The Expression asks how the emotions came to be expressed by their characteristic movements: crying for sadness, smiling for happiness, and so on. It opens with three chapters that respectively introduce the three principles of expression. For each principle I shall give Darwin's name, followed by a more humble alias that I have found helpful for keeping mental hold of the principle, and then a brief exposition with an example from the book.

2.1 Serviceable associated habits (“Old habits die hard”)

On this principle, when humans and non-human animals are in the grip of an emotion, and they act in a way that gratifies or relieves – and, in this sense, “serves” – the emotion, the action and the emotion become linked. Or to put it another way, the gratifying action becomes habitual under that emotion. One such link that Darwin identifies in humans is between a fight-primed posture and an indignant frame of mind. According to Darwin, if someone insults you, you tense up, ready to fight, like the men in the Expression's “indignation” photographs: head erect, mouth closed, eyebrows down in frowning mode, shoulders squared, chest out, arms rigid at the sides, fists clenched (Fig. 1). For Darwin, this link exists because going through with the action – taking a swing at the insulter – has in the past brought relief from the mental discomfort of indignation [6].

2.2 Antithesis (“Reverse habits come for free”)

2.3 Actions due to the constitution of the nervous system (“Extreme emotion makes for excess motion”)

Why did Darwin exclude adaptive considerations, and thus natural selection theory, so comprehensively from his treatment of expression? Several historians of science have had a crack at the problem. One is Richard Burkhardt, Jr., who wrote on it in the mid-1980s [10], and whose answer continues to circulate thanks to various endorsements, especially the one in the anthropologist Paul Ekman's generously annotated 1998 edition of the Expression [11]. For Burkhardt, the key is Sir Charles Bell's creationist view of emotional expression. In his famous study of the “anatomy and philosophy” of expression, Bell had argued not only that emotional expression in humans was discontinuous with emotional expression in animals, but that the human face had muscles specially designed for the purpose of expression [12]. Concerned to undermine Bell, Darwin “appears to have overreacted”, wrote Burkhardt. In the Expression, Darwin made impressive cases for continuity (by showing how the same principles applied equally to animals and humans) and for evolution (by supplying expressions with gradual, natural origins). But he threw the Darwinian baby out with the creationist bathwater in further denying that expressions have any communicative function, “thereby leaving himself ill-disposed to develop an idea that would later be advanced by the ethologists of the 20th century – the idea that certain expressive actions, whatever their primary origin, had been developed over time by natural selection” [13].

Undeniably Bell's creationism was a major target in the Expression. Darwin wrote as much, there and elsewhere. Furthermore, Burkhardt makes a valuable supplementary point in stressing that, by the time of the Expression, Darwin had become unprecedentedly and publicly self-critical about what he saw as a tendency in his earlier work to assume adaptedness, and so to exaggerate the explanatory scope of natural selection. If ever, then, Darwin watchers might have expected a book emphasizing non-adaptive phenomena, explained non-selectionally, it was around the time of the Expression. Even so, and as Thomas Dixon pointed out in his study of the emergence of “emotion” as a psychological category, Burkhardt's overreaction explanation takes us only so far [14]. For one thing, the Descent of Man – in which the self-criticism appears, and from which, recall, the Expression grew – is a thoroughly adaptationist and selectionist book, though no less anti-creationist than the Expression. Granted that Darwin was more prepared than ever to discuss non-adaptive evolution, we still need to know why he judged emotional expression in particular to fall among the non-adaptive phenomena – and this despite expression being, from the perspective of later Darwinians (Ekman's frustrated commentaries in his edition are instructive here), eminently suited to adaptive, selectional analysis. To put the point another way: given that creationist opponents could be found all over the scientific shop, we need to know what it was about emotional expression, as distinct from the many other human topics that Darwin handled at this time, which prompted Darwin to “overreact”.

A second explanation, of the same vintage as Burkhardt's, is due to William Montgomery [15]. He stresses the evidence of Darwin's post-Beagle notebooks of the late 1830s, in particular the M and N notebooks, dedicated to expression and other “metaphysical” topics (that is, to do with the theory of mind). In the M notebook, filled between July and early October 1838, Darwin stated what can now be identified as the first two of his three general principles of expression [16]. Indeed, it is clear from the notebooks, as it is not from the Expression, that Darwin's associationist-evolutionary approach to expression was taken over from the Zoonomia (1794–6) of his grandfather, Erasmus [17]. For Montgomery, just as significant as the intellectual character of these notebook entries is their timing. In his view, they show that Darwin's basic approach to expression was hammered out before he developed, from late 1838, his theory of natural selection. According to Montgomery, the non-Darwinian character of the Expression is thus explained by the fact that its explanatory principles predate natural selection theory and, apparently, never struck Darwin as in need of replacement in the years that followed. “These principles survived intact”, wrote Montgomery, “because they served a useful purpose in his case against Bell and because he (Darwin) faced no serious counterarguments” [18].

Again, we have here an insightful account which nevertheless falls short of solving the puzzle. Having Zoonomia in view as a major private source for Darwin on expression helps hugely in making sense of the later, public theorizing as a product of the same mind, indeed the same creative period, that produced natural selection theory. With the Erasmian roots of the Expression thus exposed, we can even, in a kind of reflexive homage, go beyond Montgomery and see the book as the result of associations formed during Darwin's medical student days at Edinburgh, when he became an admirer of Zoonomia (under the influence of his Lamarckian mentor Robert Grant [19]), and when, at the 17-year-old's first meeting as a member of the student natural history society in late 1826, he watched his sponsor deliver a paper aiming to refute Bell on expression [20]. What needs supplementing, however, is Montgomery's claim that Darwin's expression principles survived from the 1830s to the 1870s “because he faced no serious counterarguments”. In support, Montgomery notes that Darwin in the late 1850s abandoned a Lamarckian explanation of instinct in social insects for a natural-selectional explanation when he realized that the latter dealt satisfactorily with an outstanding difficulty (the origin of sterile workers). So, since Darwin never abandoned his Lamarckian explanation of expression, he must, Montgomery argues, never have encountered any similar, selection-resolvable difficulty [18].

But the comparison with the insect case is inapt. There Darwin concurred with the opinion of natural-theological writers on the adaptive value of the phenomenon, and merely switched allegiance from one design-undermining explanation (based on habit inheritance) to a different one (based on chance variation and natural selection among groups) [21]. With expression, however, Darwin dissented not just from the design explanation, but from the description of the phenomenon as adaptive. What needs explaining is his dissent on the description. One might, drawing on Montgomery, reply on his behalf that Darwin simply took over his sense of the nonadaptive origins of expression from his grandfather's work, alongside the associationist explanatory style. Up to a point, that seems right; but again, it only puts back a crucial and unanswered question, namely, why Erasmus should have been treated as authoritative on this topic but not on so many others. And in any case, it was not generally true of Darwin that he gave up on theories only when they had run into difficulties. It seems, for instance, that in 1838 he dropped his earlier theorizing on species origins for natural selection theory without there having been some outstanding problem for which natural selection provided an answer [22].

A more subtle problem with Montgomery's account, and Burkhardt's too, is that they do not really capture Darwin's positive attitude in the Expression to his own explanations. The book does not read as if Darwin is using non-selectional resources inadvertently or faute de mieux, as if he had never gotten around to seeing that natural selection theory might, in fact, be the better option. Something about Darwin's expression principles made them, in Darwin's circa 1870 perspective, the right tools for the job, in preference to natural selection theory. But what? In a 2002 article, I suggested that Darwin's deep loyalty to the vera causa methodological ideal had a bearing here [23]. To conform one's theorizing to this ideal was to show, well before giving a causal explanation, that the specified cause was real (not merely hypothetical) and adequate to producing effects of the right magnitude. Hence the premium, for the vera causa theorist, on explanations that drew upon causes whose action had been observed or, if not observed, inferred by analogy from the observed. Simply in the way Darwin structured the Expression, with his explanatory principles laid out first and argued for in a general, expression-independent way, the book is a vera causa book. But it is also so, I urged, in placing habit formation at the explanatory center. Circa 1870, it was a matter of common observation that habits form and die hard; it was equally a matter of common observation (or so the likes of Darwin believed) that these habits could be inherited. By contrast, the observational credentials for what natural selection theory required – spontaneous changes in brains, such that certain emotions became linked to certain movements – were feeble.

By way of backing this view, I cited a passage from the book's conclusion. “Some expressive movements”, wrote Darwin, “may have arisen spontaneously, in association with certain states of the mind ... and afterwards been inherited. But I know of no evidence rendering this view probable” [24]. Here Darwin explicitly considered and rejected, as evidentially wanting, the possibility that inborn, inheritable variation – the sine qua non of natural selection – had played a role in bringing into being at least some of the expressions for which he had supplied habitual origins. So taking seriously Darwin's career-long respect for well-evidenced argumentation can help us take seriously his own regard for his expression principles and their non-selectional character. But it cannot, in the end, do any better in explaining the Expression's explanatory project than Burkhardt's or Montgomery's analyses did. After all, Darwin did not in general think that selectional explanations of animal behavior failed to meet the vera causa standard (consider the insect-instinct case). His judging that selection theory nevertheless fell short for expressive behavior thus requires further explanation on the historian's part. What lay behind Darwin's judgment? Why not switch to adaptive descriptions and selectional explanations once natural selection theory – Darwin's explanatory pride and joy – was around?

New light on these questions has come in this Darwin anniversary year with the publication of Adrian Desmond and James Moore's Darwin's Sacred Cause: Race, Slavery and the Quest for Human Origins [25]. Opposition to black slavery was a tradition in Darwin's family, going back to Erasmus's day. (Darwin's other grandfather, the pottery magnate Josiah Wedgwood, manufactured the famous “Am I not a Man and a Brother?” antislavery cameo.) So much is long familiar. But no one previously has brought out the depth of Darwin's immersion in the crusading world of British antislavery or – the truly far-reaching point – the intellectual consequences of that immersion. Confronting slavery in the first half of the 19th century meant confronting the scientific arguments developed to defend it, including the view that blacks could be enslaved because they belong to a different, inferior species, derived from an aboriginally distinct stock. Against such pluralism about the human races/species, antislavery campaigners upheld the “unity of man”, arguing for the shared ancestry of the races, and their status as races within a single species. For Darwin, as Desmond and Moore show, there was never any serious doubt about which side of this debate he was on. Darwin was a unity-of-man man from the beginning. Nor did his commitment ebb after slavery was finally outlawed in the British Empire in the 1830s. Slavery in the United States became the focus, and kept antislavery passions inflamed well into the 1860s.

Desmond and Moore reconstruct the manifold impact of Darwin's antislavery attitude on his species theorizing, from his days on the Beagle voyage through to his writing of the Descent [26]. Given this time frame, the Expression falls outside their purview. Yet it looks an obvious fit for their theme, since the documenting of the unity of the human races is, for Darwin, one of the major results of the book. Mindful as ever of vera causa etiquette, he is careful to represent such documentation as a result – an incidental but welcome spin-off of an inquiry otherwise directed. He explains in his introduction that his interest in human races and emotional expression is in the first instance an interest in sorting “innate or instinctive” expressions from those which, like languages, are acquired conventionally, on the view that if lots of different races express the same emotion in the same way, the expression is probably innate. To this end, he had sent a questionnaire on expression to correspondents around the world, 36 of whom had replied. “It follows, from the information thus acquired”, wrote Darwin, “that the same state of mind is expressed throughout the world with remarkable uniformity; and this fact is in itself interesting, as evidence of the close similarity in bodily structure and mental disposition of all the races of mankind” [27].

Darwin's efforts in the Expression on behalf of the universality of human emotional expression are well remembered, not least thanks to Paul Ekman's experimental updating of Darwin's investigation. Less well remembered is that, in Darwin's view, universality was not enough to establish human racial unity. The trouble was that an emotion-expression package could come to be universal in one of two ways, with different consequences for human unity. Either the package was a common inheritance from a shared ancestor. Or it was independently derived in different human groups via natural selection, acting to preserve beneficial chance variation as it arose independently in these groups. Under the common-inheritance interpretation, humans are unified. Under the natural selection interpretation, humans are not unified. So how to choose?

For Darwin, the choice turns on the closeness of the similarity in emotional expression in the different races. And on the evidence collected, as he interprets it, the similarity is indeed close, to the point of constituting identity, sameness. Darwin put the point as follows in his conclusion:

“I have endeavoured to show in considerable detail that all the chief expressions exhibited by man are the same throughout the world. This fact is interesting, as it affords a new argument in favour of the several races being descended from a single parent-stock, which must have been almost completely human in structure, and to a large extent in mind, before the period at which the races diverged from each other. No doubt similar structures, adapted for the same purpose, have often been independently acquired through variation and Natural Selection by distinct species; but this view will not explain close similarity between distinct species in a multitude of unimportant details. Now if we bear in mind the numerous points of structure having no relation to expression, in which all the races of man closely agree, and then add to them the numerous points, some of the highest importance and many of the most trifling value, on which the movements of expression directly or indirectly depend, it seems to me improbable in the highest degree that so much similarity, or rather identity of structure, could have been acquired by independent means. Yet this must have been the case if the races of man are descended from several aboriginally distinct species. It is far more probable that the many points of close similarity in the various races are due to inheritance from a single parent-form, which had already assumed a human character.”[28].

Darwin here flags the novelty of this argument for human racial unity as well as the importance for the argument of emotional expression turning out to be beyond the reach of natural selection. Yet this passage, along with the racial element in the Expression generally, have been overlooked by commentators curious about the non-selectional character of the book. The temptation has been to treat Darwin's unity argument more or less as Darwin presented it – as a nice surprise, noticed at the end of the inquiry, and of no more interest than lots of other points Darwin raises in his conclusion. Desmond and Moore's historical tour de force makes plain that whether humans are one species or many was not, for Darwin, just another question. To have developed a new argument for human racial unity would have mattered a great deal to him. It was not something that would have caught him unawares. And although we do not know when he first understood that emotional expression could serve such an argument, or that natural selection could undermine it, a good bet is that he understood these things long before he wrote the Expression, and that his resolve to stick with non-functional descriptions and non-selectional explanations firmed up accordingly.

Above I mentioned Thomas Dixon, who wrote a pioneering analysis of the Expression historiography. There he formulated two desiderata for a good explanation of the Expression. First of all, it needs to make intelligible the division of labor between the Descent and the Expression. Second, it needs to explain why, for all Darwin's indifference and even hostility to natural selection theory in the Expression, he nevertheless grants it a role in the explanation of some emotional expressions [29]. On both counts, the human-unity explanation does rather well. To take first the question of why Darwin invokes natural selection theory when he does, consider Darwin's calling upon it to explain the evolution of the rattlesnake rattle but not the evolution of the human indignation posture. As noted previously, that posture has an obvious functionality, and would thus seem susceptible of elaboration via natural selection. Yet Darwin does not go there. Generalizing, one can say, on the new explanation, that Darwin was prepared to acknowledge a contribution from natural selection only for emotional expressions where human unity was not at stake – as it was not at stake with the rattlesnake rattle, or the various anatomical means by which frightened animals make themselves appear larger [30]. A good test of the correctness of this view is whether Darwin in the Expression deploys natural selection theory occasionally in explaining animal expressions but never in explaining human ones. And such is indeed the pattern, as William Montgomery pointed out – though he offered little insight into why the pattern holds [31].

On the Expression's relationship to the Descent, the human-unity explanation promises to revise substantially our understanding of how these books hang together. For of course, and as Desmond and Moore emphasize, Darwin in the Descent was very much concerned to defend the unity of the human races. Moreover, Darwin's Descent case for unity was just as dependent as in the Expression on some traits – external differences between the human races (skin color and so on) – being characterized as non-functional in the struggle for survival and therefore outside the domain of natural selection [32]. And again as with the Expression, Darwin in accounting for these non-functional traits introduced a distinctive set of explanatory principles, the “principles of sexual selection” (to quote Darwin's chapter title [33]), which he propounded in a general way before making the case for their real action, first among non-human animals, then finally among humans. With such commonalities in view, we can now see Darwin's human evolution project as extending over three separate books, each centered on a different set of questions and a different set of explanatory principles: one on how ancestral humans originated from non-human animals via natural selection (The Descent of Man); one on how those ancestral humans diverged into the separate but still human races via sexual selection (Selection in Relation to Sex); and a third on how, throughout all that divergence, a common legacy of emotional expressions remained constant, after initial fixing via habit-formation and -inheritance (The Expression of the Emotions in Man and Animals).

It would have made sense, as Darwin intended, for these three inquiries to be published as parts of a single book [34]. Contingencies of production (what he wrote first) and publication (what he reckoned his publisher was willing to publish in one go) no doubt contributed to the appearance of these inquiries as we have them, in a form that invites readers to see an illusory homogeneity between the first two, and an illusory heterogeneity between those two and the third. By restoring Darwin's abiding concern with human unity to our picture of him, we thus recapture Darwin's own sense of the intellectual unity across the apparently disparate parts of his argument on human evolution.

The solution proposed here to the puzzle of Darwin's Expression – that behind the steadiness of his belief in the non-functional character of emotional expression lies his commitment to human racial unity – does not, if correct, render previous solutions defunct. As we have seen, each illuminates different aspects of Darwin's text and its context. But in the Origin's 150th year, it is especially appropriate to contemplate a solution that binds that book's one, spare illustration of the tree of life with the Expression's many, extravagant ones on behalf of the same. Indeed, if in the Expression Darwin felt that he could bolster the case for common ancestry only by limiting the scope of natural selection theory, then we have a new reason to take up the recent attempt to unpick what Darwin represented in the Origin as “one long argument” [35]. An anniversary that inspires rereadings along such critical lines will have been well spent.