Félida,doubled personality,and the ‘normal state’ in late 19th-century French psychology

Introduction

‘It must be natural’, declared Bordeaux medical professor Eugène Azam in 1876–77, ‘to think that perfection is exclusive to [l’apanage de] the normal state, because it has been said to me on all sides that I am mistaken, and that Félida’s normal state could only be the one characterized by completeness of memory’ (Azam, 1877a: 395). ¹ Azam (1822–99) wrote to defend his choice to describe Félida X, a woman with an apparently doubled personality, in terms of a normal state and a condition seconde—another, second, state of being. He conceded that he differed from his principal critic of the time, British philosopher-psychologist George Croom Robertson, only over ‘the interpretation of a word’; while Croom Robertson expected a subject’s normal state to consist of ‘perfect health’, Azam had opted for a convenient comparative term, ‘due to the absence of a more suitable word’ (ibid.: 396, 397). Yet, just as pre-publication suggestions from Azam’s philosophy mentor, Ernest Bersot, had not swayed his terminological choices (ibid.: 395), nor did Croom Robertson’s ‘serious objection’ prompt Azam to seek out a ‘more suitable word’ for Félida’s state. Rather, Azam held obstinately to his naming convention over the following two decades, as he updated and revised his observation of Félida in some 24 publications and communications. The ‘normal state’ and notions of normality thus played a part in the ‘dialogue and controversy’ surrounding Félida’s case, out of which, as Jacqueline Carroy has argued, ‘physicians and philosophers developed the common language of a physiological psychology and a psychopathology à la française’ (Carroy, 1992: 76).

This article takes Carroy’s claim literally, and thereby adopts a novel vantage point from which to explore the emergence of psychology as a scientific field of enquiry in France at the end of the 19th century. That is, I interrogate the lexical and conceptual place of the normal state in debates about Félida, and trace some ways that speaking about the normal articulates with the epistemological and theoretical concerns of emerging scientific psychology. Fine analysis of the normal state can firstly help elucidate the often messy dynamics of Azam’s multiple communications about Félida; certain early usages of the ‘normal’ persist in Azam’s accounts as fixed points around which his conceptual positions otherwise appear hesitant or fluid. Beyond Azam, configurations of the normal are bound up with changing interpretations of apparently doubled mental states and, through them, with evolutions in the objects of study of psychological research.

Reciprocally, this analysis adds a further historical context to the historically specific circumstances across which Peter Cryle and Elizabeth Stephens have tracked contestation and conceptual shifts in the normal. Their genealogy leads us to expect the normal of nascent French psychology to occupy an intermediate conceptual space between the undefined, comparative normal state of mid-19th-century physiology and the ‘normal person’ of early 20th-century psychoanalysis (Cryle and Stephens, 2017: Chapters 1, 7), as researchers like Azam brought their medical training to bear on problems of pathological psychology, and in turn passed a cultural, theoretical, and material legacy to Freud and others. ² Indeed, by following the work done by the normal state through accounts of doubled personality, I suggest an alternative lexical pathway linking these configurations of normality. My pathway, though necessarily brief, notably reasserts Pierre Janet’s place in early 20th-century psychological thought, whereas Cryle and Stephens’s lineage passes through German sexology and its American counterpart. ³ My route also ‘doubles’ over some episodes mobilized by Ian Hacking (1995) for his large claims about the importance of sciences of memory in making the modern individual. But I tread a narrower textual path, and follow ways that speaking of the normal connects to psychology’s objects and methods of enquiry, and their scientific status.

That Félida’s case was central to shaping both objects and methods of scientific psychology has been demonstrated by Carroy (in French-language publications), although it is known in anglophone scholarship chiefly as an early case of double personality (Hacking, 1995: Chapter 11; Lombardo and Foschi, 2003; cf. Hajek, 2020). ⁴ It was one of the most prominent and most debated cases of the fin de siècle, running through psychological enquiry even as the preoccupations of that science shifted. Over the six months after it was first disseminated widely, in May 1876, Azam’s observation of Félida attracted responses from philosophers and physicians, elicited reports of similar cases, was abstracted in the French scientific and political press, and, in English, was summarized, discussed, and then translated in full. ⁵ It allowed proponents of a scientific psychology grounded in positive principles to challenge official spiritualist doctrines around the unity of the self (moi; Carroy, 2001: 52–9; Hacking, 1995: 163–6). Then, as the 1880s and 1890s brought the ‘golden age’ of hypnotism, Azam was recuperated as a key precursor figure, as much for his observations of Félida as for his pioneering 1860 article on hypnotism (e.g. Dumontpallier, 1889: 24). ⁶ Félida remained a prominent point of reference through to the early 20th century, including in Pierre Janet’s developing notions of a mental economy (Carroy, 2001: 64–7; Pierre Janet, 1910).

My exploration of the psychological normal state, while focusing on the intricacies of Azam’s writing, also follows these ramifications: first backwards and then forwards in time from 1876. I begin by examining the ways Azam introduces the terms normal state and condition seconde in his text, and trace likely influences on his chosen terminology. His self-conscious terminological work centres on the second term in the pair, rather than on the unproblematically scientific normal state, as Azam, like his precursors, attempts to counter suspicion over the potentially illusory nature of extraordinary psychological phenomena. My second section teases out in detail the conceptual associations of normal state, as used with respect to Félida, and situates them among configurations of the normal in neighbouring disciplinary contexts. With Félida’s normal state an individualized state, worthy of detailed observation in its own right, we see a shift away from the generalized comparative normal of 19th-century physiology, but one that does not necessarily imply any prescriptive use of the normal as healthy ideal. Associations between the normal state and healthy activity emerge, rather, in exchanges with critics such as Croom Robertson. Examining Azam’s responses to such criticism in 1877–8 allows me to elucidate the reasoning behind the fixity of his choice to label Félida’s amnesic state as normal. Ultimately, this third section reveals the ways Azam must downplay the import of additional phenomena exhibited by Félida in order to retain his comparative, binary approach to mental states. My analysis shifts in the final two sections to trace state-naming conventions in a selection of other French work on double personality from the mid-1870s to the 1910s, for which Félida’s case constitutes a significant point of comparison. In asking what became of Azam’s normal state and condition seconde, I sketch a transition in French psychological thinking from the normal state as individualized, fixed, and comparative, through increasing pairing of the normal and abnormal, to understandings of pathological characteristics as extended along a continuum bounded by the ‘normal individual’.

Before moving on, it will be helpful to evoke the messy publication history of Azam’s observation, as a guide to the sources analysed here. Azam’s major observation of Félida circulated widely in May 1876 in the pages of the Revue scientifique (Azam, 1876a), following presentations to the Académie des sciences morales et politiques that led to an official publication in the Séances et travaux de l’Académie des Sciences morales et politiques in the second half of 1877 (Azam, 1877a). But other overlapping texts also appeared in 1876–7, including verbal communications to learned societies in Bordeaux and Paris (Azam, 1876c) and accounts in provincial medical journals (Azam, 1876e). Large sections are repeated verbatim between different versions as published (see Hajek, 2020), yet there are also variations, updates, and revisions, while circular references to priority mean that no version can be counted as a definitive original. That being the case, I take the official Academic publication of Azam’s observation as my key source, and turn occasionally to other versions where relevant.

Labelling extraordinary phenomena

When Azam came to write up his observation of Félida for publication, he was conscious both of stepping into a new disciplinary space, as a physician discussing ‘an observation which belongs more to psychology than to medicine’ (Azam, 1876a: 481), and of the lexical difficulties that might entail. ‘The words of which I make use being ordinary words that I have been compelled to turn from their [usual] sense, some obscurity necessarily results’, he apologized when introducing his communication to the Académie des sciences morales et politiques (Azam, 1877a: 363). But if Azam acknowledged that he called one of Félida’s states normal ‘for want of a better word’, this was only a later response to Croom Robertson’s objections (ibid.: 396); Azam provided no particular clarification or explanation when he introduced the term in the observation proper. Normal state appears initially as a routine comparative term allowing Azam to situate other significant terms present in his observation: the unique, and confusingly relativist, terminology employed by Félida herself. ‘She has always held that the state, whichever it is, in which she is at the moment of speaking to her is the normal state, which she calls her raison [reason], by opposition to the other state that she calls her crise [fit or attack]’ (ibid.: 366). ⁷ It is Félida’s term raison that Azam signals to his readers as worthy of attention, not his own label for the state. Normal state is lexically available and unremarkable for Azam, in the same ways as in mid-century physiological discourses (Cryle and Stephens, 2017: 49–59). What is more notable is that Azam transfers such routine use of the normal into the psychological domain, in the process assuming it to be accessible to a wider, non-medical audience. By the mid 1870s, this was a fairly safe assumption, thanks to the scientific renown of Claude Bernard, whose writings on experimental method (Bernard, 1865)—grounded in his own physiological research—‘keenly interested many readers beyond the circle of savants’, making not only Bernard but physiology itself ‘almost popular currency’ (presque populaire; Patin, 1869). As Bernard and physiology entered cultivated discourse (and the Académie française), so presumably did the key physiological binary of the normal and the pathological.

That normal state was a term in routine usage contrasts markedly with Azam’s concern to introduce and define his appellation for Félida’s other state, ‘the second state that is conventionally named condition seconde’ (Azam, 1877a: 364). Although Azam presents condition seconde as ‘a term borrowed from philosophy’ when relating his observation to a predominantly medical audience (Azam, 1876c: 12), the naming convention has sufficient significance and novelty for Azam also to signal it to his academic audience—composed more of philosophers than physicians—and to gloss the term sporadically throughout the remainder of the piece, principally as ‘l’état d’accès ou de condition seconde’ (Azam, 1877a: 374, 375, 379, 385, 391). Accès can be translated as ‘fit’ or ‘sudden episode’, and was in widespread medical use to designate pathological states that manifested suddenly, often at regular intervals, like the attacks of hysteria or epilepsy. ⁸ A frequent synonym for condition seconde in Azam’s text, it helps anchor the significant term in the binary of normal state and condition seconde.

Condition seconde is no neologism or arbitrary usage on Azam’s part, however, but rather a careful choice that places Félida’s case within a series of scientific observations of apparently extraordinary neuroses. In particular, Azam deliberately and scrupulously follows the lead of Belgian physician Évariste Warlomont in lexical choice and presentation, drawing on Warlomont’s 1875 report on Louise Lateau, the famous Belgian stigmatic (Carroy, 2001: 55). Warlomont led a commission examining Lateau’s stigmata, ecstatic periods, and prolonged fasting as part of prolonged debates on these phenomena at the Académie royale de médecine de Belgique. ⁹ In the report, Lateau’s normal state appears as an uninterrogated comparison case alongside a condition seconde, which term Warlomont introduces self-consciously as a way to group together a range of altered states, including those capable of explicating Lateau’s phenomena in natural terms. He marks out the term by means of italics and quotation marks on its first appearance in the text, and proceeds to sketch its contours in a long exploration of different conditions—from brain injuries to hypnotism—that might produce ‘doublement de la vie’ (doubling of life; Warlomont, 1875: 90–114). Persistent quotation marks and the use of pleonasm—adding ‘state of’ before condition seconde—continue to signal this appellation as noteworthy, a new term of art that should not be taken for granted by his readers, in contrast to the familiarity of normal state.

What seems to appeal to Azam is the association developed by Warlomont between condition seconde and natural explanations for apparently extraordinary phenomena. Indeed, it was only after reading Warlomont’s report on Lateau that Azam felt authorized to publish his account of Félida’s case, some 16 years after first observing her condition (Azam, 1877a: 374). He was initially called upon to examine Félida in 1858–9, at which point she lapsed occasionally from her ‘ordinary state’ into a sort of ‘second life’, characterized by a pronounced change in bearing (allure) or personality, as well as alterations in her memory (ibid.: 364–7). But faced with widespread doubt from colleagues over the reality of these phenomena, Azam was dissuaded from publishing the case, though he did mention Félida in passing in his 1860 article on hypnotism (Azam, 1860; 1877a: 370–1). ¹⁰ Once Warlomont had discussed phenomena similar to Félida’s two lives in a scientifically sanctioned arena, Azam sought out Félida to supplement his 1850s observations for publication (Azam, 1877a: 374). He then framed Félida’s changes of bearing and memory in Warlomont’s terms, the already scientific stamp of condition seconde making it a ‘better word’—a likely replacement in Azam’s notes for earlier alternatives such as ‘second life’.

As for the term itself, whether coined by Warlomont or not, condition seconde represents a shift from terminological binaries found in contemporaneous accounts of altered states, including those that Warlomont appropriated to circumscribe his conceptions of doubled life and the condition seconde. ¹¹ To illustrate what he called the first form of doubling—which followed traumatic brain lesions—Warlomont reproduced an observation reported by Ernest Mesnet in 1874 and designated its subject as having entered a condition seconde. Mesnet himself made no reference to a condition seconde, however, chiefly labelling the second state as a crise (Mesnet, 1874). ¹² This sat on several occasions in a binary pair with the normal state and normal phase, among a number of other terms (ibid.: 106, 107, 111–12). Warlomont’s condition seconde thus replaced Mesnet’s crise. Now crise was also the word Félida used to describe one of her states, a choice read by Jacqueline Carroy as reflecting the term’s availability in popular milieux, as a result of the cultural pervasiveness of discourses surrounding magnétisme animal (animal magnetism/mesmerism). For if crise originally described the fit-like manifestations of Mesmer’s patients, it remained in use throughout the 19th century in relation to the gentler, somnambulism-like states more common to magnetized subjects (Carroy, 2001: 46–7). ¹³ By always glossing Félida’s crise as the more technical accès, Azam obscured possible links to magnétisme and protected his observation against potential contamination by magnétisme’s reputation for apparently supernatural explanations and fraudulent phenomena. The observation of Félida thus became exactly the kind of ‘pure fact’ needed to combat spiritualist doctrine around the unity of the self (ibid.: 62–4). ¹⁴ I contend that the same kind of ‘purification’ was in play in Warlomont’s and Azam’s choice of the self-consciously innovative condition seconde. Although normal state passed unmarked and uninterrogated, its strongly scientific connotations—associated with the eminence of Bernard’s experimental physiology—reinforced the legitimacy of both terms in the pair, and consequently emphasized the distance of double-state phenomena from the ‘regrettable [fâcheuse] promiscuity of charlatanism’ (Azam, 1893: 39).

Félida’s normal state and ordinary life

When Azam imported the scientific status of the physiological normal state into his psychological observation, he also adopted the broad comparative logic that underpinned uses of the term in the context of mid-century physiology examined by Cryle and Stephens. Just as the physiological normal state stood for the unaltered state and served to set off the ‘undisturbed functions of life’ from pathological phenomena (Cryle and Stephens, 2017: 55), Azam used normal state as a basis state against which to examine temporally circumscribed changes in Félida’s personality and mental functioning:

In the fit-like episode or condition seconde, she is more haughty [fière], more insouciant, more preoccupied by her appearance; moreover she is less industrious, but much more sensitive [sensible]; it seems that in this state she gives greater affection to those who surround her.

These differences with the normal state…(Azam, 1877a: 379)

It was particularly important for Azam’s study that he hold fixed the identity of the normal state against the condition seconde—that his terms be relational, not relativist—because Félida’s normal state was not an undifferentiated basis state. ¹⁶ Rather, it was an object of enquiry in its own right, even if this enquiry was always for the ultimate purpose of comparison with the condition seconde. Azam needed to examine both states given the kinds of change he wanted to characterize: changes in personality and in memory. Let us begin with personality, the ‘doubling of life’ (dé/doublement de la vie) in the titles of Azam’s publications. As the term life suggests, and as we see in the above quotation, what changed with Félida’s personality was her whole ‘manière d’être’, ‘her character and her affective sentiments’, and, in turn, her broad patterns of behaviour (Azam, 1877a: 376, 386). By definition, Félida’s personality was individual, or personal, to her; the normal state was then her normal state, and was marked as such by possessive pronouns in the text (e.g. ibid.: 370, 386). It followed that Azam had to determine the characteristics of Félida’s normal state, or base personality, in order to identify the changes that occurred when she entered her condition seconde. Azam indeed recognized this requirement, and enunciated it in his text. ‘Making the most of an occasion perhaps difficult to meet again, I study her with care’, he wrote of an occasion in July 1875 when he had encountered Félida in her normal state, which by that time had become highly infrequent (ibid.: 380). The normal state, in this particular psychological context, was therefore an individualized concept. If physicians like Azam, Warlomont, or Mesnet all invoked their subjects’ normal states, the only thing these states shared was a common relational logic; there was no sense of a qualitative generalized normality against which different subjects’ conditions could be examined. The physiological sense of the normal as a range of generalized healthy functionality appears exceedingly rarely in Azam’s texts, and only in his theoretical reflections (ibid.: 401) and later updates (Azam, 1877b: 578).

Similarly absent from the observations is a quantitative or frequentist dimension to the normal state. Azam did employ ordinary as a synonym for normal in the initial pages of his account, and even contrasted Félida’s condition seconde to her ‘ordinary state’ when first introducing the former term (Azam, 1877a: 364). He switched to using ‘normal state’ several pages later. Ordinary, in these early pages, carried connotations of frequency in time, of what was ‘habitual for Félida’, not of tallying qualities across different subjects. It was readily interchangeable with normal, because Félida spent most of her time in her normal state during the late 1850s, with the condition seconde lasting only several hours per day (ibid.: 367). By 1875, however, the situation had reversed, and the condition seconde was now Félida’s predominant state. In order to retain his fixed comparative binary, Azam thus had to break the conceptual link between the normal and the ordinary: ‘Today, the condition seconde is, so to speak, the ordinary life because it lasts three and four months in a row, against periods of normal life which only have a duration of three or four hours’ (ibid.: 391). It was a distinction he would maintain in subsequent writings on Félida, retaining normal as a comparative term for one particular, individualized state, and using the conceptually looser ordinary for habitual comportments and functioning, whether individual to Félida (as here) or common to people in general (as discussed below). ¹⁷

Besides being individualized, the psychological normal state was also a concept embracing the whole person, even her whole ‘life’. Félida’s normal state, as delineated by Azam, encompassed physiological functioning (‘her [sense of] taste, in the normal state, is destroyed’), character and bearing (‘naturally serious and sad’), sentiments (‘Félida is indifferent and shows little affection for those who surround her’), and memory (‘having become pregnant during her condition seconde, she was unaware of it [elle l’ignorait]…during her normal state’; Azam, 1877a: 368, 386, 381, 370). ¹⁸ Her condition seconde was, of course, similarly richly characterized. In its richness, the psychological normal state participated in the historical shift to a more holistic and complex concept of the normal that Cryle and Stephens (2017: 266–79) find in fin-de-siècle sexology and Freudian psychoanalysis.

We might then ask whether there was also a slide into a prescriptive notion of the normal state; in effect, did Azam mark certain personality traits or comportments of Félida’s normal state as more natural, or connect them to socially expected roles for lower-class women? There was perhaps a slight normative tone when Azam mentioned Félida’s industriousness in her normal state—she was a ‘good worker’ (Azam, 1877a: 364)—or rather her lack of it in the condition seconde, in which she was ‘less industrious’ and ‘more frivolous’, with Azam on one occasion associating ‘frivolity’ and ‘triviality’ (futilités; ibid.: 379, 387, 386). ¹⁹ For the very most part, however, the case reveals no clear normative position. Each state exhibited both positive and negative qualities; it was essentially the existence of two states that was problematic, such that it would have counted as a cure were Félida’s condition to have stabilized in either one of them. Félida’s husband hoped she was cured when she remained in the normal state for many months (during the hiatus in Azam’s observations from 1859 to 1875), while the rarity of the normal state in 1876 led Azam to propose that it would be ‘a sort of cure’ were the condition seconde to take over her entire life (ibid.: 377, 393).

In neither state, moreover, did Félida’s personality fall outside of the range of ordinary human behaviour:

If [in the condition seconde] she appears more attached to those who surround her, it is only in comparison with the way she is in the normal state, because, [and] I must insist on this, [nothing] we can observe in her on this point exceeds the ordinary. (Azam, 1877a: 387)

The normal state, in Azam’s texts, was distinguished by one particular characteristic, and one that was defined comparatively: the condition of Félida’s memory. During her condition seconde, Félida could remember ‘not only what has happened in previous episodes [of condition seconde], but also all her normal life, whereas…during her normal life she has no memory of what has happened in her [condition seconde]’ (Azam, 1877a: 367). The directionality of this ‘periodical amnesia’ provided another reason for Azam to study the normal state in its own right: it was only in the normal state that Félida’s ‘alterations of memory’ could be discerned. On the advice of various philosophers interested in the case (notably Victor Egger), Azam accordingly attempted to circumscribe Félida’s amnesia: he determined that it did not extend to ‘general ideas’, such as being able to read or write, but could reach no conclusion about whether habits acquired entirely in the condition seconde would be retained in the normal state (ibid.: 368, 411–12). Although the completeness of her memory made Félida’s ‘second life…by far superior to the other’ (ibid.: 367), Azam did not take that to mean that the condition seconde should properly be labelled as normal. Rather, as I have shown, he reserved the term normal as a comparative marker. And if it was Félida’s partial amnesia that most clearly characterized her normal state, the state nonetheless bore on a whole set of functions displayed by Félida as a particular individual, almost all of which fell within the conceptual scope of ordinary behaviour. As hypothesized, the psychological normal thus combined conceptual features from mid-19th-century physiology—binary comparison—with those of early 20th-century sexology and psychoanalysis—the normal as relating to a whole person. Yet there was a tension in Azam’s normal state regarding its associations with general notions of health, since Félida’s apparently complete and ordinary life was undergirded by an unquestionable diagnosis of hysteria, and very occasionally punctuated by a third, pathological state.

How can a hysteric have a normal state?

This returns us to Croom Robertson’s objections, with which I began this paper. Reporting on Azam’s observation in Mind, which he edited at the time, Croom Robertson took the opportunity to question Azam’s choice of normal to describe one of Félida’s states. For him, Azam was ‘disposed rather to make light of the want of memory in the “normal” state’, such that ‘the state which he calls “normal” becomes clearly a morbid one’. Or better, considering Félida’s pronounced physical characteristics of hysteria, it should be considered ‘that both states are more or less morbid’ (Croom Robertson, 1867b: 415). As Azam (1877a: 390) recognized in his response, the binary opposition in play in such objections was between a functionally healthy state applicable across humankind, ‘the complete state, the state of reason’, and a pathological state, ‘the sickly state’ (l’état maladif). ²⁰ And Azam was prepared to admit that both Félida’s states were largely morbid in nature, if considered in such terms (ibid.: 397). After all, when not using her name, he already referred to Félida as a sick person, as ‘ma malade’ or ‘notre malade’ (e.g. ibid.: 374–8), just as he made it eminently clear that Félida’s normal state was marked by pains and physical symptoms of hysteria, and that these came to invade her condition seconde in 1876–7 (e.g. ibid.: 409–10). Even Félida’s life before the onset of the doubling ‘must never have been perfect health’, he conceded (ibid.: 397, original emphasis).

Where for Croom Robertson the normal and the (perfectly) healthy coincided, they did not do so in Azam’s observation, as we have already seen. His normal state was entirely comparative, and the choice of ‘normal, for want of a better word’, fell upon the state that ‘most resembles [Félida’s] anterior life’ (Azam, 1877a: 397). The most appropriate ‘normal’, in other words, was the ultimate basis state preceding and perturbed by the onset of Félida’s periods of second personality. What was more, it encompassed a whole state of being, rather than any single function or faculty. Thus Azam was equally unwilling to recast his comparative binary on the basis of the (admittedly key) faculty of memory, unlike the editors of the Journal of Mental and Nervous Disease, for whom, from the perspective of ‘the defect of memory,…the “second” condition would appear to be the most nearly normal of the two’ (Azam, 1876d: 610–11).

Nonetheless, Azam did engage with his critics, and in the process—as in his observation—clung tenaciously to two related aspects of his terminological conception: his choice of which state to label as normal, and a more fundamental framing of Félida’s condition as a binary pair of states. The first he defended actively in updates to Félida’s case, responding to his critics on their chosen ground of healthiness. The second remained a matter of implicit consensus between Azam and his interlocutors; we can trace its influence through both Azam’s defensive work and the ways he construed Félida’s ‘third state’. Azam’s rebuttals proceeded by first re-emphasizing a key element from medico-physiological usage of the normal state: its anteriority, as a basis state against which to describe perturbations. In Félida’s case, the relevant perturbation was the appearance of the condition seconde, which led Azam to affirm the identity of Félida’s early life and her normal state (as observed by him). The normal state, Azam insisted, ‘is the exact representation, the continuation of Félida’s mode of existence up to the age of fifteen, that is of her ordinary life up to the observation of the illness’ (Azam, 1876b: 265; see also Azam, 1877a: 386; 1878: 195). In other words, the very first time Félida had entered the condition seconde, she had returned afterwards to her anterior state, and not to some other state of being or mind (Azam, 1877b: 579). Then, since Félida had experienced no apparent problems of memory in her early life—it was, after all, the onset of the condition that had prompted Azam’s intervention—it would not be ‘natural’ to locate the morbid processes of her amnesia in the normal state (ibid.). Azam’s normal state, initially configured as primarily comparative, was reattached in these responses to a certain sense of health, or at least of comparatively less morbidity than some other state.

But which state? Azam’s first set of theoretical speculations situated Félida’s alteration of memory in the condition seconde. ‘Forgetting’, he stressed, ‘is not necessarily provoked by an incomplete or sickly intellectual state at the moment when one tries to remember’ (Azam, 1877a: 390), as the amnesia that followed typhoid fever amply demonstrated (ibid.: 391, 397–8; Azam, 1877b: 579). Rather, remembering occurred in two stages: first ‘impression’, then ‘reproduction’—it could be the recording stage that occurred imperfectly, with events in the condition seconde not making a sufficient impression on Félida’s brain (Azam, 1877a: 390–1). But faced with the evident inconsistency between this hypothesis and his observation that Félida remembered events of one condition seconde during later periods of the same state, he subsequently shifted the defect of memory onto the short transition period between states. It would then be while transitioning back to the normal state that Félida would lose the capacity to reproduce impressions recorded during the preceding period of condition seconde (Azam, 1877b: 579–80). While, on one hand, this logic shifted the morbid burden away from Félida’s two states, on the other hand, it opened questions about the importance of the transition period and its capacity to be considered a state in its own right.

That potential received scant endorsement in Azam’s writing, even as he came to realize, ‘better than at the start of my observations’, the ‘considerable importance’ of this period (Azam, 1877a: 411), notably in explaining Félida’s amnesias—‘the initial phenomenon that sets in motion all the others’ (Azam, 1877b: 579). Azam repeatedly discussed the transition in descriptions of Félida’s changes in state. ‘A spontaneous phenomenon’, its ‘principal characteristic: the loss of consciousness’, meant it could be considered analogous to the petit mal attack of epilepsy (Azam, 1877a: 372, 382, 411; see also Azam, 1877b: 578–9). Yet he consistently labelled it a ‘period of transition’ (emphasis added), marking it as conceptually distinct from, and of secondary value to, the key ‘states’ manifested by Félida. This is especially evident compared to the way Azam referred to the analogous ‘state’ of petit mal but switched terms to ‘these attacks [accès]’ or ‘these periods’ when noting that both petit mal and the transition were of short duration (Azam, 1877a: 411). Even on the one occasion where the transition was grouped among ‘all of Félida’s states’, Azam soon clarified that it was only ‘un état…surajouté’, an auxiliary to what counted as states proper (such as sleep, in this instance; Azam, 1877b: 578–9). The very short duration of the transition period—‘almost imperceptible [insaisissable]’ by the 1870s (Azam, 1877a: 379, 411)—may account for its adjunct ontological status. Phenomena unable to be grasped during their occurrence were clearly distinct from the extended ‘life’ periods studied by Azam under a framework of normal state and condition seconde.

Something more was in play here, I propose, for Azam also characterized a rare ‘third state’ of Félida’s as similarly ‘surajouté’ to one of the fundamental states, added on in this case to the condition seconde (Azam, 1877a: 413). That is, he understood Félida’s condition as fundamentally binary, composed of the two states that could be likened to ordinary ‘lives’, and consequently minimized the import of other groupings of mental phenomena, themselves more clearly pathological. And this even if Félida exhibited these phenomena frequently or their properties remained stable from one occurrence to the next. Thus, while the normal state and condition seconde were incontrovertibly personal to Félida, and often introduced by the possessive ‘her’, the third state was only ever ‘a’ third state, the indefinite article serving to distance it conceptually from the other two. He might class it unproblematically as a ‘state’ or ‘condition’, but on its first appearance in the observation, Azam was more concerned to categorize the third state as ancillary—‘only an epiphenomenon of the attack [accès]’—than to describe its key attributes (ibid.: 368). Strikingly, of all Félida’s states, the third was unambiguously pathological in nature. It was a state ‘which resembles an attack of madness [un accès d’aliénation mentale]’, specified Azam, prompted by its more frequent manifestations in 1877 to detail it at (somewhat) greater length (ibid.: 413). Adding to a previously mentioned ‘unspeakable terror’ (ibid.: 368), Azam portrayed a Félida prey to ‘terrifying hallucinations’ full of phantoms and blood (égorgements), who recognized no one other than her husband (ibid.: 413; similarly Azam, 1877b: 579). In 1877, the third state lasted up to three hours, similar in duration to the condition seconde back in 1858, though it returned only every few months, as against most days (Azam, 1877a: 367, 413). This was no third ‘life’, in Azam’s conception, but ‘only a sort of preface or annexe’ to the condition seconde (Azam, 1877b: 579), an ‘accessory’ or ‘epiphenomenon’ (Azam, 1878: 195) to that fundamental state. His justification here was that Félida entered the third state from the condition seconde, and then returned to that same state once her terrors had ended (Azam, 1877a: 413). ²¹ Azam displayed no uncertainty as his observation structured Félida’s condition around a binary logic, even as he presented himself as hesitant to ‘seek an explanation’ for the third state (ibid.), or indeed much else in Félida’s case (see Hajek, 2020: 103–4).

Abnormal second states

In these two final sections, I trace briefly what became of Azam’s normal state, its conceptual associations, and the binary framework in which it was embedded. Space not permitting any exhaustive survey, my chain of papers follows observations of subjects claimed as comparable to Félida: two further cases of double personality from the mid 1870s (Bouchut, 1877a, 1877b; Dufay, 1876), Jules Janet’s 1880s investigations of using hypnotism to produce ‘artificial’ doubled states (J. Janet, 1888, 1889), and Pierre Janet’s 1910 reappraisal of his brother’s ‘artificial Félida’. All these works looked back to Félida’s case, for comparison, if not reinterpretation, such that they form a coherent set against which to test Carroy’s claim for the influence of Azam’s language, and from which to sketch the conceptual landscape of the normal in emerging French psychology.

As Azam’s observation of Félida became widely known, other French physicians were prompted to share accounts of analogous cases from their own practice. What might otherwise have seemed extraordinary or anomalous manifestations became authorized contributions to physiological psychology, ‘facts’ collected with a view to eventually elucidating pressing psycho-philosophical questions (Bouchut, 1877b: 414; Dufay, 1876: 71). Two physicians who (re)framed old observations in explicit relation to Félida and double personality were Charles Dufay (b. 1815), a physician turned senator, and paediatric specialist Eugène Bouchut (1818–91). Bouchut (1877a: 282), writing in the Lancette française, thus disseminated two observations of hysteric children exhibiting ‘the appearance of a second life constituting the splitting [dédoublement] of life, as M. Azam would say, or if you like, splitting [dédoublement] of the personality’. ²²

Medically trained like Azam, Bouchut and Dufay followed their Bordelais colleague in making routine use of normal state as an appropriate comparative label for their subjects’ ‘first’ or ‘ordinary’ life. They did not, however, repeat Azam’s careful naming of the altered state, neither reproducing the term condition seconde nor self-consciously proposing some other term to describe the state. Instead, they drew on a range of already accessible labels, including the medico-psychiatric accès (Dufay, 1876: 69) or the more popular crise (Bouchut, 1877a: 281, 282). Alongside these general terms, both physicians enlisted somnambulism to designate their subjects’ second states, in a move that directly invoked a certain interpretative framework. It associated the second states with nocturnal somnambulism, in the first instance—perhaps a natural step to take in relation to patients who also displayed nocturnal episodes, though not one that was cause for any reflection on Dufay’s or Bouchut’s part; Bouchut simply added the adjective diurnal to mark the distinction between the two (ibid.: 282, 283). But beyond its nocturnal form—starting to become known as ‘natural somnambulism’ around this time—somnambulism also connoted phenomena of hypnotism (or somnambulisme provoqué; artificial somnambulism) and its scientifically dubious cognate practice, magnétisme animal, whose (purportedly) clairvoyant consultants were usually known as somnambules (somnambulists; see Richet, 1875: 348, note 1, 369, 371). Connections to somnambulism were certainly present in Azam’s writing; somnambulism in all its varieties fell under the set of manifestations covered by Warlomont’s condition seconde, and Azam drew upon them as useful analogies for relating Félida’s case to existing psycho-physiological knowledge (esp. Azam, 1877a: 372, 384–5; see also Hajek, 2020: 103–4). What changed with Dufay and Bouchut was the conceptual immediacy of their largely carefree approach to terminology. This contrasted with Warlomont’s and Azam’s careful terminological work, with its aim of weakening problematic links between doubled states and the (perceived) illusions or fraud of magnétisme. Azam’s successors appear to have been much less exercised by the scientific status of their observations, perhaps precisely because Azam’s (and Warlomont’s) study had already validated double-state phenomena, including through its naming conventions. ²³ Rather, the reference to somnambulism made explicit the analogies linking second states (in doubling cases) to a broad set of other phenomena investigated through physiological psychology.

It additionally elicited associations with the pathological; Charles Richet (1875: 348, note 1) straightforwardly glossed somnambulism as ‘this neuropathy’ in his 1875 article on hypnotism, for instance. Accordingly, we find occasions when both Dufay and Bouchut opted to pair normal with pathological or abnormal when comparing their subjects’ two states. Although such occasions were relatively rare, to evoke the abnormal or pathological alongside the normal was to shift the conceptual status of the normal state in these psychological observations. Another position correspondingly opened up in the dispute between Azam and his anglophone critics, as enunciated by Dufay as he made use of the term abnormal:

In [the subject RL] like in [Félida], the amnesia belongs to the normal state, to the physiological state—to forget a dream after waking is completely normal—and not to the abnormal or pathological state, since, on the contrary, during the accès, the memory is double. (Dufay, 1876: 70)

Bouchut, too, associated the normal with generalized healthiness when he characterized his second subject (a hysterical girl) as having ‘two consciousnesses and two memories, one traditional, normal,…and the other exceptional/incidental [accidentelle], pathological’ (Bouchut, 1877a: 283). In contrast to Dufay, however, Bouchut was more concerned to mark the second state as morbid or pathological than to advocate for the relative superiority of the normal state, or the appropriateness of normal as a descriptor. He referred at times to the second state as ‘this morbid state’ or ‘a real sickly state’ (ibid.). Without employing the word abnormal, Bouchut more consistently labelled the second state as an unhealthy exception from the usual principles governing human physiological or psychological function, hence his use of traditional as a synonym for normal, and of accidentelle as linked to pathological. Accidentelle took its philosophical sense in this instance, denoting some characteristic that did not comprise a necessary part of a subject. ²⁴ For Bouchut argued that the doubled consciousness of his subjects, as pathological, was also an exception to the ‘fundamental principles of psychology’ (ibid.: 281); his theoretical concern here was to protect the doctrine of the unity of the self. Since this, in turn, ‘forms the basis of human morality and responsibility’, Bouchut (1877b: 415) thus introduced something of a normative inflection to the normal state. Like the shift to conceiving the normal in general terms of health, this normativity was not consistent or dominant in Bouchut’s text. It was only implied, since Bouchut the physician also recognized the scientific interest of pathology, even if he denied its wider applicability to humanity as a whole. Similarly, Dufay retained Azam’s interest in the individualized characteristics of his subject’s normal state as a whole state of being. What Dufay’s and Bouchut’s language reveals, then, is a gradual conceptual broadening away from a normal state defined in relation to an individual subject’s life, to a normal measured against healthy human functioning—one present more subtly in Azam’s defence of his terminological choices. If, in a sense, this was a return to ways of conceiving the normal prevalent in physiology, the interest directed by physiological psychology to the subject’s whole ‘life’ points towards more prescriptive understandings of the normal person that would emerge in 20th-century America (Cryle and Stephens, 2017: Chapters 7–8).

An ‘artificial Félida’ and the normal individual

This brings me to the case history of a woman dubbed ‘an artificial Félida’, in which we find a further imbrication of hypnotism with observations of double personality, another twist in the dispute over the healthiness of the normal state, and references to the ‘normal individual’ predominating over those to the normal state. The subject in question, known as Marceline, was studied first by Jules Janet (1889) during his time at the Pitié Hospital in Paris, before being passed to the observation and treatment of his brother Pierre (1910). When Jules Janet encountered Marceline, she was severely ill with anorexia and dysphagia (difficulty swallowing), in addition to a range of other physiological troubles, all covered by the broad diagnosis of hysteria. With the scientific and cultural prestige of hypnotism then at its height, Jules hypnotized Marceline, and found that she could and would eat without difficulty when placed in a ‘somnambulic phase’ of hypnotism (J. Janet, 1889: 473). Initially, Marceline spent only limited time in somnambulism, but after various complications, Janet discontinued his practice of ‘waking’ her, and left her to live in the second state for months at a time. ‘All in all’, he declared in his 1889 article, ‘I have created in this woman a double existence absolutely analogous to that which Dr Azam’s Félida presented naturally’ (ibid.: 475). Jules’s tag ‘artificial Félida’ (ibid.) would be echoed, albeit somewhat ironically, by Pierre in 1910, in the title of his detailed case history of Marceline. ²⁵

Reading the Janets’ articles for their configuration of the normal, it is immediately striking that the term normal state does not appear, apart from one reference to Félida in Pierre Janet’s paper. Jules Janet (1889: 474) instead framed Marceline’s two existences in terms of various interlinked terminological pairs denoting states of hypnotism: ‘(hypnotic) sleep’, ‘waking’ (la veille), ‘somnambulism’, ‘waking state’. Somnambulism, like Jules’s ‘somnambulic phase’ above, almost certainly referred to a (deeper) substate of hypnotism, along the lines of the three-state model of hypnotism developed by Jean-Martin Charcot at the Salpêtrière (see Binet and Féré, 1887: Chapter 6). These (sub)states, or the state of hypnotism considered as a whole, were a matter for detailed investigation by researchers, while the opposite terms in the binary, like waking state, remained unexamined and simply stood as qualitatively distinct. Like the normal state of physiology, they served a comparative function, and demarcated the subject’s undisturbed state from the altered state induced by the manoeuvres characteristic of hypnotism. Indeed, in the wake of Azam’s observation of Félida, normal state also had a routine presence in hypnotism discourses, whereas before 1875, the term had been effectively absent from (the few) publications on the topic (Azam, 1860; Richet, 1875). It occupied essentially the same conceptual space as waking state, standing against the state of hypnotism, although these other terms were more prevalent (e.g. Binet and Féré, 1887). That Jules Janet did not employ normal state in what was, after all, a short communication was unremarkable for the time. Nonetheless, given the analogy he drew between Marceline’s condition and Félida’s double existence, we can speculate about whether this was a deliberate choice. Did he wish to avoid conceptual mixing between the normal state in its purely comparative sense and the normal as healthy standard?

Jules could not, however, entirely avoid engaging with this ongoing question, albeit implicitly, as he worried whether leaving Marceline in her ‘second personality’ (i.e. somnambulism) for long periods would entail any ‘serious disadvantage [inconvénient]’. Concluding that it would not, he nonetheless labelled the state as ‘this abnormal personality’ (J. Janet, 1889: 475). This echoed the use of abnormal or pathological by Dufay and Bouchut, and accorded with widespread understanding of hypnotism as an experimental neurosis. It also constitutes the first instance in this chain of cases of a personality being qualified as ‘abnormal’. But, as in previous cases, Marceline’s ‘abnormal personality’ displayed apparently superior health to her waking state—after all, the fact that she could eat and generally live more fully is why Jules left her in somnambulism for months at a time. Thus, when Pierre wrote up Marceline’s case in 1910, he reopened the terminological debate and advocated for a different understanding of the normal state:

It is absurd to call a state of depression incompatible with life state 1 or natural state; it is implausible that this young woman has always been in such a state from the beginning of her life. (Pierre Janet, 1910: 525; emphasis added) ²⁶

Pierre Janet continued his redefinition of Félida’s two states by inverting the sense of the condition seconde, or what Jules had called ‘somnambulism’. The condition seconde, proposed Pierre, was effectively a return of the (putatively) healthy state of childhood, after those abnormal perturbations seen in the initial observation state (Azam’s normal state). Under this logic, the numbering of the states by Azam had been the wrong way around, such that when Félida’s life had become almost entirely filled by the condition seconde, this had simply constituted a return to ‘the normal state of her childhood’ (Pierre Janet, 1910: 526). As for Marceline, when hypnotized, ‘she regains normal functioning, the sensibility and memory that she had previously; I see no reason to distinguish this state from the natural state of her childhood’ (ibid.: 525–6). This reference to Marceline’s previous psychological qualities (sensibility and memory) inflected Janet’s usage of natural state towards an individualized conception, albeit one that had to be compatible with habitual forms of healthy functioning. Normal and natural, although conceptually very close, were nonetheless held apart in Janet’s text; both marked an anterior state against which to examine physio-psychological perturbations, but normal (and abnormal) were reserved for comparison against the general form of health, ²⁸ while natural denoted characteristics particular to a given subject.

Having firmly reattached the normal to its connotations of health, Pierre Janet was well placed to compare Marceline’s functions to some putative general standard. In turn, that he wrote more often of the normal individual than of the normal state meant his version of the normal became holistic as well as normative. A holistic sense of the normal—as relating to an individual’s whole life—had, of course, already been present in Félida’s case. But while Félida’s normal state was the object of careful investigation in Azam’s observation, Janet gave only sporadic attention to the characteristics of the ‘normal individual’ in his 1910 article, and always with regard to an explicit comparison. He did so in two distinct ways. First, when describing experiments performed to test Marceline’s respiration and (skin) evaporation rates, Janet’s comparison had a quantitative basis. It was a matter of measuring the ratio of oxygen to carbon dioxide exhaled by three members of Charles Richet’s physiology laboratory, taking the average, and reporting it alongside the number obtained for Marceline, with the reader left to count the difference between them (Pierre Janet, 1910: 355–6). If configuring the normal as quantitative average might seem a surprising conceptual and methodological departure in personality psychology, it would be better understood as a borrowing from medical forms of counting (Cryle and Stephens, 2017: Chapter 2), inspired by Janet’s own medical training, and especially by Richet’s physiological expertise. ²⁹

Moreover, counting was not a method that Janet applied to psychological properties. His second form of comparison proceeded by qualitative extrapolation along a posited continuum of functioning. In this way, the observed doubling of mental states in hysterics spoke to the psychological makeup of ‘normal individuals’. ‘When there are large oscillations of mental level in hysterics, the two mental states separate from one another.…They cease attaching together by gradations and memories, as in normal individuals’ (Pierre Janet, 1910: 528). To conceive normal (or healthy) phenomena and their pathological counterparts as differing only in degree, not in kind, was a structuring principle of physiological research, and a framework shared by Janet, and indeed most French scientific psychologists, under the heading of pathological psychology. But if Azam’s observations also presupposed that studying departures from the healthy could elucidate psychological functioning in general, Pierre Janet was able to establish a direct parallel between hysterics and the ‘normal individual’, thanks to another shift in the way he interpreted Marceline’s two states. For he abolished the binary logic held so tenaciously by Azam, and instead construed Marceline’s actual mental state at any one time as extending along a continuum, characterized by gradations in the presence of key symptoms or functions. ‘The two states…were only two extremes: Marceline was rarely entirely in one or entirely in the other, most often she was in intermediate states of depression or ascension’ (ibid.: 511). What mattered for understanding Marceline’s condition thus became her degree of depression (e.g. ibid.: 497) and fatigue (ibid.: 513)—or, alternatively, her level of alertness (ibid.: 494)—and treatment was reduced to ‘removing the symptoms [les accidents] without expressing an opinion on [her] overall state’ (ibid.: 339). In turn, problematic psychological functioning became defined by particular symptoms, not a change in an individual’s entire personality. It is easy to see how an apparently ordinary person might be susceptible to a such a pathology. This projected ‘normal person’ finally occupies, in many ways, the conceptual space that Cryle and Stephens (2017: 290) analyse as constituting a ‘convergence of the average and the healthy, of scientific quantification and medical examination’ in conceptions of the normal.

Conclusion

It is not so much our point of arrival, in Pierre Janet’s theoretical and therapeutic approach to hysteria, that matters here, as my contention that we can trace part of its development to contestation around the normal in French psychological discourses. Leading back from Janet’s ‘normal individual’ is a lexical and conceptual pathway that winds through observations of double personality from a comparative ‘normal state’, itself borrowed from physiology by Eugène Azam in writing his case history of Félida. It adds a rich strand to the genealogy of normality so carefully explored by Cryle and Stephens (2017). Concomitantly, my analysis reaffirms the centrality of phenomena of doubling (dédoublement), and of Félida’s case in particular, as conceptually productive in the early years of French psychological enquiry.

Two interlinked choices structured Azam’s account of Félida, and ramified through subsequent observations and discussions. First and most fundamentally, certainly least controversially, he framed the phenomena manifested by Félida under a binary logic of double personality or double states of consciousness. Each of the two states was characterized in rich detail, in accordance with Azam’s epistemic and textual approach to the observation—a meticulous, iterative study of one unusual woman’s life. His second choice was a seemingly routine matter of borrowing the comparative term normal state from physiology. Normal state bore a scientific stamp that could be taken for granted, unlike its opposing term, where work was needed to mark extraordinary psychological phenomena as properly scientific objects of study. But if the choice of term was straightforward, the place occupied by the normal state in Azam’s writing was conceptually complex, and was quickly critiqued by others in his field.

At issue was the extent to which a broad sense of health should be attached to the normal state, or alternatively whether it should be apprehended as purely relational, individual to a particular subject. When faced with the fact that Félida’s habitual state was no longer the one he had termed normal, and with questions over her evident pathology, Azam prioritized terminological stability. His commitment to an individualized, holistic normal, as one of a binary of states, remained firm. Around this conceptual core, he reconfigured relations between the normal and the ordinary, resituated the location of Félida’s defect of memory, and reiterated the ancillary nature of psychological states that would otherwise gain in status through his shifting explanations. Beyond any role as instigator of scientific psychological language, Azam’s work on Félida is thus most significant in showing the ways a researcher can adapt the concept of the normal state in response to the epistemic exigencies of a highly individualized object of study.

On one hand, his text helped engender a lexical shift, making the normal state readily available for use in relation to other psychological phenomena such as hypnotism. The term similarly found a place in subsequent studies of double personality. But on the other hand, the conceptual associations of Azam’s customized normal state were increasingly subject to slippage, as researchers opted to strengthen links between the normal and general biopsychological health. In observations by Dufay and Bouchut, this conceptual shift was effected by a change in the opposite term in the binary pair of mental states; they paired normal with its lexical inverse abnormal, glossed as pathological. Indeed, Azam’s condition seconde seems to have been particularly unattractive to later writers; having helped legitimize phenomena of double personality, it disappeared almost immediately from psychological debates. What persists in the psychological normal is a certain tension between the general and the particular; Dufay, for one, continued Azam’s practice of delineating his subject’s normal state in terms of her individual physiological and psychological characteristics.

The normal could no longer hold together these two usages in the case of Marceline, studied first by Jules Janet, who dubbed her an ‘artificial Félida’, and later by his brother Pierre. What would have been termed Marceline’s normal state under Azam’s scheme was so clearly ‘incompatible with life’ that Pierre Janet was prompted to flip the order of the state naming. In essence, he moved the unperturbed comparative state (Azam’s normal state) back in time, to a (hypothetical) moment when Marceline had experienced good health. The state in which she was habitually observed then became an abnormal perturbation, while her second personality was construed as a return to the original state of normal function. Janet opted for natural state over normal state to allow for some individual variation within an implied range of health. His normal finally denoted the generalized, but still holistic case: the ‘normal individual’. Perhaps more importantly for the evolution both of Janet’s therapeutic work and of psychology’s objects of study, he also dismantled Azam’s obstinately asserted binary logic. What counted in Janet’s new conception was the degree of some symptom exhibited by a psychological subject at a given moment—her depth of depression, or degree of anaesthesia—not the way symptoms might combine into a discrete complex that one might designate a ‘state’.

To return ultimately to the question of Azam’s language, we find that once again where the normal was involved, it was contestation that proved generative. Azam’s language did not so much constitute a model for later French psychological science as its complex conceptual associations, grounded in the particularities of Félida’s case, provided rich matter for adaptation to the changing preoccupations of psychological science.