures after the more regular practitioners had failed. I was informed also, that the effects produced were violent vomiting and purging; for the diuretic effects seemed to have been overlooked. This medicine was composed of twenty or more different herbs; but it was not very difficult for one conversant in these subjects, to perceive, that the active herb could be no other than the Foxglove.

(Withering 1785: 2)

With this anecdote Withering simultaneously proclaimed and reframed the therapeutic value of the remedy. The account depicted more than a sudden apprehension of the active ingredient of ‘a family receipt’, a recipe handed down from one generation to another; it also relayed his realization that diuresis accounted for its efficacy, not the ‘violent vomiting and purging’ the remedy also provoked, which had long been accepted as beneficial for dropsy. Withering bolstered this aperçu with further anecdotes: he ‘knew of a woman in the neighbourhood of Warwick’ and another ‘in Yorkshire’ who possessed a similar ‘receipt’, and relayed a ‘circumstance’ told to him by his ‘truly valuable and respectable friend, Dr. Ash’ that the Principal of Brasenose College, Oxford, had been cured of dropsy ‘by an empirical exhibition of the root of the Foxglove, after some of the first physicians of the age had declared they could do no more for him’ (Withering 1785: 3).

The source of these causal claims was lay experience,⁷ which proved rich in anecdotal remedies for dropsy in accounts that were not themselves case reports but testimonies which emerged ‘from the empirical usages and experience of the populace’ (Withering 1785: 1). At a time when such anecdotes featured in advertisements for quack remedies and false nostrums, Withering was well aware of the precarity of their credibility: ‘There are men who will hardly admit of any thing which an author advances in support of a favorite medicine, and I allow they may have some cause for their hesitation’ (Withering 1785: viii). Contrary to advertising testimonies – which were countered as ‘cases of cures never performed, and copies of affidavits never sworn to’ (Adair 1790: 256)⁸ – Withering drew attention to people harmed by the remedy in an anecdote that could not be so easily questioned or dismissed:

I recollect about two years ago being desired to visit a travelling Yorkshire tradesman. I found him incessantly vomiting, his vision indistinct, his pulse forty in a minute. Upon enquiry it came out, that his wife had stewed a large handful of green Foxglove leaves in half a pint of water, and given him the liquor, which he drank at one draught, in order to cure him of an asthmatic affection. This good woman knew the medicine of her country, but not the dose of it, for her husband narrowly escaped with his life.

(Withering 1785: 9–10)

Withering’s first anecdote contained more than a simple flash of inspiration: in positing the remedy’s active ingredient and mechanism of action, he narrowed the explanandum and refined the type of explanans that accounted for its benefits, which no longer centred on purgings. The change fuelled the need to develop a different treatment regimen, based on careful dose adjustment, which Withering set out in a hierarchy of some 150 case histories that precisely reflected his capacity to vouch for their details.⁹ Whereas the anecdote that announced the discovery invoked observations by others, Withering’s cases focused on his own observations, and how he maximized the remedy’s benefits and minimized its harms. By standardizing the use of extracts and adjusting doses, he set out the candidate ‘causal relations’ and ‘manipulable facts’ that Rachel Ankeny identifies as some of the key contributions cases make to medical knowledge (Ankeny 2014: 1009).

Eighteenth-century case reports invoked intricate descriptions taken to reflect their authors’ powers of perception, which helped underpin the verisimilitude claimed for their accounts (Da Costa 2002). Hess and Mendelsohn note that cases increasingly were treated as collections of observable data linked to specific individuals, based on information excerpted from their histories (Hess and Mendelsohn 2010; 2014). However, unlike cases, anecdotes remained as solitary micro-narratives which arose from fleeting, unformalized aspects of medical practice, demonstrated by the anecdote published in the BMJ in 1881, which turned on the perils of making inferences across cases:

The following tale has lately been reported. […] An epidemic of typhoid fever broke out in a small village in the South of France. A locksmith fell ill, and called in the local medical man, who came, prescribed, and went away. The next day, during his usual rounds, he called at the locksmith’s, and asked his wife after the health of the interesting patient. She replied: ‘Ah, sir, only imagine, whilst I went to fetch the medicine, my husband ate two pickled herrings and a dish of bean salad.’ ‘Good heavens! Then he is … .’ ‘Quite well, doctor. He went to work this morning as usual, and is as well as possible.’ ‘That is extraordinary’, exclaimed the doctor; ‘what a wonderful remedy for typhoid; I must make a note of it.’ And he accordingly entered in his note-book: ‘Typhoid fever: tried remedy, two pickled herrings and bean salad.’ Two days afterwards, a bricklayer was attacked by the same disorder. ‘Take’, said the same doctor who was consulted, ‘two pickled herrings and a dish of bean salad. I will come again to-morrow.’ To-morrow-alas! the bricklayer was dead. The doctor, taking a logical view of his experimental method, again entered in the famous note-book: ‘Typhoid fever. Remedy: pickled herrings and bean salad. Good for locksmiths, bad for bricklayers.’

(Anon. 1881: 248)

To attribute a difference in outcome to a difference in occupation, the same treatment given to the locksmith would have to have been taken by the bricklayer. But whatever the medical man prescribed for the locksmith, assuming he took it, which is not certain, the doctor assumed it played no role in his unexpected recovery. In his surprise at the locksmith’s recovery, the doctor purloined his contingent dietary likes as treatment for the next case of typhoid. As a burlesque on clinical reasoning and the doctor’s ‘experimental method’, the account relayed a scenario built on the perils of inferring causation within and across individual cases. In striving for, and falling comically short of, a generalizable possibility (Hurwitz 2017), the anecdote parodies how easily certainty and uncertainty can be elided concerning the cause of a clinical outcome. It also prefigures how unreliable patient and practitioner testimonies would come to be considered in the following century.

17.4 Quirky, Anecdotal Testimony

In the second half of the twentieth century, testimony rooted in patient and practitioner experiences came to be viewed as profoundly unreliable. The contemporary clinical researchers Murray Enkin and Alejandro Jadad defined the anecdotal as ‘any type of information informally gained, either from personal or clinical experience, one’s own or that of others, in contradistinction to evidence generated by formal research studies’ (Enkin and Jadad 1998: 963). The physician Mark Crislip put the position succinctly:

Anecdotes are how patients transmit the particulars of their disease to their health care providers. The medical history, as taken from the patient, is an extended anecdote, from which the particulars of the disease have to be extracted.

(Crislip 2008)

On this account, patients’ concerns and fears – the whole symptomatic realm – are viewed as anecdotes and placed at the very centre of practice. Steven Novella, an evidence-based neurologist, classed all ‘uncontrolled subjective observations’ as anecdotal, although in the context of discovery he argued that they may still be useful:

Many medical discoveries started as anecdotal observations. But then those observations have to be tested with controlled observations or experiments – and most anecdotal observations will turn out to be wrong or misleading, because they are quirky and uncontrolled.

(Novella 2010)

Yet the editors of The Lancet singled out anecdotes as the lingua franca of clinical exchange, information and learning, claiming that:

Clinicians learn from anecdotes – stories they heard at medical school, stories they tell each other, and stories their patients tell them. This is an efficient way to grasp new knowledge – even the most obscure hints and warnings can be made memorable if tagged to real people and actual events.

(Bignall and Horton 1995)

Less formalized than case reports, although consonant with Withering’s stance, the editors treat anecdotes as potential sources of knowledge ‘if tagged to real people and actual events’. Without some reliance on anecdotal testimony it is difficult to see how therapeutic substances could ever have been developed. Jeff Aronson, a clinical pharmacologist, observes how commonly contemporary research and development of new drugs ‘start with an anecdotal report of some sort’ (Aronson 2008: xxxi), which provokes further tests of efficacy and toxicity in animal and human tissues, Phase 1 trials, then trials in larger, more representative human populations.

In 1992, fragmentary, unshaped anecdotal accounts concerning a quirky effect of a drug prompted the pharmaceutical company Pfizer to upturn the rationale of its testing programme on sildenafil citrate, a chemical it had synthesized in the 1980s as a possible treatment for high blood pressure and angina. Six years later, the company began marketing sildenafil under the brand name Viagra (Tozzi and Hopkins 2017; Dunzendorfer 2004).

Sildenafil was the product of a ‘rational drug design program’ (Terrett et al 1996: 1819), targeted at inhibiting a family of enzymes believed to play a part in increasing thrombosis and vascular resistance. Synthesized in Pfizer’s UK Discovery Chemistry and Biology Laboratories, the drug reached Phase 1 trials in 1992 in eight healthy South Wales miners. Twenty-five years later, David Brown, a Pfizer chemist, recalled the response to a routine question at the end of the study to the whole group of volunteers: ‘Is there anything else you noticed you want to report?’

One of the men put up his hand and said, ‘Well, I seemed to have more erections during the night than normal,’ and all the others kind of smiled and said, ‘So did we.’

(Tozzi and Hopkins 2017)

Although Brown’s recollection reconstructed the occasion as a moment of revelation, Ian Osterloh, who was in charge of research in Pfizer’s laboratories, reported that: ‘At the time, no one really thought, “This is fantastic, this is great news, we’re really onto something here. We must switch the direction of this program”’. Nick Terrett, another Pfizer researcher, recalled that when asked about side effects of the drug, volunteers said: ‘“Yeah, I’ve got a headache. I feel a bit dizzy” and some added “I’ve got erections”’, but these did not register as more significant than other side effects, such as backache, throbbing and stomach upsets (Friend 2017: 480–481). Osterloh recounted Pfizer’s initial response:

None of us at Pfizer thought much of this side effect at the time. I remember thinking that even if it did work, who would want to take a drug on a Wednesday to get an erection on a Saturday? So we pushed on with the angina studies.

(Osterloh 2015)

But when intravenous sildenafil showed little cardiovascular effect in volunteers, hopes of a tangible anti-anginal benefit faded and the three-year development programme faced closure (Jackson et al. 1999). Notwithstanding the costs already incurred, reports voiced as ‘I seemed to have more erections’, which were not ‘thought much of’ by Pfizer, and likely biases from a combination of embarrassment, self-censorship and embellishment in testimonies that could have become ‘improved in the telling’ (Gross 2006: xii) – ‘all the others kind of smiled’ (Tozzi and Hopkins 2017) – the company credited the miners’ reports with sufficient warrant to commit additional funding to further investigate sildenafil.

Testimony takes place within a context framed by a variety of practices and institutions that affect both its content and the level of warrant it purports to deliver. To interpret and assess it requires sensitivity to such contextual factors. Rather than concluding that testimony is in general warranted or that only the testimony of informants who are known to be reliable is warranted, we assess testimony in light of […] not just who is talking and what she is saying, but also what is at issue, what is being assumed about the facts, the circumstances, the testifier, the audience, and the cognitive context. […] A testifier can transmit no more warrant than she has. But her audience may have epistemic resources that she lacks […] [which] is why testimony can be a vehicle for the advancement, not just the dissemination, of understanding.

(Elgin 2002: 307)

To consider the anecdotal as solely comprised of claims is to miss the cognitive and other contexts in which anecdotes arise, are received, and made sense of. To the anecdotes he heard in the 1770s, Withering deployed highly developed epistemic resources not only for the historical and biographical purposes of announcing how he happened upon the discovery, but in granting anecdotes warrant for the remedy’s therapeutic and harmful effects. In respect of the bicycle spoke in the lung anecdote, the information the patient provided proved decisive evidence against the likelihood of another faulty X-ray, which threw doubt on the reliability of the diagnosis. In regard to the sildenafil testimonies, the epistemic resources of Pfizer’s researchers appeared initially under-developed;¹⁰ it was not that they doubted the truth of the miners’ reports, but the erectile effect was unexpected and their testimonies stood outside the anti-anginal framework in which the drug had been developed. At a time when there were no oral therapies for impotence and the biochemistry of tumescence was only partially understood, the testimonies lacked a framework in which to articulate the effect, which was medically and culturally unsituated (Giere 2006; Massimi 2018).

Consistent with Fleming’s view that anecdotes can only be collected and ‘not sewn together into a single story’ (Fleming 2011: 74), Withering attempted no summation of his anecdotes. Although they carried evidential warrant, each stood alone in his Account. Pfizer attempted no summation of the anecdotal testimonies of the Welsh miners as their reports at best could only have been counted as a ‘pre-cursor to [scientific] evidence’ (DeWald 2013). Nevertheless, they proved sufficient to persuade Pfizer to turn around its research programme, a switch that would garner the company $30 billion of revenue from sildenafil sales in the first two decades of the twenty-first century (Statista 2020).

Pfizer attempted no summation of the testimony of the volunteers who subsequently joined its sildenafil trials: instead, a different type of evidence was developed that reflected the experiences of men on and off the drug. This evidence pertained to several components of the erectile process – speed of onset, hardness and duration, capacity to penetrate – gauged from answers to questionnaires and records of measuring devices in studies that recruited over 4,000 men worldwide (Eardley et al. 2002). Key to these studies was the development of an International Index of Erectile Function, a self-completable measure that took little account of emotional, inter-relational, non-erectile aspects of sex (Burnett 2020), but proved sensitive and specific enough to detect treatment- and dose-related erectile effects within and between drug and placebo (Rosen et al. 1997; Goldstein et al. 1998).

Once the biochemical and physiological mechanisms underlying tumescence had been fully elucidated, the effects of sildenafil could be fitted into a causal account of human erection (Baier 2019) and the voices of the miners could be overlaid by a set of multidimensional scores of thousands of male heterosexual performances, on and off the drug. In place of their anecdotal testimonies stood a matrix of standardized, combinable, self-reported scores, averaged for separable components of the erectile process, partitioned by age, dosage and timing, severity and cause of erectile dysfunction.

According to the philosopher Elizabeth Fricker, the paradigm of testimony is a face-to-face encounter in which expression of a knowledgeable belief is vouched for by a combination of the speaker’s trustworthiness and ‘choice of words’ (Fricker 2006: 594), a model that fits the testimony of Withering’s informants, the bicycle patient and Welsh miners. However, as we have seen, the miners’ natural language testimonies came to be revoiced in the ‘de-anecdotalized’, formal language of Pfizer’s subsequent trial participants. In place of their ‘choice of words’ – which was never elicited – were scored responses to the closed questions of impotence researchers, expressed in data points derived from Lickert scales and channelled into a quantitative construct of erectile effectiveness.¹¹ These standardized and manipulable scores spoke the quantized language of mean differences in effect size between drug and placebo, and took the place of the anecdotal accounts of the miners.

17.5 Conclusions

In examining how anecdotes handle medical materials, size up situations and gain attention, the aim has not been to advocate a presumptive right for their claims to be believed. It has been to delineate the epistemic shifts interwoven in their narratives, which confer an immediacy of engagement with what is recounted.¹² In the Greco-Roman period, anecdotes provided a view of political power profoundly at odds with authorized accounts (Fenves 2001: 153); in the Renaissance, they became ‘the principal register of the unexpected and hence of the encounter with difference’ (Greenblatt 1991: 2–3), and in questioning the social order they recruited ‘the unreliable, eccentric, and the improper’ (Patterson 1997: 160). In the eighteenth century, their ‘fragmentary, eye-witness character’ fitted them to a ‘variety of verbal practices, both oral and written, both popular and cultivated: the joke or the tall story; the jewel-like short narrative, with its witty punchline […] usually containing a moral lesson’ (Gossman 2003: 149–150).

Such tropes are evident in the dramatic and humorous nature of anecdotes in the modern era, which may overshadow their epistemic coordinates. Michael O’Donnell valued anecdotes for their quick-witted formulation of ‘uncertainties, paradoxes, life-affirming surprises and black comedy’ (O’Donnell 2013: i). Within this heterogeneity, a variety of anecdote can be delineated, which draws together observations and dialogues from scattered zones of healthcare experiences, in a cognitively and affectively rousing narrative. It opens on a situation that quickly faces challenge from a new vantage point and perspective, which introduce other views that make different sense of the situation, and lead to a degree of resolution.

Within this minimalist format, altered appearances and understandings based on first-person perspectives immediately create unanticipated effects that can severely test epistemic resources. To make sense of the mental processes believed to govern the ability to place the parts of a familiar situation in context, the psychologist Francesca Happé argued that ‘the best way to convey what is meant by the tentative and exploratory notion of central coherence is to give an anecdote’:

A clinician testing a bright autistic boy presented him with a toy bed, and asked the child to name the parts. The child correctly labelled the bed, mattress, and quilt. The clinician then pointed to the pillow and asked, ‘And what is this?’. The boy replied, ‘It’s a piece of ravioli’.

(Happé 1995: 173–174)

Happé emphasized that the boy was neither visually impaired nor joking and that his ‘clinician commented that the pillow did indeed look like a piece of ravioli’ (Happé 1991: 174). His uncanny response challenged researchers to explain how a cognitive process that correctly identified some of the accoutrements of a toy bed discounted dissimilarities in texture, function and socio-cultural context of a key component of the ensemble, allowing its place to be supplanted by a piece of pasta. The anecdote made the explanatory potential of weak central coherence the vantage point from which the boy’s response became explicable.

Walter Benjamin argued that anecdotes develop a ‘pathos of nearness’, which brings distant happenings spatiotemporally closer to us (Benjamin 1999: S1a, 3).¹³ It manifests by making apparent and proximate contours, patterns, frameworks and views of situations hitherto outside of the field of vision. Bernard Williams argued that an ‘epistemic division of labour’ (Williams 2004: 43) underpins viewpoint and perspective in the role observational position plays in the look of things. Although most of the anecdotes we have examined are not first-order accounts of sensory-based observations, but retellings of the noticings of others, Williams’s topological argument illuminates how observations made from different angles, conceptual, figurative, first-person vantage points and hypothetical positions, create different views of their objects of interest.¹⁴

In the context of a broadly observational and inferential practice such as medicine, anecdotes are the culturally mediated schemata that alert audiences to the importance of epistemic vigilance, by demonstrating adjustment of prior understanding of a situation in the light of vantage points and experiences emergent from them (Herman 1999: 25).¹⁵ By bringing observational and conceptual frameworks into close relation with each other, anecdotes bring to the fore scenarios viewed in a new light before their audiences. Those we have encountered include a pattern in bedside greetings; a radical change in diagnosis and treatment; a switch of diagnostic mindset; a change in the mode of action of a remedy; and the intimation that a tentatively voiced side effect pointed to an unrecognized therapeutic effect.

Kathryn Hunter argues that ‘medicine both scorns the anecdotal and provides for the careful reporting of single instances’ (Hunter 1991: 118). Unlike case reports, which have become highly regulated medical accounts, anecdotes remain informally patrolled schema, cast in a vernacular language that has less recourse to technical and formal terminology than cases. Where anecdotes set out their own observations and descriptions, they do so in a register more humanly voiced in the idioms of conversation and hearsay than that of cases. Epistemic switches, comprehensibility and human-centred focus are interwoven features of anecdotal narratives and contribute to their pungency. Despite the increasingly formalized nature of bioscientific discourse, anecdotes retain a continuing presence in medical culture. They demonstrate how suffused with vantage point and perspective medical understanding is and how dependent medical knowledge remains on the speakers and voices heard in healthcare.¹⁶

18 Narrative Performance and the ‘Taboo on Causal Inference’: A Case Study of Conceptual Remodelling and Implicit Causation

Elspeth Jajdelska

18.1 Introduction

Research on live storytelling illustrates how a performative storyteller holds the floor, as the audience lends them the authority to control who speaks, on which topics, when and for how long. Storytellers are also granted some authority over the organization of space, movement within it and acceptable behaviours. It is this wholeness of the social event that supports a shared attribution of underlying causes linking one event in the narrative sequence to the next (Lwin 2010; Goffman 1974). This willingness to cede, temporarily, authority to a storyteller leaves traces in written narrative, for example in readers’ willingness to trust narrative voices (on general ‘truth bias’, see Gilbert, Tafarodi and Maone 1993; in narrative in particular, see Yanal 1999). In what follows, I suggest that experience of narrative performativity in a science paper can influence readers’ trust not just in the authors’ factual authenticity (e.g., is the data accurate?) but in the more contingent interpretation of facts. I view narration as entwined with performance, such that even non-narrative episodes of performative speech in an event that is narrative overall contribute to the inference of causes. In this light, I show how causal modelling is affected by local instances in science writing of both narrative and non-narrative performativity, arguing that together they incite the reader to construct an implicit and new narrative model relating a set of cognitive concepts.¹

I use ‘model’ here in the sense developed by philosophers, psychologists and computer scientists in reference to predictive theories of mind organized around Bayesian probability. Here, perception is a continual revision of predictions through error correction, a dynamic network of shifting probabilities giving rise to implicit, and continually changing, causal relationships through a process of reconceptualization (Clark 2013).

‘Reconceptualization’ is used in the sense employed by Churchland and Boden in their accounts of creativity in the context of connectionist, and more recently neural network, approaches. In this light, concepts are not the brittle units of modular approaches to mind (Fodor), but ‘flexible, distributed representations’ (Kiefer and Pulvermüller 2012: 805). As a result they can be reconfigured by experience, or by repeated iterations of thought, as in Churchland’s example of Newton reconceptualizing the moon, from something like a ball bearing in a track to something like a projectile that has been hurled. Reconceptualization in this sense involves reconfiguring a concept’s causal profile and its relationship to the rest of the mind’s world model.

Explicit and implicit mental causal modelling of this kind is found across scientific disciplines, and the process of reconceptualization which can be involved in this, and particularly in moments of scientific ‘breakthrough’, has been identified as a key creative component of scientific thinking. Examples include the identification of the Benzene ring, or the theory of gravity, mentioned above (Boden 2004; Churchland 2012). I have argued elsewhere that narrative fiction engages readers in a comparable process of creative reconceptualization (Jajdelska 2019). At the same time, causal modelling is foundational to the comprehension and experience of narrative performance, whether live or mediated by silent reading of narrative texts. Anthropologists, for example, have shown how causal links between story episodes are largely a matter of audience/reader inference, in response to the experience of the text as a whole. As a result, a single event sequence can be associated in different performances with radically different underlying causal models (Bauman 1986; Trompf, Gough and Eckhart 1988). Narrative performance, then, has the potential to generate creative causal inference in support of scientific remodelling.

The analysis below illustrates this idea, using a theory paper in scientific psychology by two cognitive neuroscientists. Scientific psychology is a good place to start this investigation, as it has a particularly anxious relationship with causal attribution. In part, this relates to its origins, which cohered in the twentieth century around research methods ‘institutionalised across the varied communities of experimental, animal, educational, social, clinical, and applied psychologists’ (Flis 2018). But it also relates to the dual status of its practitioners as simultaneously agents (as researchers) and objects (as minds) of study (Smith 2007). Smyth, for example, has shown that, compared to other scientific disciplines, scientific psychology ‘behaves in textbooks as if there are grounds for not trusting its statements’ (Smyth 2001: 392). The current replication crisis in the discipline (Wingen, Berkessel and Englich 2020) is arguably symptomatic of these problems, as is an intense focus on methods rather than ‘ways for initial selection and identification of relevant phenomena’, which are ‘in comparison [to methods] underdeveloped’ (Flis 2019: 167) and a ‘seriously limited’ ‘conceptual analysis of psychological phenomena’ (Flis 2018: 160).

As a result, the discipline suffers from what Grosz and colleagues identify as ‘a taboo’ against what might be considered a ‘central goal of research’, that is aiming for ‘explicit causal inference’ to describe the world (Grosz, Rohrer and Thoemmes 2020). This fear of explicit causal inference can explain the ‘encyclopedic incrementalism’ of writing in the discipline, adding more and more discrete pieces of knowledge to the field. The promise that this could ‘finish’ psychology – produce comprehensive knowledge or at least ‘organized theory’ is one it ‘just could not deliver’ (Flis 2018: 31; Bazerman 1987; 1988). If the genre of the ‘APA [American Psychological Association] article’ has ended up in a cul-de-sac, the Barrett and Bar article I consider here is an interesting example of experimentation with a different genre, one that alternates with the APA conventions described by Flis, drawing on Bazerman. I argue that this paper’s elements of narrative performativity, and of performative language more generally in support of narrative inference, enable comprehension of the underlying causal claim of the paper: that emotion is a cause of perception rather than just an effect.

My analysis is informed by the work of pioneering folklorist and anthropologist of verbal art, Richard Bauman (1975; 1986; Bauman and Briggs 1990). For Bauman, performativity is a measure of how far a speaker and audience understand their relationship to be one of evaluation of the speaker by the audience, not by reference to content, accuracy or informativeness, for example, but in relation to the speaker’s ‘communicative competence’ (Bauman and Briggs 1990: 66). In this respect, performativity is embedded in the social event, and is also scalar: speech can be more or less performative, not just either/or. Texts emerging through performative speech are likely to be characterized by greater or lesser degrees of poetic patterning and meta-textual features (features that draw attention to the artful status of the text itself). In the article analysed here, I suggest that both narrative and non-narrative performative sections are quite starkly distinguished from those that are non-performative, but that the former nonetheless play a vital role in explicating the non-performative sections by supporting an implicit narrative involving causal relations.

I chose the paper by Barrett and Bar because I have learned from it, and cited it in my own work. If the analysis appears critical at any point this is in the context of a critique of a scholarly practice rather than a critique of the authors. It is also a comparatively rare, although far from unique, example of this approach of blending the performative with the APA genre.² This technique also suggests unexpected overlaps between scientific psychology and those social scientists who successfully put ‘elements into relation to each other when they appear in opposition’ in publications, such that social scientists have ‘good reason to think that they should fit together – in some way or other – rendering disparate and even oppositional matters into a narrative explanation’ (Morgan 2017: 90). In this case it is not just content that is juxtaposed in productively puzzling ways, but modes of performance, or genre.

18.2 Analysis

‘See It with Feeling: Affective Predictions during Object Perception’, by psychologist Lisa Feldman Barrett and neuroscientist Moshe Bar (Barrett and Bar 2009), reviews existing findings in the fields of neuroscience and cognitive psychology to ‘develop the hypothesis that the brain’s ability to see in the present incorporates a representation of the affective impact of those visual sensations in the past’. In some ways, the combination of title and subtitle here captures precisely the sense of a ‘taboo’ on causal inference identified above by Grosz et al. The title conveys a relationship between emotion and perception of an environmental stimulus, in which each mutually influences the other; if we ‘see with feeling’ then it is not obvious that we could see without feeling. The causal relationship between the two in this scenario therefore departs from earlier theories in both behaviourist and later cognitive traditions, which ‘assumed that affect occurred after object perception and in reaction to it’ (Barrett and Bar 2009: 1328, citing Arnold 1960). One way to figure this is as two titles for two distinctive narrations: the first more performative, the second less so. This division between two genres potentially allows for one to do the work of the reconceptualization, and the other to supply the evidence.

The authors’ moves into and out of the stance of a narrative performer throughout the paper enable a relationship with the reader which can support creative reconceptualization of emotion and perception, and thereby allow the emergence of the more global theories which Flis identifies as missing from the project of psychology writing in the APA model. It should be noted that ‘narration’ here refers to the stance of performing a narrative, although the content of the paper itself does not provide the event sequence that can give the article as a whole the status of narrative. However, this performative stance of narration, supplemented by non-narrative moments of verbal performance, is, I suggest, crucial to the implicit narrative of temporally organized causal relations between emotion and perception which the article points towards.³

The authors’ abstract in full will help with my further analysis below:

See it with feeling: affective predictions during object perception

People see with feeling. We ‘gaze’, ‘behold’, stare’, ‘gape’ and ‘glare’. In this paper, we develop the hypothesis that the brain’s ability to see in the present incorporates a representation of the affective impact of those visual sensations in the past. This representation makes up part of the brain’s prediction of what the visual sensations stand for in the present, including how to act on them in the near future. The affective prediction hypothesis implies that responses signalling an object’s salience, relevance or value do not occur as a separate step after the object is identified. Instead, affective responses support vision from the very moment that visual stimulation begins.

(Barrett and Bar 2009: 1325)

18.2.1 One Title, Two Modes

Returning to the title, it captures the contribution of local performative elements that are not explicitly narrative to the implicit narrative model developed by readers throughout the article: ‘See it with feeling: affective predictions during object perception’ (Barrett and Bar 2009: 1325). The first part of the sentence uses wordplay (‘say/see’ recalling ‘say it with feeling’, a stereotypical instruction to, for example, drama students), recalling Bauman’s characterization of performance as drawing ‘special attention to and heightened awareness of the act of expression’ (Bauman 1975: 293). It also introduces the counter-intuitive idea that there is no seeing, in the non-metaphorical sense of visual perception, without feelings. We may believe that sometimes we simply see a scene or object in a disengaged or neutral way, but that, it is suggested, is not in fact the case. The performative section of the title, therefore, is also introducing the article’s most counter-intuitive idea, one, moreover, which commits the authors to a broader theory of emotion and cognition. Performative text – that is text subject to evaluation for competence in delivery, rather than content or information – here enables a bolder claim than might otherwise seem proper to the APA article genre. Performativity may act here, then, as a creative training ground for what may be a new idea: that emotion contributes to perception rather than following it. In this respect, while the performative passages in the paper, as we will see, move in and out of narration, they cumulatively support a cohesive implicit narrative process in the reader of causal attributions in a model of mind and perception.

A different, less performative, relationship with the reader is constituted immediately after the colon: ‘affective predictions during object perception’. ‘Feeling’ and ‘affect’ both have specialized meanings in cognitive science. ‘Affect’ generally refers to ‘a state characterized by emotional feeling rather than rational thinking’, a state which generally involves ‘arousal’, or ‘corresponding bodily reaction (but not necessarily action)’. ‘Feeling’ can be used in neuroscience as a synonym for ‘emotion’, but it is often treated as ‘one component’ of emotion, ‘the proprioceptive representation’ of emotional ‘bodily changes’ (Sander and Scherer 2009). ‘Feeling’ however, unlike ‘affect’, also has a role in non-expert discourse, as a synonym for ‘emotion’, but also as a conceptual metaphor. Conceptual, or basic, metaphors are metaphors whose status we barely notice, as they are essential to everyday discourse. They take a range of verbal forms, cross languages and cultures, and emerge from fundamental aspects of embodied experience. HAPPINESS IS UP, for example, can be seen in expressions like ‘cheer up’, and relates, it is argued, to the upright, bipedal status of the human body. In the case of feeling, the everyday, non-specialist equation of feeling and emotion maps the experience of physical pain or pleasure onto mental or emotional pain or pleasure, through the basic metaphor EMOTIONAL EFFECT IS PHYSICAL CONTACT (Lakoff and Johnson 1980: 18; 50).

The first, more performative, part of the title, then, introduces readers to a new and surprising idea, but via appeal to a fundamental and familiar one. In using a term like ‘feeling’, moreover, with a place (although not quite the same place) in both expert and non-expert discourse, the abrupt disjunction between the more performative first half and the less performative second half is ameliorated. The second half is marked by a switch from the Germanic vocabulary associated in English with the everyday to the Latinate vocabulary associated with formal and abstract speech (Bar-Ilan and Berman 2007). This formal/informal contrast does not tell us how performative a stretch of discourse might be; both can be used for specific effects in both more and less performative contexts. However, formal speech does imply greater social distance between speaker and hearer (Brown and Levinson 1987). The second part of the sentence is thereby abruptly distinguished again from the social event of narrative performance implied by the first part.

This is not just a case of restating the case in discourse more suited to the evaluation of accuracy rather than to the performative criterion of competence in delivery. The case itself is somewhat altered.⁴ As discussed above, ‘affect’, unlike ‘feeling’, is specialized to psychology and neuroscience, distinct from the wider range of concepts connected by everyday uses of ‘feeling’. ‘Perception’, on the other hand, is broader in some ways than ‘seeing’, covering all modalities, not just vision, as well as processes to which we do not have conscious access. But in other ways ‘perception’ is narrower, generally excluding the easy slippage in non-expert discourse between literal and metaphorical senses of ‘see’ (such as ‘I see what you mean’). It is explicitly applied here only to ‘objects’, not, for example, to scenes. As an appendix to the imperative main clause before the colon, no verb is required, which allows this noun phrase to evade any time-locked claims. These would specify a starting point for the prediction and a causal relationship between prediction and perception, such as ‘affective predictions occur during object perception’ (my italics).

The second part of the title, then, both makes a different claim and establishes a different relationship with readers. The claim is hedged in the way Smyth characterizes as typical of scientific psychology papers, whereas the claim in the performative first half is not. Because it is less performative, it is also more accountable to its audience for accuracy. It is more formal (which is often, although not always, associated with an expectation of objectively described, accurate information). It claims that one aspect of ‘feeling’ as understood in non-expert discourse, that is ‘affect’, contributes to predictions. These predictions are described as happening during object perception; it is implicit that the process of prediction is in fact equivalent to the process of perception, with prediction/perception ending at the point where prediction error from the environment is minimal enough to be ignored (Clark 2013).

This pattern of an easy yet abrupt shift between performative (either narrative or with narrative implications) and informative relationships with the reader persists throughout the article. This might be seen as a symptom of the dual nature of science as a social practice and science as a methodology, which seeks to overcome the distortions of personal subjectivity by establishing objective facts. The oscillation between kinds of discourse is a way of keeping both aspects in play.⁵ The readers of a scientific article are both part of a social community of practice, which can bond through verbal performance, including jokes (Reimegård 2014), and committed to looking at the world, as best they can, as though from an extra-human perspective (Reiss and Sprenger 2017).

On this understanding, the reader’s relationships between more and less performative authorial voices capture an abrupt disjunction between their functions. The more performative voice instantiates science as a social practice, while the less performative one enacts science as a method. While the two can sit side by side, it might be assumed they need not, indeed should not, interact. This interpretation relies on a strict separation between the social and methodological aspects of scientific practice, an interpretation with long-standing challenges from history, philosophy and social science (e.g., see Barnes and Bloor 1982; Shapin 1994; and Latour 1987). On these views, scientific sociability and scientific methods are intimately entangled. More recently, some psychologists, confronted with the replication crisis discussed in the introduction, have turned to the social aspect of science to explain some failures in methodology (Ritchie 2020). From this perspective, the coexistence of more and less performative voices might be symptomatic of a problem. In this view, the performative voice instantiates readers as a social group sharing a broad paradigm of cognition. The less performative voice is then interpreted within that broad paradigm, leaving readers less critical of both paradigm and evidence as each is interpreted as potential confirmation for the other.

I suggest an alternative explanation for the function of the performative mode. The performative mode, I argue, does indeed engage readers in building a cognitive theory. But the theory is itself an implicit narrative embedded in the performative text, and emerges, I will suggest, from a process of reconceptualization and remodelling implicit causation in the world. It is this narrative remodelling and reconceptualization which lets them identify a causal mechanism in the data despite the ‘near taboo on causal inference’ identified by Grosz et al., above (Jajdelska 2019). Causal inference in an unfamiliar scientific paradigm, then, can be a creative act. And while performative narration in the article may be dispersed in fragments rather than performed in a sustained social event by, for example, a Homeric rhapsode or a West African griot, it can serve some of the same functions: reshaping our world models in ways that let us see otherwise opaque causal structures. Put differently, performative narration allows for the scrutiny of concepts, causal inference and broader theory building which the genre of scientific psychology writing, as discussed by Bazerman and Flis earlier, inhibits.

18.2.2 Narrative Performance and Reader Trust

The opening paragraph of the ‘Introduction’ (following the abstract) immediately instantiates a voice with a narrator’s authority to hold the floor:

Michael May lost the ability to see when he was 3 years old, after an accident destroyed his left eye and damaged his right cornea. Some 40 years later, Mr May received a corneal transplant that restored his brain’s ability to absorb normal visual input from the world (Fine et al. 2003). With the hardware finally working, Mr May saw only simple movements, colours and shapes rather than, as most people do, a world of faces and objects and scenes. […] As time passed, and Mr May gained experience with the visual world in context, he slowly became fluent in vision.

(Barrett and Bar 2009: 1325)

This is a narrative of the following event sequence in Bauman’s terms: accident; destruction or damage to eyes; transplant; simplified vision; full vision. Like the verbal performers in Bauman’s studies, in performance, the authors, and in turn hearers, attribute causal links to this sequence: the accident caused the destruction and damage; the transplant ‘restored his brain’s ability to absorb normal visual input from the world’; ‘experience with the visual world in context’ caused his later fluency ‘in vision’. These causal attributions may seem uncontroversial but they are not inevitable. The transplant could have been described as a restoration of ‘the ability of the cornea to react to rays of light’, for example, making the story one of simpler physical components of low-level cognition. ‘He slowly became fluent in vision’, could have been, for example, ‘He learned to match visual data to the data higher up in processing streams of other modalities’, again modelling vision as a mechanical process, rather than a complex skill drawing on multiple domains of knowledge and processing, such as ‘fluency’ in a language.

The authors’ narrative phrasing of causal links, then, models a world in which agents interact with their environment by acting on it, and adjusting their expectations through the feedback from their actions. In other words, narrative performance here generates a model which will support their hypothesis: ‘When the brain receives new sensory input from the world in the present, it generates a hypothesis based on what it knows from the past to guide recognition and action in the immediate future’ (Barrett and Bar 2009: 1325). The narrative performance of the story of Mr. May, then, enables readers to model the article’s hypothesis not just by providing a case study, but by modelling a specific causal structure for long enough to establish a hypothesis prior to assessing it.

This short verbal narrative gives way to material which cannot, without effort at least, be characterized as narrative. However, the authority of a narrator may persist through the use of textual features found in verbal art and performance, and it is this performative-authoritative role which supports broader narrative remodelling in these passages throughout the article. For example, the narrative to non-narrative transition sees a continued use of declarative sentences and clauses, a continuity which carries authority in the narrative domain over to the domain of neuroscience:

As time passed, and Mr May gained experience with the visual world in context, he slowly became fluent in vision. […]

What Mr May did not know is that sighted people automatically make the guesses he was forced to make with effort.

(Barrett and Bar 2009: 1325; my emphasis)

It seems unlikely here that the second sentence, which opens the paragraph after the narration ends and which does not form an event in a sequence, leads readers to believe that the authors have interviewed Mr May.⁶ More likely, they accept the claim as a continuation of the authority granted to verbal performers. In both sentences, for example, there are verbal cues to this performative status. Both begin with information available to narrators but not, typically, to readers: that time passed before Mr May experienced change, and what he did or did not know at particular points in the narrative. Compare, for example, these words from novelist Ali Smith’s third-person narrator of Autumn: ‘It is still July’. Readers have neither independent means to establish that it is July, nor any reason to question the claim. Similarly, continuing the performative narrative voice into a non-narrative paragraph might make readers less vigilant about broad, and not always consensus, positions on the nature of vision, and narrower claims about what a specific individual knew about vision, than they would be otherwise.

In the next paragraph, the speaking voice takes a step away from the narration of Michael May’s experience, but maintains the authority associated with it through declarative statements: ‘External sensations do not occur in a vacuum, but in a context of internal sensations from the body that holds the brain’. The claim here is no longer a simple one premised on a narrator’s privileged access to information (‘What did Mr May know about cognition, and when?’). Here, there is an opposition between a model of external sensations ‘in a vacuum’ compared to ‘a context of internal sensations from the body that holds the brain’. Rhetorically, the force of this is to create a binary choice between a model hard even to conceive of (external sensations in a vacuum) and a holistic model intertwining the world (‘external’), the body (‘internal’), and cognition (‘the brain’). This sense of a non-modular, dynamic system is reinforced by the selective use in this paragraph of verbs in the present continuous (-ING):

The sensory–motor patterns being stored for future use include…

In addition to directly experiencing changes in their breathing…

In addition to learning that the sounds of a voice come from moving lips …

(Barrett and Bar 2009: 1325; my emphasis)

The present continuous is associated with a range of rhetorical effects, including the narrative voice of the ‘historic present’ (Wolfson 1982). Here it marks a move away from the narrative of Mr May, but a move which preserves the authority of the performative, narrative voice, just as it preserves and develops the implicit world model established in the narrative section. Stylistic choices between declarative or present continuous sentences, or the contrast between a more and less persuasive model of cognition, can be seen as rhetorical choices. All writers must make choices of this kind, and all of those choices will have some kind of effect on the reader. But, independently of rhetorical persuasion, artful expression and narration create a performative discourse, helping the reader to conceive of the world in a way that lets them understand the causal structure underlying the hypothesis, separately from accepting the hypothesis itself. The effect of performative language, I suggest, can be to support, rather than manipulate, the reader in a creative process of reconceptualization in order to assess the hypothesis by first understanding it. In this respect, then, the elements of narrative performance in the article provide an effective way around the ‘taboo on causal inference’ problem in scientific psychological writing, but one which does not confront that problem directly and therefore might not be sustainable in the field’s discourse as a whole.

18.3 Creativity as Reconceptualization

At the heart of the article is the case for emotion and perception as having an entwined rather than a causally sequential relationship, at least at sub-personal speeds and levels of consciousness. As with modelling the relationship of mind and world as dynamic and holistic rather than linear and modular, this case requires some reconceptualization of emotion. For readers who trained in psychology under the influence of Jerry Fodor’s work, emotion might be not just habitually distinguished from perception, but by definition distinguished from perception. Fodor’s modular account of cognition drew inspiration from Turing’s identification of a minimal mechanism that can do cognitive work automatically (Turing 1936). In a modular system, processing streams for different modalities, and for language, remain distinct until processing is advanced, at which points the information from each module is blended in ‘central processes’. Modularity does not allow the kind of dynamic, holistic interaction of emotion, action and perception that Barrett and Bar outline (Fodor 1983). As Barrett and Bar explain, ‘Affective responses were ignored in cognitive science altogether’ (Barrett and Bar 2009: 1328). To even assess the coherence of their hypothesis, therefore, some readers may need first to reconceptualize (in Boden’s and Churchland’s sense described above) perception, affect and emotion. The performative, and either implicit or explicitly narrative, voice established at key points of the article, I argue, supports this reconceptualizing process in readers, in part through engaging aesthetic emotions.

Take the following extracts from the second and third paragraphs of the article:

This is how people learn that the sounds of a voice come from moving lips on a face, that the red and green blotches on a round Macintosh apple are associated with a certain tartness of taste, and that the snowy white petals of an English rose have a velvety texture and hold a certain fragrance. […] [T]hey learn that they enjoy tartly flavoured red and green (Macintosh) apples or the milder tasting yellow apples (Golden Delicious); and they learn whether or not they prefer the strong fragrance of white English roses over the milder smell of a deep red American Beauty, or even the smell of roses over lilies. (Barrett and Bar 2009: 1325)

The authors explain here that perception is contextual, and that the context includes the perceiver’s early affective responses to the target stimulus. To make this point, they draw on readers’ prior experience not only of apples, roses and lilies, but also on their capacity for aesthetic emotion, a form of emotion likely to be accessible to consciousness, not least because the stimulus itself, that is the artefact (in this case a passage of text), draws attention to itself as an object of conscious attention (Miall and Kuiken 1994; 1999). Aesthetic effects are also associated with performativity (Bauman and Briggs 1990).

The authors make a number of aesthetically directed choices here. First of all, the objects they choose (apples, roses, lilies and their colours, tastes and smells) are all ‘motifs’ which have been identified by folklorists in folk and fairy tales. Motifs migrate between different tale types with a power that is independent of narrative context (Aarne, Thompson and Uther 2004; Thompson 1955–58). This status as standalone objects, divorced in mental imagery from scenes, plays a role in the vividness with which they can be imagined, creating an undiluted focus on aesthetic affordances in relation to taste, sight and smell. The focus on sight in relation to taste and smell also has the potential to heighten vividness through synaesthetic effects (Jajdelska et al. 2010; Scarry 2001; Jajdelska 2019: 570–572).

Aesthetic emotions have been variously categorized (see, for example, Fingerhut and Prinz 2020; Brown and Dissanayake 2009; Miall and Kuiken 1994). A common thread, however, is their association with higher-order processing, that is processing which is more accessible to conscious report (Brown and Dissanayake 2009: 51–52; Fingerhut and Prinz 2020). Aestheticized verbal descriptions may both evoke emotions and make readers aware of the emotions’ source. These descriptions of apples and roses potentially enact in readers the concept of emotion as intrinsic to cognition, and thereby help them to understand the article’s otherwise counter-intuitive hypothesis.

The enactment, insofar as it occurs, would not be evidence in support of the hypothesis, which concerns not ‘emotion’ in general, but affect in particular:

We have proposed that the medial [orbitofrontal cortex] (OFC) participates in an initial phase of affective prediction (‘what is the relevance of this class of objects for me’), whereas the lateral OFC provides more subordinate-level and contextually relevant affective prediction (‘what is the relevance of this particular object in this particular context for me at this particular moment in time’).

(Barrett and Bar 2009: 1331)

Instead, the experience of mental imagery intertwining sensual experience with aesthetic emotion allows those readers for whom emotion and perception are distinct by definition to reconceptualize both and then potentially to evaluate this causal relationship.

18.4 From Reconceptualization to Causation

Reconceptualizing emotion is a first step to successfully identifying and understanding the dynamic network of causal interactions implicit both in the paper’s hypothesis and in the implicit narrative identified by performative discourse throughout. A second step is to develop an understanding of the causal links with greater precision. Here too the move between more and less performative voices may help. The more performative material metaphorically introduces agency into sub-personal processes. The most prominent example of this is the status of the brain, and/or the body as a whole, which appears in the paper’s varied discourses as at some points an objectively viewed system, lacking free will, and at others as an agent pursuing an identifiable set of goals whose parameters are established by evolution. The degree to which people can be understood as free agents is a question involving considerable philosophical effort, informed in recent years by findings in cognitive science (Vierkant 2017). The authors of the paper are not, either implicitly or explicitly, expressing a position in this debate. Instead, I suggest that they are moving between more and less performative discourses in ways that allow readers to imagine and reimagine their hypothesis in different lights, one narrative and performative and one non-narrative. In doing so, readers are supported in modelling causation through narrative discourse and then importing this into non-narrative discourse.

From the perspective of folklore, anthropology and to some extent the psychology of memory, any given narration’s underlying structure is defined not by its causal links but by its sequence of events (Aarne, Thompson and Uther 2004; Bauman 1986; Bartlett 1995). In this view, for any given event sequence underlying a narrative performance, causation is built in only at the point of narration; causation is not an intrinsic quality of the sequence itself, but emerges in the audience and only through performative narration, and a single event sequence can be narrated on different occasions with different causal links. A critical element of this process, building causality into an event sequence through narration and performance, is attributing agency.⁷ In the most influential psychological account of narrative production and comprehension of recent decades, for example, ‘change in character’s goals’, a signature aspect of agency, is one of the small number of key dimensions that hearers/readers monitor consistently in order to comprehend and recall a story (Zwaan and Radvansky 1998; Zacks, Speer and Reynolds 2009). Mythical event sequences, which explain or make sacred features of the world that arose independently of humans, attribute agency either to the rocks, rivers and mountains themselves, or to divine beings who can control them (Doty 2000: 74–76).

Where an event sequence does not feature an easily recognizable agent, then, the narration attributes agency to non-agents, as in this example, continuing from the previous section on roses and apples:

They learn whether or not they prefer the strong fragrance of white English roses over the milder smell of a deep red American Beauty, or even the smell of roses over lilies. When the brain detects visual sensations from the eye in the present moment, and tries to interpret them by generating a prediction about what those visual sensations refer to or stand for in the world, it uses not only previously encountered patterns of sound, touch, smell and tastes, as well as semantic knowledge. It also uses affective representations – prior experiences of how those external sensations have influenced internal sensations from the body. Often these representations reach subjective awareness and are experienced as affective feelings, but they need not.

(Barrett and Bar 2009: 1325)

This passage opens with the synaesthetic, and aesthetic, evocation of experiencing roses and lilies at the level of the person as agent, with the sensual processing of the moment linked to that of the past, and to high-level concepts which can potentially be expressed in language to others (‘I prefer English roses’). Having indicated for the reader the relevant event structure by evoking aesthetic emotions in the reader, the passage then renarrates the story (or story component of the wider, implicit narrative generated by performative discourse across the paper), attributing agency this time to ‘the brain’. The brain as agent has goals and priorities which are distinct from those of the person who weighs English roses against American ones. The goals of the brain as agent include ‘interpreting visual sensations’ and the method includes ‘generating a prediction’ of the world drawing on previous experiences, semantic knowledge and, critically for the paper’s hypothesis, ‘affective representations’ drawn from past experience of the relationship between external and internal sensations, all of which can operate below the level of ‘subjective awareness’.

While the person as agent has a (momentary) goal of ‘establishing preferences’, the brain as agent has the (longer-term) goal of ‘interpreting sensations’. In the second case, as the paper’s author and reader are certainly aware, ‘the brain’s goal’ is a metaphor, mapping something like, ‘has evolved in such a way that responses to past stimuli statistically shape present responses to stimuli’ onto ‘interprets current stimuli in the light of past ones’. One effect of this extended metaphor is, as with the aesthetic effects of apple and rose verbal imagery, to support readers in developing a new model not just of the causal relationships between affect and relation perception, but of the broader network of causal relationships among body, brain and world, a network in which all three are continuously and dynamically interacting in ways that cannot be captured by a unidirectional flow from stimulus to brain to action. The goal of identifying our preferred fruits and flowers, with which we are already familiar from our own conscious experience, can then be mapped onto a less familiar, sub-personal goal of making predictions to support optimal actions, and from there to a modified account of cognition more broadly. Performative discourse, then, even when not explicitly narrative, contributes to global narration across the paper through reconceptualization and its associated causal attributions.

The same movement – between recognizable personal experience and sub-personal cognition – can be seen in reverse in this passage:

With back and forth between visual cortex and the areas of the prefrontal cortex (via the direct projections that connect them), a finer level of categorization is achieved until a precise representation of the object is finally constructed. Like an artist who successively creates a representation of objects by applying smaller and smaller pieces of paint to represent the light of different colours and intensities, the brain gradually adds high spatial frequency information until a specific object is consciously seen.

(Barrett and Bar 2009: 1328)

Here, the authors open without attributing agency to the brain, using the passive voice (‘categorization is achieved’, ‘the object is finally constructed’) to avoid attributions of agency altogether. The comparison to a supposed artist creating an ever more refined representation reframes the process with an agent, and then finally the brain returns, this time as agent, one whose goal is to refine a representation, using different kinds of inputs, including affect, to the point where it enters consciousness. We start with observable data about activity in specific brain regions under specific conditions (provided through citation), and the passive voice mutes any causal claims about the causal relationship between that activity and the subsequent representation. The simile of the artist then creates explicitly a causal relationship, which is parallel, although not identical, to the implicit causal relationship in the observable data: ‘the artist creates a representation’ with the goal of making an object, where, implicitly, the activity of relevant brain regions makes a representation with the goal of supporting appropriate action. Finally the causally neutral data sequence and the causally specific simile are brought together in the brain as agent, whose goal is to make a representation available to consciousness. The shifts between different levels of more and less performative and narrative discourse allow readers not just to identify the causal relationship that the authors embed in their hypothesis, but to model processes of cognition in a way that lets them assess the plausibility of that hypothesis.

18.5 Conclusion

The article discussed in this chapter was not written in response to the current replication crisis in scientific psychology, but it does implicitly address the absence of theory and conceptual work identified by Flis as a problem in the discipline (Flis 2018: 163, and elsewhere). It does so by alternating between the genres of narrative performance and of APA approved explication. The first enables the generation of theory in the reader, and the second articulates this theory in non-performative and non-narrative ways. In this way the paper maintains the division between trustworthy ‘scientific psychology writing’ and potentially less trustworthy engagement of creative imagination in readers, but still manages to look at the key topics in theorized ways. This points to a useful role for mixed discourses in scientific writing more generally. However, whether in psychology or less troubled areas of science, an acknowledgement of the role that performative narrative discourse plays in the work of theory, challenging the sense that this kind of narrative is merely an entertaining experiment with lower status forms of discourse, a holiday from rigour, can contribute to our broader theory-building, reconceptualization and remodelling.⁸

19 Reading Mathematical Proofs as Narratives

Line Edslev Andersen

19.1 Introduction

Mathematicians sometimes emphasize the major role of inductive reasoning in mathematics (see, for example, Borwein and Bailey 2003). Results in mathematics are usually tested in reliable ways before a deductive mathematical proof of the results is produced. For example, the famous Riemann hypothesis has been verified in billions of instances. But results in mathematics are established by deductive mathematical proofs, and the Riemann hypothesis will remain just that, a hypothesis, until a deductive mathematical proof has been produced. Historically, deductive mathematical proof has become the ultimate method of justification in mathematics. For this reason, proofs play a very central role in mathematical research practice.

Mathematical proofs and narratives may seem to be opposites. Indeed, deductive arguments have been highlighted as clear examples of non-narrative sequences by narrative theorists (see, for example, Bruner 1987: 11–14; Herman 2009: 157). By contrast, scholars interested in the nature of mathematical proof have occasionally conceived of mathematical proofs as narratives (Doxiadis 2012; Robinson 1991: 269; Thomas 2007). Their accounts mainly focus on the similarities between mathematical proofs as written and narrative texts.

I similarly compare mathematical proofs with narratives but take a different approach.¹ Narrative texts are read in a quite distinct way, and I argue that mathematical proofs are often read like narrative texts by research mathematicians.² Hence, my account of mathematical proofs and narratives mainly focuses on the relationship between mathematical proofs and their readers rather than their writers.³ The argument constitutes an independent argument for conceiving of mathematical proofs as comparable with narratives, and sheds new light on the reading of mathematical proofs. My argument draws on recent empirical data on how mathematicians read proofs.

Furthermore, my account of mathematical proofs and narratives provides an account of what it means for research mathematicians to understand mathematical proofs. This account of proof understanding appears to have implications for how we should conceive of proof validation and for how proofs should be taught.

19.2 Reading Proof as Narrative

Before I say more about how proofs are read, I should say a few words about how proofs are made and presented. In recent years, philosophers have developed accounts of how we should conceive of a mathematical proof as a sequence of inferential actions performed by an agent on various objects, such as propositions, diagrams and mental images (De Toffoli and Giardino 2015; Larvor 2012; Netz 1999: 51–56; Tanswell 2017a; 2017b: 144–153). For a vivid image of a proof involving inferential actions performed on diagrams, one may watch one of the many videos on YouTube of different diagrammatic proofs of the Pythagorean Theorem (or see Tanswell 2017a: 223–225).

If a proof is a sequence of actions, we may conceive of a proof as written, as found, for example, in a research article or a textbook, as a telling of a sequence of actions. Hence, a written proof is a telling of how something happened. A written proof is a telling of a sequence of actions performed on mathematical objects by an agent with the aim of proving a given proposition (in line with Hamami and Morris 2020). This implies that a proof as written is a narrative in at least a minimal sense. A narrative is often minimally characterized as a telling of a sequence of particular events or actions involving humans or humanlike characters with particular goals (see, for example, Sanford and Emmott 2012: 1–5; Toolan 2001: 4–8). To be more precise, a written proof is a telling by the author of a sequence of actions on mathematical objects performed by an agent, usually by the author of the proof herself, with the aim of establishing a particular theorem (this is in line with Doxiadis 2012: 330–331, on Euclidean proofs).

I ask readers to have something like this picture of proofs in mind as they read on. But rather than focusing on the narrative features of written proofs, the aim of this chapter is to argue that written proofs can be and often are read like narratives by research mathematicians. A narrative is not only characterized by features of presentation. Narratives are also characterized by how they are read or heard. In fact, narratives have a quite distinct relationship with their readers or auditors. This relationship is sometimes described by narrative theorists as having two key features (see, for example, Bruner 1991: 11–13; Herman 1997).

The first key feature of the relationship between narratives and their readers or auditors is that narratives are interpreted by their readers or auditors against the background of patterns of belief and expectation with respect to how events or actions of the kinds represented in the narrative usually take place. The background patterns of belief and expectation are called scripts by theorists like David Herman and they stem from the prior experience of the readers.⁴ Hence, scripts describe standard sequences of events or actions against which a particular narrative is read. The narrative cues readers to activate the scripts.⁵

Consider, for example, the following narrative or part of a narrative: ‘John went to Bill’s birthday party. He watched Bill open his presents. John ate the cake and left’ (adapted from Schank and Abelson 1977: 39). We read this particular narrative against our birthday party script which describes standard sequences of events and actions that usually take place at birthday parties in our experience. About similar examples of narrative sequences, Herman (1997: 1051) writes: ‘I can make an astonishing number of inferences about the situations and participants – fill in the blanks of the stories, so to speak – because the sequences unfold against the backdrop of the familiar birthday-party script’. For example, when we read the narrative about John against our birthday party script, we can fill in sequences of actions such as John congratulating Bill upon arriving at his house, John giving Bill a present, and Bill tearing the wrapping paper before opening the box.

This conception of how narratives are read or heard implies that readers reconstruct a detailed story from a narrative text with less detail. It is important to note that readers reconstruct the actions behind the narrative text with the way events or actions usually occur based on the reader’s prior experience. This is to say that scripts vary across readers whose prior experience is substantially different. Different readers may also make different reconstructions when they have similar scripts available if they make different assumptions about the context of a particular narrative. Thus a reader who assumes John and Bill to be children would fill in a sequence where Bill’s mother lit candles on the cake, Bill blew out the candles, then his mother cut the cake and handed a slice to John. Another reader, assuming Bill to be an older man, might imagine there to be no candles, and Bill cutting the cake himself. An author who wanted to specify who cut the cake would need to write it out explicitly as part of the narrative sequence, not just rely on the birthday party script to do that job.

The process by which readers or auditors activate their prior experience captured in scripts when reading or hearing a narrative is sometimes described as the process by which they come to understand the narrative. I will return to the topic of understanding towards the end of this chapter.

The second key feature of the relationship between narratives and their readers or auditors is that readers or auditors are usually surprised by parts of the narratives.⁶ The surprises are surprises against the backdrop of scripts. They are breaches of the scripts. In other words, the unusual or surprising aspects of narratives are unusual and surprising against the backdrop of the scripts. When the narratives convey something unexpected relative to existing scripts, the narratives can feed into new scripts. Hence, a narrative unfolds against the backdrop of scripts but also contributes to the creation of new scripts. Existing scripts are exploited to generate new scripts.

For example, we may add to the birthday party narrative featuring John and Bill that Bill’s cat tried to lick John’s hand, attracted to the cheese that had oozed out as John bit into the cake. There is a breach of the birthday party script here, as we expect a birthday cake to be made of flour, sugar, eggs and so on, not cheese. The breach of the birthday party script may lead us to reconsider the new trend of using cheese rounds as cakes, and when we on several occasions have read or heard narratives where scripts are breached by involving cheese rounds as cakes, we may be led to new scripts about usual sequences of actions in which a cake made of cheese rounds is served.

The two key features of the relationship between narratives and their readers or auditors are closely related. Herman (1997) describes the relationship thus: ‘Stories stand in a certain relation to what their readers and auditors know, focussing attention on the unusual and remarkable against a backdrop made up of patterns of belief and expectation. Telling narratives is a certain way of reconciling emergent with prior knowledge’ (Herman 1997: 1048). Focusing on social science research, Morgan (2017) similarly emphasizes the importance of narrative in framing and resolving puzzles.

I claim that research mathematicians often read proofs as narratives. Hence, I claim that proofs are often read as narratives in addition to having the presentational features of narratives by being a telling of a sequence of actions. More specifically, proofs are often interpreted by their readers against the background of scripts. The scripts are about how different kinds of results or sub-results are usually proved, about standard sequences of proving actions; and the scripts are based on the experience with proofs of the research mathematicians.⁷ Furthermore, the readers focus attention on the breaches of the scripts, on that which appears unusual or surprising against the background of the scripts.

For example, consider the following telling of a sequence of actions aimed at proving the formula 1 + 3 + 5 + … + (2n − 1) = n², which holds for any natural number n. I begin by showing that the formula holds for n = 1. ((2 × 1) − 1) equals 1, which, in turn, equals 1² and thus the formula holds for n = 1. I now show that if the formula holds for some value n = n₀, then it also holds for n₀ + 1. Hence, I make the assumption that the formula holds for n = n₀, that is, I make the assumption that 1 + 3 + 5 + … + (2n₀ − 1) = n₀². Given this assumption, I have to show that the formula holds for n = n₀ + 1, that is, I have to show that 1 + 3 + 5 + … + (2n₀ − 1) + (2(n₀ + 1) − 1) = (n₀ + 1)². Using the assumption, I get that 1 + 3 + 5 + … + (2n₀ − 1) + (2(n₀ + 1) − 1) = n₀² + (2(n₀ + 1) − 1). And, simplifying, I get n₀² + (2(n₀ + 1) − 1) = n₀² + 2n₀ + 2 − 1 = n₀² + 2n₀ + 1 = (n₀ + 1)². In short, 1 + 3 + 5 + … + (2n₀ − 1) + (2(n₀ + 1) − 1) = (n₀ + 1)², which is what I wanted.

If we have some experience reading proofs by mathematical induction we may interpret the telling of the sequence of actions against the background of what we may call the proof by mathematical induction script, that is, against the standard sequence of actions performed in proofs by mathematical induction. We will then see that the particular sequence of actions taken in the example follows the script. In fact, in this example, we need to read the telling of the sequence of actions against the background of this script or pattern of actions in order to see that the formula has been proved. An analogous point can be made if we consider the following narrative: ‘John went over to Bill’s house. He watched Bill open his presents. John ate the cake and left’ (adapted from Schank and Abelson 1977: 39). Readers will see that the description fits the birthday party script and can fill in the important ‘detail’ that Bill is having a birthday party. In short, relying on scripts, readers recognize that a standard proof by mathematical induction is taking place without being told so explicitly and, similarly, readers recognize that a birthday party is taking place without being told so explicitly.

In the particular case of the example of the proof by mathematical induction, there are no surprises when we read the telling of the sequence of actions against the background of the proof by mathematical induction script. This is to say that the sequence of actions in the example proceeds entirely as we expect given our experience with how proofs by mathematical induction are usually carried out. But we would probably have been surprised or paid special attention if an unusual idea were used to show that the formula holds for n = n₀ + 1 or if the formula were of a kind where we would not expect the formula to be provable by mathematical induction.

It is worth emphasizing that my claim that proofs are often read as narratives is a claim about how research mathematicians read proofs. My discussion of the proof by mathematical induction thus provides a simplified illustration of how research mathematicians read proofs, since I speak about how we, who are not research mathematicians, would read the proof by mathematical induction. Most of us possess only few and simple scripts about standard sequences of actions in mathematical proofs.⁸

19.3 Evidence

In this section I draw on recent interview studies with research mathematicians about how they read proofs. The interview data is consistent with the claim that mathematicians can and often do rely on scripts when they read proofs, and that they focus attention on the breaches of the scripts, on that which appears unusual or surprising in the proofs against the background of scripts.

It is important to note that how mathematicians read proofs may well have changed substantially over time. And even if we assume that mathematicians now tend to read proofs as narratives and always have tended to read proofs as narratives, the scripts they have used will have changed over time. For example, the fact that mathematicians have used different criteria for mathematical proof across cultural and historical contexts means that they will have used different scripts. I here rely on research in the history and philosophy of mathematics which has uncovered how mathematical proofs as they occur in mathematical practice are context-sensitive. For example, the level of rigour required for mathematical proof has varied across time and discipline.

19.3.1 Reliance on Scripts

Various recent interview studies with research mathematicians suggest that mathematicians read proofs against the background of what they know from their experience with proofs. For example, these studies suggest that mathematicians have beliefs and expectations about which methods and techniques work in which situations and on which sorts of mathematical objects, and that mathematicians rely on these beliefs and expectations as they read proofs. The mathematicians seem to see recognizable patterns of action in the sense of standard sequences of proving actions in the proofs. In this sense, proofs seem to unfold against the backdrop of scripts about standard sequences of proving actions. In sum, mathematicians, as they read proofs, seem to rely on scripts about which methods and techniques work in which situations and on which sorts of mathematical objects.

For example, based on their interviews with mathematicians, Weber and Mejía-Ramos (2011: 340) suggest that mathematicians, when they read proofs, ‘might encapsulate strings of derivations into a short collection of methods and determine whether these methods would allow one to deduce the claim that was proven’. Whether the methods will work to prove a result is something they judge based on their experience. Weber and Mejía-Ramos note (2011: 340) that Konior (1993) provides further data to support their claim. Konior reports on the analyses of several hundred proofs. Konior found that a written proof often contains cues that indicate to the reader how to separate the proof into parts and what methods were being used in each part. For example, a part of a proof may begin with: ‘We have to define a one-to-one mapping g of X onto Y’ (Konior 1993: 255). In this way, a proof seems to cue readers to activate their scripts about which methodological moves work when.

Andersen (2020) has interviewed mathematicians about their proof reading practices when they act as referees for mathematics journals. Based on the interviews, she similarly suggests that mathematicians read proofs against their experience concerning which approaches work to prove different kinds of results. Mathematicians appear to have reliable intuitions based on experience about which type of approach can typically be used to prove a sub-result of this or that type. The beliefs and expectations the interviewees have about proving actions seem to correspond to what we here call scripts.

A study by Andersen, Johansen and Sørensen (2019) indicates that the scripts may in part be provided by the main text preceding a proof in an article. The study reports on interviews with a supervisor and his PhD student. In line with the interviews referenced above, the supervisor described how he studies the ‘general pattern’ or flow of a proof and sees if it is recognizable or instead raises ‘red flags’, which can be interpreted to mean that he studies whether the proof follows the scripts or there are points at which a script is breached. He pays special attention ‘if some of this pattern recognition raises a red flag or indeed gives any hint of unease or alienation’ (Andersen, Johansen and Sørensen 2019: 11). His expectations with respect to the flow of the proof are sometimes informed by the main text of the article presenting the proof. He emphasized how an article may provide examples of how a mathematical object behaves in different situations before presenting proofs establishing results about the object in question. The examples may then influence the expectations of the supervisor with respect to how the results can be proved.

The interview data is consistent with my claim that mathematicians can and often do rely on scripts based on their experience as mathematicians when they read proofs. The interview data does not shed light on the question of what concrete scripts that play a role in mathematical practice look like exactly. This is a question for future research. Note that we would expect that the parts of proofs that follow the scripts are not the parts readers focus attention on, exactly because there are no breaches of the scripts. This is supported by Andersen, Johansen and Sørensen (2019) and Andersen (2020), whose interviews suggest that mathematicians do not thoroughly read the parts of a proof that unfold the way they would expect.

19.3.2 Breaches of Scripts

Sometimes something unusual happens in a proof. In a number of interview studies, mathematicians describe how they pause and pay close attention when they read a proof and encounter something ‘surprising’ (Andersen 2020: 238) or something ‘strange’ or ‘odd’ (Weber 2008: 448). Or how they pay close attention when a part of a proof is ‘suspicious’ (Müller-Hill 2011: 307–308, 327–328) or raises a ‘red flag’ (Andersen, Johansen and Sørensen 2019: 11). My argument above offers an interpretation of the parts of proofs the mathematicians describe here. I argued that the usual moves in proofs can be interpreted as the moves that follow the scripts about standard sequences of proving actions. The unusual moves that mathematicians describe that make them pause and pay close attention when they are reading proofs should then be interpreted as the moves that do not follow the scripts about standard sequences of proving actions. The unusual moves are breaches of the scripts.

As described in section 19.2, ‘Reading Proof as Narrative’, above, when narratives convey something unexpected relative to existing scripts, the narratives feed into new scripts. In the case of mathematical proofs, we may ask how unusual moves contribute to the creation of new scripts. Moves that may be unusual at one point in time may become standard moves at a later point in time because they have been shown to work in various proofs.⁹ Before new moves can turn into standard moves, the new moves must be carefully checked in the proofs that use them. This probably involves careful attention to detail and filling intentional gaps with extra details.¹⁰ It has previously been claimed that readers of mathematical proofs commonly fill intentional gaps in the proof and thus engage in a kind of reconstruction of the proof (see, for example, Fallis 2003; Netz 2009: 71–80; Rav 1999).¹¹ It is worth adding that how the readers perceive the narrator or the author of the proof may affect how thoroughly they check unusual moves. Readers may be more thorough if the author is a PhD student than if the author is an experienced mathematician (Andersen 2017: 184–187; Inglis and Mejía-Ramos 2009; Mejía-Ramos and Weber 2014: 165–168).

When proofs are read as narratives, reading proofs really involves two kinds of reconstruction of the proofs on the part of the readers, one of which is the kind of reconstruction of proofs that has previously attracted attention from philosophers. Readers of proofs engage in the kind of reconstruction that has been discussed previously when they fail to see what is going on, when they cannot follow a step in a proof, that is, when they cannot see why B follows from A as the author claims. The readers will then insert extra steps between A and B. As just mentioned, this kind of reconstruction probably plays a role where breaches of scripts occur and in establishing new scripts. But readers also engage in a kind of reconstruction when they recognize what is going on, when they recognize a move in a proof as an instance of a script for a standard way of proving this sort of result. The details they insert in the proof are provided by the existing script. Hence, in both kinds of reconstruction details are filled in by the readers but the details are different in kind and are inserted in different parts of the proof.

19.4 Understanding Proofs

The present account of how proofs are read provides a way of thinking about mathematical understanding, more specifically the understanding of proofs.¹² Among cognitive scientists, understanding is commonly characterized as ‘a process by which people match what they see and hear to pre-stored groupings of actions that they have already experienced’ (Schank and Abelson 1977: 67; quoted in Herman 1997: 1048). In particular, coming to understand a narrative is the process by which narratives are interpreted by their readers or auditors against the background of scripts about how events or actions of the kinds represented in the narrative usually take place. Hence, the process by which readers or auditors come to understand a narrative is the process by which they use scripts to reconstruct the narrative. Consider again the narrative: ‘John went over to Bill’s house. He watched Bill open his presents. John ate the cake and left’ (adapted from Schank and Abelson 1977: 39). This narrative does not make much sense if we do not interpret the narrative against our knowledge of how birthday parties usually take place. We come to understand the narrative by reading the narrative against our birthday party script. Thus, we envision the guests arriving at Bill’s house, the guests each giving Bill a present, and Bill tearing the wrapping paper.

When the narratives convey something unexpected relative to existing scripts, the readers can fail to understand. For example, in the case of the narrative where Bill’s cat tried to lick John’s hand, the readers may fail to understand how John ended up with cheese on his hand when he ate cake, since, according to their birthday party script, a birthday cake is made of flour, sugar and eggs, not cheese. But, when the narratives convey something unexpected relative to existing scripts, the narratives can also contribute to the creation of new scripts and thus new ‘models for understanding’ (Herman 1997: 1056). Hence, when the readers see that Bill’s birthday cake is made of cheese rounds, and on a number of other occasions have read narratives where scripts are breached by involving cheese rounds as cakes, they may be led to new scripts about usual sequences of actions in which a cake made of cheese rounds is served.

If proofs are read as narratives, as I suggest, the process by which mathematicians come to understand proofs is the process by which they use scripts to reconstruct the proofs. Thus, the process by which mathematicians come to understand a proof involves recognizing the moves in the proof as instances of scripts about standard sequences of proving actions.¹³ Hence, the process by which mathematicians come to understand the proof of the formula 1 + 3 + 5 + … + (2 n − 1) = n² I gave earlier involves recognizing my moves in the proof as instances of the proof by mathematical induction script.

Consider the following useful analogy suggested by Norton Wise between coming to understand a proof and coming to understand how to frame a new roof. Coming to understand how to frame a new roof requires experience with patterns of roof framing in many particular instances. Considering only how each part of the new roof framing is placed is not enough for understanding how the stability of the whole emerges. Similarly, going through the steps of a given proof is not enough for understanding the proof but requires experience with patterns of proving in many particular instances. In other words, understanding the proof requires scripts about standard sequences of proving actions.

The present account of proof understanding can explain why mathematicians emphasize that one may have verified every logical step of a proof and still not have understood the proof. Poincaré makes this point. A mathematician may have ‘examined [the elementary] operations one after the other and ascertained that each is correct’ and still not have ‘grasped the real meaning’ of the proof (Poincaré 1958: 217–218). Feferman similarly notes that, ‘It is possible to go through the steps of a given proof and not understand the proof itself’, and adds that understanding the proof is ‘a special kind of insight into how and why the proof works’ (Feferman 2012: 372; quoted in Folina 2018: 136).

While verification is not a form of understanding, the opposite may be true. Understanding may be a form of verification. Research on proof reading tends to focus on the validation of proofs rather than the understanding of proofs. But the present account of mathematical understanding seems to suggest that there is a strong connection between the understanding and validation of proofs (in line with Dutilh Novaes 2018; and Mejía-Ramos and Weber 2014). The present account of understanding indicates that mathematicians understand proofs through action pattern recognition, which is the same kind of action pattern recognition that previous studies, based on interviews and a survey with mathematicians, suggest that mathematicians use to validate proofs (Andersen 2020; Andersen, Johansen and Sørensen 2019; Mejía-Ramos and Weber 2014; Weber and Mejía-Ramos 2011). Hence, proof understanding may be a form of proof validation.

I end this section by briefly considering how the present account of mathematical understanding is relevant to mathematics education, to the teaching of proofs. As noted by Weber and Mejía-Ramos, ‘a goal of many research programs is to lead students to think and behave more like mathematicians with respect to proof’ (2011: 330). This raises the question of how students may read and come to understand proofs in a way that is similar to how mathematicians read and come to understand proofs. Needless to say, students will always have different scripts available to them than research mathematicians do. And presumably it is tempting for students to read proofs word by word without at all engaging in the kind of reconstruction of the proof scripts that mathematicians engage in. But the picture of proof reading presented in this chapter suggests that students cannot come to understand proofs this way. Not even mathematicians come to understand proofs by reading them only word by word rather than against the backdrop of scripts. Hence, if we want to teach students to read and come to understand proofs in a way that is similar to how mathematicians read and come to understand proofs, the teaching of scripts about standard sequences of proving actions is important. For example, it is valuable to teach students about the sorts of results that can be proved by mathematical induction and the commonalities between different proofs by mathematical induction.

19.5 Conclusion

Focusing on how mathematical proofs are read, I have argued that mathematical proofs can be and often are read like narratives by research mathematicians. Mathematicians read proofs as narratives when they read proofs against the backdrop of experience-based scripts about standard sequences of proving actions. They focus attention on the breaches of the scripts, on that which appears unusual or surprising against the backdrop of the scripts. The account I have defended of how proofs are read as narratives also provides an account of how to conceive of proof understanding, which is a topic that has received very little attention in the literature. On this account of proof understanding, a process by which mathematicians come to understand proofs is the process by which they relate proofs to scripts about standard sequences of proving actions.¹⁴

20 Narrative Solutions to a Common Evolutionary Problem

John Beatty

20.1 Introduction

Sometimes, in order to understand an occurrence, we need to know what happened prior to that. (Yes, that sounds obvious, and yet … .)

And sometimes it’s not enough to know what happened immediately prior to that. We need a backstory that rewinds time to some event in the more distant past, and then takes us forward through events that (1) were not foreseeable from the starting point, and (2) were consequential for the outcome of interest, (3) in the order in which they occurred and not just any order. Such a backstory is narrative-worthy – a narrative is just right for the occasion – as I will explain later.

I’ll illustrate these points with a common problem from evolutionary biology. Or rather, I’ll rely on the common problem in order to introduce/motivate the need for a narrative solution. Evolutionary explanations sometimes (often?) invoke only circumstances contemporaneous with the phenomena to be explained – no backstory, indeed, atemporal evolutionary reasoning, as odd as that may sound. This might be satisfactory in some contexts. But it is not satisfactory in other contexts, like those I’ll discuss, where narrative-worthy backstories are called for.

20.2 Darwinian Assumptions: Successive, Slight Modifications

Flatfishes – halibut, turbot, sole, others – live horizontally/flat on the sea floor. They differ from vertical fishes in appropriate ways. Most notably, instead of having one eye on each side of their head, they have both eyes on one side, the topside, making it easier to see their prey and watch out for predators. Flatfishes are commonly coloured in a way that camouflages them against their background. Usually, just the topside is camouflaged; the unseen bottom is pigmentless (Figures 20.1 and 20.2).

Figure 20.1 Flatfish (flounder) topside

Rhombosolea leporina (Yellowbelly flounder)

Source: This illustration is licensed under the Creative Commons Attribution ShareAlike 1.0 license. The author is Dr Tony Ayling. The illustration was originally published in Tony Ayling and Geoffrey Cox, Guide to the Sea Fishes of New Zealand (Auckland: William Collins Publishers, 1982) https://commons.wikimedia.org/wiki/File:Rhombosolea_leporina_(Yellowbelly_flounder).gif.

Figure 20.2 Still life by Jan van Kessel the Elder

Jan van Kessel the Elder, 1626–79.

Source: Wikimedia Commons.

It is common to account for the traits of a species – like the lifestyle, anatomy and coloration of flatfishes – in terms of their appropriateness, the ways in which they individually, and in combination, enhance the fitness of their possessors. The possession of fitness-enhancing traits is what we expect from evolution by natural selection.

This kind of reasoning is odd, though, when you think about it: it amounts to an evolutionary account of the present that does not invoke the past, just prevailing circumstances. It’s an atemporal evolutionary account.

It’s a common enough manner of reasoning to have been called out for criticism by Stephen Gould and Richard Lewontin (1979). They dubbed it the ‘Panglossian paradigm’, after Voltaire’s Dr. Pangloss, for whom everything in the world was maximally appropriate. Thus, species are maximally adapted to their environments. But, to account for the traits of species entirely in terms of their ‘current utility’, Gould and Lewontin objected, is to pretend that adaptation does not take time. Such accounts reflect the untenable assumption of ‘immediate adaptation’ to whatever environment a species inhabits, the ‘immediate work of natural selection’.¹

Surely the targets of Gould and Lewontin’s critique did not really believe that species instantaneously adapt to their environments. But perhaps they assumed, along with John Maynard Smith, that ‘most populations have had time to come close to the optimum for the environment in which they live’ (Maynard Smith 1993: 11–12; my emphasis; but see also Maynard Smith et al. 1985). Had there been insufficient time for populations to ‘come close to the optimum for the environment in which they live’, then one could not make sense of their traits without taking into consideration the ancestral starting points from which they had not completely departed. But, as it (supposedly) happens, there’s no need to bother with the past; the present is enough.

On the Panglossian paradigm, there is another, related respect in which history is irrelevant. It has to do with the equilibrating character of an optimizing process like evolution by natural selection – or, rather, as evolution by natural selection is commonly conceived. Consider an analogy. We find a marble lying in the bottom of a bowl. How did this come about? We need only take into account the prevailing properties of the marble and the bowl, and the principles that govern this mini-universe. The past is largely irrelevant, since the marble would have rolled around and eventually come to rest there, no matter where it started. Similarly, one might think that a species (marble) evolves by natural selection in its environment (bowl) until it attains the optimum combination of traits, the equilibrium point where it then rests, no matter its starting point. One need only take into account the prevailing circumstances.

Darwin himself could not have reasoned persuasively in this manner. His case for evolution by natural selection vs special creation depended on linking the present to the past. For instance, it makes better sense to attribute imperfect but satisfactory traits – like the wonky but workable placement of flatfish eyes – to the trial and error modification of an imagined ancestor, in this case with an eye on each side, than to an all-knowing and benevolent creator who engineers each species from scratch.

Evolutionary biologists today are rarely concerned to dispatch special creation. But, insofar as they are Darwinians, they have other sceptics to contend with, and in doing so they have other reasons for looking beyond the present into the past. And here’s (at least one reason) why.

To make sense of the traits of a species, a Darwinian should be able to go back in time to an ancestor of that species, and then forward to the species in question. But, in going forward from the ancestor, the good Darwinian cannot rely on an all-at-once modification. It should be possible to specify a sequence of slight modifications that would lead to the descendant. As Darwin acknowledged:

If it could be demonstrated that any complex organ existed which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down.

(Darwin 1859: 189)

I would just add the following friendly amendment (this is after all what he meant):

If it could be demonstrated that any complex organ existed which could not possibly have been formed by numerous, successive, slight modifications, each of which increases fitness, my theory would absolutely break down.

There are times when Darwinians do not hold themselves – and are not held – to these standards, presumably on the grounds that the ancestral and gradual intermediate stages of evolution are not difficult to fathom and are perhaps not worth the worry. But there are also cases where it is not at all clear that there is a backstory that meets these standards, and the challenge is to provide one.

20.3 Plausible Orderings of Modifications

St. George Jackson Mivart, for one, challenged Darwin on his own terms (Mivart 1871).² And flatfish eyes (among other examples, see further) served him well in this regard. To give a Darwinian account of their eyes, by Darwin’s own criteria, it is not sufficient to demonstrate the usefulness of that arrangement at present. One must also propose a sequence of slight modifications leading from an ancestor with one eye on each side to descendants with both eyes on one side, each step of which increases fitness.

What’s a plausible sequence? Surely not by slight displacements of one eye through the skull to the other side! Surely it would involve slight displacements of one eye over the top of the skull to the other side. But that leaves unanswered how the initial and early displacements could have been fitness enhancing. He imagines a fish lying flat on its side with one eye in the sand. What’s the advantage of having the lower eye only slightly closer to the top of the skull? It’s still in the sand. How can the initial migrations of the eye have been anything but injurious, given the skull/eye-socket reconstructions involved?

Another instance which may be cited is the asymmetrical condition of the heads of the flat-fishes (Pleuronectidæ), such as the sole, the flounder, the brill, the turbot, &c. In all these fishes the two eyes, which in the young are situated as usual one on each side, come to be placed, in the adult, both on the same side of the head. If this condition had appeared at once, if in the hypothetically fortunate common ancestor of these fishes an eye had suddenly become thus transferred, then the perpetuation of such a transformation by the action of ‘Natural Selection’ is conceivable enough. Such sudden changes, however, are not those favoured by the Darwinian theory […] But if this is not so, if the transit was gradual, then how such transit of one eye a minute fraction of the journey towards the other side of the head could benefit the individual is indeed far from clear. It seems, even, that such an incipient transformation must rather have been injurious.

(Mivart 1871: 37–38)

Mivart generalized the problem and gave it a name: ‘the incompetency of “natural selection” to account for the incipient stages of [ultimately] useful structures’ (Mivart 1871: 23). It has since been shortened to ‘the problem of incipient stages’.³ As he put the point:

‘Natural Selection,’ simply and by itself, is potent to explain the maintenance or the further extension and development of favourable variations, which are at once sufficiently considerable to be useful from the first to the individual possessing them. But Natural Selection utterly fails to account for the conservation and development of the minute and rudimentary beginnings, the slight and infinitesimal commencements of structures, however useful those structures may afterward become.

(Mivart 1871: 23)

In addition to flatfish eyes, he illustrated the problem with other traits like the giraffe’s neck, vertebrate limbs and mimicry. And mammary glands:

Is it conceivable that the young of any animal was ever saved from destruction by accidentally sucking a drop of scarcely nutritious fluid from an accidentally hypertrophied cutaneous gland of its mother?

(Mivart 1871: 47)

For these and other reasons, Mivart inferred that new species arise not gradually, but with ‘suddenness’.

Not only are there good reasons against the acceptance of the exclusive operation of ‘Natural Selection’ as the one means of specific origination, but there are difficulties in the way of accounting for such origination by the sole action of modifications which are infinitesimal and minute, whether fortuitous or not.

Arguments may yet be advanced in favour of the view that new species have from time to time manifested themselves with suddenness, and by modifications appearing at once […] the species remaining stable in the intervals of such modifications.

(Mivart 1871: 97)

Darwin well understood and appreciated the difficulty that Mivart had raised, and in the 6th and final edition of On the Origin of Species he devoted considerable space to the problem. Referring to this and other criticisms, Darwin wrote:

A distinguished zoologist, Mr. St. George Mivart, has recently collected all the objections which have ever been advanced by myself and others against the theory of natural selection, as propounded by Mr. Wallace and myself, and has illustrated them with admirable art and force. When thus marshalled, they make a formidable array.

(Darwin 1872: 176)

Darwin responded to a variety of Mivart’s objections. But he took most seriously, and spent the most time responding to, the problem of incipient stages (1872: 177–190), including a solution to the problem of flatfish eyes. He affirmed that the trajectory of evolution involved the eye moving over the top of the skull, onto the other side. The maturation of flatfishes provided the evidence. According to August Malm (1867), flatfish larvae swim vertically and – appropriately under the circumstances – have one eye on each side of their head. But, as they develop, one eye begins to migrate towards and then over the top of the skull, to the side that becomes the topside of the horizontal, bottom-dwelling adult, as shown in the sequence in Figure 20.3. The migration of the eye during the development of individual flatfishes, together with their change in orientation from vertical to horizontal, reflects the trajectory of their evolution. Why else would flatfishes undergo that course of development, other than because they were descendants of vertical, symmetrically eyed fishes?

Figure 20.3 A depiction of eye migration in starry flounder larvae, that also illustrates Darwin’s suggested evolutionary account of the flatfish eye.

Source: Policansky (1982).

But that leaves unanswered how the initial migration of the one eye, and early extensions of that migration, could have been fitness-enhancing, which had been Mivart’s main puzzle. Darwin didn’t entirely capitulate on this, but partly/largely. He attributed the initial, slight migration of the eye, and then early extensions of that migration, not to evolution by natural selection, but rather to Lamarckian inheritance of acquired characters. He took it on authority from Malm that very young flatfishes lying on their sides on the sea floor, before eye migration is complete, strain to see with their bottom eye

[…] and they do this so vigorously that they eye is pressed hard against the upper part of the orbit [socket]. The forehead between the eyes consequently becomes, as could be plainly seen, temporarily contracted in breadth.

(Darwin 1872: 187)

Suppose this forced displacement of the eye, so as to see better, was inherited by the next generation, who also strained to see. Resulting in still further displacement of their eyes. The further, forced displacement was also inherited. This took place generation after generation until the eye made its way far enough around the skull that it was sufficiently out of the sand, at which point its migration to the other side was maintained and extended to the present state by natural selection. It was a largely Lamarckian, only partly Darwinian solution.

We thus see that the first stages of the transit of the eye from one side of the head to the other, which Mr. Mivart considers would be injurious, may be attributed to the habit, no doubt beneficial to the individual and to the species, of endeavouring to look upwards with both eyes, whilst resting on one side at the bottom. We may also attribute to the inherited effects of use the fact of the mouth in several kinds of flat-fish being bent towards the lower surface.

(Darwin 1872: 187–188)

Darwin further explained the lack of pigment on the bottom of flatfishes in terms of the Lamarckian notion that disuse of a trait, over many generations, leads to its loss (Darwin 1872: 188).

Somewhat tangentially, Lamarck had offered his own explanation of flatfish eye placement. The ancestors of flatfishes fed in very shallow waters along shorelines, he supposed, waters so shallow that they had to lie flat on their sides. ‘[T]his requirement has forced one of their eyes to undergo a sort of displacement, and to assume the very remarkable position found in the soles, turbots, dabs, etc.’ (Lamarck 1914: 120).

The all-at-once modification, as applied to flatfish evolution, and many other problematic cases, was developed in great detail by Richard Goldschmidt (1940) and had considerable influence well into the twentieth century and, in one form or another, to this day. Here is a thin version of his thinking, in connection with flatfish eyes:

In a former paper (Goldschmidt 1933) I used the term ‘hopeful monster’ to express the idea that mutants producing monstrosities may have played a considerable role in macroevolution. A monstrosity appearing in a single genetic step might permit the occupation of a new environmental niche and thus produce a new type in one step […]. A fish undergoing a mutation which made for a distortion of the skull carrying both eyes to one side of the body is a monster. The same mutant in a much compressed form of fish living near the bottom of the sea produced a hopeful monster, as it enabled the species to take to the life upon the sandy bottom of the ocean, as exemplified by the flounders. (Goldschmidt 1940:

390–391; and see 1933: 545)

The question of flatfish eyes is often posed as one that pits a Goldschmidtian (and, to be fair, Mivart-inspired) solution against a Darwinian approach – for example, in Thomas Frazzetta’s (2012) review: ‘Flatfishes, Turtles, and Bolyerine Snakes: Evolution by Small Steps or Large, or Both?’.

The problem of flatfish eyes, from a Darwinian point of view, seems to have eluded even the master Darwinian communicator, Richard Dawkins. No, Dawkins doesn’t go Lamarckian, nor Goldschmidtian. But he pulls up short of going fully Darwinian. He makes the Darwinian point that it is more reasonable to attribute asymmetric flatfish eyes to the modification of a symmetrically eyed ancestor than to special creation. Surely an intelligent designer would have created flatfishes more in the manner of skates and rays, flattened from top to bottom, with both eyes on top, rather than flattened from side to side and requiring the migration of one eye to the other side.

Even though its [the flatfish’s] evolutionary course was eventually destined to lead it into the complicated and probably costly distortions involved in having two eyes on one side, even though the skate way of being a flat fish might ultimately have been the best design for bony fish too, the would-be intermediates that set out along this evolutionary pathway apparently did less well in the short term than their rivals lying on their side.

(Dawkins 1986: 92–93)

Yes, ‘apparently’ in the lineages that beget flatfishes, lying flat on one side with a migrating eye prevailed over flattening from top to bottom. But how could the admittedly ‘complicated and probably costly distortions’ of the intermediate stages have been sufficiently advantageous to be selected for?

Interestingly, there have been recent – for the first time – fossil findings of flatfishes with intermediate stages of eye migration (Friedman 2008). But there is still no generally accepted, functional account of the fitness contributions of the early steps. The most promising clue is a fact about some flatfishes, maybe many or all, that has been known for quite a long time although not considered until recently in this connection (Olla, Wicklund and Wilk 1969; Stickney, White and Miller 1973; Friedman 2008: 211; Frazzetta 2012: 33). That is, adult flatfish (flounders in this case) sometimes use their dorsal and anal fins (what would have been the top and bottom fins in their ancestors but are right- and left-side fins in flounders) to prop themselves up, raising their heads to better see above them and lunge at their prey.

Now if early flatfishes could raise their heads in this way, and in the process raise their lower eyes out of the sand, then a slight migration of the lower eye could have allowed slightly better vision than having it face straight down and may have been selected for. And a further extension of the migration would be advantageous and selected for. And so on and so on, the entire eye migration thus being due to evolution by natural selection.

The sequence here is crucial. Head elevation for lunging is not only adaptive in combination with other flatfish traits. Its position in the sequence of evolutionary events, prior to eye migration, is what makes eye migration adaptive and hence evolutionarily possible.

An account of flatfish eye placement in terms of its current usefulness is not wrong, but it is possibly misleading, and in any case sorely incomplete, prompting the sort of objection raised by Mivart. A Darwinian explanation requires a backstory – back to an ancestor that had symmetrically placed eyes; and then forward from there, through a careful – just the right, so to speak – sequence of stages.

The problem of incipient stages – requiring a backstory – arises for various reasons. In the case of flatfish eye migration, it has to do with so-called ‘epistatic’ interactions, where the fitness contribution of a trait depends on the presence or absence of other traits. In the case of ‘sign epistasis’, the fitness contribution of a trait is positive or negative (plus ‘sign’ or minus ‘sign’) or zero depending on the presence or absence of another trait (Weinrich, Watson and Chao 2005; Weinrich et al. 2006; Poelwijk et al. 2007). For example, eye migration is fitness-enhancing in combination with head elevation, but fitness-neutral, or more likely fitness-diminishing alone. This particular kind of epistatic interaction results in there being multiple, sequential pathways to an optimal outcome, some of which are traversable by natural selection and some not. Which is to say, again, that it is not enough to attribute even highly adaptive traits to natural selection without also positing an ancestor and a carefully sequenced route from the ancestor to the descendant in question.

Consider another example of the problem of incipient stages that also points to the importance of backstory, but that arises and is resolved in a different way – different from the epistasis case. It concerns the evolution of wings. Mivart is often said to have asked ‘What use is half a wing?’ Mivart did not say that (at least not in the text regularly cited), and that does not sound like his manner of posing the more general problem. Stephen Gould (1991) had a better way of putting Mivart’s point: the ‘5 percent of a wing’ problem. How could evolution by natural selection of slight modifications lead from wingless, flightless ancestors to winged, flying descendants? How could the miniscule, incipient wing-lets have been sufficiently useful for flying in order to be favoured by natural selection? Mivart concluded, ‘It is difficult […] to believe that the Avian limb was developed in any other way than by a comparatively sudden modification of a marked and important kind’ (1871: 107). Which, again, violates Darwin’s ‘successive, slight modifications’ constraint.

The most promising solution in this case is one that Darwin himself proposed, and illustrated with the case of wings.⁴ The basic idea is that the incipient stages of the trait in question were useful in a (perhaps very different) way than the later stages. ‘In considering transitions of organs, it is so important to bear in mind the probability of conversion from one function to another’ (Darwin 1872: 183).

The structures that were eventually modified for flight might have served a variety of other functions, depending on the animal in question (e.g., insects vs birds) and depending on issues of scale. Darwin himself suggested that the thoracic wings of some insects might be modifications of parts originally related to respiration. Wings of insects and birds might, in their incipient stages, have served a variety of aerodynamic uses other than flight, like gliding and altitude control during descent. Narratives of the evolution of bird flight generally begin with the modification of four-legged ancestors into bipedal descendants, followed by modification of the forelimbs into wings. On some accounts, the initial modifications improved running and jumping (e.g., by improving balance). On other accounts, the modification followed tree climbing, and served aerodynamic uses related to descent mentioned above.

20.4 When Narratives Are Worthwhile

The evolution of flatfish eyes, wings and many other traits are narrative-worthy, in a sense I’ll now explain. But first: I’m going to rely on a fairly minimal view of narratives, namely that they relate what happened, one event at a time. In this regard I’m following the lead of narrative theorists who adopt similarly minimal views of what counts as a narrative and who concede that almost anything is narratable, but who deny that everything narratable is worth narrating. Some narratives are pointless. As William Labov famously commented:

Pointless stories are met (in English) with the withering rejoinder, ‘So what?’ Every good narrator is continually warding off this question; when his narrative is over, it should be unthinkable for a bystander to say, ‘So what?’

There are a great many ways in which the point of a narrative can be conveyed – in which the speaker signals to the listener why he is telling it. To identify the evaluative portion of a narrative, it is necessary to know why this narrative – or any narrative – is felt to be tellable.

(Labov 1972: 366, 370)

Labov’s term ‘tellable’ has become the state of the art (narrative theory) term for what I prefer to call narrative-worthy. But I like his term ‘pointless stories’. What is a narrative-worthy as opposed to a pointless story? The criteria that I offer may just be a few of many criteria for narrative-worthiness. Perhaps there are stories worth narrating that do not meet the following criteria but are worth telling for other reasons.

For starters, I’d say narratives are good for situations where we don’t know – on the basis of what has already happened, and general principles – what will happen subsequently, and we need to be told.⁵

To clarify, this does not render pointless all of those stories where the narrator begins with the ending. Most historical narratives, in both civil history and natural history, begin with the outcome, and the narrator then proceeds to tell how it came about. Rather, the criterion calls into question the need for narrating how an outcome came about, when the outcome was already foreseeable from the initial events.⁶

An example of a situation where narratives are not particularly useful – where they do not serve this basic function – involves equilibrating/optimizing processes like the one discussed earlier of a marble coming to rest in the bottom of a bowl. Why narrate its trajectory – ‘it was there, then it was there, then there’ – if we can derive from the start where the marble will end up? And regardless of where it started from.

Similarly, why narrate the evolution of a species in an environment if we ‘know’/suppose that it will eventually reach its predictable optimal state given that environment. And regardless of where it started from?

Whereas to make sense of flatfish eyes, wings, etc. in terms of evolution by natural selection, one must provide a backstory – back to a presumed ancestor, and then the sequence of stages moving forward. And these stages were hardly guaranteed by what preceded them. They could hardly be derived from past circumstances. Flatfish head elevation was hardly foreseeable from the point at which their ancestors first lay flat on their sides on the sea floor. It was certainly not foreseeable by generations of naturalists who contemplated the evolution of flatfish eye placement. Nor, if one were to start the backstory earlier (as I’ll discuss shortly) would it be predictable that predatory fishes inhabiting the sea floor would adopt the behaviour of lying flat on their sides, given that many bottom-dwelling predators (e.g., groupers) never have.

Consider the prominence of narratives in Darwin’s work, and what makes them so worthwhile. Their employment, and their value reflect in part Darwin’s view that, outside of gradual adaptation to environmental circumstances, there is nothing inevitable in the history of life.

I believe in no fixed law of development, causing all the inhabitants of a country to change abruptly, or simultaneously, or to an equal degree. The process of modification must be extremely slow. The variability of each species is quite independent of that of all others. Whether such variability be taken advantage of by natural selection, and whether the variations be accumulated to a greater or lesser amount, thus causing a greater or lesser amount of modification in the varying species, depends on many complex contingencies, – on the variability being of a beneficial nature, on the power of intercrossing, on the rate of breeding, on the slowly changing physical conditions of the country, and more especially on the nature of the other inhabitants with which the varying species comes into competition.

(Darwin 1859: 314)

[I]f we must marvel, let it be at our presumption in imagining for a moment that we understand the many complex contingencies, on which the existence of each species depends.

(Darwin 1859: 322)

These ‘many’, unforeseeable ‘complex contingencies’ need to be added to each evolutionary narrative, in the order they arise.

So, again, a good occasion for a narrative is when we don’t know what will happen next and need to be told. But that still leaves room for a lot of narratives not worth telling, pointless. The events worth including are not just those that we would not have foreseen otherwise. They should also be consequential. Otherwise what is the point of including them in the narrative?

Note that, in the last of the Darwin quotes above, he refers not only to the ‘many complex contingencies’ that arise in the course of the evolution of each species, but those contingencies ‘on which the existence of each species depends’, i.e., which are consequential for the evolution of each species. These two facts about evolutionary history correspond to two reasons why narratives are so appropriate for making sense of evolutionary outcomes: the unpredictability of the events narrated, and their consequential character.⁷

Philosopher of history William Gallie stressed the importance of the two criteria in his reader-centric view of what makes a narrative ‘worth following’. In the first paragraph below he stresses the otherwise unpredictable elements that we rely on narratives to supply. In the second paragraph, he stresses that the events in a worthwhile narrative are consequential for the outcome.⁸ Generally, in narratives,

[…] there is a dominant sense of alternative possibilities: events in train are felt to admit of different possible outcomes – particularly those events that count […] that deserve to be recorded, that could be the pivot of a good story. […] [S]ide by side with this there is the recognition that many events, or aspects of events, are predictable either exactly or approximately. But, although recognised, this predictable aspect of life is, so to speak, recessive or in shadow. It is in contrast to the generally recognised realm of predictable uniformities that the unpredictable developments of a story stand out, as worth making a story of, and as worth following. [Or, in other words, using the terminology of footnote 7 (which I said I was trying to avoid, but I just can’t help myself), the contingent (or contingent per se) developments are ‘worth making a story of’.]

[O]f [still] greater importance for stories than the predictability relation between events is the converse relation which enables us to see, not indeed that some earlier event necessitated a later one, but that a later event required, as its necessary condition [i.e., that it was contingent upon], some earlier one.

(Gallie 1964: 26; my emphasis)

I like to represent these two features of narrative-worthy stories with a branching tree of possibilities (Figure 20.4). In this world, the occurrence of event A leaves open the possibility of either B1 or B2. The occurrence of B1 leaves open the possibility of either O1 or O2 and forecloses the possibility of B2 and along with it O3 and O4. A–B1–O2 is one possible history in this world, A–B2–O4 another. There are multiple possible histories in this world; only one can come to pass.

Figure 20.4 Branching-tree representation of narrative-worthy stories

Let’s say it was A–B1–02. B1 was not derivable from A; B2 might have occurred instead. Moreover, B1 was consequential – it made a difference; had B2 occurred instead, 02 would not have occurred. In the literature on narrative theory, events like B1 are often referred to as turning points or branch points (Beatty 2016: 36–37 and references therein). As far as evolutionary narratives are concerned, ‘A’ stands for the ancestral state with which the backstory begins, and the ‘O’s’ stand for alternative evolutionary outcomes.

The diagram helps to show how the sequence A–B1–O2 counts as an explanation of O2 on the prominent ‘counterfactual difference-making’ conception of explanation. As James Woodward expresses the basic idea:

An explanation ought to be such that it enables us to see what sort of difference it would have made for the explanandum if the factors cited in the explanans had been different in various possible ways

(Woodward 2003: 11)

[A] common element in many forms of explanation, both causal and non-causal, is that they must answer what-if-things-had-been-different questions.

(Woodward 2003: 221)

The occurrence of B1 helps to explain O2, in the sense that, had B1 not occurred (had B2 occurred instead), then O2 would not have resulted. Whether B1 or B2 occurs makes a difference.

There is a case to be made that worthwhile narratives include, at least implicitly, what did not occur as well as what did, at least some of the counterfactual as well as the factual sequences of events. But I won’t press that case here (see Beatty 2016; 2017). At the very least, to see the worth of a narrative is to consider what did not happen and thereby see that there were consequential turning points, which, again, contributes to the explanatory character of the narrative.

Figure 20.5 shows a branching time representation of flatfish evolution. The acquisition of head elevation is a turning point that was not inevitable given past events, and that was consequential for the outcome. The order of events here is crucial. It is not enough to consider the three traits in question purely contemporaneously. Yes, they work well together, but that does not explain their presence. The acquisition of the trait, lying flat, made possible the evolution of head elevation for lunging, which in turn made possible the evolution of eye migration.

Figure 20.5 Branching time representation of flatfish evolution

20.5 Conclusion

In order to understand an occurrence, we sometimes need a backstory that rewinds time to some event in the more distant past, and then takes us forward through events that (1) were not foreseeable from the starting point, and (2) were consequential for the outcome of interest, (3) in the order in which they occurred and not just any order. We need a backstory that is narrative-worthy in these respects. Such a backstory is explanatory.

I’ll end with a question that may have occurred to you already, namely how far back should the backstory go? I’m not sure there is a definitive answer to this. But surely it depends in part on the question being asked, or what counts as puzzling.

Gerd von Wahlert rewound the flatfish evolution clock back beyond their ancestors’ horizontal lifestyle – the point at which the narratives above begin – to their more distant ancestors’ vertical lifestyle. According to his narrative, the ancestors then evolved to rest/‘sleep’ lying flat. (Yes, some fishes rest/sleep on their sides.)⁹ And subsequently evolved a horizontal lifestyle. He proposed this as a solution to the puzzle of flatfish eye placement, and a way to avoid Goldschmidt’s hopeful monster scenario.

The flatfish are usually cited not only as a paradigm of adaptation to benthonic life but frequently as a case of an unexplainable major evolutionary step; they are referred to by Goldschmidt as owing their origin to a ‘hopeful monster’. Analysis of their structure and their habits has, however, revealed a simpler story (Wahlert 1961). A shift from an upright to a horizontal sleeping position occurred in the symmetrical ancestors of the flatfish; sleeping on either side is done in some of the present-day symmetrical acanthopterygians, such as triggerfish and wrasses. If this sleeping position were maintained as a resting or hiding position after the animal awoke, a shift of the eye from the blind towards the upper-most side would be an advantageous modification. The shift of the eye on the blind side to the margin of the head would enable the fish to scan the waters above it with binocular vision (Wahlert 1965: 290).

But von Wahlert’s suggestion hardly solves – hardly addresses – the questionable adaptive value of the initial stages of eye migration, and that Goldschmidt (and Mivart) tried to circumvent by invoking an all-at-once transformation. On the other hand, starting with the deeper ancestral state of a vertical lifestyle, as von Wahlert does, followed by the evolution of horizontal resting, does seem a promising solution to a different puzzle, namely how flatfishes acquired a horizontal lifestyle in small steps each of which was selectively favoured. That is, they spent more and more waking time in what was previously just a resting posture, taking more and more advantage of that less conspicuous and motionless position to avoid predators and surprise prey.

And this has the elements of a worthwhile narrative. The acquisition of horizontal resting was hardly guaranteed. Indeed, von Wahlert offers no suggestion as to how it came about. His narrative discloses what was not foreseeable, a basic function of a worthwhile narrative. Moreover, once disclosed, the acquisition of horizontal resting serves as a counterfactual difference-maker in his narrative; it is consequential. Consider the counterfactual alternative: that horizontal resting was not acquired prior to acquisition of a horizontal lifestyle. It is not at all clear how the small steps from a vertical lifestyle directly to a horizontal lifestyle could be advantageous. What could be the advantage of tilting just slightly from vertical to horizontal?

As for the gradual transformation from a vertical lifestyle to horizontal resting, well, any suggestions?¹⁰

21 Just-so What?

Paula Olmos

21.1 Introduction

The recent interest shown by philosophers of science and scholars in related fields concerning the narrative qualities of our scientific explanatory practices has not sufficiently addressed a widespread reluctance to recognize narrative’s epistemological significances. On many occasions, this reluctance is marked by the derogatory use of the ‘just-so story’ label (Gould 1978; Gould and Lewontin 1979) to signify that narrative explanations – or other narrative reason-giving practices (cf. Olmos 2019) – do not meet the epistemic criteria required for scientific appraisal.

In many philosophical forums, it is now standard to present a stark opposition between allegedly genuine scientific explanations – invoking a well-established and well-delimited, ideally law-like account, which is amenable to formalization, perhaps including a causal mechanism¹ – and just-so stories – reconstructive, typically untestable, conjectures of what just-in-fact may or may not have happened to cause a particular phenomenon. However, the problem with the widespread use of this opposition is that it tends to create a strong and somewhat easy association between the noun story and the qualification just-so, so that every attempt to approach explanatory and justificatory tasks through narrative form within the sciences is easily and cursorily dismissed with the just-so story derogatory term. In the worst cases, the use of this summary label even tends to prevent further discussion, acting as a dialectical blockade.²

In this chapter, I examine the roots of the just-so charge by going back to its now classical source in Stephen J. Gould’s original paper (1978) from which it spread in the history and philosophy of science (HPS) field as a negative evaluative term. As it is well known, in choosing this denomination, Gould was inspired by Rudyard Kipling’s collected children’s tales, Just So Stories (1902), containing twelve whimsical etiological fables.³ Gould tried, thus, to convey to the general public the idea of the unscientific and boldly imaginative nature of his own target examples, namely evolutionary biological accounts.⁴

However, a careful reading of Gould’s piece shows that his concerns were neither directed to narrative explanations nor were they, in any case, ineluctably linked to their potential narrative quality. For Gould, assessing an allegedly scientific account as a just-so story, was to issue a negative evaluative judgement regarding its claim to epistemic relevance (as being a bold and so-far unwarranted conjecture) or to point out its way of presentation as avoiding further discussion and further testing (being an unfalsifiable, self-contained hypothesis).

These conditions, I claim, should not be confused or equated with the discursive and causal narrative quality of an explanatory scheme or excluded from the realm of more classically understood explanations that could also, in the mentioned senses, be just-so as well. The narrative quality of many of our scientific reason-giving practices is not, in and of itself, a way to avoid the identification of causal relations (even mechanisms, depending on how we define them; cf. Crasnow 2017) or to exclude further discussion or testing (Al-Shawaf 2019). On the contrary, it might be part of what is hypothesized of certain scientifically interesting phenomena, both in the sense of making them dependent on long-term processes (as witness the timely historicization of certain natural or social enquiries, at a certain point in their development) or on a complex, highly contextual and somewhat indeterministic causal web that’s better rendered in a narrative form.

In what follows, I will carefully examine and analyse Gould’s points and then come back to current discussions regarding the use of narratives and narrative reason-giving modes within the sciences where advocates of the epistemic relevance of the topic have felt the need to meet the challenge of the just-so charge. I identify two basic defensive approaches. One is the vindication of the use of explanatory narratives in cases where the historical, contingent and causally complex nature of the phenomena involved demands an approach that would avoid the strictures of classical models. When these conditions obtain, scientific narratives might be less just-so (i.e., less bold, less self-contained) than their too-narrowly understood mechanistic or easily formalizable rivals. This is basically what is claimed about their case studies by Crasnow (2017) and Currie and Sterelny (2017). Nonetheless, authors working along this line, usually propose some kind of collaboration or integration between these different epistemic modes and tools.

The second kind of vindication⁵ of narrative science follows, instead, an unveiling (somewhat genealogical) strategy, hinting at a deeper level of narrativity. Scholars taking this approach (Richards 1992; López Beltrán 1998; Rosales 2017) try to show how there’s a narrative – or at least a narrative kind of rationality, in W. Fisher’s sense (1989) – behind (or before) each law-like generalization or nomological explanatory formula. This kind of narrative, that depicts and delimits the scenarios in which the particular nomological expression might acquire some sense and specifically become useful for drawing scientific conclusions,⁶ is usually obscured and disregarded in its current application as a validated theory. However, it may always re-emerge when the formula comes under scrutiny as an explanatory principle (sometimes, as in Rosales’s case study, in comparison with rival theories), which makes it a crucial part of its deep understanding.⁷

As in previous contributions (Olmos 2018; 2019), I approach all these topics with the tools and conceptual framework of argumentation theory, that takes into account the argumentative nature of our discursive, explanatory and justificatory practices in terms of reason-giving, reason-asking and reason-discussing activities. The philosophers of science whose works I examine support their claims with (obviously non-demonstrative) reasons and concentrating on their argumentative moves in these discussions helps clarify their positions.⁸ But we must also take into account that their very object of study (scientific explanation and justification) is also of an argumentative, reason-giving nature. The way the grounds of this argumentative activity – i.e., scientific justificatory or forensic practice, in John Woods’s (2017: 143–144) terms – should be assessed is what is finally at stake in philosophical discussions regarding the use of narratives in science and their alleged vulnerability to the just-so charge.

A final step that the argumentative approach might help us take is based on the naturalistic assumption that scientific argumentative activities are already intrinsically normative and evaluative in nature, so that, even if philosophers might discuss the criteria for the acceptability of the concerned claims and explanantia, there is already an intra-scientific evaluative activity going on, whose most basic rule, well beyond the strictures of any aprioristic model, is the dialectical requirement of (a posteriori) openness to collective survey and discussion.⁹

A well-understood just-so charge will have more to do with possible violations of this basic rule than with the textual and formal characteristics of proposed and supported scientific arguments and explanations.

21.2 The Just-so Charge

As already mentioned, the now classical reference for the just-so charge is palaeontologist and evolutionary biologist Stephen Jay Gould’s (1978) critical article on what he saw as the excesses of certain trends in sociobiology, published in the New Scientist under the title ‘Sociobiology: The Art of Storytelling’. For Gould, the question was whether and when evolutionary scientists (sociobiologists, particularly) were being excessively speculative and overconfident with their imaginative accounts about the historical origins of the traits they studied.

Gould’s target cases included, in the first place, certain explanations of animal behaviour that he analysed as solely based on their ‘consistency with natural selection’ or ‘adaptationism’ (Gould 1978: 531). As his first example, Gould picked up David Barash’s explanation of the greater aggressiveness of male mountain bluebirds towards other males approaching their nest before rather than after eggs have been laid. According to Gould, Barash’s proposed explanans to this phenomenon (that he documented, also according to Gould, with rather scarce data) was that this behaviour was advantageous (and so adaptive) as long as it reserved aggression (a costly attitude) to periods where the male was not yet sure to have passed on his genes. This was ‘consistent with the expectations of evolutionary theory’ (Gould 1978: 531) and that was nearly all there was to it.

Gould’s criticism of this case included demands for more data and more tests, exploration of alternative (also testable) explanations and, most significantly, a call for certain control or restraint on the part of the scientist in drawing further conclusions based on his hypothesized explanans. Barash was particularly due criticism as he had gone so far as to suggest this was also a way to understand human foibles regarding adultery (Gould 1978: 531). Barash’s hasty jump from mountain bluebirds to humans brought him near the realm of sociobiology and so allowed Gould to introduce his real main target.

This was Gould’s illustration of a just-so story where the story-like quality of the negatively evaluated explanans was not really in its narrative nature (which was somewhat missing here), but rather in its far-fetched imaginative play. So Gould’s complaints were directed neither against reconstructive historicized explanations that, he assumed, were the goal of evolutionary theory in general, nor against the indication of complex (multifactorial), entangled (non-linear) and somewhat indeterministic causal webs behind a target phenomenon. On the contrary, he explicitly opposed panselectionism or panadaptationism and advocated a less rigid version of natural selection that would ‘grant a major role to other evolutionary agents (genetic drift, fixation of neutral mutations, for example)’ (1978: 531).¹⁰ Although he displayed some irony about the changing styles of evolutionary stories presented in biology – showing the workings of a favoured kind of adaptive factor as it gets theoretically trendy – he seemed to assume this feature as a rather inevitable condition of scientific development.

Now, what was exactly Gould’s problem with sociobiology in particular? Regarding the aforementioned mountain bluebirds example, Gould said that it presents ‘a perfectly plausible story that may well be true. I only wish to criticise its assertion without evidence or test, using consistency with natural selection as the sole criterion for useful speculation’ (Gould 1978: 531). This is the first important thing: it is not a question of the stark unsuitability or unacceptability of a certain methodology, in the sense of a certain way of supporting a scientific content (a kind of reason) but of its sufficiency to establish its conclusion. That, according to Gould, this was happening in sociobiological explanations of human behaviour allegedly more than in other areas of evolutionary theory could be attributed to two additional difficulties met by this particular disciplinary approach: the little observational evidence available,¹¹ and, more significantly for our purposes, the reductionistic option for one specific kind of causal explanation (biological adaptive selection) for such highly complex phenomena as human behavioural traits whose etiological history was surely more entangled than that.¹² Along with this last point, the strictly selectionist sociobiological historical explanations of human behaviour Gould had in mind could be charged with the just-so label precisely for being more reductionistically mechanistic (based on the single principle, ‘if adaptive, then genetic’) than assumedly and sophisticatedly narrative!

The derogatory label, moreover, was the more emphatically attributed by Gould as he perceived that certain ideas taken from the theories advanced by sociobiology were currently being used to uphold practical and political implications as based on what he saw as their hasty conclusions (Gould 1978: 532). So here a certain pragmatist modulation of what seemed to start as a purely epistemological concern comes to the fore.

To sum up, according to my analysis, Gould suggested three (not completely independent but yet distinguishable) criteria to be taken into account for a just-so charge and none of them has to do with either the explanatory historicization of phenomena (their being explained not by the workings of constant laws of nature but by the particular detailed history behind them in conditions that may not be repeated) nor with the suggestion that the phenomena involved might be better understood contextually, taking in account the complexities of its entangled causal web, rather than in isolation. Quite the opposite, I would remark. These three criteria amount to: (a) a charge of theoretical justificatory insufficiency; (b) a charge of unwarranted reductionism; and (c) a charge of prescriptive justificatory insufficiency. Let us analyse these three points separately.

Gould rejects the presentation of a particular historical, reconstructive explanation (not necessarily presented through a fully fledged narrative) as solely supported by its inner plausibility (understood as ‘consistency with evolutionary expectations’). This would be a charge of theoretical justificatory insufficiency and can be mitigated by additional future evidence that would still use that initial plausibility as one of the reasons adduced in favour of the supported hypothesis or by a more humble presentation that would consciously advance it as a plausible hypothesis and offer it to the scientific community, assuming that a lot of research is yet necessary to issue a judgement on it – being still a valuable contribution for some reason (e.g., its novelty).

So this criterion (a), demanding additional robustness for establishing scientific theories, is not much more than a reminder of the collective rules of scientific research, organized scepticism, public scrutiny and assumed fallibilism. This much has been acknowledged by other scholars responding to Gould’s piece: ‘The goal should not be to expel stories from science, but rather to identify the stories that are also good explanations’ (Kurzban 2012). What Gould is asking for here is just ‘more evidence’.

Gould’s complaint may be, then, argumentatively modelled as requiring for the sought-for conclusion (the assertion of the hypothetical reconstruction) a conjunction of additional arguments (Marraud 2013a: 59–62) that, significantly, does not have to drop at all the initial one. Figure 21.1 represents such a conjunction of reasons justifying an evolutionary hypothesis, presented this time through a narrative, neither solely on the basis of its narrative coherence nor disregarding such coherence’s contribution in supporting the conclusion.

Figure 21.1 Conjunction of reasons justifying an evolutionary hypothesis

I have used similar diagrams in previous works (Olmos 2019; 2020a). They are based on Marraud’s (2016) interpretation and development of Toulmin’s model (1958). In addition to the representation of co-orientated reasons (signalled by the connective ‘besides’), they combine the use of justificatory reasons (allied with connective ‘so’) and explanatory reasons (allied with connective ‘that’s why’) and, whenever needed to clarify both kinds of inferential steps, either justificatory or explanatory warrants (in Toulmin’s sense) are provided in side boxes. Gould’s concern here is structurally similar to Ian Hacking’s criticism of the sufficiency of abduction alone, i.e., of the explanatory power of a hypothesis including an existential posit, to establish a ‘realist claim’ regarding the theoretical entity posited by it (Hacking 1983: 271–272). As I have shown elsewhere (Olmos 2018: 50), Hacking’s suggestion might be argumentatively modelled as requiring for the sought for conclusion (the assertion of the hypothesis) a conjunction of arguments, specifically including further experimental evidences amounting to the detection and manipulation of the posited entity.¹³

Gould also criticized the concentration on just one kind of causal mechanism (strictly understood natural selection in terms of adaptiveness is, as could be expected, Gould’s usual suspect) on which to base a historical account of a complex phenomenon. This would be a charge of unwarranted reductionism, that is rather more serious as it might prevent rather than encourage research along other lines and easily provide a sense of overconfidence in a particular kind of explanation, precisely for its neat identification of one well-delimited responsible mechanism.

Criterion (b) is Gould’s main epistemological point – although criterion (c) may be his main motivation. It is more a caution against selectionist reductionism (i.e., panselectionism) than any other thing, which is consistent with his well-known position in evolutionary biology (Gould and Lewontin 1979). The caution works this time as the conclusion of an a fortiori argument (Marraud 2013b) based on his own reservations with panselectionism in accounting for biological traits. We could reconstruct this a fortiori rationale behind Gould’s case (see Figure 21.2).

Figure 21.2 A fortiori rationale behind the charge of unwarranted reductionism

This reconstruction, chosen for its clarity at this point in the discussion, could be much more refined if we take into account that it really works meta-argumentatively (Marraud 2013b: 10–12) as a comparison between two reason-giving acts (two explanatory accounts) that are connected by a scalar topos of the kind ‘the more …, the more …’ acting as the warrant of the comparison. This topos, makes possible that the characteristics attributed to one of these reason-giving acts (its being insufficient or inadequate in a case) be transferred in an increased measure to the compared one on the basis of some condition that places them in different positions on the comparison scale.¹⁴

So the problem this time with sociobiology’s ‘art of storytelling’ is that it has overconfidently picked up a scientifically well-defined and well-understood mechanism (the ‘selection of beneficial traits’) and used it as a guide to reconstruct (allegedly too simplistically) the causal origins of some of the most complicated and intractable phenomena available. The just-so charge arises here as a charge against misguided scientificism, not against narrative science.

Gould finally warns us against a perceived as hasty use of theoretical results from natural science to support practical (even political) decisions. This would be a charge of prescriptive justificatory insufficiency that should be weighed in its own merits and according to pragmatic reasons modulated by considerations of risk (among other things).

The significance of criterion (c) (prescriptive insufficiency) can be understood (see Figure 21.3) as based on an additional a fortiori line of reasoning: ‘if strict selection alone is not enough to account for the actual social behaviour of human beings, how could it be enough to prescribe social policies?’ I have again to thank Professor Huss for reminding me of Gould’s membership of the Harvard left-wing group ‘Science for the People’ that held strong positions against the use of scientific results in justifying oppressing policies. This is consistent with my contention that criterion (c) was Gould’s main motivation in making his case against sociobiology.

Figure 21.3 A fortiori rationale behind the charge of prescriptive insufficiency

As already said, none of these hints at assessment criteria has really much to do with the story-telling quality of assumedly narrative models of explanation. After defending the scientific credentials of evolutionary psychology (see n. 13, above), Al-Shawaf asks himself ‘why do so many people persist in the notion that evolutionary psychological hypotheses are just-so stories?’ (Al-Shawaf 2019; cf. Al-Shawaf, Zreik and Buss 2018: 9). He attributes this mainly to the inescapable fact that evolutionary psychology (as astrophysics, cosmology and geology for that matter) has a central historical component and that historicity tends to be associated with untestability.

So it seems that the noun story easily attracts the charge just-so. Even if the historicity of the phenomenon addressed by many scientific disciplines is not a contested issue, the prejudice against storied accounts remains strong enough in many forums so as to extract protestations of scientific soundness in those addressed. Scholars interested in narrative science and narrative models of scientific justificatory practice feel, therefore, compelled to answer just-so charges even if those charges, when carefully examined, have not much to do with the particular characteristics of narratives and may even be based on just the opposite traits.

21.3 Defenders of Narrative Science Meet the Just-so Charge

Among the papers included in the 2017 special issue on ‘Narrative Science and Narrative Knowing’ of Studies in History and Philosophy of Science, edited by Mary S. Morgan and M. Norton Wise (Morgan and Wise 2017), it is significantly Sharon Crasnow’s (Crasnow 2017: 6–13) and Adrian Currie and Kim Sterelny’s (Currie and Sterelny 2017: 14–21) papers that make an explicit mention of the just-so story charge and try somehow to respond to or minimize it.

These two papers stand out in the collection as taking a meta-methodological approach while analysing their respective case studies. Both engage, particularly, in an epistemological appraisal of narratively dense and detailed accounts as opposed to certain efforts to base explanations regarding historically problematic phenomena (the 1898 Fashoda colonial incident, or the evolutionary development of human cooperation, respectively) in too restrictedly understood causal mechanisms or trajectories, amounting to formal models of explanation.

For these authors, narratives help better explore and understand the very causal relations expressed by those allegedly explanatory formal formulae, their contingent nature and the alternatives available at each historical turn. Both make reference to John Beatty’s (2016) ideas about what narratives are good for, namely dealing with contingencies, alternative possibilities and the particulars of historical turning points, that are still the focus of Beatty’s own contribution to the aforementioned special issue, in which he states that: ‘Narratives are about not only what actually happened, but also what might have’ (Beatty 2017: 31).

But what both papers finally depict in their case studies is not really a situation in which a narrative account of some phenomenon opposes a rival (in the sense of theoretically divergent) narrowly mechanistic account of the same phenomenon. The point is rather that certain kinds of identified or hypothesized causal links or mechanisms are better understood (explored and discussed) under a narrative rendering than under the crystallized mode of a formal formula or strict inference licence. So, as I will emphasize in the next section, the opposition (or comparison) is not so much mechanisms vs narratives but narratives vs formalizable laws.

Sharon Crasnow’s paper focuses, in particular, on a case of historical political science, a discipline that, in principle, already accepts its narrative nature. Nevertheless, recent philosophical discussions regarding the requirement for scientific explanations to be based on causal mechanisms, emphasizing, moreover, the use of individual case studies as devices for causal process tracing, tend to be read as bringing social scientific disciplines to a point in which narratives might be dissolved in favour of a bounded search for discrete pieces of evidence that allow the operation of such allegedly well-identified, well-delimited mechanisms.

This is what Crasnow calls the ‘inferential reading of process tracing’¹⁵ that she opposes to the enriched use of narrative accounts of those same mechanisms and processes. Her claim is that narratives in political science might, in fact, help tracing causal processes and identifying mechanisms in a better way than restricted inferential-type readings, precisely because narratives focus on exploring and discussing contingencies and alternatives (Beatty 2016) and, thus, help making credible and understandable the inner and detailed workings of the very causal connections involved. Contrariwise: ‘process tracing as a search for diagnostic pieces of evidence fails to capture the way that a mechanistic account seeks to address the inter-relationship of the parts in a way that narrative elements of case studies can’ (Crasnow 2017: 8).

Crasnow acknowledges that narrative approaches to science have been challenged with the just-so charge that she equates with the notion of biased cherry picking (i.e., suppressing disconfirming evidence or biasedly selecting confirming evidence). Her suggestion to avoid both the problem and the charge is finally to add substance, detail and discussion of alternatives to narrative accounts, making them, if anything, even more narrative:¹⁶

One worry often raised about the use of case studies is the idea that it may devolve to cherry picking or just-so stories. This is indeed a concern but one that can be addressed by requiring that all of the relevant details of the case be considered and not just those that are relevant to the favored hypothesis. In order to assure that these details are addressed, alternative hypotheses – different ways that the story could have gone, different paths that could have been taken, different mechanisms through which the case can be understood – need to be explored.

(Crasnow 2017: 10–11)

The four virtues that Crasnow ultimately associates with the narrative discussion of causal links or process tracing (i.e., closure, connectivity, elimination of alternatives and examination of counterfactual options) (Cranow 2017: 10–11) seem to be doing the work of avoiding too simplified accounts based on the biased selection of a restricted kind of evidence. This is more than consistent with Robert Richards’s (1992: 41–42) suggestion that narratives (as opposed to nomological models of explanation) are the adequate vehicle for ordering and weighing (downgrading and emphasizing) the contributions of possibly many different causal links that could be invoked to account for a historically situated event.

Currie and Sterelny (2017) conduct an even bolder and more committed defence of the benefits of scientific story-telling. Their recipe, though, for obtaining such benefits without incurring just-so charges is a bit different. First of all, they are prepared to defend speculation, not anymore a vice whenever it yields the appropriate kind of Lakatosian fruitfulness (Currie and Sterelny 2017: 16). It is, precisely, on account of the value attributed to such fruitfulness that they acknowledge that Gould and Lewontin were probably right (or at least consistent) when criticizing strictly adaptionist hypotheses, because, as I have already remarked, these may tend to prevent rather than encourage research along other lines:

Gould & Lewontin’s complaints about adaptationist reasoning is in part clarified by this distinction: the charge of ‘just-so’ storytelling is in effect the charge of idle speculation: adaptationist hypotheses fail to open new investigative routes and actively discourage them (here is not the place to consider whether such a charge is plausible).

(Currie and Sterelny 2017: 17 n. 11)

A second step in Currie and Sterelny’s defence regards coherence as an epistemic virtue. Mere internal coherence, so to say, might be insufficient – although not thereby negligible in this respect – to support a historical reconstruction of the causal web leading to an explanandum-phenomenon. But, insofar as such a reconstruction is pressed (by scientific method and community) to cohere with all kinds of constraints, issuing from material discoveries, other reconstructions, general theories, etc. such extended coherence becomes a noticeable achievement. This idea may be understood as amounting to appraising consilience as a kind of master scientific virtue (Weinstein 2009) or, alternatively, as demanding from us a sufficiently flexible, assumedly multifactorial and open-ended, model of scientific argumentative assessment (Olmos 2020b) in which the contributions of different strategies (some possibly more narrative than others) may be weighed and, at least to a certain point, harmonized.

This last idea is much in line with Currie and Sterelny’s final suggestion that their defence of scientific storytelling aims more at integration than substitution. The virtues and benefits of narrative approaches should combine with the virtues and benefits of formal models and the possible shortcomings of each of them be compensated by the other. And this is so because, as they try to show (although they do not express it with these words), some kind of just-so charge could be attributed to both. There might be just-so stories but there are also just-so formal explanatory models insofar as they unwarrantedly claim to be self-standing explanations.

This is what purportedly happens in their discussed example of a neatly modelled threshold-dependent explanation of the emergence of punishment in early human communities (sociobiology again). Such a clean, self-standing explanation, leaning on the pristine comprehensibility of the mechanism invoked is exposed as a just-so attempt, insofar as it is not taking into account enough contextual constraints as the emergence of other factors leading to human cooperation. What’s missing here (according to Currie and Sterelny) is a good integrative narrative that’s lost in the decompositional strategy of formalized (narrowly mechanistic) models:

Highly complex explananda like the evolution of human cooperation are resistant to approaches which depend solely on the decomposition and abstraction which enables modellers to probe aspects of constituent dynamics in isolation. For highly complex, multi-factorial, and multi-stage causal trajectories there are no master-models to be had, and so we must instead combine narratives and models, allowing us to navigate between the trade-offs generated by complexity.

(Currie and Sterelny 2017: 20)

No wonder that Currie and Sterelny come to agree with Gould and Lewontin’s complaints. Their just-so criticism, even if it was coupled with the noun story, was not directed towards the storied character of the accounts they criticized, but to their overconfident self-standing reliance on just one supposedly well-known and well-comprehended natural mechanism.

21.4 The Narrativity behind Nomicity

A somewhat different strategy to appraise narrative models of scientific explanation and justification is the one that exploits a kind of genealogical argument based on the idea that there’s a narrative (or at least a narrative kind of rationality; cf. Fisher 1989) behind (or before) each law-like generalization or nomological explanatory formula that – even if it may be rather opaque and disregarded in its current application as a validated theory – may always re-emerge when the formula comes under scrutiny as an explanatory principle.

The point here is not that there’s a story behind its establishment that may make it more understandable or even be part of its justificatory framework. These kinds of ideas would pertain to either the history of the discovery and acceptance of particular scientific laws and theories or more generally to what I have called the narrative account of scientific experimental and research activities (Olmos 2020a; cf. Meunier’s paper in this volume on ‘research narratives’, Chapter 12). In this sense, there are recent significant case studies of how scientists themselves use a narrative rendering of their interventions and experiments (e.g., Mary Terrall’s (2017) account of Réamur and Trembley’s ‘tales of quest and discovery’ or M. Norton Wise’s (2019) work on Faraday’s series). However, this is not what I specifically want to focus on here.

The claim I want to examine is rather that any scientific law-like generalization would somehow depict and delimit the scenario of its own validity and applicability as based on considerations regarding the possibilities of isolating natural phenomena and letting them develop in a controlled setting and making them solely dependent on a relevant set of variables. Such scenarios and the assumptions that make them plausible and assessable would be narrative in the sense of describing what can be expected of either a spontaneous or a more or less controlled course of events. Invoking and exposing them in their narrative detail would be just what’s needed whenever those generalizations, instead of being just applied, are discussed and weighed against alternative ones – which is something scientists involved in original research, as opposed to science teachers and appliers of scientific current theories, are expected to do.

Several authors have defended the interest of approaching and exposing such kind of narrative ground that, on the one hand, purportedly gives support to, and, on the other, is somewhat obscured by, scientific nomological formulae. I take Alirio Rosales’s (2017) comparison between Ronald A. Fisher and Sewall Wright’s mathematical solutions (i.e., nomological models) for certain problems of population genetics in terms of the diverse narratives that not only support them but give them meaning to be fairly understandable along these lines.

An even more theoretically committed contribution in this respect is Carlos López Beltrán’s (1998) paper, centred on the combination of narrative and statistical explanations in biology and medicine. As other philosophers interested in narrative science, López Beltrán starts with the factual assumption that certain specific scientific areas and practices (his focus is on medicine and biology) make an extensive use of narrative patterns of explanation for their very particular, unique and eventful explananda (a clinical case or the evolution of a particular trait). But his most thought-provoking point is that the statistical numerical models that these same disciplines also construe still reveal their narrative warp and woof, as issuing from data collection practices whose particulars are more than present in their final presentation and effective use. López Beltrán situates statistical models midway between the particularity of the unique case and the universality of classical nomological generalizations and does so by invoking a fourth intermediary state between the unique and the statistical in the clinical-case based on typicality.¹⁷

Thus, López Beltrán predicates the genealogical and conceptual continuity of narrative and statistic explanatory strategies – against their current alleged rivalry – and even places narrativity, especially narrative cognitive capacities, as grounding both:

The continuity between these two strategies I want to expose gives certain priority to the narrative one. I want to show that, not just historically but also conceptually, the efficiency of statistical procedures is based on the either explicit or implicit use of cognitive capacities associated to narrativity. That is, the use of statistics implies a (currently nearly always occult) narrative resource and makes the same kind of explanatory work. Both strategies try to establish more or less reliable connections between strictly unrepeatable singular events and the sought for syntheses and generalizations that motivate scientific research.

(López Beltrán 1998: 275; my translation)

A major reference for López Beltrán is Robert Richards’s (1992) seminal paper, so far probably the most radical defence of the ultimately narrative character of scientific explanation in general: ‘When the barriers are down, we will see, not that historical narrative fails as a scientific explanation, but that much of science succeeds only as historical narrative’ (Richards 1992: 40).

The idea of a generalized narrative approach to scientific practice that may be just temporarily and only very superficially circumvented by relevant simplifications is very present in Richards’s radical proposal. For Richards, the narrative quality of scientific explanatory practice would be, somehow, at the bottom of any explanatory attempt in such a way that it is only when making certain simplifications and taking certain methodological decisions that some disciplines just apparently and for a limited range of phenomena succeed in leaving their narrative nature behind:

[e]volutionists cannot make many predictions of consequence. I should add physicists are not logically better off; their projected systems are usually simpler and, as far as circumstances go, dead. But they cannot more accurately predict the exact trajectory of a falling leaf on a blustery Chicago day than Darwin could have divined the rise and evolutionary development of the HIV virus.

(Richards 1992: 36–37)

Richards placed his narrative approach to explanation in opposition to law-based explanatory models, but most especially to the attempt to understand less strict patterns under the epistemological dominance of the nomological model. Instead of considering such nomological models as the successful peak from which any degree of divergence would diminish the scientific quality of an account, Richards somehow maintains that keeping in touch with the narrative roots of our scientific explanatory attempts – instead of contemplating and appraising their skeleton-like yields – will in fact improve epistemological research.

My second claim goes further: it is that all explanations of events in time are ultimately narrative in structure. This means that Hempel got it just backwards: it is not that history can offer only explanation sketches, but that nomological-deductive accounts […] provide only narrative sketches; the covering law model yields sound explanations only insofar as that skeleton can be fleshed out imaginatively with the sinew and muscle of the corresponding narrative.

(Richards 1992: 23)

According to Richards, the problem with Hempel’s nomological model as well as other equally nomologically eager models is that they assume that currently valid law-like generalizations, first, lay ready at hand and, second, simply match as objective patterns the (pre-determined as) relevant facts of the explananda they allegedly cover. However, only in very limited, artificial, textbook-like situations (insofar as the isolation of the phenomenon is ascertained) this seems to be the case. Whenever we want to explain a real event in time the explanatory work will not really be done by any prearranged formal relations between selected antecedent conditions and matching laws, but precisely by the detailed investigation of the case that would, among other things, justify their use. And for that, according to Richards, we need narratives, narrative principles and narrative cognition.

López Beltrán’s claimed continuity between scientific explanatory methodologies, striving at different ranges of applicability, as based on their common ultimate narrative nature, finally becomes a plea for a reasonable and healthy combination of approaches (1998: 277–278) that is rather in line with Currie and Sterelny’s (2017: 20) integrative proposal. Such self-assumed explanatory pluralism (cf. Mantzavinos 2016) would avoid the downright dismissal of scientific explanations solely based on their form or mode of presentation and thus be more than compatible with a more nuanced and specifically argumentative approach to explanation discussion and assessment.

21.5 Conclusion

The just-so charge is a derogatory label, a negative assessment judgement that has been often misinterpreted and hastily attributed to explanatory attempts of a narrative nature on account of their form or discursive presentation through the catch-phrase just-so story. This is misleading and rather at odds with S. J. Gould’s original introduction of the concept within epistemological discussions.

However, the academic effectiveness of the label, working as a global flaw charge and preventing, in many cases, a more careful analysis of significant epistemological suggestions, has, in many cases, forced defenders of the relevance of narrative science to meet the challenge and try to respond to it.

In this chapter, I have analysed those responses that range from assuming the methodological benefits of narrative formats (or at least of the integration of narratives with other epistemological approaches) whenever the phenomena under scrutiny meet certain conditions to the bold postulation of the ultimate narrative nature of all explanatory endeavour.

These qualified defences of narrative science constitute a contribution to contemporary discussions on explanatory pluralism and, together with other suggestions, establish the possibility of analysing scientific reason-giving practices as primarily subject to the dialectical requirement of openness to collective survey and discussion rather than to aprioristic predetermined formulae precisely aiming at circumventing it. Nothing could be more just-so.¹⁸