Daniel Barbiero
On a Path Not Taken: Roger Caillois’ The Necessity of the Mind

In 1981 there appeared La Nécessité d’esprit, a short book by Roger Caillois (1913-1978). The texts that make up La Nécessité d’esprit, translated into English as The Necessity of the Mind, were written during the brief period that Caillois, having joined as a twenty-year-old philosophy student at the École normale supérieure, was a member of the Surrealist group gathered around André Breton. Caillois adhered for a little over a year, breaking with Breton at the end of December 1934 over what he considered to be Breton’s preference for pursuing mysteries rather than taking a rigorously scientific approach to investigating the problems that interested them both. Accordingly, The Necessity of the Mind represents Caillois’ own attempt to examine two of these problems from a systematic point of view—something that Breton, a notoriously unsystematic thinker, had not done. In effect, The Necessity of the Mind was intended to show how a more rigorously thought out Surrealism could come to terms with some of the questions that concerned the movement at the time the book was written.

Parts of what would become The Necessity of the Mind first appeared as articles in the Surrealist or Surrealist-friendly journals Le Surréalisme au service de la révolution and Minotaure, and the dissident Surrealist journal Documents 34, edited by Georges Bataille. Caillois appears to have finished writing the book in 1935 but chose not to publish it during his lifetime. Despite its having been written by someone barely out of his teens, The Necessity of Mind is remarkable not only for the elegance and assurance of its composition, but for its indicating an alternative path Surrealism could have, but didn’t, take at a crucial moment in its development.

The two major points Caillois addressed in The Necessity of the Mind—the complete subtitle of which is An Analytic Study of the Mechanisms of Overdetermination in Automatic and Lyrical Thinking and of the Development of Affective Themes in the Individual Consciousness--were the suitability of automatic writing as a means for laying bare the mechanisms of the mind, and the need to theorize a way of reconciling the subjective and objective worlds, which Breton thought could be done through the evidence of significance coincidences he called “objective chance.” Indeed, the central problem facing Surrealism in the 1930s was how to reconcile an essentially Romantic fascination with the irrational and prophetic on the one hand, with a pre-existing ideological commitment to materialism on the other. In The Necessity of the Mind Caillois attempted to provide a solution based on a particular view of the mind’s workings and its relationship to language.

Caillois’ fundamental assumption was that there is an essential non-identity of thought and language. Consequently, Caillois took as his point of departure the idea that thought was prior to its expression in language, and that it was by analyzing this primary, pre-linguistic layer of thought, and not by analyzing the secondary phenomenon of language, that one could gain insight into the basic mechanisms of the mind. Thus in place of Surrealism’s experiments with automatic writing, Caillois posited experiments with something he called “automatic thinking”--essentially, letting thoughts range freely in a kind of hypnagogic or waking dream state. Caillois theorized that automatic thinking embodied a mode of thought he described as “lyrical,” which he claimed was undergirded by a complex, cognitive-affective phenomenon he termed the “ideogram.”

Lyric Thinking & the Ideogram

“Lyrical thinking” as Caillois defined it, is thinking through the “concrete, singular, and mobile nature of the realities” peculiar to an individual consciousness (p. 4). These realities in turn are couched in affectively developed representations whose complexity is “far more...than language tends to lead us to believe.” (p. 3). This is because language in its everyday, utilitarian form suppresses these individual realities in such a way that “to make the least word understood, [people] are forced to sacrifice all the particular, concrete nuances of their personal experiences to the fiduciary meaning” that just is the “abstract[], general[], and permanen[t]” meaning of the word as used in common discourse. (p. 4) Lyrical thinking, by contrast, is thinking through and with those particular, concrete nuances. It is, in other words, thinking in light of the affective and other non-semantic associations, peculiar to each individual’s history, temperament, and experiences, that permeate each individual’s thought and color his or her assimilation of language and ways of representing the world and its contents.

Underlying and facilitating lyrical thinking is the “ideogram.” Caillois initially quotes the dictionary definition of ideogram as a “sign[] presenting images of idea [sic] or of things,” but immediately goes beyond it. Unlike the dictionary definition, which pertains to graphic of linguistic signs—hieroglyphics being the commonly cited example--Caillois’ ideograms aren’t graphic elements of a written language but instead are complex mental representations amalgamated of conceptual-perceptual content as well as emotive force. Additionally, ideograms exist within intricately cross-referenced, hierarchical, and unconscious networks such that “several series of intellectual, affective, or motor representations...in theory all link up indefinitely to one another and lead to all the others without exception...[and in addition] these entangled links cease being perceived by a lucid consciousness...” An ideogrammatic concept or object thus “assumes an emblematic value...[and] is able to arouse...a certain number of emotive images” for the person for whom it has this value, and by doing so has the capacity to expand beyond its initial and immediate content to take in and integrate other emotively-charged images (p. 9). It is precisely this emotive and integrative or synthetic capacity that makes the ideogram “lyrical,” and that allows it to constitute lyrical thought as such. In sum, the ideogram is a synthetic, emotionally charged idea, image, or object that carries a symbolic value as manifested through the various associations it carries.

From the above, we might think of the ideogram as a wholly subjective phenomenon, but Caillois posited another, objective, aspect to the ideogram that would allow it to transcend itself as a content of the individual psyche. Although he defined the ideogram as a “mental representation” (p. 10), he also attributed to it a certain independence from conscious intention. What he seems to have had in mind here is that the ideogram exists as a hallucinatory or quasi-hallucinatory reality that projects itself onto the outside world just as it arises spontaneously in the mind, and as such seems to have an autonomous existence of its own. This would seem to be Caillois’ attempt to posit a way for subjective states, in the form of ideograms, to find material expression in objects or events holding a particular emotional resonance. As an example of the way subjective states might objectify themselves, he cites a utilitarian object, which he claims “always exceeds its instrumentality” at least in part by virtue of “an irrational residue” projected onto it through the “unconscious representation” of its inventor or user (p. 6). Presumably a similar claim could be made for a word or infralinguistic image, which we can imagine as picking up, at least within the user’s mind, an unconscious affective association by virtue of its use in or correlation with, say, particularly emotionally charged or otherwise personally significant circumstances. The upshot of this objectification of subjective affect is that

the ideogram, as a mental representation, very often acts...as if it were part of the outside world, and it manifests a complete autonomy with respect to the subjective will. (p. 10)

It would be just through this quasi-objective or transcendental mode of being, grounded in affective associations, that the ideogram could be held to reconcile the subjective and objective worlds.

As a consequence of his conception of the ideogram as a quasi-objective, subjectively projected node of associations, Caillois took a dissenting stance in regard to the Surrealist fascination with the significant coincidences Breton ascribed to objective chance. We can see this in his answer to a Minotaure questionnaire on significant encounters, where he opined that the coincidence of apparently causally independent chains of events is evidence that these events are in fact linked by a “subterranean interdependence” (p. 19). It was Caillois’ “working hypothes[is]” (p 21) that the encounters the Surrealists mistakenly, in Caillois’ opinion, took as miraculous really were simply the “mechanical” results of a “web of lyrical overdeterminations” (p. 20). (There is an interesting analogy here to the hidden variables some physicists hypothesized as responsible for the apparently coordinated actions of distant quantum events.) While Caillois admitted that his thinking on this point was tentative, he felt much more sure of being able to analyze the mechanism of overdetermination, which is to say associative formation, in “the immediate and ideogrammatic world of the affective imagination” (p. 21). Regarding that mechanism, he claimed that

it is not a matter of ideas but of psychic elements whose nature is quite variable: memories, images, feelings, sensations, words, concepts. These elements are associated, let us repeat, contingently or contiguously. It seems to make more sense to me to say that they can be joined together either by the partial identity of their intrinsic qualities (which accounts for their association by resemblance) or by the equally partial superimposition of their extrinsic conditions (which explains their association by contiguity). (p. 110, emphases in the original)

With the concept of the ideogram Caillois was, in effect, attempting no less than to lay the groundwork for a comprehensive view of the world as a series of correspondences forged on the basis of the associative force of affect, a force he understood as transcending the subjective source from which it was projected and hence rendering the ideogram a semi-autonomous locus of meaning. This represents a provocative—even lyrical--way of conceptualizing and accounting for the complex aggregates of thought, memory, and affect through which we experience ourselves in the world and the world in ourselves.

Automatic Thinking

If the ideogram functions as the basic element for the affective association of the internal and external worlds, it would be through what Caillois called “automatic thinking” that one could investigate the ideogram and the linkages it manifests and provokes.

Caillois defined automatic thinking as

the series of spontaneously associated representations or ideas evoked by virtue of the lyrical determinism of the ideograms and independently of any external prompting, either from the theoretical demands of practical acts, thus independently of any final conscious activity.” (p. 26—emphasis in the original).

In other words, thinking arising spontaneously on the basis of the affective and ideational associations and analogical relationships particular to an individual, as carried in that individual’s spontaneously generated stream of consciousness. He further described automatic thinking as resembling dream states to such an extent that “waking thoughts, left to their own necessity...would act exactly like condensed dream images, so that the automatic association of ideas would function according to the same mechanism of overdetermination” that operates in dreamwork (p. 23). “That being the case,’ he wrote,

I came to the conclusion that the analysis of a spontaneous associative chain, triggered by the intervention of a lyrical ideogram and left to run its own course, could well be the best source of information about the determining elements of this ideogram. (p. 23)

In effect, Caillois was suggesting that through analysis of the free association of thoughts, recollections, mental images, and the like, one could determine through which influences and experiences these came to be associated with each other, and could thereby gain insight into the unconscious mechanisms and structures of meaning underlying one’s relationship to oneself and the world.

Caillois advocated the pursuit of automatic thinking in place of the conventional Surrealist practice of automatic writing not only on its own hypothesized merits, but because in his view, the latter “had not lived up to all that we, in our enthusiasm, had believed it promised.” (p. 23) In particular, he noted the fact that automatic writing, whether of the spiritualist or Surrealist variety, tended to reflect the beliefs and other formative influences of the milieu in which it was practiced (p. 24). To be sure, Breton had recently acknowledged as much, and for the same reasons, when he admitted in 1933’s “The Automatic Message” that “the history of automatic writing in surrealism has been one of continuing misfortune.” (AB p. 137) Caillois went further and criticized Surrealism’s focus on automatic writing because it took for granted the “relationship of thought to language” (p. 24-25). While Caillois’ criticism on this point had some validity, it was also true that Breton felt that for all of its potential shortcomings, automatic writing represented a particularly developed form of automatism, and in addition generated a variety of language particularly rich in meaning and suitable for interpretation. To be sure, Caillois was not discounting the role of language in the interpretation of psychic life; rather, he felt that automatic thinking would be the more direct route to the workings of the mind because it “asks much less of language” than automatic writing, which as practiced tended to express itself in a grammar different from that of “directed thinking” (p. 26).

In an example of the experimental investigation of automatic thinking, Caillois reproduced his 27 September 1933 notes on the series of associations arising from his own obsession with the game of chess, which he asserted formed a “particularly dense ideogram” for him (p. 27). His account, which was meant to show how an ideogram could serve as the overarching and organizing theme for the chain of associations attaching to and emanating from it, lists a series of thoughts and images related to a greater or lesser degree to the recurring element of the chess game. Caillois analyzed how and why these thoughts and image came to be associated with each other and further claimed to have demonstrated the “remarkable interdependence” of thoughts and images more generally, and the associative force of emotions attaching to ideograms.

The Limits of Analysis

As Denis Hollier points out in his Afterword, The Necessity of the Mind is of a type with Breton’s Nadja, which was an autobiographical work in which Breton performed a self-analysis using the unusual events surrounding his involvement with the titular character as raw material. Hollier also notes Caillois’ debt to Freud’s self-analyses. And many parts of the book in fact do read almost like a parody of Freud’s self-analytic work, particularly The Interpretation of Dreams, which Caillois explicitly cites. Like Freud’s interpretations of his own dream images as well as of his lapses of memory and other parapraxes, Caillois’ self-analyses can involve elaborate, virtuoso displays of ingenious linkages between thoughts, images, and recollections that surely do say something about how his experiences and affective responses created a network of associations based on a hierarchical scale of personal significance. But as Freud’s work demonstrated, self-analysis carries methodological limitations that tend to qualify its conclusions. Like Freud, Caillois can overreach in extrapolating universal conclusions from his own experiences, as for example when he claims that the praying mantis functions as an ideogram likely to carry and trigger affectively charged associations in many people, a claim that may be as much a reflection of his own entomological interests as of any objectively inherent qualities of the insect. Beyond that, his doing so raises the more general problem of self-analyses’ lack of objective, which is to say separate from the person analyzing him- or herself, criteria by which to confirm or refute their conclusions, or to determine whether or not the analytic process is complete and has reached a definitive finding. Wittgenstein’s remark concerning Freud’s method applies in principle as well to Caillois’: it doesn’t “show how we know where to stop—where is the right solution” (Wittgenstein, p. 42).

In the end, Caillois’ automatic thinking, like Freud’s self-analyses, rests on a form of observation in which the observer and the observed are separated only by the narrow, and inevitably permeable, space of a willed psychological distantiation in the service of reflection. Its results are the results of interpretation rather than of experimentation in a narrowly scientific sense, and are necessarily bound to be deeply colored by subjectivity—because the subject conducting the observation and interpretation is the same as the subject of observation and interpretation. But this isn’t to invalidate Caillois’ efforts. Wittgenstein again seems relevant here, with his distinction between causes and reasons: automatic thinking may not, as Caillois had hoped, reveal the objective determinations or causes behind the associations it uncovers, but may instead suggest something like the reasons for those associations—that is, it may bring to mind experiences and our responses to them, which we would recognize as having a played a role in our having associated certain thoughts, memories, objects, and emotions in the way that we have. What constrains interpretation of this kind and brings it to an end thus aren’t objective, i.e., externally-derived, criteria but rather the sense that the point where we choose to stop makes sense as a stopping point. The criteria for judging the results of automatic thinking may be internal and intuitive, but they are not (necessarily) arbitrary. Automatic thinking may not allow us to discover an objective cause behind our associative formations, but in providing a perspective into those formations it still can allow us to discover something about ourselves. That may be all we have, but it certainly is not nothing.


Roger Caillois, The Necessity of the Mind, tr. Michael Syrotinski (Venice, CA: Lapis Press, 1990).

André Breton, “The Automatic Message,” tr. Guy Ducornet, in André Breton: What Is Surrealism? Selected Writings, ed. Franklin Rosemont (New York: Pathfinder, 1978).

Ludwig Wittgenstein, “Conversations on Freud,” in Wittgenstein: Lectures & Conversations on Aesthetics, Psychology, and Religious Belief, Compiled from Notes Taken by Yorick Smythies, Rush Rhees, and James Taylor (Berkeley: U of California Press, 1972).